EPIDEMIOLOGICAL STUDY OF NEISSERIA GONORRHEAE CO-INFECTION AMONG
MALES WITH HIV/AIDS RESIDING IN THE STATE OF MICHIGAN
By
Matthew Joseph Francis

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Epidemiology – Doctor of Philosophy
2014

ABSTRACT
EPIDEMIOLOGICAL STUDY OF NEISSERIA GONORRHEAE CO-INFECTION AMONG
MALES WITH HIV/AIDS RESIDING IN THE STATE OF MICHIGAN
By
Matthew Joseph Francis
Background: Epidemiological evidences suggest a synergistic relationship between Human
Immunodeficiency Virus (HIV) and N. gonorrheae in co-infected individuals. The objectives of
this study were; 1) to test whether age at HIV between those with a history of N. gonorrheae
infection was younger as compared to those with no history of reported N. gonorrheae infection,
2) to test whether age at HIV progression event (death or Acquired Immune Deficiency
Syndrome (AIDS)) was younger as well as time from HIV diagnosis to HIV progression event
was reduced in individuals with a history of N. gonorrheae infection as compared to those
without, 3) to investigate the geospatial distribution of HIV and N. gonorrheae in the State of
Michigan, and 4) to investigate the differences in biological markers (CD4 cell count and viral
load) in a sub-population of HIV N. gonorrheae co-infected individuals.
Methods: A retrospective cohort study from 2005-2011 was conducted in collaboration with
Michigan Department of Community Health on Michigan HIV positive males with a history of
N. gonorrheae infection between 2005-2011 as well as HIV positive males with no history of N.
gonorrheae infection who were alive at the start of 2005. A sub-group analysis of Michigan HIV
positive males with a reported N. gonorrheae infection between 2011 and 2013 to investigate
biological markers of HIV progression. Spatial, survival, and linear regression analyses were
performed on the data to evaluate the objectives of the study.

Results: Mean age at HIV diagnosis was found to be younger in individuals with pre-HIV N.
gonorrheae infections. On average males with 2 or more pre-HIV N. gonorrheae infections were
10.2 years younger than individuals with no N. gonorrheae infections, and males with only one
pre-HIV N. gonorrheae infection were 7.3 years younger at HIV diagnosis. Hazard of AIDS
increased in males with post-HIV N. gonorrheae infections, with individuals with multiple postHIV N. gonorrheae infections having increase in hazard of progression to AIDS. A longer HIV
to AIDS lag time period was observed in males with multiple post-HIV gonorrhea. Geospatial
analysis showed that the area around the City of Detroit and the Tri-County area of Oakland,
Macomb, and Wayne have the highest burden of disease due to co-infection as well as represent
the most likely clusters for HIV N. gonorrheae co-infections. CD4 cell count was found to
increase in the year during N. gonorrheae infection as well as in the years after as compared to
pre-gonorrhea levels. Similarly viral load was found to increase in the year during N. gonorrheae
infection, however, it did not remain elevated as compared to the pre-gonorrhea states.
Conclusion: This study is a first step in documenting the possible synergistic relationship
between N. gonorrheae infection and HIV. The study found that age at HIV and age at HIV
progression event were reduced (younger at event) than in those with no history of gonorrhea.
The study also identified areas of concern using geospatial analysis where resources should be
focused to try and reduce and disease. Paradoxically the time from HIV to HIV progression
event was increased in those with N. gonorrheae infections as compared to those without. Finally
the study found that CD4 cell count was increased in the time period during and after N.
gonorrheae infection, and that viral load was increased in the year during N. gonorrheae
infection, but returned to pre-gonorrhea levels in the years following.

To
My wife Jessica, without whom none of this would have been possible
AMDG

iv

ACKNOWLEDGEMENTS

First and foremost I would like to acknowledge my mentor Dr. Mahdi Saeed who has guided me
through my graduate program and has helped me at every step during the way including funding
this project. Dr. Saeed has never stopped pushing me to go above and beyond and has always
offered me wisdom and guidance in my academic, personal, and professional life. He truly
represents all that a mentor should encompass and I have been honored and blessed to have him
as mine.
I would also like to acknowledge my dissertation committee members, Dr. Sue Grady, Dr. Julie
Wirth, Dr. David Todem, and Dr. Mahdi Saeed all of whom have been instrumental in shaping
this manuscript. My committee has given me guidance, reigned me in when I have wondered
into the weeds, and have provided feedback at every step of the way. My committee has also
heard me out, and allowed me to make mistakes so that I would be able to learn from them and
master not only what it is to write a dissertation, but what it is to be a PhD in the field. Despite
grants, time zones, multiple offices and emails they have taken the time to guide me in this
process and without them I would have been lost.
I am grateful for the assistance and guidance of Kathryn Macomber and the staff of
Communicable Disease Division of Michigan Department of Community Health. Kathryn
always made time for me to pop in and talk about the data, my progress, and what I was gaining
out of writing this dissertation. She also reflected someone that I wished to emulate, a true master
of her field and a driving force behind STD reduction.
I would also like to acknowledge all of the sacrifices my family has made to allow me to follow
my dream. My wife Jessica and children, Addison Rae and Greyson, have given up time,
v

vacations and sleep to allow me to finish out my PhD and obtain this doctorate. Without their
help, support, understanding and love I would not have been able to accomplish any of this. I
would also like to thank my parents who have helped to finance my education and have been
there supporting me every step of the way. I would also like to acknowledge my mentors in my
CSTE HSIP fellowship, Dr. Pat Luedtke and Dr. Patrice Korjenek both whom have been flexible
and accommodating in allowing me to work on my dissertation while on the fellowship.
I would also like to thank The Graduate School, the College of Human Medicine, and the
Department of Epidemiology and Biostatistics for providing financial support in the form of a
dissertation completion fellowship. I would also like to acknowledge Dr. Linda Dykema and the
Michigan Department of Community Health Division of Environmental Health, who along with
Dr. Wirth gave me an opportunity and a Graduate Assistantship to apply what I had learned to
public health issues.
Finally, I would like to acknowledge the Department of Epidemiology and Biostatistics who
helped to fund this study, and for the guidance of the department chair Dr. Claudia Holzman who
helped to keep me focused as I started a fellowship, juggled a family and wrote my dissertation.

vi

TABLE OF CONTENTS

LIST OF TABLES……………………………………………………………………………….x
LIST OF FIGURES…………………………………………………………………………….xii
KEY TO ABBREVIATIONS…………………………………………………………………xiv
Chapter 1. Background and Objectives………………………………………………………...1
1.1 Human-Immunodeficiency Virus/Acquired Immune Deficiency Syndrome ………...1
Natural History…………………………………………………………………….1
Epidemiology……………………………………………………………………...5
US HIV……………………………………………………………………………6
Michigan HIV……………………………………………………………………..8
Acquired Immune Deficiency Syndrome..………………………………………10
Antiretroviral Therapy…………………………………………………………...15
1.2 Neisseria gonorrheae………………………………………………………………..18
Natural History…………………………………………………………………..18
Epidemiology……………………………………………………………………20
1.3 Co-infections…………………………………………………………………………22
HIV/Non-Gonorrheal STDs……………………………………………………...23
HIV/N. gonorrheae………………………………………………………………25
1.4 Biological Plausibility………………………………………………………………..31
Epidemiological Synergy………………………………………………………...32
Biological Markers of HIV-Progression…………………………………………34
1.5 Rationale……………………………………………………………………………..38
1.6 Objectives……………………………………………………………………………40
APPENDIX………………………………………………………………………………46
Chapter 2. Interaction of N. gonorrheae and HIV infections in Michigan Males ..…………..48
2.1 Abstract………………………………………………………………………………48
Introduction………………………………………………………………………48
Methods…………………………………………………………………………..48
Results……………………………………………………………………………48
Conclusion……………………………………………………………………….49
2.2 Introduction…………………………………………………………………………..51
2.3 Methods………………………………………………………………………………54
2.4 Results………………………………………………………………………………..58
History of gonorrhea……………………………………………………………..58
History of multiple gonorrhea episodes………………………………………….62
Pre-HIV gonorrhea……………………………………………………………….64
2.5 Conclusion…………………………………………………………………………...66
APPENDIX………………………………………………………………………………71

vii

Chapter 3. Enhancement of HIV Progression by N. gonorrheae Co-Infection in Michigan
Males ……………………………………..……………………………………………………….86
3.1 Abstract………………………………………………………………………………86
Introduction………………………………………………………………………86
Methods…………………………………………………………………………..86
Results……………………………………………………………………………86
Conclusion……………………………………………………………………….87
3.2 Introduction…………………………………………………………………………..89
3.3 Methods………………………………………………………………………………92
3.4 Results………………………………………………………………………………..95
Time to Death……………………………………………………………………98
Time to AIDS Positive Status……………………………………………………99
Time to AIDS Positive Status and Multiple N. gonorrheae Infections………..100
Post-HIV N. gonorrheae Time to AIDS………………………………………..102
Time to AIDS Positive Status and Multiple Post-HIV N. gonorrheae Infections..104
3.5 Conclusion………………………………………………………………………….106
Time to Death…………………………………………………………………..106
Time to AIDS Positive Status…………………………………………………..107
Time to AIDS Positive Status and Multiple N. gonorrheae Infections……….108
Post-HIV N. gonorrheae Time to AIDS………………………………………..110
Time to AIDS Positive Status and Multiple Post-HIV N. gonorrheae Infections…110
Limitations…………………………………………………………………………111
Conclusion…………………………………………………………………………113
APPENDIX…………………………………………………………………………..…115
Chapter 4. Geospatial Analysis of HIV associated N. gonorrheae Infections in Michigan Males
2005-2011 ……………………………………………….……………………………………...140
4.1 Abstract……………………………………………………………………………..140
Introduction……………………………………………………………………..140
Methods…………………………………………………………………………140
Results…………………………………………………………………………..140
Conclusion……………………………………………………………………...141
4.2 Introduction…………………………………………………………………………142
4.3 Methods……………………………………………………………………………..146
4.4 Results………………………………………………………………………………149
4.5 Conclusion………………………………………………………………………….154
Limitations……………………………………………………………………...157
APPENDIX……………………………………………………………………………..159
Chapter 5. Biological Markers and Geospatial Analysis of Michigan Males with HIV and N.
gonorrheae Co-Infections from 2011 to 2013 ….……………………………………………....172
5.1 Abstract……………………………………………………………………………..172
Introduction……………………………………………………………………..172
Methods…………………………………………………………………………172
Results…………………………………………………………………………..172
Conclusion……………………………………………………………………...173
5.2 Introduction…………………………………………………………………………174
viii

5.3 Methods……………………………………………………………………………..179
5.4 Results……………………………………………………………………………....184
CD4 Cell Count………………………………………………………………...185
Viral Load………………………………………………………………………186
Zip Code Cluster Analysis……………………………………………………...187
5.5 Conclusion………………………………………………………………………….188
CD4 Cell Count………………………………………………………………...188
Viral Load………………………………………………………………………190
Zip Code Cluster Analysis……………………………………………………...191
Limitations……………………………………………………………………...192
APPENDIX……………………………………………………………………………..194
Chapter 6. Conclusion………………………………………………………………………...203
6.1 Age at HIV Positive Diagnosis……………………………………………………..203
6.2 Time to HIV Progression Event…………………………………………………….205
6.3 County Level Geospatial Analysis………………………………………………….207
6.4 Zip Code Level Geospatial Analysis……………………………………………….210
6.5 Biological Markers……………………..…………………………………………...211
6.6 Conclusions…………………………………………………………………………213
6.7 Limitations………………………………………………………………………….214
REFERENCES………………………………………………………………………………..216

ix

LIST OF TABLES

Table 1.1: Percent Urethral and Rectal Gonorrhea in MSM, AIDS Cohort Study (55)…………47
Table 1.2: HIV Seroconversion Odd Ratios in MSM, AIDS Cohort Study (55)………………..47
Table 2.1: Demographic Distribution of 19,647 HIV Positive MI Males (2005-2011)…………72
Table 2.2: MDCH Non-Condensed Risk Group Distribution of HIV Positive MI Males (20052011)……………………………………………………………………………………………..73
Table 2.3: General Linear Model Adjusted and Unadjusted Means for Age at HIV Diagnosis in
Males With and Without Any History of Gonorrhea………………………………………….....74
Table 2.4: Median Age at HIV Diagnosis and Unadjusted Hazard of HIV in HIV Positive Males
in Michigan 2005-2011…………………………………………………..………………………76
Table 2.5: Multivariate Cox Regression Analysis of Hazard of HIV Diagnosis in MI Males with
a History of N. gonorrheae 2005-2011…………………………………………………………..77
Table 2.6: General Linear Model Adjusted and Unadjusted Means for Age at HIV Diagnosis in
Michigan Males with Multiple N. gonorrheae Infections……………………………………….79
Table 2.7: Multivariate Cox Regression Analysis of Hazard of HIV Diagnosis in MI Males with
a History Multiple of N. gonorrheae 2005-2011………………………………………………...80
Table 2.8: General Linear Model Adjusted and Unadjusted Means of Age at HIV Diagnosis in
Michigan Males with Pre-HIV N. gonorrheae Infections……………………………………….81
Table 2.9: Multivariate Cox Regression Analysis of Hazard of HIV Diagnosis in MI Males with
a History of Pre-HIV N. gonorrheae 2005-2011………………………………………………...82
Table 3.1: Demographic Distribution of 19,647 HIV Positive MI Males (2005-2011)………..116
Table 3.2: Distribution of HIV Positive MI Males with and without AIDS in 2005-2011…….117
Table 3.3: Risk Group Distribution of HIV Positive and AIDS Positive MI Males (20052011)……………………………………………………………………………………………118
Table 3.4: Median Age and Time from HIV Diagnosis to HIV Progression Events…………..119
Table 3.5: Unadjusted Analysis of Hazard of Death in MI Males from 2005-2011……...……120

x

Table 3.6: Multivariate Analysis of Hazard of Time from HIV to Death in MI Males with and
without a History of N. gonorrheae 2005-2011…..…………………………………………….121
Table 3.7: Unadjusted Analysis of Hazard of AIDS Positive Status in MI Males with and without
a History of N. gonorrheae 2005-2011…………………………………………………………122
Table 3.8: Multivariate Analysis of Hazard of Time of HIV Diagnosis to AIDS Positive Status in
MI Males with and without a History of N. gonorrheae 2005-2011…….……………………..123
Table 3.9: Unadjusted Analysis of Hazard of AIDS Positive Diagnosis in MI Males with and
without Multiple N. gonorrheae Infections 2005-2011…..…………………………………….124
Table 3.10: Multivariate Analysis of Hazard of Time from HIV Diagnosis to AIDS in MI Males
with and without a History of Multiple N. gonorrheae Infections 2005-2011…………………125
Table 3.11: Unadjusted Analysis of Hazard of AIDS in MI Males with and without Histories of
Post-HIV N. gonorrheae Infections 2005-2011………………………...………………………126
Table 3.12: Multivariate Analysis of Hazard of Time from HIV Diagnosis to AIDS in MI Males
with and without a History of Post-HIV N. gonorrheae Infection 2005-2011………………....127
Table 3.13: Unadjusted Analysis of Hazard of Age at AIDS in MI Males with and without a
History of Multiple N. gonorrheae Infections 2005-2011……………………………………...128
Table 3.14: Multivariate Analysis of Hazard of Time from HIV Diagnosis to AIDS in MI Males
with and without Multiple N. gonorrheae Infections 2005-2011……………………………....129
Table 4.1: Demographic Distribution of 19,647 HIV Positive MI Males (2005-2011)………..160
Table 5.1: Demographic Distribution of 743 HIV/Gonorrhea Co-Infected Michigan Males from
2011-2013………………………………………………………………………………………195
Table 5.2: Multivariate General Estimating Equation (GEE) with Repeated Subjects Model of
CD4 Cell Count in HIV Positive Males with N. gonorrheae..…………………………………196
Table 5.3: Multivariate General Estimating Equation (GEE) with Repeated Subjects Model of
Viral Load (copies/mL) in HIV Positive Males with N. gonorrheae.………………………….197
Table 5.4: Poisson Analysis of Gonorrhea/HIV Co-Infected Clusters in Michigan Males, 20112013…………………………………………………………………………………………….198

xi

LIST OF FIGURES

Figure 2.1: Kaplan-Meier Curves for Age at HIV Diagnosis in HIV Positive MI Males with and
without History of N. gonorrheae (2005-2011)………………………………………………….83
Figure 2.2: Kaplan-Meier Curves for Age at HIV Diagnosis for Multiple N. gonorrheae
Infections from 2005-2011 in MI HIV Positive Males…………………………………………..84
Figure 2.3: Kaplan-Meier Curves for Age at HIV Diagnosis in MI Males with Pre-HIV N.
gonorrheae Infections 2005-2011………………………………………………………………..85
Figure 3.1: Kaplan-Meier Curves for Survival Rate in MI Males with and without N. gonorrheae
Infection 2005-2011…………………………………………………………………………….130
Figure 3.2: Kaplan-Meier Curves for Survival from Time of HIV Diagnosis in MI Males with
and without a History of N. gonorrheae 2005-2011……………………………………………131
Figure 3.3: Kaplan-Meier Curves for Age of AIDS Positive Status in MI Males with and without
N. gonorrheae 2005-2011………………………………………………………………………132
Figure 3.4: Kaplan-Meier Curves for Time from HIV Diagnosis to AIDS Positive Status in MI
Males 2005-2011……………………………………………………………………………….133
Figure 3.5: Kaplan-Meier Curves for Age of AIDS Positive Diagnosis in MI Males with Multiple
N. gonorrheae Infections 2005-2011…………………………………………………………...134
Figure 3.6: Kaplan-Meier Curve for Time from HIV Diagnosis to AIDS Positive Status in MI
Males with a History of Multiple N. gonorrheae Infections 2005-2011……………………….135
Figure 3.7: Kaplan-Meier Curves for Age at AIDS Positive Diagnosis in MI Males with
Histories of Post-HIV N. gonorrheae Infections 2005-2011…………………………………...136
Figure 3.8: Kaplan-Meier Curves for Time from HIV Diagnosis to AIDS Positive Status in MI
Males with a History of Post-HIV N. gonorrheae Infection 2005-2011……………………….137
Figure 3.9: Kaplan-Meier Curves for Age at AIDS Positive Diagnosis in MI Males with a
History of Multiple N. gonorrheae Infections 2005-2011……………………………………...138
Figure 3.10: Kaplan-Meier Curves for Time from HIV Diagnosis to AIDS Positive Status in MI
Males with Multiple N. gonorrheae Infections 2005-2011…………………………………….139
Figure 4.1: Map of Number of HIV Positive Males per 100,000 in the State of Michigan, 20052011…………………………………………………………………………………………......161

xii

Figure 4.2: RSE (Root Mean Square Error) of Number HIV Positive Males per 100,000 in the
State of Michigan, 2005-2011…………………………………………………………………162
Figure 4.3: Map of the Number N. gonorrheae cases in HIV Positive Males from 2005-2011 in
the State of Michigan……………...…………………………..………………………………..163
Figure 4.4: Map of the Rate of N. gonorrheae per 1,000 in HIV Positive Males in the State of
Michigan 2005-2011 Adjusted Population Density.……………………………………………164
Figure 4.5: Root Mean Square Error of N. gonorrheae per 1,000 in HIV Infected Males in the
State of Michigan 2005-2011 Adjus Population Density………………………………………165
Figure 4.6: Percent of Gonorrhea Infected HIV Positive Males with Multiple N. gonorrheae
Infections in the State of Michigan, 2005-2011………………………………………………...166
Figure 4.7: Map of Percent of HIV Infected Males with Multiple N. gonorrheae Infections in the
State of Michigan, 2005-2011…………………………………………………………………..167
Figure 4.8: Getis-Ord GiZ score Cluster Analysis of N. gonorrheae Infected HIV Positive Males
in the State of Michigan, 2005-2011……………………………………………………………168
Figure 4.9: Poisson High Only Rate Cluster Detection Method of HIV Positive Males in the State
of Michigan, 2005-2011………………………………………………………………………...169
Figure 4.10: Poisson High Only Rate Cluster Detection Method of N. gonorrheae Infected HIV
Positive Males in the State of Michigan, 2005-2011…………………………………………...170
Figure 4.11: Bernouilli High Only Rate Cluster Analysis of N. gonorrheae Infected HIV Positive
Males in the State of Michigan, 2005-2011……………………………………………………171
Figure 5.1: Getis-Ord GiZ Score Cluster Analysis of N. gonorrheae/HIV Co-Infected Michigan
Males 2011-2013……………………………………………………………………………….199
Figure 5.2: Getis-Ord GiZ Score Cluster Analysis of N. gonorrheae/HIV Co-Infected Michigan
Males 2011-2013 Detroit and Tri-County Area………………………………………………...200
Figure 5.3: Poisson Cluster Analysis of Gonorrhea/HIV Co-Infections in Michigan Males from
2011 to 2013……………………………………………………………………………………201
Figure 5.4: Poisson Cluster Analysis of Gonorrhea/HIV Co-Infections in Michigan Males from
2011 to 2013 Detroit and Tri-County Area……………………………………………………..202

xiii

KEY TO ABBREVIATIONS

ACHSP

Advisory Committee for HIV & STD Prevention

AIDS

Acquired Immune Deficiency Syndrome

AmIn/AkNat American Indian, Alaskan Native, non-Hispanic
ART

Antiretroviral Therapy

ARV

Antiretroviral

Asian/HI/PI

Asian/Hawaiian, Pacific Islander non-Hispanic

CCDM

Control of Communicable Disease Manual

CCR-5

C-C Chemokine Receptor Type 5

CD4

Cluster of Differentiation 4

CD8

Cluster of Differentiation 8

CDC

Centers for Disease Control and Prevention

CI

Confidence Interval

D. C.

District of Columbia

DGI

Disseminated Gonococcal Infection

dL

Deciliter

DNA

Deoxyribonucleic Acid

ESRI

Environmental Systems Research Institute

Fem HCM

Heterosexual intercourse with female

GEE

General Estimating Equation

GIS

Geographic Information Systems

GP

Glycoprotiens

GUD

Genital Ulcerative Diseases

HAART

Highly Active Antiretroviral Therapy
xiv

HIPAA

Health Insurance Portability and Accountability Act

HIV

Human-Immunodeficiency Virus

HRadj

Adjusted Hazard Ratio

HR

Hazard Ratio

IDU

Injection Drug User

IRB

Institutional Review Board

JAMA

Journal of the American Medical Association

LC

Langerhas cells

LOS

Lipooligosaccharides

Male HCFR

Male who had sex with a female at risk for HIV

MDCH

Michigan Department of Community Health

MMWR

Morbidity and Mortality Weekly Report

MSM

Men who have sex with men

MSU

Michigan State University

Multi/Unk/Other

Multirace, Unknown, Other, non-Hispanic

N. gonorrheae

Neisseria gonorrheae

OR

Odds Ratio

PCP

Pneumocystis carinii pneumonia

PI

Protease Inhibitor

REACH

Reaching for Excellence in Adolescent Care and Health

RH

Relative Hazard

RNA

Ribonucleic Acid

ROR

Relative Odds Ratio

RR

Relative Risk

RSE

Root Mean Square Error
xv

STD

Sexually Transmitted Disease

Std Dev

Standard Deviation

T cell

Thymus cell

TB

Tuberculosis

TLR

Toll-Like Receptors

US

United States

VH

Viral Hepatitis

WHO

World Health Organization

κB

Kappa B

µL

Microliter

xvi

Chapter 1.
Background and Objectives
1.1 Human-Immunodeficiency Virus/Acquired Immune Deficiency Syndrome
Natural History
Human-Immunodeficiency Virus (HIV) is a lentivirus, a member of the retroviridae family and
is a single stranded RNA virus (1). Retroviruses use an enzyme called reverse transcriptase that
is a catalyst that transcribes double-stranded DNA from single-stranded RNA (1). There are two
species of HIV, HIV-1 and HIV-2, with HIV-1 being the most common infection and HIV-2
primarily associated with western Africa or cases linked to the area (2). HIV-1 consists of three
major groups, M, N, O, with M being the most common and is subdivided into 7 distinct
subtypes (2). HIV-2 is thought to have a lower virulence than that of HIV-1 (2). HIV is
surrounded by a lipid envelope with transmembrane glycoprotein spikes that mediate viral
adsorption to the host cell (1). Two important glycoproteins (GP) located on the viral envelope
are GP-120 and GP-41 which are important in attachment of the virus to leukocytes via the CD4
receptors on the cell surface for GP-120 and the CCR-5 receptor of macrophages and CD4 cells
for GP-41 (1).
The natural reservoir for HIV is humans and it is thought to have evolved from simian
immunodeficiency virus (2). The primary mode of transmission is person-to-person via
unprotected sexual intercourse, including penile-vaginal and penile-anal (2). Other modes of
transmission include the use of HIV infected needles and syringes, vertical transmission from
mother to infant during pregnancy, infant delivery, and by breast feeding (2). A final common
transmission mode is via transfusion with infected blood (2). Less common modes of
1

transmission include transplantation of infected organs, contact of infective secretions on
abraded skin or mucosa, as well as documented cases of transmission via pre-chewed food from
HIV-infected caregivers (2). The incubation period of HIV varies from case to case. However,
the majority of those infected will develop detectable antibodies in less than a month (2). Early
symptoms of HIV infection often include flu-like illness or mononucleosis-like illness shortly
after infection (2). Other symptoms that occur in the majority of patients include typical viral
immune response conditions such as fever, rash, inflammation of lymph nodes, and headaches
(2). Due to the flu/mononucleosis like symptoms that manifest shortly after infection, initial
infections often unnoticed (2). While the immune system starts to respond to the infection, early
clinical symptoms of latent HIV infection may not appear until years after initial infection
making it hard to pin point actual time the infection took place (2).
The period of communicability is largely unknown, but is suspected to range from shortly after
HIV onset and persists through one’s lifetime and is dependent on viral load (2). Viral load and
risk of transmission is thought to be the highest during early infection before the onset of
immune suppression (2). Use of antiretroviral (ARV) treatment has been shown to reduce viral
load in blood and genital secretions. Infected individuals who are on ARV treatment are still
considered infectious and are at risk of propagating and disseminating the virus (2).
There are four stages associated with HIV infection and they were introduced by the CDC in
2008 (3,4). One of the most important changes introduced with the implementation of HIV
stages is that Acquired Immune Deficiency Syndrome (AIDS) has been relabeled as HIV stage 3
with the goal to eliminate the term AIDS (3). Stage 1, Acute Infection, is defined as when “A
case does not have any of the conditions associated with severe HIV infection (called an AIDSdefining condition) and has >=500 CD4 cells/µL” (4). Stage 2, Clinical Latency, is defined as
2

when “A case has no AIDS-defining conditions, but the level of CD4 cells have fallen to 200499 cells/µL” (4). Stage 3 which traditionally is referred to as AIDS is defined as “Diagnosis
with any one of 26 AIDS-defining conditions which are indicative of a severe immune
deficiency, or a laboratory test demonstrating severe immune deficiency: CD4 count <200
cell/µL or CD4 percent <14%” (4). The definition of stage 3/AIDS was last changed in 1993 (2).
Stage Unknown is an intermediate stage which is defined as “A case of HIV without information
available on CD4 levels or AIDS-defining conditions” (4).
Stage 3 or AIDS positive status was first recognized in 1981 when a series of five cases of
Pneumocystis carinii pneumonia (PCP) was identified among assumingly healthy men (5). These
index cases were written about in the Morbidity and Mortality Weekly Report (MMWR)
describing men in Los Angeles. However, physicians were also seeing cases in San Francisco
and New York City (1,5). All of the men in the Los Angeles group were identified as
homosexuals and within 18 months epidemiologists and public health officials had identified all
major risk factors associated with AIDS (5). Once AIDS was identified the incidence of the
disease increased sharply from 1981 and peaked around 1993 (5). After 1993 the incidence of
AIDS began to decline with a sharp decline starting around 1996 and continuing through 2013
(5). Early AIDS cases were predominately white males, until 1996 when the trend shifted to
African Americans (5). The Control of Communicable Disease Manual (CCDM) states that the
estimated proportion of untreated HIV infected individuals that will develop AIDS is around
90% (2). In developing countries the estimated time to death after AIDS positive status is 3-5
years making the need for control of HIV progression to AIDS a major topic in reducing the
global burden of disease (2).

3

A study by Nelson et al. investigating the temporal trends of HIV in IDUs found that the relative
risk was increased (RR=2.27 CI 1.53-3.38) with the presence of any STD (105). The study also
found that MSM had an elevated risk (RR=1.90 CI 1.20-3.13) (105). In those that used
intravenous drugs on a nondaily basis but did not have sex the risk of HIV was RR=3.83 (CI
1.69-8.69) and in those that did not have sex but used daily the RR as compared to those that did
not have sex or use drugs was found to be RR=4.21 (CI 1.90-9.29) (105).
A 1995 study by O’Brien et al. explored the secular trends in CD4 counts in men and found that
from 1984 to 1991 there was no discernable change in the natural history of HIV-1 infections
(74). The study looked at the baseline CD4 counts (CD4 count at time of HIV diagnosis), a
previous seronegative visit, and scheduled post seroconversion visits and found that year of
seroconversion did not change the mean CD4 percent at any time point (74). Pre-seropositive
visits had a mean CD4 percent of 44% with a range of mean CD4 percent of 42% in 1990 to
46.7% in 1989 (74). The mean CD4 percent for visits 30 months after seroconversion was 28.3%
with a mean CD4 percent range of 23.7% in 1988 to 30.7% in 1986, however, at the 30 month
period the seroconversion years only ranged from 1984 to 1989, no 30 month data was collected
for 1990 or 1991 (74).
A longitudinal model for AIDS by Boscardin et al. investigated the relationship between HIV
viral load and CD4 found that individuals with higher baseline CD4 counts started and ended
with lower viral loads over a 12 month period (71). The study also found that in those with
higher baseline viral load have lower initial and final CD4 counts after a 12 month period as well
as a sharper rate decline (71). A 1994 paper by Zeger and Diggle observed that in the time
leading up to seroconversion there is a sudden drop in CD4 count, followed by continued CD4
decay as the disease progresses (73). Zeger and Diggle offered up a model that can be used to
4

predict an individual’s CD4 decay progression by combining the individual’s data with the
population’s curve of CD4 decay (73). A study by Saag et al. found that in a small sample group
of 68 adults and 9 children, none of the adults with CD4 cell counts of greater than 400
cells/mm3 were found to be viremic (80). Of the 17 adults with a CD4 cell count of between 200400 cells/mm3 3 adults (~18%) were found to be viremic (80). In individuals with a CD4 count
of less than 200 cells/mm3 15 of 16 were found to be viremic indicating that viral load is elevated
when CD4 was reduced (80). There were five children in the study that were either infected in
utero or during the perinatal period and all five were found to be viremic (80).
A study published in 1995 looking at the time to AIDS positive status in three groups of HIV
infected individuals found that the median time from diagnosis of HIV to AIDS in homosexual
males was 63 months, (25th percentile=46 months, 75th percentile=85 months) (62). In
homosexuals males the range time to AIDS was found to be 35-114 months (62). The study also
found that aside from hemophiliacs, homosexual males had the youngest age of HIV diagnosis
with a median age of 29 years old (62). A study published in 1997 by Alcabes et al. found that as
time from infection increased the hazard for AIDS positive status also increases with a major
jump in hazard at 4 years after seroconversion, and another jump in hazard at around 8 years
(63). The study also showed that those who were born before 1951, the older cohort, compared
to those born between 1951 and 1962, the younger cohort, had a higher rate of progression to
AIDS RH=1.76 (1.08-2.85) (63). These studies show that progression to AIDS is a complex path
with many factors that influence the progression of HIV to AIDS. Factors that influenced
progression to AIDS were found to be base line CD4, change in CD4, as well as time
progression (increase in lag time) from HIV diagnosis.
Epidemiology
5

A 2007 global HIV review by Kilmarx et al. found that the global rate of HIV was around 1%
which accounts for 33 million individuals (54). Kilmarx et al. found that 22 million of the
estimated 33 million HIV positive individuals are located in Sub-Saharan Africa up from 20.4
million in 2001 (54). The review also found that there was an estimated 2.7 million newly
infected adults and children in 2007 which was lower than the 3 million newly infected reported
in 2001 (54). In North America there was an estimated 1.2 million HIV infected individuals with
54,000 new infections in 2007 (54). Kilmarx et al. stated that although there was a rise in HIV
positive individuals as well as a rise in the newly infected, rates remained stable in North
America (54). The review also stated that in 2006 the estimated incidence per 100,000 in African
American men in the US was 115.7, whereas white males in the US have an estimated incidence
per 100,000 of 19.6 (54).
The World Health Organization (WHO) estimated in 2012 there were 35.3 million people living
with HIV with Sub-Saharan Africa being most affected region (6). In Sub-Saharan Africa it is
estimated that 1 in 20 adults in the region are infected with HIV with around 69% of the total
population living with HIV (6). The WHO also estimated that in 2012 there were a total of 2.3
million infections and 1.6 million deaths globally due to AIDS related causes (6). Roughly 9.7
million people globally in 2012 were estimated to be on ARV therapy (6). Under the 2013 WHO
ARV guidelines there are an estimated 26 million people eligible for therapy leaving a large gap
between those that are eligible and those that are receiving ARV treatment (6).
US HIV
As of 2011 the Centers for Disease Control and Prevention (CDC) estimated that there were
approximately 1.2 million U.S. citizens 13 years or older living with HIV and over 200,000

6

undiagnosed cases (7). CDC estimates that in the U.S. there are around 50,000 new cases a year,
with an estimated 47,500 new cases in 2010 (7). The majority of infections occur in Men Who
Have Sex with Men (MSM) of all racial groups and in African Americans (7, 19, 20, 22). From
2006-2009 the only group to show significant increases in HIV infections were young (13-24
year olds), black MSM, despite white MSM accounting for the largest number of new infections
(7). According to the CDC in 2009, MSM accounted for 2% of the U.S. population and make up
61% of all new infections and 49% of the 2008 living U.S. HIV population (7). Heterosexuals in
2008 accounted for 28% of U.S. individuals living with HIV and 27% of the 2009 new HIV
infections (7). African Americans in 2010 accounted for 12.6% of the US population and 45% of
all new HIV infections in 2010 (16). A study done by Goodenow et al. in 2002 stated that a
majority of men diagnosed with AIDS in their 20s or 30s were probably infected in their teens
due to the prolonged nature of HIV infections (21). Goodenow et al. also notes that many
American youths are sexually active during high school, that heterosexual exclusive males are
older than young MSM, and that by age 18 60% of men that identify themselves as bisexual have
had an STD (21).
In 2011 the CDC estimated that there were over 32,000 HIV stage 3 (AIDS positive) diagnosed
individuals in the U.S. with 24,000 of them being in adult and adolescent males (8). There were
over 1.1 million AIDS diagnoses by 2010 in the United States with 487,692 individuals living
with AIDS and 15,000 AIDS related deaths in 2010 alone (8). A 2011 study by Hughes et al.
estimated that by 2016 half of the US HIV positive population will be over the age of 50
presenting new problems in relation to the understanding of the progression of this disease and
other comorbidities (100). The top three states according to number of HIV infections in 2011
were California (n=5,973), Florida (n=5,403), and Texas (n=5,065). The top three states of AIDS
7

diagnosis were California (n=3,622), New York (n=3,574), and Florida (n=3,440) in 2011
respectively (8,9).
A study by Patrick et al. in 2012 that sought to investigate the risk factors and substance abuse
issues related to HIV/AIDS in young adults in the US cited that in 2012 there were roughly 1.2
million individuals in the US with HIV, with around 20% of HIV positive individuals being
unaware of the serostatus (99). The study consisted of over 7,000 young adults, 52% female and
around 70% white (99). The study found that in US young adults (21-30 years old) 24% of
women and 34% of males report more than 5 lifetime sexual partners (99). The study also found
that in this age group 55% report that they seldom or never use a condom which has contributed
to the increasing prevalence of HIV infection in the younger population (99). The study found
that in individuals who used illicit drugs (including marijuana) adherence to condom usage was
decreased (99). A study by Karlovsky et al. males 50 years old or older found that few males 50
years or older who are sexually active use condoms or are tested for a sexually transmitted
disease (102). The study noted that with the advent of erectile dysfunction medication this group
that was traditionally overlooked has become a new risk group for HIV and STD infections and
the spread of these diseases (102). The study found that in Florida as well as in the US as a
whole the number of HIV infections in males >50 years old has been slowly increasing from
1998 to 2004 (102). The study also found that in both the 55-64 year old age group and the 65+
year old age group that the number of cases of N. gonorrheae have been increasing from 1997 to
2002 (102). These findings represent a new risk group, sexually active adults over the age of 50,
which have been previously ignored.
Michigan HIV

8

This study will take place in the State of Michigan where in 2011 ranked 20th in HIV diagnoses
and 18th in AIDS diagnoses (9). The January 2014 Annual HIV Surveillance Analysis estimated
that there were approximately 19,800 individuals in the state of Michigan who were HIV
positive of which an estimated 15,440 (78%) were males (10). Of the 15,440 males with HIV,
men who have sex with men make up 55% of prevalent HIV cases between the combine risk
groups of MSM and injection drug user (IDU)/MSM in the state of Michigan (10). Heterosexual
contacts make up 19% of those with HIV infection, 8% were non-MSM IDU, 1% had a perinatal
exposure and 17% had an undetermined source of transmission (10). The most frequent age
group for HIV diagnosis was 30-39 year old (33%) with the combined age group of 20-29 at
32% (10). The majority of HIV diagnosis occurred in the Detroit Metropolitan Area (consisting
of Lapeer, Macomb, Monroe, Oakland, St. Clair, and Wayne counties) (64%) despite only
accounting for 43% of the total Michigan population (10). Males made up the majority of HIV
infected individuals with a reported 12,510 cases (10). Within this group 71% are MSM
(combined risk groups of MSM and MSM/IDU), 5% are heterosexuals, and 17% have an
undetermined mode of transmission (10). As of January 2014 there was an estimated 8,207 stage
3/AIDS positive individuals in the state of Michigan with a reported 428 new stage 3/AIDS
positive individuals identified in 2013 (10).
An updated version of the Annual HIV Surveillance Report was released in July 2014 and
showed little difference from the Annual HIV Surveillance Analysis released in January 2014.
The July 2014 report estimated a slightly lower total prevalence, 19,100 as compared to the
January 2014 estimate of 19,800 (10, 111). Males accounted for 78% of HIV infected individuals
with MSM (MSM and IDU/MSM) accounting for 55% of total HIV infected individuals, 71% of
HIV infected males (111). African Americans had the highest rate of HIV infection
9

(655/100,000) followed by Hispanics (148/100,000) (111). HIV remained highest in the 30-39
age group with the combined group of 20-29 year olds only slightly behind (33% and 32%
respectively) (111). Metropolitan Detroit area geographically had the highest prevalence of HIV
infections as well as the highest rate, 238/100,000 (111). The total number of individuals living
with AIDS (Stage 3 HIV) increased from 8,207 reported in January to 8,240 in the July 2014
report, with the new 2013 AIDS total at 471 (111). Of all those diagnosed with HIV up to
January 2014, 18,794 of 27,407 progressed to AIDS/Stage 3 positive status (111). Among the
AIDS positive individuals 35% were diagnosed with AIDS at the same time of their HIV
diagnosis, 8% had a “short” lag time (1-12 months after HIV diagnosis) to AIDS positive status,
and 26% had a “long” lag period (more than 12 months from HIV diagnosis) (111). Males with
HIV had a higher percent AIDS progression than females, 70% of males as compared to 64% of
females (111). Within males with HIV 37% were diagnosed as AIDS positive at the time of their
HIV diagnosis, 8% had a short lag time, and 25% of males had a long lag time to AIDS positive
status (111). In females with HIV the percent of concurrent AIDS diagnosis (HIV and AIDS
diagnosis at the same time) was 28%, 8% had a short lag time, and 28% had a long lag time
(111).
Acquired Immune Deficiency Syndrome
HIV progression has been studied from a multi-faceted vantage point. Early studies of HIV in
homosexual males were modeled on the previous knowledge of the epidemiology of hepatitis B
and other well-known Sexually Transmitted Diseases (STDs) (42). Studies like the 1987 Darrow
et al. article focused on number of sexual partners, non-steady sexual partners, drug use, and
sexual practices (42). A multicenter AIDS cohort study paper published in 1992 looking at white
men who have sex with men (MSM) found that immune response rates are predictive of
10

progression to AIDS positive status (57). The final model reached by Saah et al. included the
percent of CD4 cell 5% lower at the first visit had a relative hazard (RH) of 2.08 (1.86-2.31)
(57). Other factors that were found to be significant were a decline of 5% in the percent of CD4
between first and second visit [HR= 2.09 (1.80-2.42)], decline of 1 g/dL of hemoglobin between
first and second visit [HR=1.21 (1.02-1.44)], and the level of IgA 200 mg/dL higher at first visit
[HR=1.91 (1.21-3.00) (57)].
A 1996 study of participants from the Veterans Affairs Cooperative Study on AIDS found that
baseline CD4 counts and levels of HIV-1 RNA in the plasma are the main predictors in modeling
the risk of AIDS (72). The study found that a 75% decrease in HIV-1 RNA was associated with
an unadjusted relative risk (RR) of 0.44 (0.22-0.75) with regard to the progression of AIDS and
an adjusted RR of 0.44 (0.23-0.81), model was adjusted for CD4 cell count, ß2-microglobulin,
and treatment (72). The study also found that a 10% increase in CD4 count was associated with
an unadjusted RR of AIDS progression of 0.45 (0.27-0.77) and an adjusted RR of 0.48 (0.280.82) (72). Antiretroviral therapy was found to be significant in the univariate model for
protection of AIDS progression with a RR of 0.62 (0.43-0.92) but was not significant in the
multivariate model, RR=1.08 (0.66-1.77) (72). The study also looked at the effect of a 15%
decrease in ß2-microglobulin. However, it was neither significant in the univariate nor the
multivariate model (72). A Kaplan-Meier analysis looking at time to AIDS found that when
neither HIV-1 RNA or CD4 changed that the median survival time was around 39 months,
whereas when CD4 increased and HIV-1 RNA levels decreased only around 20% of individuals
progressed to AIDS over a 5 year (60 month) study period (72). A research letter in JAMA in
2007 captured the percent of variability due to baseline viral load in prediction of time to AIDS
and death (HIV progression) as contributing 47% to the time to AIDS and 50% of the time to
11

death making it the major contributing factor in each (88). The study was done looking at only
untreated patients (88). These findings were significant because they show how baseline levels
can be used to help to identify individuals with increased risk of AIDS or death.
Miller et al. found that as CD4 cell count decreased the rate of HIV progression events increased
(79) Miller et al. also found that when stratifying those with a CD4 cell count of at least 200
cells/mm3 by previous cell count that those with lower previous recorded cell counts had a
greater hazard of having a progression event (79).
Taylor et al. in 1989 published a study that plotted AIDS survival against CD4 percent, absolute
CD4 count, and CD4:CD8 ratio (77). The study found that as CD4 percent decreased there was
also a decrease in the median time to AIDS positive status (77). The lowest CD4 percent group
was from 0-12% and the median time to AIDS was around 9 months (77). The study also found a
similar trend when modeling time to AIDS and CD4 count, as the CD4 count decreased (in
defined groupings) there was also a reduction in AIDS free time (77). The group with the lowest
hazard of AIDS progression was in the greater than or equal to 510 cells/mm3 group with less
than 10% of individuals progressing to AIDS in 3 years (77). A final major finding of this study
was that out of the three biological markers studied, CD4 percent may be the best indicator and
have the least variability when predicting AIDS progression, however, CD4 count and CD4:CD8
ratio are also useful tools for modeling AIDS progression (77).
Yu et al. found that there is a linear relationship between CD4 count and CD4 percent (76). The
study found that a CD4% of 5% corresponds to a CD4 cell count of 45 cells/mm3 (CI 17-117
cells/mm3), CD4% of 15% corresponds to a CD4 cell count of 182 cells/mm3 (CI 64-499), and a
CD4% of 30% corresponds to a CD4 cell count of 438 cells/mm3 (CI 132-1395) (76). The study

12

also found that among three groups, homosexual/bisexual, injecting drug users, and
heterosexuals, homosexual/bisexuals have the highest CD4 cell counts in both AIDS-free and
AIDS positive groups (76). In AIDS-free individuals those that are not on zidovudine therapy or
PCP prophylaxis have the highest CD4 count in the low CD% grouping (5%) and those that are
only on zidovudine have the highest CD4 counts in the high percent grouping (30%) (76). In the
AIDS positive group CD4 counts in the 5% CD4 group were highest in those not on any therapy,
and in the 30% CD4 group, CD4 counts were highest in the zidovudine only group (76).
In 2010 looked at the risk of death or AIDs between 1979 and 2001 found that the hazard of
death or conversion to AIDS positive status remained relatively the same from 1979 to
1993(0.89 in 1979-1990 and 0.90 in 1991-1993) with a drop in the hazard in the time period of
1997 to 2001 to 0.37 when compared to the time period 1994-1996 (56). A UK study published
in 1997 found that within 10 years of HIV positive diagnosis 60% of individuals were found to
be AIDS positive (59). Within 10 years from HIV diagnosis around 48% of HIV positive males
died from any cause and that those diagnosed with HIV at an older age had a faster time to AIDS
positive status as well as a shorter survival time (59).
A 1998 study by Yerly et al. of 394 Swiss HIV-1 infected patients found that in unadjusted
models Log HIV-1 RNA was significantly associated with an increased in relative hazard (RH)
of death (75). As viral load levels rose the risk of death also increased in a dose response fashion.
When looking at CD4 cell counts an inverse relationship was found where decreases in CD4
increased the hazard of death in a dose response relationship. While the association remained
significant between CD4 count and hazard of death in a multivariate model, viral load was not
found to be a significant marker for hazard of death (75). Yerly et al. also found significant
association between viral load, CD4 count and ß2-microglobulin in the univariate analysis (75).
13

A 2009 study looking at risk of death in HIV individuals that deferred treatment was compared
to those that started treatment early and found in the 351-500 CD4 cell count group deferral of
antiretrovirals resulted in a RR of 1.69 (CI 1.26-2.26) in a model that included sex, age and
baseline CD4 count (78). In the group that had a CD4 cell count of over 500, deferral of
antiretroviral therapy, however, baseline CD4 count did not (78). In a second model that
included viral load found in the 351-500 CD4 cell count group that deferral of antiretroviral
therapy increased risk of death, whereas baseline CD4 count [RR=0.74 (CI 0.55-1.00)] and
baseline HIV RNA level [RR=1.11 (CI 0.96-1.28)] where not found to be significant risk factors
for death (78). The CD4 cell count of more than 500 found similar results with regard to deferral
of antiretrovirals, CD4 count and viral load when HIV RNA levels were added (78). A major
finding in this study was that longer deferment of antiretroviral therapy was associated with an
increase in risk of death, whereas baseline CD4 count and baseline viral load were not found to
be associated with risk of death suggestive of the fact that the progression of CD4 count is more
important than the baseline level (78).
A 1996 paper by Turner et al found that among those diagnosed with AIDS women had a longer
median survival time in months than men as well as the younger age group as defined by those
less than 40 years old (70). When AIDS severity (based on judgment by expert panel) was
characterized into three categories, based off AIDS-defining diagnosis and complications seen
within 3 months of AIDS diagnosis, median time to death decreased as severity increased (70).
The administration and use of Zidovudine, a nucleoside analog reverse-transcriptase inhibitor,
increased the median time of survival time (70).
A study published in the Lancet in 2000 investigating the mean time to AIDS and death by age
of seroconversion found that without treatment as age at HIV infection increased, the time to
14

AIDS positive status decreased (65). In those infected from ages 15-24 years old the median time
to AIDS was 11 years while the time to AIDS in those infected with HIV after 65 years old was
5 years (65). Mortality was also seen to increase with age (65).
Antiretroviral Therapy
Antiretroviral therapy (ART) has been shown in multiple studies to reduce viral load, increase
the time to AIDS, and in many cases has been shown to decrease the rate of CD4 destruction and
facilitate the resurgence of CD4 counts in the body. An early concern of ART and Highly Active
Antiretroviral Therapy (HAART) was that those on the treatment would engage in risky sexual
history believing that they were not at risk for transmission (67). A 2012 study by Fu et al. found
that those that started HAART treatment have a 75% reduction in unprotected sex (OR=0.25
0.19-0.32) (67). Those that started HAART did not see a reduction in sexual activity, there was
however a decrease in the use of injection drugs (67). A negative finding that Fu et al. found was
that in some drug users the activity of needle-sharing increased by around 2 fold (OR=1.99 1.572.52) (67). Fu et al. also noted that the changes lasted for at least 5 years regardless of adherence
to HAART (67). Thus the early concerns of engagement of risky sexual behaviors did not hold
true, however, in some injection drug users needle-sharing did increase.
Using patients from the Swiss HIV cohort researchers looked at viremia and progression to
AIDS or death in 2674 individuals that started HAART between 1995-1998 (82). The study
found that within a year, over 90% of individuals who started treatment during the study period
were not viremic (defined as a viral load of less than 400 copies/mL). Individuals on more drugs
were less likely to become viremic (82).

