IMPACT OF PROTOCOL CHANGES IN MICHIGAN NEWBORN SCREENING FOR
CONGENITAL HYPOTHYROIDISM ON SCREENING PERFORMANCE METRICS
By
Steven J. Korzeniewski

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
EPIDEMIOLOGY
2011

ABSTRACT
IMPACT OF PROTOCOL CHANGES IN MICHIGAN NEWBORN SCREENING FOR
CONGENITAL HYPOTHYROIDISM ON SCREENING PERFORMANCE METRICS
By
Steven J. Korzeniewski
Objective: This study was conducted to evaluate changes in newborn screening for congenital
hypothyroidism operations in Michigan.
Study Design & Participants: This population-based retrospective cohort analysis includes Michigan
resident infants born and screened in Michigan from 1/1/1994 to 6/30/2010. Newborn screening
laboratory and follow-up data managed by the Michigan Department of Community Health are used
in this study.
Methods: The primary exposure is the method of dried blood spot testing: 1) thyroxine (T4) with
backup thyrotropin (TSH) testing, 2) tandem T4 and TSH testing; 3) primary TSH testing without
serial testing; and 4) primary TSH testing plus serial testing for births weighing <1,800g. Outcomes
of interest include screening performance metrics: detection rate, positive predictive value, false
positive rate, sensitivity, and specificity. The proportion of cases later determined to have transient
disease is also investigated among those detected since 10/1/2003 that were followed-up after age
three years and underwent diagnostic re-evaluation. Logistic regression analysis is used to investigate:
a) whether the detection and false positive rates changed significantly as dried blood spot testing
methods changed, and b) to investigate predictors of transient CH among cases followed-up and reevaluated after age three years.
Results: T4 testing is as likely to detect CH as is primary TSH testing without serial testing. Primary
TSH testing yields fewer false positive determinations. Primary TSH testing yields fewer false
positive determinations. The addition of serial testing among infants born weighing < 1,800g to the

primary TSH testing protocol significantly increased the detection rate; in this period, 14% of
detected cases were identified by retest and three of five cases weighing <1,800g were detected by
retest. Primary TSH screening is more susceptible to false negative screening determinations and is
incapable of detecting the 1-3 cases of central hypothyroidism expected annually in Michigan.
Tandem T4 and TSH testing increased the detection rate by 80% and 20% relative to primary T4
backup TSH and primary TSH plus serial testing approaches respectively; however, this approach
increased the number of false positives by eight fold relative to primary TSH plus serial testing. At
three year follow-up, one of four cases was no longer receiving thyroid hormone medication. One of
five cases re-evaluated had stopped treatment without medical supervision.
Discussion: Our results indicate that newborn screening programs should consider tandem T4 and
TSH testing or primary TSH plus serial testing for infants at risk of later rising TSH for detecting
congenital hypothyroidism; however, a benchmark balance between sensitivity and specificity is
necessary to determine the optimal testing strategy. Further efforts are necessary to standardize the
process of diagnosis and terminologies used to characterize cases across newborn screening
programs to make comparisons between and within programs over time more meaningful;
guidelines for newborn screening, pediatricians, and pediatric endocrinologists are necessary to
support this effort. Our findings also indicate that long term follow-up should be conducted for all
CH cases, not only those eligible for a trial off thyroid hormone supplements, to ensure treatment
compliance in hopes of avoiding potential brain damage among cases. As diagnostics and the
nomenclature of congenital hypothyroidism are standardized across newborn screening programs,
cost-benefit analyses are necessary to support recommended program operations.

DEDICATION

This dissertation is dedicated to my lovely wife, Stephanie, for supporting me through many
sleepless nights endured in finishing this work, teaching me every day the true meaning and feeling
of love, and immeasurably brightening my existence. I also dedicate this work to my children,
Akasha and Cassious, as well as our new baby due to arrive in September, for the endless inspiration
they provide. Special thanks also goes to Jack and Lori Moore, my parents-in-law, for their interest
in my work, the support they have given, the model they provide as parents, and most importantly
the beautiful family they have raised, particularly their amazing daughter, the love of my life,
Stephanie. Thanks also to my parents, Kathy Potter, Kim Korzeniewski and his wife also named
Kim.

iv

ACKNOWLEDGEMENT
I would like to thank my dissertation guidance committee members, Dr. Violanda
Grigorescu, Dr. David Todem, Dr. Gretchen Birbeck and Dr. Nigel Paneth for their support and
guidance in preparing this dissertation.
Dr. Grigorescu played a critical role in this dissertation and in my overall career. She
introduced me to the interesting topic of newborn screening when she hired me as the newborn
screening epidemiologist in 2007. Dr. Grigorescu facilitated access to the data used in this
dissertation. She also supported my promotion from epidemiologist to manager of the Maternal and
Child Health Epidemiology Section of the Michigan Department of Community Health which gave
me the experience necessary to further advance in my career. Dr. Grigorescu’s passion for public
health continues to serve as an inspiration and I thank her for her support of my work and for the
education and friendship she has provided me over the years.
I would like to thank Dr. David Todem for taking the time to discuss statistical methods and
for his diligence and accommodation in answering my many questions in preparing this dissertation.
Thanks also to Dr. Gretchen Birbeck for her willingness to participate, even while working in Africa,
and for her always prompt and diligent review of my work.
I would like to give special acknowledge to the many years of guidance and support provided
by my primary advisor and mentor, Dr. Nigel Paneth. His willingness to recognize my curiosity and
interest as academic potential absolutely changed my life. The fact that someone of his stature
appreciated my work and thoughts gave me the confidence to pursue post-graduate education and a
career in epidemiology. Dr. Paneth’s support extended far beyond the confines of an academic
advisory relationship. While working four jobs, pursuing two master degrees, and virtually raising
two children alone as I endured significant personal turmoil, Dr. Paneth was always there to offer a
kind word, advice about important life decisions, and bits of knowledge gleaned from his own life.
v

He even helped me financially to avoid dropping out of school when I had fallen short of a
semester’s tuition several years ago. It is truly difficult to express how important his role has been
and how honored and immeasurably grateful I am to have him as a mentor, role model and friend.
I would also like to thank Dr. Bill Young for his engagement in thinking about newborn
screening and opportunities for improvement. Dr. Yong’s cutting edge thoughts about newborn
screening, his tremendous experience and his interest in science were tremendously important in my
learning about congenital hypothyroidism and newborn screening. Work initiated with Dr. Yong led
me to pursue this dissertation. I thank Dr. Yong for his humble manner in always helping me to
understand how newborn screening came about and where it should go in terms of operational
changes. I also thank Dr. Yong for his willingness to serve as an ‘informal dissertation committee
member’.
Special thanks also go to Karen Andruszewski and Mary Kleyn of the Michigan Department
of Community Health’s Newborn Screening Follow-up Program for their assistance in providing
data and helping me to understand the nuances therein. Their help in completing this dissertation
was absolutely invaluable and appreciated.

vi

TABLE OF CONTENTS
LIST OF TABLES ........................................................................................................................................... ix
LIST OF FIGURES ......................................................................................................................................... xi
LIST OF ACCRONYMS ............................................................................................................................... xii
KEY DEFINITIONS ................................................................................................................................... xiii
CHAPTER 1: BACKGROUND .................................................................................................................... 1
Introduction: ....................................................................................................................................................... 1
Definition & Physiology of Congenital Hypothyroidism: ........................................................................... 1
Etiology of Congenital Hypothyroidism: ....................................................................................................... 5
Diagnosis of Congenital Hypothyroidism: ..................................................................................................... 6
Outcomes of Congenital Hypothyroidism: .................................................................................................... 9
Treatment of Congenital Hypothyroidism: .................................................................................................... 9
Epidemiology of Congenital Hypothyroidism: ...........................................................................................10
Controversies in Newborn Screening for Congenital Hypothyroidism: .................................................14
Dried blood spot testing for Congenital Hypothyroidism: ...................................................................15
Treatment for Congenital Hypothyroidism .............................................................................................20
Long Term Follow-Up and Confirmation of Disease Permanence.....................................................24
CHAPTER 2 METHODS .............................................................................................................................26
Study Design & Participants: .........................................................................................................................26
Research Questions & Hypotheses: ..............................................................................................................27
Newborn Screening for Congenital Hypothyroidism in Michigan:..........................................................29
Data: ...................................................................................................................................................................32
Analysis: .............................................................................................................................................................35
Full Cohort ...................................................................................................................................................35
Sub-Cohort 1: Infants Born Weighing < 1,800g (1/1/2004-6/30/2010) ..........................................35
Sub-Cohort 2: Cases Detected Since 10/1/2003 & Followed-up After Age 3 Years .......................36

vii

CHAPTER 3: COMPARISON OF DRIED BLOOD SPOT TESTING PROTOCOLS IN
MICHIGAN NEWBORN SCREENING FOR CONGENITAL HYPOTHYROIDISM,
1/1/1994-6/30/2010 ......................................................................................................................................38
Results:...............................................................................................................................................................38
Limitations: .......................................................................................................................................................41
Discussion: ........................................................................................................................................................42
CHAPTER 4: IMPACT OF SERIAL TESTING ON CONGENITAL HYPOTHYROIDISM
NEWBORN SCREENING PERFORMANCE METRICS AMONG MICHIGAN RESIDENT
BIRTHS WEIGHING <1,800g, 1/1/2004-6/30/2010 ...........................................................................52
Results:...............................................................................................................................................................52
Limitations: .......................................................................................................................................................53
Discussion: ........................................................................................................................................................54
CHAPTER 5: RESULTS OF FOLLOW-UP OF CONGENITAL HYPOTHYROIDISM CASES
DETECTED BY MICHIGAN NEWBORN SCREENING AFTER AGE 3 YEARS ....................61
Results:...............................................................................................................................................................61
Limitations: .......................................................................................................................................................64
Discussion: ........................................................................................................................................................65
Chapter 6: Conclusions and Future Directions ...........................................................................................73
APPENDIX- APPROVALS FOR REPRINTS .........................................................................................79
BIBLIOGRAPHY ...........................................................................................................................................84

viii

LIST OF TABLES

TABLE 1: RADIONUCLIDE IMAGE AND ULTRASONOGRAPHY FINDINGS IN CONGENITAL
HYPOTHYROIDISM (RASTOGI & LAFRANCHI, 2010) ............................................................................ 8
TABLE 2: GEOGRAPHIC VARIATION IN BIRTH PREVALENCE OF CONGENITAL HYPOTHYROIDISM ...10
TABLE 3: CONGENITAL HYPOTHYROIDISM DETECTION BY SELECTED POPULATION SEGMENT, NEW
YORK STATE NEWBORN SCREENING, 2000-2003 (HARRIS & PASS, 2007)......................................14
TABLE 4: BIRTH PREVALENCE OF ATYPICAL HYPOTHYROIDISM (LOW T4 & LATE RISING TSH) BY
BIRTH WEIGHT, MASSACHUSETTS NEWBORN SCREENING, 1993-1996, (MANDEL ET AL., 2000)
....................................................................................................................................................................18
TABLE 5: SCREENING PERFORMANCE METRICS, UNITED STATES, 2006 (PASS & NETO, 2009) ..........21
TABLE 6: DRIED BLOOD SPOT TESTING PROTOCOLS, CUTOFF VALUES AND ASSOCIATED
DETERMINATIONS APPLIED BY MICHIGAN NEWBORN SCREENING FOR CONGENITAL
HYPOTHYROIDISM, 1977-2010 ..............................................................................................................37
TABLE 7: NEWBORNS SCREENED BY SELECTED DEMOGRAPHIC & PERINATAL CHARACTERISTICS &
BY DRIED BLOOD SPOT TESTING METHOD, MICHIGAN NEWBORN SCREENING, 1/1/19946/30/2010 ................................................................................................................................................46
TABLE 8: NEWBORNS SCREENED POSITIVE BY SELECTED DEMOGRAPHIC & PERINATAL
CHARACTERISTICS & BY DRIED BLOOD SPOT TESTING METHOD, MICHIGAN NEWBORN
SCREENING, 1/1/1994-6/30/2010.......................................................................................................47
TABLE 9: CONGENITAL HYPOTHYROIDISM CASES DETECTED BY SELECTED DEMOGRAPHIC &
PERINATAL CHARACTERISTICS & BY DRIED BLOOD SPOT TESTING METHOD, MICHIGAN
NEWBORN SCREENING, 1/1/1994-6/30/2010...................................................................................48
TABLE 10: CONGENITAL HYPOTHYROIDISM CASES DETECTED BY NEWBORN SCREENING RESULT &
BY DRIED BLOOD SPOT TESTING METHOD, MICHIGAN NEWBORN SCREENING, 1/1/19946/30/2010 ................................................................................................................................................49
TABLE 11: NEWBORN SCREENING RESULTS & PERFORMANCE METRICS, MICHIGAN, 1/1/19946/30/2010 ................................................................................................................................................50
TABLE 12: LIKELIHOOD OF CONGENITAL HYPOTHYROIDISM DETECTION AND FALSE POSITIVE
SCREENING DETERMINATION BY DRIED BLOOD SPOT TESTING PROTOCOL, MICHIGAN
NEWBORN SCREENING, 1/1/1994-6/30/2010...................................................................................51
ix

TABLE 13: INFANTS SCREENED, SCREENED POSITIVE & DIAGNOSED BY SELECTED DEMOGRAPHIC
CHARACTERISTICS AMONG CHILDREN BORN WEIGHING <1,800G BEFORE AND AFTER
IMPLEMENTATION OF SERIAL TESTING, 1/1/2004-6/30/2010 .......................................................56
TABLE 14 (TABLE 14 CONTINUED): INFANTS SCREENED, SCREENED POSITIVE & DIAGNOSED BY
SELECTED DEMOGRAPHIC CHARACTERISTICS AMONG CHILDREN BORN WEIGHING <1,800G
BEFORE AND AFTER IMPLEMENTATION OF SERIAL TESTING, 1/1/2004-6/30/2010..................57
TABLE 15: NEWBORN SCREENING PERFORMANCE METRICS BEFORE & AFTER IMPLEMENTATION OF
SERIAL TESTING AMONG INFANTS BORN WEIGHING <1,800G, MICHIGAN, 1/1/20046/30/2010 ................................................................................................................................................58
TABLE 16: CHARACTERISTICS OF CONGENITAL HYPOTHYROIDISM CASES DETECTED BY INITIAL VS.
RETEST DRIED BLOOD SPOT SCREEN, MICHIGAN NEWBORN SCREENING, 1/1/20046/30/2010 ................................................................................................................................................59
TABLE 17: LIKELIHOOD OF OVERALL CONGENITAL HYPOTHYROIDISM DETECTION, DETECTION BY
RETEST & FALSE POSITIVE DETERMINATION FOLLOWING IMPLEMENTATION OF SERIAL
TESTING, MICHIGAN NEWBORN SCREENING, 1/1/2004-6/30/2010.............................................60
TABLE 18: PROPORTION OF CONGENITAL HYPOTHYROIDISM CASES DETECTED BY MICHIGAN
NEWBORN SCREENING FROM 10/1/2003-12/31/2007 FOLLOWED-UP AT OR BEYOND AGE
THREE YEARS BY SELECTED DEMOGRAPHIC/PERINATAL CHARACTERISTICS .............................67
TABLE 19: PROPORTION OF CONGENITAL HYPOTHYROIDISM CASES DETECTED BY MICHIGAN
NEWBORN SCREENING FROM 2004-2007 FOLLOWED-UP AT OR BEYOND AGE THREE YEARS BY
NEWBORN SCREENING RESULT ............................................................................................................68
TABLE 20: METHOD OF TRIAL OFF THYROID HORMONE MEDICATION OR REASON FOR LACK OF
TRIAL, CONGENITAL HYPOTHYROIDISM CASES FOLLOWED-UP AFTER AGE 3 YEARS, MICHIGAN
NEWBORN SCREENING ...........................................................................................................................69
TABLE 21: RESULTS OF TRIAL OFF THYROID HORMONE MEDICATION AMONG CONGENITAL
HYPOTHYROIDISM CASES FOLLOWED-UP AFTER AGE 3 YEARS BY SELECTED DEMOGRAPHIC &
PERINATAL CHARACTERISTICS, MICHIGAN NEWBORN SCREENING ...............................................70
TABLE 22: RESULTS OF TRIAL OFF THYROID HORMONE MEDICATION AMONG CONGENITAL
HYPOTHYROIDISM CASES FOLLOWED-UP AFTER AGE 3 YEARS BY NEWBORN SCREENING &
CONFIRMATORY TESTING RESULTS, MICHIGAN NEWBORN SCREENING ......................................71
TABLE 23: PREDICTORS OF TRANSIENT DETERMINATION AMONG CHILDREN DIAGNOSED WITH
CONGENITAL HYPOTHYROIDISM AND FOLLOWED-UP BY MICHIGAN NEWBORN
SCREENING AFTER AGE 3 YEARS, 2004-2007 ......................................................................................72
x

LIST OF FIGURES

FIGURE 1: HYPOTHALAMIC-PITUITARY-AXIS ................................................................................................ 2
FIGURE 2: INCIDENCE OF CONGENITAL HYPOTHYROIDISM IN THE UNITED STATES, 1987-2002
(HARRIS ET AL., 2007) ) [FOR INTERPRETATION OF THE REFERENCES TO COLOR IN THIS AND
ALL OTHER FIGURES, THE READER IS REFERRED TO THE ELECTRONIC VERSION OF THIS
DISSERTATION] .........................................................................................................................................11
FIGURE 3: DATA SOURCES USED TO POPULATE THE FINAL ANALYTIC FILE ........................................33
FIGURE 4: THREE YEAR FOLLOW-UP STUDY FLOW DIAGRAM .................................................................62

xi

LIST OF ACCRONYMS
AAP
American Academy of Pediatrics
CH
Congenital Hypothyroidism
CI
Confidence Interval
CLSI
Clinical and Laboratory Standards Institute
DIT
Diiodotyrosine
DOB
Date of Birth
FGR
Fetal Growth Ratio
FN
False Negative
FP
False Positive
g
Grams
HPT
Hypothalamic-Pituitary-Thyroid
IQ
Intelligence Quotient
LBW
Low Birth Weight
LIMS
Laboratory Information Management System
Lost to Follow-up LTFU
MCIR
Michigan Care Improvement Registry
MDCH
Michigan Department of Community Health
MIT
Monoiodotyrosine
NBS
Newborn Screening
NBW
Normal Birth Weight
NICU
Neonatal Intensive Care Unit
OR
Odds Ratio
PEAC
Pediatric Endocrinology Advisory Committee
SCBU
Sick Child Birthing Unit
Se
Sensitivity
SGA
Small for Gestational Age
Sp
Specificity
T3
Triiodothyronine
T4
Thyroxine
TBG
Thyroxine-Binding Globulin
TN
True Negative
TP
True Positive
TPN
Total Parenteral Nutrition
TRH
Thyrotropin Releasing Hormone
TSH
Thyroid Stimulating Hormone
US
United States
VLBW
Very Low Birth Weight

xii

Central Hypothyroidism
Classic or Primary
Hypothyroidism
Confirmed Transient CH
rate
Congenital Hypothyroidism

