This is to certify that the

thesis entitled

Vaccine Safety - A Review and A Study
of Newborn Hepatitis B Vaccine Coverage
During a Period of Changing Recommendations

presented by

Brian Joseph Biroscak

has been accepted towards fulfillment
of the requirements for

MS degree in _Ep:l.demiology

 

Major professor

Date September 26, 2002

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

LIBRARY

Michigan State

University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

Apt; 0 3 ma

 

032005

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue.p65-p.15

 

VACCINE SAFETY — A REVIEW AND A STUDY OF NEWBORN HEPATITIS B
VACCINE COVERAGE DURING A PERIOD OF CHANGING
RECOMMENDATIONS
By

Brian Joseph Biroscak

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Epidemiology

2002

ABSTRACT
VACCINE SAFETY — A REVIEW AND A STUDY OF NEWBORN HEPATITIS B
VACCINE COVERAGE DURING A PERIOD OF CHANGING
RECOMMENDATIONS
By

Brian Joseph Biroscak

Thanks to vaccines, many once-prevalent devastating diseases are now rarely seen
in the United States. Despite the remarkable success of the US. National Immunization
Program, individual safety concerns about vaccines have become a leading topic of
public debate.

Temporary changes in the hepatitis B vaccination schedule due to concerns about
thimerosal used as a preservative in that vaccine led to the Michigan Thimerosal-Impact
Study - an observational study designed to assess the impact of recommendation changes
on the proportion of newborns receiving the hepatitis B vaccine before hospital
discharge. This study focused on births where the mother’s hepatitis B surface antigen
(HBsAg) status was unknown, because infants born to women who are not screened for
HBsAg are at greater risk of perinatal infection (compared to screened mothers).
Furthermore, this study targeted births where the mother’s HBSAg status was unknown
rather than positive because the Michigan Department of Community Health has a
perinatal hepatitis B coordinator that already tracks births to HBsAg-positive women to
ensure these infants are treated accordingly. This study’s results indicate a need to renew
efforts to begin hepatitis B vaccination at birth, especially for infants at higher risk of

infection.

ACKNOWLEDGEMENTS

I would like to thank the Official members of my thesis committee for all of their
time and effort in helping to refine this manuscript. The official members of my thesis
committee were: Dr. Michael Collins, Michigan State University; Dr. Anthony Fiore,
Centers for Disease Control and Prevention; and Dr. Wilfried Karmaus, Michigan State
University. Thanks also to Dr. Thomas Saari, University of Wisconsin, for participating
in thesis discussions, reviewing drafts, and sharing slides for the thesis defence.

I would especially like to thank Dr. Michael Collins, who not only served as the
chairperson of my thesis committee, but also was my academic advisor. His guidance and

sense of humor made this entire journey a pleasurable one — thank you, Dr. Collins.

iii

TABLE OF CONTENTS

LIST OF TABLES ........................................................... : ........................................ v
INTRODUCTION ..................................................................................................... 1
Overview of the remainder of this thesis ...................................................... 6
VACCINE SAFETY CONCERNS ........................................................................... 7
Polio vaccine ................................................................................................. 9
Yellow fever vaccine ..................................................................................... 12
Whole-cell pertussis vaccine ......................................................................... 15
Measles-mumps—rubella (MMR) vaccine ...................................................... l8
HEPATITIS B VIRUS (HBV) VACCINE ............................................................... 24
Background ................................................................................................... 24
Hepatitis B Vaccine - Safety Concerns ........................................................ 26
Demyelinating disease ....................................................................... 26
Thimerosal preservative .................................................................... 27
The Michigan Study ...................................................................................... 39
Introduction ....................................................................................... 39
Methods ............................................................................................. 42
Results ............................................................................................... 45
Discussion ......................................................................................... 50
CONCLUSION ......................................................................................................... 57
REFERENCES .......................................................................................................... 60

iv

LIST OF TABLES

TABLE 1. Average annual birth rate (1998-1999) of participating hospitals, by
hospital. ..................................................................................................................... 46

TABLE 2. Results of maternal HBSAg status upon medical record review, by
recommendation period. ............................................................................................ 47

TABLE 3. Demographic and prenatal care characteristics of 1,201 Michigan
women by HBSAg status, 1999-2000 ........................................................................ 48

TABLE 4. Proportion of Michigan infants born to women of unknown HBSAg
status who were vaccinated before discharge or within 12 hours, by time period... 49

INTRODUCTION

With the exception of safe water, no other health intervention has impacted the
improvement of the human condition as much as immunization.l During the past century,
the average life expectancy of US. citizens has increased by 30 years.2 Thanks to
vaccines, many once-prevalent devastating diseases are now rarely seen in the United
States.

In 1796, Edward Jenner inoculated James Phipps with cowpox and called the
procedure vaccination.3 Within a few years of Jenner’s first scientific demonstration of
the use of vaccination against smallpox, people throughout the world embraced the
practice. Generations of people have died or become disfigured from smallpox since the
disease first appeared in northeastern Africa about 10,000 years before the appearance of
Christ.4 This disease that historically brought down empires today has been eradicated
worldwide. The last indigenous case of smallpox occurred in 1977 (Somalia); global
eradication was declared two years later by the World Health Organization (WHO).5

Protection from vaccine-preventable diseases, such as smallpox, is readily
obtained through immunization. While disease management is necessary for the
maintenance of a healthy population, the optimal goal is to prevent a disease from ever
occurring. In order to understand the basis for recommendations of vaccine usage, it’s
important to understand the basics of vaccines themselves.

There are two primary mechanisms for acquiring immunity to infectious diseases
— active and passive. Active immunity is protection produced by a person’s own immune
system.6 It relies on the ability of the host to generate an immune response following

exposure to foreign antigens. Passive immunity is the other type of protection, and it is

conferred to a host through products produced by an animal or human, usually via
injection (e.g., hepatitis B immunoglobulin).6 Individuals can acquire immunity against
vaccine-preventable diseases through immunization, where the goal is to elicit an
immune response that mimics natural infection (i.e., active immunity). There are two
basic types Of vaccines: live attenuated and inactivated.

Vaccines are referred to as being attenuated if the bacteria or Viruses they contain
have been rendered nonpathogcnic (e.g., MMR vaccine and yellow fever vaccine).3
These live attenuated organisms must then replicate within the host to induce a protective
immune response. Inactivated vaccines refer to either viruses or bacteria that are killed
or components of the microorganism that are extracted and purified (e.g., polio vaccine
and pertussis vaccine).3 Receipt of inactivated vaccines cannot cause the disease they’re
meant to provide protection against. Several doses are usually required, though, to boost
the specific antibody level in the host. Microorganisms can also be genetically altered to
produce either live attenuated or inactivated vaccines, a process which can result in
recombinant vaccines (e.g., hepatitis B vaccine).6

Immunization programs can be Viewed broadly as serving two purposes. First,
vaccines help to prevent individuals from contracting the diseases that they’ve been
inoculated against. Second, vaccination of many individuals within a community helps to
establish herd immunity. Herd immunity is a level of immunity in a population that is
sufficient to prevent epidemics of a given communicable disease and prevent acquisition
of disease by those who can not be immunized (e. g., too young, immunocompromised,
concurrent disease states, etc.). Herd immunity is closely related to the basic

reproductive rate, which Giesecke defines as:

. .the average number of persons directly infected by an infectious case during his entire
infectious period, when he enters a totally susceptible population.”7

The higher the basic reproductive rate is for an infectious disease, the greater the
proportion of a community will have to be immunized to achieve herd immunity and
prevent epidemicity. Exemptors (i.e., individuals who refuse vaccination or persons
unable to access or receive a vaccine) rely on herd immunity to protect them from
vaccine-preventable diseases. The individual freedom of whether or not to vaccinate
Should be balanced with public health responsibilities, because the decision has both
personal as well as public consequences.8

Individuals who refuse vaccination may do so for several reasons, including but
not limited to: religious or philosophical grounds, medical reasons (e.g., allergies and/or
compromising underlying health conditions), or safety concerns. Forty—eight States offer
nonmedical exemptions (religious, philosophical) to state immunization laws, in spite of
the fact that legislative mandates are key to ensuring that children receive recommended
vaccinations in the US.9 It has been suggested that in many states, it’s easier for parents
to obtain an exemption than it is to get their children immunized.9 Despite the remarkable
success of the US. National Immunization Program, individual safety concerns about
vaccines have become a leading topic of public debate.

How does one define “safe”? If safe is defined as, “free from damage, danger, or

”'0 then vaccines are not safe. All vaccines, just like all preventive treatments used

injury
in medicine, have possible side effects.

We as American citizens are almost prisoners of our own success. As the

incidence of vaccine-preventable diseases has continued to decline, the questions

regarding vaccine safety issues have increased. Public concerns have shifted from the
fear of getting disease to health risks from receiving vaccines. Because vaccines are
typically given to healthy persons, a higher standard of safety is generally expected than
is true for, say - prescription drugs to treat chronic conditions. When one considers the
fact that many vaccines in this country are part of a recommended childhood
immunization schedule, public tolerance of adverse reactions related to these products is
even lower, because now one is considering a degree of risk to healthy youngsters.

Unusual health effects thought to be associated with vaccines may be true adverse
reactions or associated temporally with vaccination purely by coincidence. To assure that
true adverse reactions are detected as quickly as possible, the Vaccine Adverse Event
Reporting System (VAERS) collects reports of adverse immunization events in the
United States. VAERS was created in 1990 to unify the national effort for the collection
of vaccine-associated adverse events.“ This passive, open surveillance system places no
restrictions on who may submit reports, or when they can submit them. While jointly
administered by the Centers for Disease Control and Prevention (CDC) and the Food and
Drug Administration (FDA), a private contractor is responsible for data collection and
standardization of VAERS reports.ll Approximately one-fifth of the annual reports
received by VAERS are deemed serious enough (e.g., death, hospitalization, disability,
etc.) to be followed-up by a health professional. The number of annual reports to this
system now totals approximately 10,000 per year,ll which exceeds the current reported
incidence of most vaccine-preventable childhood diseases combined.12

Case reports of adverse events are not definitive in assessing vaccine causality,

. . . . . . 3
because of methodologrc weaknesses such as underreportrng and mdrvrdual brasm'

Recognizing the need to improve the study of vaccine safety, the CDC began work on the
Vaccine Safety Datalink (VSD) project — a large linked database of computerized
vaccination and medical records.12 The problems of underreportin g or recall bias are
reduced by the VSD project, because the databases are generated routinely and patient
records number in the millions. Data collection has now been expanded to include those
from infancy through adulthood, bringing the total to about 2% of the US. population.M

In June 2001, the CDC announced the setting up of the Clinical Immunization
Safety Assessment (CISA) Centers Network.‘5 The CISA design incorporates desirable
attributes of both the VAERS and VSD projects, in an effort to improve the
understanding of vaccine safety issues. CISA Centers will combine the detail Of
individual, patient level adverse events seen by clinicians with the Standardization of
routine reporting. Beyond enhancement of the understanding of known adverse vaccine
reactions, the CISA project will evaluate newly hypothesized relationships originating
from both VAERS data and the VSD project.

There is also a federal program designed to compensate families for the adverse
effects of vaccines. In 1987, Congress created the National Vaccine Injury
Compensation Program (NVIP) fund.l6 Because some companies at the time were
threatening to quit the vaccine business, the federal government set up the program as a
way to protect manufacturers from lawsuits brought about by injured vaccine recipients.
Although the fund has paid out $1.3 billion to 1,700 families since 1988, the program has
been labeled by some as being more combative than the court cases it was designed to
avoid.'7 This program, however, has been credited with substantially reducing the annual

number of lawsuits directed toward vaccine manufacturers.

Overview of the remainder of this thesis

Surveillance systems and compensation funds are some of the initiatives the US.
immunization program uses to deal with the detection and provision of remedies for
vaccine safety issues. The next section of this document will describe several of the most
widely publicized vaccine safety concerns affecting the United States and the entire
globe. This discussion will lead into a review of the history and safety profile of hepatitis
B vaccine, highlighted by a detailed discussion of the scientific plausibility of the alleged
association between the vaccine’s preservative (thimerosal) and adverse neurologic
events.

Temporary changes in the hepatitis B vaccination schedule due to concerns about
thimerosal used as a preservative in that vaccine led to the Michigan Thimerosal—Impact
Study — an observational study designed to assess the impact of recommendation changes
on the proportion of Michigan newborns receiving the hepatitis B vaccine before hospital
discharge. The results of this study and other examples to date will lay the groundwork
for the conclusion of this manuscript, including recommendations to consider when

making changes to proven immunization schedules.

VACCINE SAFETY CONCERNS

AS was already mentioned, vaccines are not totally free from possible negative
effects. All vaccines have the potential to cause adverse events. A vaccine adverse event
refers to any harmful reaction, large (anaphylaxis) or small (sore limb), that occurs
following receipt of the vaccine. Vaccine adverse reactions are broadly categorized into
three general groups — local, systemic, and allergic.'8 Local reactions are typically the
least severe yet most frequent, and they usually manifest as pain, swelling, and/or redness
at the site of injection. Systemic reactions are more nonspecific and generally include
symptoms such as fever, malaise, and headache. Probably the most severe of the general
adverse events is allergic reaction, which rarely results in a dangerous immune system
response called anaphylaxis. The most frequently occurring signs and symptoms of
anaphylaxis include: urticaria and angioedema; dyspnea and wheezing; and dizziness
and hypotension. '9

Aside from anaphylaxis, some parents have voiced concerns that the increase in
vaccine number and doses given to infants might be responsible for various allergic
diseases. Anderson et al. performed an ecologic analysis using national and local
immunization rates for several routinely administered vaccines to examine their
relationship with symptoms of atopic disease in children.20 Outcomes were measured as
the prevalence of allergic rhinoconjunctivitis, atopic eczema, and wheezing in the past .12
months. Immunization rates for tuberculosis, diphtheria, tetanus, and pertussis (DTP),
and measles vaccine were sought that corresponded to the approximate year of birth for

study groups.

No significant associations were found between symptoms of atopic disease and
vaccination rates using national figures. Negative significant associations were found at
the local level between DTP and wheezing, allergic rhinoconjunctivitis, and atopic
eczema and between measles vaccine and rhinoconjunctivitis and atopic eczema for the
13- to 14-year Old group. Controlling for per capita gross national product (GNP) only
slightly decreased Significant associations.