15

Some studies have found that use of protease-inhibitor-containing highly active antiretroviral
therapy (HAART) did not guarantee a reduction in viral load with some studies reporting
upwards of 49% of individuals on HAART did not have viral suppression (81). Kaufmann et al.
showed that while a group of 101 HIV-1 patients on HAART were predominately viremic after
at least a 3 month study, that HAART did increase their CD4 counts despite not reducing or
suppressing viral load counts (81). Kauffmann et al. point out that while HAART does not easily
reduce viral load, which until better drugs become available the benefit of increase in CD4 count
should not be over looked (81). A second study by Kaufmann et al. showed that the longer HIV
is left untreated, the harder it is to reach normal CD4 cell count level (84). At the end of 4 years
nearly all patients reached a cell count of greater than 200 cells/µL and 74% reaching a cell
count of greater than 500 cells/µL (84).
A 2001 study from the 1996-1997 Aquitaine cohort found that in patients starting a protease
inhibitor during the first 4 months the change in CD4 cell count did not influence the rate of
opportunistic infections in participants (87). After the initial 4 month period (120 days) change in
CD4 count was found to reduce the rate of opportunistic infections, for every 50 cells/mm3 there
was a reduction in opportunistic infections by 60% (87). The study also found that the median
change in CD4 cell count was around 111 cells/mm3 over 18 months (87). An important finding
of the study was that while protease inhibitors increased CD4 cell counts, there is an early delay
in the ability of the immune system to reduce opportunistic infections during the initial CD4 cell
count ramp up (87).
A 1997 study of eight patients by Autran et al. found that in 3 of 8 patients on HAART viral load
was reduced to non-detectable levels over a year period (83). CD4 cell count rose with treatment
of HAART during the first 2 weeks of treatment with the remaining year of treatment increasing
16

CD4 counts at a slower pace (83). A 2003 study published in the Lancet looking at survival and
progression to AIDS in those on HAART found 26% of the 7740 participants died from the 22
cohorts looked at (85). Hazard of death was compared to the pre-1997 data, and showed that
there was a decrease in hazard of death (HR=0.47 0.39-0.56) in 1997, HAART usage 22%, and a
further decrease (HR=0.16 0.12-0.22) in 2001, HAART usage 57% (85). This study along with
the one mentioned in the previous paragraphs show the difference in the range of effectiveness of
ARV cocktails.
A study looking at HIV care among African Americans found that in the US in 2010 44% of new
infections and 44% of those living with HIV are African Americans, despite African Americans
only making up 12% of the US population (110). The study also found that around 75% of
African Americans who are HIV positive can be linked back to care, 10% shy of the National
HIV/AIDS goal mark of 85% (110). The study also found that while 3/4th of those with HIV
could be linked to care, only 48% retained care and only 46% were prescribed ART (110). This
resulted in only 35% of African Americans achieving viral suppression (110). The study also
found that in African Americans less than 25 years old, their levels of care and viral suppression
were lower than for African Americans 25 years or older at every stage of the continuum (110).
The article concluded by speculating that the reason for the gaps in care of HIV may not be only
access to care issues but also barriers due to poverty and social stigma that must be addressed
and overcome for public health to make an impact (110).
A study from 1994 modeling the CD4 counts over time in AIDS positive individuals taking the
drug zidovudine found that after an early spike in CD4 counts, around week 10 of treatment CD4
counts rapidly decline, obtaining baseline levels by week 20 and reaching an asymptote around
week 90 (61). One of the major limitations of the study is that as time progressed, individuals
17

dropped off due to death and missing observations (61). However, when looking at only
survivors the CD4 curves look similar to the total population CD4 curve and show that in the
sample AIDS population CD4 counts dropped below 10 cells/µL at around one year (61). A
study looking at individuals diagnosed with AIDS in San Francisco from 1993-1996 found that
hazard of death after AIDS positive status increased with infection of opportunistic infection
(RH=1.97, 1.8-2.2) (64). The use of antiretroviral therapy (ART) was found to be protective of
mortality with ART before AIDS diagnosis and post AIDS protease inhibitor (PI) treatment
having the greatest impact on survivability (64).
Antiretrovirals are not the only drugs used to treat the progression of HIV to AIDS in positive
individuals. A study by Longini et al. looking at the use of zidovudine and aerosolized
pentamidine (a common antibiotic used to prevent Pneumocystis carinii pneumonia now known
as Pneumocystis jiroveci pneumonia) found that starting treatment prior to AIDS can increase the
time between disease progression stages (higher CD4 counts for longer) and keep individuals out
of the AIDS positive category longer (86). The study showed that prompt treatment with
zidovudine and antibiotics prolong the time to AIDS positive status (86).
Research shows that not all HAART or ARVs work the same, and it may not be appropriate to
group them together (have you ever been on ARVs). There is a need to study the patterns of
reduction in HIV progression among those with the disease on the basis of their treatment. This
may give insight into variances between treatments in similar individuals.
1.2 Neisseria gonorrheae
Natural History

18

Neisseria gonorrhea is a gram-negative bacterium that is paired, bean-shaped, and has the ability
to attach to each other along the flat sides of the cell (1). Gonorrhea is a sexually transmitted
disease (STD) and has deep roots in early civilizations easily dating back to ancient Greece (1).
The major virulence factor of N. gonorrheae is the presence of pili and other surface molecules
that facilitate the bacteria attaching to each other and assist in the invasion and infection of host
tissue (1). The pili also aid in the defense of the bacteria against macrophages and neutrophils,
and N. gonorrheae creates a protease that attacks IgA on mucosal surfaces (1).
N. gonorrheae can cause urethritis (inflammation of the urethra), epididymitis (inflammation of
the epididymis, tube between the testicles and the vas deferens), proctitis (inflammation of the
rectum), cervicitis (inflammation of the cervix), Bartholinitis (formation of a cyst in the
Bartholin gland), pelvic inflammatory disease, and pharyngitis (inflammation of the pharynx)
(2). In males symptoms start 2-7 days after infection and typically include discharge of pus from
the urethra as well as painful urination (dysuria) (2). Rectal infection may include discharge, anal
itching, soreness, bleeding, and painful bowel movements, whereas infections of the throat may
manifest in sore throat but is typically asymptomatic (11). Although considered rare, some
strains of N. gonorrheae can also cause arthritis-dermatitis syndrome or gonorrhea arthritis due
to what is called Disseminated Gonococcal Infection (DGI) which is the spread of the bacteria to
the joints and blood and can be a life threatening condition (2, 11). It is estimated that
approximately 10% of males with N. gonorrheae are asymptomatic (1).
Sexual contact is the main source of N. gonorrheae and incubation is usually 1 to 14 days, but
can be longer (2). The period of communicability varies in individuals and is largely dependent
on if the individual receives treatment (2). In individuals who do not seek treatment the period of
communicability can extend for months, whereas treatment has been shown to end risk of
19

communicability in hours (2). Untreated gonorrhea in males can develop into epididymitis,
inflammation of the tube that connects the testicles and the vas deferens, and can lead to
infertility (11). While N. gonorrheae infection can be treated with antibiotics, prevention of
outbreaks has proven difficult and the spread of drug-resistant strains have introduced a new
problem for public health departments (11, 39). While many areas in the U.S. and throughout the
globe are experiencing outbreaks of N. gonorrheae infections, prevalence of the disease is found
to be highest in communities with lower social economic status (2).
Epidemiology
In 2011 WHO estimated that roughly 88 million of the estimated 448 million curable STD
infections reported that year were due to N. gonorrheae (12). Gonorrhea is one of the most
reported infectious diseases in the United States (15). In the United States in 2012, there were
over 334,000 gonorrhea cases resulting in an increased rate of 4.1% from the previous year (13).
In the U.S. the prevalence among females is greater than males; however, the rate in males is
increasing at a higher rate than in women (13). The rates were highest in 2012 among teens and
young adults with the highest age group among men 20-24 years, while the highest overall
increase was in males 30-34 (13). Where age has been reported, 95% of all gonorrhea fall
between 15-44 years old with the age group of 30-34 years having the highest increase in rate
from 2011 (13). The only age group to see a decrease in the rate of gonorrhea in 2012 was seen
in the 15-19 years age group (13). Manhart el al. found that risk of N. gonorrheae infection
decreases with age and is increased in African Americans (26). The rate in African Americans in
the United States was around 15 times higher than in Caucasians despite an overall decrease in
African Americans and a 22% increase in Caucasians (13). The top three states/areas in 2012 for
reported gonorrhea cases were Washington D.C. (388.7/100,000), Mississippi (230.8/100,000),
20

and Louisiana (194.0/100,000) respectively (13). A study published in 2014 that reviewed
ciprofloxacin resistance and N. gonorrheae infection incidence rates from 1991-2006 and
estimated there are 820,000 cases of gonorrhea each year and which have a direct medical cost of
$162 million (109).
A study by Weinstock et al. looking at STDs in American youths found that the age group of 1524 year olds make up a quarter of the sexually active population in the US (48). In 2000 48% of
the estimated 19 million STDs occurred in 15-24 year olds (48). The study also found that in
2000 approximately 60% of N. gonorrheae infections were in 15-24 year olds (48). A study by
Gunn et al. investigating sexual gender preferences found that in San Diego between 1997 and
2003 around 16% of MSM who were tested for N. gonorrheae were positive (52).
A geographic study of gonorrheal infections in Baltimore, MD found that over a two year period
there were over 6,000 infections of which 9% were those with repeat infections (68). The study
found those with reoccurring infections, who were more likely to be younger women in the 1524 year old age group (68). The study also found that in census tracts with higher infections rates
per 1,000 the number of repeat infections was also higher with the center of the city being the
area with the highest rates and highest occurrences of re-infections (68). While repeated N.
gonorrheae infections are not overwhelmingly common, (<10%), this study was able to show
that they do occur and they are not equally spread across the city.
As of the 2010 Census, Michigan has the 8th largest population in the US. In 2011 there were
over 13,000 cases of gonorrhea reported in the state of Michigan with the highest rate reported in
the 20-24 age group (14). The 20-24 age group accounts for 6.7% of the population of Michigan,
yet accounts for 34% of all gonorrhea cases (14). The overall rate in African Americans in the

21

state of Michigan was roughly 24 times higher than the rate in Caucasians (14). Males account
for 41% of all 2011 gonorrhea cases in the state of Michigan (14). Of the reported 13,070
gonorrhea cases 2% (n=259) were also HIV positive (14). In 2012 the State of Michigan reported
a rate of 127.4/100,000 (13).
1.3 Co-infections
While researchers studying STDs have been aware that there is a significant amount of coinfection periods in those that have a reported infection, the epidemiology and exploration of the
possible synergistic relationship that exists before, during and after co-infection has not
adequately been investigated (93). It was estimated using 2006 US dollars that the annual direct
medical cost of STDs in the US was around $15 billion with over half being spent on the
diagnosis and treatment of non-HIV STDs (89). The 1997 Advisory Committee for HIV & STD
Prevention (ACHSP) investigated the relationship between HIV and STDs and recommended
that the early diagnosis and treatment of treatable STDs should be a major focus of all HIV
prevention programs at all levels of public health (24). The committee also stated that in areas
where it is believed that STDs assist in the transmission of HIV that screening and treatment
programs should be expanded where they currently exist or created where there are gaps (24).
The committee also suggested that public health integration between private and public health
agencies should take joint responsibility for the implementation and evaluation of the proposed
strategies (24). The concluding message that the 1997 committee left with was that a major step
in the prevention of HIV is the early diagnosis and treatment of curable STDs (24).
A 1998 study by Chene et al. found that higher CD4 counts were associated with lower rate of
opportunistic infections (66). Study participants were prescribed at least one protease inhibitor

22

and the study found that for each 50% increase of CD4 at baseline the risk of death or
opportunistic infection was reduced by 23% (66). A 50% increase in CD4 cell count during the
follow-up period (median follow-up time 230 days) was associated with a 9% reduction in the
risk of opportunistic infection or death (66). Study participants that increased their CD4 counts
at second visit had fewer opportunistic infections than those that decreased in CD4 count (66).
HIV/Non-Gonorrheal STDs
The need to understand the co-infection dynamics of HIV and other STDs is an important aspect
in developing and implementing effective outreach and prevention programs (93). In a study by
Zhang et al. in 2007 looking at STD co-infections in HIV infected males 94% of study
participants had evidence of at least one co-infection, with 50% (8/16) HIV positive MSM being
infected with syphilis (25). Another study by Buchacz et al. that found in men during
syphilis/HIV co-infection periods viral load increased and CD4+ cell count decreased (49). The
study also found that after treatment for syphilis viral concentrations fell and CD4 cell counts
rose (49). A 2005 study by Kahn et al. estimated that in adolescents entering juvenile facilities
5.9% of males were infected with chlamydia and 1.3% of males were infected with gonorrhea.
Of those infected with N. gonorrheae 51% of males were also infected with Chlamydia
trachomatis (15).
A study done by Huhn et al. using Chicago and Illinois STD/HIV surveillance data found that of
the 43,517 patients seen in 2002 around 13% received a positive HIV test (96). The study also
found that in 2002 there were 308 new cases of HIV and that within that population the rate of
co-infection (STD infection at time of HIV diagnosis) was 23% (96). The study found that the
risk of HIV/STD co-infection was highest in MSM and those over the age of 30 (96). The

23

majority of STD infections were found in males, and when looking at STDs regardless of HIV
status the age group with the greatest infection rate was the 20-24 year olds (96). Syphilis had the
highest unadjusted OR (11.0 CI 7.7-15.8) for co-STD infection in newly diagnosed HIV
individuals, with gonorrhea coming in second with an OR of 2.2 (CI 1.4-3.3) (96).
In a review of HIV/AIDS in Africa Corbett et al. found that a major effect of STD infection was
the rise in HIV transmission (94). The review outlines that during co-infection periods HIV can
increase the morbidity and fatality rate of the co-infecting pathogen in immunosuppressed
individuals (94). There was an increase in the incidence of surrounding pathogens due to the fact
that infection more often results in symptomatic presentation of the infection, resulting in more
cases coming in for treatment and diagnosis (94). The review also speculates that increased
transmission of the non-HIV pathogen may be due to an increase in the virulence of the disease
(94). On the other side of the coin, the effect of other pathogens on HIV is that there is a decrease
in the survival rate due not only to the elevated virulence of the non-HIV pathogen but also due
to the possible acceleration of HIV progression via immune system stimulation (stimulation of
immune response offers more targets for HIV infection resulting in a higher viral load) (94).
Another effect that STDs have on HIV is that there is an elevated state of infectivity/transmission
resulting from the immune response of STDs that cause genital inflammation (94). The authors
make a note to point out that unlike counties like the US, the majority of African HIV
transmission is through heterosexual encounters, not MSM intercourse (94).
The review goes on to speculate that some of the problems with STDs in Africa can be traced
back not only to behavioral factors (condom usage) but also to social factors such as poor health
service delivery and that it is really a combination of the two (94). A need for HIV control
measures in Africa to include improved STD diagnosis and treatment, with the aim to reduce the
24

prevalence of STDs in Africa and hopefully thereby reduce the HIV enhancement that can cause
higher rates of transmission (94). The review goes on to outline some of the research priorities
that need to be considered with regard to HIV and STD prevention. One area is that there is a
need for combined intervention practices, mass treatment as well as syndromic treatment, and
that these need to be targeted at populations with high rates of curable STDs (94). Another
research area that needs to be considered are trials that seek to test interventions on high risk
groups, as well as trials aimed at keeping adolescents and young HIV naïve individuals HIV and
STD free (94). Other areas for consideration are on increasing STD treatment coverage,
monitoring of treatment, development of treatment algorithms, and assessment of herpes
treatment options and HIV transmission (94).
HIV/N. gonorrheae
Despite recent multiple outbreaks of N. gonorrheae as well as the emergence of antibiotic
resistant gonorrhea and the fact that N. gonorrheae is one of the most reported diseases in the
United States, the relationship between HIV and gonorrhea has yet to be fully explored. Studies
show that for many HIV/N. gonorrheae co-infected individuals partner service data is not
available (33). It has been suggested that HIV and N. gonorrheae share an epidemiological
synergistic relationship (24). Research shows that in HIV infected males that are co-infected with
N. gonorrheae viral RNA is increased in the semen roughly 10-fold and drop to pre-gonorrhea
infection levels after the start of antibiotic treatment (24). The literature also suggests that there
is at least a two-fold increase in the risk of men spreading HIV in their ejaculate who are coinfected with N. gonorrheae than in men that are not currently co-infected (31). Mayaurd and
McCormick state that among HIV infected males there can be an upwards of an 8-fold increase

25

in the viral load in and an increase in viral load found in the vaginal secretions of HIV infected
women as compared to individuals with no co-infections (40).
Reports show that in the early years of the HIV epidemic N. gonorrheae infections were on the
decline in the United States until the start of its increase in 1990 in MSM sexual networks (34,
39). A study by Fox et al. reported that in 1998 rates of N. gonorrheae infection increased
dramatically in 21 states (34). N. gonorrheae HIV co-infection in males is documented as
peaking in 2005 and by 2007 had returned to levels observed in 1997 (35). Studies show that
among HIV infected males the rate of N. gonorrheae infection is 20 times higher than those who
are HIV negative (35). A 12 year trend in gonorrhea and sexual risks in MSM from 1990 to 2001
found that in HIV infected men the initial gonorrhea increase was between 1995 and 1997, with
a brief decline in gonorrhea/HIV co-infections from 1997 to 1998 (37). The second period
according to Rietmeijer was from 1999 to 2000 (37).
A multicenter AIDS cohort study of four sites (Chicago, Los Angeles, Pittsburg,
Baltimore/Washington DC) in 1987 found that lifetime history of gonorrhea was between 4067% depending on site with Los Angeles and Chicago having the highest rates (58). The 1987
article also found that the percent of white MSM with HIV was higher in Los Angeles and
Chicago (51% and 43%) respectively than in the other two sites (58). A separate study using the
multicenter AIDS cohort study published in 1987 found that N. gonorrheae infection increased
the risk of HIV infection in males (60). The Chmiel et al. study found that risk of HIV infection
increases with the frequency of rectal N. gonorrheae infection (60). The unadjusted relative odd
ratio (ROR) for 1-3 infections was 3.55 and increased to 13.49 in the 4+ N. gonorrheae
infections (60). The multivariate model showed that the ROR for 1-3 infections was 1.65 (CI
1.39-1.96) and the ROR for 4 or more infections was 3.89 (CI 2.45-6.20) (60). The model also
26

looked at the grouping of lifetime infection with other STDs which included urethral gonorrhea,
syphilis, urethritis, and genital herpes and found an increase in HIV seroconversion that
increased with re-occurrence of STD infection (60). The occurrence of one lifetime infection
resulted in a multivariate ROR of 1.27 (CI 1.04-1.56) as compared to zero lifetime infections,
two infections increased the ROR to 1.48 (CI 1.19-1.83), and three or more infections increased
the ROR to 1.93 (CI 1.48-2.52) (60).
Table 1.1 shows the percent urethral and rectal gonorrhea in MSM from a separate AIDS cohort
study (55). Average age of sexually activity in MSM was 17 in whites and 15 in Hispanic-whites
and African Americans (55). Among whites and Hispanic-whites the rate of individuals who
have a history of sexual partners with AIDS was 20% and 23% respectively and lower in African
Americans who had lifetime history of partners with AIDS of 13% (55). The study also
constructed a multivariate analysis, and odd ratios of HIV seroconversion appear in table 1.2,
white males 26-30 year olds had the highest risk of HIV seroconversion, and those with a history
of rectal gonorrhea an OR of 1.59 (CI 1.28-1.97). While this study did not look at co-infection
status it was able to show that lifetime history of N. gonorrheae infection is associated with an
increase in risk of HIV seroconversion (55).
A study done in 2001 looking at female HIV positive individuals found that 7% of women with
HIV are infected with N. gonorrheae (36). In 2005 Stekler et al. found when looking at multiple
STIs, partner infections, and co-infection states found that among MSM in those with N.
gonorrheae 33% of their partners were also co-infected with N. gonorrheae and 3% were
infected with HIV (41). In heterosexual males that had N. gonorrheae around 28% had
gonorrhea co-infected partners and no HIV infections were detected (41). A separate study found
that in men infected with N. gonorrheae 7.9% were co-infected with HIV (48).
27

A study published in 1997 by Royce et al. that focused on the transmission patterns of HIV
found that as early in the research as 1997 gonorrhea (and chlamydia) were seen to increase the
risk of HIV infection in males from 60% to 340% (44). The study also reported that local
infection which consisted of inflammation or ulceration of the rectum, reproductive track, or
mouth was associated with increased viral load in genital secretions, risk of transmission, as well
as risk of HIV infection in naïve individuals (44). The study also points out that immune
activation in general may increase all three (viral load, transmission, infection) as well although
not at the rate of local infection (44).
A study done by Tzeng et al. investigating US HIV positive military personnel showed that 50%
of HIV positive individuals also had a STD diagnosis reported (17). Tzeng et al. found that
within the HIV positive population who had a history of a STD infection, 32% only had pre-HIV
STD infections, 47% only had STD infections after HIV diagnosis, and 20% had both pre and
post HIV STD infections (17). Among study participants gonorrhea was the most common preHIV infection with 144 individuals having 182 incidents of N. gonorrheae infections (17). N.
gonorrheae incident rate in the 10-year period leading up to diagnosis of HIV were shown to
increase and decrease in the 10-year post diagnosis period (17). A study by Torian et al. that
investigated risk factors associated with HIV positive diagnosis found that in MSM N.
gonorrheae was twice as common in HIV positive males as compared to non-seroconverted
MSM (18). Increased risk of HIV positive seroconversion with MSM who had a diagnosis of N.
gonorrheae as well as that MSM who are infected with N. gonorrheae were less likely to be
tested for HIV than those with other or no reported STDs (18).
A study by Lafferty et al. that looked at N. gonorrheae and HIV and found that being HIV
positive increased risk of gonorrhea infection (OR=2.3) (23). Gonorrhea was higher in MSM
28

(11%) than in heterosexuals (6%) and that all but one of 66 males infected with urethral N.
gonorrheae showed signs of urethritis and 46% of all men with rectal infections presented with
symptoms (23). A 2006 review looking at surveillance reports, published studies, and ad hoc
publications by Dougah et al. found that in England and Wales 32% of N. gonorrheae infected
males in 2004 were HIV positive (43). From 1994-1999 in Denmark the rate of gonorrhea was 6
times higher in HIV positive MSM and that in Paris a clinic reported that from 1999-2001 1/3 of
patients with gonorrhea were co-infected with HIV (43). In 2004 in 9 US cities the reported rates
of gonorrhea in HIV positive MSM was higher than it was in negative or unknown HIV status
MSM (43). An article in 1992 by Wasserheit stated that in HIV/N. gonorrheae co-infection
periods that the presentation of gonorrhea in HIV infected individuals was usually atypical (did
not further define atypical), that there was both an increase in pelvic inflammatory disease as
well as disseminated gonococcal infection, increased resistance to classical antibiotic treatments
resulting in an increase in treatment failures (95).
A 2005 review done by Risbud and the National AIDS Research Institute found that HIV viral
shedding is enhanced during STD/HIV co-infection periods by the damaging epithelial cells and
causing the accumulation of HIV vulnerable cells (47). The article also notes that STDs are the
single major factor in the spread of HIV (47). The review also noted a study done by Wasserheit
et al. that found that while both ulcerative and non-ulcerative increase HIV transmission, nonulcerative STDs infections such as N. gonorrheae and may be due to the higher incidence of
these infections in the public as compared to ulcerative STDs (47).
A study done in Sweden from 1990 to 2004 found that among MSM in the study infected with
gonorrhea, 10% were co-infected with HIV (50). This was a steep decline in the percent of N.
gonorrheae HIV co-infected MSM from the estimated 50% in 1991 (50). In MSM 13% had
29

multiple re-infections of N. gonorrheae with 5% of re-infections occurring in the same year (50).
Of the 10% with co-infections, 25% of those (22 of 88) were concurrently diagnosed (50). While
the percent of co-infections decreased from the start of the study, the rate of concurrent
infections increased from 11% between 1990 and 1997 to 32% from 1998 to 2004 (50). The
study also found that from 1996 to 2004 the rate of gonorrhea in Swedish MSM increased from
2.4 per 100,000 to 6.3 to 100,000 (50).
A study done by Do et al. looking at N. gonorrheae and HIV co-infections found that gonorrhea
incidence rates were higher in younger adults, MSM, blacks, and those that have not progressed
to AIDS yet (51). As age increased there was a significant decrease in risk of N. gonorrheae coinfection status from RR=18.2 in 13-19 year olds to RR=3.7 in 25-29 year olds and 1.9 in 35-44
year olds with those 45 years and older as a reference (51). Use of antiretroviral therapy was not
found to be significant (51). The Reaching for Excellence in Adolescent Care and Health
(REACH) project found that 8.5% of youths were positive for N. gonorrheae infection and while
the data trended that those with HIV had higher rates of infection it was not statistically
significant (53). For males in the REACH project 7.9% of HIV positive individuals were coinfected with N. gonorrheae while 6.9% of HIV negative males were infected with N.
gonorrheae (53).
A Minnesota study by the Minnesota Department of Health found that of the 2,315 HIV positive
individuals in the study 1.3% had at least one reported diagnosis of an STD (92). The study
found that gonorrhea incidence was higher in the HIV infected individuals as compared to the
rest of the MN population (92). Four of the 24 HIV positive individuals (17%) with gonorrhea
had 2 or more gonorrhea infections from 1993 to 1994, all of these were male (92). Median time
to STD infection from the diagnosis of HIV was 3 years (92). This study is a key paper in
30

STD/HIV co-infections because it was one of the first (if not the first) to use surveillance data to
tract STDs in a HIV positive population and shows that state health departments with limited
resources can still conduct meaningful investigations to access the burden of disease in there
population using record linkage techniques between existing databases (92). The paper still
stands as an important piece as we move to the integration of health based systems for disease
tracking, prediction and modeling. The study also points out that an important factor in HIV
prevention is the assessment of sentinel events, such as STD infections, as a high risk behavior
where intervention may be possible (92).
1.4 Biological Plausibility
There are 4 main types of sexually transmitted diseases which include systemic infections
without mucosal disease, genital ulcers, mucosal inflammation (which includes N. gonorrheae),
and those that cause changes in epithelial cells (29). In men, STDs in the inflammation category
cause an increase in HIV viral load in the urethra and in semen and is thought to have a greater
effect on viral load than chlamydia (29). Research has shown that the major target cells of an
early HIV infection are the CD4 T cells (1, 27, 28). During bacterial infections the immune
system releases T cells to help fight off the infection, these T cells can then serve as targets for
HIV to infect and replicate resulting in an increased viral load during bacterial co-infection
periods (24, 27- 29, 31).
A study looking at host response to co-infection of HIV and N. gonorrheae done by Montano et
al. states that during periods of co-infection the risk of transmitting or acquiring HIV is increased
(30). N. gonorrheae infections cause an intense inflammatory response in the host due to the
immune response of a bacterial agent and during co-infection periods more genes activated that

31

are associated with immune response, infection and T-cell stimulation as compared to those with
only HIV (30).
Epidemiologic Synergy
A key paper on epidemiologic synergy in HIV and STDs is the 1992 Wasserheit paper entitled
“Epidemiological Synergy: Interrelationships between Human Immunodeficiency Virus
Infection and Other Sexually Transmitted Diseases”. The main premise of the article was the
investigation of role of STDs in the transmission and progression of HIV as well as the role of
HIV in the natural history of non HIV STDs (95). To paraphrase the article, epidemiological
synergy is when two causes of the disease work seemingly together to increase the virulence of
both (95). Three key relationships were outlined between HIV and non HIV STDs by Wasserheit
(95). The first relationship is that there is a period of increased transmission of HIV during STD
co-infection periods (95). The second theorized relationship is that there is an accelerated
progression of HIV (shortening of time to AIDS or time to death) during periods of STD coinfection (95). The third key relationship when exploring epidemiological synergy is that there is
an alteration in the presentation, diagnosis, and treatment response of the co-infection non HIV
STD in individuals who are HIV positive (95).
Due to the nature of transmission of HIV and other STDs being sexual contact or of intravenous
drug use, STD infection could also serve as a proxy for the measurement of other know risky
behaviors such as use of prostitution or engagement is anonymous sex rather than a definitive
causal link to HIV seroconversion (95). Wasshereit points out that the control of these other
known risk factors for HIV transmission need to be controlled for to positively identify STDs as
a biological risk factor for HIV transmission (95). The author does acknowledge that the

32

measurement of these other risk factors is often imperfect and heavily dependent self-reported
histories and difficult to properly validate (95). A key shift in the article is the acknowledgement
of while STDs that cause genital ulcers are more commonly studied, nonulcerative infections
such as gonorrhea and chlamydia make up a larger burden of disease and if there exists a
synergistic relationship between STDs and HIV that these nonulcerative infections need to be
studied as well (95).
Another key point in the Wasserheit article is that there is a hard to define gap between infection
with HIV and when HIV is diagnosed (95). This is important to note because it has a great effect
on what one would consider the baseline levels of CD4 and viral load to be (95). Those with low
baseline levels of CD4 or high baseline levels of viral load concentrations may be in a state of
elevated HIV progression due to an unaccounted immune response at time of testing (95).
Working from the original STD synergy papers, Fleming and Wasserheit conclude that non HIV
STDs can facilitate the shedding of HIV in the genital tract (increased viral load) which then
promotes the transmission of HIV and the infectivity of the genital secretion (97). The review
also found that in those that are HIV negative, that STDs increase the susceptibility of HIV by
the release of inflammatory cells in the infected area as well as disruption of mucosal barriers via
lesions and ulcers which are infection targets for HIV, also found by Rottingen et al. and Mayaud
and McCormick (38, 40, 97). The review speculates that N. gonorrheae infections most likely
promote HIV transmission through the typical immune response to a bacterial infection in the
genital area, the release of inflammatory immune cells that are targets for viral entry and
reproduction (97). Another finding of the review is that the natural history changes for
nonulcerative STDs such as N. gonorrheae may be less than those of genital ulcerative diseases
(GUD) (97). However, nonulcerative infections may increase the risk in the receptive partner as
33

opposed to both insertive and receptive partners in those with a GUD (97). The three key points
for HIV prevention programs to focus on that the review found were improving access to STD
clinical services, promoting STD related health care behaviors, and establishing a joint HIV/STD
surveillance system (97).
A review of the synergistic relationships between STDs and HIV by Rottingen et al. explored the
possible effects of STDs during possible periods of HIV infection (38). The systemic review also
found that in co-infected individuals the risk of HIV transmission was increased and that
concurrent HIV/STI co-infection periods decrease the time to AIDS positive status (38).
Rottingen et al. also examined the relationship between HIV infection and the epidemiology of
STDs (38). The review found in individuals with HIV the risk for acquiring an STD during an
infectious encounter is increased as compared to those without HIV as well as that in those who
are HIV positive the risk for recurrence from latent infection is greater (38). Mayaud and
McCormick state that HIV can modify the duration, severity, and response to treatment of certain
STDs, but cite that this is most common in viral infections (40).
Biological Markers of HIV-Progression
An editorial review by Bentwich et al. in 2000 states that since all co-infections will cause
immune responses that it is plausible that these co-infection states enhance HIV infection,
increase HIV viral load, and increase the rate of progression of disease in those individuals with
STD/HIV co-infection states (45). Bentwich et al. diagrams that co-infection starts via activation
of cytokines such as tumor necrosis factor-alpha and helper T cells 1 and 2 and activation of
chemokines and related receptors co-infections states increase the susceptibility of HIV infection
in naïve individuals (45). The review also diagrams that by activation of cytokines and

34

chemokines as well as cell anergy (inactivation of a lymphocyte following an antigen encounter
(46)) and apoptosis (programmed cell death) as well as HIV replication all increase the
progression of HIV to AIDS (45). The authors acknowledge that ulcerative infections are
associated with increased risk of HIV infection and increased viral load concentrations in those
infected with HIV for both men and women but state that inflammation may be the main factor
in HIV enhancement in co-infected individuals (45).
A review of co-infection literature by Bafica et al. states that HIVs enhanced viral expression
during co-infection states may be due to the microbial-induced immune response that occurs
during these periods (90). Bafica et al. also states that when the article was written (2004) tolllike receptors (TLR) were the major innate immune response for microbial infections (90). In
most human cells, when stimulated with TLR-2, 4, or 9 that there is an increase in the viral
replication of HIV and cytokine-mediated induction of HIV long terminal repeats [long identical
sequences of DNA found at the retrotransposons or proviral DNA end and used to insert genes
into host (91)] via nuclear factor κB activation occur during co-infection (90).
N. gonorrheae has been shown to attract CD4+ lymphocytes to infective areas disrupting
epithelial and mucosal barriers and increasing possible susceptibility to HIV (24). N. gonorrheae
enhanced HIV infective potential of activated dendritic cells and during co-infective state HIV’s
ability to replicate in dendritic cells is enhanced and is thought to be due to an interaction
between peptidoglycan and TLR-2 (27). Studies also show that N. gonorrheae promotes HIV
infectivity of CD4+ T cells (27). Ding et al. found that along with peptidoglycan (a TLR2
agonist), Pam3CSK4 and Pam3C-Lip (gonococcus derived synthetic lipopeptide) also increase
HIV’s ability to infect CD4+ T cells, whereas TRL4 agonists where not shown to increase
infection potential (28, 32). Interleukin 2 was also found to be a vital component in the TLR235

mediated HIV enhancement (28). N. gonorrheae enhanced the HIV infection of CD4 cells early
infection via TLR-2 activation which not only increased the susceptibility of CD4 infection but
also enhanced the viral load associated with HIV infected CD4 cells (28). It has also been shown
that when monocytes differentiate into macrophages they become more susceptible to HIV
infection (32). Liu et al. also found that lipooligosaccharides (LOS) which are components of the
outer membrane of N. gonorrheae, have the ability to stimulate TLR4 which has been shown to
protect macrophages against HIV infection (32).
A review by Mayer and Venkatesh from 2012 found that individuals with HIV who were also
co-infected with an STD were more likely to spread HIV to their partners than those who were
only HIV positive (101). The review states that the biological plausibility of bacterial STD
enhancement of HIV is largely backed by the fact that during a bacterial co-infection
inflammation causes an increase in HIV expression and that this increase in concentration can be
seen to decrease by treatment of the bacterial STD (101). After treatment of the bacterial STD,
HIV viral levels in the genitals do not return to levels at a pre-bacterial infection state, this is an
important concept because it shows that while treatment has the ability to reduce the sudden
increase in viral load, treatment does not fully negate the damage done during a co-infection
period (101). The review also found that N. gonorrheae enhanced HIV infection by activation of
TLR-2 which can both increase susceptibility of CD4 cells to HIV infection and enhance HIV
replication in those CD4 cells already infected (101). The review also noted that N. gonorrheae
co-infection periods increased the ability of HIV to infect dendritic cells which then present the
virus to other susceptible immune cells (101).
A study by Ghys et al. found that HIV positive women with N. gonorrheae co-infection were
more likely to shed HIV than in women who were not co-infected (98). Treatment of the STD
36

was found to decrease shedding within one week (98). The study also found that in co-infected
women, not only was there an increase in viral shedding, but the viral load was higher making
the risk of transmission greater (98).
A 2004 study by Kaul et al. looking at antibiotic prophylaxis in female HIV naïve Kenyan sex
workers and their risk of HIV infection found that while the rate of HIV between those treated
and those on a placebo did not differ at the end of 2 years, however, the risk of developing HIV
in those with a diagnosed N. gonorrheae infection was increased (103). Prophylactic antibiotics
were given to female sex workers as prevention against bacterial STDs and to reduce excess risk
of HIV infection due to bacterial/HIV interactions. This study is important for multiple reasons.
First this study shows that those infected with N. gonorrheae have a higher risk of developing
HIV then those that are uninfected (103). Second the study shows that prophylactic antibiotic
treatments have the ability to reduce the rate of STDs in endemic areas (103). The final key thing
that the study found was that despite the reduction in STDs the rate of HIV was unaffected (103).
This finding shows that reduction in of the spread of STDs may not be enough and while those
on the treatment have protection from bacterial STDs of their partners, their partners could still
be infected and therefore have higher viral loads thus increasing the chance of HIV infection
(103).
N. gonorrheae has also been shown to induce defensins which have also been show to enhance
HIV (104). In a study by Klotman et al. defensins, antimicrobial peptides thought to play a role
in mucosal host immune responses, were found to promote HIV during viral entry into the cell
(104). Two defensins, 5 and 6, were created during the immune response to N. gonorrheae and
were found to promote HIV infection of immune cells (104). The study also found that in men
with N. gonorrheae infection acute urethritis (inflammation of the urethra) is common, which
37

would lead to the creation of defensins 5 and 6, which in turn would promote HIV infection of
susceptible immune cells, as opposed to women who typically have an asymptomatic rate of 5080% (104).
A study by Ogawa et al. found that Langerhas cells (dendritic cells of the skin and mucosa) are
susceptible to HIV infection after sexual intercourse (106). The study found that antagonists of
TLR 1/2 [Pam3CSK4 and defensins (107)] and TLR 2/6 [Pam2CSK4 (108)] enhanced both
susceptibility and viral replication in monocytes created from Langerhas cells (LC) (106). The
study also looked at other TLR antagonists, but did not find any associations (106). A theory
presented by the article to explain the synergistic relationship between N. gonorrheae and HIV is
that gram positive bacteria such as N. gonorrheae may enhance the sexual transmission of HIV
by ligation of TLR2 (106). The study also notes that during N. gonorrheae infections LC are
commonly found in the genital mucosal epithelium and move to the lymph nodes during
infections (106).
1.5 Rationale
As of the 2010, Census Michigan has the 8th largest population in the US. It is estimated that in
2012 there were 20,600 persons living with HIV in the state of Michigan, with 63% of HIV cases
residing in the Detroit Metro Area. Black males make up the largest group of those living with
HIV in the state of Michigan. In 2012, there were 12,770 reported cases of N. gonorrheae
infections in the state of Michigan, with the highest rate of infected being persons in the age
range of 20-24 years old. This age group represents 6.7% of Michigan population but accounts
for 1/3 of all gonorrheal infections in the state of Michigan. The rates of infection with N.
gonorrheae were higher in African Americans than in Whites (655 per 100,000 vs 68 per

38

100,000 respectively as of July 2014) and 43% of all cases were males. In 2011, 259 of the
13,000 reported cases were co-infected with HIV.
Evidence suggests that there is a synergistic relationship between HIV-1 and infection with N.
gonorrheae resulting in non-classical manifestations of these diseases in co-infected individuals
(17, 18, 23, 24, 31, 36, 40, 41, 44, 48, 51, 55, 58, 60, 92). This study aims to explore the
possibility of a synergistic relationship between gonorrhea and HIV via reduction in age at HIV
diagnosis in those with a history of gonorrhea, reduction in age at AIDS and death in those with
a history of gonorrhea, the geospatial distribution of co-infections in Michigan, the change in
CD4 cell count (spike during infection then fall to levels lower than pre-gonorrheal state) and the
increase in HIV viral load. N. gonorrheae infection will be used as a proxy for risky sexual
behavior as data on other STDs, bacterial and viral infections during this time period were
unavailable. Attention has focused primarily on co-infection with ulcerating STI’s in recent
years. However, non-ulcerating STI’s such as N. gonorrheae account for a significant coinfection burden in the HIV/AIDS community and there is a need for research to explore the
dynamic relationship between these co-infections. Prevention efforts should be aimed at those
with the highest risk for infection or death to reduce the burden of disease. Among its objectives,
this study aims to identify the trends associated with time to infection, survival rates, and
geographic clustering of the co-infection paradigm between HIV and N. gonorrheae.
A retrospective cohort of HIV positive men in Michigan from 2005 to 2011 with N.
gonorrheae was created by the Michigan Department of Community Health (MDCH). HIV
infected men without any reported cases of N. gonorrheae between 2005 and 2011 were selected
as the control group by MDCH. Males with HIV that had previously been infected with N.
gonorrheae prior to 2005 were considered ineligible.
39

There is a need to explore the co-infection relationship between HIV and N. gonorrheae and
little is known about the impact of N. gonorrheae in facilitating acquisition of HIV or about the
change in the survival rates in those with histories of N. gonorrheae infections. With limited
funding and resources available to control the spread of STD’s in sexual networks the need for
well-structured interventions are required to make the greatest impact. The geographical analysis
of STD co-infections will help to identify the spatial patterns of at risk populations where
resources should be spent to have the highest chance of success in disrupting the cycle of
transmission.
1.6 Objectives
Objective 1: To test whether infection with N. gonorrheae increases the hazard (reduces the age
at diagnosis) of developing HIV in Michigan males, with N. gonorrheae infection being a proxy
for engagement in risky sexual practices.
Objective 1.1: Using time to event analysis, test whether individuals with a history of N.
gonorrheae infections are younger at HIV diagnosis than those with no history of N.
gonorrheae infections in Michigan males.
Objective 1.2: Using time to event analysis test whether multiple N. gonorrheae
infections increase the hazard (younger at HIV diagnosis) of HIV infection in Michigan
males.