Detection Rate
False Positive Rate (FPR)
Hyperthyrotropinemia
Hypothyroxinemia of
Prematurity
Lost to follow-up
Positive Predictive Value
(PPV)
Proportion of False Positive
Screens (PFP)
Rate of self-cessation of
therapy
Rate of trial off therapy
Sensitivity (Se)
Specificity (Sp)
Suspected Transient CH rate
Transient Hypothyroidism

KEY DEFINITIONS
characterized by persistently low T4 concentrations and low or
normal TSH concentrations, often associated with other pituitary
hormone deficiencies
characterized by persistently low T4 concentrations and elevated
TSH concentrations
Proportion of cases exhibiting normal thyroid function[Number
exhibiting normal thyroid function/total number followed-up]
an umbrella term referring to congenital thyroid disorders often
characterized by pathologically low concentrations of T4 that may or
may not be accompanied by elevated concentrations of thyroid
stimulating hormone (TSH), otherwise known as thyrotropin.
Number of screens performed to yield 1 detected case [Total
Number Screened/Total Number of Cases]
Number of false positives divided by number absent disease
[FP/(FP+TN), or 1-Specificity]
characterized by normal T4 concentrations and persistently elevated
TSH concentrations, often referred to as asymptomatic
hypothyroidism
characterized by low T4 concentrations and normal TSH
concentrations, not associated with central hypothyroidism,
common among premature infants, commonly resolves with
maturity
the proportion of children not followed-up among those eligible for
follow-up[Number not Followed-up/Total Eligible for Follow-up]
Proportion of infants with a positive screen that are confirmed as
having CH[TP/(TP+FP)]
Proportion of screens having false positive determinations[FP/Total
Screens]
Proportion of cases that stopped therapy of their own
accord[Number stopped therapy/total number followed-up]
Proportion of cases eligible for a trial off therapy that have
undergone such a trial[Number trialed/Number followed-up eligible
for trial ]
Proportion with disease that are correctly screened
positive[TP/(TP+FN)]
Proportion absent disease that are correctly screened
negative[TN/(FP+TN)]
Proportion of cases followed that are suspected to be
transient[Number Untreated at Follow-up/total number followedup]
characterized by having abnormal screening values (low T4 and
elevated TSH) but normal serum T4 and TSH concentrations on a
subsequent serum tests at 1 to 2 months of age or later among
premature infants
xiii

CHAPTER 1: BACKGROUND

Introduction:
Newborn screening (NBS) for congenital hypothyroidism (CH) began in the mid 1970s
following the development of a radioimmunoassay capable of measuring thyroxine (T4) in dried
blood spotted on filter paper.[1-5] Based on findings from the first million infants screened, the
NBS Committee of the American Thyroid Association recommended broad establishment and
expansion of NBS programs for CH In 1977.[6] By 1992, it was estimated that 50 million infants
were screened annually for CH worldwide.[7] While NBS for CH has been implemented in virtually
all industrialized and in many developing nations over the past several decades, there has been a lack
of emphasis on assessing and evaluating screening system components.[8] Accordingly, debates
persist about the ideal CH screening algorithm, how best to screen for the disease in premature and
sick newborns, and which children need particular attention and long term treatment based on the
screening results. This study focuses on Michigan NBS for CH and investigates the impact of
changes in program operations over time including alterations to dried blood spot testing protocols
comprising four distinct screening periods and initiation of long term follow-up of detected cases.
Definition & Physiology of Congenital Hypothyroidism:
CH is often used as an umbrella term referring to congenital thyroid disorders often
characterized by pathologically low concentrations of T4 that may or may not be accompanied by
elevated concentrations of thyroid stimulating hormone (TSH), otherwise known as thyrotropin.
The primary function of the thyroid gland is to combine iodine with the amino acid tyrosine to
manufacture thyroid hormones, T4 and triiodothyronine (T3). Iodide is taken into the thyroid
follicular cells by an active transport system and then oxidized to iodine by thyroid peroxidase. Next,
organification occurs when iodine is attached to tyrosine molecules attached to thyroglobulin,
1

forming monoiodotyrosine (MIT) and diiodotyrosine (DIT). The coupling of 2 molecules of DIT
forms T4. The coupling of one molecule of MIT and one molecule of DIT forms T3. T3 is the
metabolically active hormone that is produced primarily by cells when T4 is deiodinated by two
deiodinase enzymes in the thyroid gland. While the thyroid gland produces all T4, approximately
80% of T3 is produced through conversion of T4 in various tissues of the body when more T3 is
needed. Both T3 and T4 bind to thyroxine-binding globulin (TBG) for transport in the blood.
Thyroglobulin, with T4 and T3 attached, is stored in the follicular lumen of the gland. TSH activates
the enzymes needed to cleave T4 and T3 from thyroglobulin and also stimulates T4 production..
Thyroid hormones act on processes such as neuronal migration and differentiation,
myelination and synaptogenesis which are essential for proper neurodevelopment and are also
involved in maintenance of normal
physiological function including bone
maturation.[9, 10] Thyroid hormone
concentrations are regulated by the
hypothalamic-pituitary-thyroid (HPT) axis
as portrayed in Figure 1.[10] The pituitary
gland stimulates the thyroid to produce
hormones by releasing TSH. The
hypothalamus regulates the pituitary gland
in response to low thyroid hormone
concentrations (negative feedback) by stimulating

Figure 1: Hypothalamic-Pituitary-Axis

production of thyrotropin releasing hormone

it via

(TRH) to produce TSH. Thus, when the levels of T4 and T3 are low, the hypothalamus secretes
TRH which in turn stimulates the pituitary gland to release TSH which then stimulates the thyroid
2

gland to increase secretion of T4.[11] TSH secretion is inhibited by dopamine, somatasotatin and
high doses of corticosteroids.

The fetal thyroid is capable of producing thyroid hormone around 10-12 weeks of gestation
and blood levels are thought to reach term concentrations by 18-20 weeks of gestation. Prior to 1820 weeks of gestation, the fetus is dependent on transplacental passage of thyroid hormones;
afterwards, in non-hypothyroid fetuses, the fetal HPT axis is believed to function independently of
the maternal HPT axis.[12] Hypothyroid fetuses rely on transplacental passage of thyroid hormones
throughout gestation.
At birth, full term (gestation > 37 weeks) neonates experience a surge of TSH reaching
greater than 50 mU/L during the first 24 hours of life as a physiological response to a cold
environment followed by a decrease to below 10 mU/L during the first few postnatal weeks.[13]
However, many preterm infants (gestation < 37 weeks) and/or sick newborns exhibit weak or no
TSH surge at birth due to immaturity of the HPT axis.[14-19] Van Wassenaer et al. examined
thyroid hormone concentrations during the first eight weeks of life among 100 infants born at less
than 30 weeks gestation and reported that TSH increases from day three onward and from day
seven onward in sick and healthy infants respectively and stabilizes after the 28th day life in both
groups.[18] Estimates of the frequency of late rising TSH (occurring beyond the first 48 hours of
life) range from 1:18,000 to 1:67,226 live births, 1:202 to 1:400 among very premature (gestation <
32 weeks) and/or very low birth weight infants (birth weight < 1,500g).[14, 16, 20-22]
Pathologically low concentrations of thyroid hormone are often associated with underlying
structural abnormalities. In iodine sufficient nations, an estimated 85% of CH cases is characterized
by thyroid dysgenesis (abnormality in thyroid gland development), the remaining 10%-15% are
characterized by dyshormonogenesis (defects in the synthesis of T4) or, in rare cases, by defects in
3

peripheral thyroid hormone transport, metabolism or action.[23] Thyroid dysgenesis involves
ectopia (abnormal formation of thyroid tissue) or aplasia/hypoplasia (full or partial absence of the
thyroid gland). Defects involving thyroid hormone transport, metabolism or action most often
involve hypopituitarism, the loss of function in an endocrine gland due to failure of the pituitary
gland to secrete hormones which stimulate that gland's function; this is sometimes referred to as a
‘peripheral’ defect in that the problem originates outside of the thyroid.
CH is often sub-classified by distribution of thyroid hormones. Infants with persistently low
T4 concentrations and elevated TSH concentrations are referred to as primary or classic CH. Central
hypothyroidism refers to infants with persistently low T4 concentrations and low or normal TSH
concentrations. Central CH is most often associated with other pituitary hormone deficiencies,
particularly in the presence of features such as hypoglycemia or micropenis and undescended testes
in males which are associated with growth hormone, ACTH and gonadotropin deficiencies.[24]
When T4 concentrations are normal and TSH concentrations are persistently elevated, the disease is
referred to as hyperthyrotropinemia or asymptomatic hypothyroidism.[25]
CH is classically defined as a permanent condition requiring lifelong treatment, although
transient forms of the disease have also been identified. Yet they remain ill-defined, as recently
reviewed by Parks et al.[26] The American Academy of Pediatrics (AAP) defines transient
hypothyroidism as abnormal screening values (low T4 and elevated TSH) but normal serum T4 and
TSH concentrations on a subsequent serum tests,, although NBS programs apply broad and varying
definitions, and many fail to differentiate suspected transient from permanent types of CH.[27]
When T4 concentrations are low, TSH concentrations are normal, and central hypothyroidism has
been ruled out, the disease is referred to as hypothyroxinemia of prematurity, as it is commonly
found among preterm infants and usually resolves with maturity.[28] Transient hypothyroxinemia

4

has also been reported in term infants and in rare cases the disorder may persist until approximately
10 years of age.[27]
Etiology of Congenital Hypothyroidism:
Globally, the most common cause of thyroid hormone insufficiency during gestation is lack
of dietary iodine, referred to as endemic cretinism.[29] Unlike CH, brain damage associated with
endemic cretinism likely occurs in utero due to maternal thyroid hormone deficiencies. Iodization of
table salt has largely addressed the issue of iodine insufficiency among most developed nations,
although recent data suggest dietary iodine intake has decreased in several countries including the US
in recent years.[30] While many authors refer to underlying morphologies as etiologies of CH,
structural defects lie in the causal pathway of the disease and are more appropriately referred to as
mediators. The underlying causes of thyroid dysgenesis and dyshormonogenesis associated with
primary permanent CH remain largely unknown; although an estimated 10%-20% of identified cases
exhibit genetic etiologic factors including mutation in the TTF-2 gene and mutations in genes
encoding transcription factors important in thyroid gland development.[24, 31] The etiology of
central CH is similarly unknown, although on rare occasions, isolated genetic defects including
mutations in the TSH β subunit or TRH receptor genes have been identified as etiologic factors.[24]
Better recognized are causes of transient CH including transplacental passage of thyrotropin
receptor blocking antibodies or antithyroid drugs used to treat maternal hyperthyroidism, iodine
deficiency, and iodine excess; on rare occasions transient dyshormonogenesis can also be caused by
DUOX2 mutations.[26, 32] Hypothyroxinemia is thought to be caused by prematurity or delayed
maturation of the HPT axis, increased TSH response to thyroid releasing hormone, presence of
antithyroid antibodies or thyroperoxidase (an enzyme involved in the production of T4 or T3), or
thyrotropin receptor gene sequence variations.[33, 34] Dopamine infusion and environmental

5

chemicals including certain polyhalogenated aromatic hydrocarbons have also been identified as
potential causes of transient CH, although the role of environmental chemicals is less clear.[35-43]
Diagnosis of Congenital Hypothyroidism:
Globally, most CH cases are detected clinically, since only an estimated 25%-33% of the
overall birth population is screened for the disease due to the lack of NBS programs in many
developing nations.[24] Infants with CH usually appear unaffected at birth. Clinical signs associated
with CH include gestation beyond 42 weeks, macrosomia, neonatal hyperbilirubinemia lasting more
than three weeks, lethargy, feeding difficulties, constipation, umbilical hernia, macroglossia (large or
muscular appearing tongue), and cold or mottled skin.[7, 44, 45] CH is also associated with other
congenital anomalies including trisomy-21.[46-52] However, clinical signs of CH are not specific to
the disease meaning that diagnosis absent NBS is often delayed beyond the first several weeks or
months of life. Most industrialized nations use some form of NBS to detect and later diagnose CH.
NBS for CH began in Quebec, Canada in response to reports showing that while children
treated within three months of birth developed normally intellectually, only one third of newborns
were diagnosed during that period.[53-55] Dussault and Laberge developed a radioimmunoassay for
measuring T4 from the eluate of filter paper blood spots in 1971 and early 1972, although initial
reports of this work were not received well by the academic community.[56] The first report of their
work was published in French in 1973, and in 1974 screening for CH was incorporated with NBS
for phenylketonuria (PKU) and tyrosinemia in Quebec in 1974.[57] Dussault et al. soon developed
radioimmunoassays to measure TSH and TBG from filter paper blood to reduce the number of false
positive results.[58, 59] Initially, most NBS programs in North America, Australia, New Zealand and
Israel and some in Europe conducted primary T4 and backup TSH testing among infants having T4
concentrations less than the 10th centile.[5, 60] Most European and Japanese NBS programs

6

employed a primary TSH test strategy.[61-63] Most NBS make referrals for confirmatory testing
based on TSH concentrations relative to age adjusted cutoffs.
While infants with significantly elevated dried blood spot TSH concentrations are expected
to have permanent primary CH, confirmatory testing is usually based on serum tests of venipuncture
blood samples combined with some measure of binding proteins (i.e., T3 resin uptake) used to
differentiate free (active) from total T4.[24, 64] Blood samples for confirmatory testing are ideally
obtained around 2-3 weeks of life when the upper range of TSH falls to approximately 10 mU/L.
Reference ranges for free T4, total T4 and TSH concentrations measured in serum at 2-4 weeks of
life are approximately 10-26 pmol/L, 90-206 nmol/L and less than 10 mU/L respectively.[22]
Infants having two or more serum TSH concentrations >20 mU/L are expected to have permanent
primary CH.[13] If a defect in thyroid hormone synthesis is suspected, perchlorate washout testing
is sometimes performed to test the ability of the thyroid to transform iodine into organically bound
iodine.[65] Other tests including scintigraphy and ultrasound are also useful during the process of
diagnosing CH.
Evaluation of the thyroid gland via ultrasound imaging and radioisotope scanning is useful in
identifying underlying morphologies and differentiating transient from permanent disease. Normal
sized and anatomically positioned (eutopic) thyroid glands are more likely among transient
conditions.[26] Newborns with permanent CH are most likely to have thyroid gland agenesis, an
ectopic gland located along the path of embryonic descent of the thyroid, or a hypoplastic gland,
although eutopic thyroid glands are also found among permanent cases of dyshormonogenesis.[26,
66] Isotope scanning is better able to detect ectopic tissue, whereas ultrasound is better able to
portray abnormalities of thyroid volume and morphology including thyroid aplasia.[67]
Ultrasonography can also identify large thyroid glands, suggestive of dyshormonogenesis.[24] Chang
et al. recently compared ultrasound and scintigraphic findings and reported a significant correlation
7

(p<.001) between measures of thyroid volume, concluding that use of both scintigraphy and
ultrasound is preferable to either test individually. Results of thyroid imaging findings in CH are
reported in Table 1.
Table 1: Radionuclide Image and Ultrasonography Findings in Congenital Hypothyroidism
(Rastogi & LaFranchi, 2010)
Defect
Radionuclide image
Ultrasonography
Aplasia
Hypoplasia
Ectopia
TSH β mutations
TSH receptor inactivating mutation
Trapping error
Beyond trapping error
Maternal TRB-Ab

No uptake
↓uptake
↓ uptake, ectopic
No uptake
↓ uptake
↓ or no uptake
↑ uptake
↓ or no uptake

Absent gland
Small, eutopic
Ectopic gland (hypoplastic)
Eutopic gland (hypoplastic)
Eutopic gland
Eutopic gland
Eutopic, large gland
Eutopic gland

Imaging findings alone are unable to conclusively differentiate transient from permanent CH
as eutopic thyroid glands are found in both scenarios. A trial off therapy consisting of a period
during which thyroid hormone supplements are not provided or provided at a significantly lower
dose and thyroid function is monitored by measuring TSH and T4 to determine if treatment remains
necessary is recommended around age three years for all suspected cases other than those with
thyroid aplasia to confirm the diagnosis of permanent CH.[27, 68-70] Infants having a low serum
free T4 and elevated TSH concentration at the conclusion of the trial are confirmed as having
permanent CH.[24] Some have also suggested that molecular analysis be conducted to identify
genetic causes of CH during the process of medical management.[71]
While guidelines exist to assist clinicians in the management of CH, there is no consensus
case definition or universal testing strategy applied in diagnosing the disorder.[69, 72] Many
diagnosed infants do not undergo thyroid imaging or radioisotope scanning to identify the
underlying thyroid morphology because the information gained is not required to manage CH. It is
8

unknown how many cases undergo therapy cessation around age three years to determine disease
permanence due to the lack of long term follow-up in most NBS programs and the absence of
pediatric endocrinology clinical guidelines.[26]
Outcomes of Congenital Hypothyroidism:
Pathologically low concentrations of thyroid hormone during critical stages of development
can cause severe mental retardation and skeletal growth abnormalities and several studies have also
associated CH with deaf-mutism, spastic diplegia and extrapyramidal rigidity.[9, 54, 73-80] Grosse
and Van Vliet recently reviewed population based studies of the outcomes of CH prior to the NBS
era and concluded that 8%-29% of cases detected clinically (~1:25,000 births) developed intellectual
or learning disability as indicated by either low intelligence quotient (IQ) scores or requirement of
special schooling.[81] The mean IQ scores observed among clinically detected cases of CH in studies
included in their review ranged from 82 to 87, representing a 20-25 point loss in IQ relative to their
potential assuming an expected score of 105-110. Gorsse and Van Vliet’s review further reported
that the percentage of clinically detected cases with height below the 10th centile ranged from 19% to
31%.[81-83] Alternatively, normal development is expected in most CH cases detected by NBS and
treated shortly after birth; deleterious outcomes in the context of NBS are associated with maternal
thyroid disease, late onset and an inadequate dosage of thyroid hormone substitution, a poor socialeconomic environment and lack of treatment compliance.[46, 74, 81, 84-97]
Treatment of Congenital Hypothyroidism:
CH is treated by thyroid hormone supplementation with levothyroixne. The initial goal of
treatment is to normalize T4 concentrations within two to three weeks and TSH concentrations
within one month.[72, 98] Target values during the first year of life are 10-16 µg/dl for serum T4,
18-30 pmol/L for free T4 and less than 5 mU/L for TSH.[27, 99] Initial dosages of levothyroxine

9

reported in past studies range from 5 to 15 µg/kg/day, although the European Society for Paediatric
Endocrinology and the AAP recommend a starting dose of 10-15 mcg/kg/day.[99-101]
Epidemiology of Congenital Hypothyroidism- All Birth Weights
Prior to NBS, the prevalence of CH was thought to range from 1:7,000 to 1:10,000 live
births.[1] NBS programs around the world initially reported detection rates ranging from 1:3,000 to
1:4,000 infants screened [102-105], although unexplained international variation in the birth
prevalence of CH has been observed. (Table 2)
Table 2: Geographic Variation in Birth prevalence of Congenital Hypothyroidism
First Author
Year
Location
Infants
Cases
Detection
Screened Detected
rate
(N)
(N)
Harris[106]
2002
United States
3,655,755
1,608
1: 2,274
Rendon-Macias[107]
2000-2004 Mexico
2,777,292
1,286
1:2,160
Deladoey[108]
1990-2005 Quebec
1,303,341
424
1:3,074
Mikelsaar[109]
1998
Estonia
20,021
7
1:2,860
Ray[110]
1979-1993 Scotland
1,513,600
344
1:4,400
Jones[66]
1994-2003 Scotland
54,473
17
1:3,084
AlJurayyan[111]
1989-1995 Saudi Arabia
1,007,350
306
1:3,292
Saudi Arabia,
AlJurayyan[112]
1990-1995
30,810
22
1:1,400
Najran Province
Fernandeziglesias[113] 1995
Spain
91,529
27
1:3,400
Inoue[114]

1979-1990

Japan

270,597

46

1:5,900

Connelly[115]
Skordis[116]
Stranieri[117]
Ordookhani[118]

1977-1997
1990-200
2003-2004
1998-2002

Victoria, Australia
Cypriot, Greece
Brazil
Iran

704,723
109,800
66,337
35,067

199
61
7
25

1:3,541
1:1,800
1:9,448
1:1,403

Within the US, the birth prevalence of CH also varies considerably by state as depicted in
Figure 2. NBS programs in the US have recently reported an increase in the birth prevalence of CH
from 1:3,985 in 1987 to 1:2,274 in 2002 not fully explained by changes in laboratory methods or
potential misclassification of transient disease.[26, 119-121] Analysis of US data from 1991-2000
revealed an increase in the birth prevalence of CH only among white infants, although state specific
10

differences by sex, race, ethnicity and birth weight have been reported.[122] California observed an
increase in the birth prevalence of CH among Hispanic newborns from 2000-2007. Texas observed
an increase among males that varied by race and ethnicity from 1992-2006. The increase in the birth
prevalence of CH observed in Massachusetts was primarily among low birth weight newborns who
had a delayed rise in TSH.