Mass immunization of children against a variety of communicable diseases is
only one of several factors that affect the immune system early in life, although this
practice has increased along with various allergic diseases Since the 19508.20 This
international ecologic study by Anderson et a1. does not support the idea that
immunization is responsible for increases in atopic disease among children. Although
ecologic studies can’t exclude associations at the individual level, they may help correct
for random variation encountered by values taken at the individual level. The Institute of
Medicine’s (lOM) Immunization Safety Review Committee recently examined five
studies looking at multiple vaccinations and their potential to induce hypersensitivity
reactions, and the committee concluded there was not enough evidence to accept or reject
a causal relationship.2|

Vaccines are given to children at multiple points during their development, and
coincidental adverse events of many types can happen throughout that time. There are
several vaccines that have received considerable publicity in recent years for their alleged
associations with adverse events. Some of these safety concerns involve risks that,
although rare, appear to be factual in nature, including: fears about live polio vaccine

producing rare paralysis and yellow fever vaccine associated with recent deaths. Other

vaccines have gained notoriety because of their purported risks and the associated
damage done to immunization coverage, including: measles-mumps-rubella (MMR)
vaccine associated with autism-spectrum disorder and whole-cell pertussis vaccine
leading to adverse neurologic events.

Polio vaccine

The first outbreaks of poliomyelitis were reported in Europe during the early
18003.22 Throughout the nineteenth century, epidemics of polio were described in
temperate countries of the Northern Hemisphere each summer and autumn.22 Probably
the most well—known American citizen to become paralyzed by polio was President
Franklin Delano Roosevelt, who used a wheelchair for the remainder of his life after
contracting the disease.

Polio is caused by a virus and is highly contagious. Poliovirus is a member of the
genus Enterovirus, and there are three serotypes (P1, P2, and P3).23 The mode of
transmission for poliovirus is primarily through the fecal-oral route. After entering
through the mouth and multiplying at the site of implantation, the virus enters the blood
stream via lymphoid tissue and can go on to infect cells of the central nervous system
(CNS).22

The majority of all polio infections (greater than 90%) are subclinical, although
these infected persons still shed virus in their stool.23 Less than 1% of all poliovirus
infections result in flaccid paralysis, where the disease selectively destroys the motor
neurons of the Spinal cord and brain. Asymmetrical muscle weakness and diminished

deep tendon reflexes are key features of the illness.23

Transmission of wild poliovirus stopped in the US. in 1979, and vaccination has
also led to elimination of polio throughout the Western Hemisphere since 1991 (Peru).23
In the early vaccine era, the incidence of wild virus cases dramatically decreased
following the introduction of Salk inactivated polio vaccine (IPV) in 1955.6 This decline
continued following licensure of Sabin oral live attenuated vaccine (OPV) in 1961-1962.6

While the Salk IPV provided excellent individual protection from paralytic
poliomyelitis, persons who got IPV received less intestinal immunity than OPV
recipients. This made IPV recipients more susceptible to local wild poliovirus replication
in the gut, with reduced impact on the continued excretion of wild strains of polio in stool
resulting in continued exposure to those in the community22 (the eIPV — enhanced IPV —
vaccine now used worldwide and in the USA is a more potent inducer of intestinal
immunity than the original Salk vaccine). The World Health Organization (WHO)
recommends the use of OPV alone in developing countries where wild virus is still
endemic.23

One adverse event that follows receipt of oral poliovirus vaccine, though rarely, is
known as vaccine—associated paralytic poliomyelitis (VAPP).22 VAPP induced by OPV
occurred in about 1 out of every 750,000 first doses of the vaccine,24 and the paralysis
that results is identical to that caused by wild poliovirus. The occurrence of VAPP
following receipt of oral vaccine is increased in infants if followed by multiple injections
of antibiotics (commonly used in some countries to treat infants with febrile illness).25
The IPV form, which cannot cause paralysis (because it doesn’t contain live Virus), is the

preferred vaccine for routine immunization of children in the United States where wild

virus has been eliminated.26

10

Because of the low cost, ease of administration, and capability to provide
immunity to susceptible contacts through secondary spread, WHO has continued to
support an OPV-only polio eradication policy in many parts of the world.22 This would
indicate that Organization officials have opted for the m of possibly eradicating
wild-virus polio cases globally, despite the rare (but known) Lsk associated with oral
poliovirus vaccine.

This risk has been evident recently in two regions of the Western Hemisphere that
are quite distant from each other. From July 2000-September 2001, an outbreak of
poliovirus infection due to a mutant strain of vaccine-derived oral poliovirus (OPV-I)
resulted in 21 cases of poliomyelitis on the island of Hispaniola (Dominican Republic
and Haiti).27 The outbreak occurred in areas of low vaccine coverage (Haiti, 20-32% and
Dominican Republic, 73-82%), and only one patient had received at least three doses of
OPV. After these cases associated with circulating vaccine-derived poliovirus (cVDPV)
were discovered, all polioviruses under investigation worldwide were required to undergo
additional testing to distinguish wild virus from cVDPV. These new prospective testing
requirements also detected three cases of VAPP associated with cVDPV in the
Philippines during March-July 2001.28 No other paralytic cases attributable to cVDPV
have been reported since then.

It is believed that the viruses may have been replicating for a while within certain
individuals with deficient immune systems living in those communities, with reversion of
the circulating strains to augment neurovirulence and improve transrrrrissibility.28'30 The

longer an attenuated virus can replicate in the gut, the more time it has to potentially

11

revert away from the parent strain (OPV) and towards the wild-type virus.29 An
inadequately vaccinated or immunodeficient host may then pick up the mutated strain.

Poor vaccination rates are one of the most important causes of cVDPV. These
episodes of VAPP due to reverted strains are especially noteworthy since both areas
(Hispaniola and the Philippines) are believed to have been free of wild-type poliovirus
since 1991. The occurrence of VAPP in association with cVDPV in these regions of the
world raises the question: what is the appropriate immunization schedule (OPV, IPV, or
combined) in developing countries that have few cases of wild-type polio? While a
switch to IPV would eliminate the risk of vaccine-associated paralysis in individuals
receiving the vaccine, it might not be economically feasible in developing countries. In
addition, a switch to IPV would remove the benefit that comes from vaccine virus being
introduced into the environment following administration of OPV, which helps to
“vaccinate” persons who may not have directly received the vaccine themselves.
Yellow fever vaccine

Early in the nineteenth century, the means of transmission of yellow fever were
unclear. Some believed that the disease could be transmitted from one person to another,
until Stubbins Firth exposed himself both orally and parentally to fluids of patients dying
of yellow fever.3 "32 After Firth didn’t fall ill with yellow fever, he concluded the disease
was not communicable. When the US. occupied Cuba during the Spanish-American
War, a study commission (led by Walter Reed) established that Aedes aegypti mosquitoes
transmitted yellow fever.”33 Reed and his commission also determined that yellow fever

was caused by a virus, the first disease to be labeled as such.

12

Yellow fever is an infectious viral disease of the genus F lavivirus.3 4 After an
incubation period of three to six days, typical attacks are characterized by fever, malaise,
nausea, and vomiting. Some cases experience a brief remission period followed by
hemorrhagic symptoms and liver failure, with up to 40% of individuals succumbing to
the disease.34 Patients who are fortunate enough to recover from yellow fever experience
lasting immunity.

In 1937 Theiler and Smith developed an attenuated 17D version of the yellow
fever Virus, and all the currently produced vaccines are based on this.3 "35 The current
vaccine substrains are designated 17DD and l7D-204. The disease is especially
problematic in areas densely populated with vector mosquitoes and many susceptible
humans. A single injection of vaccine is effective in almost 99% of recipients,34 and its
administration is recommended for inhabitants of, and travelers to, many regions of South
America and Africa.

Until recently, the most serious side effects thought to be associated with yellow
fever vaccine (for recipients older than 9 months) were severe allergic reactions.34 But the
Lancet of July 14, 2001 collated reports of 7 cases of multiple organ system failure
(MOSF) (including 6 deaths) following yellow fever vaccination in Brazil, the USA,
and Australia.”38 The clinical manifestations differed among the cases, and there were
no obvious correlations between the three reports. Vaccines produced by different
manufacturers had been used on each continent.

In 1998, Brazil included yellow fever vaccine in their national program of
childhood immunization in response to one of the largest epizootics in history.36

Beginning in 1998, the country registered 192 human cases of yellow fever, almost half

13

of which (46%) died.”39 Brazil uses the 17DD substrain of yellow fever vaccine. The
two Brazilian fatalities reported in the Lancet (ages 5 and 22) that occurred following
yellow fever immunization represent two deaths in over 85 million vaccinations in that
country between 1990 and 2000.36 These two patients experienced significant organ
damage (e.g., liver necrosis and hemorrhaging) that was confirmed as vaccine derived,
based on serum and tissue samples. This demonstration that 17DD vaccine has residual
Viscerotropism resembling wild-type infection represents a complication not previously
reported.40

The four U.S. cases (three deaths and one severe illness) involved patients of
advanced age, all of whom were older than 63 years.37 Yellow fever vaccine is
recommended for individuals traveling from the United States to areas where the disease
is endemic. The United States uses the 17D-204 substrain of yellow fever vaccine.
Three of the reports of severe illness were received at the CDC in 1998, while the
additional case had been reported to VAERS in 1996.37 These cases showed less organ
damage than the Brazilian fatalities and more involvement of the CNS. The timing
between vaccination and illness, clinical features, and isolation of vaccine virus from
bodily fluids led the authors to suggest the possibility of a causal association between
yellow fever vaccine and illness in these patients.37

The one Australian case of MOSF was 56 years old.38 Australia also uses the
l7D-204 substrain of yellow fever vaccine, but from a different manufacturer than the
United States uses. The man suffered from extensive hemorrhages, and his liver was
found to have large areas of necrosis.38 His death was attributed to vaccine-derived

yellow fever by virus isolation from multiple tissues.

14

It is worth mentioning that all seven adverse events occurred in a small number of
vaccine recipients among many others who received the same vaccine. A commentary
published in the same issue Of the Lancet suggested that now is the time for investigating
the pathophysiology of wild—type yellow fever virus as compared to attenuated Viruses, in
addition to host susceptibility factors.4| Despite the recent adverse events, an Advisory
Committee on Immunization Practices (ACIP) working group suggested that persons
traveling to areas where yellow fever is known to occur should still be vaccinated.42
However, providers are cautioned to carefully screen potential recipients based upon
where they’re planning to travel, to ensure that their destination warrants them getting the
vaccine. Two of the persons who succumbed to vaccine-associated MOSF were planning
to travel to regions where yellow fever transmission had never even been reported.42
Whole-cell pertussis vaccine

The DTP vaccine is a combined vaccine used to prevent three diseases —
diphtheria, tetanus, and pertussis. Pertussis (widely known as whooping cough) is an
acute infectious disease caused by the bacterium Bordetella pertussis.43 In non-
immunized populations, pertussis is among the most lethal diseases of infants and young
children, with an estimated 300,000 deaths per year.44

B. pertussis produces a toxin that is responsible for the clinical, respiratory
features of the disease. The incubation period of pertussis is usually 7-20 days.43
Pertussis-associated neurologic complications are more common among infants, and
encephalopathy and seizures have been reported from 0.2% and 1.4%, respectively, of all

cases.“ Pertussis is highly communicable with a secondary attack rate of 80% among

15

susceptible household members Via contact with respiratory secretions or aerosol
droplets.”47

Whole-cell pertussis vaccine was developed in the mid—19308 and combined as
DTP in the mid-19403.44 Whole-cell DTP vaccines had a fairly high rate of mild and
severe side effects, with febrile seizures reported in l of every 1,750 doses and
encephalopathy occurring very rarely.44 Concerns about the safety of DTP led to
development of an “acellular” pertussis vaccine (DTaP). The acellular pertussis vaccine
has been the one in use in the United States since 1996.44

Anti-vaccine movements that targeted whole-cell pertussis vaccine provide
another example of harm done to immunization coverage when imprudent decisions are
made. Adverse neurologic events were first associated with receipt of whole-cell
pertussis vaccine in 1933.47‘48 Since that time, several review committees and large-scale
studies have examined the issue and arrived at varying conclusions.

Gale et a1. conducted a population-based case-control study (as a part of a
feasibility assessment) to evaluate adverse whole-cell pertussis vaccine effects in Oregon
and Washington.49 Prospective surveillance was established from August 1987 through
July 1988 using a blinded panel of clinicians to identify all cases of serious acute
neurological illness (acute encephalopathy, infantile spasms, and complex febrile
seizures). The authors observed no increased risk of onset of serious neurological disease
in the seven days after DTP exposure. Even with a prospective surveillance system of

218,000 children, the authors admitted that the statistical power would only be able to

detect significant odds ratios of at least 2.5 for all incident cases.49

16

Following the study by Gale et al., a 1991 Institute of Medicine report concluded
that the available evidence suggested a causal relationship between the receipt of DTP
and acute encephalopathy.50 The Vaccine Safety Datalink Working Group subsequently
used VSD records to examine the relation between DTP vaccine and seizures, as well as
the outcomes among children with seizures.5| Using criteria similar to Gale et al. for
classifying episodes of seizure, the VSD Working Group found that receipt of DTP
vaccine was associated with an increased risk of febrile seizures only on the day of
vaccination (RR, 5.70; 95% CI, 1.98 to 16.42).