Objective 1.3: Using time to event analysis test whether prior multiple N. gonorrheae
infections increase the hazard (younger at HIV diagnosis) of HIV infection in Michigan
males.
40

Gonorrhea is a treatable and easily identified STD, if multiple infections reduce the time
to HIV+ status then health departments may want to aim intervention resources to
educate and reduce STD burden in those individuals in high risk groups. If multiple
infections reduce time to HIV+ then intervention should be aimed at the minimum
number of infections that will increase the lag time to HIV+ status. From the public
health point of view, gonorrhea pre-HIV+ status may indicate possible predictive social
behaviors that may be intervened to prevent progression of individual to infection with
HIV. These behaviors include non-routine screening of STDs, condom use, multiple
partners, or increased frequency of anonymous sexual partners. Infection with N.
gonorrheae may act as an easily identifiable flag for risky sexual behavior that would
reduce the time to HIV positive status in those that have a history of N. gonorrheae
infections.

Objective 2: To test if N. gonorrheae co-infection increases the hazard (younger at event) of an
HIV progression event (death or AIDS) due to possible synergistic effects related to co-infection
status.

Objective 2.1: Using survival analysis, test whether age at death in HIV infected men that
have N. gonorrheae infection occurs at a younger age than in those without a recorded
incidence of N. gonorrheae in the state of Michigan.

41

Objective 2.2: Using survival analysis, test whether time from HIV diagnosis to death in
HIV infected men that have N. gonorrheae infection is reduced as compared to those
without a recorded incidence of N. gonorrheae infection in the State of Michigan.

Objective 2.3: Using time to event analysis test whether history of gonorrhea increases
the hazard (younger at AIDS) of AIDS progression in HIV infected males in Michigan.

Objective 2.4: Using time to event analysis, test whether the time from HIV diagnosis to
AIDS in those that have N. gonorrheae infections is reduced as compared to those with
no reported N. gonorrheae infections in HIV infected males in Michigan.

Objective 2.5: Using time to event analysis test whether co-infection status increases the
hazard (younger at AIDS) of AIDS progression in those with multiple recorded N.
gonorrheae infections in HIV infected males in Michigan.

Objective 2.6: Using time to event analysis, test whether the time from HIV diagnosis to
AIDS is reduced in those with multiple recorded N. gonorrheae infections in HIV
infected males in Michigan.

Objective 2.7: Test whether hazard of AIDS progression is increased (younger at AIDS)
in co-infected individuals with post-HIV N. gonorrheae infections in HIV infected
Michigan males.

42

Objective 2.8: Test whether time to AIDS status from HIV diagnosis is reduced in coinfected individuals with post-HIV N. gonorrheae infections in HIV infected Michigan
males.

Objective 2.9: Test whether the hazard of AIDS progression is increased (younger at
AIDS) in co-infected individuals with multiple post-HIV N. gonorrheae infections in
HIV infected Michigan males.

Objective 2.10: Test whether time to AIDS status from HIV diagnosis is reduced in coinfected individuals with multiple post-HIV N. gonorrheae infections in HIV infected
Michigan males.

The public health importance of this objective is to test the hypothesized synergistic
relationships between N. gonorrheae and HIV resulting in a more pathological state in
the individual and reducing their life expectancy through progression of the disease, to
either death or AIDS positive status. Public health action can then be aimed to educate
and inform individuals about the dangers of co-infection states and the need to engage in
safe sex practices even after HIV diagnosis and the need for routine screening of STDs.

Objective 3: To identify locations in Michigan where N. gonorrheae and HIV co-infections are
elevated in order to target future STD intervention.

43

Objective 3.1: Explore the spatial patterns and clusters of HIV infection in males in
Michigan.
Objective 3.2: Explore the spatial patterns and clusters of N. gonorrheae infection in HIV
infected males in Michigan.
Objective 3.3: Explore the spatial patterns and clusters of multiple N. gonorrheae
infections in males affected with HIV
Public health importance of this aim to assess if there are geographical boundaries that
constitute sexual networks that are statistically more prone to N. gonorrheae HIV coinfections in HIV positive males in the state of Michigan from 2005 to 2010. The
identification of these sexual networks is crucial to assisting public health agencies to
better focus their resources in reducing the burden of disease in these areas. Clusters can
be identified as focal points in the effort to reduce the burden of disease due to coinfection in these areas. Public health education can focus on informing HIV+ individuals
in these areas about the importance of safe sex to reduce the risk of N. gonorrheae coinfection. Areas that have higher occurrences of multiple infections may be of greater
public health concern than other areas, and funding for education and prevention
programs may be put to better use by being tailored for this population as opposed to a
wider sweeping public health message.
Objective 4: To test whether biological markers of HIV progression are altered during coinfection states in HIV infected Michigan males.
Objective 4.1: Characterize the change in HIV viral levels before, during, and after N.
gonorrheae co-infection.
44

Objective 4.2: Characterize the change in CD4 cell counts before, during, and after N.
gonorrheae co-infection.
Objective 4.3: Explore the spatial patterns and clusters of N. gonorrheae in those that are
infected with HIV in Michigan from 2011-2013.
Change in laboratory markers such as CD4 cell count and HIV viral load are important
indicators of HIV disease progression and due to the synergistic relationship of N.
gonorrheae and HIV may be altered during and after co-infection periods. These changes
in the natural history of HIV are important markers in the progression to AIDS or death
in the individual. Investigation of these changes may lead to better understanding of the
burden that this places on the host during co-infection periods and allow for departments
of public health to educate HIV positive individuals on the risk of STDs and the burden
to their already overtaxed immune system. This data set also allows for examination of
zip code level GIS mapping and the ability to explore ART and its effects on disease
progression during co-infection periods.

45

APPENDIX

46

Table 1.1: Percent Urethral and Rectal Gonorrhea in MSM, AIDS Cohort Study (55)
Group

Urethral Gonorrhea

Rectal Gonorrhea

White

47%

27%

Hispanic-Whites

57%

34%

African Americans

66%

27%

Table 1.2: HIV Seroconversion Odd Ratios in MSM, AIDS Cohort Study (55)
Group

Urethral Gonorrhea

Rectal Gonorrhea

White

1.83

1.59

(95% CI 1.51-2.23)

(95% CI 1.28-1.97)

Hispanic-Whites

Not Significant

Not Significant

African Americans

Not Significant

2.44
(95% CI 1.07-5.59)

47

Chapter 2.
Interaction of N. gonorrheae and HIV infections in Michigan Males
2.1 Abstract
Introduction
Evidences suggest a synergistic relationship between HIV and N. gonorrheae infections resulting
in abnormal epidemiological patterns of these diseases in infected individuals. The goal of this
study is to investigate the role of age at HIV diagnosis in individuals with and without N.
gonorrheae co-infection.
Methods
A retrospective cohort study of 19,647 HIV infected males in the state of Michigan from 20052011 was designed in collaboration with Michigan Department of Community Health. Cases of
N. gonorrheae HIV co-infections consisted of males with HIV infection in Michigan with a
reported infection with N. gonorrheae between 2005 and 2011 and were diagnosed with HIV by
2011. Controls were Michigan males with no history of N. gonorrheae infections who were
diagnosed with HIV by 2011. Linear regression and Cox proportional hazard models were
constructed to assess N. gonorrheae and age at HIV diagnosis.
Results
Mean age at HIV diagnosis in males with no history of gonorrhea was 36.8 years old (95% CI
36.6-36.9) whereas there was decrease in mean age at HIV diagnosis of 8.4 years (95% CI 7.89.1) in those with a history of gonorrhea. A dose response was seen in the multiple N.
gonorrheae infections, with one reported gonorrheal infection event having a mean reduction in
48

age at HIV diagnosis of 7.3 years (95% CI 6.5-8.2), mean reduction of 9.3 years (95% CI 7.810.9) in those with two gonorrhea events, a mean reduction of 10.1 years (95% CI 7.8-12.4) in
those with three reported gonorrhea events, a mean reduction of 11.8 years (95% CI 8.4-15.2) in
those with four N. gonorrheae infections, and a mean reduction of 13.9 years (95% CI 10.9-16.8)
in those with five or more N. gonorrheae infections as compared to those with no history of
gonorrhea. Those with only one previous N. gonorrheae infection prior to HIV diagnosis had a
mean reduction in age at HIV diagnosis of 7.3 years (95% CI 5.7-9.0) as compared to those with
no history of gonorrhea, and a mean reduction of 10.2 years (95% CI 7.7-12.7) in those with 2 or
more pre-HIV N. gonorrheae infections. Resulting median age of HIV diagnosis was 36 for no
history of gonorrhea and decreased to 26 (HRadj=2.08 95% CI 1.95-2.23) in individuals with a
reported history of gonorrhea from 2005 to 2011 in the state of Michigan. A dose response was
also observed with hazard increasing as number of reported N. gonorrheae infections increased.
The study also found that in those with pre-HIV gonorrhea the median age of HIV diagnosis also
decreased. In males with one reported gonorrheal event, the median age of diagnosis was 26
(HRadj=1.85 95%CI 1.57-2.17) and in those with two or more pre-HIV N. gonorrheae
infections, the median age at HIV diagnosis was 23.5 (HRadj=2.98 95%CI 2.33-3.80).
Conclusion
Michigan males, at HIV diagnosis who had history of N. gonorrheae infections were younger
than males diagnosed with HIV but without history of N. gonorrheae infections. Males with
multiple N. gonorrheae infections had an increased hazard (lower median age) at HIV diagnosis
than those without reported histories of N. gonorrheae infections. Recurrent incidents of pre-HIV
N. gonorrheae infections were associated with a younger age at HIV diagnosis among males in
Michigan. HIV prevention efforts need to consider intervention steps aimed at first occurrence of
49

an STD, in this case N. gonorrheae, as an important event in the reduction of HIV infection
among populations at risk for the disease.

50

2.2 Introduction
Evidences suggest that there is a synergistic relationship between HIV and N. gonorrheae
infections resulting in non-classical manifestations of these diseases in co-infected individuals.
Attention has focused primarily on co-infection with ulcerating STI’s in recent years. However,
non-ulcerating STI’s such as N. gonorrheae infections account for a significant burden of coinfection in the HIV/AIDS community and therefore the investigation of the dynamic
relationship between these co-infections is warranted.
Despite multiple recent outbreaks of N. gonorrheae as well as the emergence of antibiotic
resistant strains of N. gonorrheae and the fact that N. gonorrheae is one of the most reported
diseases in the United States, the relationship between HIV and gonorrhea has not been fully
delineated (15). In 2011, the WHO estimated that roughly 88 million of the estimated 448
million curable STD infections reported that year were due to N. gonorrheae (12). In the United
States in 2012, there were over 334,000 N. gonorrheae cases resulting in an increased rate of
4.1% from the previous year (13).
As of the 2010 Census Michigan has the 8th largest population in the US. In 2012, there were
12,770 reported cases of N. gonorrheae infections in the state of Michigan, with the highest rate
of infected being persons in the age range of 20-24 years old. This age group represents 6.7% of
Michigan population but accounts for 1/3 of all N. gonorrheae infections in the state of
Michigan. The rates of infection with N. gonorrheae were higher in African Americans than in
Whites, and 43% of all cases were males. In 2011, 259 of the 13,000 reported cases were coinfected with HIV.

51

Based on the January 2014 Annual HIV Surveillance Analysis estimates, there were
approximately 19,800 individuals in the state of Michigan who were HIV positive. Of this group,
an estimated 15,440 (78%) were males (10). Of the 15,440 males with HIV, men who have sex
with men accounted for 55% of prevalent HIV cases between the combined risk groups of MSM
and injection drug user (IDU)/MSM in the state of Michigan (10). Heterosexual contacts
accounted for 19% of those with HIV infection and 17% had an undetermined source of
transmission (10). The most frequent age group for HIV diagnosis was 30-39 year old (33%)
with the combined age group of 20-29 at 32% (10). The majority of HIV diagnosis was among
males in the Detroit Metropolitan Area (64%) despite the fact that the population of Detroit
Metropolitan accounts only for 43% of the total Michigan population (10). Males made up the
majority of HIV infected individuals with a reported 12,510 cases (10). Within this group 71%
are MSM (combined risk groups of MSM and MSM/IDU), 5% are heterosexuals, and 17% have
an undetermined mode of transmission (10).
Wasserhiet et al. reported that epidemiological synergy of HIV and STDs occurs “If coinfection
with HIV prolongs or augments the infectiousness of individuals with STDs, and if the same
STDs facilitate transmission of HIV, these infections may greatly amplify one another” (95).
Reported research indicated that the major target cells of an early HIV infection are the CD4 T
cells (1, 27, 28). During bacterial infections the immune system releases T cells to help fight off
the infection, these T cells can then serve as targets for HIV to infect and replicate resulting in an
increased viral load during bacterial co-infection periods (24, 27- 29, 31).
A study done by Huhn et al. using Chicago and Illinois STD/HIV surveillance data found that of
the 43,517 patients seen in 2002 around 13% received a positive HIV test (96). Syphilis had the
highest unadjusted OR=11.0 (95%CI 7.7-15.8) for co-STD infection in newly diagnosed HIV
52

individuals, with gonorrhea coming in second with an OR of 2.2 (95%CI 1.4-3.3) (96). In 1987,
a multicenter AIDS cohort study of four sites (Chicago, Los Angeles, Pittsburg,
Baltimore/Washington DC) found that lifetime history of gonorrhea was between 40-67%
depending on site with Los Angeles and Chicago having the highest rates (58). Another early
study of HIV and STDs looking at the AIDS cohort sites found that in MSM (defined as
homosexual and bisexual males) 47% of whites had a history of urethral gonorrhea and 27% had
a history of rectal gonorrhea (55). Among Hispanic-whites, 57% had a history of urethral
gonorrhea and 34% had a history of rectal gonorrhea and among African Americans the rate of
urethral gonorrhea was 66% and the rate of rectal gonorrhea 27% (55). The study also found that
the average age of sexual activity in MSM was 17 in whites and 15 in Hispanic-whites and
African Americans (55). The study also constructed a multivariate analysis and found that 26-30
year old white males had the highest risk of HIV seroconversion and those with a history of
rectal gonorrhea had an OR of 1.59 (95%CI 1.28-1.97). Those with a history of urethral
gonorrhea had an OR of 1.83 (95%CI 1.51-2.23) (55). Among Hispanic-whites, gonorrhea was
not found to be significantly associated with HIV seroconversion (55). In African Americans,
history of rectal gonorrhea was associated with an OR of 2.44 (95%CI 1.07-5.59) and urethral
gonorrhea was not found to be significant (55). While this study did not look at co-infection
status, it was able to show that lifetime history of N. gonorrheae infection is associated with an
increased risk for HIV seroconversion (55).
A separate study using the multicenter AIDS cohort study published in 1987 found that N.
gonorrheae infection increased the risk of HIV infection in males (60). The Chmiel et al. study
found that risk of HIV infection increases with the frequency of rectal N. gonorrheae infection
(60). The multivariate model showed that the ROR for 1-3 infections was 1.65 (1.39-1.96) and
53

the ROR for 4 or more infections was 3.89 (2.45-6.20) (60). The model also looked at the
grouping of lifetime infection with other STDs which included urethral gonorrhea, syphilis,
urethritis, and genital herpes and found an increasing dose response in HIV seroconversion that
increased with re-occurrence of STD infection (60).
A study by Royce et al. found that gonorrheal and chlamydial infections increased the risk of
HIV infection in males by 60-340% (44). Increased viral load and increased risk of HIV
infection in areas where inflammation due to immune response occurred was also seen (17, 30,
38, 44, 97, 106). A study looking at HIV positive US soldiers found that N. gonorrheae infection
the most common among those with a history of STD infection prior to HIV seroconversion,
with some individuals having multiple pre-HIV infections (17).
This study aims to test the hypothesis that N. gonorrheae increases the hazard of HIV
seroconversion in HIV naïve individuals. Risk of seroconversion is measured primarily in this
study by age at HIV diagnosis since there is no way to tell the actual date of HIV exposure. This
study also uses N. gonorrheae as a proxy for risky sexual behavior. The first objective of this
study is to test if those with a history of N. gonorrheae (both pre and post HIV diagnosis)
between 2005 and 2011 have an increased hazard (younger age at HIV diagnosis) as compared to
other HIV infected Michigan males with no reported N. gonorrheae infections. The second
objective of this study is to assess the dose response associated with multiple reported N.
gonorrheae infections with the increased hazard in Michigan males. The third and final objective
of the study is to assess the relationship between N. gonorrheae and HIV when only looking at
pre-HIV N. gonorrheae infections in HIV positive Michigan males.
2.3 Methods

54

A retrospective cohort of HIV positive males in Michigan were selected as the study population.
Cases of N. gonorrheae-HIV co-infection were defined as HIV positive males with a history of
reported N. gonorrheae between January 1st 2005 and December 31st 2011. HIV infected males
with N. gonorrheae prior to 2005 were excluded from the study population. All HIV positive
males without a history of N. gonorrheae during the study time period that were alive as of
January 1st 2005 were included in the study as controls. In collaboration with Michigan
Department of Community Health (MDCH) a linkage structure between HIV and N. gonorrheae
data was created for this study. Records with missing first name, last name, or date of birth were
excluded from the study population prior to the matching of the data. A total of 19,647 HIV
infected males in the state of Michigan were included in this study. Ethical approval was
obtained from the Institutional Review Board (IRB) at Michigan State University (MSU) and
was reciprocal with the MDCH IRB. To comply with patient confidentiality and privacy a
professional employee of HIV Surveillance and Body Art Unit HIV/STD/VH/TB Epidemiology
handled the creation of the record linkage structure between HIV datasets and N. gonorrheae
data sets. The match was conducted between the two registries using LinkPlus. All data used in
this study were de-identified according to HIPAA guidelines on public health information. Study
staff also did not contact or attempt to contact any participants of the study.
Table 2.1 shows the demographic distributions of study participant. HIV cases were defined
according to the CDC case definition of HIV and AIDS defined according to CDC 1993
guidelines. HIV infection was defined as HIV-NA=HIV infection, non-stage 3 and AIDS=stage
3 HIV infection. Acute and Latent, stage 1 and 2, differentiated in this study. Date of birth, date
of death, N. gonorrheae infection, as well as HIV and AIDS diagnosis were all limited to year of
the event. Vital status was coded as 1=alive, 2=deceased, 9=unknown. Ethnicity was classified
55

into unique non-over lapping groups, 1.White= White non-Hispanic, 2.Black= Black nonHispanic, 3.Hispanic=Hispanic of any race, 4.Asian/HI/PI=Asian/Hawaiian, Pacific Islander
non-Hispanic, 5.Am In /Ak Nat=American Indian, Alaskan Native, non-Hispanic, or
9.Multi/Unk/Other=Multirace, Unknown, Other, non-Hispanic. Due to low rates of
Asian/Hawaiian, Pacific Islanders (N=0) and American Indian, Alaskan Native (N=1) in the N.
gonorrheae group these racial groups were included in the 9.Multi/Unk/Other grouping to avoid
analytical complications. Marital status while largely missing was categorized as A=Married and
Separated, S=Single and never married, M=Married, N=Not otherwise specified, Missing or
U=Unknown, W=Widowed; Risk was condensed and categorized as A.MSM=Male-Male Sex,
B.IDU=Injecting Drug Use, C.MSM/IDU=Male-Male Sex and Injecting Drug Use, D. Blood
Recipient=Received Blood, E1.(Male HCFR)=Male who had sex with a female at risk for HIV,
F. Perinatal=Child exposed by mother (further condensed to risk group G.Unk:Other due to low
response, N=3), G.Unk:Other=Unknown/Other. Individual risk categories were also coded and
were analyzed as “Yes” and “Not Yes” to handle missing data as suggested by MDCH. In those
with perinatal risks mother’s HIV status was also coded, although largely missing, as 1=Refused
HIV Testing, 2=Known Uninfected after Birth, 3=Know HIV+ before pregnancy, 4=Known
HIV+ during pregnancy, 5=Known HIV+ at time of delivery, 6=Known HIV+ sometime before
birth, 7=Known HIV+ sometime after birth, 8=HIV+ with time unknown, 9=Unknown and is
mainly used for individuals with risk group F. Perinatal where they were infected via the mother.
History of gonorrhea was coded by MDCH as match=’N’ for no history of gonorrhea and
match=’Y’ for reported history of gonorrhea. The history of gonorrhea model was run using
three different strategies to assess possible co-variates included in the data. The initial model
used the MDCH defined condensed risk groups and ethnicity. Because MSM was divided into
56

MSM and MSM/IDU in the MDCH condensed risk group schema, a second model was ran to
assess the variable sex with male. This model include history of gonorrhea, ethnicity, sex with
male, sex with female, idu, received clotting factor, received a transfusion, received a transplant,
heterosexual sex with an IDU, heterosexual sex with a hemophiliac, heterosexual sex with
transfusion recipient, heterosexual sex with transplant recipient, heterosexual sex with person
living with HIV, and worked in health care setting. A separate final model was ran that created
“presumed” sexual orientation groups such as presumed homosexual (yes to sex with male, no to
sex with female) presumed bisexual (yes to both males and females). This model also included
history of gonorrhea, ethnicity, idu, received clotting factor, received a transfusion, received a
transplant, heterosexual sex with an IDU, heterosexual sex with a hemophiliac, heterosexual sex
with transfusion recipient, heterosexual sex with transplant recipient, heterosexual sex with
person living with HIV, and worked in health care setting.
Number of total gonorrheal events was constructed by count number of years of reported N.
gonorrheae infections an individual had. The cut point of 5 or more was selected because of the
large drop off in individuals with greater than 5 recorded N. gonorrheae infections between 2005
and 2011. Kaplan-Meier curves and a cox regression model were analyzed to determine
reduction in age at HIV diagnosis in those with multiple N. gonorrheae infections. This model
included number of gonorrheal infections, ethnicity, and MDCH condensed age groups.
Alternative explorations of sexual strategies were not preformed after investigation returned no
significant improvement in the history of gonorrhea model.
The variable of pre-HIV gonorrhea was constructed by taking the year of HIV diagnosis and
subtracting the year of N. gonorrheae infection for each individual’s infections. Those that had a
result greater than 0 (0 was not included because it could not be determined if gonorrhea was
57

before, concurrent, or after HIV diagnosis). These were then counted to see how many pre-HIV
infections individuals had. The cut point was set at two or more because very few individuals
had greater than 2 pre-HIV gonorrhea events.
Statistical analysis of the data was done using SAS 9.3 (SAS Institute, Inc. Cary, North
Carolina). Descriptive statistics were preformed to assess distribution of demographic and risk
characteristics. Primary analysis was done using general linear model to investigate mean age at
HIV diagnosis. Time to event analysis was used to investigate change in age at HIV diagnosis
and Kaplan-Meier curves were created to explore unadjusted trends. Event was defined as HIV
diagnosis and time was defined as age of HIV diagnosis. Adjusted and unadjusted models were
ran for each of the three study objectives.
2.4 Results
History of gonorrhea
Between 2005 and 2011 there were 903 HIV infected males in Michigan with a reported
infections with N. gonorrheae. Demographic information can be seen in Table 2.1. Of the 903
individuals with gonorrhea, 216 (23.93%) had at least one N. gonorrheae infection prior to their
diagnosis with HIV. In HIV infected males with a history of gonorrhea, 117 (19.60%) were
White, 674 (74.64%) were Black, 23 (2.55%) were Hispanic, 1 (0.11%) was American Indian,
and 28 (3.10%) were reported as Multiracial/Unknown/Other. The majority of N. gonorrheae
cases did not report marital status 889 (98.45%).
In the distribution of MDCH defined risk groups for HIV infected males with a history of N.
gonorrheae infection, 697 (77.19%) of males were MSM, 21 (2.33%) were IDU, 21 (2.33%)
were MSM/IDU, 20 (2.21%) of HIV infected males reported having sex with a female at risk for
58

HIV, and 144 (15.95%) had other or unknown risks. Of the 903 HIV positive males that had a
reported history of N. gonorrheae infection 390 (43.19%) progressed to AIDS/Stage 3 HIV
positive status and 20 (2.21%) died. Table 2.2 shows the distribution of non-condensed MDCH
risk groups.
The mean age at HIV diagnosis in those with no history of gonorrhea was 36.8 years old. Those
with any reported history of N. gonorrheae infection were 8.4 years younger HIV diagnosis
resulting in a mean age of HIV diagnosis of 28.3 years. Three multivariate models were also
constructed. Table 2.3 shows the results from the unadjusted and adjusted models. The first
model adjusted for ethnicity and MDCH condensed risk groups. History of gonorrhea infection
was associated with reduction of age at HIV diagnosis by 7.2 years in the adjusted model. As
compared to Whites, Blacks had a reduction in age at HIV diagnosis of 2.3 years, Hispanics by
2.6 years, and those that were grouped in the Multiracial/Unknown/Other saw a reduction of 2.6
years. As compared to MSM condensed risk group, all other risk groups, aside from Blood
Recipients were older in age at HIV diagnosis. Blood Recipients had a reduction of 2.2 years as
compared to MSMs. Those that fell into the IDU condensed risk group increased age at HIV
diagnosis by 6.6 years, those that fell into the MSM/IDU condensed risk group had increased age
of 0.9 years, Male who had sex with a female at risk for HIV had an increased age at HIV
diagnosis of 3.6 years, and those that fell into the unknown/Other condensed risk group had an
increased age at HIV diagnosis of 5.3 years as compared to MSMs.
The second model investigated a different risk group structure than the MDCH condensed risk
groups. In the adjusted model, at HIV diagnosis the age of individuals with history of gonorrhea
infection was found to be 7.2 years younger than the group without gonorrheal infection. Those
who were Black were 2.1 years younger at HIV diagnosis compared to whites, Hispanics were
59

younger by 2.4 years, and those that were grouped in the Multiracial/Unknown/Other were 2.5
years younger when compared to whites. Individuals with male partners were 4.5 years younger
at HIV diagnosis as compared to those who did not have male sex partners. Those that had
female sex partners were 0.9 years older at HIV diagnosis as compared to those with who did not
have female partners. Those who were identified as IDU were 1.5 years older at HIV diagnosis
as compared to the non-IDU group. Those that had sex with a HIV infected female were 2.0
years younger at HIV diagnosis as compared to those who did not have sex with a HIV positive
female.
The third model investigated a second sexual orientation structure where presumed sexual
orientation was constructed via sex partner data. Orientation groups included presumed
homosexual, presumed heterosexual (reference group), presumed bisexual, and unknown. The
model also included IDU and sex with HIV infected female. History of gonorrhea infection was
found to reduce age at HIV diagnosis by 7.1 years in the adjusted model. As compared to
Whites, those who were Black had a reduction in age at HIV diagnosis of 2.1 years, Hispanics by
2.4 years, and those that were grouped in the Multiracial/Unknown/Other saw a reduction of 2.5
years. Presumed bisexuals had a reduction of 4.0 years as compared to presumed heterosexuals,
presumed homosexuals had a reduction 5.4 years, and those that had unknown orientation had a
reduction of 0.02 years.
Table 2.4 shows the median age at HIV diagnosis as well as the unadjusted hazard ratio for
history of N. gonorrheae infection. In HIV positive males with a history of gonorrhea (any
reported case of gonorrhea from 2005 to 2011 regardless of HIV status at infection time) had a
median age of HIV diagnosis of 26. In HIV positive males without a history of reported
gonorrhea the median age at HIV diagnosis was 36. Figure 2.1 shows the Kaplan-Meier
60

unadjusted curves for HIV infected males with a history of N. gonorrheae infection and those
without a history of infection. There was an increase in hazard, hazard ratio (HR) 2.24 (95%CI
2.09-2.39), in those with a reported history of gonorrhea between 2005 and 2011. This increase
in hazard equates to a reduction in the median time to HIV diagnosis in these individuals (26
years of age at diagnosis for those with gonorrhea compared to 36 years of age in those with no
history of gonorrhea).
Multivariate models were also analyzed for history of N. gonorrheae infection and age at HIV
diagnosis. The first model used the MDCH condensed risk groups. No individual fell into more
than one risk group and the final model included history of gonorrhea, ethnicity and condensed
risk group. The adjusted hazard ratio (HRadj) for history of gonorrhea was 2.08 (95%CI 1.952.23). As compared to whites all other reported racial groups had an increase adjusted hazard as
seen in Table 2.5. An increase in age of HIV diagnosis (decrease in hazard) was seen in MDCH
condensed risk groups of IDU, Blood Recipient, Male who had sex with a female at risk for HIV,
and Other/Unknown as compared to the risk group MSM. MSM/IDU was not observed to have a
significant reduction in hazard (reduced age at HIV diagnosis) HRadj=0.95 (95%CI 0.90-1.01 pvalue=0.08). Two other multivariate models were also tested to look at the effect of reported
sexual orientation. The first of these models still consists of history of N. gonorrheae infection as
well as ethnicity, but the risk groups were not modeled in a condensed version and individuals
were able to fall into multiple groups. Variables introduced in this section were sex with males,
sex with females, IDU, and heterosexual intercourse with someone with HIV. The main affect,
history of N. gonorrheae infection remained the same HRadj=2.08 (95%CI 1.95-2.23). Ethnicity
was also found to be significant with all groups having elevated hazard as compared to whites.
Sex with males was found to increase hazard of HIV infection as well as heterosexual
61

intercourse with an HIV infected partner. Both IDU and sex with females were found to be
protective (increased age of HIV diagnosis). Sex with males and sex with females are not
mutually exclusive groups; an individual can belong to either or both. The final multivariate
model considered sorted sexual orientation to Presumed Heterosexual, Presumed Homosexual,
Presumed Bisexual and Unknown. This analysis also contained history of N. gonorrheae
infection, ethnicity, IDU, and heterosexual intercourse with HIV infected individual. History of
N. gonorrheae was again found to be significant, HRadj=2.08 (95%CI 1.94-2.22) and ethnicity
was also found to be significant. Presumed Homosexual was found increase hazard, as well as
Presumed Bisexual as compared to Presumed Heterosexual males. Unknown status was found to
be protective. Since HRadj or mean reduction of age at HIV diagnosis did not change with the
reformulation of risk groups, the MDCH condensed risk group structure was deemed to be the
most useful model for determining possible co-variates because it allowed for the most
parsimonious model.
History of multiple gonorrhea episodes
Mean age at HIV diagnosis in those with no history of gonorrhea was 36.8 years. In those with
only one N. gonorrheae infection mean age at HIV diagnosis was reduced by 7.3 years, those
with two N. gonorrheae infections had a reduction in age at HIV diagnosis of 9.3 years, those
with three N. gonorrheae infections had a reduction in age at HIV diagnosis of 10.1 years, in
those with four N. gonorrheae infections reduction in age at HIV diagnosis was 11.8 years, and
in those with five or more N. gonorrheae infections reduction in age at HIV diagnosis was 13.9
years. A multivariate model was also constructed that adjusted for race and MDCH condensed
risk group. Results for unadjusted and adjusted models are shown in table 2.6. In those with only
one N. gonorrheae infection mean age at HIV diagnosis was reduced by 5.1 years, those with
62

two N. gonorrheae infections had a reduction in age at HIV diagnosis of 10.1 years, those with
three N. gonorrheae infections had a reduction in age at HIV diagnosis of 6.2 years, in those with
four N. gonorrheae infections reduction in age at HIV diagnosis was 10.0 years, and in those
with five or more N. gonorrheae infections reduction in age at HIV diagnosis was 10.9 years. As
compared to Whites, those who were Black had a reduction in age at HIV diagnosis of 2.2 years,
Hispanics by 2.6 years, and those that were grouped in the Multiracial/Unknown/Other saw a
reduction of 2.6 years. As compared to MSM condensed risk group all other risk groups aside
from Blood Recipient was an increase of age at HIV diagnosis. Blood Recipients was a reduction
of 2.2 years as compared to MSMs. Those that fell into the IDU condensed risk group increased
age at HIV diagnosis by 6.6 years, those that fell into the MSM/IDU condensed risk group
increased age by 0.9 years, Male who had sex with a female at risk for HIV had an increase age
at HIV diagnosis of 3.6 years, and those that fell into the unknown/Other condensed risk group
had an increase in age at HIV diagnosis of 5.3 years as compared to MSMs.
In those with multiple reported N. gonorrheae infections between 2005 and 2011 a dose response
with regard to decrease of age at HIV diagnosis was observed. Figure 2.2 shows the KaplanMeier curve for individuals with multiple gonorrhea events as compared to those HIV positive
males with no reported events. Table 2.7 contains the unadjusted and adjusted HR for multiple N.
gonorrheae infections and HIV hazard. A logrank p-value of <0.0001 was observed indicating
that not only was the reduction in median age of HIV diagnosis significant, the order, 1 vs 2 vs 3
was also significant. In HIV positive males with only one reported gonorrhea event (N=573) the
median age of HIV diagnosis was observed to be 27 years old. In those with two reported N.
gonorrheae infections (N=173) the median age of HIV diagnosis was observed to be 26 years
old. In those with 3 reported gonorrhea events (N=76) the median age of HIV diagnosis was 23.5
63

years old, in those with 4 gonorrhea episodes (N=35) the median age of HIV diagnosis was 22,
and in those with 5 or more N. gonorrheae infections between 2005 and 2011 the median age of
HIV diagnosis was observed to be 21 years old. Hazard also increased as incidence of gonorrhea
increased as compared to those with no reported histories of gonorrhea. In those with only one
reported infection of N. gonorrheae the HR=1.96 (95%CI 1.80-2.13), in those with two reported
gonorrhea episodes HR=2.58 (95%CI 2.22-2.99), the hazard ratio in those with three gonorrhea
events was 2.72 (95%CI 2.17-3.40), in those with four N. gonorrheae infections HR=3.64 (2.615.07), and in those HIV positive males with five or more gonorrheal episodes HR=7.26 (5.439.70).
A multivariate modeling approach was also used for multiple N. gonorrheae infections. Since the
previous model testing did not show any benefit to breaking down of sexual history into
heterosexual, homosexual, bisexual and unknown, only models including the MDCH condensed
risk groups were used. Those with no history of gonorrhea were selected as the reference group.
In those with 1 reported N. gonorrheae infection the HRadj=1.87 (95%CI 1.72-2.03). The hazard
increased to HRadj=2.29 (95%CI 1.97-2.66) in those with a history of 2 N. gonorrheae infections.
Those with 3 infections had a HRadj=2.30 (95%CI 1.83-2.88), 4 N. gonorrheae infections
HRadj=3.38 (95%CI 2.43-4.71) and in those with 5 or more reported N. gonorrheae infections
HRadj=6.34 (95%CI 4.74-8.48). Hazard of HIV diagnosis also increase across ethnicities as
compared to Whites. Hazard decreased in all MDCH condensed risk groups as compared to
MSM except for in the MSM/IDU group.
Pre-HIV gonorrhea

64

The study also investigated the hazard of HIV infection associated with only prior N. gonorrheae
infections. Of the 903 HIV positive males in Michigan that had a history of N. gonorrheae
infection only 216 had an event prior to their diagnosis with HIV. Of the 216 with a pre-HIV
gonorrheal episode 152 (70%) had only one reported infection and 64 (30%) had two or more N.
gonorrheae infections up to a maximum of 12 pre-HIV reported N. gonorrheae infections. A
similar dose response was also seen in figure 2.3 when only looking at pre-HIV N. gonorrheae
infections. Concurrent infections defined as those N. gonorrheae infections that happened in the
same year as HIV diagnosis were not included in this analysis since only year of infection/year
of diagnosis were given.
Mean age at HIV diagnosis in those with no history of gonorrhea was 36.8 years old. In those
with only one pre-HIV N. gonorrheae infection the mean reduction in age at HIV diagnosis was
7.3 years, and 10.2 years in those with 2 or more pre-HIV N. gonorrheae infections. A
multivariate model that was adjusted for MDCH condensed risk and race. Table 2.8 shows the
results of the unadjusted and adjusted models. After adjusting for MDCH risk groups and race
the mean reduction in age at HIV in those with only one pre-HIV N. gonorrheae infection was
6.4 years, and 9.0 years in those with two or more pre-HIV N. gonorrheae infections. As
compared to Whites, those who were Black had a reduction in age at HIV diagnosis of 2.2 years,
Hispanics by 2.6 years, and those that were grouped in the Multiracial/Unknown/Other saw a
reduction of 2.6 years. As compared to MSM condensed risk group all other risk groups aside
from Blood Recipient was an increase of age at HIV diagnosis. Blood Recipients was a reduction
of 2.2 years as compared to MSMs. Those that fell into the IDU condensed risk group increased
age at HIV diagnosis by 6.6 years, those that fell into the MSM/IDU condensed risk group
increased age by 0.9 years, Male who had sex with a female at risk for HIV had an increase age
65

at HIV diagnosis of 3.6 years, and those that fell into the unknown/Other condensed risk group
had an increase in age at HIV diagnosis of 5.3 years as compared to MSMs.
The median age in those with no reported N. gonorrheae infections was 36 years old. In those
with only one pre-HIV infection the median age at HIV diagnosis was observed to be 26 years
old, and in those with two or more gonorrhea events the median age of HIV diagnosis was 23.5
years. Table 2.9 contains the adjusted and unadjusted HR for pre-HIV gonorrhea. The unadjusted
HR of one pre-HIV N. gonorrheae infection compared to no history was HR=1.89 (95% CI 1.612.21). In those with two or more gonorrhea events the unadjusted HR=3.12 (95%CI 2.18-4.47).
A multivariate model was also constructed for multiple pre-HIV N. gonorrheae infections. The
model included ethnicity and MDCH condensed risk groups. The HRadj for those with one
reported N. gonorrheae infection was HRadj=1.85 (95%CI 1.57-2.17) and HRadj=2.98 (95%CI
2.33-3.80) for those with two or more pre-HIV N. gonorrheae infections. Hazard increased for
other racial groups as compared to Whites and decreased in the MDCH condensed risk groups
for all group aside from MSM/IDU.
2.5 Conclusion
From 2005 to 2011 there were roughly 19,647 HIV infected males that did not have gonorrhea
prior to 2005. Of these HIV infected males 903 (~5%) were diagnosed with gonorrhea during the
study period. Blacks accounted for 54% of HIV infected males and 75% of HIV infected males
with a history of gonorrhea. Data were collected, coded and linked by MDCH staff. MDCH had
pre-constructed risk groups, MSM, IDU, MSM/IDU, Blood Recipient, Perinatal and
Other/Unknown. Due to low responses of perinatal exposure, perinatal was included in
Other/Unknown. MDCH also supplied a non-condensed version of the collect risk groups. These
66

were analyzed in the first objective, History of Gonorrhea, and sexual orientation was recoded
first by allowing for multiple sex partner responses (male or female) and then by sexual
orientation of presumed heterosexual (female partners, no males), presumed homosexuals (male
partners, no female), presumed bisexual (male and female partners) and unknown (no males no
females). Neither of these explorations of sexual preference added any new insight into the
pathway between N. gonorrheae and HIV and were later omitted from other models. Ethnicity
was also defined by MDCH and recoded due to low values of Asian and American Indians as
separate racial groups.
Prevention efforts for gonorrhea including education, screening, and treatment are well
established yet not necessarily easily implemented. Evidence suggests that N. gonorrheae may
share a synergistic relationship with HIV. The co-infection between N. gonorrheae and HIV is
an understudied relationship that causes these two diseases to express variant epidemiological
patterns.
The time to HIV diagnosis, measured by age at diagnosis, was reduced in those with a history of
gonorrhea as compared to those that have no history of N. gonorrheae infection. History of
gonorrhea included any reported infection between 2005 and 2011 independent of current HIV
status at time of N. gonorrheae infection. This study found that there was an increase in the
hazard of HIV diagnosis in those with a history of gonorrhea.
The study observed a 10 year median (8 year mean) difference in the age of HIV diagnosis in
those with gonorrhea as compared to those that had no reported history. Ethnicity was also found
to significantly increase hazard of HIV diagnosis. Three models were constructed to assess
sexual orientation and other risky activities. The first model used MDCH specified risk groups.