Figure 2: Birth Prevalence of Congenital Hypothyroidism in the United States, 1987-2002
(Harris et al., 2007) [For interpretation of the references to color in this and all other figures,
the reader is referred to the electronic version of this dissertation]
Central CH is less common than primary CH; Dutch studies report a birth prevalence of
central CH of 1:20,263 infants (95% CI: 1:12,976-1:33,654).[123, 124] The National Newborn
Screening Information System reports an overall US birth prevalence of central CH of 1:158,268,
although many NBS programs are either not designed to detect the disease or fail to differentiate it
from primary CH making the estimate invalid.[125] It is similarly difficult to determine the birth
prevalence of transient CH since definitions and estimates vary significantly by geographic location.

11

Kempers et al.’s analysis of Dutch NBS data revealed a birth prevalence of transient CH of
1:12,000.[126] Victorian NBS reports a transient CH detection rate of 1:23,750.[127] An earlier
Canadian follow-up study of 56 CH cases estimated only 1%-2% of cases detected by NBS are
transient.[128] Two Iranian studies reported the birth prevalence of transient CH to range from
1:1,114 to 1:5,845 infants screened.[129, 130] Gaudino et al.’s analysis of CH cases with gland in
situ detected by a French NBS program revealed 38% were transient, amounting to a birth
prevalence of 1:10,383.[131] However, if a case definition involving an initial positive test followed
by a subsequent negative test result were used, the birth prevalence of transient CH would be
significantly greater, particularly among premature infants screened by a primary T4 testing method
given their high rate of false positive screening determinations.
The birth prevalence of transient CH in the US is less clear due to the lack of long-term
follow-up. Kemper et al. analyzed MarketScan Commercial Claims and Encounters data and
MarketScan Multi-State Medicaid data from 2001-2006 and revealed that 38% of more than 700 CH
cases no longer exhibited claims for thyroid replacement therapy after 36 months and may have
represented transient CH.[132] Texas NBS identified only four instances of a change from
permanent to transient CH from 2004-2006, although their study relied on passive reporting by
pediatric endocrinologists to determine transient etiology.[26] Eugster et al.’s clinic based follow-up
study of 33 CH cases detected by Indiana NBS who either had anatomically normal thyroid glands
on imaging, no thyroid imaging, or relatively mild T4 and TSH concentration abnormalities despite
imaging results indicating thyroid agenesis revealed that 36% had a transient state.[68]
Similar follow-up studies conducted outside of the US that did not relate observed cases to a
population base report varying estimates of the proportion of transient versus permanent CH
detected by NBS. Three studies of children with CH having thyroid glands in situ conducted in Italy
revealed varying distributions of transient disease. Corbetta et al.’s analysis of 59 CH cases revealed
12

22% were transient upon three year follow-up.[133] Costa et al.’s clinic based follow-up study of 23
infants revealed that nearly half were transient.[134] Weber et al’s study determined only one of 31
CH cases evaluated were transient.[135] Nair et al.’s clinic based follow-up study of 36 CH cases
with gland in situ detected by an Indian NBS program reported that half were transient.[136]
Tamam et al.’s follow-up study of 182 cases detected by Turkish NBS revealed that 30% had
transient CH.[137] In the largest follow-up study identified during comprehensive literature review,
Panaoutsopoulos et al. reported that 24 of 413 cases detected by Greek NBS were determined to be
transient at three year follow-up.[138] More recently, a Chinese study reviewed by Parks et al.
followed up157 CH cases who required low maintenance levels of levothyroxine around age three
years reported that 90% of the children’s TSH concentrations eventually normalized.[139]
Demographic & Perinatal Differences in Birth Prevalence
Harris et al.’s report of New York state’s NBS program findings indicates that CH is more
common among white infants, those born earlier in gestation and/or at lower birth weights,
multiparous infants, females and among infants born to older mothers relative to non-white infants,
those born later in gestation and/or at greater birth weights, singleton infants, males and infants
born to younger mothers respectively.[106] (Table 3) The birth prevalence is particularly elevated
among LBW and/or premature infants. Historically, CH has been characterized by sexual
dimorphism with the ratio of diagnosed females to males being 2:1; this ratio is greater among
infants diagnosed specifically with thyroid aplasia or ectopy.[32, 46, 66, 108, 140, 141] An Italian
NBS study reported female to male ratios of 2:1 and 0.5:1 among CH cases having thyroid
dysgenesis and those diagnosed as having transient hypothyroidism respectively.[46] A Scottish NBS
study also reported a female to male sex ratios of 2:1 and 1:1 among infants diagnosed with definite
and transient hypothyroidism respectively.[66] Parks et al. recently analyzed US data from the
National Newborn Screening and Genetics Resource Center for the period 1993-2000 and reported
13

a female to male ratio of 1.56:1, suggesting that misclassification of transient disease as permanent
CH likely attenuated the ratio.[26] No US study has evaluated the sex ratio among transient relative
to permanent CH cases identified by NBS due to the lack of long term follow-up.

Table 3: Congenital Hypothyroidism Detection by Selected Population Segment, New York
State Newborn Screening, 2000-2003 (Harris & Pass, 2007)
Population Segment
Number of infants
Number of
Detection
Screened (N)
infants
Rate
Diagnosed (N)
Male
Female
Asian
Hispanic
Race/Hispanic
Ethnicity
White
Blacks
Single
Single vs. multiple
Twins
births
Triplets
<1500 g
Birth weight
1500 to <2500 g
P2500 g
<20 years
20–29 years
Mother’s age
30–39 years
>39 years
Total

551,974
520,271
59,987
204,293
551,701
192,074
1,025,547
42,914
4022
39,086
78,225
954,454
80,251
481,213
464,420
41,168
1,072,784

Gender

313
325
59
131
304
101
581
49
7
28
92
518
45
285
277
31
638

1:
1:
1:
1:
1:
1:
1:
1:
1:
1:
1:
1:
1:
1:
1:
1:
1:

1,763
1,601
1,017
1,559
1,815
1,902
1,765
876
575
1,396
850
1,843
1,783
1,688
1,677
1,328
1,681

Controversies in Newborn Screening for Congenital Hypothyroidism:
After more than 30 years of operation, a plethora of controversies persist about how to
conduct many of the components of NBS for CH including screening, diagnosis,
treatment/management, evaluation and long-term follow-up.[65, 142, 143] NBS programs employ
14

various screening methods and differ in their definitions of CH and its sub-classifications, as well as
in the diagnosis and management of newborns with positive screening results (i.e., whether
ultrasound or other tests are used). NBS programs also collect and report different types of
information about diagnosed cases, varying by how long they follow identified cases and whether
they play a role in confirming disease permanence.[26, 65, 121] This is in part due to the lack of
quality assurance activities or comparative effectiveness evaluation involving other than NBS
laboratory operations.[8, 144, 145] Guidelines have been published including those of the Council of
Regional Networks for Genetic Services illustrating the need for cost-effectiveness evaluation and
high quality NBS operations[65], although few studies demonstrating cost-effectiveness and/or
quality have been published. Therrell et al. recently (2010) published the NBS Performance
Evaluation Assessment Scheme to outline high quality indicators, although it is unknown whether
any existing NBS program exemplifies high performance on each of the metrics.[8] Absent
consistent evidence of ideal NBS operations, inter-program variation persists as do questions about
how NBS should operate, particularly in regard to CH. Questions remain about dried blood spot
testing methods including whether primary TSH, primary T4, or a combination of both is the ideal
strategy. Controversy exists about whether serial testing is necessary to detect CH among infants at
risk for later rising TSH concentrations. Questions also remain about which infants benefit from
thyroid hormone treatment, how long some cases should be treated and what this means in regard
to NBS program operations. Furthermore, specific pediatric endocrinology clinical guidelines are
needed in addition to improvement in screening methods.
Dried blood spot testing for Congenital Hypothyroidism:
Initially, TSH testing methods were thought to be more specific and less sensitive in
detecting CH relative to the primary T4 testing strategy.[56] While the sensitivity of TSH assays
improved over time, their primary detractions are: a) an inability to detect rare cases of central CH
15

having low T4 and normal TSH concentrations, and b) an inability to detect infants having a late rise
in TSH occurring beyond the time of initial NBS specimen collection. While T4 testing methods are
capable of detecting some but not all cases of central CH and are not impacted by late rising TSH so
long as referrals for confirmatory testing are made based on T4 alone, they also detect disorders
other than primary CH including thyroid-binding-globulin deficiency and hypothyroxinemia of
prematurity which inflates the number of false positive screening results. Further, in T4/TSH testing
programs, 90% of infants do not receive a TSH screen meaning infants with normal T4 and elevated
TSH concentrations are not referred for confirmatory testing. The primary detraction associated
with tandem T4 and TSH testing for all infants is the additional associated costs of using two
methods, although little evidence of cost-effectiveness, or lack thereof, has been published.[146]
The debate over whether T4 or TSH testing methods or a combination of both are
preferable has persisted since the advent of NBS for CH. Shortly after developing the first TSH
assay used in NBS for CH in Quebec, Canada, Dussault and Morissette simultaneously measured T4
and TSH concentrations from 93,000 consecutive filter paper blood samples over a six month
period and reported that the T4 assay had greater precision and reproducibility than either of the
two TSH kits commercially available at the time.[147] Their study reported similar frequencies of
detected cases and false negative results for each method although one infant with secondary
hypothyroidism was detected by the T4 and not the TSH testing method. Mandel et al. conducted a
retrospective analysis of infants born in Massachusetts from 1993-1996 who underwent both T4 and
TSH testing for CH and concluded that 13% of cases were not detected by the initial TSH screen
due to delayed rise in TSH.[22] More recently, Zamboni et al. investigated detection of CH during a
period of tandem TSH and T4 testing for all infants in northeast Italy from 1989-2001 (N= 745,258)
and reported that an additional 21 CH cases detected (1:35,485 infants screened) via primary T4
testing would have been missed had primary TSH testing been the only method employed.[148]
16

Cases detected by the T4 method and missed by the TSH method included four cases of central CH
and 17 diagnosed as primary CH with delayed rise in TSH within the second month of life. Zamboni
et al. also noted that 10 cases of central CH were not detected by either T4 or primary TSH
screening methods.
Alternatively, Wang et al. analyzed receiver operating characteristic (ROC) curves of primary
T4 and primary TSH testing methods among 2,198 apparently healthy infant and 117 primary CH
patient samples.[149] Their study reported a preferable area under the ROC curve for primary TSH
testing relative to T4 backup TSH testing methods and estimated that approximately 8% of CH
cases misclassified by T4 backup TSH testing strategy as ‘normal’ are appropriately detected by
primary TSH testing. Wang et al. noted similar results from prior investigations in Oregon, Texas
and Missouri in which 4%-10% of CH cases identified by the primary TSH screening method would
have been missed by T4 backup TSH testing strategies.[150-152] Pass and Carmago report that 131
NBS programs worldwide employ a primary TSH testing strategy and 75 use the T4 backup TSH
method.[31] However, in the US, most NBS programs employ a T4 backup TSH testing method,
although several have recently switched to a primary TSH testing strategy or a combined T4 and
TSH testing method for all infants screened to detect CH.[153]
The impact of serial testing protocols must also be considered in comparing dried blood
spot testing protocols used in NBS for CH. Few NBS programs in the US make referrals for
confirmatory testing based on low T4 concentrations alone, meaning most suffer from an inability
to detect CH among newborns having late rising TSH in initial blood specimens collected during the
first few days of life.[123] Serial testing protocols involve retesting selected infants later in life to
account for difficulties in NBS among premature and/or sick newborns; however, there is no
universally accepted method for serial NBS. In 2009, the Clinical and Laboratory Standards Institute
(CLSI) published a guideline for NBS for preterm, LBW and sick newborns recommending a first
17

screen upon admission to a sick child birthing unit (SCBU), again at 48-72 hours of life, and once
more at discharge or day 28 of life; the latter screen is indicated primarily for infants [I assume born]
younger than 34 weeks and/or weighing <2,000g.[154] While the CLSI guideline states that their
recommendations “..provide the most reliable screening for the infant requiring the fewest
specimens possible..(pg.19)”, there is little evidence supporting this statement. Prior studies of serial
testing protocols have applied both varying inclusion criteria and different testing regiments and no
study identified during comprehensive literature review objectively compared different serial testing
protocols.
Mandel et al.’s study of tandem T4 and TSH testing among Massachusetts infants born in
1993-1996 retested those with low initial T4 concentrations at two weeks of life or at discharge.
Their study identified 13% of all cases across all birth weights after the initial screen; of those
detected, 4 were normal birth weight (NBW- >2,500g), 4 were low birth weight (LBW - <2,500g),
and 10 were very low birth weight (VLBW- < 1,500g).[22] (Table 4) Larson et al. analyzed data from
the New England combined T4 and TSH testing NBS program in which all infants born weighing
less than 1,500g were retested at 2, 6 and 10 weeks of age or until reaching a weight of 1,500g;
infants admitted to the neonatal intensive care unit (NICU) were also retested at two weeks of
life.[21] Their study reported that 1:400 VLBW and 1:75,000 non-VLBW infants were diagnosed as
CH with delayed TSH elevation. Similar results have been reported outside of the US.
Table 4: Birth Prevalence of Atypical Hypothyroidism (low T4 & late rising TSH) by Birth
Weight, Massachusetts Newborn Screening, 1993-1996, (Mandel et al., 2000)
Birth Weight
Total Births
Atypical CH
Birth prevalence
NBW (>2,500g)
311,281
4
1:
77820
LBW (1,500-2,499g)
16,899
4
1:
4225
VLBW (<1,500g)
3,238
10
1:
324
Silva et al. recently analyzed data from a Brazilian NBS program that employed a primary
TSH testing method and retested dried blood spot samples collected on the 5th, 10th and 30th days
18

of life from newborns born at gestational ages less than 32 weeks and/or those that were born
VLBW from 2004-2006. Their study reported that 66% of TSH elevations were detected after the
first screen and that only one of 11 cases was detected on the initial screen of blood samples
conducted around the 5th day of life.[155] Tylek-Lemanska et al. evaluated data from a Polish NBS
program employing a primary TSH testing method with repeat testing after four weeks for LBW
infants and identified one case of elevated TSH (>10 mlU/L) per 202 LBW infants retested after
four weeks of age; of those identified as having elevated TSH, 84% were treated.[16] Alternatively,
Vincent et al. analyzed Quebec NBS data from 1993-1994 when tandem T4 and TSH screening
methods were used and repeat NBS was conducted for all VLBW infants and concluded that serial
testing was unnecessary based on their finding of a single case of late rising TSH later diagnosed as
CH.[19] However, Vincent et al.’s study was under-powered (n=465) to detect CH among children
having a late rise in TSH and their findings were likely impacted by delayed initiation of NBS.
Madison and LaFranchi reviewed the overall controversies regarding T4 versus TSH
screening approaches and reported that while both methods detect 98%-100% of cases, the
proportion of false positive screening results (calculated as the number of false positive results
divided by the total number of screens and often referred to as the false positive rate in NBS) was
significantly greater in T4 screening programs (1.44%) compared to TSH screening programs
(0.45%).[156] Alternatively, Hertzberg et al. noted that during 1991-2000, US laboratories that used
TSH assays reported a 24% greater birth prevalence of CH than those that used T4 assays
suggesting greater sensitivity for primary TSH testing protocols.[121] Pass and Neto compared
screening performance metrics among US NBS programs employing four separate protocols and
noted significant variation in each, both between and within protocol groupings.[31] (Table 5)
While it is relatively well understood that TSH testing for CH yields fewer false positive screening
results than T4 testing strategies, the issue of sensitivity is less clear. Absent a clear and consistently
19

applied operational case definition, it is difficult to compare the birth prevalence of CH generated by
different NBS programs or within the same program over time.
Early studies that observed equivalent sensitivity of T4 and TSH testing strategies reported
detection rates ranging from 1:3,000 to 1:7,000, suggesting that they applied a classical CH case
definition.[147, 148] Studies that report increased sensitivity via application of primary TSH testing,
serial testing methods or reduced initial cutoff values have noted elevated detection rates (~1:2,000)
suggesting that they likely include milder forms of CH, particularly HT or subclinical CH, although
many NBS programs fail to differentiate HT from primary CH making it difficult to validate this
assumption.[26, 133, 157-159] It remains controversial whether detection of HT or subclinical
hypothyroidism is a useful application of NBS as it is unclear if these cases benefit from treatment.
Thus, it is difficult to weigh the benefit of increased sensitivity associated with different dried blood
spot testing protocols that identify mild cases of CH.
New methods of detecting CH in dried blood spots have also been developed and some are
currently used in Europe. In 2009, Chace et al. reported a method of measuring T4 in dried blood
spots using tandem mass spectrometry, an extremely cost-effective method of testing single blood
spots for multiple conditions.[160] Similarly, Pass and Neto recently reported development of
multiplex systems capable of simultaneously measuring T4, TSH, free T4 and TBG.[31] Some
programs are also employing bioluminescence enzyme-linked immunosorbent assays to test TSH in
dried blood spots.[161] Others have suggested that TRH be tested in NBS for CH; however, little
evidence of the comparative effectiveness of new strategies relative to traditional testing methods
has been reported.[162, 163]
Treatment for Congenital Hypothyroidism
The AAP currently recommends treating HT patients if the condition persists after six weeks
of age, although little evidence exists about the cognitive outcomes of permanent or transient forms
20

of the disease.[27, 164, 165] A Swedish retrospective study identified six children with untreated
subclinical CH by analyzing dried blood spots collected during the first week of life for NBS for
PKU and tested them using the Griffiths test of developmental quotient at age five years.[166]

Table 5: Screening Performance Metrics, United States, 2006 (Pass & Neto, 2009)
Protocol/State
Infants
Screened
Dx
PPV Detection
FPR
Screened
Positive Cases
rate (1:)
N
%
N %
T4 backup TSH
Arizona
98,605 1141 1.2
0 N/A N/A
1141 1.2
Colorado
69974 296 0.4
19
6.4
3683 277 0.4
Massachusetts
73156 981 1.3
36
3.7
2032 945 1.3
Indiana
89,888 1849
2
28
1.6
3210 1821 2.0
New York^
247,901 4539 1.8
217
5.3
1037 4300 1.7
Texas
394,629 7688 1.9
239
2.8
1819 7451 1.9
New Jersey
111,290 2648 2.3
79
2.9
1408 2569 2.3
TSH backup T4
Missouri
44,621 123 0.3
16
13
2788 107 0.2
Oklahoma
51,126
39 0.07
15 38.5
3408
24 0.0
Oregon
46,881 990 2.1
23
2.3
2086 967 2.1
Virginia
104,475 2513 2.4
38
1.5
2749 2475 2.4
T4 and TSH
West Virginia
21,356 883 1.3
24
2.7
889 465 2.2
Georgia
145,074 3326 2.3
26
0.8
5580 3300 2.3
Kansas
41,293 849 2.1
40
4.7
1032 309 0.7
Louisiana
61,456 524 0.8
16
3
3849 508 0.8
South Carolina
55,530 1529 2.8
27
1.8
2057 1502 2.7
TSH only
Arizona
38,845 1546 3.9
13
0.8
2984 1533 3.9
Iowa
39,529 629 1.5
30
4.8
1318 588 1.5
New Mexico
28,538 196 0.6
15
7.6
1902 181 0.6
Minnesota
72,278 191 0.3
35 18.3
2605 159 0.2
California
656,090 789 0.1
301 38.2
1847 488 0.1
Note: PPV= Positive Predictive Value; Dx= Diagnosed; ^=2005 data reported; % represents
proportion of total screens. Table modified from Pass & Neto, 2009 version.