Despite the published associations between DTP vaccine and neurologic
complications, public health authorities around the world should carefully weigh the risks
and benefits of using DTP vaccine. This controversy may be less of an issue in the US,
where acellular (DTaP) vaccine has been recommended for use Since 1996.44 However,
nearly all developing countries still use DTP vaccine because of its lower cost.43

An estimated 45 million cases of pertussis occur annually, and case-fatality rates
in developing countries can reach 15%.52 If these figures are coupled with estimates of
adverse events attributable to DTP vaccine, one begins to understand why the decision to
completely discontinue pertussis vaccination is imprudent. For example, the 1991 IOM
report only translated into an excess number of acute encephalopathy cases in the range
of 0-10.5 cases per million doses of DTP administered.50 The risk of febrile seizures on
the day of vaccination reported by the VSD study only equated to approximately 6-9
additional febrile seizures for every 100,000 recipients.5 l Furthermore, the VSD Working

Group conducted a follow-up analysis and found that these children were at no greater

l7

risk of subsequent seizures, epilepsy or learning, behavioral, or psychiatric disorders than
other children with febrile seizures (in the absence of vaccination).51

Another way of examining the importance of DTP immunization is to study the
experience of nations where pertussis vaccine coverage declined. Gangarosa et al.
compared the pertussis experience of two groups of countries: Group I (Hungary, the
former East Germany, Poland, and the US), where high coverage with DTP vaccine was
maintained; and Group 11 (Sweden, Japan, the UK, the Russian Federation, Ireland,
Italy, the former West Germany, and Australia), where anti—DTP movements disrupted
vaccination.S 3 This study showed overall trends indicating that pertussis incidence was
10-100 times lower in countries where high vaccine coverage was maintained versus
countries belonging to Group II. In nations where vaccine coverage was breached, the
reasoning behind each story is varied, including: a loss of confidence in vaccine efficacy
(Sweden);54 fear due to fatalities following vaccination (Japan); attitudes, knowledge, and
practices of providers (Italy and the former West Germany); and concern about potential
nonfatal adverse events (the Russian Federation, the United Kingdom, Ireland, and
Australia). It is worth mentioning that the commotion raised in the UK. Spread to Ireland
and Australia because of their common ties,53 providing further evidence that fear about
vaccines can be contagious.
Measles-mumps-rubella (MMR) vaccine

Measles, mumps, and rubella are all viral infectious diseases.”57 Measles and
mumps Viruses are paramyxoviruses, while the rubella virus belongs to the family
Togaviridae. There is only one antigenic type of measles Virus, and its incubation period

averages 10—12 days.55 Diarrhea is the most commonly reported complication of measles

18

(8%), while the most serious nonfatal sequelae is encephalitis. Acute encephalitis occurs
in about 0.1% of cases, and subacute sclerosing panencephalitis (SSPE) is reported in
five to ten cases per million measles infections.58

The measles-mumpS-rubella (MMR) vaccine is a combined vaccine used to
prevent all three diseases. Each vaccine has been available as a single antigen
preparation. Fever is the most commonly reported adverse event following MMR
immunization (5%-15%).” While encephalopathy is only noted in less than one per
million doses (estimated to be the same as the background rate of encephalitis due to all
causes),” MMR vaccine has been publicly blamed for another CNS-related disorder —
autism spectrum disorder (ASD).

In February 1998, the Lancet published an early report by Andrew Wakefield et
al. proposing that MMR vaccination might cause autism, possibly by a mechanism
involving damage to the gut.” The report was based on a case-series of 12 children
referred to a London pediatric gastroenterology department, with a history of intestinal
problems and presentation of developmental disorders. In eight cases, the child’s
behavioral problems were pinned upon receipt of MMR vaccine. The average interval
from immunization to first notice of behavioral symptoms was six days.”

Wakefield et al. described a hypothesis known as the “opioid excess” theory of
autism, which postulates that an exogenous influence impairs the cerebral function of
patients.” The theory proposes that disruption of the gut wall (presumably caused by
persistent measles vaccine-strain virus infection) increases intestinal permeability. Gut-
derived peptides are increasingly absorbed, initiating a process that leads to disruption of

normal brain development.”

19

Autism is a complex and severe developmental disorder marked by cognitive and
neurobehavioral deficits."0 The term “autism-Spectrum disorder” (ASD) refers to a
continuum of impairments in which patients vary in the severity of their symptoms. The
reported time course Of developmental problems determines whether a patient’s autism is
described as early onset or regressive. The majority of autism cases seem to be early
onset, presumably due to some sort of brain injury.” Regressive autism occurs less often,
and follows what is thought to be normal, early development. The two types of
presentation (early onset and regressive) are typically not diagnosed until the second year
of life.62 ASD appears to exhibit a strong genetic component, also.”'”

The provocative study by Wakefield et al. is marred with shortcomings. The
clinical investigation of patients involved the collection of histories from a variety of
sources, including parents and health visitors.” Eleven of the twelve cases were
investigated by the same clinician (J.A. Walker-Smith), who is also cited in the
references section of Wakefield’s paper as having previously published work on a related
hypothesis.”64 The investigators should have been blinded for aspects such as clinical
assessments. In addition, four of the children had their behavioral assessments done
elsewhere, and these results were used to determine their diagnosis for this study.”

There was not any patient selection, other than the case—series of 12 children
referred to the department, and the underlying population was not clear. A case-series
approach is sometimes useful for generating new hypotheses, but it shouldn’t be used for
testing hypotheses and stating conclusions. The report announced that eight out of twelve
patients had their autistic behavioral problems temporally linked with MMR

vaccination.” Parents or the child’s physician did this linking, but we’re told nothing

20

about the querying procedure or how often a troubled parent rather than a physician made
the association.

The Wakefield article set off a flurry of correspondence in the next month’s
Lancet.”7| All of the published letters echoed the same sentiment: a genuine fear that the
published study results would damage vaccine uptake and lead to a resurgence of
measles, mumps, and rubella in the United Kingdom (UK). Three of the seven

05. 7, 0
letters 6 7

cited the negative pertussis-vaccine movement in Europe as foreshadowing
of what can happen when an article with such potential for harm is published.
Wakefield’s research was widely reported in the media and generated much
public concern. The British Medical Journal subsequently reported that UK. parents
believed the MMR vaccine to be even more dangerous than natural measles virus.72 In
the same Lancet issue as the aforementioned letters, Wakefield73 and some of the other
study’s authors74 printed their replies. The two letters”74 had very different tones. The
accompanying authors claimed that they were emphatic about not calling for a change to
existing immunization policy,74 although this did little to soothe frightened parents.
Wakefield, in regards to the criticism his work received from the public health
community of the United Kingdom, stated:
“. . .the clinical researcher’s obligation is to test hypotheses of disease pathogenesis on the
basis of the story as it is presented to him. . .listen to the patient or the patient’s parent,
and they will tell you the answer.”73
These five pages of heated correspondence ended with an editor’s reply, which asked the

question:

6 ' 3 ' ‘ 997-
‘. . .are Wakefield and colleagues observations reproducrble? 3

21

From a deductionist’s viewpoint, the more appropriate question might be: is Wakefield
and colleagues’ hypothesis refutable?
The latter question was quickly addressed through a succession of studies and

-79 . . .
76 A group of researchers in leand traced 31 children who

review committees.
developed gastrointestinal symptoms after MMR vaccination (apart from within the first
hour), out of about three million vaccinees.76 Diarrhea was the most commonly reported
symptom, and no children developed ASD. Working backwards from patients known to
have ASD in North East Thames, Taylor and colleagues showed that, although the known
number of cases of ASD had been increasing since the late 19703, there was not a notable
increase after the introduction of MMR vaccine a decade later.77 They also Showed that at
age 2 years, MMR vaccination coverage among the identified ASD cases was almost
identical to other children in the same birth cohort.

Two separate reviews of the issue have been conducted in the US, one by the
Institute of Medicine78 and another at an American Academy of Pediatrics conference.”
Both reviews resulted in the same conclusion: the available evidence does not support a
causal relationship between MMR vaccine and ASD at the population level. However,
that MMR vaccine could contribute to ASD in a small percentage of children cannot
currently be ruled out.

A high-profile campaign was announced in early 2001 in the UK. to publicize the
lack of causal evidence regarding the issue and reassure parents that MMR vaccine was
safe.80 However, Andrew Wakefield had also opined that the Single antigen preparation

of each vaccine should be used instead of the combined vaccine81 and gave already

concerned parents something else to consider. In addition to press conferences and

22

interviews, Wakefield has made two trips across the Atlantic to appear before the US.
Senate.”83 Wakefield and John O’Leary have presented fragmented results in
Washington that raise more questions than they answer. It was understandable how these
clinicians got a spot on the agenda - Congressman Dan Burton is the chair of the House
Committee on Government Reform. and his grandson was diagnosed with regressive
autism.

Despite their efforts, the United Kingdom’s Department of Health may be losing
the battle. It was already reported that, prior to Wakefield’s fame, the coverage of MMR
vaccine was subtly decreasing;8’4 but this subtleness turned overt following a 12% drop in
the number of children receiving the vaccine.85 There have also been reports Of mumps
outbreaks in Northern Ireland and England.”86 In the United Kingdom, trust in
government announcements is low, and it’s not improving with a govemment-imposed
gag order that prohibits officials (including Prime Minister Tony Blair) from disclosing
whether their children have received the MMR vaccine.87 In February 2002, it was

reported that uptake of the MMR vaccine in London is only 73 percent.88

HEPATITIS B VIRUS (HBV) VACCINE
Background

Transmissible illnesses that caused jaundice have been reported since antiquity.
The earliest recognition that blood or blood products could transmit some forms of
jaundice was in 1885, when cases of what was likely hepatitis B were identified
following the use of glycerinated human lymph as part of smallpox vaccination.”90 The
hepatitis B virus (HBV) contains multiple antigens, including hepatitis B surface antigen
(HBSAg), hepatitis B core antigen (HBcAg), and hepatitis B 8 antigen (HBeAg).89'9'
Identification of HBSAg in human serum indicates that the person is potentially
infectious. In addition, HBV is relatively resistant to decontamination - it can retain
infectivity after remaining on inanimate surfaces for at least one month at room
temperatureggm The incubation period of HBV infection usually ranges from 45-180
days."1

Part of the difficulty in controlling the spread of HBV is its tendency to cause
asymptomatic infections, especially among infants and children.89‘9' Even though many
individuals will be unaware that they’re infected, the most serious complications won’t
usually occur unless a person develops chronic infection."2 The risk of chronic infection
is highest among infants and young children, and decreases with increasing age.89'93
Chronic infection is responsible for most HBV-related morbidity and mortality - an
estimated 5,000 liver disease deaths occur each year in the US. because of chronic HBV
infection.89

A plasma-derived vaccine was licensed for protection against hepatitis B virus in

the US. in the early 19803.9l Due in part to fears that the vaccine could become

24

contaminated with other viruses that might be found in human blood, a recombinant
vaccine was licensed in the mid-19803, making it the first of its kind in the United
States.”2 The efficacy of hepatitis B vaccine ranges from 80% to 100% in those who
follow the recommended schedule of doses."2 Because HBV can cause primary
hepatocellular carcinoma,89 this vaccine is the first vaccine that prevents cancer, and
deaths among children due to hepatocellular cancer have significantly decreased in
Taiwan following universal vaccination.94

The initial strategy for use of hepatitis B vaccine in the United States was based
on vaccinating persons at highest risk for disease (e.g., intravenous drug abusers,
promiscuous hetero-lhomosexuals, and health care workers).95‘% However, surveillance
data indicated that 30-40% of hepatitis B cases had no identifiable risk factors, and
vaccination coverage among high risk groups was very low.%

Although a relatively small proportion of the estimated 200,000 to 300,000 HBV
infections per year occurred in children during the 1980’s, infections prior to age 5
accounted for 20-30% of all chronic cases, thereby perpetuating the endemicity of HBV
(albeit at a low level).97 In 1991, the Advisory Committee on Immunization Practices
(ACIP) developed a comprehensive strategy for controlling the virus, which was based
on universal screening for HBsAg of all pregnant women and universal vaccination of all
infants.”8 The 1991 recommendations called for all infants to begin the vaccine series
within the first 6 months of life, and noted that beginning the vaccination series shortly
after birth, before discharge from the hospital, was safe and effective.98 This universal
immunization strategy emphasized the importance of US. birthing hospitals in

preventing vertical transmission of the virus from mothers to newborns. Theoretically, a

25

sustained program of universal infant vaccination would complement programs aimed at
preventing horizontal transmission among adolescents and hi gh-risk adults. As these
vaccinated infants grew to adulthood, the pool of susceptible hosts would be low enough
to interrupt transmission of the disease. However, widespread use of HBV vaccine has
been criticized for its alleged association with demyelinating diseases, as well as because
of the vaccine’s preservative — thimerosal.
Hepatitis B Vaccine - Safety Concerns
Demyelinating disease

Around the same time that the CDC released its universal immunization strategy,
concerns were being voiced that hepatitis B vaccine might be associated with adverse
neurologic events.”100 Guillain-Barre’ syndrome was reported significantly more often
than expected after administration of plasma-derived hepatitis B vaccine in one study.”
However, an Institute of Medicine panel review, as well as other reviewers, found
insufficient evidence to support or reject a causal association.”’"102

In 1991, a Belgian group reported in a case-series that CNS demyelination
occurred after vaccination with recombinant hepatitis B vaccine.‘00 Their findings were
based on only two patients (one already known to have multiple sclerosis (M8)), and the
findings appeared as a commentary in the Lancet. A larger study (25 cases) reported an

103 although sufficient

association between hepatitis B vaccine and MS in France,
evidence to support a causal association was not established there, either.102 More
recently, a brief report from Greece suggested a causal link between hepatitis B vaccine

and leukoencephalitits.104 This conclusion was based upon a case report of two separate

episodes in the same patient. Only sketchy details were released about the patient’s

26

neurologic history - a questionnaire had been administered, and no preexisting conditions
could be demonstrated.

An Institute of Medicine committee currently reviewed the hypothesized
association between hepatitis B vaccine and demyelinating neurological disorders, such
as multiple sclerosis and Guillain-Barre’ syndrome.105 A review of the epidemiological
evidence led the committee to reject the notion of a causal relationship between hepatitis
B vaccine and multiple sclerosis; however, not enough evidence currently exists to
comment on the relationship between the vaccine and other demyelinating disorders.

Thimerosal preservative

 

Thimerosal has been used as an antibacterial preservative in many childhood
vaccines since the 19303, and in some hepatitis B vaccines until 2000.106 Preservatives
are often used when vaccines are supplied in multi-dose vials. Repeated entry of syringes
into such vials allows for the possible introduction of bacteria and fungi. Multi—dose
vials are preferred in certain settings because they generally are less expensive to
purchase and store. '06

In 1976 there was a formal Food and Drug Administration (FDA) assessment and
review of any potential harm from repeated exposure to thimerosal. At that time it was
concluded that dangerous quantities of mercury were unlikely to result from receipt of
biological products (such as thimerosal-containing vaccines) over the course of a

lifetime.10(”'07

During the next couple of decades, more vaccines containing thimerosal as
a preservative were gradually added to the US. childhood immunization schedule.

The 1997 FDA Modernization Act called for a review of heavy metal exposures

to humans from all potential sources, including foods, work place, environment, and

27

biologics. Thimerosal was also scrutinized given the increased number of vaccines
recommended for children, along with advances made in the understanding of the effects
of exposure to mercury.'06 This reassessment indicated that, depending on the vaccines
used, some infants might have been receiving total doses of mercury that exceeded
Environmental Protection Agency (EPA) guidelines - it is important to note that these
guidelines were set according to data from studies on the effects of methylmercury.
Ethylmercmjv is the actual metabolite of thimerosal that is found in the human body, not
methylmercury. In addition, only the EPA guidelines were found to be surpassed by
certain immunization schedules, whereas the Agency for Toxic Substances and Disease
Registry (ATSDR), World Health Organization (WHO), and FDA guidelines were not
exceeded.106

To assess the scientific plausibility regarding exposure to thimerosal and adverse
neurologic effects, the author searched the literature for reports of such adverse events in
humans and animals. This plausibility assessment has two components: (1) an
examination of the causal relationship between the exposure and adverse events, and (2) a
discussion of any pathogenic mechanisms that support the alleged associations.
Causal Relationship

Much of what is known about mercury toxicity comes from poisoning episodes
over the previous 50 years or so. The consumption of contaminated fish in Minamata,
Japan during the late 19503 and early 1960s provided evidence of methylmercury’s
toxicity.108 A strange nervous disorder began afflicting villagers in and around the area.