67

In this model an individual only fit into one of the 5 condensed risk groups. MSM was selected
as the reference population and hazard was found to decrease in all other groups, aside from
MSM/IDU which was found to not be significant. The second model was constructed to better
examine the association of MSM individuals and their age at HIV diagnosis. The model showed
that those that reported to have sex with men had an increased hazard, and those that reported sex
with females had a decrease in hazard. The final model was constructed to gain better
understanding of presumed sexual orientation gathered from available data. Individuals were
coded as heterosexual if they only responded yes to having sex with females, homosexual if they
only responded yes to having sex with males, bisexual if they responded yes to both male and
female and unknown if the response was no in both males and females. The model found that
both presumed homosexuals and presumed bisexuals had an increase in hazard (lower age of
HIV diagnosis) than presumed heterosexual individuals. However, neither of these
reformulations of the MDCH condensed risk added any new information to the associations
between gonorrhea, MSM, and HIV and due to the amount of unknown orientations, did not
respond yes to having sex with men or women, MDCH’s condensed risk group structure was
deemed to be the most useful.
A possible synergistic relationship between N. gonorrheae and HIV was observed in this study.
The study also observed that N. gonorrheae infection prior to HIV diagnosis significantly
increased hazard of diagnosis for males. This relationship was also seen to follow a dose
response pattern in both overall history of gonorrhea and in pre-HIV only gonorrhea with
multiple N. gonorrheae re-infections increasing the hazard, reducing the age at HIV diagnosis.
While overall history of gonorrhea allowed for insight into risky sexual behaviors, pre-HIV
gonorrhea allowed for better understanding of the temporal relationship. Median age of HIV
68

diagnosis ranged from 27 years old in those with only one reported N. gonorrheae infection to 21
years old to individuals with five or more reported infections. In individuals with one pre-HIV
infection the median age of HIV diagnosis was 26, and 23.5 in those with two or more pre-HIV
infections as compared to 36 in gonorrhea free males. These increases in hazard represent
possible intervention points for HIV reduction strategies. Resources may best be used in focusing
on individuals that come in with their first case of gonorrhea and counseling them on safe sex
practices aimed not only at reducing their risk of HIV but also aimed at reducing risk of N.
gonorrheae re-infection. Resources should also be aimed at educating males that every N.
gonorrheae infection increases their hazard of HIV diagnosis.
The study also found that 37% (N=330) of those with gonorrhea between 2005 and 2011 had
more than one occurrence of the infection. This represents a key group to intervene and provide
prevention strategies to help to reduce re-infection rates. The study found a 6 year age reduction
in age of HIV diagnosis in those with 5 or more infections as compared to those with only one
reported infection. In the fight against HIV and AIDS in Michigan, any deferment in the time to
HIV or progression to AIDS is a victory. This study showed that risky sexual behavior, as
measured by reported gonorrhea, results in a reduction in the time an individual remains HIV
free. The study also found that as number of N. gonorrheae infections in an individual increases
so does their hazard of being diagnosed with HIV.
Limitations of this study include only being able to gain access to data on gonorrhea. Bacterial
STDs promote HIV in infecting white blood cells (1, 24, 27-29, 31) and other STDs that were
not included could have had an influence on HIV seroconversion. Gonorrhea infection was used
as a proxy for risky sexual behavior and a marker for individuals being in sexual networks where
STDs were present. Because this study could not control for the effect of other STDs on age of
69

HIV diagnosis, the resulting synergistic relation is only speculation driven largely by the effect
of pre-HIV gonorrhea median age decreases. Another limitation of this study is that both HIV
diagnosis and gonorrhea date are reported only in years. N. gonorrheae infections that occurred
in the same year as diagnosis were excluded from the pre-HIV analysis due to not being able to
place them temporally before, at, or after HIV diagnosis. A final limitation is that since it is very
hard to pin point the actual date of HIV infection the study really measures the age at diagnosis
of HIV and the actual date of infection could be a year or more prior.
N. gonorrheae and other STIs are reportable diseases, and resources should be given to studying
their epidemiological patterns in co-infected individuals. Identification of possible risk factors
that may allow for intervention are important tools in trying to prevent the spread not only of
HIV but other STDs as well. Education and counseling efforts should be focused on those with
their first STD as well as those who are HIV naïve and have multiple reported STD infections as
a HIV prevention tool. Education and counseling has to walk the fine line between
destigmatization of STDs (so individuals get tested and do not feel ashamed if they are positive)
and warning individuals about the dangers of these diseases. As individuals within public health
we must change the paradigm from “you just have gonorrhea”, to the sexual behaviors that lead
to you getting gonorrhea puts you at risk for HIV and other sexually transmitted diseases which
may not be curable.

70

APPENDIX

71

Table 2.1: Demographic Distribution of 19,647 HIV Positive MI Males (2005-2011)

Variable

No N. gonorrheae history
N=18744

N. gonorrheae history
N=903

7635 (40.73%)
9988 (53.29%)

177 (19.60%)
674 (74.64%)

Hispanic
Asian
American Indian

711 (3.79%)
62 (0.33%)
40 (0.21%)

23 (2.55%)
1 (0.11%)

Multiracial/Unknown/
Other

308 (1.64%)

28 (3.10%)

Married
Divorced

1317 (7.03%)
1268 (6.76%)

2 (0.22%)
2 (0.22%)

Single
Other/Unknown/
Widower/Missing
Risk Group

4966 (26.49%)
11193 (59.72%)

10 (1.11%)
889 (98.45%)

MSM
IDU
MSM/IDU
Blood Recipient
Male HCFR
Other/Unknown
AIDS Positive
Deceased

11018 (58.78%)
2675 (14.27%)
1338 (7.14%)
268 (1.43%)
770 (4.11%)
2675 (14.28%)
13767 (73.45%)
8906 (47.51%)

697 (77.19%)
21 (2.33%)
21 (2.33%)
20 (2.21%)
144 (15.95%)
390 (43.19%)
20 (2.21%)

Ethnicity
White
Black

Marital Status

72

Table 2.2: MDCH Non-Condensed Risk Group Distribution of HIV Positive MI Males
(2005-2011)

Risk Group

No N. gonorrheae history
N=18744

N. gonorrheae history
N=903

Sex with Males

12356 (65.92%)

718 (79.51%)

Sex with Females

9192 (49.04%)

412 (45.63%)

Injection Drug Use (IDU)

4013 (21.41%)

42 (4.65%)

Received Clotting Factor

7 (0.04%)

1 (0.11%)

Received a Transfusion

331 (1.77%)

4 (0.44%)

Received a Transplant

10 (0.05%)

-

1372 (7.32%)

26 (2.88%)

Heterosexual Sex with a
Hemophiliac

27 (0.14%)

-

Heterosexual Sex with a
Transfusion Recipient

36 (0.19%)

-

Heterosexual Sex with a
Transplant Recipient

13 (0.07%)

-

Heterosexual Sex with a Person
Living with HIV

1610 (8.59%)

58 (6.42%)

Worked in a Health Care Setting

762 (4.07%)

9 (1.00%)

Heterosexual Sex with an IDU

73

Table 2.3: General Linear Model Adjusted and Unadjusted Means for Age at HIV
Diagnosis in Males With and Without Any History of Gonorrhea
N. gonorrheae
History

No Infection/
Baseline
History of gonorrhea
Ethnicity
White
Black
Hispanic
Multiracial/Unknown/
Other*
Risk Group
MSM
IDU
MSM/IDU
Blood Recipient
Male HCFRǂ
Other/Unknown†
Sex with Males
Sex with Females
IDU (Not Condensed)
Sex with HIV Female
Presumed
Heterosexual
Presumed
Homosexual

Mean
Unadjusted
Age at HIV
Diagnosis
36.8 years old
(36.6-36.9)
28.4 years old
(27.5-29.1)

Change in
AdjustedM1
Mean Age at
HIV Diagnosis
36.3 years old
(36.0-36.5)
7.2 years
(-8.5--5.9)

Change in
AdjustedM2
Mean Age at
HIV Diagnosis
40.4 years old
(40.0-40.8)
7.2 years
(-8.5--5.8)

Change in
AdjustedM3
Mean Age at
HIV Diagnosis
41.1 years old
(40.7-41.5)
7.2 years
(-8.4--5.8)

37.1 years old
(36.9-37.4)
36.4 years old
(35.9-37.0)
35.9 years old
(34.9-37.0)
35.5 years old
(34.3-36.9)

-

-

-

34.7 years old
(34.5-34.9)
40.9 years old
(40.3-41.5)
35.6 years old
(34.8-36.3)
33.7 years old╫
(32.3-34.7)
37.9 years old
(37.0-38.9)
39.4 years old
(38.8-40.0)
34.8 years old
(34.2-35.3)
37.3 years old
(36.8-37.8)
39.1 years old
(38.6-39.6)
35.8 years old
(35.2-36.5)
39.5 years old
(39.2-39.7)
34.5 years old
(33.9-35.1)

-

-

-

2.3 years
(-2.6--1.9)
2.6 years
(-3.4--1.9)
2.6 years
(-3.6--1.6)

2.1 years
(-2.4--1.8)
2.4 years
(-3.2--1.7)
2.5 years
(-3.4--1.5)

2.1 years
(-2.4--1.8)
-2.4 years
(-3.2--1.7)
2.5 years
(-3.4--1.5)

-

-

-

6.6 years
(6.2-7.0)
0.9 years
(0.3-1.4)
2.2 years
(-3.4--1.0)
3.5 years
(2.9-4.3)
5.3 years
(4.8-5.7)
-

-

-

-

-

-

-

-

-

-

-

-

4.5 years
(-4.9--4.2)
0.9 years
(0.6-1.3)
1.5 years
(1.1-1.9)
2.0 years
(-2.5--1.5)
-

-

-

-

74

-

1.5 years
(1.1-1.9)
1.9 years
(-2.5--1.4)
-

5.4 years
(-5.8--5.0)

Table 2.3 (cont’d)
Presumed Bisexual
Unknown Orientation

35.3 years old
(34.6-35.9)
39.9 years old╫
(39.1-40.7)

M1

-

-

-

-

Model 1
Model 2
M3
Model 3
*
Includes Asian/Hawaiian, Pacific Islanders and American Indian, Alaskan Native
M2

ǂ

Male who had sex with a female at risk for HIV
Includes Perinatal
╫
Not statistically significant
†

75

-

4.0 years
(-4.5--3.6)
0.02 years╫
(-0.6-0.5)

Table 2.4: Median Age at HIV Diagnosis and Unadjusted Hazard of HIV in HIV Positive
Males in Michigan 2005-2011
N. gonorrhea History

Median Age of HIV
Diagnosis

Unadjusted Hazard
Ratio

No N. gonorrhea Infection

36 years old

Reference Group

History of N. gonorrhea Infection

26 years old

2.24 (2.09, 2.39)

1 N. gonorrhea Infection

27 years old

1.96 (1.80, 2.13)

2 N. gonorrhea Infections

26 years old

2.58 (2.22, 2.99)

3 N. gonorrhea Infections

23.5 years old

2.72 (2.17, 3.40)

4 N. gonorrhea Infections

22 years old

3.64 (2.61, 5.07)

5 or More N. gonorrhea Infections

21 years old

7.26 (5.430, 9.70)

1 Pre-HIV N. gonorrhea Infection

26 years old

1.90 (1.62, 2.23)

23.5 years old

3.20 (2.51, 4.09)

Multiple Pre-HIV N. gonorrhea
Infections

76

Table 2.5: Multivariate Cox Regression Analysis of Hazard of HIV Diagnosis in MI HIV Positive Males with a History of N.
gonorrheae 2005-2011
Variable
History of N.
gonorrheae
No Infection
History of gonorrhea
Ethnicity
White
Black
Hispanic
Multiracial/Unknown/
Other*
Risk Group
MSM
IDU
MSM/IDU
Blood Recipient
Male HCFRǂ
Other/Unknown†
Sex with Males
Sex with Females
IDU (Not Condensed)
Sex with HIV Female
Presumed
Heterosexual
Presumed
Homosexual
Presumed Bisexual
Unknown Orientation

Unadjusted
HR

1
2.24
1
1.07
1.10
1.17

95% CI

Adjusted
HR M1

2.09-2.39

1
2.08

1.04-1.11
1.02-1.19
1.06-1.28

1
1.25
1.22
1.27

95%CIMI

Adjusted
HR M2

1.95-2.23

1
2.08

1.21-1.29
1.13-1.32
1.16-1.40

1
1.24
1.22
1.27

Adjusted
HR M3

95%CIM3

1.95-2.23

1
2.08

1.94-2.22

1.20-1.28
1.13-1.31
1.15-1.40

1
1.25
1.22
1.27

1.21-1.29
1.13-1.31
1.16-1.40

1
0.64
0.98
0.80
0.75
0.65
1.51
0.85
0.81
1.09
1

0.62-0.67
0.93-1.04
0.70-0.90
0.70-0.81
0.62-0.67
1.47-1.56
0.83-0.87
0.79-0.84
1.04-1.15

1
0.60
0.95
0.88
0.71
0.62
-

1.55

1.49-1.60

-

-

1.69

1.63-1.76

1.39
0.93

1.34-1.45
0.88-0.98

-

-

1.43
0.98

1.37-1.49
0.93-1.04

0.57-0.62
0.90-1.01
0.78-0.99
0.66-0.76
0.59-0.64

77

1.54
0.90
0.94
1.22
-

95%CIM2

1.49-1.59
0.87-0.92
0.91-0.98
1.15-1.28

0.95
1.21
1

0.91-0.99
1.15-1.27

Table 2.5 (cont’d)
M1

Model 1,M2 Model 2,M3 Model 3,* Includes Asian/HI/PI AmIn/AkNat ,ǂ Male who had sex with a female at risk for HIV,† Includes Perinatal

78

Table 2.6: General Linear Model Adjusted and Unadjusted Means for Age at HIV
Diagnosis in Michigan Males with Multiple N. gonorrheae Infections
Variable

History of N.
gonorrheae
No Infection/
Baseline
1 Infection
2 Infections
3 Infections
4 Infections
5 or More
Infections
Ethnicity
White
Black
Hispanic
Multiracial/Unknown/
Other*
Risk Group
MSM

Mean
Unadjusted Age
at HIV
Diagnosis

Change in
Adjusted Mean
Age at HIV
Diagnosis

36.8 years old
(36.6-36.9)
29.5 years old
(28.4-30.4)
27.5 years old
(25.7-30.4)
26.7 years old
(24.2-29.1)
25.0 years old
(21.4-28.5)
22.9 years old
(19.8-26.0)

36.2 years old
(36.0-36.5)
5.1 years
(-6.9--3.2)
10.1 years
(-12.9--7.3)
6.2 years
(-10.2--2.1)
10.0 years
(-14.4--5.5)
10.9 years
(-15.6--6.2)

37.1 years old
(36.9-37.4)
36.4 years old
(35.9-37.0)
35.9 years old
(34.9-37.0)
35.5 years old
(34.3-36.9)

-

2.2 years
(-2.5--1.9)
2.6 years
(-3.4--1.9)
2.6 years
(-3.6--1.6)

34.7 years old
(34.5-34.9)
IDU
40.9 years old
6.6 years
(40.3-41.5)
(6.2-7.0)
MSM/IDU
35.6 years old
0.9 years
(34.8-36.3)
(0.3-1.4)
Blood Recipient
33.7 years old╫
2.2 years
(32.3-34.7)
(-3.4--1.0)
Male HCFRǂ
37.9 years old
3.6 years
(37.0-38.9)
(2.9-4.3)
Other/Unknown†
39.4 years old
5.2 years
(38.8-40.0)
(4.8-5.7)
*
Includes Asian/HI/PI AmIn/AkNat ,ǂ Male who had sex with a female at risk for HIV,† Includes
Perinatal
79

Table 2.7: Multivariate Cox Regression Analysis of Hazard of HIV Diagnosis in MI HIV
Positive Males with a History Multiple of N. gonorrheae 2005-2011
Variable

Unadjusted HR

95% CI

Adjusted
HR

95%CI

History of N.
gonorrheae
No Infection
1
1
1 Infection
1.96
1.80-2.13
1.87
1.72-2.03
2 Infections
2.58
2.22-2.99
2.29
1.97-2.66
3 Infections
2.72
2.17-3.40
2.30
1.83-2.88
4 Infections
3.64
2.61-5.07
3.38
2.42-4.71
5 or More
7.26
5.43-9.70
6.34
4.74-8.48
Infections
Ethnicity
White
1
1
Black
1.07
1.04-1.11
1.24
1.20-1.28
Hispanic
1.10
1.02-1.19
1.22
1.13-1.32
Multiracial/Unknown/ 1.17
1.06-1.28
1.27
1.16-1.40
Other*
Risk Group
MSM
1
1
IDU
0.64
0.62-0.67
0.60
0.57-0.62
MSM/IDU
0.98
0.93-1.04
0.95
0.90-1.01
Blood Recipient
0.80
0.70-0.90
0.88
0.78-0.99
ǂ
Male HCFR
0.75
0.70-0.81
0.71
0.66-0.76
Other/Unknown†
0.65
0.62-0.67
0.62
0.59-0.64
*
ǂ
Includes Asian/HI/PI AmIn/AkNat , Male who had sex with a female at risk for HIV,† Includes
Perinatal

80

Table 2.8: General Linear Model Adjusted and Unadjusted Means of Age at HIV Diagnosis
in Michigan Males with Pre-HIV N. gonorrheae Infections
Variable

History of N.
gonorrheae
No Infection/
Baseline
1 Pre-HIV Infection
2 or More Pre-HIV
Infections
Ethnicity
White
Black
Hispanic
Multiracial/Unknown/
Other*
Risk Group
MSM

Mean
Unadjusted Age
at HIV
Diagnosis

Change in
Adjusted Mean
Age at HIV
Diagnosis

36.8 years old
(36.6-36.9)
29.5 years old
(27.6-31.2)
26.6 years old
(23.9-29.2)

36.2 years old
(36.0-36.5)
6.4 years
(-8.0--4.8)
9.0 years
(-11.5--6.6)

37.1 years old
(36.9-37.4)
36.4 years old
(35.9-37.0)
35.9 years old
(34.9-37.0)
35.5 years old
(34.3-36.9)

-

2.2 years
(-2.6--1.9)
2.6 years
(-3.4--1.9)
2.6 years
(-3.6--1.6)

34.7 years old
(34.5-34.9)
IDU
40.9 years old
6.6 years
(40.3-41.5)
(6.2-7.0)
MSM/IDU
35.6 years old
0.9 years
(34.8-36.3)
(0.3-1.4)
Blood Recipient
33.7 years old╫
2.2 years
(32.3-34.7)
(-3.4--1.0)
Male HCFRǂ
37.9 years old
3.6 years
(37.0-38.9)
(2.9-4.3)
Other/Unknown†
39.4 years old
5.2 years
(38.8-40.0)
(4.8-5.7)
*
Includes Asian/HI/PI AmIn/AkNat ,ǂ Male who had sex with a female at risk for HIV,† Includes
Perinatal

81

Table 2.9: Multivariate Cox Regression Analysis of Hazard of HIV Diagnosis in MI HIV
Positive Males with a History of Pre-HIV N. gonorrheae 2005-2011
Variable

Unadjusted HR

95% CI

Adjusted
HR

95%CI

History of Pre-HIV
N. gonorrheae
No Infection
1
1
1 Pre-HIV Infection 1.90
1.62-2.23
1.85
1.57-2.17
2 or More Pre-HIV
3.20
2.51-4.09
2.98
2.33-3.80
Infections
Ethnicity
White
1
1
Black
1.07
1.04-1.11
1.21
1.18-1.25
Hispanic
1.10
1.02-1.19
1.23
1.14-1.33
Multiracial/Unknown/ 1.17
1.06-1.28
1.24
1.12-1.67
*
Other
Risk Group
MSM
1
1
IDU
0.64
0.62-0.67
0.60
0.58-0.63
MSM/IDU
0.98
0.93-1.04
0.96
0.91-1.02
Blood Recipient
0.80
0.70-0.90
0.87
0.77-0.98
Male HCFRǂ
0.75
0.70-0.81
0.72
0.67-0.77
†
Other/Unknown
0.65
0.62-0.67
0.61
0.59-0.64
*
Includes Asian/HI/PI AmIn/AkNat ,ǂ Male who had sex with a female at risk for HIV,† Includes
Perinatal

82

Figure 2.1: Kaplan-Meier Curves for Age at HIV Diagnosis in HIV Positive MI Males with
and without History of N. gonorrheae (2005-2011).

Trend: Logrank p <.0001

History of Gonorrhea

83

Figure 2.2: Kaplan-Meier Curves for Age at HIV Diagnosis for Multiple N. gonorrheae
Infections from 2005-2011 in MI HIV Positive Males.

Trend: Logrank p <.0001

Number of
N. gonorrheae
Infections

+

84

Figure 2.3: Kaplan-Meier Curves for Age at HIV Diagnosis in MI Males with Pre-HIV N.
gonorrheae Infections 2005-2011

Trend: Logrank p <.0001

History of
Pre-HIV
Gonorrhea

85

Chapter 3.
Enhancement of HIV Progression by N. gonorrheae Co-Infection in Michigan Males
3.1 Abstract
Introduction
Evidence suggests that there is a synergistic relationship between N. gonorrheae and HIV with
regard to progression of HIV infection, towards AIDS and death. N. gonorrheae co-infection in
HIV infected individuals has been shown to both increase the hazard associated with progression
to AIDS and death. The objective of this study is to test whether N. gonorrheae increases the
hazard (younger at event) associated with progression to AIDS and death, as well as increases
the hazard (shorter lag time) of an HIV progression event from HIV diagnosis.
Methods
A retrospective cohort study of 19,647 HIV infected males in the State of Michigan from 20052011 was created through collaboration with Michigan Department of Community Health. Cases
consisted of males in Michigan with a reported incidence of N. gonorrheae infection between
2005 and 2011 and were diagnosed with HIV by 2011. Controls were Michigan males with no
history of N. gonorrheae infections who were diagnosed with HIV by 2011. Time to event
analysis was used to determine hazard of progression to AIDS and death.
Results
The median age at death in HIV positive males with no history of N. gonorrheae infection was
found to be 41 years as compared to 38.5 years individuals with a history of gonorrhea. The time
period from HIV to death was longer in those with gonorrhea than in those with no history, 5
86

years and 3 years respectively. The median age at AIDS in individuals with no history of
gonorrhea was 38 years as compared to 31 in those with a history of gonorrhea. Those of AIDS
positive status with only one reported infection with N. gonorrheae had a median age of AIDS
positive status of 32 years, 29 years in those with two N. gonorrheae infections, 32 years in those
with three N. gonorrheae infections, 21 years in those with 4 N. gonorrheae infections, and 24
years in those with 5 or more N. gonorrheae infections. Those with gonorrhea, but no post-HIV
gonorrheal episode had a median age of AIDS positive status of 29 years old. In those with a
history of post-HIV gonorrhea as well as in those with only one post-HIV N. gonorrheae
infection the median age of AIDS positive status was 32 years, in those with two post-HIV N.
gonorrheae infections the median age was 30.5 and those with 3 or more post-HIV infections the
median age was 26. Time from HIV to AIDS positive status was longer in those with a history of
N. gonorrheae as compared to those without. The median time from HIV to AIDS in those with
no history of N. gonorrheae infection was less than one year. In those with a history of
gonorrhea, as well as those with 1-4 N. gonorrheae infections the median time from HIV to
AIDS was 1 year. In those with 5 or more N. gonorrheae infections the median time was 2 years.
In individuals with gonorrhea, but no post-HIV gonorrhea the median time between HIV and
AIDS was less than one year. In both history of post-HIV gonorrhea as well as in males with 1-2
reported post-HIV N. gonorrheae infections the median time from HIV to AIDS was 2 years,
and in those with three or more post-HIV infections the median time was 3 years.
Conclusion
A decrease in age at a progression to AIDS or death was observed in individuals with a history of
N. gonorrheae infection and persisted when looking at post-HIV N. gonorrheae infections. An
observed increase in the lag time between HIV diagnosis and AIDS positive status was also
87

found. Infection with N. gonorrheae may reduce the hazard of time from HIV to progression
event; however, more research is needed to assess possible confounders that may be influencing
this potential paradoxical relationship.

88

3.2 Introduction
Epidemiological evidence suggests that there is a synergistic effect between co-infection with N.
gonorrheae and infection with HIV resulting in an increase in the progression of HIV in these
individuals. Wasserheit (1992) explored the synergistic relationship between STDs and HIV.
Wasserheit proposed that there were 3 key relationships between these diseases with the second
being the theorized accelerated progression of HIV to AIDS or death in individuals with STD coinfections (95). Wasserheit also noted that STD infection can serve as a useful proxy for difficult
to measure variables such as anonymous sex or use of sex workers (95). A review article also
focusing on the epidemiological synergistic relationship between STDs and HIV by Rottingen et
al. (2001) found that individuals with periods of STD/HIV co-infection have a decrease in time
to AIDS positive status. However, while many studies investigated viral load and CD4 cell count
during co-infection periods, the actual effect N. gonorrheae and other STDs on age of AIDS
diagnosis, age at death and time from HIV to one of these progression events has not been well
documented. There is a need to study not only laboratory markers of HIV progression in coinfected individuals, but also progression events such as death and AIDS positive status in these
individuals.
N. gonorrheae is one of the most reported diseases in the United States with multiple outbreaks
of the disease occurring yearly; however, its effect of HIV during periods of co-infection is not
fully understood (15). In 2011, WHO estimated that roughly 88 million of the estimated 448
million curable STD infections reported that year were due to N. gonorrheae (12). In the United
States in 2012, there were over 334,000 N. gonorrheae cases resulting in an increased rate of
4.1% from the previous year (13).

89

The World Health Organization (WHO) estimated in 2012 that there were 35.3 million people
living with HIV and a total of 2.3 million infections and 1.6 million deaths due to AIDS related
causes (6). Roughly 9.7 million people in 2012 were estimated to be on ARV therapy (6). Under
the 2013 WHO ARV guidelines, there are an estimated 26 million people eligible for therapy
leaving a large gap between those that are eligible and those that are receiving ARV treatment
(6).
In 2011, the CDC estimated that there were over 32,000 HIV stage 3 (AIDS positive) diagnosed
individuals in the U.S. with 24,000 of them being in adult and adolescent males (8). There were
over 1.1 million AIDS diagnoses by 2010 in the United States with 487,692 individuals living
with AIDS and 15,000 AIDS related deaths in 2010 alone (8). The top three states of AIDS
diagnosis were California (n=3,622), New York (n=3,574), and Florida (n=3,440) in 2011
respectively (8,9). The state of Michigan in 2011 ranked 20th in HIV diagnoses and 18th in AIDS
diagnoses (9).
As of the 2010 Census Michigan has the 8th largest population in the US. In 2012, there were
12,770 reported cases of N. gonorrheae infections in the state of Michigan. In 2011, 259 of the
13,000 reported cases were associated with HIV co-infection. The highest rate of N. gonorrheae
infection occurred in individuals between 20 and 24 years old. This age group represents 6.7% of
Michigan population but accounts for 1/3 of all N. gonorrheae infections in the state of
Michigan. The rates of infection with N. gonorrheae were higher in African Americans than in
Whites, and 43% of all cases were males.
As of January 2014, there was an estimated 8,207 Stage 3/AIDS positive individuals in the State
of Michigan with a reported 428 new Stage 3/AIDS positive individuals identified in 2013 (10).

90

The total number of individuals living with AIDS (Stage 3 HIV) increased from 8,207 reported
in January to 8,240 in the July 2014 report, with the new 2013 AIDS total at 471 (111). Of all
those diagnosed with HIV up to January 2014, 18,794 of 27,407 progressed to AIDS/Stage 3
positive status (111). Among the AIDS positive individuals 35% were diagnosed with AIDS at
the same time of their HIV diagnosis, 8% had a “short” lag time (1-12 months after HIV
diagnosis) to AIDS positive status, and 26% had a “long” lag period (more than 12 months from
HIV diagnosis) (111). Males with HIV had a higher percent AIDS progression than females,
70% of males as compared to 64% of females (111). Within males with HIV, 37% were
diagnosed as AIDS positive at the time of their HIV diagnosis, 8% had a short lag time, and 25%
of males had a long lag time to AIDS positive status (111). In females with HIV the percent of
concurrent AIDS diagnosis (HIV and AIDS diagnosis at the same time) was 28%, 8% had a
short lag time, and 28% had a long lag time (111).
N. gonorrheae infections were thought to be on a decline at the early start of HIV in the US
(circa 1986), however, gonorrhea started its re-emergence in the MSM population around 1990
and reached its peak in 2005 (34). Rietmeijer et al. note that N. gonorrheae infection of HIV
positive males can be broken into two time periods with a decline between them (37). The first
increase in HIV positive males occurred between 1995 and 1997 with a decline between 1997
and 1998 (37). This was then followed in 1999 by a second period of N. gonorrheae/HIV coinfections that lasted until 2000 (37).
A study conducted to study HIV and STDs in US military personnel of around 2,000 HIV
positive individuals, found that among HIV positive personnel, 50% also had a history of a STD
(17). The study found that of those with a history of STD infection, 47% were reported to only
have had a STD infection after HIV diagnosis and 20% having both pre and post HIV STD
91

infections (17). Syphilis was the most common post-HIV STD infection while N. gonorrheae
accounted for 24% (N=157) of post-HIV STD infections (17). The study also found that
individuals had multiple gonorrhea events with a total of 186 reported N. gonorrheae infections
in 157 individuals (17). A study by the Minnesota Department of Health found that of 1.3% of
HIV positive individuals in the study had at least one reported diagnosis of an STD (92). Among
MN HIV positive individuals, gonorrhea incidence was higher than it was in the HIV negative
population (92). 17% of HIV positive individuals with gonorrhea had 2 or more N. gonorrheae
infections from 1993 to 1994, all of which were male (92). Other studies have reported as high as
a 20 fold increase in gonorrhea in HIV infected men than in those that are HIV negative (35).
In a review of the synergistic relationship between STDs and HIV Rottingen et al. suggest that
along with susceptibility of the HIV infection in naïve individuals the HIV/STD co-infection
period decreases the time to AIDS positive status (38). A study by Mayaud and McCormick also
found that STDs both ulcerative and inflammatory effect the progression of HIV to AIDS (40).
An editorial review by Bentwich et al. also states that during HIV co-infections the activation of
cytokines and chemokines, anergy and apoptosis, as well as increased HIV replication all
influence the progression of HIV to AIDS (45).
3.3 Methods
A retrospective cohort of HIV positive males in Michigan were selected as the study population.
Cases were defined as HIV positive males with a history of reported N. gonorrheae between
January 1st 2005 and December 31st 2011. HIV infected males with N. gonorrheae prior to 2005
were excluded from the study population. All HIV positive males without a history of N.
gonorrheae during the study time period that were alive as of January 1st 2005 were included in

92

the study as controls. A linkage structure between HIV and N. gonorrheae data for this study was
created in collaboration with Michigan Department of Community Health (MDCH). Records
with missing first name, last name, or date of birth were excluded from the study population
prior to the matching of data. A total of 19,647 HIV infected males in the state of Michigan were
included in this study. Ethical approval was obtained from the Institutional Review Board (IRB)
at Michigan State University (MSU) and was reciprocal with the MDCH IRB. An MDCH
employee at the HIV Surveillance and Body Art Unit HIV/STD/VH/TB Epidemiology created
the record linkage structure between HIV datasets and N. gonorrheae data sets to comply with
patient confidentiality and privacy. The match was conducted between the two registries using
LinkPlus. All data used in this study were de-identified according to HIPAA guidelines on
public health information. Study staff also did not contact or attempt to contact any participants
of the study.
HIV infection was defined as HIV-NA=HIV infection, Non-Stage 3 and AIDS=Stage 3 HIV
infection. Acute and Latent, Stage 1 and 2, differentiated in this study. Date of birth, date of
death, N. gonorrheae infection, as well as HIV and AIDS diagnosis were all limited to year of
the event. Vital status was coded as 1=alive, 2=deceased, 9=unknown. Ethnicity was classified
into unique non-over lapping groups, 1.White= White non-Hispanic, 2.Black= Black nonHispanic, 3.Hispanic=Hispanic of any race, 4.Asian/HI/PI=Asian/Hawaiian, Pacific Islander
non-Hispanic, 5.Am In /Ak Nat=American Indian, Alaskan Native, non-Hispanic, or
9.Multi/Unk/Other=Multirace, Unknown, Other, non-Hispanic. Due to low rates of
Asian/Hawaiian, Pacific Islanders (N=0) and American Indian, Alaskan Native (N=1) in the N.
gonorrheae group these racial groups were included in the 9.Multi/Unk/Other grouping to avoid
analytical complications. Marital status while largely missing was categorized as A=Married and
93

Separated, S=Single and never married, M=Married, N=Not otherwise specified, Missing or
U=Unknown, W=Widowed; Risk was condensed and categorized as A.MSM=Male-Male Sex,
B.IDU=Injecting Drug Use, C.MSM/IDU=Male-Male Sex and Injecting Drug Use, D.Blood
Recipient=Received Blood, E1.(Male HCFR)=Male who had sex with a female at risk for HIV,
F.Perinatal=Child exposed by mother (further condensed to risk group G.Unk:Other due to low
response, N=3), G.Unk:Other=Unknown/Other. Individual risk categories were also coded and
were analyzed as “Yes” and “Not Yes” to handle missing data as suggested by MDCH. Mothers
HIV status was also coded, although similar to marital status largely missing as 1=Refused HIV
Testing, 2=Known Uninfected after Birth, 3=Know HIV+ before pregnancy, 4=Known HIV+
during pregnancy, 5=Known HIV+ at time of delivery, 6=Known HIV+ sometime before birth,
7=Known HIV+ sometime after birth, 8=HIV+ with time unknown, 9=Unknown and is mainly
used for individuals with risk group F. Perinatal where they were infected via the mother. Table
3.1 shows the descriptive statistics for HIV positive males in MI from 2005-2011.
Statistical analysis of the data was performed using SAS 9.3 (SAS Institute, Inc. Cary, North
Carolina). Descriptive statistics were preformed to assess distribution of demographic and risk
characteristics. Time to event analysis was used to investigate change in age at AIDS diagnosis
as well as time from HIV to AIDS and Kaplan-Meier curves were created to explore unadjusted
trends. Event was defined as AIDS diagnosis and time was defined as age of AIDS diagnosis.
Adjusted and unadjusted models were run time from HIV to event models. Only unadjusted
models were constructed for age at HIV progression event. Survival analysis and Kaplan-Meier
curves were created to investigate age at death as well as time from HIV diagnosis to death in
those with and without a history of N. gonorrheae infection.

94

As stated above death was coded by two variables, the first variable was vital status,
yes/no/unknown, and year of death for those that died. The variable of time from HIV diagnosis
to death was constructed by taking the year of death and subtracting the year of HIV diagnosis.
Due to low numbers of deaths in gonorrhea positive HIV males in this study population, no
multiple N. gonorrheae infection models were ran on the data. Gonorrhea was coded as
match=’N’ for no history of gonorrhea and match=’Y’ for reported history of gonorrhea by
MDCH.
Current Diagnostic Status was coded as HIV-NA=HIV infection, Non-Stage 3 and AIDS=Stage
3 HIV infection and only year of diagnosis was recoded for HIV and AIDS. History of gonorrhea
was ran as match=’N’ for no history of gonorrhea and match=’Y’ for reported history of
gonorrhea and a model looking at multiple N. gonorrheae infections was also constructed.
Number of total gonorrhea events was constructed by count number of years of reported N.
gonorrheae infections an individual had. The cut point of 5 or more was selected because of the
large drop off in individuals with greater than 5 recorded N. gonorrheae infections between 2005
and 2011. Post-HIV gonorrhea events were constructed by taking the year of gonorrhea and
subtracting the year of HIV diagnosis, positive values constituted post-HIV infections, 0
indicated same year of diagnosis, and negative numbers indicated pre-HIV gonorrhea. All
concurrent years (where HIV diagnosis and N. gonorrheae infection were equal) were not
included in the Post-HIV variable. Post-HIV infections were tabulated and a cut point of 3 or
more was used for analysis.
3.4 Results

95

In Michigan from 2005-2011, there were 903 HIV positive males that had a reported history of
N. gonorrheae infection. Of these individuals, 390 (43%) progressed to AIDS positive status by
2011. Of the remaining HIV positive males in Michigan that were alive as of January 1st 2005,
who had no reported gonorrhea, 13,767 (73%) progressed to AIDS positive status by 2011 and
48% died as compared to only 2% in those with a reported history of gonorrhea. Demographic
distributions are shown in Table 3.1. MDCH supplied both a condensed single variable to
determine high risk sexual behavior as well as individual binary components. In HIV positive
males that had no history of N. gonorrheae infection 59% identified themselves as MSM and 7%
identified as MSM as well as an injection drug user (IDU). Presumed heterosexual IDU
accounted for 14% of the gonorrhea free HIV males. 4% of gonorrhea free HIV males were
males that have had sex with a female that was HIV positive and the remaining 14% were
categorized as other or unknown. In HIV positive males with a history of gonorrhea, 77%
identified themselves as MSM with another 2% identifying themselves as MSM/IDU. 2% of
HIV positive males with gonorrhea were IDU, 2% were males that have had sex with a HIV
positive female, and 16% were other or unknown.
Table 3.2 shows the demographic distributions for those who have progressed to AIDS positive
status compared to those that have not progressed to AIDS as of 2011. Of those HIV positive
males in MI that have progressed to AIDS positive status, 40% were White, 54% were Black and
about 4% were Hispanic. Most individuals did not give marital status (51%), of those that did the
majority were single. In the MDCH condensed sexual risk groups MSM accounted for 60% of
AIDS positive males with 7% identifying themselves as MSM/IDU and 15% as only IDU. 2%
identified themselves as blood recipients, 4% of AIDS positive males had sex with a female that
was positive for HIV, and 12% fell into the other or unknown risk group. In HIV positive males
96

that had not progressed to AIDS positive status by 2011 39% were White, 55% were Black and
4% were Hispanic. The majority of individuals did not report marital status (89%) and of those
that did the majority were single. In the MDCH condensed risk groups 60% of non-AIDS
positive HIV males were MSM with 6% identifying themselves as MSM/IDU, and 10%
identifying themselves as only IDU. 4% were males that have had sex with a HIV infected
female and 20% fell into the other or unknown risk group.
The non-condensed distribution of high risk factors by AIDS status can be seen in Table 3.3. Of
the total study population 14,157 of 19,647 HIV positive males in Michigan progressed to AIDS
positive status by 2011. Individuals may fall into multiple categories of the non-condensed risk
groups as opposed to the MDCH condensed grouping, which only allows for an individual to
exist in one hierarchical category. Of these individuals who progressed to AIDS positive status
67% had sex with men, 51% had sex with women, and 23% were IDU. A total of 5 individuals
who progressed to AIDS received clotting factor, 10 received a transplant, and 287 (2%)
received a transfusion. 7% of AIDS positive males identified themselves as having heterosexual
intercourse with an IDU and 9% reported that they have heterosexual intercourse with a person
living with HIV. Approximately 4% worked in a health care setting and 390 (2.75%) had a
history of N. gonorrheae between 2005 and 2011. In those that had not yet progressed to AIDS
positive status 65% reported they had sex with males, 44% reported they had sex with females,
and 16% were IDU. None of the non-AIDS positive HIV males received a transplant, 3 received
clotting factors and less than 1% received a transfusion. 6% reported that they have had
heterosexual intercourse with an IDU and 8% reported having heterosexual intercourse with a
female living with HIV. Of the non-AIDS positive HIV males less than 3% were health care
workers and 513 (9%) had a history of N. gonorrheae infection reported to the state. Table 3.4
97

shows the median age of progression event (AIDS or death) and the median time from HIV
diagnosis to a progression event.
Time to Death
Figure 3.1 shows the Kaplan-Meier curve for the survival rate of those with and without a history
of N. gonorrheae infection. Only 20 of those with gonorrhea died between 2005 and 2011 as
compared to 47.51% of those with no reported N. gonorrheae infections. The study found that
those with no history of N. gonorrheae infection had a median age at death of 41 years old as
compared to 38.5 years old in those with reported gonorrhea. Despite the low number of death
events in the gonorrhea group a multivariate analysis was ran to determine hazard of death in
those with gonorrhea. Table 3.4 shows the unadjusted hazard ratios for death in those with and
without a history of N. gonorrheae infection. Unadjusted HR for those with gonorrhea as
compared to those with no history of gonorrhea was HR=0.01 (95% CI 0.06-0.15) as compared
to those with no history of gonorrhea. Time from HIV diagnosis to death was also analyzed.
Figure 3.2 shows the Kaplan-Meier survival curve for time from HIV diagnosis to death in those
with and without a history of gonorrhea. The median time form HIV diagnosis to death was 3
years in those with no history of gonorrhea and 5 years in those who have had gonorrhea. Table
3.6 shows the unadjusted and adjusted HR for time from HIV diagnosis to death. The
multivariate model consisted of history of N. gonorrheae infection, ethnicity, MDCH condensed
risk group, and age at HIV diagnosis categorized by less than 20 years old, 20-29, 30-39, 40-49,
and 50 years old and older. History of a reportable N. gonorrheae infection was seen to have a
protective effect of death with a HRadj=0.06 (95% CI 0.04-0.09). The only racial grouping to
have an increasing effect was black with an increase in hazard of HRadj=1.05 (95% CI 1.00-1.10)

98

as compared to Whites. Both Hispanic and Multiracial/Other/Unknown had HRadj= 0.83 (95% CI
0.74-0.95) and HRadj=0.63 (95% CI 0.52-0.75) respectively. Those that identified as MSM/IDU,
IDU, or a blood recipient all had increased hazard of death as compared to MSM. IDU had a
HRadj=1.56 (95% CI 1.47-1.65), MSM/IDU condensed risk group individuals had a HRadj=1.23
(95% CI 1.14-1.33) and those that fell into the blood recipient condensed risk group had an
increase in HRadj=2.28 (95% CI 1.98-2.62). Both males that had heterosexual intercourse with a
female living with HIV and those that fell into the other/unknown condensed risk group had
HRadj=0.75 (95% CI 0.67-0.84) and HRadj=0.78 (95% CI 0.73-0.84) respectively. Age of HIV
diagnosis was also found to be a significant predictor of time to death. In those that were
diagnosed with HIV in their 20s as compared to those diagnosed before they turned 20
HRadj=2.17 (95% CI 1.72-2.73), those that were diagnosed in their 30s had a HRadj=2.86 (95%
CI 2.27-3.60), those that were diagnosed in their 40s had a HRadj=3.33 (95% CI 2.64-4.20) and
those that were diagnosed at 50 years or older had an increased HRadj=4.43 (95% CI 3.50-5.61)
as compared to those diagnosed with HIV prior to turning 20.
Time to AIDS Positive Status
Figure 3.3 shows the Kaplan-Meier time-to-event curve for age at AIDS positive status in those
with and without a reported history of N. gonorrheae infection. Median age at AIDS positive
diagnosis was seen to be 38 years old in those with no reported history of gonorrhea, and 31
years old in those with a history of N. gonorrheae infection. Table 3.7 shows the unadjusted HR
for age at AIDS positive status in those with and without N. gonorrheae infection. Unadjusted
HR of AIDS for males with a history of N. gonorrheae was HR=1.17 (95% CI 1.06-1.30) as
compared to those with no history of gonorrhea. Time from HIV diagnosis to AIDS positive
status was also analyzed.
99