21

None of the children with subclinical CH had an IQ below 70, nor did any require special education
services; however, they had an average decrement of about 7 IQ points and also exhibited a greater
frequency of behavioral problems compared to six euthyroid children. Similarly, Azizi et al.’s
historical cohort study comparing global intelligence and psychomotor performance at age 9 years
among 19 children having transient hyperthyrotropinemia relative to euthyroid children born at the
same time reported an 8 point decrement in mean IQ among affected children.[167] Leonardi et al.
followed-up 44 newborns transient hyperthyrotropinemia cases that did not receive treatment after
approximately 5 years of age and reported that 31% had subclinical hypothyroidism.[168]
McDermott and Ridgway’s review of adult onset hyperthyrotropinemia concluded that treatment
was both beneficial and cost-effective, yet these findings are not generalizable to newborns.[169]
Alternatively, several other small studies have reported normal mental and physical
development among untreated HT cases.[170-173] Miki et al. reported normal mental and physical
development among 16 cases of transient hyperthyrotropinemia, yet identified signs of mild
pituitary-thyroid dysfunction in later childhood.[171] Cody et al. conducted a retrospective study of
children with hyperthyrotropinemia attending a single clinic over a 20 year period; all of the children
had normal growth and development yet only one child was treated at age one year, four had a
transient form of the disease which resolved within 3-18 months, and three continued to have
persistently elevated TSH concentrations at ages 5, 9 and 17 years.[172] Kohler et al. re-evaluated 61
children with transient neonatal elevations of TSH and reported that all had normal thyroid function
and physical development at ages 6-14 years.[173] Several investigators have also reported potential
harm associated with treatment of hyperthyrotropinemia patients. Zung et al. conducted a
multicenter study of transient (n=18) and persistent (n=25) hyperthyrotropinemia patients and
reported that 79% and 55% respectively experienced elevated concentrations of free T4 during
treatment, concluding that iatrogenic hyperthyroidism was an issue of concern.[25] An earlier
22

follow-up study of 36 hyperthyrotropinemia cases similarly noted that treatment was stopped in
three children due to iatrogenic hyperthyroidism.[174] Salerno et al.’s study of 30 young adults with
CH detected by NBS also raised the issue of cardiovascular abnormalities potentially related to longterm levothyroxine therapy relative to age and sex matched controls, although further study is
necessary.[175]
It is also unclear whether cases of hypothyroxinemia should be treated. While NBS programs
have traditionally considered positive screening results associated with hypothyroxinemia of
prematurity as false positives, evidence suggests poor outcomes and a potential benefit from
treatment for such cases. Reuss et al. reported that after adjustment for gestational age and other
perinatal and neonatal factors the odds of disabling cerebral palsy were more than fourfold greater
(OR 4.4; CI: 1-18.6) and the mental-development scores were nearly 7 points less among children
who had severe hypothyroxinemia, defined as blood T4 values greater than 2.6 standard deviations
below the mean for New Jersey newborns, compared to those who did not have the disease.[176]
Ongoing research is investigating whether premature infants suffering from hypothyroxinemia
benefit from thyroid hormone supplementation.[70, 177] While further research is ongoing to
assess the long-term neurodevelopmental effects of treatment, if the results determine benefit then
hypothyroxinemia may warrant treatment and management in the context of NBS.
It is clear that primary CH requires lifelong treatment, although debate persists about the
ideal initial dosage. Higher doses (10-15 µg/kg/day) of levothyroxine have been associated with a
quicker time to normalize thyroid hormone concentrations (within 3 weeks) compared to 8
µg/kg/day (within 6-8weeks), although the association between higher initial dosage and cognitive
outcome remains controversial.[178-180] Higher doses have also been associated with hyperactivity,
delinquency, aggression and poorer attention in school aged children.[180, 181] Ng et al. recently
published a Cochrane review of high versus low dose of initial thyroid hormone replacement for CH
23

and concluded that there is inadequate evidence to suggest that a high doses are more
beneficial.[182] However, the single trial (published in two parts) that met Ng et al.’s inclusion
criteria reported favorable outcomes associated with higher levothyroxine starting dose including
earlier thyroid hormone normalization and improved full scale intelligence quotient, although
adverse effects were not reported.[183, 184] Determination of whether certain types of CH should
be treated or which dosages are ideal requires long term follow-up.
Long Term Follow-Up and Confirmation of Disease Permanence
While long term follow-up of cases detected by NBS has received progressively more
attention over the past decade, few programs are designed to follow children beyond the first few
months of life.[26, 143, 185] Accordingly, it is unknown whether many CH cases detected by NBS
are permanent considering that few have documented therapy cessation trials as recommended.
Texas NBS relies on passive reporting by pediatric endocrinologists to monitor changes in the
diagnosis of CH from permanent to transient, although less than a quarter of cases are followed up
to age three years.[26] Indiana NBS reports following up CH cases whom a permanent requirement
for exogenous thyroid hormone has not been established and conducting trials of therapy cessation
after age three years, although no information on the proportion of eligible cases followed has been
published.[186] Three year follow-up and therapy cessation trials to confirm the diagnosis of
permanent CH appear more common outside of the US, although few programs, if any, do so in a
standardized manner and achieve follow-up of a majority of detected cases.[118, 129, 133, 134, 136]
There is no clear standard for conducting treatment cessation trials or interpreting their
results to confirm permanence of CH. The duration of treatment cessation varies from 3-4 weeks to
10 months among the few published studies and the frequency and timing of follow-up thereafter
among children whose thyroid function normalizes after cessation of thyroid hormone therapy also
varies.[68, 128, 139] The AAP briefly reports that therapy cessation should be conducted for CH
24

cases after age three years for 30 days if no permanent etiology was found and no increase in TSH
was observed after the newborn period, yet the guideline cites only a single study conducted by
Eugster et al.[27] Eugster et al. advise a more rigorous protocol for confirming CH permanence
including: a) a trial off therapy for four weeks after age three for cases without an identified
permanent cause; b) after 4 weeks, re-measurement of thyroid hormone concentrations and conduct
of thyroid ultrasounds; c) follow-up of abnormal ultrasound results with a Te-99m thyroid scan and
follow-up of abnormal thyroid function tests with a perchlorate washout test; and follow-up of
children with normal results for an additional 12 months.[68]
The concept of long term follow-up in the context of NBS is also expanding beyond the
first few years of life. The US Secretary of Health and Human Services’ Advisory Committee on
Heritable Disorders and Genetic Diseases in Newborns and Children defines long term follow-up as
providing assurance and provision of quality chronic disease management, condition-specific
treatment, and age-appropriate preventive care throughout the lifespan of affected individuals
through coordination of a medical home, evidence-based treatment, continuous quality
improvement and new knowledge discovery.[143] However, a system demonstrating successful long
term follow-up across the lifespan has yet to be realized and many physicians may be unprepared to
assist in managing follow-up care for children with positive NBS results, suggesting an even greater
need for NBS programs to coordinate such care.[187]

25

CHAPTER 2 METHODS
Study Design & Participants:
In this dissertation, a population-based retrospective cohort design is applied in studying
outcomes and implications of changes in MI NBS for CH program operations from 1/1/19946/30/2010; the primary exposure of interest is the method of dried blood spot testing defined based
on infant date of birth (DOB) as follows: 1) T4 backup TSH testing (DOB: 1/1/1994-12/31/1997),
2) tandem T4 & TSH testing for all infants, no serial testing (DOB: 1/1/1998-9/30/2003); 3)
primary TSH testing, no serial testing (DOB: 10/1/2003-2/28/2007); 4) primary TSH testing, serial
testing for births weighing <1,800g (DOB: 3/1/2007-6/30/2010). MI resident infants born and
screened in MI are included in this study; out-of-state infants born within the state are followed-up
and diagnosed elsewhere and are accordingly excluded from consideration. The period of
observation is restricted to beyond 1994 due to inadequate availability of data capable of
representing the period 1977-1993. Results of the overall comparative analysis are reported in
chapter 3.
Two sub-cohorts are also studied. First, in sub-cohort 1, outcomes of serial testing for CH
are investigated among children born weighing less than 1,800g during primary TSH testing periods
(DOB: 10/1/2003- 6/30/2010); results are reported in chapter 4. Second, in sub-cohort 2, the
proportion of cases later determined transient is investigated among those detected since 2004 and
followed-up after age three years; this cohort includes followed cases whose clinician either engaged
in diagnostic re-evaluation or reported why they had not or would not conduct a diagnostic reevaluation (i.e., if thyroid imaging was indicative of permanent disease). Cases that were followed-up
but did not undergo diagnostic re-evaluation for a specified reason were excluded from sub-cohort 2
because it was unable to be determined whether the case experienced transient or permanent CH.

26

Results of the sub-cohort 2 analysis are reported in chapter 5. Chapter 6 provides conclusions and
details future directions focusing on the need for additional research and program changes based on
our findings.
Research Questions & Hypotheses:
R.1

Do NBS for CH performance metrics (Detection Rate, False Positive Rate (FPR), False

Negative Rate (FNR), Positive Predictive Value (PPV), Sensitivity, and Specificity) differ during four
periods defined by method of dried blood spot testing as follows: T4 backup TSH testing (DOB:
1/1/1994-12/31/1997); tandem T4 & TSH testing for all infants, no serial testing (1/1/19989/30/2003); primary TSH testing, no serial testing (10/1/2003-2/28/2007); primary TSH testing,
serial testing for births weighing <1,800g (3/1/2007-6/30/2010)?
H1.a: During tandem T4 and TSH testing, the T4 screen will have greater sensitivity and
less specificity compared to the TSH screen.
H1.b: The majority of false positive screening determinations resulting from the initial T4
screen will be observed among LBW and/or premature infants.
H1.c: Tandem T4 and TSH testing will have greater sensitivity and less specificity than
both primary TSH testing protocols.
H1.d: The FNR will be greater during primary TSH testing periods relative to the tandem
TSH and T4 testing period.
H1.e: Sensitivity will be greater and specificity will be lesser during the primary TSH
plus serial testing period compared to the primary TSH testing absent serial testing
period among the overall population and among births weighing < 1,800g.
H1.f: Among primary TSH testing periods, fewer cases will be detected among births
weighing 1,800g-2,499g after the addition of serial testing among births weighing <
1,800g.
27

R.2

Are more cases of central hypothyroidism detected in Michigan during primary T4 testing

periods compared to the primary TSH testing absent T4 testing periods?
H2: T4 testing will detect a greater number of central hypothyroidism cases than primary
TSH testing.
R.3

Do diagnostic re-evaluation methods differ by physician and do they adhere to guidelines

published by the AAP?
H3: Child’s age at time of diagnostic re-evaluation, duration of time off treatment,
whether or not treatment is stopped or dosage reduced, and/or reason for not conducting
a trial off medication during diagnostic re-evaluation will differ by physician.
R.4

At what rate to CH cases stop taking thyroid hormone medication without medical

supervision?
H4: Greater the 10% of cases’ treatment will be stopped absent medical supervision.
R.7

Do results of medication cessation (confirmed transient or permanent CH) differ by whether

treatment was stopped with or without medical supervision?
H7: Cases whose treatment is stopped without medical supervision will be more likely to
be determined transient CH compared to those whose treatment was stopped under
medical supervision during diagnostic re-evaluation.
R.8

Among detected cases, how many are later determined to have had transient CH?
H8: Greater than 30% of cases whose diagnosis is re-evaluated will be determined to
have had transient CH.

R.9

Do initial TSH and retest analyte concentrations and demographic/perinatal characteristics

differ among cases determined to have transient CH compared to those considered to have
permanent CH?

28

H9: Cases considered to have had transient CH will have lesser initial and retest analyte
concentrations and will be more likely to be: black, born LBS and/or premature, small
for gestational age (SGA), a twin birth and/or admitted to the NICU relative to cases
initially detected by the first screen.
Newborn Screening for Congenital Hypothyroidism in Michigan:
NBS began in MI in 1965 shortly after Dr. Robert Guthrie developed a bacterial inhibition
assay to detect PKU in newborn blood spotted on filter paper. The program expanded to include
screening for CH in 1977 as recommended by the American Thyroid Association. NBS for CH and
several other conditions was mandated in Michigan in 1987 by Public Act 14 which also designated
the state laboratory as the sole testing facility and set a fee to fund the program along with additional
components including follow-up, medical management and quality assurance. The program,
conducted by the Michigan Department of Community Health (MDCH), includes the state
laboratory and Follow-up Program which contracts four medical management centers located
outside of the MDCH.
The state laboratory is housed in the Division of Chemistry and Toxicology within the
Bureau of laboratories and is tasked with testing blood spots and establishing newborn reference
ranges that maximize sensitivity and specificity in detecting 49 conditions now included in the NBS
panel. The Follow-up Program, housed in the Division of Genomics, Perinatal Health and Chronic
Disease Epidemiology within the Bureau of Epidemiology, coordinates follow-up and medical
management of infants having positive screening determinations. The primary objective of the
Follow-up Program is early initiation of treatment, although it is also responsible for education,
training, quality assurance and data management. Medical management of infants screened positive
for CH is performed under contract from the NBS Follow-up Program by the Endocrine Follow-up
Program, located in the Department of Pediatrics at the University of Michigan. The Endocrine
29

Follow-up Program is responsible for ensuring appropriate diagnostic evaluation and treatment of
infants having positive dried blood spot screening determinations for CH or congenital adrenal
hyperplasia.
As indicated in table 6, MI NBS for CH initially used a T4/TSH testing strategy in which
infants having T4 concentrations below the 10th centile underwent TSH testing; referrals for
confirmatory testing were made based on the TSH concentration. In 1998, the dried blood spot
testing protocol was changed to include TSH and T4 testing for all infants to improve
sensitivity/specificity while retaining the ability to detect central CH; during this period, referrals for
confirmatory testing were made if either T4 or TSH concentrations exceeded fixed cutoffs. Around
2000, interpretation of test results was modified to include a borderline positive determination. In
2003, T4 testing was removed from the NBS protocol due to an unacceptable number of false
positive determinations (3%-5% of all screens), identification of few cases of central CH and
experiences suggesting that central CH would likely be detected clinically in the process of
hypopituitarism work-up. March 1st, 2007, a serial testing protocol was implemented as follows.
Infants born weighing less than 1,800g are re-screened at two and four weeks of life; if the final
specimen is consistent with continuous transfusions and/or total parenteral nutrition (TPN), testing
is repeated after 72 hours from discontinuation of transfusion and/or TPN and again 90 days after
the last transfusion if there is no documented negative screen for hemoglobinopathies.
[Insert Table 6 here]
Upwards of 99% of live births in Michigan undergo dried blood spot screening.[188] Infants
having positive or strong-positive dried blood spot screening determinations are immediately
referred for confirmatory testing. Borderline positive initial screening determinations undergo a
repeat dried blood spot screen for CH; if the result of the re-test is other than negative, then the
infant is referred for confirmatory testing. Confirmatory testing involves clinical evaluation by a
30

pediatric endocrinologist and one or more thyroid function tests consisting of serum TSH and
usually free T4 measurement; tests including thyroid ultrasound or others useful in diagnosing CH
may also be conducted, although this varies by physician. Pediatric endocrinologists involved in the
NBS Pediatric Endocrinology Advisory Committee (PEAC) anecdotally report that if serum TSH
concentrations do not fall below 10 mU/l after approximately four weeks of age then treatment is
initiated. Treated infants are recorded as confirmed cases of CH by the NBS Follow-up Program.
Thereafter, cases may be managed by pediatric endocrinologists involved in the NBS process,
endocrinologists not involved with the NBS program or by primary care physicians.
In 2007, the NBS Follow-up Program initiated follow-up of CH cases detected by NBS after
age three years to determine whether disease permanence had been confirmed and if so, the results.
Initially, borderline cases, defined as those having pre-treatment serum TSH concentrations below
the 15th centile calculated among all cases identified from 2004-2007, were targeted for three year
follow-up. In 2008, follow-up was expanded to include all CH cases detected by Michigan NBS
beginning with those born in 10/01/2003. Three year follow-up was initiated by mailing a brief
survey to the endocrinologist recorded in NBS Follow-up Program data. The survey was developed
by the author and the NBS Follow-up Program manager in collaboration with the PEAC and was
pilot tested among several pediatric endocrinologists. Information collected during three year
follow-up included whether the targeted physician was still providing care to the child; if so,
whether the child was still being treated for CH and if not, the reason why; whether a trial off
therapy had been or would be conducted and if not, the reason why. If the surveyed endocrinologist
was not currently providing care to the child, they were asked to provide contact information for the
last known care provider and another survey was mailed to that provider. If the child’s current care
provider was unknown or incorrectly identified, then NBS Follow-up Program staff extracted
current physician contact information from the Michigan Care Improvement Registry (MCIR); if so,
31

the physician was contacted by phone by the NBS program manager and the survey was conducted
by phone. MCIR, initially designed as a child immunization registry, is updated basically in
accordance with the timing of immunizations and provided a near universal source of current
provider contact information.
Data:
Data used to populate the analytic file used in this dissertation are portrayed in Figure X.
Demographic and perinatal information collected on the NBS card and managed by the state
laboratory were used to identify and characterize infants screened in 1/1/1994-6/30/2010. NBS
Follow-up and medical management data were extracted from different sources before and after
10/1/2003, when the current laboratory information system (LIMS- PerkinElmer Life Sciences,
Inc.) was implemented. Archived data managed by the NBS Follow-up Program were used to
identify and characterize infants born in 1/1/1994-9/30/2003 who received a positive initial blood
spot screen determination for CH; these data included screening analyte concentrations and
determinations and other test results generated during the process of medical management. LIMS
data including screening analyte concentrations and determinations and NBS Follow-up Program
medical management data were used to characterize infants screened 10/1/2003-6/30/2010.
Medical management data managed by the NBS Follow-up Program includes diagnostic and
treatment information. LIMS data were linked to NBS Follow-up Program data by a unique patient
identification number found in both laboratory and follow-up program data. Reports actively and
passively ascertained from pediatric endocrinologists by the NBS Follow-up Program are used in
this study to identify false negative screening results.
Results of three year follow-up of CH cases detected by Michigan NBS since 10/1/2003 that
are recorded by the Michigan NBS Follow-up Program are also used in this study. Survey activities
stopped temporarily while transitioning responsibilities from the NBS Follow-up Program to the
32

Endocrine Follow-up Program in 2009. Upon resuming three year follow-up survey activities, the
decision was made to start with children born in 2007 meaning follow-up was attempted for only
three cases born in 2006.