During the second half of the 19503, approximately 6 percent of the children born in that

28

area developed cerebral palsy.108 In utero exposure to methylmercury was blamed,
because these infants had not eaten contaminated fish.

During the early 19703, barley and wheat grain treated with methylmercury was
distributed to Iraqi farmersmg This mercury was presumably added to help stave off
fungal contamination. Reported symptoms were similar to those of the Japanese
epidemic. Blood mercury levels of infants born just prior to the epidemic were attributed
to poisoning via breast milk. In utero exposure, breast milk, or both were implicated as
the methods of poisoning for infants born during or after the epidemic.

Bakir et al. studied 15 mother-infant pairs in Iraq by carrying out serum analyses
with repeated clinical examination.l '0 Six of the affected infants showed irritability and
excessive crying, with four of them having severely affected mental power. Although this
Iraqi study had a very small sample size, the authors concluded that exposure to
methylmercury in the third trimester of pregnancy could lead to damaging effects for the
infant.110 However, there was much variability between when the infants were examined.

Studies from the Faroe Islands found subtle cognitive deficits that were associated
with methylmercury levels previously considered safe.l '1 Increased methylmercury
exposure from maternal consumption of pilot whale meat was indicated by mercury
concentrations in cord blood and maternal hair. Of the original cohort of 1,022
consecutive births, 917 of the children underwent sophisticated neuropsychometric
testing.l '1 The most pronounced neuropsychological dysfunctions were in the domains of
language, attention, and memory. These associations remained even after the exclusion of

children with maternal hair mercury concentrations at the highest end of the spectrum.

29

The same group of authors published another study on the Faroese cohort, but this
time including several other biomarkers of methylmercury exposure.l '2 This time they
analyzed hair from the children at ages 12 months and 7 years, and blood at 7 years was
collected also. The same neuropsychologic tests were chosen as before. Results were
presented as the change in test performance at age 7 years associated with a doubling of
the mercury concentration for the exposure biomarkers.

The cord-blood concentration generally showed greater regression coefficients
and lower P values than did the other biomarkers (significance range: <0.001-0.049).l '2
The 1-year hair concentrations of children were significantly associated with three of the
neuropsychologic test scores, whereas the two 7-year biomarkers did not present much of
an association. The cord-blood concentrations were the best predictors of decrements in
the domains of language, attention, and memory.1 '2 Since no exposure markers were
obtained between ages 1 and 7 years, the time at which postnatal damage occurred could
not be specified.

This North Atlantic population was again the subject of study in another
prospective cohort venture.l ‘3 A cohort of 182 births was assembled during a 12-month
period. The following samples were collected for exposure assessment of the infant:
maternal serum at week 34, cord blood, maternal hair, and transition milk. For the
neurologic examination of the newborn, a technique was used at approximately two
weeks of age designed to assess functional abilities and the stability of behavioral status
during the exam.‘13 The cord-blood mercury concentration showed a negative association
with neurologic function, as assigned by a neurologic optimality score (NOS). A lO-fold

increase in mercury was associated with an NOS decrease of 2 points. It is worth

30

mentioning that neonatal assessment tools have not always been highly predictive of later
functioning. It is also unclear whether this study’s 2-point difference in NOS suggests a
true difference in morbidity. The predictive value of measuring these outcomes at two
weeks of age is unknown to this author, as well.

Stajich conducted a repeated cross-sectional study among newborns in 1997-
1998.114 All of the newborns had been admitted to the Neonatal Intensive Care Unit
(NICU) at the Grady Health System in Georgia. Baseline mercury concentrations were
measured in the serum of these infants. It was found that vaccination against hepatitis B
significantly increased serum mercury levels in both term and preterm infants (P <
0.01).1 '4 Post-vaccination mercury levels were significantly higher among extremely low
birthweight (< 1000 gm) and very low birthweight (< 1500 gm) preterm babies compared
to term infants (P < 0.01). This was the first published study of its kind. There were 15
subjects in total (10 preterm vs. 5 term infants). All of the newborns were NICU
admitted, so it’s unclear if findings could be generalized to well infants. Most
disappointingly, they did not measure the neurologic function of these infants. Therefore
it is unknown what the clinical significance of their data represents. This shortcoming is
even more frustrating given that their exposure of interest was actually ethylmercury via
thimerosal, and not methylmercury like all other previous studies.

A prospective cohort study in Greenland attempted to examine the effects of
methylmercury in utero.1 '5 Pregnant women were invited to participate when they entered
birth clinics at the beginning of labor. Methylmercury concentrations were measured via
maternal serum and cord blood. The outcomes of interest were gestational length and

birthweight of the infant. Gestational length was not shown to be associated with blood

31

mercury concentrations.”5 Blood mercury of the infant was negatively associated more
strongly with birthweight than was maternal blood mercury. Any conclusions drawn from
these analyses are limited by the fact that one half of the eligible births were missed.

Another study that assessed similar outcomes was conducted in Tagum,
Philippines, where tons of methylmercury had been continually dumped into the river
system from mining ores.H6 This cross-sectional study was conducted as part of a
prospective cohort study evaluating the long-term effects of mercury exposure. Total
mercury concentrations were measured via maternal blood, breast milk, cord blood, and
meconium. The prevalence of mercury in meconium was investigated for its relationship
to neonatal head circumference. A negative relationship between levels of mercury in
cord blood and a smaller head circumference suggested an effect of total mercury on head
growth (P = 0.0469 for the association).116 Similar to the previous Greenland study,
certain assumptions must be made on the utility of these two studies. The importance of
birthweight and head circumference decrements related to increased mercury levels
remains to be assessed by future neurodevelopmental evaluation.

The fact that several studies have found significant associations between mercury
exposure and adverse effects does not resolve the causal debate, especially since the
specific exposures measured have not traditionally been the thimerosal metabolite,
ethylmercury. However, not all of the studies to date have found an association between
exposure to mercury and adverse neurologic effects. Davidson et al. conducted a
prospective cohort study with an inception cohort of mother-child pairs in the Republic of
Seychelles.l '7 The cohort consisted of 711 pairs of mothers and children. Prenatal

mercury exposure of the infant was assessed by measuring the concentration of total

32

mercury in maternal hair presumably representing levels during pregnancy. Postnatal
exposure was similarly determined by measuring total mercury from a segment of the
child’s hair at 5.5 years of age. At the same time, nurses blinded to exposure status
evaluated each child with a test battery consisting of six age—appropriate
neurodevelopmental tests. This study also measured and controlled for caregiver IQ and
the quality of the home environment.

None of the global tests employed indicated deleterious effects of mercury
exposure.l '7 The authors commented that results from this study might be pertinent for
the US, where dietary intake of fish is similar. Seychelles methylmercury levels are
much higher than in the US. because they eat more fish, not because they consume a few
species of fish with extremely high concentration of mercury.1 '7 Therefore, potential
deficits from methylmercury should be seen in the Seychelles before such effects would
be seen here. Given that the authors used somewhat global scales to measure the
outcome, this study may be less sensitive to certain subtleties than the domain-specific
assessments made in the Faroese cohorts. Recall that the studies from the Faroe Islands
found significant associations whereas the Seychelles cohort did not.

The inconsistency of results about the effects of methylmercury exposure on child
development led to a panel review by the National Academy of Sciences.l ’8 The review
concluded that both the Seychelles and Faroe studies were well-carried out, and that a
variety of differences might account for the inconsistencies. Based upon the panel’s
suggestions, the authors of the Seychelles study reanalyzed their data.”9 They used the
same statistical procedures as reported earlier, except that age at testing and tester were

included as additional covariates in the regression analyses. In addition, the raw test

33

scores were used this time around. The reanalysis confirmed their previous findings.
They reasserted that consumption of methylmercury-containing fish had no association
with adverse events in child development at 5.5 years of age.1 19
In addition to the aforementioned cohort and cross-sectional studies, much of the
literature on vaccine-associated neurologic disorders is generated as case reports.
Fenichel assessed the neurologic risk of immunization with thimerosal-containing

7
’0‘ None of

vaccines using data from the Vaccine Injury Compensation Program (VICP).
the thimerosal-containing vaccines that were reviewed demonstrated a consistent
association with any of the neurologic conditions discussed in this manuscript thus far.
Niu et al. used data from VAERS and the VSD project to compare the safety experience
of thimerosal-containin g hepatitis B vaccines at the time.120 Not only were there no
reports of the previously discussed neurologic conditions, but also, no unexpected serious
events appeared to be caused by the thimerosal-containing vaccines.

Niu again made use of VAERS data to assess neonatal deaths after receipt of
thimerosal-containing hepatitis B vaccines, but this time the number of years of data
coverage was increasedm In this review of autopsies, no mention was made of causes of
death related to CNS damage. In addition, the comparative safety study by Niu et al.
only followed subjects for up to 30 days post-administration of the thimerosal-containing
vaccines.’20 This may not be sufficient time for the occurrence or reporting of certain
neurologic abnormalities.

The fact that inconsistency is a feature of the literature regarding mercury

exposure and adverse neurologic effects does not resolve the debate about whether

mercury metabolites from thimerosal cause neurologic deficits or disorders. Noteworthy,

34

[1F

 

the exposure of interest to our discussion (ethylmercury via thimerosal) has not been
adequately examined, as compared to methyl- and total-mercury. As was already
demonstrated in the description of the literature to date, many different outcomes have
been used to try and measure the effects of mercury compounds. Various neurologic
decrements sought out have included: neurologic abnormalities (e.g., delays in motor
function), cognitive deficits (e.g., performance on attention, language, and memory tests),
global IQ testing, and neurodevelopmental disorders (e. g., ASD and speech or language
delay). The timing of when these adverse effects occur in relation to the exposure of
interest is another consideration when sorting out the causal puzzle.

Temporality (often referred to as ‘time-order’) is a causal criterion that requests
the putative cause precedes the effect in time. In general, the most valid study design for
the proper establishment of temporality is a prospective cohort study. All of the studies
described thus far from the islands of Faroe and Seychelles, as well as in Greenland, were
conducted prospectively using cohorts of mothers and their infants. However, there are
several problems with interpreting these findings that need to be discussed yet.

First, in their studies of infants from the Faroe Islands, Grandjean et al. found
their most significant associations with neurologic impairment using cord-blood

- 3 . .
”I ” Cord blood 13 more representative of exposure to mercury

measures of mercury.
near delivery than of long-term maternal exposure during pregnancy.122 Maternal hair
concentration is a better representation of repeated prenatal exposures. Second, it is
unclear what the most relevant window of exposure is to study. We have demonstrated

that mercury compounds can reach the child in utero; but even when Davidson et al.

analyzed mercury concentrations using both prenatal (maternal) and postnatal (infant)

35

markers, no association was found with adverse effects for children.1 '7 This discussion
also points out the limitations of the relevance and stability of using biomarkers to
measure exposures.

In terms of a dose-response assessment, one of the Faroe Island studiesl '2
reported noticing adverse effects in children with a doubling of the mercury concentration
for the exposure biomarkers. The relevance of a two-fold increase in exposure to
ethylmercury, let alone methylmercury, is unclear. Similarly, Grandjean et al. reported a
decreased neurologic optimality score with a lO-fold increase in the cord-blood mercury
concentration. Such dramatic jumps between exposure categories need to be refined.
Nevertheless, relationships that don’t express a dose-response curve only refute those
hypotheses that are specific enough to predict them.123
Pathogenic Mechanisms

In 2001, the IOM’s Immunization Safety Review Committee commissioned a
background paper on the toxicity of ethylmercury (L. Magos), as a part of their scientific
plausibility assessment on thimerosal-containing vaccines and adverse neurologic
events.124 In regards to the decomposition of methyl- and ethylmercury, methylmercury is
more stable; and the decomposition rate of ethylmercury is greater than that of
methylmercury (both in vivo and in vitro).124 The consequence of these different
decomposition rates is that ethylmercury’s neurotoxic potential declines faster than
methylmercury’s. In addition to differing stability, a methylmercury molecule is 6%

smaller than a molecule of ethylmercury, which facilitates transport through the blood-

brain barrier for methylmercury.‘24

36

Based on the current evidence, L. Magos stated that, “At equal exposure the risk
of neurotoxicity is higher from methylmercury than from ethylmercury...” In fact, there
doesn’t seem to currently be a best hypothesis (if any) regarding the neurotoxic potential
of ethylmercury. No direct studies in either animals or humans of thimerosal exposure of
a level similar to those from vaccination can currently establish a sound biologic model.

Limited animal studies have looked at the toxicity of thimerosal (ethylmercury).
Doses of 1 to 6 ug/kg/day in adult squirrel monkeys resulted in detectable inorganic
mercury within the brain.’25 However, no histopathological changes were observed. The
literature is much more abundant with studies that examined methylmercury exposure in
animals. Gunderson et al. studied crab-eating macaque infants, where the cases had
mothers who were dosed with 50—70 ug/kg/day of methylmercury.126 Animals were
administered different sets of visual recognition memory tests. Infant crab-eating
macaques exposed prenatally to methylmercury performed at a significantly lower level
than did controls (p < 0.02).'26

The effects of in utero methylmercury exposure were also studied using 7- to 9-
year old Macaca fascicularis ..127 The subjects were surviving offspring, infants of
females that had been exposed during pregnancy to doses of methylmercury ranging from
0—90 ug/kg/day. Monkeys exposed in utero to methylmercury showed no deficits in
performance when compared to age-matched controls, as measured by performance on a
spatial delayed alternation task.127

Difficulties in the interpretation of these animal studies are obvious. Varying
sources of mercury, exposure doses, and outcome measures adds even more uncertainty

to the utility of results. The toxicity of low doses of methylmercury and

37

ethylmercury/thimerosal has been assumed to be similar. Given the inconsistencies
presented thus far of studies examining just methylmercury, these prior assumptions
involving ethylmercury need to be tested. Nevertheless, the supposition that
ethylmercury’s actions are analogous to methylmercury’s has driven the bulk of thinking
behind thimerosal-related recommendation and policy changes.