Time from HIV to AIDS positive status in those with and without a history of N. gonorrheae
infection was also analyzed. The median time from HIV diagnosis to AIDS positive status was
less than 1 year in those with no history of gonorrhea and 1 year in those with gonorrhea. Figure
3.4 shows the Kaplan-Meier time-to-event curve for time from HIV diagnosis to AIDS positive
status in those with and without gonorrhea. A multivariate model was constructed to analyze
time from HIV diagnosis to AIDS positive status and includes history of gonorrhea, ethnicity,
MDCH condensed risk groups, and age at HIV diagnosis. Table 3.8 shows the unadjusted and
adjusted HR for progression to AIDS from time at HIV diagnosis. History of N. gonorrheae
infection was seen to be protective with a HRadj=0.63 (95% CI 0.57-0.70). Being Black as
compared to White slightly increased hazard of AIDS progression, HRadj=1.06 (95% CI 1.021.10). Only blood recipient was found to increase hazard as compared to falling into the MSM
condensed group, HRadj=1.26 (95% CI 1.10-1.43). Both risk groups of males who have had sex
with a HIV positive female and the group other/unknown were found to be protective as
compared to MSM, with HRadj=0.85 (95% CI 0.78-0.93) and HRadj=0.82 (95% CI 0.77-0.86)
respectively. Age at HIV diagnosis was also found to increase the hazard of AIDS progression
from HIV diagnosis with a similar dose-response relationship that was seen in the time from HIV
to death analysis. In those that were diagnosed with HIV in their 20s as compared to those that
were diagnosed before they turned 20, the HRadj=1.31 (95% CI 1.15-1.49), in those that were
diagnosed in their 30s HRadj=1.68 (95% CI 1.47-1.91), in those diagnosed in their 40s
HRadj=1.78 (95% CI 1.56-2.04) and in those that were diagnosed with HIV at age 50 or over
HRadj=1.97 (95% CI 1.71-2.26).
Time to AIDS Positive Status and Multiple N. gonorrheae Infections

100

History of multiple N. gonorrheae infections was also analyzed in relation to age AIDS positive
status. Figure 3.5 shows the Kaplan-Meier time-to-event curve for those with multiple N.
gonorrheae infections and progression to AIDS. Median age of AIDS positive status was seen to
be 38 years old as reported in the previous section. Those with only one reported N. gonorrheae
infection between 2005 and 2011 had a median age of AIDS positive progression of 32 years, in
those with two reported N. gonorrheae infection the median age was observed to be 29 years,
those with three N. gonorrheae infections the median age of AIDS positive status was 32 years
old, in those with four reported N. gonorrheae infections the median age of AIDS positive status
was 21, and in those with 5 or more N. gonorrheae infections the median age was observed to be
24 years old. Table 3.9 shows the unadjusted HR for multiple N. gonorrheae infections and age
at AIDS positive status. Unadjusted HR for one gonorrheal infection was HR=1.01 (95% CI
0.89-1.15), in males with two N. gonorrheae infections HR=1.45 (95% CI 1.17-.181) as
compared to those with no history of gonorrhea. In HIV positive males with three N. gonorrheae
infections HR=1.55 (95% CI 1.11-2.16), in males with four N. gonorrheae infections HR=2.16
(95% CI 1.30-3.59), and in HIV positive males with five or more N. gonorrheae infections
HR=2.47 (95% CI 1.54-3.97) as compared to those with no history of gonorrhea. Time from HIV
to AIDS in those with multiple N. gonorrheae infections was also modeled.
Time from HIV diagnosis to AIDS positive status was also looked at in those with a history of
multiple N. gonorrheae infections. Figure 3.6 shows the Kaplan-Meier time-to-event curve for
time from HIV diagnosis to AIDS positive status in individuals with a history of multiple N.
gonorrheae infections. The median time from HIV diagnosis to AIDS positive status in those
males without a reported history of gonorrhea was less than one year. In those with only one
reported event of gonorrhea the median time to AIDS positive status was 1 year, as was the
101

median time from HIV diagnosis to AIDS positive status in those with 2, 3, and 4 reported N.
gonorrheae infections. In those that had 5 or more N. gonorrheae infections the median time
from HIV diagnosis to AIDS positive status was 2 years. Table 3.10 shows the unadjusted and
adjusted HR for time from HIV diagnosis to AIDS in those with a history of multiple N.
gonorrheae infections. A multivariate model was constructed that included number of N.
gonorrheae infection, ethnicity, condensed risk group, and age of HIV diagnosis. In those with
only one N. gonorrheae infection the HRadj=0.60 (95% CI 0.53-0.68), in those with two N.
gonorrheae infections HRadj=0.68 (95% CI 0.54-0.85), those that had three N. gonorrheae
infections had a HRadj=0.71 (95% CI 0.51-0.99), in those with four reported N. gonorrheae
infections had a HRadj=0.72 (95% CI 0.44-1.20), and in those with five or more reported N.
gonorrheae infections the HRadj=0.58 (95% CI 0.36-0.94) as compared to those with no history
of gonorrhea. Blacks had a slightly higher HR than Whites, HRadj=1.06 (95% CI 1.02-1.10). As
compared to MSM only blood recipients had a higher HR, HRadj=1.26 (95% CI 1.10-1.43).
There was no statistical difference between IDU or MSM/IDU when compared to MSM,
HRadj=0.96 (95% CI 0.91-1.01) and HRadj=0.96 (95% CI 0.90-1.03) respectively. Male HCFR
and those that were categorized into the other/unknown risk group had a HRadj=0.85 (95% CI
0.78-0.93) and HRadj=0.82 (95% CI 0.77-0.86) respectively. Age was found to follow a dose
response with time. In those that were diagnosed in their 20s as compared to those that were
diagnosed before age 20 the HRadj=1.31 (95% CI 1.15-1.50), in those that were diagnosed with
HIV in their 30s the HRadj=1.68 (95% CI 1.47-1.92), in those diagnosed with HIV in their 40s
HRadj=1.79 (95% CI 1.56-2.04) and in those that were diagnosed in their 50s or above the
HRadj=1.97 (95% CI 1.72-2.26).
Post-HIV N. gonorrheae Time to AIDS
102

The study was also able to look at those who had post-HIV N. gonorrheae infections and analyze
the hazard of development to AIDS. Figure 3.7 shows the Kaplan-Meier time-to-event curve for
individuals with post-HIV gonorrhea and their time to AIDS positive status. The median time to
AIDS positive status in those individuals that did not have a post-HIV gonorrhea event (pre or
concurrent with HIV diagnosis only) was 29 years old. In those with a history of post-HIV
gonorrhea the median age of AIDS was 32 years old. Those that did not have a reported N.
gonorrheae infection had a median age at AIDS positive diagnosis of 38 years old. Table 3.11
shows the unadjusted HR for individuals with and without a post-HIV gonorrhea event.
Unadjusted HR in those with any history of post-HIV N. gonorrheae was HR=1.26 (95% CI
1.13-1.41) as compared to those with no history of gonorrhea.
Time from HIV diagnosis to AIDS positive status was also analyzed in those with and without a
post-HIV N. gonorrheae infection. Figure 3.8 shows the Kaplan-Meier time-to-event curve for
time from HIV to AIDS in those with a post-HIV N. gonorrheae infection. The median time
from HIV diagnosis to AIDS positive status in those with post-HIV gonorrhea was 2 years as
compared to less than 1 year for both those with no post-HIV gonorrhea and those without a
reported N. gonorrheae infection. Table 3.12 shows the unadjusted and adjusted HR for time
from HIV diagnosis to AIDS positive status in those with a reported post-HIV N. gonorrheae
infection. A multivariate model that adjusted for ethnicity, MDCH condensed risk group and age
of HIV diagnosis. As compared to those with no history of gonorrhea those with a post-HIV N.
gonorrheae infection had a HRadj=0.68 (95% CI 0.61-0.77). Blacks had a slightly higher hazard
than Whites, HRadj=1.05 (95% CI 1.01-1.09). As compared to those in the MSM condensed risk
group, blood recipients had a HRadj=1.27 (95% CI 1.11-1.45). There was no statistical
significance between MSM and those that either identified themselves as IDU or those that
103

identified as both MSM and IDU. Males that had sex with HIV positive females had a
HRadj=0.86 (95% CI 0.78-0.93) as compared to MSMs, and those that fell into the
other/unknown had a HRadj=0.82 (95% CI 0.77-0.86). Similar to the other time from HIV
diagnosis to AIDS positive status, there was a dose response with regard to age at HIV diagnosis.
In those that were diagnosed in their 20s as compared to those that were diagnosed with HIV pre
20s, the HRadj=1.34 (95% CI 1.17-1.53), in those that were diagnosed in their 30s the HRadj=1.73
(95% CI 1.52-1.97), in those that were diagnosed in their 40s the HRadj=1.84 (95% CI 1.612.10), and in those that were diagnosed at 50 years of age or older the HRadj=2.03 (95% CI 1.772.33).
Time to AIDS Positive Status and Multiple Post-HIV N. gonorrheae Infections
The study also looked at multiple post-HIV N. gonorrheae and hazard of progression to AIDS.
Figure 3.9 shows the Kaplan-Meier time-to-event curve for those with a history of multiple postHIV N. gonorrheae infections and age of AIDS positive status. The median age of AIDS positive
status in those with no history of N. gonorrheae infection was 38 years old. In those with
gonorrhea, but no post-HIV N. gonorrheae infections the median age of AIDS positive status
was 29 years old. Those that had only one occurrence of post-HIV gonorrhea had a median age
of AIDS positive status of 32 years, in those with 2 post-HIV N. gonorrheae infections the
median age of AIDS positive status was 30.5 years, and in those with 3 or more post-HIV N.
gonorrheae infections the median age of AIDS positive status was 26 years. Table 3.13 shows
the unadjusted HR for multiple post-HIV N. gonorrheae infections and hazard of AIDS
progression from time of HIV diagnosis. Unadjusted HR in males with a history of gonorrhea,
but no post-HIV gonorrhea was HR=0.95 (95% CI 0.77-1.17) as compared to those with no
history of gonorrhea. Unadjusted HR in males with only one post-HIV N. gonorrheae infection
104

was HR=1.10 (95% CI 0.95-1.27), in those with two post-HIV N. gonorrheae infections
HR=1.55 (95% CI 1.22-1.97), and in those with three or more post-HIV N. gonorrheae
infections HR=2.06 (95% CI 1.51-2.79) as compared to those with no history of gonorrhea. Time
from HIV to AIDS in those with multiple post-HIV gonorrheal infections was also analyzed.
The study also looked at multiple post-HIV N. gonorrheae and hazard of time from HIV
infection to AIDS. Figure 3.10 shows the Kaplan-Meier time-to-event curve for those with a
history of multiple post-HIV N. gonorrheae infections and the time from diagnosis of HIV to
AIDS positive status. The median time to AIDS positive status from HIV diagnosis in
individuals with no history of N. gonorrheae was less than 1 year. In individuals with gonorrhea,
but who did not have post-HIV N. gonorrheae infections the median time from HIV diagnosis to
AIDS positive status was also less than 1 year. For HIV positive males with one or two post-HIV
N. gonorrheae infections the median time to AIDS from HIV diagnosis was 2 years, and in those
who had 3 or more post-HIV N. gonorrheae infections the median time was 3 years. Table 3.14
shows the unadjusted and adjusted HR for multiple post-HIV N. gonorrheae infections and
hazard of AIDS progression from time of HIV diagnosis. A multivariate model was constructed
adjusting for ethnicity, MDCH condensed risk group, and age of HIV diagnosis. Those who had
no history of post-HIV gonorrhea had a HRadj=0.51 (95% CI 0.41-0.63) as compared to those
with no history of gonorrhea. In those with only on post-HIV N. gonorrheae infection
HRadj=0.64 (95% CI 0.55-0.74), HRadj=0.77 (95% CI 0.60-0.97) in those with two post-HIV N.
gonorrheae infections, and in those with 3 or more post-HIV N. gonorrheae infections
HRadj=0.72 (95% CI 0.53-0.98). Blacks had an increase in hazard as compared to Whites,
HRadj=1.06 (95% CI 1.02-1.09). Hispanics as compared to Whites had a HRadj=1.07 (95% CI
0.98-1.17) and those in the Multiracial/Unknown/Other ethnic grouping had a HRadj=0.94 (95%
105

CI 0.84-1.06). As compared to MSMs those in the Blood Recipient risk group had a HRadj=1.26
(95% CI 1.10-1.43). Both Male HCFR and those in the Other/Unknown risk groups were found
to have a protective effect (increase lag time from HIV diagnosis to AIDS positive status) as
compared to MSMs, HRadj=0.85 (95% CI 0.78-0.93) and HRadj=0.82 (95% CI 0.77-0.86)
respectively. Both IDU and MSM/IDU were not found to be significant as compared to MSMs,
HRadj=0.96 (95% CI 0.91-1.01) and HRadj=0.96 (95% CI 0.90-1.03) respectively. Age of HIV
diagnosis was found to follow a similar dose-response pattern as it did in previous analysis. As
compared to those that were diagnosed before they turned 20 years old, those diagnosed in their
20s had a HRadj=1.31 (95% CI 1.15-1.50), in those that were diagnosed with HIV in their 30s the
HRadj=1.68 (95% CI 1.48-1.92), in those diagnosed in their 40s HRadj=1.79 (95% CI 1.57-2.04),
and in those that were diagnosed at 50 years of age or older with HIV HRadj=1.97 (95% CI1.722.26).
3.5 Conclusion
Among the 903 HIV infected males in MI from 2005-2011 that were included in this study,
43.19% progressed to AIDS positive status by 2011 and only 2.21% died. In those that had no
history of gonorrhea, by 2011 73.45% progressed to AIDS positive status and 47.51% had died.
Time to Death
The median age of death in those that had no history of gonorrhea was 41 years old as compared
to 38.5 years old in those who had a reported N. gonorrheae infection between 2005 and 2011.
When looking at the unadjusted model history of gonorrhea seems to be protective with regard to
death even though the median age (in those that died) was lower in the gonorrhea group.
However, far more individuals that never had gonorrhea died (48%) than in those with a history
106

of gonorrhea (2%). The median time from HIV diagnosis to death in those with no history of
gonorrhea was 3 years as compared to those with a history of gonorrhea who had a median time
from HIV diagnosis to death of 5 years. This seemingly protective effect of N. gonorrheae
infection was also observed in the multivariate model. One of the most likely causes of this
paradox to the epidemiological synergistic model proposed by other researchers is that those
with reported N. gonorrheae infections may have had more exposure to the health care system,
and less time between seeing a medical provider than those who did not have any reportable N.
gonorrheae infections. This may possibly explain why in each time from HIV diagnosis analysis
having a history seemingly extends time to AIDS positive status, even though age at AIDS is
reduced in all versions of gonorrhea history. The study was able to show that one of the largest
factors in time from HIV diagnosis to death was age of HIV diagnosis, and that in all models it
has a dose-response effect of increasing hazard as age increases.
Blacks had a marginal increase in hazard (reduction in time from HIV diagnosis to death) as
compared to Whites. Both ethnic groups of Hispanic and those that fell into the
Multiracial/Unknown/Other had reductions in hazard (longer lags from HIV diagnosis to death)
as compared to Whites. Using the MDCH condensed risk group of MSM as the reference group
those that fell into the IDU, MSM/IDU, and Blood Recipients all had reduced lag times from
HIV to death. Those that fell into the Male HCFR and those in the Other/Unknown risk groups
had an increase in lag time from HIV diagnosis to death as compared to the reference group of
MSM.
Time to AIDS Positive Status

107

The median age of AIDS positive status in those that had no history of gonorrhea was 41 years as
compared to the median age in those with a history of N. gonorrheae infection which was 38.5
years old. The unadjusted hazard model also found that history of N. gonorrheae infection
increased hazard of AIDS. The median time from HIV diagnosis to AIDS positive status in those
with no history of gonorrhea was less than one year as compared to those with a history of
gonorrhea who had a median time between HIV diagnosis and AIDS positive status of one year.
Those of Black ethnicity were found to have an increase in hazard as compared to Whites. Both
Hispanic and Multiracial/Other/Unknown were not statistically significant with Hispanic
trending toward increase in hazard of AIDS positive status and Multiracial/Other/Unknown
trending toward a decrease in hazard as compared to Whites. Neither IDU nor MSM/IDU were
found to be statistically significant as compared to MSM, however, both trended toward
reduction in hazard. Blood Recipients were found to have an increase in hazard, and Male HCFR
and those in the Other/Unknown risk group were found to reduce hazard (increase age at AIDS
positive status) as compared to MSM. Age again was found to increase hazard of a progression
event, in this case AIDS positive status. There was a significant increase in hazard with each
increase of age at HIV diagnosis as compared to those who were diagnosed with HIV before they
turned 20.
Time to AIDS Positive Status and Multiple N. gonorrheae Infections
The median age at AIDS positive status in those with no history of N. gonorrheae infection was
observed to be 38 years old. Those with a single reported history of gonorrhea had a median age
of AIDS positive status of 32 years old which is 6 years earlier than in those without a history of
gonorrhea. Median age in those with two N. gonorrheae infections is 29, in those with three N.
gonorrheae infections the median age is 32 years, in those with 4 N. gonorrheae infections the
108

median age of AIDS positive status is 21 years and 24 years in those with 5 or more N.
gonorrheae infections. This trend in reduction in age of AIDS positive status shows that there is
a possible increase in the virulence of HIV in these males due to the presence of having N.
gonorrheae infections. Each infection may cause a spike in antibody production leading to
increase sites for HIV replication. Each N. gonorrheae infection represents a possible
intervention opportunity to get individuals to be aware that they are at risk for progressing to
AIDS positive status and an opportunity to determine where their viral load and CD4 levels are.
The unadjusted hazard analysis also showed that there was an increase in hazard of progression
to AIDS in individuals as their re-infection with N. gonorrheae increases. Time from HIV
diagnosis to AIDS positive status had similar results as it did with the binary history of N.
gonorrheae infection model, with there being a reduction in risk with the increase of N.
gonorrheae infections. Median time from HIV diagnosis to AIDS positive status in those with no
history of gonorrhea was less than 1 year as compared to those with 1-4 N. gonorrheae infections
having a median time to infection of 1 year and in those with 5 or more the median time from
HIV diagnosis to AIDS positive status was 2 years. The multivariate model also found similar
results with hazard decreasing in a linear trend with the increase of N. gonorrheae infections.
Blacks were found to have increased hazard of AIDS as compared to Whites. Neither IDU nor
MSM/IDU were found to significantly decrease hazard as compared to MSM. Those that fell
into the Blood Recipient group were found to have an increase in hazard of AIDS progression as
compared to MSM. Both Male HCFR and those in the Other/Unknown risk groups were found to
have reductions in hazard. As seen in the previous time from HIV models hazard increased with
increase in age of HIV diagnosis meaning that there was a shorter time from HIV to AIDS in

109

those that were diagnosed at an older age with HIV as compared to those that were diagnosed at
a younger age.
Post-HIV N. gonorrheae Time to AIDS
The median age of AIDS positive status in those that had no history of gonorrhea was 38 years
old as compared to those with post-HIV gonorrhea who had a median age of AIDS positive
status of 32 years. The unadjusted hazard model also showed an increase in hazard in those with
a history of post-HIV gonorrhea. For time from HIV to AIDS positive status those with no
history of gonorrhea had a median time from HIV diagnosis to AIDS positive status of less than
1 year as compared to those with a post-HIV N. gonorrheae infection who had a median time
from HIV to AIDS positive status of 2 years. The multivariate model shows this seemingly
protective relationship as well with hazard decreasing in those with a post-HIV N. gonorrheae
infection as compared to those with no history of infection. Blacks have a slight increase in
hazard as compared to Whites and Blood Recipients have an increase in hazard as compared to
MSMs. Neither IDU nor MSM/IDU were found to statistically reduce hazard as compared to
MSMs. Both Male HCFR and those in the Other/Unknown risk group had decreases in hazard. A
similar dose-response with regard to hazard and age of HIV diagnosis was also found in the
model with hazard increasing as age of HIV diagnosis increased.
Time to AIDS Positive Status Multiple Post-HIV N. gonorrheae Infections
The median age of AIDS positive status in those that had no history of gonorrhea was 38 years
old, and in those with a history of gonorrhea, but no history of post-HIV gonorrhea (gonorrhea
reported after HIV diagnosis) the median age of AIDS positive status was 29 years old, almost a
10 year difference. In those with a history of post-HIV N. gonorrheae the median age of AIDS
110

positive status does increase as compared to those with no post-HIV gonorrhea, but decreases
(increase in hazard) as compared to those with no reported history of N. gonorrheae infection.
The median age of AIDS positive status in those with only one post-HIV gonorrhea infection is
32 years, in those with 2 post-HIV infections the median age is 30.5 years old and in those with 3
or more post-HIV infections the median age of AIDS positive status is 26 years old, a 12 year
difference from those that remained gonorrhea free.
The unadjusted hazard model also showed an increase in hazard in those with post-HIV
gonorrhea. Hazard increased with occurrence of post-HIV N. gonorrheae infections as compared
to those with no history of N. gonorrheae infection. The median time from HIV diagnosis to
AIDS positive status in those with no history of gonorrhea was less than one year as well as in
those with a history of gonorrhea, but no post-HIV gonorrhea. The median time from HIV
diagnosis to AIDS positive status in those with one or two post-HIV N. gonorrheae infections
was 2 years and the median time between HIV and AIDS positive status in those with 3 or more
post-HIV N. gonorrheae infections was 3 years. The multivariate model also shows a reduction
in hazard for the increase in post-HIV N. gonorrheae infections. There was a slight increase in
hazard in Blacks as compared to Whites as well as in those that were in the risk group Blood
Recipients as compared to MSMs. Neither IDU nor MSM/IDU was found to be statistically
significant, and Male HCFR and those in the Other/Unknown were found to have a reduction in
hazard as compared to MSMs. Again a dose-response with hazard and age of HIV diagnosis was
seen in this model as well, with hazard increasing as age of HIV diagnosis increased.
Limitations

111

One of the major limitations of this study in that the interval of time limited the number of deaths
as well as the number of individuals that progressed to AIDS in the N. gonorrheae group. This
was most seen in the time to death modeling. Another limitation that hampered statistical
analysis was that all dates were only recorded as years. This had the greatest impact on time from
HIV diagnosis to AIDS positive status, but also did not allow us to use con-current
HIV/gonorrhea diagnoses in the data (HIV and gonorrhea reported in the same year). The use of
actual dates would have allowed for time from HIV to AIDS to be modeled in days not years
which may have given better distinction between the different break downs of history of
gonorrhea. Another major limitation is that the data available did not allow for the adjustment of
other bacterial or viral STDs during the progression to AIDS positive status. While this study is a
first step in understanding co-infection paradigms and in acknowledging that STDs, especially
HIV do not occur in vacuums to one another, this study did not have information on the
frequency of other STDs in the progression to AIDS or death.
It is also important to note that the data used in this study was obtained from a system designed
to collect data on reportable diseases, such as STDs and HIV and that it was not a study that was
expressly designed to capture the difference in time between HIV and AIDS. Surveillance data is
a very useful tool but it does have its limitations as compared to a true retrospective cohort study
or a case-control study. Data on ARV was not included in this section of the study, but was
included in a subgroup study done on HIV positive males with N. gonorrheae co-infection
between 2011 and 2013 in chapter 5. In this subgroup the question ever been on ARV, along
with the date range was reported and linked. Only 8% of individuals responded in the positive
(ever having been on ARV treatment) and among them 71% started the same year or after their
N. gonorrheae infection which may account for some of the paradoxical findings presented in
112

this study. It is also important to note that absence of a positive response does not constitute a
definite no (never having been on ARV treatment) and the data dictionary instructed to analyze
as “Yes” and “Not Yes”. This is an important distinction from an epidemiologic view because it
expresses that not all the missing or not “yes” are actually no (individuals who have never been
on ARV). While in the GEE models presented in chapter 5 ARV was never found to be
significant, it did have a major influence over both CD4 cell count and HIV viral load, two key
components of HIV disease progression.
The final major limitation is that we do not know how long a N. gonorrheae infection lasted in
an individual. Since these are all reported cases, we only capture N. gonorrheae infections
caught on a screening or due to treatment of symptoms that forced the male to get test, we have
no way of knowing the actual rate of N. gonorrheae infections in the population. This is true for
most STDs as they often, in men go undiagnosed and untreated.
Conclusion
This study was able to find that there was an increase in hazard (progression event at younger
ages) in those with a history of N. gonorrheae infection as compared to those with no reported
history. The study also found that despite a younger age of progression even, the time between
HIV diagnosis and a progression event (death or AIDS) was longer in those that had a history of
N. gonorrheae than in those with no history. The study also found that these relationships were
modified by repeated N. gonorrheae infections, and multiple N. gonorrheae infections were
found to both decrease the age of a progression event, and increase the time between HIV
diagnosis and a progression even such as AIDS or death. The study found that as multiple N.
gonorrheae infections, both total and post-HIV, occurred in an individual their hazard of AIDS

113

increased. Multiple N. gonorrheae infections also increased the lag time (reduced hazard)
between the diagnosis of HIV in an individual and their diagnosis as being AIDS positive. A
possible explanation may be that those with gonorrhea have more visits to a health care
professional and therefore may have better management over their HIV compared to individuals
that have HIV and then only are possibly seen at routine screenings. The data did not have the
reported usage of ART or HAART which may have also influenced the results.
More research is needed, especially looking into ARV rates in Michigan, as well as mechanisms
to increase or those that decrease the time from HIV to AIDS positive status. The chapter 5
subgroup, which only consisted of HIV positive males with a history of co-infection, did not
allow for testing of number of laboratory tested performed on an individual as a proxy for access
to health care due to these being no controls (HIV males with no history of N. gonorrheae coinfection. Further research using dates, aside from years is also need to better understand the
progression of laboratory markers of HIV progression pre, during, and post N. gonorrheae
infection.

114

APPENDIX

115

Table 3.1: Demographic Distribution of 19,647 HIV Positive MI Males (2005-2011)

Variable

No N. gonorrheae history
N=18744

N. gonorrheae history
N=903

7635 (40.73%)
9988 (53.29%)

177 (19.60%)
674 (74.64%)

Hispanic
Asian
American Indian

711 (3.79%)
62 (0.33%)
40 (0.21%)

23 (2.55%)
1 (0.11%)

Multiracial/Unknown/
Other

308 (1.64%)

28 (3.10%)

Married
Divorced

1317 (7.03%)
1268 (6.76%)

2 (0.22%)
2 (0.22%)

Single

4966 (26.49%)

10 (1.11%)

11193 (59.72%)

889 (98.45%)

11018 (58.78%)
2675 (14.27%)
1338 (7.14%)
268 (1.43%)
770 (4.11%)
2675 (14.28%)

697 (77.19%)
21 (2.33%)
21 (2.33%)
20 (2.21%)
144 (15.95%)

13767 (73.45%)
8906 (47.51%)

390 (43.19%)
20 (2.21%)

Ethnicity
White
Black

Marital Status

Other/Unknown/Widower/Missing
Risk Group
MSM
IDU
MSM/IDU
Blood Recipient
Male HCFR
Other/Unknown
AIDS Positive
Deceased

116

Table 3.2: Distribution of HIV Positive MI Males with and without AIDS in 2005-2011

Variable

HIV (Non-Stage 3)
N=5490

AIDS Positive
N=14157

2134 (38.87%)
3002 (54.68%)

5678 (40.11%)
7660 (54.11%)

Hispanic
Asian
American Indian

217 (3.95%)
21 (0.38%)
17 (0.31%)

517 (3.65%)
41 (0.29%)
24 (0.17%)

Multiracial/Unknown/
Other

99 (1.8%)

237 (1.67%)

Married
Divorced

93 (1.69%)
115 (2.09%)

1226 (8.66%)
1155 (8.16%)

Single

365 (6.65%)

4611 (32.57%)

4917 (89.57%)

7165 (50.64%)

3285 (59.84%)
547 (9.96%)
305 (5.56%)
38 (0.69%)
234 (4.26%)
1081 (19.69%)

8430 (59.55%)
2149 (15.18%)
1054 (7.45%)
230 (1.62%)
556 (3.93%)
1738 (12.28%)

Ethnicity
White
Black

Marital Status

Other/Unknown/Widower/Missing
Risk Group
MSM
IDU
MSM/IDU
Blood Recipient
Male HCFR
Other/Unknown

117

Table 3.3: Risk Group Distribution of HIV Positive and AIDS Positive MI Males (20052011)

Risk Group

HIV (No Progression to
AIDS)
N=5490
3590 (65.39%)

9484 (66.99%)

Sex with Females

2417 (44.03%)

7187 (50.77%)

Injection Drug Use (IDU)

852 (15.52%)

3203 (22.62%)

Received Clotting Factor

3 (0.05%)

5 (0.04%)

Received a Transfusion

48 (0.87%)

287 (2.03%)

Received a Transplant

-

10 (0.07%)

338 (6.16%)

1060 (7.49%)

Heterosexual Sex with a
Hemophiliac

5 (0.09%)

22 (0.16%)

Heterosexual Sex with a
Transfusion Recipient

6 (0.11%)

30 (0.21%)

Heterosexual Sex with a
Transplant Recipient

3 (0.05%)

10 (0.07%)

Heterosexual Sex with a Person
Living with HIV

432 (7.87%)

1236 (8.73%)

Worked in a Health Care Setting

143 (2.60%)

628 (4.44%)

History of N. gonorrheae

513 (9.34%)

390 (2.75%)

Sex with Males

Heterosexual Sex with an IDU

118

AIDS Positive
N=14157

Table 3.4: Median Age and Time from HIV Diagnosis to HIV Progression Events
N. gonorrheae History

No N. gonorrheae
History
N. gonorrheae History

Median Age of Progression
Event
Death
41 years old

Median Time from HIV Diagnosis to
Progression Event
3 years

38.5 years old

5 years

AIDS Positive Status
<1 year

No N. gonorrheae
History
N. gonorrheae History

38 years old
31 years old

1 year

1 N. gonorrhea
Infection
2 N. gonorrhea
Infections
3 N. gonorrhea
Infections
4 N. gonorrhea
Infections
5 or More N.
gonorrhea Infections

32 years old

1 year

29 years old

1 year

32 years old

1 year

21 years old

1 year

24 years old

2 years

No Post HIV
gonorrhea
Post HIV gonorrhea
2 Post HIV N.
gonorrheae infections
3 or more Post HIV N.
gonorrheae infections

29 years old

< 1 year

32 years old
30.5 years old

2 years
2 years

26 years old

3 years

119

Table 3.5: Unadjusted Analysis of Hazard of Death in MI Males from 2005-2011
Variable
Unadjusted HR
95% CI
History N.
gonorrheae
No Infection
1
History of gonorrhea 0.10
0.06-0.15
Ethnicity
White
1
Black
1.13
1.08-1.18
Hispanic
0.90
0.79-1.02
Multiracial/Unknown/ 0.68
0.57-0.82
Other*
Risk Group
MSM
1
IDU
1.27
1.20-1.34
MSM/IDU
1.25
1.16-1.35
Blood Recipient
1.75
1.52-2.01
ǂ
Male HCFR
0.70
0.62-0.79
Other/Unknown†
0.62
0.58-0.67
*
ǂ
Includes Asian/HI/PI AmIn/AkNat , Male who had sex with a female at risk for HIV,† Includes
Perinatal

120

Table 3.6: Multivariate Analysis of Hazard of Time from HIV to Death in MI Males with
and without a History of N. gonorrheae 2005-2011
Variable

Unadjusted HR

95% CI

Adjusted
HR

95%CI

History N.
gonorrheae
No Infection
1
1
History of gonorrhea 0.05
0.03-0.07
0.06
0.04-0.09
Ethnicity
White
1
1
Black
1.05
1.00-1.09
1.05
1.00-1.10
Hispanic
0.82
0.72-0.93
0.83
0.74-0.95
Multiracial/Unknown/ 0.58
0.49-0.70
0.63
0.52-0.75
*
Other
Risk Group
MSM
1
1
IDU
1.88
1.78-1.98
1.56
1.47-1.65
MSM/IDU
1.30
1.21-1.40
1.23
1.14-1.33
Blood Recipient
2.00
1.74-2.29
2.28
1.98-2.62
Male HCFRǂ
0.83
0.74-0.93
0.75
0.67-0.84
†
Other/Unknown
0.86
0.80-0.92
0.78
0.73-0.84
Age Group (Age of
HIV Diagnosis)
Less than 20 Years 1
1
Old
20-29 Years Old
2.17
1.72-2.72
2.17
1.72-2.73
30-39 Years Old
3.10
2.74-3.89
2.86
2.27-3.60
40-49 Years Old
3.82
3.04-4.80
3.33
2.64-4.20
Over 50 Years Old
4.89
3.88-6.16
4.43
3.50-5.61
*
ǂ
Includes Asian/HI/PI AmIn/AkNat , Male who had sex with a female at risk for HIV,† Includes
Perinatal

121

Table 3.7: Unadjusted Analysis of Hazard of AIDS Positive Status in MI Males with and
without a History of N. gonorrheae 2005-2011
Variable
Unadjusted HR
95% CI
History N.
gonorrheae
No Infection
1
History of gonorrhea 1.17
1.06-1.30
Ethnicity
White
1
Black
1.06
1.03-1.10
Hispanic
1.10
1.01-1.21
Multiracial/Unknown/ 1.01
0.90-1.13
Other*
Risk Group
MSM
1
IDU
0.73
0.69-0.76
MSM/IDU
0.95
0.89-1.01
Blood Recipient
1.07
0.94-1.22
ǂ
Male HCFR
0.75
0.68-0.81
Other/Unknown†
0.62
0.59-0.66
*
ǂ
Includes Asian/HI/PI AmIn/AkNat , Male who had sex with a female at risk for HIV,† Includes
Perinatal

122

Table 3.8: Multivariate Analysis of Hazard of Time of HIV Diagnosis to AIDS Positive
Status in MI Males with and without a History of N. gonorrheae 2005-2011
Variable

Unadjusted HR

95% CI

Adjusted
HR

95%CI

History N.
gonorrheae
No Infection
1
1
History of gonorrhea 0.56
0.50-0.62
0.63
0.57-0.70
Ethnicity
White
1
1
Black
1.00
0.96-1.03
1.06
1.02-1.10
Hispanic
1.02
0.93-1.11
1.07
0.98-1.17
Multiracial/Unknown/ 0.88
0.79-0.99
0.94
0.84-1.06
*
Other
Risk Group
MSM
1
1
IDU
1.09
1.04-1.15
0.96
0.91-1.01
MSM/IDU
1.01
0.94-1.07
0.96
0.90-1.03
Blood Recipient
1.51
1.01-1.31
1.26
1.10-1.43
Male HCFRǂ
0.91
0.84-1.00
0.85
0.78-0.93
†
Other/Unknown
0.87
0.83-0.92
0.82
0.77-0.86
Age Group (Age of
HIV Diagnosis)
Less than 20 Years 1
1
Old
20-29 Years Old
1.33
1.17-1.52
1.31
1.15-1.49
30-39 Years Old
1.73
1.52-1.96
1.68
1.47-1.91
40-49 Years Old
1.82
1.60-2.08
1.78
1.56-2.04
Over 50 Years Old
1.99
1.74-2.28
1.97
1.71-2.26
*
ǂ
Includes Asian/HI/PI AmIn/AkNat , Male who had sex with a female at risk for HIV,† Includes
Perinatal

123

Table 3.9: Unadjusted Analysis of Hazard of AIDS Positive Diagnosis in MI Males with
and without Multiple N. gonorrheae Infections 2005-2011
Variable
History N.
gonorrheae
No Infection
1 N. gonorrhea
Infection
2 N. gonorrhea
Infections
3 N. gonorrhea
Infections
4 N. gonorrhea
Infections
5 or More N.
gonorrhea Infections

Unadjusted HR

95% CI

1
1.01

0.89-1.15

1.45

1.17-1.81

1.55

1.11-2.16

2.16

1.30-3.59

2.47

1.54-3.97

124

Table 3.10: Multivariate Analysis of Hazard of Time from HIV Diagnosis to AIDS in MI
Males with and without a History of Multiple N. gonorrheae Infections 2005-2011
Variable

Unadjusted HR

95% CI

Adjusted
HR

95%CI

History N.
gonorrheae
No Infection
1
1
0.54
0.48-0.62
0.60
0.53-0.68
1 N. gonorrhea
Infection
0.60
0.48-0.74
0.68
0.54-0.85
2 N. gonorrhea
Infections
0.60
0.43-0.83
0.71
0.51-0.99
3 N. gonorrhea
Infections
0.61
0.37-1.01
0.72
0.44-1.20
4 N. gonorrhea
Infections
0.48
0.30-0.77
0.58
0.36-0.94
5 or More N.
gonorrhea Infections
Ethnicity
White
1
1
Black
1.00
0.96-1.03
1.06
1.02-1.10
Hispanic
1.02
0.93-1.11
1.07
0.98-1.17
Multiracial/Unknown/ 0.88
0.79-0.99
0.94
0.84-1.06
Other*
Risk Group
MSM
1
1
IDU
1.09
1.04-1.15
0.96
0.91-1.01
MSM/IDU
1.01
0.94-1.07
0.96
0.90-1.03
Blood Recipient
1.51
1.01-1.31
1.26
1.10-1.43
ǂ
Male HCFR
0.91
0.84-1.00
0.85
0.78-0.93
Other/Unknown†
0.87
0.83-0.92
0.82
0.77-0.86
Age Group (Age of
HIV Diagnosis)
Less than 20 Years 1
1
Old
20-29 Years Old
1.33
1.17-1.52
1.31
1.15-1.50
30-39 Years Old
1.73
1.52-1.96
1.68
1.47-1.92
40-49 Years Old
1.82
1.60-2.08
1.79
1.56-2.04
Over 50 Years Old
1.99
1.74-2.28
1.97
1.72-2.26
*
Includes Asian/HI/PI AmIn/AkNat ,ǂ Male who had sex with a female at risk for HIV,† Includes
Perinatal

125

Table 3.11: Unadjusted Analysis of Hazard of AIDS in MI Males with and without
Histories of Post-HIV N. gonorrheae Infections 2005-2011
Variable
History N.
gonorrheae
No Infection
Post-HIV History of
gonorrhea

Unadjusted HR

1
1.26

95% CI

1.13-1.41

126

Table 3.12: Multivariate Analysis of Hazard of Time from HIV Diagnosis to AIDS in MI
Males with and without a History of Post-HIV N. gonorrheae Infection 2005-2011
Variable

Unadjusted HR

95% CI

Adjusted
HR

95%CI

History N.
gonorrheae
No Infection
1
1
Post-HIV History of 0.60
0.54-0.68
0.68
0.61-0.77
gonorrhea
Ethnicity
White
1
1
Black
1.00
0.96-1.03
1.05
1.01-1.09
Hispanic
1.02
0.93-1.11
1.06
0.97-1.17
Multiracial/Unknown/ 0.88
0.79-0.99
0.94
0.84-1.05
Other*
Risk Group
MSM
1
1
IDU
1.09
1.04-1.15
0.97
0.92-1.02
MSM/IDU
1.01
0.94-1.07
0.97
0.91-1.03
Blood Recipient
1.15
1.01-1.31
1.27
1.11-1.45
ǂ
Male HCFR
0.91
0.84-1.00
0.86
0.78-0.93
Other/Unknown†
0.87
0.83-0.92
0.82
0.77-0.86
Age Group (Age of
HIV Diagnosis)
Less than 20 Years 1
1
Old
20-29 Years Old
1.33
1.17-1.52
1.34
1.17-1.53
30-39 Years Old
1.73
1.52-1.96
1.73
1.52-1.97
40-49 Years Old
1.82
1.60-2.08
1.84
1.61-2.10
Over 50 Years Old
1.99
1.74-2.28
2.03
1.77-2.33
*
Includes Asian/HI/PI AmIn/AkNat ,ǂ Male who had sex with a female at risk for HIV,† Includes
Perinatal

127

Table 3.13: Unadjusted Analysis of Hazard of Age at AIDS in MI Males with and without a
History of Multiple N. gonorrheae Infections 2005-2011
Variable
History N.
gonorrheae
No Infection
No Post HIV
gonorrhea
Post HIV gonorrhea
2 Post HIV N.
gonorrheae infections
3 or more Post HIV
N. gonorrheae
infections

Unadjusted HR

95% CI

1
0.95

0.77-1.17

1.10
1.55

0.95-1.27
1.22-1.97

2.06

1.51-2.79

128

Table 3.14: Multivariate Analysis of Hazard of Time from HIV Diagnosis to AIDS in MI
Males with and without Multiple N. gonorrheae Infections 2005-2011
Variable

Unadjusted HR

95% CI

Adjusted
HR

95%CI

History N.
gonorrheae
No Infection
1
1
No Post HIV
0.45
0.37-0.55
0.51
0.41-0.63
gonorrhea
Post HIV gonorrhea
0.60
0.50-0.67
0.64
0.55-0.74
2 Post HIV N.
0.67
0.53-0.85
0.77
0.60-0.97
gonorrheae infections
3 or more Post HIV
0.60
0.44-0.81
0.72
0.53-0.98
N. gonorrheae
infections
Ethnicity
White
1
1
Black
1.00
0.96-1.03
1.06
1.02-1.09
Hispanic
1.02
0.93-1.11
1.07
0.98-1.17
Multiracial/Unknown/ 0.88
0.79-0.99
0.94
0.84-1.06
Other*
Risk Group
MSM
1
1
IDU
1.09
1.04-1.15
0.96
0.91-1.01
MSM/IDU
1.01
0.94-1.07
0.96
0.90-1.03
Blood Recipient
1.15
1.01-1.31
1.26
1.10-1.43
ǂ
Male HCFR
0.91
0.84-1.00
0.85
0.78-0.93
Other/Unknown†
0.87
0.83-0.92
0.82
0.77-0.86
Age Group (Age of
HIV Diagnosis)
Less than 20 Years 1
1
Old
20-29 Years Old
1.33
1.17-1.52
1.31
1.15-1.50
30-39 Years Old
1.73
1.52-1.96
1.68
1.48-1.92
40-49 Years Old
1.82
1.60-2.08
1.79
1.57-2.04
Over 50 Years Old
1.99
1.74-2.28
1.97
1.72-2.26
*
Includes Asian/HI/PI AmIn/AkNat ,ǂ Male who had sex with a female at risk for HIV,† Includes
Perinatal

129

Figure 3.1: Kaplan-Meier Curves for Survival Rate in MI Males with and without N.
gonorrheae Infection 2005-2011

Trend: Logrank p <.0001

History of Gonorrhea

130

Figure 3.2: Kaplan-Meier Curves for Survival from Time of HIV Diagnosis in MI Males
with and without a History of N. gonorrheae 2005-2011

Trend: Logrank p <.0001

History of Gonorrhea

131

Figure 3.3: Kaplan-Meier Curves for Age of AIDS Positive Status in MI Males with and
without N. gonorrheae 2005-2011

Trend: Logrank p <.0001

History of Gonorrhea

132

Figure 3.4: Kaplan-Meier Curves for Time from HIV Diagnosis to AIDS Positive Status in
MI Males 2005-2011

Trend: Logrank p <.0001

History of Gonorrhea

133

Figure 3.5: Kaplan-Meier Curves for Age of AIDS Positive Diagnosis in MI Males with
Multiple N. gonorrheae Infections 2005-2011

Trend: Logrank p <.0001

Number of
N. gonorrheae
Infections

+

134

Figure 3.6: Kaplan-Meier Curve for Time from HIV Diagnosis to AIDS Positive Status in
MI Males with a History of Multiple N. gonorrheae Infections 2005-2011

Trend: Logrank p <.0001

Number of
N. gonorrheae
Infections

+

135

Figure 3.7: Kaplan-Meier Curves for Age at AIDS Positive Diagnosis in MI Males with
Histories of Post-HIV N. gonorrheae Infections 2005-2011

Trend: Logrank p <.0001

History of post-HIV
Gonorrhea

N

136

Y

Figure 3.8: Kaplan-Meier Curves for Time from HIV Diagnosis to AIDS Positive Status in
MI Males with a History of Post-HIV N. gonorrheae Infection 2005-2011

Trend: Logrank p <.0001

History of post-HIV
Gonorrhea

N

137

Y

Figure 3.9: Kaplan-Meier Curves for Age at AIDS Positive Diagnosis in MI Males with a
History of Multiple Post-HIV N. gonorrheae Infections 2005-2011

Trend: Logrank p <.0001

Post HIV N. gonorrheae

138

Figure 3.10: Kaplan-Meier Curves for Time from HIV Diagnosis to AIDS Positive Status in
MI Males with Multiple Post-HIV N. gonorrheae Infections 2005-2011

Trend: Logrank p <.0001

Post HIV N. gonorrheae

139

Chapter 4.
Geospatial Analysis of HIV associated N. gonorrheae Infections in Michigan Males 20052011
4.1 Abstract
Introduction
Sexual networks are a key factor in understanding the proliferation of sexually transmitted
diseases. The purpose of this study is to identify geographic boundaries and clusters at the county
level of HIV infected individuals with N. gonorrheae co-infection in Michigan from 2005-2011.
In order for public health to most effectively use its resources, prevention efforts must be
targeted at locations with the highest burden of disease.
Methods
A retrospective cohort of 19,647 HIV infected Michigan males from 2005-2011was linked to N.
gonorrheae infection data through collaboration with the Michigan Department of Community
Health. Geospatial analysis of HIV prevalence and N. gonorrheae incidence was constructed to
determine areas with increased burden of disease. Spatial clusters of N. gonorrheae, HIV, and
HIV/N. gonorrheae co-infection areas were also analyzed using both Bernoulli and Poisson
models.
Results
Detroit and the tri-county area of Macomb, Wayne, and Oakland were identified as potential hot
spots of N. gonorrheae infected HIV positive males. Bernouilli high rate detection analysis
identified Oakland County as the most likely cluster with a RR=1.44 (p-value=0.007). The
140

analysis also identified the secondary high rate cluster of the city of Detroit, Huron, Macomb,
Sanilac, and Lapeer counties had a RR=1.26 (p-value=0.036) as opposed to the rest of the state.
The final cluster consisted of cases in Gratiot and Clinton counties, RR=2.22 but was not found
to be statistically significant (p-value=0.974).
Conclusion
The identification of Detroit and the surrounding counties as hot spots for N. gonorrheae
infected HIV positive males suggests that the burden of disease due to the co-infection of these
two STDs is not distributed equally across the state of Michigan. These hotspots represent
potential areas for intervention and prevention programs to have the potential for greatest impact
on disease reduction in the state.