Demographic/Perinatal Characteristics and Data
representing births screened from 1/1/1994-6/30/2010,
managed by the State Laboratory
Dried Blood Spot Screening Results & Determinations
representing infants born in 10/1/2003-6/30/2010, managed
by the State Laboratory LIMS System
Archived Dried Blood Spot Screening Results/Determinations
& Medical Management Data Representing infants screened
positive in 1/1/1994-9/30/2003, managed by MDCH NBSFU
Medical Management Data Representing infants screened
positive in 10/1/2003-6/30/2010, managed by MDCH
NBSFU

Analytic
File

Passively reported false negative data, managed by MDCH
NBSFU
Three Year Follow-up Survey Results, managed by MDCH
NBSFU

Figure 3: Data Sources Used to Populate the Final Analytic File
Demographic and perinatal information extracted to characterize infants included their sex,
race, ethnicity (Hispanic), gestational age (weeks), birth weight (grams) and whether they were
admitted to a NICU. Gestational age was categorized as <28 weeks, 28-31 weeks, 32-36 weeks and
(>37 weeks. Birth weight was categorized as <750g, 750g-1,499g), 1,500g-2,499g and >2,500g. Fetal
growth ratio (FGR) was calculated by dividing the observed birth weight by mean birth weight for
33

gestational age based on percentile distributions published by Alexander et al.[189] Infants having
FGRs at or below the 10th centile are categorized as small for gestational age (SGA); otherwise,
infants are considered normal for gestational age (NGA). Prior to 1997, NICU admission and
gestational age were not recorded on the NBS card and are accordingly not reported during this
period.
NBS information extracted to characterize infants included specimen collection date, dried
blood spot T4 and TSH concentrations, serum TSH and free T4 concentrations, ultrasound imaging
and/or radioisotope scanning results, and treatment initiation date. Infants who received positive
dried blood spot screening determinations and received treatment at the conclusion of confirmatory
testing performed by the Endocrine Follow-up Program are considered as having CH by the
Michigan NBS program. Treated infants are sub-classified by dried blood spot TSH concentration,
ultrasound and/or radio isotope scan results (when available) and pre-treatment serum TSH and
FT4 concentration to identify children most likely to have permanent CH. Prior literature indicates
that newborns having initial dried blood spot TSH concentrations at or greater than 50 uIU/mL,
imaging results indicating thyroid dysgenesis, a serum TSH concentration greater than or equal to 20
mU/L and/or a serum FT4 concentration less than 1ng/dL are considered at greater risk of
permanent CH than infants having dried blood spot TSH concentrations < 50 uIU/mL, normal
imaging results, a serum TSH concentration < 20 mU/L and/or a serum FT4 concentration greater
than or equal to 1ng/dL.[13, 26, 64]
Results of three-year follow-up are used to differentiate transient from permanent CH cases
among those followed-up. Cases that were not receiving thyroid hormone supplements at follow-up
are considered transient; those whose clinician sent their normal lab values while not receiving
thyroid hormone treatment to the MDCH NBS Follow-up Program are considered confirmed
transient cases, those whose clinician indicated that their lab values were normal but did not yet
34

transmit those values to the MDCH NBS Follow-up Program are considered suspected transient
cases.
Analysis:
Full Cohort
Descriptive and analytical techniques are used to investigate the impact of NBS protocol
changes. Descriptive techniques include tabulation and trending of newborn characteristics by NBS
outcomes of interest (reported in Table 7) during four periods defined by method of NBS based on
their DOB as follows: T4 backup TSH testing (DOB: 1/1/1994-12/31/1997), tandem T4 & TSH
testing for all infants, no serial testing (DOB: 1998-2003); primary TSH testing, no serial testing
(DOB: 2004-2/28/2007); primary TSH testing, serial testing for births weighing <1,800g (DOB:
3/1/2007-6/30/2010).
Logistic regression analysis is used to investigate whether the detection rate and rate of false
positive screening determinations changed significantly over time as NBS program operations
changed after adjusting for differences in the distribution of selected newborn demographic and
perinatal characteristics. Differences in the overall risk of CH as well as the risk of false positive
determination are investigated.
Sub-Cohort 1: Infants Born Weighing < 1,800g (1/1/2004-6/30/2010)
Implications of serial testing are similarly investigated among infants born in 1/1/20046/30/2010 weighing less than 1,800g and characteristics of cases detected after having an initial
negative screen are compared to those detected by an initial positive screen. Newborn characteristics
are tabulated by NBS outcomes of interest during two periods again defined by dried blood spot
testing method: 1) primary TSH testing, no serial testing (DOB: 2004-2/28/2007); primary TSH
testing, plus serial testing for births weighing <1,800g (DOB: 3/1/2007-6/30/2010). Crude and
adjusted Logistic regression models are used to investigate the association between serial testing and
35

overall detection of CH. Predictors of detection after having a negative first screen are also
investigated during the serial testing period (3/1/2007-6/30/2010) using crude and adjusted logistic
regression models.
Sub-Cohort 2: Cases Detected Since 10/1/2003 & Followed-up After Age 3 Years
Three-year follow-up findings are reported by selected newborn characteristics. Outcomes of
interest of the three year follow-up project are as follows: whether the child was followed-up and if
so, whether they were currently being treated; if treated at follow-up, the dosage used; if untreated at
follow-up, the reason for absence of treatment including whether a trial off therapy had been
conducted; if a trial off therapy had been conducted, how it was conducted and its results including
those of thyroid function tests; if a trial off therapy had not been conducted, whether and when one
would be conducted; if a trial off therapy had not been or would not be conducted, the reason why.
Cases lost to follow-up are compared to those followed. Permanent cases are also compared to
transient cases. Crude and adjusted logistic regression models are used to investigate predictors of
transient CH among cases diagnosed and followed-up. Cases no longer receiving thyroid hormone
medication at follow-up are considered transient CH in this study.

36

Table 6: Dried Blood Spot Testing Protocols, Cutoff Values and Associated
Determinations Applied by Michigan Newborn Screening for Congenital
Hypothyroidism, 1977-2010
Time
Test
T4
Age at
TSH
Determination
Period
Method (mug/dl ) Specimen (uIU/mL)
Collection
T4
1977<10th
follow>20
Positive
1997
Centile
up TSH
All Ages
>20
Positive
Tandem
23-49
Borderline Positive
< 5.0
1998T4 and
>49
Strong Positive
2003
TSH
≥ 50
Strong Positive
< 24 hours
< 50
Early specimen - needs
repeat NBS
< 33
Negative
24-36
33-49
Borderline Positive
hours
≥ 50
Strong Positive
< 25
Negative
2004Primary
37hoursN/A
25-49
Borderline Positive
2010
TSH*
6days
≥ 50
Strong Positive
< 13
Negative
13-49
Borderline Positive
7-31 days
≥ 50
Strong Positive
≤ 10
Negative
> 31 days
> 10
Refer for serum testing
Note: ^ Daily percentiles were used to refer children for confirmatory testing from 19771997. Serial testing was implemented for infants born weighing <1,800g March 1st, 2007. .

37

CHAPTER 3: COMPARISON OF DRIED BLOOD SPOT TESTING PROTOCOLS IN
MICHIGAN NEWBORN SCREENING FOR CONGENITAL HYPOTHYROIDISM,
1/1/1994-6/30/2010
Results:
Descriptive Characteristics of Births Screened
Greater than two million infants are included in this population-based retrospective cohort
study. Table 7 reports the distributions of demographic and perinatal characteristics across the four
exposure periods defined by dried blood spot testing procedure. The population characteristics did
not meaningfully differ among the periods; although, due to the large sample size, observed
differences were statistically significant.
[Insert Table 7 Here]
Descriptive Characteristics of Births Screened Positive
The distribution of demographic and perinatal characteristics among infants screened
positive for CH are reported by dried blood spot testing protocol in Table 8. White infants were
proportionally more common among positive screens during the tandem T4 and TSH testing
compared to other periods. While the sex distribution among infants screened positive differed
statistically across periods, the actual differences were negligible. Twins were proportionally more
common among infants screened positive during the T4 and backup TSH testing and Primary TSH
plus serial testing relative to other periods of observation. Infants born SGA were proportionally
more common among positive screens during the primary TSH plus serial testing period compared
to both tandem T4 and TSH and primary TSH without serial testing periods. The proportion of
infants having positive screens who were admitted to the NICU after birth was greater during the
primary TSH plus serial testing period compared to prior protocols.
[Insert Table 8 Here]

38

Descriptive Characteristics of Detected Cases
As indicated in Table 9, among detected cases, the distributions of race, multiparous births,
and those born SGA did not statistically differ by dried blood spot testing protocol. The expected
sexual dimorphism of more female than male cases was reversed only during the tandem T4 and
TSH testing period; otherwise, more females than males were diagnosed as having CH. Detection
among LBW and/or premature infants as well as among those admitted to the NICU was increased
in both primary TSH testing periods relative to the T4 and backup TSH and the tandem T4 and
TSH testing periods. While CH detection increased among children born weighing less than 1,500g
following the introduction of serial testing for children born weighing <1,800g, the detection rate
among moderately LBW children (1,500g-2,499g) decreased.
[Insert Table 9 Here]
While pre-treatment serum TSH and free T4 concentrations varied substantially before and
after 10/1/2003, these data are not comparable because the time to treatment was significantly
shorter during primary TSH testing periods meaning cases would be expected to have greater serum
TSH concentrations compared to those whose serum TSH concentrations were measured
later.[Table 10] The median number of days from birth to treatment during each exposure period is
as follows: in 1/1/1994-12/31/1997, 18 days; in 1/1/1998-9/28/2003, 22 days; in 10/1/20032/28/2007, 11 days; and in 3/1/2007-6/30/2010, 15 days. Across all exposure periods, few cases
(~12-25%) underwent any form of thyroid imaging. In total, seven cases of central CH were
detected; one by the primary T4 backup TSH testing protocol and the remaining six by the tandem
T4 and TSH testing protocol.
[Insert Table 10 Here]
Screening Performance Metrics

39

Table 11 reports screening performance metrics by dried blood spot testing protocol. During
the T4 backup TSH testing period the detection rate, positive predictive value, and specificity were
each less than observed during primary TSH testing periods, both with and without serial testing;
alternatively, the FPR was more than twofold greater during the primary T4 relative to primary TSH
testing periods. The greatest rate of detection was observed during the tandem T4 and TSH testing
period (1:1,271); however, the FPR (4.45%) was far greater than during other periods of
observation. Accordingly, the PPV and specificity were significantly less during the tandem T4 and
TSH testing period compared to others observed in this study. Primary TSH testing protocols were
more specific, yielding greater PPVs; however, the detection rate observed during primary TSH
testing periods is less than was observed during the tandem T4 and TSH testing period. Overall,
primary TSH testing with serial testing for infants born weighing less than 1,800g yielded fewer false
positive results, a greater PPV and greater overall detection than either primary TSH testing without
serial testing or primary T4 backup TSH testing protocols. However, the two false negative results
observed during this study occurred during the primary TSH plus serial testing period.
[Insert Table 11 Here]
Odds of Detection & False Positive Screen Determination
Overall, after adjusting for potentially confounding factors, Tandem T4 and TSH testing and
primary TSH plus serial testing for infants born weighing <1,800g are associated with a 90% and
58% increase in the odds of detection respectively compared to primary T4 backup TSH testing. In
comparing the two primary TSH testing periods to the primary T4 backup TSH testing period, the
rate of detection was significantly increased only after introducing serial testing to the primary TSH
screening protocol. While tandem T4 and TSH testing was associated with a near twofold increase
in detection compared to primary T4 testing, it was also associated with a near threefold increase in
the rate of false positives compared to primary T4 testing. Alternatively, the FPR was significantly
40

reduced during both primary TSH testing periods relative to the primary T4 backup TSH and
tandem testing periods in both crude and adjusted models. The strongest predictor of detection and
false positive screening determination observed in this study was birth weight which was inversely
associated with both outcomes (detection and false positive determination).
[Insert Table 12 Here]
To compare the tradeoffs of tandem TSH and T4 testing verse primary TSH plus serial
testing for infants born weighing <1,800g, we applied the detection rates and FPRs to a hypothetical
birth population of 125,000 infants and estimated that an additional 297 false positive
determinations would be incurred for each additional case detected if Michigan were go switch from
primary TSH plus serial testing back to tandem T4 and TSH testing for all births.
Limitations:
This study is limited by missing data. While missing demographic and perinatal information
appeared to occur at random meaning it is unlikely to have introduced differential misclassification,
the lack of thyroid imaging results inhibited our ability to estimate whether cases more likely to have
transient CH were more likely to be detected during any of the observed exposure periods. We were
also unable to investigate ethnic, maternal age or gestational length related differences in screening
performance metrics across protocols due to missing data.
Our findings are also negatively impacted by reliance on passive reporting to identify false
negative screening results. It is possible that false negatives have occurred yet were not reported to
the MDCH NBS Follow-up Program; accordingly, our results per false negative determinations
should be interpreted with caution. However, the MDCH and the PEAC have a strong relationship
and meet often to discuss cases with the greater community of pediatric endocrinologists. Thus, it is
unlikely that any known false negatives have not been reported, although some may have escaped
clinical detection.
41

It is also possible that observed differences in the rate of detection between protocols is
confounded by unmeasured changes in diagnostic practice or changes in laboratory operations
(including kit changes). Given that the greatest rate of detection was observed during the tandem T4
and TSH testing period and that this was the only period during which the number of male cases
was greater than the number of female cases, it is possible that more cases that are likely to be
transient CH were diagnosed.
This study is also limited by the lack of universal long term follow-up beyond age three years
for infants diagnosed with CH by NBS given that we are unable to differentiate true permanent
from transient CH for a majority of cases. However, long term follow-up for CH cases beyond age
three years is now standard practice for infants born after 2007 (results of follow-up activities to date
are reported in chapter 5).
Discussion:
This study finds that unless serial testing is conducted, the T4 testing strategy is equally as
likely to detect CH as the primary TSH testing method in a population characterized by a high birth
prevalence of CH relative to the historically known detection rate of approximately 1:4,000.
However, primary TSH testing yields fewer false positive determinations making it preferable in
terms of program operating efficiency.
The addition of serial testing among infants born weighing < 1,800g to the primary TSH
testing protocol significantly increases the rate of detection relative to the primary T4 backup TSH
testing approach. During the primary TSH plus serial testing period, 14% of all detected cases were
identified by retest after having an initial negative screen which is comparable to Mandel et al.’s
earlier findings, although their serial testing protocol was not restricted to infants born >1,800g.[22]
It is potentially concerning that the detection rate decreased slightly among moderately LBW
(1,500g-2,499g) children following the implementation of serial testing for births weighing <1,800g.
42

The decline may reflect that clinicians who previously used their clinical judgment to select infants
for retesting now apply a fixed serial testing cutoff based on the MDCH serial testing protocol
inclusion criteria (<1,800g) meaning marginally LBW cases exhibiting later rising TSH may remain
undetected.
Prior literature supports the possibility of cases born weighing >1,800g remaining
undetected due to later rising TSH and the lack of serial testing in Michigan. Four of the 18 infants
detected by retest in Mandel et al.’s study had normal birth weights (>2,499g).[22] Hyman et al.
reported that 6/13 infants identified in their study as having a late rise in TSH were born weighing
greater than 1,500g and concluded that all infants admitted to the NICU should be subjected to
serial NBS.[14] Larson et al. reported in their study of infants admitted to neonatal intensive care
that for every 20 CH cases born weighing greater than 1,500g that had an elevated initial TSH
screen, one additional case was detected with delayed TSH elevation.[21] In this study, the two false
negative determinations occurred among NBW infants that were admitted to the NICU at birth.
Thus, had serial testing been conducted among all births transferred to the NICU, the observed false
negative determinations would have been screened positive by a later retest.
While serial testing significantly increased detection in our study, the downside is that
primary TSH screening approaches appear more susceptible to false negative screening
determinations and remain incapable of detecting the one to three cases of central CH expected in
Michigan each year. Our findings indicate that testing T4 in tandem with TSH increases the CH
detection rate by approximately 80% relative to the primary T4 backup TSH testing approach and
by 20% relative to the primary TSH plus serial among infants born weighing <1,800g approach.
Tandem testing is also capable of detecting the one to three cases of central CH expected in
Michigan each year.