Because no guidelines existed for ethylmercury exposure, the FDA’s 1998 risk
assessment from thimerosal in vaccines used the guidelines for safe exposure to
methylmercury as a guide.‘06 The finding that potential exposure to mercury from
recommended childhood vaccines (during the first six months of life) could exceed the
EPA guidelines, in addition to their literature review of whether thimerosal actually
posed a true health risk, led the FDA to begin talks with manufacturers in April 1999
about developing thimerosal-free vaccinesm’ The FDA planned to send a formal letter to
vaccine manufacturers in July 1999 regarding the topic, at which time the public would
become aware of the issuem’

Fearing potential damage to the public’s trust in the immunization system,
representatives of several organizations involved in US. immunization policy were
quickly summoned to discuss an appropriate plan of action. Disagreements were
widespread and heated between attendees of a June 30, 1999 meeting.'28 Some
participants believed that all thimerosal-containin g vaccines should have been
immediately removed from the market, whereas others doubted the accuracy of the
FDA’s evidence that receipt of thimerosal from vaccines was harmful. Over the course of
the July 4, 1999 holiday weekend, AAP and USPHS officials agreed to release a joint

statement that the birth dose of the hepatitis B vaccine be temporarily delayed for infants

38

of HBsAg-negative status mothers.’28 However, the timing and speed with which the
recommendations were developed led some public health officials to wonder if the joint
statement had been interpreted correctly by physicians and hospital staff, which
eventually led to the Michigan study on the impact of thimerosal recommendations on
infant hepatitis B immunization among infants born to women of unknown HBsAg status.
The Michigan Study
Introduction

Women who are not screened for hepatitis B surface antigen (HBsAg) are more
likely to be HBsAg-positive compared to women who receive prenatal screening,129 and
their infants are less likely to be vaccinated appropriately.130 For infants born to women
of unknown HBsAg status who are actually HBsAg-positive, the consequences are
potentially serious and long lasting. The risk of perinatal infection for an infant born to
an HBsAg-positive woman who is also hepatitis B e anti gen-positive (rates of hepatitis B
e antigen positivity average between 20-30%, depending on maternal origin97) is as high
as 90%.93‘97'131 As many as 90% of perinatally infected infants develop chronic infection,
and up to 25% of HBV-infected newborns will subsequently die of the consequences of
chronic liver disease during adulthood.’32

Screening all pregnant women for HBsAg and immunizing all infants with
hepatitis B vaccine are the cornerstones of efforts to prevent perinatal and early
childhood hepatitis B virus transmission. Vaccination beginning at birth (the birth dose)
is recommended for all children born to HBsAg-positive women and women whose
HBsAg status is unknown and is the preferred schedule for all infants."8 Hospitals and

practitioners who want to provide the birth dose only to infants born to women who are

39

HBsAg-positive and HBsAg-unknown must carefully track HBsAg screening status,
rapidly assess the infant’s risk of early childhood infection, and ensure that vaccine is
provided to infants at risk within 12 hours of birth.

By 1999, many birthing hospitals had recognized the potential for error in
tracking the HBsAg status of pregnant women and the difficulty in assessing each
infant’s risk of early childhood infection. These hospitals established routine policies and
practices to prevent perinatal infection by initiating the hepatitis B vaccine series before
hospital discharge for all newborn infants, providing a safety net for infants born to
HBsAg-positive women who were not identified through screening, whose HBsAg status
is unknown to the birthing hospital, or whose HBsAg status is misidentified.

The secure position of the birth dose as part of the vaccination strategy towards
eliminating hepatitis B was threatened in July of 1999, when the AAP and the USPHS
jointly recommended reducing infant exposure to thimerosal. Specific recommendations
were made to postpone the first hepatitis B vacCine dose until two to six months of age
for infants born to HBsAg-negative women.133 Recommendations for infants born to
HBsAg-positive women, or to women whose HBsAg status was unknown, did not
change.'33 These temporary changes in the recommended routine hepatitis B vaccination
schedule were made because of the flexibility of the hepatitis B schedule for infants born
to HBsAg-negative women, and were to be discontinued when preservative-free hepatitis
B vaccines were licensed. The rationale at the time was that any potential harm that
might result from exposure to mercury in the vaccine would be less when the newborn

was larger and the neurological system was more fully developed. The recommendations

40

received wide publicity, and were transmitted to AAP members via fax 134 and by a
website posting.

By mid-September 1999, adequate supplies of preservative-free hepatitis B
vaccine were available for all newborn infants in the United States. The USPHS then
advocated a return to previous infant hepatitis B vaccination practices, including
administering the first dose of hepatitis B vaccine to all newborn infants in hospitals that
had discontinued the practice.‘35 But by August 1999, the National Immunization
Program was already reporting data that indicated that the recommendation changes had
not been correctly interpreted (CDC, unpublished data, 1999). Several reports suggested
that the change in recommendations led to disruptions in vaccination practices that could
potentially have an impact on coverage rates of infants born to unscreened and even
HBsAg-positive women. Surveys of hospitals in Wisconsin,'36 Chicago,'37 and
Colorado138 demonstrated significant decreases in the number of hospitals that offered
universal hepatitis B vaccination of all newborns after the recommendation changes, and
substantial increases in the percentage of hospitals that did not routinely vaccinate infants
born to HBsAg-positive womenm”I38 Vaccine coverage of infants less than 5 days old
declined 28% in Oregon, and vaccination of infants < 1 month old declined 50% in
Oklahoma during May — June 2000, compared to May — June of 1999.139 A national
survey sample of 773 hospitals in December 1999 indicated similar trends.140 In
Michigan, an unvaccinated infant born at a hospital that had suspended its birth dose
policy died from fulminant hepatitis B in December 1999. Although the mother had
been tested prenatally and was positive for HBsAg, her lab results had been erroneously

reported to the hospital to be “hepatitis-negative”."’9'141

41

A group of researchers (including the author) conducted a study among newborns
in Michigan to assess the impact of the disruption in routine hepatitis B immunization of
newborns born to women whose HBsAg status was unknown. We focused on births
where the mother’s HBsAg status was unknown, because infants born to unscreened
mothers are at greater risk of perinatal infection (compared to screened mothers).
Furthermore, we decided to focus on births where the mother’s HBsAg status was
unknown rather than positive because the Michigan Department of Community Health
has a perinatal hepatitis B coordinator that already tracks births to HBsAg-positive
women to ensure these infants are treated accordingly. Infants born to women that are not
screened do not receive this type of case management, so the likelihood of missing
vaccination is much higher.

Methods

We addressed the primary hypothesis that infants born to mothers who were not
screened for hepatitis B surface antigen were less likely to receive timely
immunoprophylaxis after the recommendations changed in July 1999, and that these
changes persisted during 2000, long after hepatitis B vaccine that did not have thimerosal
as a preservative became widely available.

We used a standardized form to collect demographic information, maternal
HBsAg status, and infant hepatitis B vaccination among Michigan infants born during
three, 60-day time periods. The first time period, March 1 to April 30, 1999 (T1), was
used to assess the baseline proportion of infants born to women of unknown HBsAg
status before the recommendation changes in July 1999. The second time period, July 15

to September 15, 1999 (T2), provided data from the time immediately after the Joint

42

Statement was issued to the approximate time when preservative-free hepatitis B vaccine
first became available, and the USPHS requested that hospitals resume their previous
newborn immunization practices. The third time period, March 1 to April 30, 2000 (T3),
was chosen to represent a time frame when preservative-free hepatitis B vaccine was
widely available, and hospitals should have had an opportunity to resume their previous
vaccination practices. During 2000 and 2001, all MI birthing hospitals were contacted to
determine current and previous policies for offering the first dose of hepatitis B vaccine
during the delivery admission.

Infants born during the three time periods were identified by obtaining birth
records from the Newborn Screening Division of the Michigan Department of
Community Health (MDCH). Michigan newborn screening cards, which have been
required by law for all Michigan infants since 1987, contain a field requesting the HBsAg
status of the infant’s mother (“Mother Hepatitis 8 Surface Antigen Tested? Don ’t Know

; N0 _; YES _; Date ; Result: Positive _; Negative _").

 

The schema we employed was a two-phase approach. The first phase was a pilot
study. We requested hospital charts for the mother and newborn from 200 births with
maternal HBsAg status of “Don’t Know”. The requested charts were evenly divided
between five birthing hospitals in the state of Michigan (i.e., approximately 40 births per
hospital). The births at each hospital were evenly distributed across the three time frames
of the study (e. g., 13/14/13). Our request to review 200 births yielded information on
only 156 births (78%), and few of these women (11%) turned out to truly have unknown

HBsAg statuses.

43

Based upon the results of a similar venture that had just been completed in
Oregon in which hepatitis B vaccine coverage levels were measured before and after July
1999 (Corwith-Jensen H, et a1, Abstract 653, National Immunization Conference,
Denver, CO, 2002), we calculated an initial sample size estimate. Prior to the thimerosal
announcement, 81% of infants in the Oregon study were vaccinated before hospital
discharge. Immediately following the Joint Statement, this figure dropped to 4%; and one
year later when thimerosal-free vaccine was widely available, only 62% of infants born to
women of unknown HBsAg status were being vaccinated before discharge. Using these
vaccination figures, the initial sample size estimate for the Michigan study was 146 births
to women that were truly unscreened. The adjusted sample size estimate that accounted
for the small proportion of unscreened mothers located during the pilot phase was 1,340
births (146/11%). This was the number of births to be reviewed in order to find an
adequate number of newborns born to unscreened mothers. However, during the pilot
phase only 88% of the birth records were suitable for use (due to missing charts,
mismatched matemal-infant pairs, etc.). Thus in order to protect from decreased power,
1,521 (1,340/88%) births needed to be requested to compensate for low pull rates and a

lack of cooperation at the hospitals.

 

To improve efficiency of data collection efforts, we excluded infants born at any
birthing hospital that reported greater than 40% of its births to MDCH as maternal
HBsAg status “Don ’t Know” or missing, or had <15 births for which maternal HBsAg
status was reported as “Don ’t Know After all of our exclusion criteria were
implemented, 19 hospitals remained. Four of these were excluded because of prior

difficulty getting into these hospitals for data abstraction (per MDCH). So from the 15

44

hospitals that remained (n=4,179 total births in the three exposure periods [this total
includes multiple births to one mother]), we requested access to the medical records of
1,520 (36%) randomly selected maternal-infant pairs. Appointments were scheduled for
MDCH staff to visit hospitals and abstract data from subjects’ records.

The exposure status of an infant was based upon whether or not he/she was born
during T1, T2 or T3. The outcome of interest was whether or not an infant received the
first dose of the hepatitis B vaccination series before he/she was discharged from the
hospital. When hour of vaccination was also specified, infants were further classified
according to whether or not they were treated within 12 hours of birth (the recommended
interval). The same chart abstraction form was used for recording information on both
exposure status and outcome. Though data on exposure and outcome were gathered
simultaneously, we analyzed and interpreted this data as for a cohort study, since the time
order is clear (vaccination status is subsequent to period of delivery).

After the completion of data entry, every tenth birth was systematically reviewed
for discrepancies between database tables and abstraction forms. Data tables were
subsequently imported into SAS/STAT (Cary, NC) for appropriate analyses. PROC
UNIVARIATE was also used to check’for outlying data values. Results were first
analyzed using the FREQ procedure in SAS to calculate chi-square statistics (x2) and p-
values. In the presence of small expected values, Fisher’s exact test was used.

Results

Nineteen hospitals met the inclusion criteria, and after four hospitals were

excluded (see above), 13 agreed to participate (Table 1). All of these hospitals reported

that, before the Joint Statement was released, they had a policy to offer a birth dose of

45

hepatitis B vaccine to all infants. After the Joint Statement, all 13 hospitals stopped these
policies. By the time of follow up in 2000-2001, 9 of 13 had resumed the previous policy
of offering hepatitis B vaccine to all newborns, and 2 more hospitals reported that they
planned to resume in late 2001 or 2002.

TABLE 1. Average annual birth rate (1998-1999) of participating hospitals, by
hospital.

 

Hospital #01 #02 #03 #04 #05 #06 #07 '

 

Average annual 1,681 779 3,074 625 1,861 3,024 4,655

birth rate

 

Hospital #08 #09 #10 #11 #12 #13 Avg.

 

Average annual 3,879 3,310 2,239 4,509 5,954 6,089 3,206

birth rate

 

 

 

 

 

 

 

 

 

 

Of 68,776 total Michigan births during these three time periods (T1, T2, and T3),
9,206 had maternal HBsAg status recorded as “Don ’t Know". We requested 1,520 births

for review that were evenly distributed across T1 (3/1/99 — 4/30/99), T2 (7/15/99 —

 

1""

9/15/99) and T3 (3/1/00 — 4/30/00). Hospital personnel allowed access to the maternal
and infant medical records for 89% of requested births (1,355/1,520). After the
completion of data collection, 89% of these births could be used for subsequent analyses
(1,201/1,355). The remaining 154 were excluded because some or all of the screening
and vaccination information was missing from the chart. Of these 1,201 births, 216

(18%) were to women whose HBsAg status was truly unknown at the time of hospital

46

discharge. For the remaining 985 infants, maternal HBsAg status actually was recorded
in the chart (HBsAg (+)=5 mothers and HBsAg (-)=980 mothers). It appeared that
maternal hepatitis B screening status was independent of the time frames of the study
(Pearson Chi-Square Statistic [x2]=4.15, p-value=0. 13) [Table 2].

TABLE 2. Results of maternal HBsAg status upon medical record review, by
recommendation period.

 

 

 

 

T1 * T2* T3* Total P-value
Number of 418 389 394 1,201
charts (34.8) (32.4) (32.8)
reviewed (% )
Maternal Screened (% ) 343 308 334 985
HBsAg (34.8) (31.3) (33.9) (82.0)
Unknown (%) 75 81 60 216 0.13
(34.7) (37.5) (27.8) (18.0)

 

 

 

 

 

 

 

 

 

*Denotes the three exclusive time periods under study: T1 = March-April 1999,
T2=July 15-September 15, 1999, and T3=March-April 2000.

The median age of the 216 women of unknown HBsAg status was 26.6 years
(range 14-43 years), compared to 26.9 years for the women of known HBsAg status
(range 14-57 years). Table 3 summarizes other demographic and prenatal care
characteristics. Maternal characteristics did not differ significantly by time periods of the

study.

47

 

 

TABLE 3. Demographic and prenatal care characteristics of 1,201 Michigan
women by HBsAg status, 1999-2000.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Characteristic Number (%)
HBsAg status - unknown HBsAg status - known
(n=216) (n=985)
Race
White 85 (39.4) 450 (45.7)
Black 120 (55.6) 438 (44.5)
Other 11 (5.0) 97 (9.9)
First child 52 (24.1) 258 (26.2)
First prenatal visit during 33 (35.1) 549 (55.7)
first trimester*
Greater than or equal to 5 32 (39.5) 669 (67.9)
prenatal visits*
Received screening during 6 (2.8) 66 (6.7)
admission for delivery ‘
Delivery time period
March 1-April30, 1999 75 (34.7) 343 (34.8)
July 15-September 15, 1999 81 (37.5) 308 (31.3)
March 1-April30, 2000 60 (27.8) 334 (33.9)

 

 

 

 

 

*Data unavailable for many women.