141

4.2 Introduction
Sexually transmitted diseases (STDs) do not exist in a vacuum in relation to each other.
Epidemiologic evidence suggests that there is a synergistic relationship between N. gonorrheae
and HIV causing atypical progression of the diseases in co-infected individuals. It is important
from a disease control standpoint identify areas that have higher than expected rates of coinfected individuals and to aim intervention efforts at these locations to help to prevent the
possible increase in disease burden due to individuals having both diseases. While geospatial
analysis is still an under-utilized epidemiologic tool for tracking and preventing the spread of
STDs, geographic boundaries have played an important role in our understanding of STDs. From
dating back to 1546 when Fracastoro wrote a poem that gave rise to calling the “French disease”
syphilis after the shepherd in the poem who insulted Apollo and was cursed with the disease to
the first MMWR on HIV that pointed out the initial cases being in Los Angeles with physicians
also seeing the disease in New York and San Francisco geography has played a vital role in
understanding STD transmission patterns (112, 1, 5). Geospatial analysis is an important tool in
understanding the geographic underpinnings of where diseases occur and where to focus
prevention and intervention efforts.
Despite the fact that N. gonorrheae is one of the most reported diseases in the United States, the
relationship between HIV and gonorrhea has yet to be fully explored and little attention has been
given to the geospatial relationship between the two diseases (15). In 2011 WHO estimated that
roughly 88 million of the estimated 448 million curable STD infections reported that year were
due to N. gonorrheae (12). In the United States in 2012, there were over 334,000 N. gonorrheae
cases resulting in an increased rate of 4.1% from the previous year (13). As of the 2010 Census
Michigan has the 8th largest population in the US. The January 2014 Annual HIV Surveillance
142

Analysis estimated that there were approximately 19,800 individuals in the state of Michigan
who were HIV positive of which an estimated 15,440 (78%) were males (10). The majority of
HIV diagnosis occurred in the Detroit Metropolitan Area (64%) despite only accounting for 43%
of the total Michigan population (10). Areas that are affected by poverty and that lack health
infrastructure are at a greater risk for spread of HIV and STDs, as well as have worse outcomes
associated with these diseases (31).
In a geospatial study investigating chlamydia clusters in networks of STDs, researchers
acknowledged that STDs are a behavior-based disease and that clusters represented distinct
transmission networks (unique chlamydial genotypes) within larger sexual networks (113). The
study found a total of 10 geographic clusters within the Canadian province of Manitoba (113).
Researchers found that clusters were characterized by the socioeconomic factors attributed to the
area with little to no geographic overlap of chlamydial genotypes (113). The study showed that
even within larger sexual networks, distinct sexual networks existed characterized by their
unique chlamydia genotype (113). A HIV study by Tanser et al. in 2009 found that in the rural
population of KwaZulu-Natal, South Africa that there was variation in local rates of HIV, with
the highest rates attributed to a three cluster zone along the National Road, with lower rate
clusters being located inland away from the National Road (114). The study is important because
it was able to show that in an area where HIV was believed to be “ubiquitous” there actually
existed localized epidemics that shared a unique geographic characteristic, their proximity to the
National Road (114). The authors go on to state that there needs to be a shift in the paradigm of
interventions aimed at rural South African communities, focusing not on entire population but on
targeted high risk socio-geographic communities (114).

143

A separate article by Wand and Ramjee also looking at KwaZulu-Natal, South Africa found
similar results when looking at sexually active women (116). The study area was slightly south
of the study area used in the Tanser et al. paper and found that hotspot of HIV tend to border
major roadways in the area, aside from the lowest cluster (RR of excess HIV=2.4) which
occurred in a more outlying area (116). This study along with Tanser et al. shows that even in
areas where disease is considered endemic, hotspots of disease occur and it takes proper use of
geographic spatial analysis to detect where resources need to be focused to reduce the burden of
disease in these areas (116).
A 2010 article by Westercamp et al. found opposite results when studying sexual risk behaviors
in young Kenyan men (115). The study did not find clusters, nor did it find an association
between residency and risk behavior (115). Results showed that STDs were evenly distributed
among urban and rural communities in the Kisumu district of Kenya (115). Even distribution of
STDs means that prevention strategies should be aimed at centralized areas as well as areas
where individuals are most likely gather. Like the Tanser et al. article this paper is important
because it shows that STD distribution is not a one size fits all paradigm and that in order to
understand the geospatial distribution of STDs in a community, researchers need to map disease
patterns and not rely on previous assumptions.
A study by Jennings et al. that mapped gonorrhea in Baltimore, MD states that STIs (STDs) are
not equally distributed across a population, for example the city of Baltimore and thus mapping
of disease clusters may give insight into disease trends that extend past sexual orientation, age
and race, and help to identify geographic boundaries associated with the disease after controlling
for race/ethnicity (69). The purpose of the study was to identify high risk transmission areas and
hopefully interrupt the transmission of the disease in these high risk areas (69). The study found
144

that in Baltimore the highest burden of disease occurred in the African American population and
that sexual networks in the Baltimore areas were heavily segregated based on race/ethnicity (69).
The study also found that N. gonorrheae infections tended to be focused around the center of the
city, and eight high transmission clusters were identified in the city, indicating that risk of N.
gonorrheae infection can be varied across geographic boundaries (69).
In another study of gonorrhea in Baltimore researchers focused on cases counts, rates, and
percent of repeaters in the Baltimore area and found that of the 6,108 gonorrhea cases recorded
in the 2 year study period, 9% were in individuals with multiple N. gonorrheae infections (68).
These individuals tended to be younger females, and the study found that those labeled as
repeaters tended to cluster geographically closer to each other than those that only had one
reported N. gonorrheae infection in the study period (68). Similar geospatial trends were found
in this study as were found in the Jennings et al paper, with the highest burden of disease focused
toward the center of the city (68).
There is a current need to explore the relationship between HIV and N. gonorrheae co-infections.
With the limited funding and resources available to control the spread of STI’s in sexual
networks the need for well-structured interventions are required to make the greatest impact. The
geographical analysis of STI co-infections will help to identify high risk populations where
resources should be spent to have the highest chance of success in disrupting the cycle of
transmission. This study aims to investigate the geographical distribution of clusters of male
individuals diagnosed with HIV with and without gonorrhea residing in the state of Michigan
between 2005 and 2011.

145

4.3 Methods
The Michigan Department of Community Health (MDCH) constructed a retrospective cohort of
HIV infected males in Michigan from 2005 to 2011. Records of reported N. gonorrheae
infections were linked by MDCH staff using LinkPlus. An employee of HIV Surveillance and
Body Art Unit HIV/STD/VH/TB Epidemiology created the record linkage structure between the
HIV datasets and the N. gonorrheae data sets to comply with patient confidentiality and privacy
(HIPPA 2014). All data used in this study were de-identified according to HIPAA guidelines on
public health information. Study staff also did not contact or attempt to contact any participants
of the study. All data collected by county and state employees prior to the start of the study.
Ethical approval was obtained from the Institutional Review Board (IRB) at Michigan State
University (MSU) and was reciprocal with the MDCH IRB.
Cases were defined as HIV positive males with a history of reported N. gonorrheae between
January 1st 2005 and December 31st 2011. HIV infected males with N. gonorrheae prior to 2005
were excluded from the study population. Controls were defined as all HIV positive males
without a history of N. gonorrheae during the study time period that were alive as of January 1st
2005. Records with missing first name, last name, or date of birth were excluded from the study
population prior to the matching of the data. A final dataset of 19,647 HIV infected males in the
State of Michigan were included in this study.
HIV infection was defined as HIV-NA=HIV infection, non-stage 3 and AIDS=stage 3 HIV
infection. Acute and Latent, stage 1 and 2, were not differentiated in this study. Ethnicity was
classified into unique non-over lapping groups, 1.White= White non-Hispanic, 2.Black= Black
non-Hispanic, 3.Hispanic=Hispanic of any race, 4.Asian/HI/PI=Asian/Hawaiian, Pacific Islander

146

non-Hispanic, 5.Am In /Ak Nat=American Indian, Alaskan Native, non-Hispanic, or
9.Multi/Unk/Other=Multirace, Unknown, Other, non-Hispanic. Due to low rates of
Asian/Hawaiian, Pacific Islanders (N=0) and American Indian, Alaskan Native (N=1) in the N.
gonorrheae group these racial groups were included in the 9.Multi/Unk/Other grouping to avoid
analytical complications associated with extremely small numbers. Marital status while largely
missing was categorized as A=Married and Separated, S=Single and never married, M=Married,
N=Not otherwise specified, Missing or U=Unknown, W=Widowed; Risk behaviors were
condensed and categorized as A.MSM=Male-Male Sex, B.IDU=Injecting Drug Use,
C.MSM/IDU=Male-Male Sex and Injecting Drug Use, D.Blood Recipient=Received Blood,
E1.(Male HCFR)=Male who had sex with a female at risk for HIV, F.Perinatal=Child exposed
by mother (further condensed to risk group G.Unk:Other due to low response, N=3),
G.Unk:Other=Unknown/Other.
Descriptive statistics were preformed to assess the distribution of demographic and risk
characteristics using SAS 9.3 (SAS Institute, Inc. Cary, North Carolina). The MDCH provided
dataset did not contain any cases or controls from Keweenaw County. The geospatial analysis
was performed using the 83 counties in the State of Michigan as well as the city of Detroit. Data
collected from MDCH also contained individuals who were either unknown or labeled as
“prison” diagnosed. These unknown and prison diagnosed were removed from the geospatial
analysis (n=1,097). The State of Michigan has over 30 state level prisons located in various
counties throughout the state and one federal prison in the City of Milan. Chloropleth maps (map
that uses colors, shading, or symbols within geographic boundaries to represents differences
between areas) were created for frequency of N. gonorrheae, rate of HIV per 100,000; rate of N.
gonorrheae per 1,000 and percent multiple gonorrhea, using ArcGIS v.10.2 (Environmental
147

Systems Research Institute (ESRI) 2014). The rates per 100,000 and 1,000 were calculated using
the 2006-2008 American Community Survey (ACS) data where available, and 2012
Demographic and Housing data when counties were not included in the ACS. The population of
Wayne County was adjusted by removing people who resided in the City of Detroit so both
geographies could be mapped independently. Spatial cluster-analyses were performed using two
methods: 1) Getis-Ord statistics to access global and local High and Low clusters of prevalence
rates in ArcMap 10; and 2) Bernouilli (0/1) and Poisson distributed methods using SaTScan
V9.1.1 to detect high and low prevalence rates. Getis-Ord Gi Hot Spot analysis is a cluster
detection analysis that measures each area in context to its neighbors. If a map unit (county or
Detroit) has high values that are contiguous to its adjoining neighbors then the area constitutes a
‘hot spot’. The cluster is then compared to all map units and a Z score is calculated. Clusters with
higher rates of disease will have higher Z scores and clusters with lower than average rates of
disease have lower Z scores. The GiZ scores were mapped at the county level along with the city
of Detroit. A GiZ score of 0 represents a map unit that does not differ from the overall mean. An
increase in GiZ score represents greater distance from the overall mean (increase in N.
gonorrheae/HIV co-infection). A negative GiZ score represents areas with a lower than average
rate of N. gonorrheae/HIV co-infection.
In SaTScan the Bernouilli method cases were defined as HIV positive males with a history of N.
gonorrheae infection between 2005 and 2011 and controls were defined as HIV positive male
between 2005 and 2011 with no reported history of N. gonorrheae infections. Bernouilli high
only analysis was conducted to best identify areas with the greatest need for prevention efforts
which include screening, treatment and education. For the Poisson method the cases were
defined as described for the Bernoulli method and population data were gathered from the 2006148

2008 American Community Survey (ACS) data where available, and 2012 Demographic and
Housing data when counties were not included in the ACS (available online at
http://factfinder2.census.gov/faces /nav/jsf/pages /index.xhtml) at the county level. Coordinate
data was constructed using ArcGIS via the “Calculate Geometry” function, and mapped as
latitude and longitude. Bernouilli and Poisson cluster-detection methods consist of a gradually
scanning window that moves across the study frame (county and city of Detroit centroids)
calculating the number of observed and expected observations (cases) inside the scanning
window of each map unit. Maximum likelihood estimates are calculated to determine areas
where clustering was least likely (p-value less than 0.05) to be due to chance. SaTScan has the
ability to document and rank multiple possible clusters of high rates.
County population data was adjusted for males 0-65 years old from the 2006-2008 American
Community Survey (ACS) data where available, and 2012 Demographic and Housing data when
counties were not included in the ACS. (available online at http://factfinder2.census.gov/faces
/nav/jsf/pages /index.xhtml). HIV positive males with a history of gonorrhea before 2005 were
excluded from the study as well as those with missing first name, last name, or date of birth due
to record linkage issues. Geospatial analysis of HIV and N. gonorrheae clusters were done using
Poisson method scanning for high rate areas. No time precision options were used and map areas
were scanned for high rates with no clusters centered over other clusters.
4.4 Results
Between 2005 and 2011 903 HIV positive males in the State of Michigan had a reported N.
gonorrheae infection. This accounted for 1,588 total N. gonorrheae infections in the study
period. Females were not included in the study due to low rates of HIV positive females with a

149

history of N. gonorrheae between 2005 and 2011 in the State of Michigan. Demographic
information for both HIV males with and without gonorrhea can be seen in Table 4.1. In HIV
infected males with a history of gonorrhea, 117 (19.60%) were White, 674 (74.64%) were Black,
23 (2.55%) were Hispanic, 1 (0.11%) was American Indian, and 28 (3.10%) reported as
Multiracial/Unknown/Other. The majority of N. gonorrheae cases did not report marital status
889 (98.45%). In the distribution of MDCH defined risk groups for HIV infected males with a
history of N. gonorrheae infection 697 (77.19%) of males were MSM, 21 (2.33%) IDU, 21
(2.33%) were MSM/IDU, 20 (2.21%) of HIV infected males reported having sex with a female
at risk for HIV, and 144 (15.95%) had other or unknown risks. Of the 903 HIV positive males
that had a reported history of N. gonorrheae infection 390 (43.19%) progressed to AIDS/Stage 3
HIV positive status and 20 (2.21%) died. 37% of HIV positive men in Michigan with gonorrhea
had more than one N. gonorrheae infection from 2005-2011.
There were 18,744 HIV positive males that were alive at the start of 2005, or diagnosed between
2005 and 2011 that did not have a reported case of gonorrhea. In HIV positive males with no
history of N. gonorrheae infection 7,635 were White (40.73%), 9,988 were Black (53.29%), 711
were Hispanic (3.79%), 62 identified as Asian (0.33%), 40 identified as Native American
(0.21%) and 308 (1.64%) were grouped into the Multiracial/Missing/Unknown ethnicity
category. Response rate for marital status was slightly higher than in HIV men with gonorrhea,
with 1,317 (7.03%) responding as married, 1,268 (6.76%) as Divorced, 4,966 (26.49%) as Single
and 11,193 (59.72%) falling into the Other/Unknown/Widower/Missing Marital status category.
The distribution of MDCH condensed risk groups showed that MSM account for the majority
(58.78%) of HIV positive males with no history of gonorrhea. There were 2,675 individuals that
identified as IDU (14.27%) and 1,338 that identified as both MSM and IDU (7.14%). Males that
150

have had sex with a female living with HIV accounted for 4.11% of HIV positive males with no
history of gonorrhea and there were a total of 268 (1.43%) Blood Recipients and 2,675 (14.28%)
that fell into the Unknown/Other MDCH condensed risk group.
Figure 4.1 shows the geospatial distribution of the rate of HIV infected males per 100,000 in the
State of Michigan between 2005 and 2011. Areas with the highest rates of infection (82-1,032
per 100,000) tend to be in the southern part of the state with the exception of Baraga County in
the northern part of the Upper Peninsula. Detroit along with the tri-county area that surrounds it
all fall into the highest rate per 100,000 group for HIV infected males. Other high counties
include Genesee (Flint, MI), Saginaw County, Ingham, Kent and Washtenaw counties. Rates
tend to fall off further north in the state and become unstable. County population data was
adjusted for males 0-65 years old from the 2006-2008 American Community Survey (ACS) data
where available, and 2012 Demographic and Housing data when counties were not included in
the ACS. 47 counties were found to have unstable rates, cases less than 20 or population less
than 50.
Figure 4.3 shows the geospatial distribution of gonorrhea cases in HIV positive males in the
State of Michigan from 2005 to 2011. No data was available (either no cases or missing) for
Keweenaw county in the Northern tip of the Upper Peninsula. While this chloropleth map does
not show clusters, the areas with the highest frequency of gonorrhea cases in HIV positive males
are the City of Detroit, Oakland County, and Wayne County (minus Detroit). Other high
frequency counties are Macomb, Genesee, Washtenaw, Ingham and Kent. A slightly different
trend emerges in Figure 4.4 that maps gonorrhea/HIV positive males per 1,000. Rate per 1,000
shows the highest counties to Oakland, Ogemaw, Roscommon, Mecosta, Gratiot, Clinton, Eaton,
and Sanilac. Detroit and Macomb County fell into the 45-58 per 1,000 along with 6 other
151

counties. 24 counties were found to have unstable rates, cases less than 20 or population less than
50.
The next series of maps display the rates of repeated N. gonorrheae infection in HIV positive
male between 2005 and 2011. Figure 4.6 shows the percent of HIV individuals with N.
gonorrheae who had multiple re-infections of the bacteria between 2005 and 2011. While Detroit
and the tri-county area have some of the highest counts of HIV positive males infected with N.
gonorrheae they do not fall into the highest division of percent of repeatedly infected HIV males.
Low incident counties like Grand Traverse, Lenawee, Calhoun, Clinton, and Gratiot have fewer
individuals getting infected with N. gonorrheae, but these individuals are more prone to being reinfected. While Detroit, Macomb, Wayne, and Oakland do not fall into the highest infection rate
group, it is important to note that their number of individuals of repeat N. gonorrheae cases are
higher. This represents two distinct pathways to intervention. In counties with low total
gonorrhea, but high reinfection rates prevention strategies aimed at improvement of partner
services may have the greatest impact, where as in areas with high incidence of gonorrhea and
mid to high re-infection rates, prevention strategies may need to target sexual networks and focus
on partner tracking methods. Figure 4.7 shows the rate of multiple N. gonorrheae infections in
HIV positive males. This map shows that the overall percent of HIV infected males who are
multiple N. gonorrheae infections is low with most counties that have re-infected individuals at
less than 3% of total HIV infected males.
Cluster analysis was done in multiple phases using three different techniques to try and isolate
potential hotspots of HIV/N. gonorrheae co-infections. The first cluster analysis performed used
Getis-Ord Gi Hot Spot detection to identify areas of HIV/N. gonorrheae co-infections. Figure 4.8
shows the standard deviation of Getis-Ord GiZ scores for N. gonorrheae infected HIV positive
152

males. Four counties had GiZ scores whose standard deviations greater than 2.58 which were St.
Clair, Shiawassee, Clinton and Gratiot. Areas that had standard deviations between 1.96 and 2.58
included Macomb, Oakland, Lapeer, Isabella, and Midland represent the second set of hot spots.
The standard deviation grouping included the city of Detroit, Wayne, Ingham, Saginaw,
Montcalm, Gladwin, and Iosco counties which had standard deviations between 1.65 and 1.96.
Two distinct clusters were detected, one consisted of eight counties around Gratiot and Clinton,
and the other being located around Detroit and the Tri-County area. No areas of low N.
gonorrheae infections in HIV positive males were detected using the Getis-Ord GiZ scores.
The second stage consisted of independently mapping HIV and N. gonorrheae clusters using the
discrete Poisson method. Figure 4.9 shows the Poisson cluster analysis of HIV positive males in
Michigan. This cluster analysis used high onlyrate analysis, with a limit radius of 20 miles
(32.2km) and a maximum spatial cluster size of 50% of the population at risk. Detroit was found
to be the most likely cluster and had an increase in relative rate (RR=10.26 p-value <0.0001).
The City of Detroit represented the only high risk cluster identified. Figure 4.10 shows the
Poisson cluster analysis of N. gonorrheae in HIV infected males in Michigan from 2005 to 2011.
A similar high only rate Poisson analysis was performed for N. gonorrheae. Detroit represented
the most likely cluster and as with the HIV cluster analysis the only area with an increased
relative risk was Detroit, RR=12.66 (p-value <0.0001).
The final cluster analysis performed used Bernouilli methods of N. gonorrheae infected HIV
males in Michigan. Figure 4.11 shows the high only Berouilli analysis of N. gonorrheae infected
HIV males in Michigan. Oakland County represented the most likely cluster with a RR=1.44 (pvalue=0.007). A secondary cluster consisted of the city of Detroit, Huron, Macomb, Sanilac, and
Lapeer county had a RR=1.26 (p-value=0.036) and was also statistically significant. A third
153

cluster consisting of Clinton and Gratiot counties was also detected but was not found to be
statistically significant (p-value=0.974).
4.5 Conclusion
Geospatial analysis at the county level shows that most of the burden of disease of HIV is
centered around Detroit and the tri-county area of Macomb, Wayne, and Oakland counties. HIV
rates per 100,000 were also elevated across the southern part of the state with a high rate per
100,000 extending from Detroit and the tri-county area to Allegan, Kent, and Kalamazoo
counties on the Western edge of the state. There was also a high rate county, Baraga, located in
the Upper Peninsula.
HIV cluster analysis using the Poisson methods shows that the only real hotspot in the state is the
City of Detroit when using high only rate detection methods. The use of two geospatial analysis
techniques, rate and cluster detection, allow for epidemiologists and public health officials to
address the HIV problem at multiple fronts. While the City of Detroit represents the only hot
spot in the state for HIV, HIV rates are increased in multiple other counties including the
surrounding tri-county area and across the Southern part of the state. These counties represent
areas of concern and intervention efforts should be targeted at the county level as well as in the
City of Detroit.
The frequency of HIV positive men with a history of N. gonorrheae infection is highest in
Detroit and the tri-county area and decreases as it moves toward the center and Western part of
the state. This may be largely due to the higher population of HIV infected males in the area and
may not represent a complete picture of burden of disease in these areas. N. gonorrheae infected
per 1,000 in HIV positives males gives a slightly better picture of where the real burdens of
154

disease occur in Michigan. The N. gonorrheae infected per 1,000 HIV positive males shows that
while Detroit had some of the highest occurrences of gonorrhea in the state, it does not have the
corresponding highest rate. Rates of N. gonorrheae HIV infected males decrease as they move
west from Detroit and the tri-county area and are noticeably low in the northern part of the state
aside from Roscommon and Ogemaw counties.
Rates of multiple N. gonorrheae were analyzed by those that had at least one reported N.
gonorrheae infection between 2005 and 2011 and by total number of HIV positive males. Both
maps show relatively similar results with Detroit and the tri-county area being lower than
expected for rates of multiple N. gonorrheae infections. These maps are interesting for a number
of reasons, mainly that while Detroit, Wayne County, Macomb County, and Oakland County fall
in the middle of the rate distributions some of the smaller counties like Gratiot, Clinton, and
Lenawee have high rates of individuals with repeat N. gonorrheae infections. These areas
represent where there may be greatest potential and need for partner based services to try and
prevent individuals from being re-infected.
Figure 4.3 along with Figure 4.4 show that one of the major areas of concern when investigating
N. gonorrheae/HIV co-infections in MI males is the tri-county area and the City of Detroit due to
their increased frequency of N. gonorrheae infection as well as their increased rates of N.
gonorrheae/HIV positive males. These two maps offer different ways of looking the
gonorrhea/HIV problem from a public health stand point. While areas like Detroit and Wayne
County have some of the highest raw numbers of HIV positive males with a history of
gonorrhea, other counties have higher rates of the disease and may see their need for prevention
funding as great as Detroit and Wayne County.

155

Geospatial analysis and cluster detection show that the City of Detroit and the tri-county area
represent high risk areas of HIV/N. gonorrheae co-infections. Using three different cluster
detection approaches this study was able to identify possible hot spots of HIV/N. gonorrheae coinfections in the State of Michigan. As stated previously Poisson based cluster detection found
that there was only one true hot spot for N. gonorrheae/HIV co-infections in MI males. From this
analysis prevention should be focused on Detroit and counties in close proximity to the City of
Detroit. A second cluster technique used Getis-Ord GiZ scores to determine where there was an
increase in rate of co-infections and found that there appeared to be another possible grouping
that was not flagged on the Poisson method approach. The Getis-Ord approach revealed that
similar to the Poisson approach that Detroit, the tri-county area, and counties in close proximity
to Detroit all had increased rates as seen by their GiZ score, but the analysis also revealed that
there may be another smaller cluster near the center of the state focusing around Clinton, Gratiot,
and Shiawassee counties. This is important because it shows that the burden of disease is not just
limited to Detroit and the surrounding areas, that there are increased rates of N. gonorrheae/HIV
co-infections in other parts of the state as well forming their own unique little cluster.
The final cluster detection approach was the Bernouilli method. In the Bernouilli analysis, using
only high rate detection three hot spots were identified. Oakland County again had its own high
rate cluster, but the secondary cluster of Detroit, Macomb, Lapeer, Sanilac, and Huron counties
was statistically significant this time representing a unique cluster as compared to Oakland and
the rest of the state. The final cluster detected that was not found to be statistically significant
was that of Gratiot and Clinton county which were two of the counties identified in the Getis-Ord
technique that gave each individual map unit a GiZ score as compared to the Poisson and
Bernouilli approaches that relied on neighbor level data as well.
156

These maps show that there is unequal distribution across Michigan with regard to N.
gonorrheae infection among HIV positive males and this is primarily centralized among Detroit
and the neighboring counties. The purpose of this study was to identify areas where there is an
increase in N. gonorrheae among HIV infected males since it has been hypothesized that risk of
transmitting both diseases is increased in those with co-infections. This study identifies possible
clusters in the state of Michigan and helps both epidemiologists and public health officials plan
targeted interventions to stop the spread of these diseases.
Limitations
A major limitation was that only county of diagnosis, Detroit, Prison, or Unknown were reported
so all inferences are applied to the county level, even though the entire county may not be at the
same risk. Further studies should seek to gain smaller geographical units, such as zip codes or
census block level data to explore the possibility of unequal distribution among the counties,
specifically among those in the hot spots. This would require alternate approaches to analysis
including automated zone mapping to censor units with low numbers. Prisons could also be
mapped to represent their locations. Given the high number of prison diagnosed HIV cases, it
would be important to know if there is equal distribution across all prisons in the system, or if
prisons have the same rates as the county they are located in. Another limitation is that location
can only be reported as location of residence at time of diagnosis, and not residence at time of
infection or location at infection. Due to the nature of STDs, their social underpinnings, and the
difficulties in not only identifying from whom an infection may have spread, but when infection
occurred, only residence at diagnosis can be measured by public health. It also remains a key
limitation in not only the spatial analysis of STDs, but also in the epidemiological study of them
as well. A final limitation of this study is that N. gonorrheae is often asymptomatic in males and
157

aside for a handful of strains that develop into systemic infections, Disseminated Gonococcal
Infection (DGI), infections can resolve without treatment (1, 2, 11). Due to this there is a
possibility that some N. gonorrheae infections go undiagnosed and unreported skewing the
actual rate and rate of re-infections to be lower than they truly are in a community. This could
possibly be negated with an increase in men’s health initiatives aimed at routine testing for all
sexually active males especially in males who fall into high risk groups.

158

APPENDIX

159

Table 4.1: Demographic Distribution of 19,647 HIV Positive MI Males (2005-2011)

Variable

No N. gonorrheae history
N=18744

N. gonorrheae history
N=903

7635 (40.73%)
9988 (53.29%)

177 (19.60%)
674 (74.64%)

Hispanic
Asian
American Indian

711 (3.79%)
62 (0.33%)
40 (0.21%)

23 (2.55%)
1 (0.11%)

Multiracial/Unknown/
Other

308 (1.64%

28 (3.10%)

Married
Divorced

1317 (7.03%)
1268 (6.76%)

2 (0.22%)
2 (0.22%)

Single

4966 (26.49%)

10 (1.11%)

11193 (59.72%)

889 (98.45%)

11018 (58.78%)
2675 (14.27%)
1338 (7.14%)
268 (1.43%)
770 (4.11%)
2675 (14.28%)

697 (77.19%)
21 (2.33%)
21 (2.33%)
20 (2.21%)
144 (15.95%)

13767 (73.45%)
8906 (47.51%)

390 (43.19%)
20 (2.21%)

Ethnicity
White
Black

Marital Status

Other/Unknown/Widower/Missing
Risk Group
MSM
IDU
MSM/IDU
Blood Recipient
Male HCFR
Other/Unknown
AIDS Positive
Deceased

160

Figure 4.1: Map of Number HIV Positive Males per 100,000 in the State of Michigan, 20052011

*
*

*

*

*
*

*
*

*

*

*

*
* *

*

Legend
HIV Rates per 100,000
Detroit
Counties
HIV per 100,000
32-85
85-139
139-245
245-435
*435-2589
- Unstable Rate
*- Unstable Rate
No Data Available

*
*

*
*
*

*

* *
* *

*
*

* * *
* * *
*
* *
* * *
*
*
*
*

*
*

*

*

HIV rate per 100,000 was calculated from 2006-2008 ACS and 2012 Demographic and Housing
estimates: http://factfinder2.census.gov/faces/nav/jsf/pages /index.xhtml. County population was
adjusted for males 0-65 years old. Detroit and the tri-county area of Macomb, Wayne, and
Oakland, along with most of the counties in the lower half of the state all have relatively high
rates of HIV, as compared to the northern section of the Lower Peninsula and the Upper
Peninsula, with the exception of Baraga County (however rate is considered unstable). Counties
marked with an “*” represent areas with unstable rates (less than 20 cases of HIV or population
less than 50). A total of 18,550 HIV positive males were able to be geocoded at the county level.

161

Figure 4.2: RSE (Root Mean Square Error) of Number HIV Positive Males per 100,000 in
the State of Michigan, 2005-2011
*
*
*

*
*

*

*
*

*

*

*

*
*

*

Legend
RSE HIV per 100,000
0-10
10-20
20-40
40-60
60-100
*- Unstable Rate
Detroit
Counties
No Data Available

* *
*
* * * *
* * * *

*
*
*

* *

* *

*

*

* *
*

*

*

*
*

*

*

*

*

*
*
*

*

*
*
*

*
*
*

*

*

*

Root Mean Square Error (RSE) of rate of HIV per 100,000. Calculated from 2006-2008 ACS
and 2012 Demographic and Housing estimates: http://factfinder2.census.gov /faces/nav/jsf
/pages /index.xhtml. County population was adjusted for males 0-65 years old. RSE was
calculated by Standard Error/Rate*100, Standard Error=Rate/Square Root (Cases). Counties
marked with an “*” represent areas with unstable rates (less than 20 cases of HIV or population
less than 50).

162

Figure 4.3: Map of the Number N. gonorrheae cases in HIV Positive Males from 2005-2011
in the State of Michigan.

Legend
Number of Gonorrhea
Detroit
Counties
Gonorrhea Cases
1-20
20-47
47-82
82-140
140-415
No Data Available

Number of N. gonorrheae represents individuals and not unique infections of N. gonorrheae.
Detroit and the tri-county area of Macomb, Wayne, and Oakland all have increased HIV positive
N. gonorrheae infected individuals than the rest of the state. The majority of cases seem to be
located near Detroit, with numbers of individuals reducing the further away from Detroit a
county is. A total of 879 HIV positive males with gonorrhea were able to be geocoded

163

Figure 4.4: Map of Rate of N. gonorrheae per 1,000 in HIV Infected Males in the State of
Michigan 2005-2011 Adjusted Population Density

Legend
NG per 1,000 HIV
Detroit
Counties
NG per 1,000 HIV Males
0
0-28
28-45
45-58
58-250
No Data Available
*
-Unstable
*- UnstableRate
Rate

*
*

*

*

*

*

* *
*
*

*
*
*

*

*

*

*

*

*

*

*

*

*

*

*

*
*

*

*
*

*

*

*

*

*

*
*

*

*

*

*

*
*

*

*

*

*

*

*

*

*

Rate of N. gonorrheae in HIV positive males was calculated using the 2010 US census at:
http://factfinder2.census.gov/faces/nav/jsf/pages /index.xhtml. Oakland County, along with
several others appears to have the highest rates of HIV/N. gonorrheae co-infected males. The
city of Detroit and Macomb County, despite having higher than most occurrences of co-infected
males do not fall into the highest group. Counties marked with “*” represent unstable rates (less
than 20 cases or population less than 50). A total of 879 HIV positive males with gonorrhea were
able to be geocoded.

164

Figure 4.5: Root Mean Square Error of N. gonorrheae per 1,000 in HIV Infected Males in
the State of Michigan 2005-2011 Adjus Population Density

Legend
RSE Gonorrhea per 1,000
0-10
10-30
30-50
50-70
70-100
*- Unstable Rate
Detroit
Counties
No Data Available

*
*

*

*

*

*

*

*

*

*
*

*

* *
* *
*

*

*

*
*

*

*
*

*

*

*
*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*
*

*

*

*
*

*

*
*

Root Mean Square Error (RSE) of rate of gonorrhea per 1,000 HIV positive males. Calculated
from 2006-2008 ACS and 2012 Demographic and Housing estimates:
http://factfinder2.census.gov /faces/nav/jsf /pages /index.xhtml. County population was adjusted
for males 0-65 years old. RSE was calculated by Standard Error/Rate*100, Standard
Error=Rate/Square Root (Cases). Counties marked with an “*” represent areas with unstable
rates (less than 20 cases of HIV or population less than 50).

165

Figure 4.6: Percent of Gonorrhea Infected HIV Positive Males with Multiple N. gonorrheae
Infections in the State of Michigan, 2005-2011

Legend
Percent Multiple Infection
Detroit
Counties
Percent of Gonorrhea
Infected HIV+ Males with
Multiple N. gonorrheae
Infections
16.7-25.0
25.0-31.7
31.7-42.6
42.6-42.9
42.9-100.00
No Data Available
Highest rates of repeated infections in HIV males with a history of N. gonorrheae infections do
not occur in the city of Detroit or in the tri-county area, but rather in counties with lower
populations of HIV infected males such as Gratiot, Clinton, and Lenawee counties. Detroit,
Wayne County, Oakland County, and Macomb County all still have relatively high rates of HIV
positive males with a history of N. gonorrheae having multiple N. gonorrheae infections. A total
of 879 HIV positive males with gonorrhea were able to be geocoded.

166

Figure 4.7: Map of Percent of HIV Infected Males with Multiple N. gonorrheae Infections
in the State of Michigan, 2005-2011

Legend
Percent Multiple Infection
Detroit
Counties
Percent HIV+ Males with
Multiple N. gonorrheae
Infections
0.44-0.85
0.85-1.37
1.37-1.99
1.99-2.76
2.76-7.14
No Data Available
Gratiot and Clinton counties have the highest rates of HIV positive males with multiple N.
gonorrheae infections. Detroit and the tri-county area also have increased rates of HIV positive
males with multiple N. gonorrheae infections. A total of 327 individuals with repeated N.
gonorrheae infections were able to be geocoded.

167

Figure 4.8: Getis-Ord GiZ score Cluster Analysis of N. gonorrheae Infected HIV Positive
Males in the State of Michigan, 2005-2011

Legend
Getis Ord HIV/Gonorrhea
Detroit
Counties
NG HIV GiZ Score
1.65-1.65 Std. Dev
1.65-1.96 Std. Dev.
1.96-2.58 Std. Dev.
>2.58 Std. Dev
No Data Available

Getis-Ord GiZ score found that counties located near the city of Detroit (Oakland, Wayne,
Macomb, Lapeer and St. Clair have higher rates of HIV positive males with a history of N.
gonorrheae as well as a select group of counties near Gratiot, Clinton and Shiawassee counties.
Counties and areas in the orange to red range are considered to be hot spots and represent a GiZ
score of greater than or equal to 1.65 standard deviations above the average rate. A total of
18,550 HIV positive males were able to be geocoded at the county level.

168

Figure 4.9: Poisson High Only Rate Cluster Detection Method of HIV Positive Males in the
State of Michigan, 2005-2011

Legend
Poisson High Only Cluster
HIV
Detroit
Counties
CLUSTER
Cluster 1 RR=10.26
p-value <0.0001
Observed=7792
Expected=1223.73

Poisson high only rate cluster detection of HIV clusters found that Detroit was the most likely
cluster with a RR=10.26 (p-value <0.0001). The observed number of cases in Detroit was 7,792
whereas the expected number of cases for the city was 1,223.73. A total of 18,550 HIV positive
males were able to be geocoded at the county level.

169

Figure 4.10: Poisson High Only Rate Cluster Detection Method of N. gonorrheae Infected
HIV Positive Males in the State of Michigan, 2005-2011

Legend
Poisson High Only Cluster
HIV/Gonorrhea
Detroit
Counties
CLUSTER
Cluster 1 RR=12.66
p-value <0.0001
Observed=415
Expected=57.99
Poisson method high only cluster detection found Detroit was the most likely cluster with a
RR=12.66 (p-value <0.0001). Observed number of HIV males with gonorrhea was 415 whereas
the expected number was 57.99. A total of 879 HIV positive males were able to be geocoded.

170

Figure 4.11: Bernouilli High Rate Only Cluster Analysis of N. gonorrheae Infected HIV
Positive Males in the State of Michigan, 2005-2011

Legend
Bernouilli High Only Cluster
HIV/Gonorrhea
Detroit
Counties
CLUSTER
Cluster 1 RR=1.44
p-value=0.007
Observed=140
Expected=101.88
Cluster 2 RR=1.26
p-value=0.036
Observed=466
Expected=415.32
Bernouilli analysis was performed investigating only high rate clusters. Oakland County
represented the most likely cluster with a RR=1.44 (p-value=0.007). A secondary cluster
consisted of the city of Detroit, Huron, Macomb, Sanilac, and Lapeer county had a RR=1.26 (pvalue=0.036) A third cluster consisting of Clinton and Gratiot counties was also detected,
RR=2.22, but was not found to be statistically significant (p-value=0.974). Cases consisted of
879 HIV positive males with a history of gonorrhea, and controls consisted of 18,550 HIV
positive males that were able to be geocoded at the county level.

171

Chapter 5.
Biological Markers and Geospatial Analysis of Michigan Males with HIV and N.
gonorrheae Co-Infections from 2011 to 2013
5.1 Abstract
Introduction
Epidemiological evidence suggests that there is a synergistic relationship between HIV and N.
gonorrheae with regard to changes in the natural history of these diseases in individuals who are
co-infected. This study aims to investigate the changes in biological markers, CD4 cell count and
HIV viral load, during co-infection periods. Geospatial analysis was also done in order to
investigate areas with high rates of co-infections at the zip code level.
Methods
A retrospective study of 743 males in the State of Michigan from 2011-2013 with a history of
HIV and N. gonorrheae was created in collaboration with the Michigan Department of
Community Health. Laboratory data extending back to time of HIV diagnosis were linked to
HIV and STD datasets, and longitudinal analysis was preformed to assess changes in CD4 cell
count and viral load due to N. gonorrheae infection. Geospatial cluster analysis at the zip code
level was also performed using Getis-Ord GiZ score as well as Poisson high only cluster
detection method.
Results
Both CD4 cell count and HIV viral load were found to be increased during the year time period
in which a N. gonorrheae infection was reported by MDCH. There was low overall reporting of
172

antiretroviral (ARV) usage, however, in those that did report ever being on ARV treatment, 71%
started in the same year or after their N. gonorrheae infection. Cluster analysis showed that there
is not equal geographic distribution of gonorrhea/HIV co-infections in the State of Michigan
within counties or in the City of Detroit. Poisson cluster analysis showed that Detroit based zip
codes had a RR=12.71 (p-value<0.0001), and that zip codes from the surrounding counties also
had increased in risk, Wayne and Northern Monroe RR=3.65 (p-value<0.0001) and Oakland,
Livingston, and Washtenaw RR=2.71 (p-value<0.0001).
Conclusion
Infection with N. gonorrheae during HIV positive status causes a change in both CD4 cell count
as well as in viral load, however, without better knowledge of ARV treatment, total individuals
ever receiving ARV, ARV cocktail usage, and length of time on ARV, these estimates only serve
as stepping stones to investigating the change in laboratory indicators of HIV progression in coinfected individuals. Geospatial cluster analysis showed that there is not uniform distribution of
co-infected individuals in the State of Michigan within counties and the State of Detroit. Public
Health and Epidemiology need to target areas with the highest risk to make the most of the
limited resources available to preventing the spread and acquisition of HIV and other preventable
STDs.