43

In this study, during the tandem testing period, 8% of cases correctly detected by primary T4
testing would have been missed by primary TSH testing; alternatively, a surprising 72% of cases
detected by primary TSH testing would have been missed by primary T4 screening alone, a far
greater proportion than previously estimated (4%-10%).[149-152] This finding suggests that many
treated children had hyperthyrotropinemia yet were classified as CH and received treatment. The
reversal of the expected sex ratio during the tandem T4 and TSH testing period further suggests that
a greater number of cases unlikely to have permanent CH were diagnosed as CH during this period,
perhaps explaining why the observed birth prevalence was among the greatest worldwide. Nearly
70% (n=16) of children detected by primary T4 testing alone weighed <750g at birth; alternatively,
nearly 80% of children detected by TSH alone were NBW.
Our study also finds that while tandem testing identified the greatest number of cases, the
FPR (4.5%) represents an eight fold increase in the number of false positive determinations relative
to primary TSH plus serial testing. Operating costs associated with the additional laboratory
resources as well as those necessary to manage confirmatory testing among the significantly
increased number of false positive determinations deters many NBS programs from conducting
tandem testing. Accordingly, our findings indicate that primary TSH plus serial testing is a preferable
testing strategy for programs unwilling to bear the burden of additional laboratory costs and false
positive determinations created by tandem testing.
Recently developed techniques capable of measuring T4 via tandem mass spectrometry
(MS/MS) could sharply reduce the laboratory costs associated with tandem testing.[160] Our
findings indicate that if Michigan added T4 to their MS/MS screening panel it would likely detect 13 additional cases per year, far more than many of the rarer conditions currently screened for using
this technique. However, the additional false positives would be greater than yielded by other
conditions screened for by MS/MS; but this too may soon no longer be an issue. Emerging evidence
44

suggests that many children having hypothyroxinemia of prematurity, now considered false positives
in NBS for CH, may actually benefit from T4 supplementation.[70, 177] If ongoing trials of T4
supplementation continue to suggest a beneficial effect of treatment for cases of transient
hypothyroxinemia, tandem T4 and TSH testing would clearly be preferable to primary TSH plus
serial testing given that the number of treatable cases identified would increase significantly as many
false positives would then be considered treatable cases.

45

Table 7: Newborns Screened by Selected Demographic & Perinatal Characteristics & by Dried Blood Spot Testing Method,
Michigan Newborn Screening, 1/1/1994-6/30/2010
Tandem T4 and
TSH, No Serial
TSH, Plus Serial
T4 backup TSH
TSH
Testing
Testing
Population Segment
(1/1994-12/1997)
(1/1998-9/2003)
(10/2003-2/2007)
(3/2007-6/2010)
N
%
N
%
N
%
N
%
386459
73.37
511212
72.23
281854
71.9
253137
70.59
White
Race
Black
103361
19.62
129240
18.26
73688
18.8
70611
19.69
Other
36929
7.01
67327
9.51
36474
9.3
34845
9.72
Female
260569
48.79
364919
48.75
211349
48.85
193165
48.85
Sex
Male
273445
51.21
383625
51.25
221266
51.15
202298
51.15
No
526879
97.04
729276
96.63
417431
96.49
381470
96.46
Multiple Birth
Yes
16066
2.96
25446
3.37
15184
3.51
13993
3.54
<750
2249
0.43
4410
0.58
1199
0.28
144
0.29
750-1499g
643.7
1.23
8293
1.1
4690
1.1
4352
1.12
Birth Weight
1500-2499g
35301
6.73
47864
6.35
28308
6.65
25778
6.62
>2,500g
480752
91.62
693477
91.97
391351
91.96
358355
91.97
<28 weeks
2327
0.57
2184
0.57
Gestational Age
28-37 weeks
40212
9.85
36115
9.42
>37 weeks
No
Yes
No
Yes

365499
89.57
344903
90.01
318936
90.44
295320
90.23
SGA
33698
9.56
31964
9.77
387543
89.58
353966
89.51
NICU Admission
45072
10.42
41497
10.49
Total
542945
100
754722
100
432615
100
395463
100
Note: Missing data are as follows: race, n= 140,620; sex, n=22,140; birth weight, n=31,786; gestational age, n=36,843; NICU admission
and gestational age data were not recorded on the NBS card prior to 10/1/2003. Percentages reported are column based.

46

Table 8: Newborns Screened Positive by Selected Demographic & Perinatal Characteristics & by Dried Blood Spot Testing
Method, Michigan Newborn Screening, 1/1/1994-6/30/2010
Tandem T4 and
TSH, No Serial
TSH, Plus
T4 backup TSH
TSH
Testing
Serial Testing
Population Segment
p-value
(1/1994-12/1997)
(1/1998-9/2003)
(10/2003-2/2007)
(3/2007-6/2010)
N
%
N
%
N
%
N
%
6303
63.22
21678
67.76
2142
64.13
1356
61.72
White
Race
Black
3131
31.4
7723
24.14
951
28.47
635
28.9 <.0001
Other
536
5.38
2592
8.1
247
7.4
206
9.38
Female
3734
40.71
14303
43.02
1622
45.04
1098
46.23
Sex
<.0001
Male
5439
59.29
18947
56.98
1979
54.96
1277
53.77
No
9139
87.5
32416
94.89
3466
96.25
2227
93.77
Multiple Birth
<.0001
Yes
1306
12.5
94.89
5.11
135
3.75
148
6.23
<750
703
7.97
929
2.77
32
0.91
72
3.08
750-1,499g
1657
18.8
1429
4.26
95
2.7
182
7.79
Birth Weight
<.0001
1,5002,499g
2110
23.93
3115
9.28
327
9.3
258
11.04
>2,500g
4346
49.3
28101
83.7
3061
87.08
1824
78.08
<28 weeks
49
1.47
117
5.14
28-37
Gestational Age
<.0001
weeks
437
13.1
421
18.49
>37 weeks
2851
85.44
1739
76.37
No
2506
87.32
1630
83.98
SGA
0.0005
Yes
364
12.68
311
16.02
No
8362
80.06
29770
87.15
2891
80.28
1611
67.83
NICU
<.0001
Admission
Yes
2083
19.94
4390
12.85
710
19.72
764
32.17
Total
10445
20.65
34160
67.54
3601
7.12
2375
4.7
Note: Missing data are as follows: race, n= 3,092; sex, n=2,367; birth weight, n= 2,346; gestational age, n= 363. Gestational age data were
not recorded on the NBS card prior to 10/1/2003. Percentages reported are column based. P-values represent the test for differences in
the proportion of infants in each population segment by exposure period.
47

Table 9: Congenital Hypothyroidism Cases Detected by Selected Demographic & Perinatal Characteristics & by Dried Blood
Spot Testing Method, Michigan Newborn Screening, 1/1/1994-6/30/2010
Tandem T4 and TSH, No Serial
TSH, Plus
T4 backup TSH
TSH
Testing
Serial Testing
Population Segment
p-value
(1/1994(1/1998(10/2003(3/200712/1997)
9/2003)
2/2007)
6/2010)
N
%
N
%
N
%
N
%
153
65.38
401
71.35
130
67.01
154
67.84
White
Race
0.3515
Black
52
22.22
84
14.95
36
18.56
41
18.06
Other
29
12.39
77
13.7
28
14.43
32
14.1
Female
119
55.61
265
46.9
122
54.22
143
55.21
Sex
0.0399
Male
95
44.39
300
53.1
103
45.78
116
44.79
No
223
92.53
563
94.78
206
91.56
239
92.28
Multiple Birth
0.2777
Yes
18
7.47
31
5.22
19
8.44
20
7.72
<750
6
2.94
23
4.04
5
2.28
10
3.89
750-1,499g
15
7.35
18
3.16
8
3.65
26
10.12
Birth Weight
0.001
1,500-2,499g
29
14.22
55
9.65
33
15.07
30
11.67
>2,500g
154
75.49
474
83.16
173
79
191
74.32
<28 weeks
5
2.34
16
6.43
Gestational Age
<.0001
28-37 weeks
40
18.69
51
20.48
>37 weeks
169
78.97
182
73.09
No
142
82.56
159
79.1
SGA
0.43
Yes
30
17.44
42
20.9
No
226
93.78
518
87.21
171
76
182
70.27
NICU Admission
<.0001
Yes
15
6.22
76
12.79
54
24
77
29.73
Total
241
18.27
594
45.03
225
17.06
259
19.64
Note: Missing data area as follows: race, n=111; sex, n= 68; birth weight, n= 70; gestational age, n= 303. Percentages reported are column
based. P-values represent the test for differences in the proportion of infants in each population segment by exposure period. Fisher test
used when >35% of cell counts were below 5.
48

Table 10: Congenital Hypothyroidism Cases Detected by Newborn Screening Result & by Dried Blood Spot Testing Method,
Michigan Newborn Screening, 1/1/1994-6/30/2010
T4 backup TSH
Population Segment

Tandem T4 and
TSH

TSH, No Serial
Testing

TSH, Plus Serial
Testing

(1/1994-12/1997)
N
%

(1/1998-9/2003)
N
%

(10/2003-2/2007)
N
%

(3/2007-6/2010)
N
%

p-value

Initial T4 Screen Result
63.14
437
74.32
0.0014
>5 mug/dl
149
<5 mug/d
87
36.86
151
25.68
Initial TSH Screen Result
97
40.25
298
50.17
57
25.33
95
36.68
<.0001
<50 uIU/mL
144
59.75
>50 uIU/mL*
296
49.83
168
74.67
164
63.32
Confirmatory Testing Results
125
77.64
375
80.47
47
22.27
72
29.15
<.0001
Serum TSH < 20 mU/L
Serum TSH > 20 mU/L
36
22.36
91
19.53
164
77.73
175
70.85
108
54.55
Serum FT4 < 1pmol/L
22
14.47
82
17.98
108
54.55
<.0001
Serum FT4 > 1 pmol/L
130
85.53
374
82.02
90
45.45
90
45
Thyroid Imaging Result
53
21.90
60
10.1
28
12.44
36
13.9
Abnormal
<.0001
Normal
11
4.50
9
1.52
5
2.22
2
0.77
Not Imaged
177
73.40
525
88.38
192
85.33
221
85.33
Central Hypothyroidism
<.0001
No
240
99.60
588
99.00
225
100
259
100
Yes
1
0.40
6
1.00
0
0
0
0
Total (row % reported)
241
18.27
594
45.03
225
17.06
259
19.64
Note: missing data are as follows: tsh, n=11; t4, n=11; serum free T4, n=278; serum TSH, n=235. P-values represent the test for
differences in the proportion of infants in each population segment by exposure period. Fisher test used when >35% of cell counts were
below 5.
49

Table 11: Newborn Screening Results & Performance Metrics, Michigan, 1/1/19946/30/2010
TSH,
T4
Tandem
TSH, No
Newborn Screening Results/
Plus
backup
T4 and
Serial
Performance Metrics
Serial
TSH
TSH
Testing
Testing
True Positives (N)
241
594
225
259
False Positive (N)
10,213
33,578
3,376
2,116
True Negative (N)
532,491
720,550
429,014
393,088
False Negative (N)
0
0
0
2
Population Screened (N)
542,945
754,722
432,615
395,463
Detection Rate
1: 2253
1: 1271
1: 1923
1: 1527
Positive Predictive Value (%)
2.31
1.74
6.25
10.91
Sensitivity (%)
100
100
100
99.20
Specificity (%)
98.12
95.55
99.22
99.46
False Positive Rate (%)
1.88
4.45
0.78
0.54

50

Table 12: Likelihood of Congenital Hypothyroidism Detection and False Positive Screening Determination by Dried Blood Spot
Testing Protocol, Michigan Newborn Screening, 1/1/1994-6/30/2010
CH Detection
False Positive Screen
Population Segment
Crude
Adjusted
Crude
Adjusted
OR LCL UCL
OR LCL UCL
OR
LCL
UCL
OR
LCL UCL
Serial Testing
T4 backup TSH
Tandem T4 and TSH
1.77
1.53
2.06
1.89
1.61
2.22
2.43
2.38
2.48
2.85
2.78
2.92
TSH, No Serial Testing
1.17
0.98
1.41
1.19
0.98
1.46
0.41
0.40
0.43
0.48
0.46
0.50
TSH, Plus Serial Testing
1.48
1.24
1.76
1.58
1.30
1.91
0.28
0.27
0.29
0.33
0.32
0.35
Race
White
Black
0.97
0.83
1.12
0.86
0.73
1.00
1.54
1.50
1.57
1.37
1.34
1.40
Other
1.62
1.37
1.91
1.53
1.28
1.82
0.91
0.88
0.94
0.87
0.84
0.91
Sex
Male
Female
1.12
1.00
1.25
1.09
0.97
1.22
0.78
0.76
0.79
0.76
0.74
0.77
Multiple Birth
No
Yes
2.08
1.68
2.58
1.23
0.96
1.57
2.10
2.03
2.18
1.09
1.04
1.14
Birth Weight
<750
9.52
7.04
12.89
5.81
6.33
12.26
12.04
11.41
12.70
10.47
9.87
11.11
750-1,499g
5.48
4.28
7.02
5.81
4.46
7.56
8.38
8.06
8.70
8.48
8.13
8.85
1,500-2,499g
2.08
1.75
2.47
2.08
1.72
2.52
2.24
2.17
2.30
2.17
2.10
2.24
>=2,500g
Note: Missing data are as follows: race, n= 140,620; sex, n=22,140; birth weight, n=31,786.

51

CHAPTER 4: IMPACT OF SERIAL TESTING ON CONGENITAL
HYPOTHYROIDISM NEWBORN SCREENING PERFORMANCE METRICS AMONG
MICHIGAN RESIDENT BIRTHS WEIGHING <1,800g, 1/1/2004-6/30/2010

Results:
Slightly more than 17,000 infants born weighing <1,800g were screened by primary TSH
testing for CH in Michigan during 1/1/2004-6/30/2010; half were exposed to a serial testing
designed to detect CH among children characterized by later rising TSH concentrations. The
distribution of selected demographic and perinatal characteristics did not statistically differ among
the population of infants screened, those screened positive for CH, or among those diagnosed with
CH following the implementation of the serial testing protocol. (Tables 13 & 14)
[Insert Tables 13 & 14]
While the characteristics of cases did not appreciably differ following the implementation of
serial testing, the number of cases detected increased significantly, as did the number of false
positive determinations. Table 16 reports the screening performance metrics before and after
implementation of the serial testing protocol. [Table 15] Following implementation of serial testing,
the detection rate nearly tripled and the FPR nearly doubled.
[Insert Table 15]
The proportions of children born weighing <1,500g and of those having gestational lengths
<28 weeks were greater among cases detected by a retest compared to those detected on the first
screen. [Table 16] The proportion of multiple births was also greater among cases detected by retest
compared to those detected on the initial screen. Otherwise, cases detected by retest did not
appreciably differ compared to those detected by the first screen.
[Insert Table 16]
52

Overall, the likelihood of CH detection increased nearly threefold (OR 2.77; CI 1.5-5.1)
following the implementation of serial testing; the rate of detection by retest increased by 11
fold.[Table 17] Retest detection was significantly more likely among multiparous births. While, as
expected, cases detected by retest had lesser initial TSH concentrations compared to those detected
on the first screen, they did not differ in regard to pre-treatment serum TSH and free T4
concentrations. Female cases were less likely to be detected by retest, although the association was
not statistically significant. The false positive rate increased by nearly 80% after the implementation
of serial testing; LBW/premature infants and those born SGA were most likely to receive a false
positive screening result compared to infants of normal growth or gestational length. The increase in
false positive determinations accordingly resulted in reduced specificity during the serial testing
period.
[Insert Table 17]
Limitations:
We suggest interpreting our findings per Hispanic ethnicity with caution as the data may
have been under-recorded. Race may also be mis-recorded as it is often assigned rather than
obtained by parental report. Due to the absence of thyroid scans via ultrasound or scintigraphy for
all cases we are unable to report our results by type of thyroid abnormality. We observed an extreme
rise in CH birth prevalence in the first half of 2010; we suggest this rise be interpreted with caution
as it is unexplained and may not be comparable to birth prevalence estimates obtained over the
course of an entire year. Gestational age may have been misclassified as few infants are expected to
be born weighing less than 1,800g at gestational ages beyond 37 weeks. Our study is also impacted
by missing race, sex and gestational age data; although missing data appears to occur at random.

53

Discussion:
Our findings indicate that serial NBS for CH increased detection by nearly threefold among
Michigan resident infants born weighing <1800g. Overall, three of five cases were detected by retest
during the serial testing period observed in this study. The detection rate observed among infants
born weighing < 1,500g observed in this study (1:182) is nearly 80% greater than reported by
Mandel et al (1:324) and more than twofold greater than reported by Larson et al. (1:400).[21, 22]
Accordingly, our findings suggest that the impact of serial NBS on detection in a program
characterized by earlier initial specimen collection (24-36hrs. of life) is greater than reported among
programs characterized by later specimen collection (>36hrs. of life). While serial testing
significantly increased detection, the FPR also increased by nearly 80%, although this amounted to
only an additional 125 false positive screening determinations representing a negligible increase in
NBS program operating costs.
While this and other studies support the necessity of serial testing for primary TSH testing
programs, others have suggested that serial testing may be unnecessary. Vincent et al. analyzed data
from the Province of Quebec from 1993-1994 when repeat NBS was conducted for all VLBW
infants and concluded that serial testing was unnecessary based on their finding of a single case of
late rising TSH identified in an infant that was not later diagnosed as CH.[19] However, Vincent et
al.’s study was likely under-powered to detect CH among children having a late rise in TSH and their
findings were impacted by delayed initiation of NBS. Korada et al. suggest that repeat NBS for CH
may be unnecessary if the initial TSH cutoff value were reduced to 6mU/L; although their findings
are impacted by later specimen collection and were also associated with a 126% increase in false
positives.[190] Had Michigan employed a 6mU/L initial cutoff rather than serial testing, 30% of the
cases detected by retest in this study would have been missed meaning our findings indicate that

54

serial testing is preferable to lowering of initial test cutoff values in NBS for CH among
premature/LBW infants.

55

Table 13: Infants Screened, Screened Positive & Diagnosed by Selected Demographic Characteristics among Children Born
Weighing <1,800g Before and After Implementation of Serial Testing, 1/1/2004-6/30/2010
Screened
Positives screens
Detected Cases
ppp-value
Pre
Post
Pre
Post
Pre
Post
Population
value
value
Segment
N
%
N
%
N
%
N
%
N
%
N
%
Race
White
4430 55.0 4459
54.7 0.89
86 58.9 157
55.5
10 66.7 28
65.1
0.37
0.50
Black
2933 36.4 3001
36.8
53 36.3 102
36.0
3 20.0 13
30.2
Other
685
8.5
692
8.5
7
4.8
24
8.5
2 13.3
2
4.7
Ethnicity
0.94
0.28
Hispanic
450
5.3
431
5.0 0.52
7
4.4
14
4.6
1
5.9
0
0.0
Non-Hispanic 8119 94.8 8130
95.0
152 95.6 294
95.5
16 94.1 44 100.0
Sex
0.40
0.65
Female
4279 50.8 4259
50.3 0.54
80 51.0 141
46.8
9 52.9 20
46.5
Male
4148 49.2 4207
49.7
77 49.0 160
53.2
8 47.1 23
53.5
Multiple Birth
0.78
0.12
No
5962 69.6 6112
71.4 0.01
116 73.0 221
71.8
14 82.4 27
61.4
Yes
2607 30.4 2449
28.6
43 27.0
87
28.3
3 17.7 17
38.6
Total
8569 100 8561 49.48
159 100 308 100.0
17 100 44
100
Note: Missing data are as follows- race, n= 930; sex, n=237. . P-values represent the test for differences in the proportion of infants in
each population segment by exposure period. Fisher test used when >35% of cell counts were below 5.