Of infants born to women of unknown HBsAg status during T1, 14/74 (19%)
received the vaccine within 12 hours of birth, compared to 1/79 (1%) infant born during
T2 (pr-0.0002). While only 8/58 (14%) of infants born during T3 received the vaccine
within 12 hours of birth, this was not significantly different from the proportion during
T1. Four infants vaccinated before discharge did not have a time of vaccination recorded.

For most infants born to unscreened women during T2, vaccination was not only
delayed beyond 12 hours, but also did not occur before discharge. Of infants born to

unscreened women during T2, 6/81 (7%) received a dose of hepatitis B vaccine

48

before hospital discharge, compared to 40/75 (53%) infants born during T1
(x2=39.50, p<0.0001) [Table 4]. However, the proportion of infants vaccinated before
hospital discharge during T3 (34/60 [57%]) did not significantly differ from T1.

TABLE 4. Proportion of Michigan infants born to women of unknown HBsAg
status who were vaccinated before discharge or within 12 hours, by time period.

 

 

 

 

 

Time Vaccinated before P-value Vaccinated within 12 hours P-value
Period discharge, n/total (%) of birth, n/total (%)

T1 40/75 (53) Ref. 14/74* (19) Ref.
T2 6/81 (7) <0.0001 1/80* (1) <0.001
T3 34/60 (57) N3§ 8/58* (14) Ns§

 

 

 

 

 

* Time of vaccination not available for one infant born during T1, one infant born
during T2, and 2 infants born during T3.
NS§ = not statistically significant

When the analysis was restricted to infants who received vaccine before
discharge, a larger proportion (14/39 [36%]) in T1 received the vaccine within 12 hours
(as would be recommended if the mother were known to be antigen positive), compared
to T2 (1/5 [20%]) and T3 (8/32) [25%], but these differences in proportions were not
statistically significant. We also analyzed the data with the outcome defined as the
hospitals having offered the vaccine (even if the parents refused it). Reanalysis using this
outcome yielded no qualitative difference, as only five infants were different with regard
to the two outcomes, including two born during Tl, one during T2 and two during T3.
Discussion

The primary goal of our study was to quantify the impact of changing
recommendations on the vaccination of infants against hepatitis B. Specifically, we were
most interested in those newborns whose mothers’ HBsAg status was unknown. Women

who have not been screened or have not received prenatal care demonstrate a higher

49

 

prevalence of chronic HBV infection.‘29 Therefore, babies born to unscreened mothers
are at higher risk of infection as compared to those born to screened women, and as a
group these infants must be considered as ‘high risk.’

Previous studies have shown that many hospitals discontinued policies that
provided for hepatitis B vaccination of all newborns after the Joint Statement was issued,
and that these policies were not promptly reinstated in some hospitals, even after
preservative-free hepatitis B vaccine became available. Our study demonstrates that a
significant decline in timely hepatitis B vaccine coverage occurred among infants born to
women of unknown HBsAg status immediately after the Joint Statement was released. In
Michigan, timely hepatitis B vaccination of infants born to women of unknown HBsAg
status virtually ceased within our study hospitals in July and September of 1999, despite
specific language in the Joint Statement indicating that hepatitis B vaccination practices
for these infants had not changed. Changes in vaccination coverage during the three time
periods coincided with changes in routine vaccination policies. This study provides
further evidence that rapid policy changes associated with evolving, largely theoretical
concerns about the safety of thimerosal were misinterpreted.

Our study revealed that 18% of the births reviewed involved mothers whose
HBsAg status was unknown. This figure is higher than the 13% that was reported from
the state of Michigan’s “Assessment of Progress Toward Goals to Prevent Perinatal HBV
Transmission,” (unpublished results, MDCH, 1998). However, that survey was based
upon all pregnant women in the state, whereas our study is much more limited in its
scope. Our target population included only those infants born in Michigan across several

study periods whose mothers were unlikely, based upon maternal HBsAg information

50

from newborn screening cards, to have been screened for hepatitis B. Although,
information from screening cards indicating that a mother’s HBsAg status was unknown
was likely a random process at each hospital (i.e., varied by who filled out the card on
what day), rather than being due to factors specific to these women. Nevertheless, the
proportion of unscreened mothers presented above (13%) is not meant to be compared
directly to our study’s findings.

The time periods of our study were not related to the rigor with which mothers
were screened for hepatitis B. The study periods (which varied by vaccine
recommendations at the time) were, however, significantly associated with infant
vaccination proportions. Newborns of mothers with unknown HBsAg status were much
less likely to be vaccinated during the time period after the Joint Statement was released
(T2). Infants born soon after thimerosal-free vaccine became widely available (T3) were
also less likely to be vaccinated compared to infants born during Tl, although the
difference in proportions vaccinateddid not reach the level of statistical significance.
However, given the fact that vaccination of these newborns was already quite low during
T1 (53%), it is not surprising that we were unable to show a significant difference one
year later (T3).

An important aspect of our study worth noting is that we actually conducted chart
review to confirm maternal screening status and infant vaccination, whereas previous
surveys relied on interview responses from hospital personnel to generate their
conclusions. Because we did not assess vaccination data beyond the year 2000, our

vaccine coverage data may not reflect the current situation. Our results should be

51

interpreted as indicating that hospitals were slow to return to practices in place before the
Joint Statement, even if they may well have returned to them eventually.

It is unknown whether our findings reflect a nationwide reduction during the alert
period in first dose hepatitis B vaccine coverage for infants born to women of unknown
HBsAg status. We only examined birth records at a subset of Michigan hospitals that
consistently filled out information on maternal HBsAg status on newborn screening
cards. These hospitals may not have been representative of all Michigan birthing
hospitals or those in other states. By virtue of the exclusion criteria we employed, small
rural hospitals (usually staffed by family practice physicians) would be underrepresented
in our study; and hospital size, location, and makeup of the hospital attending staff
strongly influence birthdosing practices.I36 However, hospitals that were unable or
unwilling to adequately complete newborn screening cards would probably not be more
likely to successfully screen and provide timely vaccination to infants born to unscreened
women. In addition, a similar pattern and magnitude in reduction of coverage for infants
born to women of unknown HBsAg status was observed in Oregon (Corwith-Jensen H, et
al, Abstract 653, National Immunization Conference, Denver, CO, 2002).

Several explanations for the observed effects on newborn vaccine coverage are
possible. First, there may have been confusion about which infants should have
vaccination deferred, despite the specific language used in the Joint Statement. In a
Colorado survey done in early 2000 of persons who identified themselves as being
responsible for nursery vaccination policy, 71% learned of the Joint Statement
recommendations from colleagues or the news media, rather than from the health

138

department or a professional society. The precise language used to delineate which

52

infants should have vaccine deferred may have been lost in summaries provided by
colleagues or the news media.

Secondly, practitioners and hospitals may have considered the theoretical risk
from thimerosal exposure to exceed the risk of perinatal infection due to missed
immunoprophylaxis for infants born to women of unknown HBsAg status. The haste
with which the Joint Statement was developed and publicized may have led practitioners
and hospitals to conclude that reducing exposure to thimerosal in vaccines was a public
health emergency for all infants.

Finally, practitioners and other delivery hospital personnel may not have been
aware that infants born to women of unknown HBsAg status were no longer being
routinely immunized. Practitioners may have assumed that prenatal screening could
successfully identify all HBsAg-positive women, and grown accustomed to the safety net
that universal hepatitis B vaccination of newborns provided. Formerly (before the Joint
Statement - T1), each nursery probably had a routine involving the ascertainment of
hepatitis B information about every birth, when practically all newborns were receiving
the birth dose of vaccine. But after the thimerosal alert, the absence of such a routine
probably allowed for oversights that wouldn’t have happened during times of near-
universal birth dosing, resulting in newborns being discharged without any hospital staff
ever checking the maternal record for hepatitis B status.

Misinterpretation of public health and professional society recommendations
regarding thimerosal might have been avoided if a more deliberate process had been
followed as recommendations were developed. The need for urgent changes in

vaccination policy, without review by the Advisory Committee on Immunization

53

 

Practices (ACIP) and the National Vaccine Advisory Committee (NVAC)’28, was not
supported by the existing data. At the time the Joint Statement was issued, no harm from
the mercury in thimerosal-containing vaccines had been demonstrated, and there was no
consensus about the risk of exposure to the quantities of mercury-containin g compounds
present in vaccines.

Since the universal immunization strategy was released in 1991, the number of
children born to HBsAg-positive women in the US. has risen substantially.142 It is well

known that timely immunoprophylaxis provides the best means for interrupting

 

transmission of hepatitis B virus. For example, the 1998 US. birth cohort would have
experienced 6,800 perinatal infections and 18,700 more infections due to horizontal
transmission by 10 years of age if no vaccination had taken place.142 Infants born to
women of unknown HBsAg status should be treated as though the mother is a chronic
carrier of hepatitis B.

Inadequate immunoprophylaxis of infants born to HBsAg-positive women has
been associated with a failure to document maternal screening results in the delivery
room prior to birth.143 Based upon our results from babies delivered by mothers with

unknown HBsAg statuses, we can only speculate on what happened to babies of HBsAg-

 

positive mothers. Before the Joint Statement, many hospitals relied upon provision of i;
hepatitis B vaccine to all infants at birth as a safety net, so that infants whose mothers

were unscreened but HBsAg-positive would automatically get immunoprophylaxis

without the need to wait for lab results and make a decision at delivery. After the Joint

Statement, these hospitals would have returned to a policy of making the decision on a

case-by-case basis, thus introducing the opportunity for error. In addition, it seems likely

54

that hospitals that missed vaccinating unscreened mothers’ newborns would have a
greater potential to miss administering vaccine to some infants born to HBsAg-positive
women also. In fact, this sort of error was documented in the case of the Michigan infant
alluded to earlier.

The recommendation changes that occurred in July 1999 were not based on a
proven causal relationship. The effects of exposure to methylmercury were assumed to
hold true for ethylmercury, which is the metabolite of thimerosal preservative. Available
data does indicate, however, that the toxicological properties and pattern of tissue
disposition for ethylmercury compounds are qualitatively similar to those of the
methylmercury compounds.”"‘”5 The recommendation changes in July 1999 were the
result of a cautious approach when making alterations to proven immunization programs.
The theoretical risk posed by thimerosal-preservative in vaccines was viewed as being
greater than the known benefits of vaccinating newborns prior to discharge. However, the
detrimental effects seen on infant vaccination coverage in Michigan should be viewed as
a caution when changes to existing immunization schedules are considered in the future.

Beginning hepatitis B vaccination at birth for all infants has been the preferred
schedule advocated by the AAP since their policy was first published in Pediatrics in
1992,146 as it was in 1999 before the Joint Statement. The ACIP recently stated its
preference for the birth dose schedule. Our data indicates that there is a need to renew
efforts to begin hepatitis B vaccination at birth, especially for infants at higher risk of
infection. Although hepatitis B vaccine coverage of infants born to unscreened women in
Michigan returned to near baseline (57% by discharge, 14% within the first 12 hours of

life) by mid 2000, physicians and birthing centers must increase vaccine coverage for

U]
Q)!

 

 

A.-

these newborns beyond this unacceptably low level. Efforts to improve hepatitis B
vaccine birth dose coverage in Michigan are underway, and include provision of free
vaccine and educational resources to hospitals and staff. A recent report from MDCH
indicated that 83 of 102 (81.4%) birthing hospitals in Michigan have now implemented or
reinstated policies to offer hepatitis B vaccine to all newborns prior to discharge.'47
Hopefully these policies will help provide better protection for Michigan newborns

against hepatitis B, especially among those at highest risk of infection.

56

CONCLUSION

The examples of damage done to immunization coverage discussed in this thesis
should serve as a warning when dealing with vaccine safety issues. In the case of Andrew
Wakefield and colleagues’ findings, the warning pertains to responsibility — the
responsibility researchers have to apply appropriate criteria for determining causality
when publishing their results. In response to criticism his efforts have received relating
MMR vaccination to autism, Wakefield stated:
“Assumptions of vaccine safety, based upon inadequate safety trials and dogma
contribute largely to confusion and public loss of confidence in vaccination.”73

Ironically, Wakefield’s own crusade has been at the forefront of this confusion,
and he has not wavered despite several studies that have refuted his conclusions at the
group level.”79 The editor of the Lancet defended the joumal’s decision to publish the
controversial report, mentioning the fact that the journal did run a commissioned
commentary in the same issue by Chen and DeStefano.I48 However, common sense tells
us which of the two references (Wakefield or Chen & DeStefano) a media journalist is
going to pick up on.

Whereas health officials in the United Kingdom have backed the safety of the
MMR vaccine, other examples of diminished vaccination coverage can be attributed to
public health policymakers. Unlike the MMR controversy, it appears as though a causal
relationship may exist between whole-cell pertussis vaccine (DTP) and certain adverse,
neurologic events (e. g., acute encephalopathy).50 Despite such findings, the estimated 45

million cases of pertussis that occur annually are more than enough reason to be cautious

when examining immunization policy. Although anti-vaccine movements have had some

57

beneficial effects (e. g., promoting interest in funding research for improving vaccine
safety and surveillance of adverse events), the DTP controversy provides a strong
international example of imprudent policy decisions leading to outbreaks of disease.53 For
example, following the elimination of whole-cell pertussis vaccine use in Japan, a
pertussis epidemic occurred with more than 13,000 cases and 41 deaths.53

Although it’s not currently known whether thimerosal-related policy changes led
to a nationwide rise in perinatal infections with hepatitis B virus, the Michigan study
_ presented here demonstrated decreased vaccine coverage among a group of hi gh-risk
children in this state. Abrupt changes in established vaccination recommendations,
originally intended only to pertain to low risk children (i.e., infants whose mothers’ were
known to be HBsAg-negative), led to decreased coverage among higher risk children.
The unintended impact of the thimerosal controversy on hepatitis B vaccine coverage of
infants born to women of unknown HBsAg status should serve to remind public health
and professional associations that changes in established recommendations, especially if
they occur without timely communication and education of health-care providers, may
result in unexpected changes in vaccination practices. Anticipating the need for public
health to respond to future safety controversies, the IOM report on thimerosal-containin g
vaccines also called for “. . .a review and assessment of how public health policy
decisions are made under uncertainty” as well as research on how to improve strategies
used to communicate rapid changes in vaccine policy.’24

The need may also currently exist for research on how to boost public confidence
in US. immunization policy. One hundred years ago, this confidence was easily obtained

r49

by making vaccination the popular thing to do. In London at the time, there was public

58

""7“" . I'm-r

 

 

apathy regarding the need for smallpox vaccination. But once a member of the local stock
exchange donned a red ribbon on the vaccinated arm (indicating to others not to press
against the arm), the act of smallpox vaccination became fashionable, and “a thousand
ribboned arms may be met in five minutes anywhere in the city?”49 However, it will take
more than a simple fashion statement for public health officials today to regain the level

of public support for immunization that existed just a few years ago.