173

5.2 Introduction
Evidence suggests that there is a synergistic relationship between HIV and infection with N.
gonorrheae resulting in non-classical manifestations of these diseases in co-infected individuals.
Attention has focused primarily on co-infection with ulcerating STI’s in recent years. However,
non-ulcerating STI’s such as N. gonorrheae account for a significant co-infection burden in the
HIV/AIDS community and there is a need for research to explore the changes in laboratory level
data, such as CD4 cell count and viral load, in those who are co-infected (45, 47).
Despite multiple recent outbreaks of N. gonorrheae as well as the emergence of antibiotic
resistant gonorrhea and that N. gonorrheae is one of the most reported diseases in the United
States, the relationship between HIV and gonorrhea has yet to be fully explored (15). In 2011
WHO estimated that roughly 88 million of the estimated 448 million curable STD infections
reported that year were due to N. gonorrheae (12). In the United States in 2012, there were over
334,000 N. gonorrheae cases resulting in an increased rate of 4.1% from the previous year (13).
As of the 2010 Census Michigan has the 8th largest population in the US. In 2011 there were over
13,000 cases of gonorrhea reported in the State of Michigan with the highest rate reported in the
20-24 age group (14). The 20-24 age group accounts for 6.7% of the population of Michigan, yet
accounts for 34% of all gonorrhea cases (14). The overall rate in African Americans in the State
of Michigan was roughly 24 times higher than the rate in Caucasians (14). Males account for
41% of all 2011 gonorrhea cases in the State of Michigan (14). Of the reported 13,070 gonorrhea
cases 2% (n=259) were also HIV positive (14).
The January 2014 Annual HIV Surveillance Analysis estimated that there were approximately
19,800 individuals in the State of Michigan who were HIV positive of which an estimated
174

15,440 (78%) were males (10). Of the 15,440 males with HIV, men who have sex with men
make up 55% of prevalent HIV cases between the combine risk groups of MSM and injection
drug user (IDU)/MSM in the State of Michigan (10). Heterosexual contacts make up 19% of
those with HIV infection and 17% had an undetermined source of transmission (10). The most
frequent age group for HIV diagnosis was 30-39 year old (33%) with the combined age group of
20-29 at 32% (10). The majority of HIV diagnosis occurred in the Detroit Metropolitan Area
(64%) despite only accounting for 43% of the total Michigan population (10). Males made up the
majority of HIV infected individuals with a reported 12,510 cases (10). Within this group 71%
are MSM (combined risk groups of MSM and MSM/IDU), 5% are heterosexuals, and 17% have
an undetermined mode of transmission (10).
Wasserhiet et al. states that epidemiological synergy of HIV and STDs occurs “If coinfection
with HIV prolongs or augments the infectiousness of individuals with STDs, and if the same
STDs facilitate transmission of HIV, these infections may greatly amplify one another” (95). A
second point the article makes is that there is a hard to define gap between infection with HIV
and when HIV is diagnosed (95). This is important to note because it has a great effect on when
baseline levels of CD4 and viral load would be collected (95). Those with low baseline levels of
CD4 or high baseline levels of viral load concentrations may be in a state of elevated HIV
progression due to an unaccounted immune response at time of testing (95). The article also
notes that while there is value in testing the presence of absence of a given STD and its influence
on HIV, the number of infections for each STD may also be an important risk factor, that at the
time of the article were being overlooked (95).
Another key paper in the epidemiological synergy literature between HIV and STDs is the 1999
Fleming and Wasshereit paper that reviewed the scientific data on these co-infections (97). The
175

article found that non HIV STDs can facilitate the shedding of HIV in the genital tract (increased
viral load) which then promotes the transmission of HIV and the infectivity of the genital
secretion (97). Mayaud and McCormick state that HIV can modify the duration, severity, and
response to treatment of certain STDs, but cite that this is most common in viral infections (40).
The article also states that STDs can influence the risk of HIV transmission in naïve individuals
as well as enhance the viral load of those infected with HIV (40). The article states that among
HIV infected males there can be an upwards of an 8-fold increase in the viral load in semen as
compared to the control group, as well as an increase in viral load found in the vaginal secretions
of HIV infected women (40).
A review of co-infection literature by Bafica et al. states that HIVs enhanced viral expression
during co-infection states may be due to the microbial-induced immune response that occurs
during these periods (90). Bafica et al. also states that when the article was written (2004) tolllike receptors (TLR) were the major innate immune response for microbial infections (90). The
review also found that in most human cells, when stimulated with TLR-2, 4, or 9 that there is an
increase in the viral replication of HIV (90). The review goes on to state that another key
component of HIV enhancement during co-infection periods may be due to cytokine-mediated
induction of HIV long terminal repeats [long identical sequences of DNA found at the
retrotransposons or proviral DNA end and used to insert genes into host (91)] via nuclear factor
κB activation which occurs during times of co-infection (90).
N. gonorrheae infection has been shown to attract CD4+ lymphocytes to infective areas
disrupting epithelial and mucosal barriers and increasing possible susceptibility to HIV (24). A
study by Ding et al noted that N. gonorrheae infection enhanced the HIV infection of CD4 cells
early infection via TLR-2 activation which not only increased the susceptibility of CD4 infection
176

but also enhanced the viral load associated with HIV infected CD4 cells (28). An editorial review
by Bentwich et al. in 2000 states that since all co-infections will cause immune responses, that it
is plausible that these co-infection states enhance HIV infection, increase HIV viral load, and
increase the rate of progression of disease in those individuals with STD/HIV co-infection states
(45). Bentwich et al. diagrams that co-infection starts via activation of cytokines such as tumor
necrosis factor-alpha and helper T cells 1 and 2 and activation of chemokines and related
receptors co-infections states increase the susceptibility of HIV infection in naïve individuals
(45). Bentwich also states that non-ulcerative bacterial infections such as N. gonorrheae are
more prevalent in HIV positive individuals than genital ulcerative bacterial infections such as
Haemophilus ducreyi and Treponema pallidum more commonly referred to as syphilis (45). The
authors acknowledge that ulcerative infections are associated with increased risk of HIV
infection and increased viral load concentrations in those infected with HIV for both men and
women but state that inflammation may be the main factor in HIV enhancement in co-infected
individuals (45). The review goes on to state that gonorrhea may increase HIV transmission by
recruitment of CD4+ cells at that infected site and that treatment of the co-infecting STD will
reduce the transmission potential of HIV (45).
A review by Mayer and Venkatesh from 2012 found that even after treatment of the bacterial
STD, HIV viral levels in the genitals do not return to levels at a pre-bacterial infection state, this
is an important concept because it shows that while treatment has the ability to reduce the sudden
increase in viral load, treatment does not fully negate the damage done during a co-infection
period (101). The review also found that N. gonorrheae enhanced HIV infection by activation of
TLR-2 which can both increase susceptibility of CD4 cells to HIV infection and enhance HIV
replication in those CD4 cells already infected (101). The review also noted that N. gonorrheae
177

co-infection periods increased the ability of HIV to infect dendritic cells which then present the
virus to other susceptible immune cells (101).
A conflicting study found that in HIV infected males that are co-infected with N. gonorrheae
viral RNA is increased in the semen roughly 10-fold and drop to pre-gonorrhea infection levels
after the start of antibiotic treatment (24). The literature also suggests that there is at least a twofold increase in the risk of men spreading HIV in their ejaculate who are co-infected with N.
gonorrheae than in men that are not currently co-infected (31). A 2005 review done by Risbud
and the National AIDS Research Institute found that HIV viral shedding is enhanced during
STD/HIV co-infection periods by the damaging epithelial cells and causing the accumulation of
HIV vulnerable cells (47). The article also notes that STDs are the single major factor in the
spread of HIV (47). The review also noted a study done by Wasserheit et al. that found that while
both ulcerative and non-ulcerative increase HIV transmission, non-ulcerative STDs infections
such as N. gonorrheae and may be due to the higher incidence of these infections in the public as
compared to ulcerative STDs (47).
A study modeling the CD4 counts over time in AIDS positive individuals taking the drug
zidovudine found that after an early spike in CD4 counts, around week 10 of treatment CD4
counts rapidly decline, obtaining baseline levels by week 20 and reaching an asymptote (part of a
curve line where the distance between the curve and the line approaches zero as both move
toward infinity) around week 90 (61). One of the major limitations of the study is that as time
progressed, individuals dropped off due to death and missing observations (61). However, when
looking at only survivors the CD4 curves look similar to the total population CD4 curve and
show that in the sample AIDS population CD4 counts dropped below 10 cells/µL at around one
year (61).
178

Yu et al. found that there is a linear relationship between CD4 count and CD4 percent (76). The
study found that a CD4% of 5% corresponds to a CD4 cell count of 45 cells/mm3 (CI 17-117
cells/mm3), CD4% of 15% corresponds to a CD4 cell count of 182 cells/mm3 (CI 64-499), and a
CD4% of 30% corresponds to a CD4 cell count of 438 cells/mm3 (CI 132-1395) (76). The study
also found that among three groups, homosexual/bisexual, injecting drug users, and
heterosexuals, homosexual/bisexuals have the highest CD4 cell counts in both AIDS-free and
AIDS positive groups (76). In AIDS-free individuals those that are not on zidovudine therapy or
pneumocystis carinii pneumonia (PCP) prophylaxis have the highest CD4 count in the low CD%
grouping (5%) and those that are only on zidovudine have the highest CD4 counts in the high
percent grouping (30%) (76). In the AIDS positive group CD4 counts in the 5% CD4 group were
highest in those not on any therapy, and in the 30% CD4 group, CD4 counts were highest in the
zidovudine only group (76). Zidovudine works well in keeping CD4 cell counts elevated,
however, in individuals who are in the lowest CD4 grouping, the drug does not seem as
effective.
5.3 Methods
A retrospective cohort of HIV positive males in Michigan with a history of N. gonorrheae
between 2011 and 2013 were selected as the study population. Through collaboration with
Michigan Department of Community Health (MDCH) a linkage structure between HIV and N.
gonorrheae datasets was created for this study. Records with missing first name, last name, or
date of birth were excluded from the study population prior to the matching of the data.
Laboratory data for these individuals was also obtained through MDCH and included all CD4
cell count/percent and viral load data available for an individual. An employee of HIV
Surveillance and Body Art Unit HIV/STD/VH/TB Epidemiology created the record linkage
179

structure between HIV datasets, N. gonorrheae datasets and laboratory datasets to comply with
patient confidentiality and privacy. The matches were conducted between the registries using
LinkPlus. All data used in this study were de-identified according to HIPAA guidelines on public
health information. Study staff also did not contact or attempt to contact any participants of the
study. Ethical approval was obtained from the Institutional Review Board (IRB) at Michigan
State University (MSU) and was reciprocal with the MDCH IRB.
HIV cases were defined according to the CDC case definition of HIV and AIDS according to
CDC 1993 guidelines. HIV infection was defined as HIV-NA=HIV infection, non-stage 3 and
AIDS=stage 3 HIV infection. Acute and Latent, stage 1 and 2, were not differentiated in this
study. Date of birth, date of death, N. gonorrheae infection, as well as HIV and AIDS diagnosis
were all limited to year of the event. Vital status was coded as 1=alive, 2=deceased, 9=unknown.
Ethnicity was classified into unique non-over lapping groups, 1.White= White non-Hispanic,
2.Black= Black non-Hispanic, 3.Hispanic=Hispanic of any race, 4.Asian/HI/PI=Asian/Hawaiian,
Pacific Islander non-Hispanic, 5.Am In /Ak Nat=American Indian, Alaskan Native, nonHispanic, or 9.Multi/Unk/Other=Multirace, Unknown, Other, non-Hispanic. Due to low rates of
Asian/Hawaiian, Pacific Islanders (N=0) and American Indian, Alaskan Native (N=1) in the N.
gonorrheae group these racial groups were included in the 9.Multi/Unk/Other grouping to avoid
analytical complications. Marital status while largely missing was categorized as A=Married and
Separated, S=Single and never married, M=Married, N=Not otherwise specified, Missing or
U=Unknown, W=Widowed; Risk was condensed and categorized as A.MSM=Male-Male Sex,
B.IDU=Injecting Drug Use, C.MSM/IDU=Male-Male Sex and Injecting Drug Use, D. Blood
Recipient=Received Blood, E1.(Male HCFR)=Male who had sex with a female at risk for HIV,
E2. (Fem) HCM=Heterosexual contact with a female, F. Perinatal=Child exposed by mother, G.
180

Unk:Other=Unknown/Other. For CD4 analysis MDCH risk groups were condensed to
A.MSM=MSM and MSM/IDU and G. Unk:Other=All other risk groups due to low numbers in
the other groups. For viral load analysis F. Perinatal was analyzed as its own group along with
MSM, and Other/Unknown.
The dataset also contained limited information on antiretroviral (ARV) therapy. A binary
variable indicating if ARV were ever taken was coding as ‘Y’ and ‘U’ and was analyzed as yes
and not yes. Along with this were two variables that indicated date of ARV start and date of
ARV end. While most of the data for individuals who had responded yes to ever taken ARV was
complete, day month and year, there were 5 individuals with missing information on ARV start
date. Four individuals were missing the day of the month that they started ARV therapy. In these
individuals start date was set to the last day of that month. This was done to allow for usage of
the ARV information in the longitudinal analysis of the data while trying to adjust for bias
caused by assigning a day to the unknown field. One individual only had year of ARV start,
ARV start date was set to December 31st of that year. Two individuals were missing days of the
month for ARV end dates, end dates were set to the first of the month. In individuals with
perinatal exposure to HIV, no date of HIV positive status was recorded. For these individuals
year of HIV positive status was set to birth year.
Biological marker data included date of collection (MM/DD/YYYY), collection day (DD),
collection month (MM), and collection (YY) as four distinct variables. Result data was captured
using several variables as well. Results were mainly numeric data (aside for “UNDET”) stored as
character data, which was transformed to numeric data using SAS 9.3. A variable label
results_units was the main identifier of what type of laboratory data the result was. PCT=Percent
CD4 count, CNT=CD4 Cell Count, and C/ML=Viral Copies per mL. Lab data also contained a
181

result interpretation variable that identified results as < (below limit), = (within limit) and >
(above limit). The final variables that captured results data were VL=Viral load (collected as
character, converted to numeric), CD4=CD4 cell count (collected as character, converted to
numeric) and CD4PCT=CD4 percent (collected as character, converted to numeric). Biological
marker data also contained N. gonorrheae infection year for study participants.
Age at HIV was a constructed variable that approximated age at HIV by using year of HIV
diagnosis and year of birth. A flag variable for time on ARV treatment was also created for this
study. The flag variable was constructed so ARV would only equal 1 while the individual was
actually on treatment, as opposed to the MDCH ARV indicator that only measured if someone
had ever been on ARV treatment. ARV flag was constructed by using data of lab collection (date
the viral load/CD4 cell count was collected) and was equal to 0 if collection occurred before start
of ARV treatment, after the end of ARV treatment, or where individuals did not have start or end
dates, ARV=’U’, from the MDCH dataset. ARV flag was set to 1 when collection date occurred
on or between the ARV start and end date, and when it occurred after ARV start and no end date
was recorded. As stated before N. gonorrheae infection was limited to only the year of infection
and analyzed for viral load as N. gonorrheae during collection year=1 and not during collection
year=0. This variable was constructed using collection year, all data where collection year was
equal to year of N. gonorrheae infection, gonorrhea=1, else gonorrhea=0. For CD4 cell count,
longitudinal analysis was able to capture before, during, and after N. gonorrheae infection so
gonorrhea was coded as 0=collected before N. gonorrheae infection, 1= same year as N.
gonorrheae infection, and 2=collected in the year/s after N. gonorrheae infection.
Longitudinal analysis was run using SAS 9.3 (SAS Institute, Inc. Cary, North Carolina).
Descriptive statistics were performed to assess distribution of demographic and risk
182

characteristics. Primary analysis was done using general estimating equation (GEE) with
repeated subject adjustment to investigate change in CD4 and viral load in HIV positive males
with a history of N. gonorrheae infection between 2011 and 2013. Adjusted (collection date,
ethnicity, MDCH condensed risk group, age at HIV, and ARV usage) and unadjusted models
were ran for each of the biological marker variables.
Spatial cluster-analyses were performed using two methods: 1) Getis-Ord statistics to access
global and local High and Low clusters of prevalence rates in using ArcGIS v.10.2
(Environmental Systems Research Institute (ESRI) 2014).; 2) Poisson distributed methods using
SaTScan V9.1.1 to detect high prevalence rates. Getis-Ord Gi Hot Spot analysis is a cluster
detection analysis that measures each area in context to its neighbors. If a map unit (zip code) has
high values that are contiguous to its adjoining neighbors then the area constitutes a ‘hot spot’.
The cluster is then compared to all map units and a Z score is calculated. Clusters with higher
rates of disease will have higher Z scores and clusters with lower than average rates of disease
have lower Z scores. The GiZ scores of were mapped at the zip code level. A GiZ score of 0
represents a map unit that does not differ from the overall mean. An increase in GiZ score
represents greater distance from the overall mean (increase in N. gonorrheae/HIV co-infection).
A negative GiZ score represents areas with a lower than average rate of N. gonorrheae/HIV coinfection.
Coordinate data was constructed using ArcGIS via the “Calculate Geometry” function, and
mapped as latitude and longitude. Poisson cluster-detection methods consist of a gradually
scanning window that moves across the study frame (Michigan zip codes) calculating the
number of expected and observed observations (cases) inside the scanning window of each map
unit. Cases were Michigan HIV positive males with a history of N. gonorrheae infection between
183

2011 and 2013. Maximum likelihood estimates are calculated to determine areas where
clustering is least likely (p-value less than 0.05) to be due to chance. SaTScan has the ability to
document and rank multiple possible clusters of both high and low rates. Geospatial analysis of
HIV and N. gonorrheae clusters were done using Poisson method scanning for high only rate
areas. No time precision options were used and map areas were scanned for high only rates with
no clusters centered over other clusters. A maximum radius of 20 miles (32.2 km) was used as
well as the limitation of no more than 50% of population at risk in a cluster. Zip code
populations were taken from the 2008-2012 American Community Survey 5-year Estimate
available at the US Census website (http://factfinder2.census.gov/faces/tableservices
/jsf/pages/productview.xhtml?pid=ACS_12_ 5YR_S0101&prodType=table). Zip code
populations were adjusted for age (0-65) and sex (males only) using the estimated percent for 5
year age group blocks.
5.4 Results
Table 5.1 shows the demographic distribution for the 743 HIV positive males in the State of
Michigan with a reported history of N. gonorrheae infection between 2011 and 2013. Blacks
represent the largest ethnic group, accounting for 78% of co-infected individuals, with Whites
accounting for 15%, Hispanics accounting for 2% and Multiracial/Unknown/Other accounting
for 4%. In HIV positive males who had a N. gonorrheae infection between 2011 and 2013
around 66% progressed to AIDS positive status by 2013, with only 1 death in co-infected
participants. MDCH condensed risk group showed that MSM accounted for the majority of the
population, 77%. Other MDCH risk group categories had relatively low numbers. Only 6
individuals (<1%) fell into the IDU category, 5 (<1%) were categorized as MSM/IDU, 1
individual was grouped in the Blood Recipient group, 15 (2%) males identified as having
184

heterosexual intercourse with a woman living with HIV, and 45 (6%) identified themselves as
heterosexual. There were 31 individuals in the study population with perinatal exposure to HIV
(4%). The mean age at HIV diagnosis was 25 years old.
CD4 Cell Count
CD4 cell count was modeled using a longitudinal approach that accounted for repeated measures
in study participants. Table 5.2 shows both the collection date adjusted and multivariate change
in CD4 cell count. A multivariate model was constructed using N. gonorrheae infection as a
discrete variable with 3 levels, collection date prior to N. gonorrheae infection, collection date
occurring in same year as N. gonorrheae infection, and collection date after N. gonorrheae
infection. The multivariate model was also adjusted for race, MDCH risk group (MSM vs
Other), age at HIV diagnosis and ARV usage at time of collection. ARV usage was not found to
be a statistically significant predictor of CD4 cell count in this group, however, due to what is
known about the relationship between CD4 and ARV usage, ARV was retained in the model.
The model observed that as date of collection progressed forward in time there was a mean
reduction in CD4 cell count of -0.06 (p-value=0.005). As compared to the reference state of CD4
collected prior to N. gonorrheae infection, CD4 collected during the same year saw a mean
increase in cell count by 110 cells (p-value <0.0001), and collections that occurred in the years
after a N. gonorrheae infection saw a mean increase in difference of 138 cell as compared to the
reference state (p-value =0.0001).
As compared to MSM (MSM and MSM/IDU) HIV infected males that fell into one of the other
condensed risk groups had a mean increase in CD4 cell count of 90 (p-value=0.011). Blacks
were the only ethnic group to have a statistically significant change in mean CD4 cell count as

185

compared to Whites. Blacks had a reduction in CD4 cell count of 64 (p-value=0.036) as
compared to the reference group of Whites. Hispanics also had a reduction in CD4 cell count, but
this was not statistically significant when compared to Whites. Age of HIV diagnosis reduced
CD4 cell count by 5 per year increase in age (p-value=0.002).
Viral Load
Viral load was modeled using a longitudinal approach that accounted for repeated measures in
study participants. Table 5.3 shows both the adjusted and unadjusted change in viral load. A
multivariate model was constructed using N. gonorrheae infection as a discrete variable with 2
levels, collection date occurring in same year as N. gonorrheae infection, and collection date
prior/after N. gonorrheae infection. The multivariate model was also adjusted for race, MDCH
risk group (MSM, Perinatal and Other), age at HIV diagnosis and ARV usage at time of
collection. ARV usage was not found to be a statistically significant predictor of viral load in this
group, however, due to what is known about the relationship between viral load and ARV usage,
ARV was retained in the model. The model observed that as date of collection progressed
forward in time there was a mean decrease in viral load of -1.13 copies/mL (p-value=0.001). As
compared to the reference state of viral load collected in years with no N. gonorrheae infection,
viral load copies/mL collected during the same year saw a mean increase in copies/mL by 1631
copies/mL (p-value <0.033).
As compared to MSM, those that were placed in the perinatal MDCH condensed risk group had
a reduction in mean viral load of 16,374 (p-value<0.0001), those in the Unknown/Other group
did not have a statistically different mean viral load than MSM. Age at HIV diagnosis had a
mean reduction of 260 copies/mL (p-value<0.0001). As with CD4 cell count only those in the

186

Black ethnicity group had a statistically significant change in viral load as compared to Whites.
Despite ARV being not statistically significant there was a decreasing viral load trend for
measurements taken during their reported treatment time. The estimated mean reduction due to
ARV treatment was 2,038 copies/mL (p-value=0.306)
Zip Code Cluster Analysis
The State of Michigan consists of approximately 900 unique zip codes according to the 2010 US
census. Figure 5.1 shows the map of the Getis-Ord GiZ score of HIV/gonorrhea co-infections
between 2011 and 2013 in males in the State of Michigan. Using GiZ score cluster analysis two
“hot spots” were detected, both in close proximity to the City of Detroit. Figure 5.2 is a zoomed
in view of the 2 “hot spots” identified by the Getis-Ord GiZ score analysis. The main cluster
consists of around 48 zip codes that account for most of Detroit as well as the Northern adjacent
zip codes. This primary cluster is largely situated within the North East corner of Wayne County,
but extends north to the Southern Edges of Oakland and Macomb counties. The second cluster
that was detected by Getis-Ord GiZ score analysis is situated slightly east of center of the county
of Oakland, a relatively short distance from the primary cluster.
A second cluster analysis was performed using Poisson high only cluster analysis techniques.
Maximum spatial cluster size was set to 50.0% of the population at risk as well as in a circle with
a 32.2 km (20 miles) radius. Figure 5.3 shows the 5 clusters detected in the Poisson cluster
analysis and Figure 5.4 shows a zoomed in version of this map. The most likely cluster includes
zip codes that constitute the City of Detroit as well as a number of zip codes to the north that
extend into parts of Oakland and Macomb counties. Table 5.4 shows the RR and respective pvalue of Poisson identified clusters. Cluster 1 consists of 44 zip codes with a RR=12.71 (p-

187

value<0.0001). A secondary cluster consisting of 38 zip codes located in Wayne and northern
Monroe counties and had a RR=3.65 (p-value<0.0001). Cluster 3 is located in Washtenaw,
Livingston, and Oakland as well as 5 zips in the North West corner of Wayne County and
consists of 49 total zip codes with a RR=2.71 (p-value<0.0001). Cluster 4 consists of 3 zip codes
in Kalamazoo with a RR=3.02 which was found to not be statistically significant (pvalue=0.971). The final cluster that was detected using Poisson high only analysis was located
Genesee County and consisted of 2 zip codes in the center of the county, cluster 5 had a RR=2.78
and was not found to be statistically significant (p-value=0.998).
5.5 Conclusion
CD4 Cell Count
This study found that there is an increase in CD4 cell count in the year during a N. gonorrheae
infection as well as in the years following gonorrheal infection as compared to the time from
HIV diagnosis to first N. gonorrheae infection. The study also found that there was an overall
decrease in CD4 count as time progressed in the study population which is characteristic of being
HIV positive. Using the MDCH condensed risk group categories, those that were MSM (MSM
and MSM/IDU) had a lower CD4 cell count as compared to the rest of the population. The study
also found there is a lower CD4 count in those that are diagnosed with HIV at an older age as
compared to those that are younger, which may help to explain why time from HIV to AIDS is
so dependent on the age of the individual, with those that are diagnosed with HIV at an older age
typically having a shorter time to AIDS positive status as seen in Chapter 3. The study also found
that while both those that identified their ethnicity as Black or Hispanic had a reduction in mean
CD4 cell count, only Blacks were statistically significant.

188

Finally, ARV usage was not a statistically significant factor in CD4 progression, which may be
largely due to two reasons. First, there were very few individuals in the study population that
reported that they had ever been on ARV treatment, and not all of these individuals remained on
ARV treatment to the end of the study. Second, ARV data is not defined as yes and no, it has to
be analyzed as yes and not yes. Individuals that did not have a “Y” for ARV are actually
unknown and may be on ARV treatment. Without knowing if individuals are on ARV treatment
and for how long they have been on treatment ARV as a measurement will remain incomplete.
Despite this ARV was included in the model because of the known relationship between taking
ARV and changes in CD4 cell count.
These findings may help to shine light on a previous seemingly paradoxical relationship between
N. gonorrheae infection and time from HIV to AIDS positive status. In a previous chapter,
Chapter 3, there was a greater lag time between HIV and AIDS in individuals with a history of
N. gonorrheae infection as compared to those without a history of gonorrhea. This spike in CD4
cell count during and after N. gonorrheae infection may play a role in that paradoxical
relationship and may represent a possible intervention time in the natural history of an
individual’s disease progression, however, it would also increase the number of CD4 cells HIV
could infect. Since this data was collected from MDCH reportable disease database, individuals
that had a confirmed N. gonorrheae infection would have gotten treatment for the infection and
thus brought back into medical care; where as those without a history of N. gonorrheae infection
may not have the same access to medical care. This is speculation and there is need for subgroup
analysis in both the initial HIV/gonorrhea study population (2005-2011) as well as the laboratory
subgroup (HIV/gonorrhea 2011-2013) population. There is also a need for further research into
all STDs and reported infections an individual had from onset of HIV to assess the change in
189

CD4 due to various bacterial, viral, parasitic, and fungal infections. Finally there is a need to gain
better information on ARV usage in this population and address any barrier to ARV access or
compliance.
Viral Load
Viral load (copies/mL) increased in the year during N. gonorrheae infection as compared to the
time periods before and after N. gonorrheae infection. This coincides with what other studies
have found and what researchers like Wasshereit and Flemming have outlined in the STD
synergistic papers (24, 28, 40, 45, 90, 95, 97, 101). This finding is also biologically supported by
the finding in the previous section in that during the year of a N. gonorrheae infection there is
also an elevation of CD4 in the body, these CD4 act as replication sites for HIV allowing the
virus to have more host cells in which to make viral copies.
The study also found that Blacks had a statistically significant increase in viral load as compared
to Whites. The study also observed that in the multivariate model viral load decreased with
collection date, representing a decreasing trend in viral load as time progressed for these
individuals. Being older at age of HIV diagnosis was also found to reduce the mean viral load in
individuals. This study was unable to investigate the cause of these findings, but they may be due
to ARV treatment or HAART in these individuals. As with CD4 cell count ARV was not found
to be a statistically significant predictor of viral load, but was included in the model because of
its known ability to facilitate in the regulation of viral load in HIV positive individuals. The
study also found that 42 of 59 (71%) individuals that reported they were ever on ARV started in
the same year or after their N. gonorrheae infection. While ARV reported usage was low in this
study (~8%) it is important to note that lack of a positive response is not the same as a no, and

190

that in fact some of these individuals may have been on ARV treatment. The fact that 71% of
those that responded that they were ever on ARV having started ARV in the same year or after a
N. gonorrheae infection may help to explain some of the seemingly paradoxical observations
found in this study, such as CD4 cell count not returning to baseline or lower level in the years
after a N. gonorrheae infection and the increased lag time from HIV to AIDS seen in Chapter 3.
Zip Code Cluster Analysis
This subgroup of HIV/gonorrhea co-infected individuals represented a unique opportunity to
investigate spatial patterns of disease at the zip code level, and to better identify areas in the State
of Michigan were efforts to reduce N. gonorrheae infections and disrupt HIV transmission
should be targeted. Both cluster techniques found that the City of Detroit represented the most
likely “hot spot” with regard to HIV/gonorrhea co-infections. Both techniques also found that
there was an area in the proximity of central Oakland County that also had elevated risk of
HIV/N. gonorrheae co-infection as compared to the rest of the state. These maps reflect what
was also seen at the county level of geospatial analysis and aid in confirming that the City of
Detroit and neighboring areas need to be the primary target for interventions aimed at stopping
the spread of STDs, specifically N. gonorrheae and N. gonorrheae/HIV positive males.
The Getis-Ord GiZ score analysis showed that the areas with the highest standard deviation from
the mean of the state was located in and around the City of Detroit with a small subcluster being
located just right of center in Oakland County. This analysis was able to show that these areas
have higher than average rates and that intervention services may have the most impact in these
areas. The Poisson cluster analysis method was also able to show that the area around the City of
Detroit was a “hot spot”, it unlike the Getis-Ord analysis, the Poisson analysis was able to

191

visually show how the rates of disease seem to diffuse from the primary cluster (Detroit) and
spread to other areas, such as Wayne, Oakland, and parts of Livingston and Washtenaw counties.
These maps as with the maps in the previous chapter show that the burden of HIV/gonorrhea coinfection is not even distributed among the state, and by using zip code level data, the study
found that it is not evenly distributed among or within the tri-County area of Macomb, Wayne,
and Oakland. The use of zip codes allows for epidemiologists and public health officials to
consider where the best area for intervention and prevention efforts should be placed in the city
as well as across the affected counties.
Limitations
The data for this study were collected for surveillance purposes and not primarily for
epidemiological or biostatistical analysis. There is a certain level of missing/incomplete data that
is inherent in using data not specifically designed for the research question as opposed to
designing a study to specifically capture key variables. The lack of information on ARV is a key
limitation of this study and a key area to focus on to make these data more meaningful.
Obtaining better response rates on ARV treatment will allow for better modeling of key
laboratory factors such as viral load and CD4 cell count, allowing epidemiologists and public
health officials to model the actual disease progression in an individual by adjusting for a
treatment that is designed to reduce viral load and there by either maintain or improve CD4 cell
count. Another key limitation is that N. gonorrheae infection was only recorded as year of
infection, and no data such as history of other STDs or other bacterial reportable disease during
HIV was collected. The lack of information of other STDs is important when looking at spikes in
CD4 and viral load in the period of time prior to N. gonorrheae infection. This again is a

192

limitation to using surveillance data and should be considered when developing future integrated
models. A final limitation of the study was that only data on co-infected individuals was
collected. The study was not able to model the change in CD4 and viral load in individuals with
only HIV infections as compared to those with a history of N. gonorrheae infection. Having a
control group may have allowed for better inferences from the data as well as a way to control
for potential bias due to other STD or reportable infections individuals may have had in the time
leading up to their N. gonorrheae infection.

193

APPENDIX

194

Table 5.1: Demographic Distribution of 743 HIV/Gonorrhea Co-Infected Michigan Males
from 2011-2013

Variable

Ethnicity
White
Black
Hispanic
Multiracial/Unknown/
Other

ARV Treatment Status
(Ever Been on)
Yes
Unknown

HIV/N. gonorrheae CoInfected
N=743
109 (14.67%)
583 (78.47%)
18 (2.42%)
33 (4.44%)

59 (7.94%)
684 (92.06%)

Risk Group
MSM
IDU
MSM/IDU
Blood Recipient
Male HCFR
Fem HCM
Perinatal
Other/Unknown

569 (76.58%)
6 (0.81%)
5 (0.67%)
1 (0.13%)
15 (2.02%)
45 (6.06%)
31 (4.17%)
71 (9.56%)

AIDS Positive
Deceased

246 (33.11%)
1 (0.13%)

Mean Age at HIV (Std Dev)

25.35 (9.59)

195

Table 5.2: Multivariate General Estimating Equation (GEE) with Repeated Subjects Model
of CD4 Cell Count in HIV Positive Males with N. gonorrheae
Variable

Mean Collection Date
Adjusted CD4 Cell Count

1843.2 (898.2-2788.2)
1967.1 (1904.4-2029.7)
1998.0 (1917.2-2078.8)

Mean Multivariate Adjusted
CD4 Cell Count
Baseline:
1818.9 (937.4-2700.4)
Reference
109.9 (56.0-163.8)
137.9 (68.4-207.5)

1273.247 (550.6-1995.9)
1253.8 (1197.1-1310.4)
1253.7 (1153.9-1353.6)
1353.7 (1240.2-1467.2)

Reference
64.4 (-124.7--4.1)
40.6 (-133.4-52.2)
78.6 (-36.7-193.8)

1201.3 (523.9-1878.7)
1296.5 (1221.7-1371.3)

Reference
90.4 (20.9-159.9)

Gonorrhea
Pre-Gonorrhea Period
During Gonorrhea Period
Post Gonorrhea Period
Ethnicity
White
Black
Hispanic
Multiracial/Unknown/
Other*
Risk Group
MSM
Other/Unknown

Per Unit Increase

Baseline:
1338.1 (610.4-2065.9)
4.4 (-7.7--1.2)

ARV
Unknown/Not Yes
Yes

1284.1 (557.0-2011.1)
1317.8 (1202.9-1432.7)

Age at HIV Diagnosis

Collection Date
Per Unit Increase

-

-

4.9 (-8.1--1.8)

Reference
47.3 (-84.7-179.4)
-

0.06 (-0.11--0.02)

196

Table 5.3: Multivariate General Estimating Equation (GEE) with Repeated Subjects Model
of Viral Load (copies/mL) in HIV Positive Males with N. gonorrheae
Variable

Mean Collection Date
Adjusted Viral Load
(copies/mL)

Mean Multivariate Adjusted
Viral Load (copies/mL)

Pre/Post Gonorrhea
During Gonorrhea

23029.0 (9339.7-36718.3)
24624.4 (23124.4-26124.7)

Baseline:
36441.7 (22845.7-500376)
Reference
1630.8 (129.8-3131.8)

Ethnicity
White
Black
Hispanic
Multiracial/Unknown/
Other*

18509.9 (4740.2-32279.6)
22347.2 (20298.8-24395.5)
20563.7 (17121.1-24089.7)
20563.7 (17121.1-24006.4)

Reference
2181.6 (59.6-4303.6)
759.6 (-5153.1-3633.8)
799.3 (-2532.7-4131.2)

Gonorrhea

Risk Group
MSM
Perinatal
Other/Unknown

23206.4 (10214.1-36198.7)
14629.0 (11953.1-17301.8)
23055.1 (20661.7-25448.6)

Per Unit Increase

Baseline:
26056.3 (10068.8-42043.8)
195.3 (-299.5--91.0)

ARV
Unknown/Not Yes
Yes

19005.0 (5559.2-32450.8)
16647.2 (12627.8-20666.5)

Age at HIV Diagnosis

Collection Date
Per Unit Increase

Reference
16374.4 (-20152.1--12596.6)
719.6 (-1685.8-3125.0)

-

-

260.0 (-351.8--168.2)

Reference
2037.61 (-5940.2-1865.0)

-

-

1.1 (-1.8--0.4)

197

Table 5.4: Poisson Analysis of Gonorrhea/HIV Co-Infected Clusters in Michigan Males,
2011-2013
Cluster ID
1
2
3
4
5

Geographic
Description
Detroit, Southern
Oakland and Macomb
Wayne and Monroe
Oakland, Livingston,
Washtenaw
Kalamazoo
Genesee

Relative Risk (RR)

p-value

12.71

<0.0001

3.65
2.71

<0.0001
<0.0001

3.02
2.78

0.971
0.998

198

Figure 5.1: Getis-Ord GiZ Score Cluster Analysis of N. gonorrheae/HIV Co-Infected
Michigan Males 2011-2013

Legend
Getis Ord
Gonorrhea/HIV
GiZ Score
1.65-1.65 Std. Dev
1.65-1.96 Std. Dev
1.96-2.58 Std. Dev
>2.58 Std. Dev
Detroit
Counties

Getis-Ord GiZ Score found that the zip codes representing the City of Detroit had higher rates of
N. gonorrheae/HIV co-infected males than the rest of the state. This “hot spot” extends in to the
lower zip codes of Oakland and Macomb counties. There is a second “hot spot” located in the zip
codes that make up the center right area of Oakland County. A total of 700 HIV positive males
with gonorrhea were able to be geocoded.

199

Figure 5.2: Getis-Ord GiZ Score Cluster Analysis of N. gonorrheae/HIV Co-Infected
Michigan Males 2011-2013 Detroit and Tri-County Area

Legend
Getis Ord
Gonorrhea/HIV
GiZ Score
1.65-1.65 Std. Dev
1.65-1.96 Std. Dev
1.96-2.58 Std. Dev
>2.58 Std. Dev
Detroit
Counties

Getis-Ord GiZ Score distribution zoomed in to show “hot spot” located in the City of Detroit that
extends into the Southern parts of Oakland and Macomb counties. The map also shows a
secondary “hot spot” located just east of the center of Oakland County. The City of Detroit is
represented by the shaded region in the North East Corner of Wayne County. A total of 700 HIV
positive males with gonorrhea were able to be geocoded.

200

Figure 5.3: Poisson Cluster Analysis of Gonorrhea/HIV Co-Infections in Michigan Males
from 2011 to 2013

Legend
Bernouilli High Only Cluster
HIV/Gonorrhea
Detroit
Counties
CLUSTER
Cluster 1 RR=12.71
p-value<0.0001
Observed=408
Expected=7.17
Cluster 2 RR=3.65
p-value<0.0001
Observed=237
Expected=88.08
Cluster 3 RR=2.71
p-value<0.0001
Observed=271
Expected=135.20
Poisson cluster analysis of Michigan gonorrhea/HIV infected males from 2011 to 2013. Cluster 1
which is mainly made up of the zip codes comprising the City of Detroit as well as zip codes in
the Southern adjacent edges of Oakland and Macomb counties represents the most likely cluster
with a RR=12.71 (p-value<0.0001). Cluster 2 represents another elevated risk area, RR=3.65 (pvalue=<0.0001) which is located in the remaining zip codes of Wayne County as well as in the
Northern part of Monroe County. Cluster 3 is comprised of the lower half of Oakland County,
parts of Livingston and Washtenaw counties and has a RR=2.71 (p-value<0.0001). Cluster 4 is
located in the middle of Kalamazoo County, between Van Buren and Calhoun counties, and has
a RR=3.02, however, this was not found to be statistically significant (p-value=0.971). Cluster 5
is located in Genesee County and has a RR=2.78 (p-value=0.998) which was not statistically
significant. A total of 700 HIV positive males with gonorrhea were able to be geocoded.

201

Figure 5.4: Poisson Cluster Analysis of Gonorrhea/HIV Co-Infections in Michigan Males
from 2011 to 2013 Detroit and Tri-County Area

Legend
Bernouilli High Only Cluster
HIV/Gonorrhea
Detroit
Counties
CLUSTER
Cluster 1 RR=12.71
p-value<0.0001
Observed=408
Expected=7.17
Cluster 2 RR=3.65
p-value<0.0001
Observed=237
Expected=88.08
Cluster 3 RR=2.71
p-value<0.0001
Observed=271
Expected=135.20
Poisson cluster analysis, zoomed into to focus on Cluster 1-3 which are mainly located in the
Detroit and Tri-County area, as well as extending into Livingston, Washtenaw, and Monroe
counties and represent areas of increased risk of gonorrhea/HIV infection. Clusters 4 and 5,
located in Kalamazoo and Genesee counties respectively were not found to have statistically
significant elevations in gonorrhea/HIV co-infections. A total of 700 HIV positive males with
gonorrhea were able to be geocoded.