56

Table 14 (Table 14 continued): Infants Screened, Screened Positive & Diagnosed by Selected Demographic Characteristics
among Children Born Weighing <1,800g Before and After Implementation of Serial Testing, 1/1/2004-6/30/2010
Screened
Positives screens
Detected Cases
pppPre
Post
Pre
Post
Pre
Post
Population
value
value
value
Segment
N
%
N
%
N
%
N
%
N
%
N
%
Birth Weight
<750
1111 13.0 1137 13.3
32 20.1
72 23.4
5
29.4 10 22.7 0.439
0.67
0.58
750-1,499g
4313 50.3 4333 50.6
94 59.1 182 59.1
7
41.2 26 59.1
1,500-1,799g
3145 36.7 3091 36.1
33 20.8
54 17.5
5
29.4
8 18.2
Gestational Age
<28 weeks
2057 25.2 2119 25.6
48 32.0 115 38.7
5
29.1 16 37.2
0.74
0.32
0.5806
28-36 weeks
5735 70.2 5752 65.6
98 65.3 172 57.9
11
64.7 26 60.5
>37 weeks
383
4.7
394
4.8
4
2.7
10
3.4
1
5.9
1
2.3
SGA
0.78
0.92
0.41
No
5098 70.6 5165 70.8
83 62.4 159 61.9
10
71.4 23 59.0
Yes
2120 29.4 2126 29.2
50 37.6
98 38.1
4
28.6 16 41.0
NICU
0.03
0.21
0.53
No
545
6.4
478
5.6
11
6.9
13
4.2
0
0
1
2.3
Yes
8024 93.6 8083 94.4
148 93.1 295 95.8
17 100.0 43 97.7
Total
8569
100 8561
159
100 308 100
17
100 44
100
Note: Missing data are as follows- gestational age, n=690; SGA, n=2617. P-values represent the test for differences in the proportion of
infants in each population segment by exposure period. Fisher test used when >35% of cell counts were below 5.

57

Table 15: Newborn Screening Performance Metrics Before & After Implementation of Serial
Testing among Infants Born Weighing <1,800g, Michigan, 1/1/2004-6/30/2010
Exposure
Detection Positive
False Positive Sensitivity Specificity
Period
Rate
Predictive Value
Rate
Pre- Serial
Testing
1:571
9.68%
1.60%
100%
98.40%
Post-Serial
Testing
1:203
13.64%
3.10%
100%
96.90%

58

Table 16: Characteristics of Congenital Hypothyroidism Cases Detected by Initial vs.
Retest Dried Blood Spot Screen, Michigan Newborn Screening, 1/1/2004-6/30/2010
Detected by 1st Screen
Detected by 2nd Screen
Population Segment
p-value
N
C%
R%
N
C%
R%
Repeat NBS Protocol
Pre
14
56.0
82.4
3
8.3
17.7 <.0001
Post
11
44.0
25.0
33
91.7
75.0
Race
White
17
70.8
44.7
21
61.8
55.3
0.6207
Black
5
20.8
31.3
11
32.4
68.8
Other
2
8.3
50.0
2
5.9
50.0
Sex
0.4603
Female
13
54.2
44.8
16
44.4
55.2
Male*
11
45.8
35.5
20
55.6
64.5
Multiple Birth
0.0199
No
21
84.0
51.2
20
55.6
48.8
Yes
4
16.0
20.0
16
44.4
80.0
Gestational Length
<28 weeks
6
24.0
28.6
15
42.9
71.4
0.3104
28-37 weeks
18
72.0
48.7
19
54.3
51.4
>37 weeks
1
4.0
50.0
1
2.9
50.0
Birth Weight
<750
6
24.0
40.0
9
25.0
60.0
0.5525
750-1,499g
12
48.0
36.4
21
58.3
63.6
1,500-1,799g
7
28.0
53.9
6
16.7
46.2
SGA
0.5922
No
14
66.7
42.4
19
59.4
57.6
Yes
7
33.3
35.0
13
40.6
65.0
NICU Admission
0.4098
No
1
4.0
100
0
0.0
0.0
Yes
24
36.0
40.0
36
100.0
60.0
Initial Screen
TSH <50 uIU/mL
16
64.0
30.8
36
100.0
69.2 <.0001
TSH >50 uIU/mL
9
36.0
100
0
0.0
0.0
Confirmatory Testing
0.407
Serum TSH < 20 mU/L
11
50.0
44.0
14
38.9
56.0
Serum TSH > 20 mU/L
11
50.0
33.3
22
61.1
66.7
Serum FT4 < 1pmol/L
14
66.7
46.7
16
53.3
53.3
0.341
Serum FT4 > 1 pmol/L
7
33.3
33.3
14
46.7
66.7
Total
25
100
41.98
36
100
59.02
Note: Missing=SGA, n=8; GA, n=1; sex, n=1; race, n=3; serum TSH, 3; serum free T4, n=10. Pvalues represent the test for differences in the proportions by first & retest detection.
59

Table 17: Likelihood of Overall Congenital Hypothyroidism Detection, Detection by Retest
& False Positive Determination Following Implementation of Serial Testing, Michigan
Newborn Screening, 1/1/2004-6/30/2010
Population
Segment

Overall Detection
OR

LCL

UCL

Retest Detection
OR

LCL

UCL

False Positive
OR

LCL

UCL

Serial Testing
No
Yes
2.77
1.50
5.10
11.05
3.39
36.03
1.78
1.42
2.24
Race
White
Black
0.63
0.35
1.13
0.78
0.38
1.63
0.95
0.76
1.19
Other
0.68
0.24
1.90
0.61
0.14
2.62
0.64
0.16
2.60
Ethnicity
Non-Hispanic
Hispanic
0.31
0.04
2.22
0.53
0.07
3.84
0.98
0.46
2.08
Sex
Male
Female
0.92
0.55
1.52
0.78
0.41
1.51
0.97
0.78
1.21
Multiple Birth
No
Yes
1.17
0.68
1.99
1.91
0.99
3.69
0.81
0.63
1.04
Birth Weight
<750
3.22
1.53
6.77
4.18
1.49
11.76
3.53
2.50
5.00
750-1,499g
1.83
0.97
3.49
2.53
1.02
6.27
2.41
1.80
3.24
1,500-1,799g
Gestational Age
<28 weeks
1.56
0.91
2.68
2.18
1.11
4.29
1.73
1.37
2.17
28-37 weeks
>37 weeks
0.80
0.19
3.32
0.78
0.10
5.82
0.61
0.30
1.24
SGA
No
Yes
1.47
0.84
2.56
1.48
0.72
3.03
1.49
1.19
1.87
Note: Missing data are as follows: race, n= 930; sex, n=237; gestational age, n=690; SGA, n=2621.

60

CHAPTER 5: RESULTS OF FOLLOW-UP OF CONGENITAL HYPOTHYROIDISM
CASES DETECTED BY MICHIGAN NEWBORN SCREENING AFTER AGE 3 YEARS

Results:
Follow-up after age three years was attempted for 152 CH cases detected by Michigan NBS
in 10/1/2003-12/31/2007; follow-up was attempted for only 3 cases detected in 2006 because
survey activities stopped temporarily while transitioning responsibilities from the NBS Follow-up
Program to the Endocrine Follow-up Program. As indicated in Figure 4, 68 (44.7%) cases were
LTFU, diagnostic re-evaluation was in process for 8 (5.3%), and diagnostic re-evaluation was not
completed, in-process or planned for unknown reasons for 4 (2.6%) cases, leaving a final study
population of 72 cases having undergone some degree of diagnostic re-evaluation to determine
whether the child had transient or permanent CH.
Cases that were LTFU were marginally more likely to be born LBW or premature, admitted
to the NICU at birth, and were more likely to have pre-treatment serum FT4 values < 1pmol/L;
although none of the differences reached statistical significance given an alpha of 5%. (Tables 18 &
19) The majority of cases successfully followed-up were white (67%), singleton births (95%), born
NBW (90%) and/or at full term (82%), were normal for gestational age at birth (90%), and were not
admitted to the NICU at birth. Slightly more males (52%) were followed-up compared to females
(48%). Few followed-up cases underwent thyroid imaging (13%); of those that underwent thyroid
imaging (n=11), nine exhibited a thyroid abnormality. A minority of followed cases had initial TSH
concentrations measured in dried blood spots <50 uIU/mL (27%), or pre-treatment serum TSH
concentrations < 20 mU/L. Nearly half (47%) of the cases followed-up had pre-treatment serum
FT4 concentrations < 1pmol/L.
[Insert Tables 18 & 19]

61

Figure 4: Three Year Follow-up Study Flow Diagram
As indicated in Figure 4, of the 72 cases followed-up and known to have undergone some
form of diagnostic re-evaluation, 34 (47%) were deemed ineligible for a trial off thyroid hormone
supplements and were considered to have permanent CH. Of the 34 cases deemed ineligible for a
trial off thyroid hormone supplements, reasons for ineligibility reported by clinicians included that
thyroid imaging depicted permanent CH (n=7, 21%) and that the child’s thyroid hormone
supplement dosage had increased over time (n=27, 79%) representing a permanent need for
treatment. Of note, 3 of the 27 children whose diagnosis was considered permanent based on an
increase in thyroid hormone supplement dosage had their dosages increased at ages 5 weeks, 6 and 9
months; age at time of thyroid hormone supplement dosage increase was not reported for the
remaining 24 cases deemed permanent CH by this criterion. While the overall number of cases
62

confirmed permanent based on abnormal thyroid imagine results was small (n=7), they were more
likely to be female (57%), whereas cases confirmed permanent CH based on a medication increase
were more likely to be male (63%). (Table 20)
A trial during which patients were either not taking thyroid hormone supplements or taking
them at a lower dosage was conducted among 38 (53%) cases. Of those having undergone a trial off
therapy, nearly 40% (n=15) were conducted absent medical supervision when children’s families
decided to no longer provide treatment. Unsupervised cessation of thyroid hormone
supplementation was more likely among racial minorities, cases born LBW, and cases admitted to
the NICU at birth compared to the distribution of these characteristics in the overall population of
cases re-evaluated.
[Insert Table 20]
While all cases whose family stopped medication of their own accord were reported to have
normal lab values off treatment, thyroid hormone supplementation was re-instated for 20 of the 23
cases that underwent a trial off therapy under medical supervision.(Table 21) Normal lab values were
transmitted to the NBS Follow-up Program confirming transient CH for 3 of 15 cases that
underwent a trial off therapy absent medical supervision. Treatment was more likely to be resumed
among white cases, those born LBW and/or premature, and those that were admitted to the NICU
after birth compared to the distribution of these characteristics in the overall population of cases reevaluated; the association between reinstatement of treatment and birth weight was the only to reach
statistical significance. The rate of treatment reinstatement did not vary significantly by NBS results.
(Table 22)
[Insert Tables 21 & 22]
Table 23 reports the odds of treatment cessation among all cases that underwent diagnostic
re-evaluation, including those deemed ineligible for a trial off treatment based on medication dosage
63

increase or abnormal thyroid imaging results. The crude odds of treatment cessation at diagnostic reevaluation was elevated among non-white cases, those born LBW, and those admitted to the NICU
after birth relative to cases of white race, those born NBW, and those not admitted to the NICU
respectively. After adjustment, the only factor significantly predictive of treatment cessation was
black race (OR: 9.86; CI: 1.82*53.31) Clinically, the likelihood of treatment cessation was appreciably
elevated among LBW cases and those admitted to the NICU at birth.
[Insert Table 23]
Limitations:
This study is limited by significant LTFU; approximately half of the cases for whom followup was attempted were unable to be located. Given the rarity of CH and the characteristically low
response rate associated with physician surveys, this study represents a major success in public
health programming considering that more than half of the targeted cases were successfully
followed-up and that the final study population did not statistically differ from the population of
cases LTFU. However, it is noteworthy that cases that were LTFU may have been more likely to
stop treatment, thus creating a selection bias.
Our results are impacted by missing data. The dearth of thyroid imaging results particularly
impacted our ability to predict transient CH among treated cases. Survey respondents were also
unlikely to report exact dates and lab values or specify their method of diagnostic re-evaluation
and/or follow-up regiment. Due to the survey design, our results are potentially subject to response
bias. Race information was also missing for an appreciable number of records. However, missing
information did not appear to be systematically distributed and thus should not have lead to
differential misclassification. Results pertaining to Hispanic ethnicity should be interpreted with
caution because this factor is often reported or collected incorrectly.

64

Discussion:
One of four CH cases detected by Michigan NBS that were followed-up and underwent any
form of diagnostic re-evaluation is no longer receiving thyroid hormone medication, indicating that
while the rate of treatment for CH is 1:1,527 births screened, the birth prevalence of CH
maintaining a sustained need for treatment after age three years is closer to 1:2,035 births screened.
However, our study may have underestimated the proportion of cases no longer requiring treatment
after age three years for the following reasons. First, children that were LTFU may have been more
likely to stop treatment. Second, children that underwent a trial off treatment under medical
supervision may have had insufficient time for their thyroid hormones to normalize and thus could
have been inappropriately categorized as permanent. Alternatively, discrepant results of trials off
therapy conducted with and without medical supervision could be explained by the fact that cases
more likely to have transient CH were also more likely to stop taking their medication without
medical supervision in which case our estimate the proportion of cases having transient disease
would be internally valid. Third and finally, 27 cases were considered as having permanent CH based
on an increase in thyroid hormone medication dosage, at times occurring as early as 5 weeks to 6
months of age, some of which may not still require medication.
Our study also found that one of five cases re-evaluated had stopped treatment without
medical supervision. While normal lab values were reported for children removed from treatment by
their families, this represents a significant threat to public health given that poor treatment
compliance is a predictor of brain damage. Cases that were removed from treatment by their families
without medical supervision did not differ in terms of their NBS and confirmatory testing results; it
is possible that some cases LTFU may have stopped treatment without medical supervision suffered
poor outcomes not observed in this study.

65

Due to differences in study design, our findings are not directly comparable to prior
investigations of CH medication cessation after age three years. Generally, we observed a smaller
proportion of cases that no longer required treatment after age three years compared to Kemper et
al.’s claims-based analysis and Eugster et al.’s clinic based follow-up study of 33 CH euthyroid cases
which estimated that approximately 38% and 36% of CH cases are no longer treated after age three
years respectively.[68, 191] However, among cases who underwent a trial off therapy in our study,
nearly half did not resume treatment, greater than the rate observed among euthyroid cases included
in Eugster et al.’s study.

66

Table 18: Proportion of Congenital Hypothyroidism Cases Detected by Michigan Newborn
Screening from 10/1/2003-12/31/2007 Followed-up at or Beyond Age Three Years by
Selected Demographic/Perinatal Characteristics
Lost to FollowFollowed-up
up
Population Segment
p-value
N
N
Race
White
46
64.65
56
66.67
0.34
Black
9
13.24
17
20.24
Other
13
19.12
11
13.10
Sex
0.35
Female
37
55.22
40
47.62
Male
30
44.78
44
52.38
Multiple Birth
0.69^
No
66
97.06
80
95.24
Yes
2
2.94
4
4.76
Birth Weight
<750
4
5.97
1
1.23
0.06^
750-1,499g
6
8.96
2
2.47
1,500-2,499g
7
10.45
5
6.17
>2,500g
50
74.63
73
90.12
Gestational Length
<28 weeks
10
14.71
7
8.33
0.08
28-37 weeks
13
19.12
8
9.52
>37 weeks
45
66.18
69
82.14
Small for Gestational Age
0.36
No
43
84.31
62
89.86
Yes
8
15.69
7
10.14
NICU Admission
No
48
70.59
70
83.33
0.06
Yes
20
29.41
14
16.67
Total
68
100
84
100
Note: ^Fisher Exact Test Used. Missing data are as follows: Sex (n=1), Birth Weight (n=4), Small
for Gestational Age (n=32). P-values represent the test for differences in the proportion of infants
in each population segment by follow-up.

67

Table 19: Proportion of Congenital Hypothyroidism Cases Detected by Michigan Newborn
Screening from 2004-2007 Followed-up at or Beyond Age Three Years by Newborn
Screening Result
LTFU
Followed
pPopulation Segment
value
N
%
N
%
Dried Blood Spot TSH
TSH <50 uIU/mL

24

35.29

23

27.38

TSH >50 uIU/mL

44

64.71

61

72.62

22
45

32.84
67.16

29
51

36
63.75

39

63.93

37

47.44

Serum TSH
Serum TSH < 20 mU/L
Serum TSH > 20 mU/L
Serum FT4
< 1pmol/L

0.29

0.66

0.053
> 1 pmol/L
22
32.35
41
48.81
Thyroid Image Result
Normal
1
1.47
2
2.38
Abnormal
12
17.65
9
10.71
0.47
No thyroid image
55
80.88
73
86.90
Total
68
100
84
100.00
Note: Missing data are as follows: TSH, n=3; Serum TSH, n=5; Serum FT4, n=13. P-values
represent the test for differences in the proportion of infants in each population segment by followup.

68

Table 20: Method of Trial Off Thyroid Hormone Medication or Reason for Lack of Trial,
Congenital Hypothyroidism Cases Followed-up after Age 3 Years, Michigan Newborn
Screening
Diagnosis
Trial Off Medication
Reason for Lack of Trial
RePopulation
Patient
Physician
Abnormal
Medication
Evaluated
Initiated
Initiated
Thyroid Scan
Increase
Segment
N
%
N
%
N
%
N
%
N
%
Race
White
48
66.7
5
33.3
17
73.9
3
42.9
23
85.2
Black
15
20.8
7
46.7
4
17.4
2
28.6
2
7.4
Other
9
12.5
3
20.0
2
8.7
2
28.6
2
7.4
Sex
Female
38
52.8
10
66.7
14
60.9
4
57.1
10
37.0
Male
34
47.2
5
33.3
9
39.1
3
42.9
17
63.0
Twin
No
69
95.8
14
93.3
22
95.7
7
100.0
26
96.3
Yes
3
4.2
1
6.7
1
4.3
0
0.0
1
3.7
Birth Weight
<750
1
1.4
1
6.7
0
0.0
0
0.0
0
0.0
750-1,499g
1
1.4
1
6.7
0
0.0
0
0.0
0
0.0
1,5002,499g
4
5.7
0
0.0
2
9.1
0
0.0
2
7.7
>2,500g
64
91.4
13
86.7
20
90.9
7
100.0
24
92.3
Gest. Age
<28 weeks
5
6.9
1
6.7
3
13.0
0
0.0
1
3.7
28-37
weeks
8
11.1
2
13.3
1
4.3
1
14.3
4
14.8
>37
weeks
59
81.9
12
80.0
19
82.6
6
85.7
22
81.5
SGA
No
55
91.7
11
91.7
16
84.2
5
100.0
23
95.8
Yes
5
8.3
1
8.3
3
15.8
0
0.0
1
4.2
LBW
No
64
88.9
13
86.7
20
90.9
7
100.0
24
92.3
Yes
6
8.3
2
13.3
2
9.1
0
0.0
2
7.7
NICU
No
61
84.7
10
66.7
20
87.0
7
100.0
24
88.9
Yes
11
15.3
5
33.3
3
13.0
0
0.0
3
11.1
Total
72
100
15
100
23
100
7
100
27
100
Note: Missing data are as follows: Sex (n=8), Birth Weight (n=4), Small for Gestational Age
(n=32).