 

 

10.

11.

REFERENCES

Plotkin SL, Plotkin SA. Short History of Vaccination. In: Plotkin S, Orenstein W,
editors. Vaccines. 3rd ed. Philadelphia: WB. Saunders Company; 1999. p. 1-12.

Centers for Disease Control and Prevention. Achievements in public health, 1900-
1999: impact of vaccines universally recommended for children — United States,
1990-1998. MMWR l999;48(RR-12):243-248.

Centers for Disease Control and Prevention. Principles of Vaccination. In:
Atkinson W, Wolfe C, Hermiston S, et al., editors. Epidemiology and Prevention
of Vaccine-Preventable Diseases. 6‘h ed. Atlanta: Public Health Foundation; 2001.
p. 1-8.

Barquet N, Domingo P. Smallpox: The Triumph Over the Most Terrible of the
Ministers of Death. Annals of Internal Medicine 1997;127(8(Part 1)):635-42.

. American Public Health Association. Smallpox. In: Chin J, editor. Control of

Communicable Diseases Manual. 17‘h ed. Baltimore: United Book Press, Inc.;
2000. p. 455-457.

Strathdee SA and Loughlin AM. Vaccines — Past, Present, and Future. In: Nelson
KE, Williams CM, Graham NM, editors. Infectious Disease Epidemiology.
Gaithersburg: Aspen Publishers, Inc.; 2001. p. 255-282.

Giesecke J. Mathematical models for epidemics. In: Giesecke J. editor. Modern
Infectious Disease Epidemiology. New York: Oxford University Press; 1994. p.
109—123.

Feikin D, Lezotte D, Hamman R, et al. Individual and Community Risks of

Measles and Pertussis Associated with Personal Exemptions to Immunization.
J AMA 2000;284:3145-50.

Rota J S, Salmon DA, Rodewald LE, et al. Processes for Obtaining Nonmedical
Exemptions to State Immunization Laws. Am J Public Health 2001;91:645-648.

Webster’s NewWorld Dictionary. Third College Edition. New York: Prentice
Hall; 1994. Safe; p. 1181.

Centers for Disease Control and Prevention. Vaccine Safety. In: Atkinson W,
Wolfe C, Hermiston S, et al., editors. Epidemiology and Prevention of Vaccine-
Preventable Diseases. 6‘h ed. Atlanta: Public Health Foundation; 2001. p. 265-
276.

60

 

 

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Chen RT, Glasser JW, Rhodes, PH, et al. Vaccine Safety Datalink Project: A New
Tool for Improving Vaccine Safety Monitoring in the United States. Pediatrics
1997;99:765-73.

Chen RT. Special methodological issues in pharrnacoepidemiology studies of
vaccine safety. In: Strom BL, editor. Pharmacoepidemiology. Sussex, United
Kingdom: John Wiley & Sons; 1994.

Chen RT, DeStefano F, Davis RL, et al. The Vaccine Safety Datalink:
immunization research in health maintenance organizations in the USA. Bulletin
of the World Health Organization 2000;78(2): 186-194.

Program announcement 01112, Clinical Immunization Safety Assessment Centers
(CISA), published in the Federal Register dated June12, 2001, Vol. 66, No. 113,
p. 31649-31652.

Health Resources and Services Administration. Background Information on
VICP. Available from: URL: http://www.hrsa.gov/osp/vicp/abdvic.htm

Zwillich T. Controversy erupts over US vaccine-liabilities fund. Reuters Health
Information [online] 2001 Nov 01. Available from: URL:
h@://\xtw'w.reutershealth.com

 

Centers for Disease Control and Prevention. General Recommendations on
Immunization. In: Atkinson W, Wolfe C, Hermiston S, et al., editors.
Epidemiology and Prevention of Vaccine-Preventable Diseases. 6th ed. Atlanta:
Public Health Foundation; 2001. p. 9-25.

Lieberman P. Anaphylaxis and Anaphylactoid Reactions. In: Middleton E, Ellis
EF, Yunginger JW, et al. Allergy - Principles & Practice. Volume H. St. Louis:
Mosby; 1998. p. 1079-92.

Anderson HR, Poloniecki JD, Strachan DP, et a1. Immunization and Symptoms of
Atopic Disease in Children: Results From the International Study of Asthma and
Allergies in Childhood. Am J Public Health 2001;91:1126-29.

Institute of Medicine. Immunization Safety Review: Multiple Immunizations and
Immune Dysfunction. 2002 February 20. Available from: URL:
http://www.nagedu/books/030908328 l/html/

Centers for Disease Control and Prevention. Poliomyelitis. In: Atkinson W, Wolfe
C, Hermiston S, et al., editors. Epidemiology and Prevention of Vaccine-
Preventable Diseases. 6‘h ed. Atlanta: Public Health Foundation; 2001. p. 85-100.

61

23.

26.

27.

28.

29.

31.

32.

33.

34.

American Public Health Association. Poliomyelitis. In: Chin J, editor. Control of
Communicable Diseases Manual. 17th ed. Baltimore: United Book Press, Inc.;
2000. p. 398-405.

. The Children’s Hospital of Philadelphia. Polio Vaccine. Vaccine Education

Center — Resources for Healthcare Professionals. Available from: URL:
http://www.vaccinechopedu/each vaccine2.shtml

 

. Strebel PM, Ion-Nedelcu N, Baughman AL, et al. Intramuscular injections within

30 days of immunization with oral poliovirus vaccine — a risk factor for vaccine-
associated paralytic poliomyelitis. J Engl J Med 1995;332:500-6.

Centers for Disease Control and Prevention. Notice to Readers: Recommended
Childhood Immunization Schedule — United States, 2002. MMWR
2002;51(02):31-33.

Centers for Disease Control and Prevention. Public Health Dispatch: Update:
Outbreak of Poliomyelitis — Dominican Republic and Haiti, 2000-2001. MMW R
2001;50(39):855-6.

Centers for Disease Control and Prevention. Public Health Dispatch: Acute
Flaccid Paralysis Associated with Circulating Vaccine-Derived Poliovirus —
Philippines, 2001. MMWR 2001;50(40):874-5.

Kemper CA. Recent Outbreak of Oral Poliovirus-1 Vaccine-Derived Viral
Infection in Haiti and the Dominican Republic. Infectious Disease Alert 2001;83-
5'.

. Centers for Disease Control and Prevention. Progress Toward Global Eradication

of Poliomyelitis, 2001. MMWR 2002;51(12):253-6.

Nelson KE. Early History of Infectious Disease: Epidemiology and Control of
Infectious Diseases. In: Nelson KE, Williams CM, Graham NM, editors.
Infectious Disease Epidemiology. Gaithersburg: Aspen Publishers, Inc.; 2001. p.
3-16.

Rosen G. A History of Public Health. Baltimore: Johns Hopkins University Press;
1993.

Reed W, Carroll J. The Prevention of YF. Med Rec. NY. 1901;60:641-649.
American Public Health Association. Yellow fever. In: Chin J, editor. Control of

Communicable Diseases Manual. 17‘h ed. Baltimore: United Book Press, Inc.;
2000. p. 553-558.

62

l,
i;

 

35

37.

39.

40.

41.

42.

43.

44.

45.

46.

47.

. Theiler M and Smith HH. Use of yellow fever virus modified by in vitro

cultivation for human immunization. J Exp Med 1937;65:787-800.

. Vasconcelos PFC, Luna EJ, Galler R, et al. Serious adverse events associated with

yellow fever 17DD vaccine in Brazil : A report to two cases. Lancet
2001;358(9276):91-97.

Martin M, Tsai TF, Cropp B, et al. Fever and multisystem organ failure associated
with 17D-204 yellow fever vaccination: A report of four cases. Lancet
2001;358(9276):98-104.

. Chan RC, Penney DJ, Little D, et al. Hepatitis and death following vaccination

with 17D-204 yellow fever vaccine. Lancet 2001;358(9276):121-122.

Vasconcelos PFC, Rosa APAT, Rodrigues SG, et al. Yellow Fever in Para’ State,
Amazon Region of Brazil, 1998-1999 : Entomologic and Epidemiologic Findings.
Emerging Infectious Diseases 2001;7(3):565-569.

Monath TP. Yellow Fever. In: Plotkin S, Orenstein W, editors. Vaccines. 3rd ed.
Philadelphia: W.B. Saunders Company; 1999. p. 815-998.

Marianneau P, Georges-Courbot MC, and Deubel V. Rarity of adverse effects
after 17D yellow-fever vaccination. Lancet 2001;358(9276):84-85.

Centers for Disease Control and Prevention. Notice to Readers: Fever, Jaundice,
and Multiple Organ System Failure Associated With 17D-Derived Yellow Fever
Vaccination, 1996-2001. MMWR 2001;50(30):643-5.

American Public Health Association. Pertussis. In: Chin J, editor. Control of
Communicable Diseases Manual. 17th ed. Baltimore: United Book Press, Inc.;
2000. p. 375-379.

Centers for Disease Control and Prevention. Pertussis. In: Atkinson W, Wolfe C,
Hermiston S, et al., editors. Epidemiology and Prevention of Vaccine-Preventable
Diseases. 6‘h ed. Atlanta: Public Health Foundation; 2001. p. 67-83.

Deen JL, Mink CA, Cherry JD, et al. Household contact study of B. pertussis
infections. Clin Infect Dis 1995 ;21 :1211-1219.

Mortimer EA Jr. Pertussis and its prevention : a family affair. J Infect Dis
1990;161:473-479.

Chotani RA, Roghmann MC, and Trish MP. Nosocomial Infections. In: Nelson

KE, Williams CM, Graham NM, editors. Infectious Disease Epidemiology.
Gaithersburg: Aspen Publishers, Inc.; 2001. p. 357-408.

63

 

48.

49.

50.

51.

52.

53.

55.

56.

57.

58.

59.

60.

Madsen T. Vaccination against whooping cough. JAMA 1933;101:187-8.

Gale JL, Thapa PB, Wassilak SGF, et al. Risk of serious acute neurological illness
after immunization with diphtheria-tetanus-pertussis vaccine. JAMA
1994;271(l):37-41.

Institute of Medicine. Adverse Effects of Pertussis and Rubella Vaccines — Health

Promotion and Disease Prevention. Washington, DC: National Academy Press;
September 1991.

Barlow WE, Davis RL, Glasser JW, et al. The risk of seizures after receipt of
whole-cell pertussis or measles, mumps, and rubella vaccine. N Engl J Med
2001;345(9):656-61.

World Health Organization. Vaccines, Immunization, and Biologicals — Pertussis.
Available from: URL: http://www.who.int/vaccines-
surveilIance/deseasedesc/RSquertus.htm

 

 

Gangarosa EJ, Galazka AM, Wolfe CR, et al. Impact of Anti-Vaccine Movements
on Pertussis Control : The Untold Story. Lancet 1998;351(9099):356-361.

. Romanus V, Jonsell R, and Bergquist SO. Pertussis in Sweden after the cessation

of general immunization in 1979. Pediatr Infect Dis J 1987;6z364-371.

American Public Health Association. Measles. In: Chin J, editor. Control of
Communicable Diseases Manual. 17th ed. Baltimore: United Book Press, Inc.;
2000. p. 330-335.

American Public Health Association. Mumps. In: Chin J, editor. Control of
Communicable Diseases Manual. 17‘h ed. Baltimore: United Book Press, Inc.;
2000. p. 353—355.

American Public Health Association. Rubella. In: Chin J, editor. Control of

Communicable Diseases Manual. 17’h ed. Baltimore: United Book Press, Inc.;
2000. p. 435-440.

Centers for Disease Control and Prevention. Measles. In: Atkinson W, Wolfe C,
Hermiston S, et al., editors. Epidemiology and Prevention of Vaccine-Preventable
Diseases. 6’h ed. Atlanta: Public Health Foundation; 2001. p. 117-140.

Wakefield AJ, Murch SH, Anthony A, et al. Ileal-lymphoid-nodular hyperplasia,
non-specific colitis, and pervasive developmental disorder in children. Lancet

1998;351:637-41.

American Psych Assoc (APA). Diagnostic & Stat. Manual of Mental Disorders.
4‘h edition. Washington, DC: APA; 1994.

64

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

Dalton R and Forman MA. Pervasive Developmental Disorders and Childhood
Psychosis. In: Behrman RE, Kliegman RM, and Jenson HB, editors. Nelson
Textbook of Pediatrics. 16th ed. Philadelphia: W.B. Saunders Company; 2000. p.
87-88.

Bernard S, Enayati A, Redwood L, et al. Autism: a novel form of mercury
poisoning. Medical Hypothesis 2001;56(4):462-471.

Bailey A, Phillips W, Rutter M. Autism: towards an integration of clinical,
genetic. neuro-psychological, and neurobiological perspectives. J Child Psychol

Psychiatry 1996;37(1):89-126.

Walker-Smith JA, Andrews J. Alpha-1 antitrypsin, autism and celiac disease.
Lancet 1972;2z146-51.

Lee JW, Melgaard B, Clements CJ, et al. Autism, inflammatory bowel disease,
and MMR vaccine. Lancet 1998;351:905.

Black D, Prempeh H, Baxter T. Autism, inflammatory bowel disease, and MMR
vaccine. Lancet 1998;351:905-6.

Beale AJ. Autism, inflammatory bowel disease, and MMR vaccine. Lancet
1998;351:906.

O’Brien SJ, Jones IG, and Christie P. Autism, inflammatory bowel disease, and
MMR vaccine. Lancet 1998;351:906-7.’

Payne C and Mason B. Autism, inflammatory bowel disease, and IVIMR vaccine.
Lancet 1998;351:907.

Bedford H, Booy R, Dunn D, et a1. Autism, inflammatory bowel disease, and
MR vaccine. Lancet 1998;351:907.

Lindley KJ and Milla PJ. Autism, inflammatory bowel disease, and MMR
vaccine. Lancet 1998;351:907-8.

Begg N, Ramsey M, White J, et al. Media dents confidence in MMR vaccine.
BMJ 1998;316:561.