202

Chapter 6.
Conclusion
6.1 Age at HIV Positive Diagnosis
The co-infection between N. gonorrheae and HIV may cause these two diseases to express
variant epidemiological patterns and therefore warrants further evaluation. Prevention efforts for
gonorrhea including education, screening, and treatment are well established yet not necessarily
easily implemented. In this dissertation, age at HIV diagnosis, age at AIDS, time from HIV to
AIDS, the geospatial patterns, and the change in biological markers were investigated in
Michigan males with and without a history of gonorrhea.
A synergistic relationship between N. gonorrheae and age at HIV diagnosis was observed. The
study found that when looking at individuals with a reported history of gonorrhea between 2005
and 2011 had a lower age at HIV diagnosis than males with no reported history of gonorrhea,
29.1 years old (in the adjusted model) compared to 36.3 years old in those with no N. gonorrheae
infection. This represents a possible opportunity to intervene and attempt to prevent HIV by
reaching out to males with N. gonorrheae infections in populations. Tracking N. gonorrheae
reported infections may lead to a window of opportunity for identifying males in high risk sexual
networks before they are exposed to HIV and become infected. N. gonorrheae infection is a
useful marker for risky sexual behaviors for various reasons including the short incubation time
(2-7 days in males (2)) as well as well-known symptoms including dysuria and discharge of pus
from the urethra (2).
The study also found that when looking at an individual’s complete reported N. gonorrheae
infection history between 2005 and 2011 there was a dose response with regard to number of N.
203

gonorrheae infection in the time period and reduction of age at HIV diagnosis. This finding was
important for several reasons. These findings again show that a useful addition to the techniques
used in HIV surveillance may be the identification of other STDs including bacterial infection as
markers for high risk groups. The dose response also shows that reducing the number if
infections as well as interrupting the infection of treated individuals by members of their sexual
network is important in not only controlling the burden of disease cause by gonorrhea but also
HIV as the more N. gonorrheae infections an individual has the younger they are when they are
diagnosed with HIV. Individuals with only one reported N. gonorrheae infection were found to
have a mean reduction in age of HIV diagnosis of 5.1 years (31.1 years old 95% CI 29.3-33) in
the adjusted model as compared to males with no reported history of N. gonorrheae infection
(36.2 years old), and age at HIV diagnosis was reduced by 10.9 years (25.3 years old 95% CI
20.6-30.0) in those with a reported history of 5 or more N. gonorrheae infections between 2005
and 2011. These findings are important because they show that intervening early in someone’s
STD history, at someone’s first N. gonorrheae infection as opposed to their fifth would may help
in increasing the age at when they may be diagnosed with HIV by almost 5 years. The study also
found that 37% (N=330) of those with gonorrhea between 2005 and 2011 had more than one
occurrence of the infection. This represents a key group to intervene and provide prevention
strategies to help to reduce re-infection rates.
The study also modeled pre-HIV only gonorrhea and the age at HIV diagnosis and found a
similar trend as compared to looking at any N. gonorrheae infection regardless of their state of
HIV. In those with no history of any N. gonorrheae infection (neither pre or post HIV infection)
the mean age at HIV diagnosis was 36.2 years old (in the adjusted model) and in those with only
one pre-HIV N. gonorrheae infection there was a reduction of 6.4 years (29.8 95% CI 28.2-31.4)
204

and in those with 2 or more pre-HIV N. gonorrheae infections the mean reduction in age at HIV
diagnosis was 9 years (26.2 years old 95% CI 24.7-29.6). This along with the two previous
analysis show that intervening in areas with N. gonorrheae outbreaks, or the STD is common
help aid in reducing the burden of disease due to HIV in those communities. These pre-HIV
infections represent real opportunities to intervene when individuals come in with their first
infection. Education and HIV prevention services should target at risk males defined as those
males with their first N. gonorrheae infection, or even their first reported STD infection.
6.2 Time to HIV Progression Event
Very few deaths in HIV positive males with a history of N. gonorrheae between 2005 and 2011
were observed in the State of Michigan. In individuals with no history of N. gonorrheae
infection the median age at death was 41 years old and the median time from HIV diagnosis to
death was observed to be 3 years. In individuals with a N. gonorrheae infection the median age
at death was observed to be 38.5 years, 2.5 years younger, however the median time from HIV
diagnosis to death was observed to be 5 years, 2 years longer than those with no history of N.
gonorrheae infection. This paradox, younger at age of event with longer time to event from HIV
diagnosis was observed in time to AIDS as well and may be related to the findings in the
laboratory markers of HIV progression where information on CD4 cell count, viral load, and
ever being on ARV treatment were analyzed. Results from that analysis showed that while only
around 8% of individuals reported that they ever were on ARV treatment, over 70% were
observed to start ARV treatment in the same year or after their infection with N. gonorrheae.
This may have led to the increase in the time from HIV infection to HIV progression event;
however, more research is needed to understand the rate of ARV usage, length of ARV

205

treatment, variances in ARV cocktails on laboratory markers of HIV progression and any
barriers to ARV and medical access Michigan HIV positive males may face.
Age at AIDS positive status as well as time from HIV to AIDS was also analyzed in this study.
The study found when looking at any reported history of N. gonorrheae infection that individuals
with no N. gonorrheae infection had a median age of AIDS positive status of 38 years old, with a
median time from HIV to AIDS of less than 1 year. In individuals with any history of N.
gonorrheae infection between 2005 and 2011 the median age at AIDS positive status was
observed to be 31 years old (HR=1.17 95% CI 1.06-1.30) with a median time from HIV to AIDS
positive status of 1 year (HRadj=0.63 95% CI 0.57-0.70). A dose response similar to that found
in age of HIV diagnosis was also observed in individuals with multiple N. gonorrheae infection.
The median age of AIDS positive status in individuals with only one N. gonorrheae infection
was 32 years old while the median age of AIDS positive status in individuals with 5 or more N.
gonorrheae infections was 24 years old. HR=1.01 (95% CI 0.89-1.15) for one N. gonorrheae
infection as compared to those with no gonorrhea and HR=2.47 (95% CI 1.54-3.97) in those with
5 or more N. gonorrheae infections. A minor inverse dose response was found when looking at
time from HIV diagnosis to AIDS positive status. Those with only one infection had a median
time from HIV to AIDS of 1 year and those with 5 or more infections had a median time of 2
years. The adjusted HR also showed a minor inverse dose response with one N. gonorrheae
infection HRadj=0.60 (95% CI 0.53-0.68) as compared to those with no N. gonorrheae infection
history and HRadj=0.58 (95% CI 0.36-0.94) in those with 5 or more N. gonorrheae infections.
While the biological markers of HIV progression may help to understand these paradoxical
findings, more complete data on medical history is needed to fully understand the increase in
time from HIV to AIDS and death.
206

The study also modeled post-HIV N. gonorrheae infections and age at AIDS positive status as
well as the time from HIV diagnosis to AIDS positive status. In individuals who had a N.
gonorrheae infection, but did not have a post-HIV N. gonorrheae infection the median age at
AIDS positive status was 29 years old (HR=0.95 95% CI 0.77-1.17) as compared to 38 years old
in those with no N. gonorrheae infection history showing that post-HIV gonorrhea may have a
more significant impact on HIV progression than pre-HIV infections. In those with only one
post-HIV N. gonorrheae infection the median age of AIDS positive status was 32 years old
(HR=1.10 95% CI 0.95-1.27). In those with two post-HIV N. gonorrheae infections the median
age of AIDS positive status was 30.5 years (HR=1.55 95% CI 1.22-1.97) and in those with three
or more post-HIV N. gonorrheae infection the median age was 26 years old (HR=2.06 95% CI
1.51-2.79). Again a seemingly paradoxical finding with regard to time from HIV to AIDS was
observed with 1-2 post-HIV N. gonorrheae infections having a time from HIV to AIDS of 2
years as opposed to less than one year in those with no post or no history of reported N.
gonorrheae infections and 3 years in those with 3 or more post-HIV N. gonorrheae infections.
Such observations may be attributed, in part, to: i) the fact that individuals with gonorrhea are
frequently treated with antimicrobials that can delay or slow the impact of infections that are
commonly end fatally in AIDS patients and ii) infection with N. gonorrheae may elicit a type
immune response that can slow the progression of the HIV infection. Additionally, as stated
previously this may be driven by what is happening with the laboratory markers for HIV
progression as well as by the finding that in those that reported being on ARV, 71% started
within the same calendar year or in the years after their N. gonorrheae infection.
6.3 County Level Geospatial Analysis

207

A chloropleth map of HIV rate per 100,000 in Michigan males age 0-65 was constructed and
showed that the areas with the highest rate per 100,000 in Michigan were Detroit as well as the
Tri-County area and a select handful of counties such as Genesee, Ingham, and Kalamazoo. This
map is important because it gives us a basis for the underlying distribution of HIV in the State of
Michigan males. The study was then able to map the number of N. gonorrheae infections per
county and show the underlying distribution of N. gonorrheae in HIV infected males. Again
Detroit and the Tri-County area have the highest number of N. gonorrheae infections. While
these two maps are very useful they do not show the complete picture of HIV/N. gonorrheae coinfections in Michigan males.
A map of gonorrhea per 1,000 in HIV positive males was also constructed to see if counties that
had a high rate of HIV per 100,000 as well as elevated number of N. gonorrheae infections also
had high rates of N. gonorrheae infections in HIV positive males. The study found that while
Detroit and the Tri-County area had high rates, such as Oakland that had one of the highest, other
counties with low HIV rates and fewer over all N. gonorrheae infections the rate of N.
gonorrheae in HIV males was elevated. This map shows that while overall numbers may be low,
in some counties there is a high rate of co-infected individuals which is an area of concern, as
well as possible opportunities for education and intervention. The study was also able to map the
percent of HIV/N. gonorrheae co-infected individuals with multiple N. gonorrheae infections in
the State of Michigan. This map presents an opportunity to investigate which counties may need
to focus on partner therapy to reduce the burden of disease on individuals how have had multiple
N. gonorrheae infections within the study period of 2005 to 2011. While Detroit and the TriCounty area have mid to high levels of repeated infections smaller counties like Calhoun, Van
Buren, and Lapeer have a higher percentage of individuals that may be at risk for re-occurring
208

infections. A similar map looking at percent of HIV positive males with a history of multiple N.
gonorrheae infections was also created. This showed similar results to that of the percent of
HIV/N. gonorrheae co-infected individuals with multiple N. gonorrheae infections. Counties
like Gratiot and Clinton have the highest percent of HIV men with multiple N. gonorrheae
infections, and Detroit and the Tri-County area also have mid to high rates. These maps show
that while Detroit and the Tri-County area is an important area to focus on treatment and
prevention, other counties may also have problems and need to evaluate how to control and
prevent males from having re-occurring infections.
Both Poisson high only cluster analysis of HIV as well as the high only cluster analysis of
HIV/N. gonorrheae co-infections show that only Detroit had an increase in relative risk,
RR=10.26 and RR=12.66 respectively, as well as being the most likely cluster. A second method
of cluster or “hot spot” analysis, Getis-Ord GiZ scoring, showed that there are two distinct “hot
spots” in the state with regard to HIV/N. gonorrheae co-infections. Getis-Ord analysis detected a
cluster group of Detroit, Wayne, Macomb, Oakland, Lapeer and St. Clair as well as a separate
cluster representing a group of counties near Gratiot, Clinton and Shiawassee counties. The
Getis-Ord analysis offers a different way to detect clusters that uses the surrounding neighbors as
reference points to detect “hot spots”.
The third and final cluster analysis that was used was the Bernoulli high only cluster analysis. In
the Bernoulli high only analysis three distinct high RR clusters were detected. Both the previous
clusters of Oakland and the cluster of Detroit, Huron, Macomb, Sanilac and Lapeer counties
were found to be statistically significant and together present one of the “hot spots” detected by
the Getis-Ord approach. The third cluster detected using Bernoulli high only consisted of Gratiot

209

and Clinton counties, similar to the second Getis-Ord “hot-spot”, but was not found to be
statistically significant.
Geospatial analysis of disease distribution and detection of disease clusters is an important tool
in understanding the underlying distribution and proliferation of the disease in a given
community. This study was able to show that there is not uniform co-infection of HIV
individuals with N. gonorrheae, and offers some insight into where education and prevention
programs should be aimed to gain the most benefit. Using this data it is clear that Detroit and the
Tri-County area are the “hot spots” for HIV/N. gonorrheae co-infections, but as the other maps
show, such as the Getis-Ord and Bernoulli maps, there are other counties and areas where while
the overall number of individuals may not be as high as a place like Detroit, there is still an
important population that is co-infected and needs to be addressed.
The previous findings from age at HIV diagnosis and age at AIDS positive status show that in
male with a history of N. gonorrheae the age of HIV and eventually AIDS is reduced as
compared to individuals who remain N. gonorrheae free. These maps offer important insight into
where N. gonorrheae interventions should be placed to reduce the number of HIV infected males
with gonorrhea, as well as where these individuals are more likely to have multiple infections
thus increasing their hazard for HIV and AIDS.
6.4 Zip Code Level Geospatial Analysis
In order to better determine where education and prevention services should be targeted the study
was also able to do geospatial analysis at the zip code level using a separate population of HIV
positive males who had a N. gonorrheae infection between 2011 and 2013. Population level data
on the zip codes were adjusted for a male only population and for an age range of 0-65 years old
210

using the 2008-2012 American Community Survey 5-year Estimates from the US Census. Two
cluster detection techniques were used, Getis-Ord and Poisson. The Getis-Ord GiZ Score
analysis found two “hot spots” one being the City of Detroit and some of the zip codes that
bordered the Northern edge of the city and extended into Oakland and Macomb counties and a
separate distinct “hot spot” located in the center of Oakland County. These findings matched up
well with the previous Getis-Ord county findings, but were able to identify at the zip code level
where in a county like Oakland the burden of disease due to HIV/N. gonorrheae co-infections is
located. The Poisson analysis found 5 distinct clusters with the most likely cluster being the City
of Detroit and the Northern zip codes, closely approximating what was found using the GetisOrd approach. A second cluster was also detected that consisted of Wayne and Monroe county
zip codes and was found to have a statistically significant increase in RR. A third cluster which
included Oakland, Livingston, and Washtenaw counties was also identified and had a statistically
significant increase in relative risk. Two other clusters, which included zip codes in Kalamazoo
County and Genesee County had an increase in relative risk, but were not found to be
statistically significant. Zip code level analysis allowed for better detection of possible clusters as
well as visually represented the reduction in HIV/N. gonorrheae co-infection as the distance
from Detroit increases. Working at the zip code level will allow for epidemiologists and public
health officials to better target at risk areas with prevention and education strategies.
6.5 Biological Markers
CD4 cell count was found to increase in the year of N. gonorrheae infection as well as remain
higher than pre-gonorrhea states in the years after. A rise in CD4 cell count in HIV positive
males during the year that coincided with N. gonorrheae infection would most likely be due to
the CD4 cell response of a bacterial STD. CD4 remaining elevated from pre-gonorrheal state
211

may be due to a number of factors including re-entry into public health/access to health care as
well as starting ARV therapy. The study found that among the 8% that reported ever being on
ARV treatment, 71% started in the same year or after their N. gonorrheae infection. This finding
may also help shed light on the seemingly paradoxical findings of increased time from HIV
diagnosis to HIV progression event (AIDS or death) as seen in Chapter 3. This increase and lack
of decrease in the years after may be an important factor in understanding the progression of the
disease as well as how re-occurring reported infections, and thus re-occurring contact with public
health or a health care provider may alter the natural history of HIV.
Viral load was also found to increase sharply in the year of a N. gonorrheae infection, but did not
remain elevated in the years post. This again may be related to access to medical care or the
initiation of ARV treatment post infection. This finding does however show that in the year
during a N. gonorrheae infection there is increased viral load in an individual making them at a
higher risk for transmitting the disease to others. This is an important finding because it shows
there in HIV positive males; HIV is enhanced during the time period surrounding co-infection.
More research is needed and there is a need to run these types of models with either the date of
onset of gonorrhea symptoms, or the date gonorrhea was confirmed and reported to gain a better
understanding of the rise and fall of Laboratory markers such as CD4 cell count and HIV viral
load in the time period leading up to N. gonorrheae infection, during treatment, and post
infection. There is also a need to collect better data on ARV usage and when individuals start and
end. Finally since ARV treatment relies on a cocktail of drugs, it is also important to know what
series of drugs individuals are on to assess if there are variances in ARV treatment around coinfection periods.

212

6.6 Conclusions
This study showed that risky sexual behavior, as measured by reported gonorrhea, results in a
reduction in the time an individual remains HIV free. The study also found that as number of N.
gonorrheae infections in an individual increases so does their hazard of being diagnosed with
HIV. The study also showed that this trend persisted when looking at pre-HIV gonorrhea, as well
as when looking at age of death and AIDS positive status. A seemingly paradoxical finding of
increased time from HIV diagnosis to HIV progression event was also observed. This finding
along with the increase and lack of decrease in CD4 cell count in the laboratory markers analysis
may represent possible unmeasured confounding most likely due to the lack of individuals
reporting if they are or have been on ARV treatment. These findings could also be influenced by
the re-entry into HIV care as represented by reporting of a N. gonorrheae infection. It should be
noted that for time from HIV diagnosis to HIV progression event all time was measured as the
year an event occurred and the difference in times was around 1 year. Using full dates,
mm/dd/yyyy for this type of analysis in the future and calculating days instead of years, along
with better information on ARV treatment status may help to better define this time window. The
study also found that during the year of N. gonorrheae infection was recorded viral load
increased representing a time period of increased risk for spreading the disease.
Finally the study was able to map out possible clusters and “hot spots” at the county and in a
smaller population at the zip code level. These maps show that the major area of concern for
HIV/N. gonorrheae co-infections in the state of Michigan in males from 2005-2013 is in the
Detroit and Tri-County areas, with a few small (non -statistically clusters in other counties).
These maps along with the information gained in the previous chapters allow for better
placement of intervention efforts in high risk areas.
213

6.7 Limitations
Limitations of this study include only being able to gain access to data on gonorrhea. Bacterial
STDs promote HIV in infecting white blood cells (1, 24, 27-29, 31) and other STDs that were
not included could have had an influence on HIV seroconversion. Gonorrheal infection was used
as a proxy for risky sexual behavior and a marker for individuals being in sexual networks where
STDs were present. Because this study could not control for the effect of other STDs on age at
HIV diagnosis, time from HIV to AIDS, age at AIDS, or changes in laboratory markers the
resulting synergistic relationship is only the start of exploring HIV and STD co-infections.
Another limitation of this study HIV diagnosis, AIDS positive status, N. gonorrheae infection
and death were all only coded as year of event. In analysis such as time from HIV to HIV
progression event using days or weeks may have been more appropriate. A final limitation
regarding the non-spatial analysis sections is that HIV diagnosis is not a proxy for HIV infection,
and other than HIV screening tests, there is little data to determine length of time an individual
may be been incubating HIV and not aware of their infection status.
A major limitation in the geospatial analysis section is that locations represent where the
individual was living and may not reflect where they were infected with either HIV or N.
gonorrheae. While the study set 20 mile windows for cluster analysis individuals may travel
much further distances and become infected out of their zip code of residency or even out of
their county of residence. Another limitation is that a number of infections were diagnosed in the
prison system, but this does not necessarily mean infection happened in prison, and therefore
place of residence would have been an acceptable substitute for prison diagnosed. A final
limitation of the geospatial analysis only a small subset of the data had zip code level data. While
county level data is a great tool, the study was able to show that similarly to state level, co214

infections are not equally distributed across a county. Having more years of zip code level data
may have been able to give a time series analysis of the trends in co-infections over the last
decade or half decade.
N. gonorrheae and other STIs are reportable diseases, and resources should be given to studying
their epidemiological patterns in co-infected individuals. Identification of possible risk factors
that may allow for intervention are important tools in trying to prevent the spread not only of
HIV but other STDs as well. Education and counseling efforts should be focused on those with
their first STD as well as those who are HIV naïve and have multiple reported STD infections as
a HIV prevention tool. Education and counseling has to walk the fine line between
destigmatization of STDs (so individuals get tested and do not feel ashamed if they are positive)
and warning individuals about the dangers of these diseases. As individuals within public health
we must change the paradigm from “you just have gonorrhea”, to the sexual behaviors that lead
to you getting gonorrhea puts you at risk for HIV and other sexually transmitted diseases which
may not be curable.

215

REFERENCES

216

REFERENCES

1. Talaro, Kathleen Park (2005). Foundations in Microbiology (5th ed). New York, NY: McGraw
Hill pp 566-569, 766-776
2. Heymann, David L Editor (2008). Control of Communicable Diseases Manual (19th ed.).
Washington, DC: American Public Health Association pp 1-9.
3. Schneider, E., S. Whitmore, et al. (2008). "Revised surveillance case definitions for HIV
infection among adults, adolescents, and children aged <18 months and for HIV infection and
AIDS among children aged 18 months to <13 years--United States, 2008." MMWR Recomm
Rep 57(RR-10): 1-12.
4. MDCH, Michigan Department of Community Health, Quarterly HIV Surveillance Report,
Michigan April 2013. Available at:
http://www.michigan.gov/documents/mdch/April_2013_ALL_418616_7.pdf [Accessed on
October 7, 2013]
5. Centers for Disease, C. and Prevention (2001). "HIV and AIDS--United States, 1981-2000."
MMWR Morb Mortal Wkly Rep 50(21): 430-434.
6. WHO, World Health Organization, HIV/AIDS. Available at: http://www.who.int/hiv/en/
[Accessed on October 7, 2013]
7. CDC, Centers for Disease Control and Prevention. HIV in the United States. Available at:
http://www.cdc.gov/hiv/resources/factsheets/us.htm [Accessed on February 28, 2012 ]
8. CDC, Centers for Disease Control and Prevention. Statistics Overview. Available at:
http://www.cdc.gov/hiv/statistics/basics/index.html. [Accessed on May 11, 2014]
9. KFF, The Henry J. Kaiser Family Foundation. Estimated Numbers of AIDS Diagnosis, All
Ages. Available at: http://kff.org/hivaids/state-indicator/estimated-number-of-aids-diagnosesadults-and-adolescents/. [Accessed on May 11, 2014]
10. MDCH, Michigan Department of Community Health. Annual HIV Surveillance Report,
Michigan January 2014. Available at:
http://www.michigan.gov/documents/mdch/January_2014_ALL_446611_7.pdf. [Accessed on
February 10, 2014]
11. CDC, Centers for Disease Control and PreventioN. gonorrheae- CDC Fact Sheet. Available
at: http://www.cdc.gov/std/gonorrhea/stdfact-gonorrhea.htm [Accessed on February 28, 2012]

217

12. WHO, World Health Organization. Emergence of Multi-Drug Resistant Neisseria
gonorrhea- Threat of Global Rise in Untreatable Sexually Transmitted Infections. Available at:
http://whqlibdoc.who.int/hq/2011/WHO_RHR_11.14_eng.pdf?ua=1 [Accessed on May 14,
2014]
13. CDC, Centers for Disease Control and Prevention. 2012 Sexually Transmitted Disease
Surveillance: Gonorrhea. Available at: http://www.cdc.gov/std/stats12/gonorrhea.htm [Accessed
on May 15, 2014]
14. MDCH, Michigan Department of Community Health, 2012 Epidemiologic Profile of HIV in
Michigan. Available at:
http://www.michigan.gov/documents/mdch/2012_Epi_Profile_FULL_403522_7.pdf [Accessed
on October 7, 2013]
15. Kahn, R. H., D. J. Mosure, et al. (2005). "Chlamydia trachomatis and Neisseria gonorrhoeae
prevalence and Co-infection in adolescents entering selected US juvenile detention centers,
1997-2002." Sex Transm Dis 32(4): 255-259]
16. Herbst, J. H., T. M. Painter, et al. (2014). "Evidence-Based HIV/STD Prevention
Intervention for Black Men Who Have Sex with Men." MMWR Surveill Summ 63: 21-27.
17. Tzeng, J, et al. (2013). "Epidemiology of Sexually Transmitted Infections among Human
Immunodeficiency Virus Positive United States Military Personnel." Journal of Sexually
Transmitted Diseases. 2013 Vol 2013
18. Torian, L. V., H. A. Makki, et al. (2002). "HIV infection in men who have sex with men,
New York City Department of Health sexually transmitted disease clinics, 1990-1999: a decade
of serosurveillance finds that racial disparities and associations between HIV and gonorrhea
persist." Sex Transm Dis 29(2): 73-78.
19. Millett, G. A., J. L. Peterson, et al. (2006). "Greater risk for HIV infection of black men who
have sex with men: a critical literature review." Am J Public Health 96(6): 1007-1019.
20. Wolitski, R. J., R. O. Valdiserri, et al. (2001). "Are we headed for a resurgence of the HIV
epidemic among men who have sex with men?" Am J Public Health 91(6): 883-888.
21. Goodenow, C., J. Netherland, et al. (2002). "AIDS-related risk among adolescent males who
have sex with males, females, or both: evidence from a statewide survey." Am J Public Health
92(2): 203-210.
22. Hall, H. I., R. Song, et al. (2008). "Estimation of HIV incidence in the United States." JAMA
300(5): 520-529
23. Lafferty, W. E., J. P. Hughes, et al. (1997). "Sexually transmitted diseases in men who have
sex with men. Acquisition of gonorrhea and nongonococcal urethritis by fellatio and implications
for STD/HIV prevention." Sex Transm Dis 24(5): 272-278.
218

24. (1998). "HIV prevention through early detection and treatment of other sexually transmitted
diseases--United States. Recommendations of the Advisory Committee for HIV and STD
prevention." MMWR Recomm Rep 47(RR-12): 1-24.
25. Zhang, X., C. Wang, et al. (2007). "Risk factors of HIV infection and prevalence of coinfections among men who have sex with men in Beijing, China." AIDS 21 Suppl 8: S53-57.
26. Manhart, L. E., S. O. Aral, et al. (2004). "Influence of study population on the identification
of risk factors for sexually transmitted diseases using a case-control design: the example of
gonorrhea." Am J Epidemiol 160(4): 393-402.
27. Zhang, J., G. Li, et al. (2005). "Neisseria gonorrhoeae enhances infection of dendritic cells by
HIV type 1." J Immunol 174(12): 7995-8002.
28. Ding, J., A. Rapista, et al. (2010). "Neisseria gonorrhoeae enhances HIV-1 infection of
primary resting CD4+ T cells through TLR2 activation." J Immunol 184(6): 2814-2824.
29. Galvin, S. R. and M. S. Cohen (2004). "The role of sexually transmitted diseases in HIV
transmission." Nat Rev Microbiol 2(1): 33-42
30. Montano, M., M. Rarick, et al. (2006). "HIV-1 burden influences host response to coinfection with Neisseria gonorrhoeae in vitro." Int Immunol 18(1): 125-137.
31. Mushayabasa, S., J. M. Tchuenche, et al. (2011). "Modeling gonorrhea and HIV cointeraction." Biosystems 103(1): 27-37.
32. Liu, X., A. Mosoian, et al. (2006). "Gonococcal lipooligosaccharide suppresses HIV
infection in human primary macrophages through induction of innate immunity." J Infect Dis
194(6): 751-759.
33. Taylor, M. M., J. A. Schillinger, et al. (2013). "Gonorrhea infections diagnosed among
persons living with HIV/AIDS: identifying opportunities for integrated prevention services in
New York City, Washington, DC, Miami/Dade County, and Arizona." J Acquir Immune Defic
Syndr 64(1): 115-120.
34. Fox, K. K., C. del Rio, et al. (2001). "Gonorrhea in the HIV era: a reversal in trends among
men who have sex with men." Am J Public Health 91(6): 959-964.
35. Stenger, M, et al. (2009). "Trends in Neisseria gonorrhoeae Incidence Among HIV-Negative
and HIV-Positive Men in Washington State, 1996-2007." Public Health Reports 2009
Supplement 2, Vol 124, P18-23
36. Peralta, L., S. J. Durako, et al. (2001). "Correlation between urine and cervical specimens for
the detection of cervical Chlamydia trachomatis and Neisseria gonorrhoeae using ligase chain

219

reaction in a cohort of HIV infected and uninfected adolescents." J Adolesc Health 29(3 Suppl):
87-92.
37. Rietmeijer, C. A., J. L. Patnaik, et al. (2003). "Increases in gonorrhea and sexual risk
behaviors among men who have sex with men: a 12-year trend analysis at the Denver Metro
Health Clinic." Sex Transm Dis 30(7): 562-567.
38. Rottingen, J. A., D. W. Cameron, et al. (2001). "A systematic review of the epidemiologic
interactions between classic sexually transmitted diseases and HIV: how much really is known?"
Sex Transm Dis 28(10): 579-597
39. Donovan, B. (2004). "Sexually transmissible infections other than HIV." Lancet 363(9408):
545-556
40. Mayaud, P. and D. McCormick (2001). "Interventions against sexually transmitted infections
(STI) to prevent HIV infection." Br Med Bull 58: 129-153
41. Stekler, J., L. Bachmann, et al. (2005). "Concurrent sexually transmitted infections (STIs) in
sex partners of patients with selected STIs: implications for patient-delivered partner therapy."
Clin Infect Dis 40(6): 787-793
42. Darrow, W. W., D. F. Echenberg, et al. (1987). "Risk factors for human immunodeficiency
virus (HIV) infections in homosexual men." Am J Public Health 77(4): 479-483
43. Dougan, S., B. G. Evans, et al. (2007). "Sexually transmitted infections in Western Europe
among HIV-positive men who have sex with men." Sex Transm Dis 34(10): 783-790
44. Royce, R. A., A. Sena, et al. (1997). "Sexual transmission of HIV." N Engl J Med 336(15):
1072-1078
45. Bentwich, Z., G. Maartens, et al. (2000). "Concurrent infections and HIV pathogenesis."
AIDS 14(14): 2071-2081
46. Schwartz, R. H. (2003). "T cell anergy." Annu Rev Immunol 21: 305-334.
47. Risbud, A. (2005). "Human immunodeficiency virus (HIV) & sexually transmitted diseases
(STDs)." Indian J Med Res 121(4): 369-376.
48. Weinstock, H., S. Berman, et al. (2004). "Sexually transmitted diseases among American
youth: incidence and prevalence estimates, 2000." Perspect Sex Reprod Health 36(1): 6-10.
49. Buchacz, K., P. Patel, et al. (2004). "Syphilis increases HIV viral load and decreases CD4
cell counts in HIV-infected patients with new syphilis infections." AIDS 18(15): 2075-2079.
50. Berglund, T., T. Asikainen, et al. (2007). "The epidemiology of gonorrhea among men who
have sex with men in Stockholm, Sweden, 1990-2004." Sex Transm Dis 34(3): 174-179

220

51. Do, A. N., D. L. Hanson, et al. (2001). "Risk factors for and trends in gonorrhea incidence
among persons infected with HIV in the United States." AIDS 15(9): 1149-1155.
52. Gunn, R. A., C. J. O'Brien, et al. (2008). "Gonorrhea screening among men who have sex
with men: value of multiple anatomic site testing, San Diego, California, 1997-2003." Sex
Transm Dis 35(10): 845-848
53. Vermund, S. H., C. M. Wilson, et al. (2001). "Sexually transmitted infections among HIV
infected and HIV uninfected high-risk youth in the REACH study. Reaching for Excellence in
Adolescent Care and Health." J Adolesc Health 29(3 Suppl): 49-56.
54. Kilmarx, P. H. (2009). "Global epidemiology of HIV." Curr Opin HIV AIDS 4(4): 240-246.
55. Easterbrook, P. J., J. S. Chmiel, et al. (1993). "Racial and ethnic differences in human
immunodeficiency virus type 1 (HIV-1) seroprevalence among homosexual and bisexual men.
The Multicenter AIDS Cohort Study." Am J Epidemiol 138(6): 415-429.
56. Babiker, A., J. Darbyshire, et al. (2002). "Changes over calendar time in the risk of specific
first AIDS-defining events following HIV seroconversion, adjusting for competing risks." Int J
Epidemiol 31(5): 951-958
57. Saah, A. J., A. Munoz, et al. (1992). "Predictors of the risk of development of acquired
immunodeficiency syndrome within 24 months among gay men seropositive for human
immunodeficiency virus type 1: a report from the Multicenter AIDS Cohort Study." Am J
Epidemiol 135(10): 1147-1155.
58. Kaslow, R. A., D. G. Ostrow, et al. (1987). "The Multicenter AIDS Cohort Study: rationale,
organization, and selected characteristics of the participants." Am J Epidemiol 126(2): 310-318.
59. (1998). "The AIDS incubation period in the UK estimated from a national register of HIV
seroconverters. UK Register of HIV Seroconverters Steering Committee." AIDS 12(6): 659-667
60. Chmiel, J. S., R. Detels, et al. (1987). "Factors associated with prevalent human
immunodeficiency virus (HIV) infection in the Multicenter AIDS Cohort Study." Am J
Epidemiol 126(4): 568-577
61. De Gruttola, V. and X. M. Tu (1994). "Modelling progression of CD4-lymphocyte count and
its relationship to survival time." Biometrics 50(4): 1003-1014
62. Operskalski, E. A., D. O. Stram, et al. (1995). "Human immunodeficiency virus type 1
infection: relationship of risk group and age to rate of progression to AIDS. Transfusion Safety
Study Group." J Infect Dis 172(3): 648-655.
63. Alcabes, P., P. Pezzotti, et al. (1997). "Long-term perspective on the prevalent-cohort biases
in studies of human immunodeficiency virus progression." Am J Epidemiol 146(7): 543-551

221

64. Schwarcz, S. K., L. C. Hsu, et al. (2000). "Impact of protease inhibitors and other
antiretroviral treatments on acquired immunodeficiency syndrome survival in San Francisco,
California, 1987-1996." Am J Epidemiol 152(2): 178-185.
65. (2000). "Time from HIV-1 seroconversion to AIDS and death before widespread use of
highly-active antiretroviral therapy: a collaborative re-analysis. Collaborative Group on AIDS
Incubation and HIV Survival including the CASCADE EU Concerted Action. Concerted Action
on SeroConversion to AIDS and Death in Europe." Lancet 355(9210): 1131-1137.
66. Chene, G., C. Binquet, et al. (1998). "Changes in CD4+ cell count and the risk of
opportunistic infection or death after highly active antiretroviral treatment. Groupe
d'Epidemiologie Clinique du SIDA en Aquitaine." AIDS 12(17): 2313-2320
67. Fu, T. C., R. P. Westergaard, et al. (2012). "Changes in sexual and drug-related risk behavior
following antiretroviral therapy initiation among HIV-infected injection drug users." AIDS
26(18): 2383-2391
68. Bernstein, K. T., F. C. Curriero, et al. (2004). "Defining core gonorrhea transmission
utilizing spatial data." Am J Epidemiol 160(1): 51-58.
69. Jennings, J. M., F. C. Curriero, et al. (2005). "Geographic identification of high gonorrhea
transmission areas in Baltimore, Maryland." Am J Epidemiol 161(1): 73-80.
70. Turner, B. J., L. Markson, et al. (1996). "Estimation of survival after AIDS diagnosis: CD4 T
lymphocyte count versus clinical severity." J Clin Epidemiol 49(1): 59-65
71. Boscardin, W. J., J. M. Taylor, et al. (1998). "Longitudinal models for AIDS marker data."
Stat Methods Med Res 7(1): 13-27.
72. O'Brien, W. A., P. M. Hartigan, et al. (1996). "Changes in plasma HIV-1 RNA and CD4+
lymphocyte counts and the risk of progression to AIDS. Veterans Affairs Cooperative Study
Group on AIDS." N Engl J Med 334(7): 426-431.
73. Zeger, S. L. and P. J. Diggle (1994). "Semiparametric models for longitudinal data with
application to CD4 cell numbers in HIV seroconverters." Biometrics 50(3): 689-699.
74. O'Brien, T. R., D. R. Hoover, et al. (1995). "Evaluation of secular trends in CD4+
lymphocyte loss among human immunodeficiency virus type 1 (HIV-1)-infected men with
known dates of seroconversion." Am J Epidemiol 142(6): 636-642.
75. Yerly, S., T. V. Perneger, et al. (1998). "A critical assessment of the prognostic value of
HIV-1 RNA levels and CD4+ cell counts in HIV-infected patients. The Swiss HIV Cohort
Study." Arch Intern Med 158(3): 247-252.
76. Yu, L. M., P. J. Easterbrook, et al. (1997). "Relationship between CD4 count and CD4% in
HIV-infected people." Int J Epidemiol 26(6): 1367-1372.

222

77. Taylor, J. M., J. L. Fahey, et al. (1989). "CD4 percentage, CD4 number, and CD4:CD8 ratio
in HIV infection: which to choose and how to use." J Acquir Immune Defic Syndr 2(2): 114-124
78. Kitahata, M. M., S. J. Gange, et al. (2009). "Effect of early versus deferred antiretroviral
therapy for HIV on survival." N Engl J Med 360(18): 1815-1826.
79. Miller, V., A. Mocroft, et al. (1999). "Relations among CD4 lymphocyte count nadir,
antiretroviral therapy, and HIV-1 disease progression: results from the EuroSIDA study." Ann
Intern Med 130(7): 570-577.
80. Saag, M. S., M. J. Crain, et al. (1991). "High-level viremia in adults and children infected
with human immunodeficiency virus: relation to disease stage and CD4+ lymphocyte levels." J
Infect Dis 164(1): 72-80.
81. Kaufmann, D., G. Pantaleo, et al. (1998). "CD4-cell count in HIV-1-infected individuals
remaining viraemic with highly active antiretroviral therapy (HAART). Swiss HIV Cohort
Study." Lancet 351(9104): 723-724.
82. Ledergerber, B., M. Egger, et al. (1999). "Clinical progression and virological failure on
highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. Swiss HIV
Cohort Study." Lancet 353(9156): 863-868.
83. Autran, B., G. Carcelain, et al. (1997). "Positive effects of combined antiretroviral therapy on
CD4+ T cell homeostasis and function in advanced HIV disease." Science 277(5322): 112-116.
84. Kaufmann, G. R., M. Bloch, et al. (2002). "The extent of HIV-1-related immunodeficiency
and age predict the long-term CD4 T lymphocyte response to potent antiretroviral therapy."
AIDS 16(3): 359-367
85. Porter, K., A. Babiker, et al. (2003). "Determinants of survival following HIV-1
seroconversion after the introduction of HAART." Lancet 362(9392): 1267-1274.
86. Longini, I. M., Jr., W. S. Clark, et al. (1993). "Effect of routine use of therapy in slowing the
clinical course of human immunodeficiency virus (HIV) infection in a population-based cohort."
Am J Epidemiol 137(11): 1229-1240.
87. Binquet, C., G. Chene, et al. (2001). "Modeling changes in CD4-positive T-lymphocyte
counts after the start of highly active antiretroviral therapy and the relation with risk of
opportunistic infections: the Aquitaine Cohort, 1996-1997." Am J Epidemiol 153(4): 386-393
88. Mellors, J. W., J. B. Margolick, et al. (2007). "Prognostic value of HIV-1 RNA, CD4 cell
count, and CD4 Cell count slope for progression to AIDS and death in untreated HIV-1
infection." JAMA 297(21): 2349-2350.

223

89. Guttmacher Institute. Facts on Sexually Transmitted Infections in the United States.
Available at: http://www.guttmacher.org/pubs/2009/06/09/FIB_STI_US.pdf [Accessed on
January 10, 2014]
90. Bafica, A., C. A. Scanga, et al. (2004). "Influence of coinfecting pathogens on HIV
expression: evidence for a role of Toll-like receptors." J Immunol 172(12): 7229-7234.
91. S. Modrow et al., Molecular Virology, DOI 10.1007/978-3-642-20718-1_8, Springer-Verlag
Berlin Heidelberg 2013
92. Belongia, E. A., R. N. Danila, et al. (1997). "A population-based study of sexually
transmitted disease incidence and risk factors in human immunodeficiency virus-infected
people." Sex Transm Dis 24(5): 251-256.
93. Cohen, M. S. and W. C. Miller (1998). "Sexually transmitted diseases and human
immunodeficiency virus infection: cause, effect, or both?" Int J Infect Dis 3(1): 1-4.
94. Corbett, E. L., R. W. Steketee, et al. (2002). "HIV-1/AIDS and the control of other infectious
diseases in Africa." Lancet 359(9324): 2177-2187.
95. Wasserheit, J. N. (1992). "Epidemiological synergy. Interrelationships between human
immunodeficiency virus infection and other sexually transmitted diseases." Sex Transm Dis
19(2): 61-77.
96. Huhn, G. D., A. F. McIntyre, et al. (2008). "Factors associated with newly diagnosed HIV
among persons with concomitant sexually transmitted diseases." Sex Transm Dis 35(8): 731-737.
97. Fleming, D. T. and J. N. Wasserheit (1999). "From epidemiological synergy to public health
policy and practice: the contribution of other sexually transmitted diseases to sexual transmission
of HIV infection." Sex Transm Infect 75(1): 3-17.
98. Ghys, P. D., K. Fransen, et al. (1997). "The associations between cervicovaginal HIV
shedding, sexually transmitted diseases and immunosuppression in female sex workers in
Abidjan, Cote d'Ivoire." AIDS 11(12): F85-93
99. Patrick, M. E., P. M. O'Malley, et al. (2012). "HIV/AIDS risk behaviors and substance use by
young adults in the United States." Prev Sci 13(5): 532-538.
100. Hughes, A. K. (2011). "HIV knowledge and attitudes among providers in aging: results
from a national survey." AIDS Patient Care STDS 25(9): 539-545.
101. Mayer, K. H. and K. K. Venkatesh (2011). "Interactions of HIV, other sexually transmitted
diseases, and genital tract inflammation facilitating local pathogen transmission and acquisition."
Am J Reprod Immunol 65(3): 308-316.

224

102. Karlovsky, M., B. Lebed, et al. (2004). "Increasing incidence and importance of HIV/AIDS
and gonorrhea among men aged >/= 50 years in the US in the era of erectile dysfunction
therapy." Scand J Urol Nephrol 38(3): 247-252.
103. Kaul, R., J. Kimani, et al. (2004). "Monthly antibiotic chemoprophylaxis and incidence of
sexually transmitted infections and HIV-1 infection in Kenyan sex workers: a randomized
controlled trial." JAMA 291(21): 2555-2562.
104. Klotman, M. E., A. Rapista, et al. (2008). "Neisseria gonorrhoeae-induced human defensins
5 and 6 increase HIV infectivity: role in enhanced transmission." J Immunol 180(9): 6176-6185.
105. Nelson, K. E., D. Vlahov, et al. (1995). "Temporal trends of incident human
immunodeficiency virus infection in a cohort of injecting drug users in Baltimore, Md." Arch
Intern Med 155(12): 1305-1311.
106. Ogawa, Y., T. Kawamura, et al. (2009). "Gram-positive bacteria enhance HIV-1
susceptibility in Langerhans cells, but not in dendritic cells, via Toll-like receptor activation."
Blood 113(21): 5157-5166.
107. Funderburg, N. T., J. K. Jadlowsky, et al. (2011). "The Toll-like receptor 1/2 agonists
Pam(3) CSK(4) and human beta-defensin-3 differentially induce interleukin-10 and nuclear
factor-kappaB signalling patterns in human monocytes." Immunology 134(2): 151-160.
108. Drake, M. G., S. E. Evans, et al. (2013). "Toll-like receptor-2/6 and Toll-like receptor-9
agonists suppress viral replication but not airway hyperreactivity in guinea pigs." Am J Respir
Cell Mol Biol 48(6): 790-796.
109. Chesson, H. W., R. D. Kirkcaldy, et al. (2014). "Ciprofloxacin resistance and gonorrhea
incidence rates in 17 cities, United States, 1991-2006." Emerg Infect Dis 20(4): 612-619.
110. Whiteside, Y. O., S. M. Cohen, et al. (2014). "Progress along the continuum of HIV care
among blacks with diagnosed HIV- United States, 2010." MMWR Morb Mortal Wkly Rep
63(5): 85-89.
111. MDCH, Michigan Department of Community Health. Annual HIV Surveillance Report,
Michigan July 2014. Available at:
http://www.michigan.gov/documents/mdch/July_2014_full_report_465192_7.pdf [Accessed on
9/13/14]
112. Nelson, KE., Williams CM., Graham NMH. Infectious Disease Epidemiology Theory and
Practice. Jones and Bartlett Publishers 2005, Sudbury, MA pp 4.
113. Wylie, J. L., T. Cabral, et al. (2005). "Identification of networks of sexually transmitted
infection: a molecular, geographic, and social network analysis." J Infect Dis 191(6):
899-906.
225

114. Tanser, F., T. Barnighausen, et al. (2009). "Localized spatial clustering of HIV infections in
a widely disseminated rural South African epidemic." Int J Epidemiol 38(4): 1008-1016.
115. Westercamp, N., S. Moses, et al. (2010). "Spatial distribution and cluster analysis of sexual
risk behaviors reported by young men in Kisumu, Kenya." Int J Health Geogr 9: 24.
116. Wand, H. and G. Ramjee (2010). "Targeting the hotspots: investigating spatial and
demographic variations in HIV infection in small communities in South Africa." J Int
AIDS Soc 13: 41.

226