Wakefield AJ. Autism, inflammatory bowel disease, and MMR vaccine. Lancet
1998;351:908.

Murch S, Thomson M, and Walker-Smith JA. Autism, inflammatory bowel
disease, and MMR vaccine. Lancet 1998;351:908.

65

. c t- .ft‘tx 1:3

 

 

 

75.

76.

77.

78.

79.

80.

81.

82.

84.

85.

86.

87.

Horton R. Autism, inflammatory bowel disease, and MR vaccine. Lancet
1998;351:908-9.

Peltola H, Patja A, Leinikki P, et al. No evidence for measles, mumps, and rubella
vaccine-associated inflammatory bowel disease or autism in a l4-year prospective
study. Lancet 1998;351(9112):1327-28.

Taylor B, Miller E, Farrington CP, et al. Autism and measles, mumps, and rubella
vaccine: no epidemiological evidence for a causal association. Lancet
1999;353:2026-29.

Institute of Medicine. Immunization Safety Review: Measles-Mumps-Rubella
Vaccine and Autism. 2001. Available from: URL:
http://www.nap.edu/books/0309074479/html/

Halsey NA, Hyman SL, and the Conference Writing Panel. Measles-Mumps-
Rubella Vaccine and Autistic Spectrum Disorder: report From the New
Challenges in Childhood Immunizations Conference Convened in Oak Brook,
Illinois, June 12-13, 2000. Pediatrics 2001;107(5):1-23.

Ramsay S. UK starts campaign to reassure parents about MMR-vaccine safety.
Lancet 2001;357(9252):290.

Birmingham K, and Cimons M. Reactions to the MMR immunization scare.
Nature 1998;4z478-479.

Anonymous. Measles, MMR, and autism: The confusion continues. Lancet
2000;355(9213):1379.

. Vastag B. Congressional Autism Hearings Continue - No Evidence MMR

Vaccine Causes Disorder. JAMA 2001;285(20):2567-69.

CDR Weekly — Communicable Disease Report 1998;8(5):43.

Woodman R. UK: Mumps outbreak after parents shun vaccine. Reuters Health
Information [online] 2001 Aug 06. Available from: URL:

http://www.reuter3health.com

Fears grow as mumps cases rise. BBC News 2001 Aug 30. Available from: URL:
httfl/news.bbc.co.uk/hi/english/health/newsid 1516000/1516706.stm

Bennett R. MMR Gag Angers Health Officials. Financial Times 2001 Dec 21 :p.1.
Available from: URL: www.ft.com

66

 

 

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

Meikle J. Measles Outbreak Raises Fears Over Parents Shunning MMR Vaccine.

Guardian Unlimited (UK) 2002 Feb 02:P.5. Available from: URL:
www.guardian.co.uk

 

Nelson KB and Thomas DL. Viral Hepatitis. In: Nelson KE, Williams CM,
Graham NM, editors. Infectious Disease Epidemiology. Gaithersburg: Aspen
Publishers, Inc.; 2001. p. 567-610.

Lurman A. Eine icterus epidemic. Berl Klin Worchenschr. 1885;22:20-23.

American Public Health Association. Viral Hepatitis B. In: Chin J, editor. Control
of Communicable Diseases Manual. l7lh ed. Baltimore: United Book Press, Inc.;
2000. p. 243-251.

Centers for Disease Control and Prevention. Hepatitis B. In: Atkinson W, Wolfe
C, Hermiston S, et al., editors. Epidemiology and Prevention of Vaccine-
Preventable Diseases. 6‘h ed. Atlanta: Public Health Foundation; 2001. p. 208-
230.

Stevens CE, Neurath RA, Beasley RP, et al. HBeAg and Anti-HBe Detection by
Radioimmunoassay: Correlation With Vertical Transmission of Hepatitis B Virus
in Taiwan. Journal of Medical Virology 1979;32237-241.

Chang MH, Chen CJ, Lai MS, et al. Universal hepatitis B vaccination in Taiwan
and the incidence of hepatocellular carcinoma in children. Taiwan Childhood
Hepatoma Study Group. N Engl J Med 1997;336:1855-9.

Centers for Disease Control and Prevention. Guidelines for Prevention of
Transmission of Human Immunodeficiency Virus and Hepatitis B Virus to
Health-Care and Public-Safety Workers. MMWR 1989;38(S-6):3-37.

Alter MJ, Hadler SC, Margolis HS, et al. The Changing Epidemiology of
Hepatitis B in the United States — Need for Alternative Vaccination Strategies.
JAMA 1990;263(9):1218-1222.

Margolis HS, Alter MJ, and Hadler SC. Hepatitis B: Evolving Epidemiology and
Implications for Control. Seminars in Liver Disease 1991;11(2):84-92.

Centers for Disease Control and Prevention. Hepatitis B Virus: A Comprehensive
Strategy for Eliminating Transmission in the United States Through Universal
Childhood Vaccination: Recommendations of the Immunization Practices
Advisory Committee (ACIP). MMWR 1991;40(RR-13):1-19.

Shaw FE, Jr., Graham DJ, Guess HA, et al. Postmarketing Surveillance for

Neurologic Adverse Events Reported After Hepatitis B Vaccination. American
Journal of Epidemiology 1988;127:337-52.

67

100. Herroelen L, De Keyser J, Ebinger G. Central-nervous-system
demyelination after immunization with recombinant hepatitis B vaccine. Lancet
1991;338:1174-75.

101. Institute of Medicine. Adverse Events Associated with Childhood
Vaccines: Evidence Bearing on Causality. 1993. Available from: URL:
http://books.napfiedu/catalog/2138.html

102. Fenichel GM. Assessment: Neurologic risk of immunization. Neurology
1999;52:1546-1552.

103. Gout O, Theodorou I, Liblau R, et al. Central nervous system
demyelination after recombinant hepatitis B vaccination: report of 25 cases.
Neurology 1997;48(Suppl 3):A424.

104. Konstantinou D, Paschalis C, Maraziotis T, et al. Two Episodes of
Leukoencephalitis Associated with Recombinant Hepatitis B Vaccination in a
Single Patient. Clinical Infectious Diseases 2001;33:1772-3.

105. Institute of Medicine. Immunization Safety Review: Hepatitis B Vaccine
and Demyelinating Neurological Disorders. 2002. Available from: URL:
http://www.nap.edu/books/O309084695/html/index.html

106. Ball LK, Ball R, and Pratt RD. An assessment of thimerosal use in
childhood vaccines. Pediatrics 2001;107:1147-1154.

107. Gibson S. Memorandum from Assistant Director, Bureau of Biologics,
FDA to Director, Bureau of Biologics, FDA entitled, “Use of thimerosal in
biologics production.” February 27, 1976.

108. Harada M. Minamata disease: methylmercury poisoning in Japan caused
by environmental pollution. Crit Rev Toxicol 1995;25:1-24.

109. Koos BJ, Longo LD. Mercury toxicity in the pregnant woman, fetus, and
newborn infant. Am J Obstet Gynecol 1976;126(3):390-406.

110. Bakir F, Rustam H, Tikriti S et al. Clinical and epidemiological aspects of
methylmercury poisoning. Postgraduate Medical Journal 1980;56:1-10.

l 11. Grandjean P, Weihe P, White RF et al. Cognitive deficit in 7-year-old

children with prenatal exposure to methylmercury. Neurotoxicol Teratol
1997;19(6):417-28.

68

112. Grandjean P, Budtz-Jorgensen E, White RF et al. Methylmercury

exposure biomarkers as indicators of neurotoxicity in children aged 7 years. Am J
Epidemiol 1999;149:301-305.

113. Steuerwald U, Weihe P, Jorgensen P] et a1. Maternal seafood diet,

methylmercury exposure, and neonatal neurologic function. J Pediatr
2000;136:599-605.

114. Stajich GV, Lopez GP, Harry SW et al. Iatrogenic exposure to mercury
after hepatitis B vaccination in preterm infants. J Pediatr 2000;136:679-81.

115. Foldspang A, Hansen JC. Dietary intake of methylmercury as a correlate
of gestational length and birthweight among newborns in Greenland. Am J
Epidemiol 1990;132:310-17.

116. Ramirez GB, Cruz CV, Pagulayan O et al. The Tagum Study 1: analysis
and clinical correlates of mercury in maternal cord blood, breast milk, meconium,
and infants’ hair. Pediatrics 2000;106:774-781.

117. Davidson PW, Myers GJ, Cox C. Effects of prenatal and postnatal
methylmercury exposure from fish consumption on neurodevelopment: outcomes
at 66 months of age in the Seychelles Child Development Study. JAMA
1998;280:701-707.

118. National Research Council. Toxicological Effects of Methylmercury.
Washington, DC: National Academy Press; 2000.

119. Davidson PW, Kost J, Myers GJ. Methylmercury and neurodevelopment:
reanalysis of the Seychelles child development study outcomes at 66 months of
age. JAMA 2001;285:1291-1293.

120. Niu MT, Rhodes P, Salive M. Comparative safety of two recombinant
hepatitis B vaccines in children: data from the Vaccine Adverse Event Reporting
System (VAERS) and Vaccine Safety Datalink (VSD). J Clin Epidemiol
1998;51(6):503-510.

121. Niu MT, Salive ME, Ellenberg SS. Neonatal deaths after hepatitis B
vaccine. Arch Pediatr Adolesc Med 1999;153:1279-1282.

122. Pless R, Risher JF. Mercury, infant neurodevelopment, and vaccination. J
Pediatr 2000;136:571-3.

123. Rothman K] and Greenland S. Causation and Causal Inference. In:
Rothman K] and Greenland S, editors. Modern Epidemiology. Second edition.
Philadelphia, PA: Lippincott, Williams & Wilkins;l998:7-28.

69

 

. .AI_...

124. Immunization Safety Review: Thimerosal Containing Vaccines and
Neurodevelopmental Disorders 2001. Available from: URL:
http://www.nap.edu/catalog/10208.html

125. Blair AM, Clark B, Clarke AJ et al. Tissue concentrations of mercury
after chronic dosing of squirrel monkeys with thimerosal. Toxicology
1975;32171-176.

126. Gunderson VM, Grant KS, Burbacher TM. The effect of low-level
prenatal methylmercury exposure on visual recognition memory in infant crab-
eating macaques. Child Development 1986;57:1076-1083.

127. Gilbert SG, Burbacher TM, Rice DC. Effects of in utero methylmercury
exposure on a spatial delayed alternation task in monkeys. Toxicol Appl
Pharmacol 1993;123:130-136.

128. Freed GL, Andreae MC, Cowan AE, et al. The Process of Public Policy
Formulation: The Case of Thimerosal in Vaccines. Pediatrics 2002;109(6):1153-
1159.

129. Silverrnan NS, Darby MJ, Ronkin SL, Wapner RJ. Hepatitis B prevalence
in an unregistered prenatal population: implications for neonatal therapy. JAMA
1991; 266:2852-5.

130. Pierce RL, Smith S, Rowe-West B, et a1. Hepatitis B maternal screening,
infant vaccination, and infant prophylaxis practices in North Carolina. Arch
Pediatr Adolesc Med 1999;153(6):619-623.

131. Xu ZY, Liu CB, Francis DP, et al. Prevention of perinatal acquisition of
hepatitis B virus carriage using vaccine: Preliminary report of a randomized,
double-blind placebo-controlled and comparative trial. Pediatrics 1985;76:713-
718.

132. Beasley RP, Hwang L-Y. Epidemiology of hepatocellular carcinoma. In:
Vyas GN, Dienstag JL, Hoofnagle JH, editors. Viral hepatitis and liver disease.
New York: Grune & Stratton, 1984:209-24.

133. Centers for Disease Control and Prevention. Notice to Readers:
Thimerosal in Vaccines: A Joint Statement of the American Academy of

Pediatrics and the Public Health Service. MMWR 1999; 48(26):563-565.

134. American Academy of Pediatrics. PedCommzAAP Member Alert. July 15,
1999.

70

135. Centers for Disease Control and Prevention. Notice to Readers:
Availability of Hepatitis B Vaccine That Does Not Contain Thimerosal as a
Preservative. MMWR l999;48(35):780-782.

136. Hurie MB, Saari TN, Davis JP. Impact of the Joint Statement by the
American Academy of Pediatrics/US Public Health Service on Thimerosal in
Vaccines on Hospital Infant Hepatitis B Vaccination Practices. Pediatrics 2001;

107:755-758.

137. Oram RJ, Daum RS, Seal JB, Lauderdale DS. Impact of recommendations
to suspend the birth dose of hepatitis B virus vaccine. JAMA 2001;285:1874-9.

138. Brayden RM, Pearson KA, Jones JS, Renfrew BL, Berman 8. Effect of
thimerosal recommendations on hospitals’ neonatal hepatitis B vaccination
policies. J Pediatr 2001; 1382752-5.

139. Centers for Disease Control and Prevention. Impact of the 1999
AAP/USPHA joint statement on thimerosal in vaccines on infant hepatitis B
vaccination practices. MMWR 2001;50:94-7.

140. Clark SJ, Cabana M, Malik T, Hussain Y, Freed GL. Hepatitis B
vaccination practices in hospital newborn nurseries before and after changes in
vaccination recommendations. Arch Pediatr Adolesc Med 2001;155:915-20.

141. Michigan Department of Community Health. Three-month-old baby dies
of acute hepatitis B. Michigan Immunization Update 2000; 7: 1-2.

142. Armstrong GL, Mast EE, Wojczynski M, et al. Childhood Hepatitis B
Virus Infections in the United States Before Hepatitis B Immunization. Pediatrics
2001;108(5):]123-28.

143. Bimbaum JM and Bromberg K. Evaluation of prophylaxis against
hepatitis B in a large municipal hospital. Am J Infect Control 1992;20:172-6.

144. Berlin M. Mercury. In: Friberg L, Nordberg GF, Vouk VB, editors.
Handbook on the Toxicology of Metals. Second Edition. Amsterdam: Elsevier,
1990:387-435.

145. Clarkson TW. The Three Modern Faces of Mercury. Environmental
Health Perspectives 2002;110(3uppl 1): 11-23.

146. Committee on Infectious Diseases. Universal Hepatitis B Immunization.
Pediatrics 1992;89(4):795-799.

147. Michigan Department of Community Health. More Hospitals Giving the

Birth Dose of Hepatitis B Vaccine to All Their Newborns. Local Liaison Report
to the Local Health Departments 2002;118.

71

 

 

148.

149.

Chen RT and DeStefano F. Vaccine adverse events: causal or
coincidental? Lancet 1998;351:611-612.

Making Vaccination Fashionable. JAMA 1902;38:1444-1445, 1456.

72