69

Table 21: Results of Trial Off Thyroid Hormone Medication among Congenital
Hypothyroidism Cases Followed-up After Age 3 Years by Selected Demographic &
Perinatal Characteristics, Michigan Newborn Screening
Underwent
Treatment
Treatment Not
Trial
Resumed
Resumed
Population Segment
P-value
N
%
N
%
N
%
Race
White
22
57.9
15
75
7
38.9
0.10
Black
11
28.9
3
15
8
44.4
Other
5
13.2
2
10
3
16.7
Sex
0.67
Female
24
63.2
12
60
12
66.7
Male
14
36.8
8
40
6
33.3
Multiple Birth
0.22
No
36
94.7
20
100
16
88.9
Yes
2
5.3
0
0
2
11.1
Gestational Age
<28 weeks
4
10.5
3
15
1
5.6
0.70
28-37 weeks
3
7.9
1
5
2
11.1
>37 weeks
31
81.6
16
80
15
83.3
SGA
0.33
No
27
87.1
15
93.8
12
80
Yes
4
12.9
1
6.3
3
20
LBW
0.046
No
33
89.2
19
100
14
77.8
Yes
4
10.8
0
0
4
22.2
NICU
0.12
No
30
78.9
18
90
12
66.7
Yes
8
21.1
2
10
6
33.3
Total
38
100
20
100
18
100
Note: Missing data are as follows: Birth Weight (n=4), Small for Gestational Age (n=7). P-values
represent the test for differences in the proportion of infants in each population segment by whether
treatment was resumed.

70

Table 22: Results of Trial Off Thyroid Hormone Medication among Congenital
Hypothyroidism Cases Followed-up After Age 3 Years by Newborn Screening &
Confirmatory Testing Results, Michigan Newborn Screening
Underwent
Treatment
Treatment
Trial
Resumed
Not
Population Segment
Resumed
N

%

N

%

N

%

pvalue

Dried Blood Spot Screen Result
0.21
TSH <50 uIU/mL
13
34.2
5
38.46
8
61.54
TSH >50 uIU/mL
25
65.8 15
60.00 10
40.00
Confirmatory Testing Results
0.88
Serum TSH < 20 mU/L
19
51.4 10
52.63
9
47.37
Serum TSH > 20 mU/L
18
48.6
9
50.00
9
50.00
Serum FT4 < 1pmol/L
10
27.0
5
50.00
5
50.00
0.92
Serum FT4 > 1 pmol/L
27
73.0 14
51.85 13
48.15
Image result- Eutopic Thyroid
1
2.6
1 100.00
0
0.00
0.63
Image result- Thyroid Dysgenesis
2
5.3
1
50.00
1
50.00
No thyroid image
35
92.1 18
51.43 17
48.57
Total
38
100 20
52.63 18
47.4
Note: Missing data are as follows: TSH, n=3; Serum TSH, n=1; Serum FT4, n=1. P-values
represent the test for differences in the proportion of infants in each population segment by whether
treatment was resumed.

71

Table 23: Predictors of Transient Determination among Children Diagnosed with
Congenital Hypothyroidism and Followed-up by Michigan Newborn Screening after Age 3
years, 2004-2007
Diagnosis
Transient CH
RePopulation Segment
Evaluated
Crude
Adjusted
N
R%
OR
LCL UCL
OR
LCL UCL
Race
White
48
85.71
Black
15
88.24 6.69
1.84 24.39 9.86
1.82 53.31
Other
9
81.82 2.93
0.59 14.52 5.10
0.86 30.19
Sex
Female
38
95.00 2.15
0.71
6.57
Male
34
77.27
Gestational Length
<28 weeks
5
71.43 0.73
0.08
7.09
28-37 weeks
8
100.00 0.98
0.18
5.38
>37 weeks
59
85.51
LBW
No
64
88.90
Yes
6
8.30 7.14
1.18 43.11 2.77
0.18 42.26
NICU Admission
No
61
87.14
Yes
11
78.57 4.70
1.22 18.05 5.20
0.78 34.71
Dried Blood Spot Screen
Result
TSH <50 uIU/mL
18
78.26 3.36
1.06 10.69 1.35
0.30
6.02
TSH >50 uIU/mL
54
88.52
Confirmatory Testing
Results
Serum TSH < 20 mU/L
24
80.00 2.50
0.82
7.61
Serum TSH > 20 mU/L
45
88.24
Serum FT4 < 1pmol/L
37
100.00
Serum FT4 > 1 pmol/L
37
90.24 2.60
0.80
8.48
Total
72
85.71
Note: Missing data are as follows: Sex (n=8), Birth Weight (n=4), Small for Gestational Age
(n=32).

72

Chapter 6: Conclusions and Future Directions

Comparison of Dried Blood Spot Testing Methods
Our findings suggest that different cases are detected by primary T4 compared to primary
TSH approaches to NBS for CH. While each test independently detected an equivalent number of
cases, when they are combined they detect a significantly greater number of cases, many more by the
TSH screen compared to the T4 screen. It appears that each test (T4 and TSH) falsely screens
different infants positive based on the same evidence; while the FPR is comparable between each
test method independently, when they are combined, the FPR increases significantly. To make a
final conclusion about whether the additional false positive determinations generated by tandem T4
and TSH testing is acceptable in the context of increased detection, guidelines detailing a benchmark
balance between sensitivity and specificity are needed. Alternatively, the overall mission of NBS may
need to be revisited to determine the ideal approach to testing.
If the goal of NBS is to detect all cases and ensure prompt provision of treatment, then
tandem T4 & TSH screening is the ideal approach; it detects a greater number of cases, including
infants having central CH. If the mission of NBS is to detect the most cases as efficiently as possible in
the context of financial resources, then primary TSH testing plus serial testing for infants at risk of
later rising TSH may be the ideal approach. If the latter is true, then cost-benefit analyses are
necessary to support NBS operational decisions. This research would need to determine the number
of false positives acceptable to detect an additional case of CH; costs saved by early detection of
potentially ‘borderline’ cases that would be picked up as sensitivity increases as well those related to
early detection of central CH would need to be estimated. Regarding central CH, our findings do not
support the contention that it is detected clinically absent NBS considering that no central CH cases
were detected after T4 testing stopped and that the two false negative results observed in our study
73

were primary CH. Thus, the estimated costs savings per detection of central CH would likely be
significant.
In determining the most cost-effective means of NBS for CH, it may be beneficial to also
investigate a combined approach including aspects of tandem TSH and T4 in addition to serial
testing to maximize detection while conserving the number of false positive determinations.
Additionally, further efforts are necessary to identify the ideal target population for serial testing and
similarly, the acceptable balance between false and true positives. Future studies should also
determine whether serial testing increases detection among primary T4 or tandem T4 and TSH test
programs.
Need for Standardization & Guidelines
To make future research more meaningful, further effort is necessary to standardize the
process of diagnosing and classifying CH, similar to efforts done in developing cerebral palsy
surveillance systems in Europe. Otherwise, absent a consistent and operational case definition, it is
difficult to meaningfully compare screening performance metrics both between and within NBS
programs over time. Guidelines are necessary to:
•

illustrate the number and type of tests needed to make the initial CH diagnosis, particularly
thyroid imaging by ultrasound and/or radioisotope scanning;

•

recommend cutoffs for confirmatory analyte testing adjusted for the time of the test and
other pertinent characteristics associated with potentially immature thyroid development or
function;

•

standardize the nomenclature used to describe CH, ensuring that treated cases are
consistently sub-classified into primary CH, central CH, hyperthyrotropinemia,
hypothyroxinemia of prematurity and/or potentially transient CH in accordance with
specified ranges of thyroid hormone concentrations and thyroid imaging results;
74

•

recommend follow-up regiments adjusted to the type of case identified including those
focused towards potentially transient CH; and

•

detail specifically when and how to confirm the diagnosis of CH after age three years.
Few cases in this study underwent any form of thyroid imaging. While it is true that thyroid

imaging does not augment the course of treatment for CH, it does allow for more precision in
predicting whether a child will have a transient or permanent need for treatment. If parents were
informed that their child may or may not require lifelong treatment, it would perhaps reduce the
possibility of them taking their children off treatment without medical supervision. Those informed
that their child may be able to stop treatment later on under medical supervision would be more
inclined to follow-up with their physician to determine when/how they should stop treatment.
Those informed that their child would likely require lifelong treatment would better understand that
they are not to stop treatment without medical supervision. In short, thyroid imaging would provide
a better opportunity to discuss prognosis and emphasize the importance of treatment compliance
and diagnostic re-evaluation after three years of age.
Greater clarity about other types of tests and how to interpret their results coupled with an
improved nomenclature of CH would further allow for standardization between and within NBS
programs, yet would preserve the opportunity of clinicians to initiate treatment based on clinical
judgment alone if necessary. If classic primary CH were defined in operational terms and that
diagnosis were applied consistently across programs, inter-program and intra-program variation over
time would be minimized and more meaningful comparisons could be made. Improved classification
would not preclude clinicians from rendering treatment based solely on clinical judgment; however,
it would allow for such decisions to be monitored and evaluated apart from treatment of classic
primary CH. For instance, it may be correct that children with marginal T4 concentrations and
elevated TSH concentrations benefit from thyroid hormone treatment; although, if this is the case
75

then it should be demonstrated and such cases should be classified separate from classic CH as not
to distort birth prevalence estimates. In short, the current process of treating infants for CH based
primarily on clinical judgment has led to considerable and unexplained variation which precludes
investigators from effectively learning from comparison of NBS performance data.
Greater specificity in CH typology would not only allow a better understanding of the
tradeoffs of different approaches to screening and whether we’re truly observing a rising birth
prevalence as recently reported, it would also allow us to investigate different follow-up regiments
for cases classified as more likely to be transient CH. Several of the cases determined transient at
follow-up stopped treatment before age three. Accordingly, for particular cases, it may be possible to
conduct diagnostic re-evaluation earlier. However, further research is also necessary to identify the
optimal time for diagnostic re-evaluation, which cases should be re-evaluated using a trial off thyroid
hormone supplements, and whether and when treatment should be resumed for cases undergoing a
trial off thyroid hormone supplements.
While current AAP guidelines recommend therapy cessation after age three years if no
permanent etiology was found and no increase in TSH was observed after the newborn period, the
method of determining ‘permanent etiology’ and separately what constitutes important changes in
TSH in terms of degree and timing need further clarification.[27] Similarly, when to reinstate
treatment in conducting a trial off therapy should be operationally defined. There is no clear
standard for conducting treatment cessation trials or interpreting their results to confirm
permanence of CH. The duration of treatment cessation varies from 3-4 weeks to 10 months among
the few published studies and the frequency and timing of follow-up thereafter among children
whose thyroid function normalizes after cessation of thyroid hormone therapy also varies.[68, 128,
139] In our study, all children removed from treatment by their families were reported to have
normal thyroid function off treatment, yet virtually all re-evaluated by a physician resumed treatment
76

suggesting that clinicians may prematurely reinstate treatment, although without a clear guideline
detailing when to reinstate treatment the determination of permanent CH is again made solely based
on clinical judgment which is subject to significant variation. Our findings also indicate that long
term follow-up should be conducted for all CH cases, not only those eligible for a trial off therapy,
given that one of five cases followed-up after age three years in our study had been removed from
treatment absent medical supervision.
Conclusion
In sum, our results indicate that NBS programs should consider tandem T4 and TSH testing,
or primary TSH plus serial testing for infants at risk of later rising TSH. A benchmark balance
between sensitivity and specificity is necessary to make the decision between tandem TSH and T4
testing and primary TSH plus serial testing methods. Further efforts are necessary to standardize the
process of diagnosis and terminologies used to characterize cases across NBS programs; guidelines
are necessary to support this effort.

77

APPENDIX

78

APPENDIX- APPROVALS FOR REPRINTS

Approval for Table 1:
BioMed Central Open Access Charter states:
The author(s) or copyright owner(s) irrevocably grant(s) to any third party, in advance and in
perpetuity, the right to use, reproduce or disseminate the research article in its entirety or in part, in
any format or medium, provided that no substantive errors are introduced in the process, proper
attribution of authorship and correct citation details are given, and that the bibliographic details are
not changed.

79

Approval for Figure 2 and Table 3:

80

Approval for Table 4:
From: Ballen, Karen [mailto:KBallen@liebertpub.com]
Sent: Monday, April 11, 2011 11:08 AM
To: Steven Korzeniewski
Subject: RE: reprint inquiry
Dear Steven:
Copyright permission is granted free of charge.
Kind regards,
Karen Ballen
Manager, Reprints and Permissions
-----Original Message----From: Steven Korzeniewski [mailto:skorzeniewski@mpro.org]
Sent: Friday, April 08, 2011 4:07 PM
To: Ballen, Karen
Subject: reprint inquiry
Hello,
I would like to include a table from Mandel, S. J., R. J. Hermos, et al. (2000). "Atypical
hypothyroidism and the very low birthweight infant." Thyroid. 10(8): 693-695 in my dissertation and
notice that there is no request option for inclusion in dissertations. May I use the table of interest at
no charge, or does it actually cost $87?
I look forward to your reply, thank you.

Steven J. Korzeniewski, PhD-Candidate, MSc, MA,
Director, Statistical Analysis Resource Group (SARG),
Chief Scientific Officer,
MPRO- Michigan's Quality Improvement Organization
22670 Haggerty Rd Ste. 100, Farmington Hills, MI 48335
Telephone: (248) 465-7365
Email: skorzeniewski@mpro.org

81

Approval for Table 5:
License Number
2644570211342
License date
Apr 08, 2011
Licensed content
Elsevier
publisher
Licensed content
Endocrinology and Metabolism Clinics of North
publication
America
Licensed content title
Update: Newborn Screening for Endocrinopathies
Licensed content author Kenneth A. Pass, Eurico Carmago Neto
Licensed content date
December 2009
Licensed content volume 38
number
Licensed content issue
4
number
Number of pages
11
Type of Use
reuse in a thesis/dissertation
Portion
figures/tables/illustrations
Number of
1
figures/tables/illustrations
Format
both print and electronic
Are you the author of this No
Elsevier article?
Will you be translating?
No
Order reference number
Title of your
IMPACT OF NEWBORN SCREENING FOR
thesis/dissertation
CONGENITAL HYPOTHYROIDISM PROTOCOL
CHANGES ON SCREENING PERFORMANCE
METRICS IN MICHIGAN: SUCCESSES,
OPPORTUNITIES FOR QUALITY IMPROVEMENT
AND PUBLIC HEALTH IMPLICATIONS
Expected completion date Apr 2011
Estimated size (number 110
of pages)
Elsevier VAT number
GB 494 6272 12
Permissions price
0.00 USD
VAT/Local Sales Tax
0.0 USD / 0.0 GBP
Total
0.00 USD

82

BIBLIOGRAPHY

83

BIBLIOGRAPHY

1.

Alm, J., A. Larsson, and R. ZetterstrÖM, CONGENITAL HYPOTHYROIDISM IN
SWEDEN Incidence and Age at Diagnosis. Acta Pædiatrica, 1978. 67(1): p. 1-3.

2.

Klein, A.H., A.V. Agustin, and T. Foley, SUCCESSFUL LABORATORY
SCREENING FOR CONGENITAL HYPOTHYROIDISM. The Lancet, 1974.
304(7872): p. 77-79.

3.

Dussault, J.H., et al., Preliminary report on a mass screening program for
neonatal hypothyroidism. The Journal of pediatrics, 1975. 86(5): p. 670-674.

4.

Klein, A.H. and T.P. Foley, Screening for hypothyroidism. The Journal of
pediatrics, 1975. 87(4): p. 667-668.

5.

LaFranchi, S.H., et al., Neonatal hypothyroidism detected by the Northwest
Regional Screening Program. Pediatrics., 1979. 63(2): p. 180-91.

6.

Fisher, D.A., et al., Recommendations for screening programs for congenital
hypothyroidism. Report of a committee of the American Thyroid Association. Am
J Med., 1976. 61(6): p. 932-4.

7.

Delange, F., Neonatal screening for congenital hypothyroidism: Results and
perspectives. Hormone Research, 1997. 48(2): p. 51-61.

8.

Therrell, B.L., et al., Newborn Screening System Performance Evaluation
Assessment Scheme (PEAS). Seminars in Perinatology, 2010. 34(2): p. 105-120.

9.

Horn, S. and H. Heuer, Thyroid hormone action during brain development: More
questions than answers. Molecular and Cellular Endocrinology, 2010. 315(1-2):
p. 19-26.

10.

Zoeller, R.T., S.W. Tan, and R.W. Tyl, General background on the hypothalamicpituitary-thyroid (HPT) axis. Critical Reviews in Toxicology, 2007. 37(1-2): p. 1153.

11.

The Thyroid Gland. Handbook of Clinical Pediatric Endocrinology. 2008:
Blackwell Publishing Ltd. 84-98.

12.

Jain, V., et al., Congenital Hypothyroidism. 2008, All India Institute of Medical
Sciences: New Delhi.

13.

Lifshitz, F., Pediatric Endocrinology. Clinical Pediatrics. Vol. 8. 1996, New York:
Marcel Dekker, Inc.

14.

Hyman, S.J., et al., Late rise of thyroid stimulating hormone in ill newborns.
Journal of Pediatric Endocrinology & Metabolism, 2007. 20(4): p. 501-510.
84

15.

Rapaport, R., Evaluation of thyroid status of infants in intensive care settings:
Recommended an extension of newborn screening. Journal of Pediatrics, 2003.
143(5): p. 556-558.

16.

Tylek-Lemanska, D., M. Kumorowicz-Kopiec, and J. Starzyk, Screening for
congenital hypothyroidism: the value of retesting after four weeks in neonates
with low and very low birth weight. Journal of Medical Screening, 2005. 12(4): p.
166-169.

17.

Paul, D.A., et al., Low serum thyroxine on initial newborn screening is associated
with intraventricular hemorrhage and death in very low birth weight infants.
Pediatrics, 1998. 101(5): p. 903-907.

18.

VAN WASSENAER, A.G., et al., Thyroid Function in Very Preterm Infants:
Influences of Gestational Age and Disease. Pediatric Research, 1997. 42(5): p.
604-609.

19.

Frank, J.E., et al., Thyroid function in very low birth weight infants: effects on
neonatal hypothyroidism screening. J Pediatr., 1996. 128(4): p. 548-54.

20.

Hunter, M.K., et al., Follow-up of newborns with low thyroxine and nonelevated
thyroid-stimulating hormone-screening concentrations: Results of the 20-year
experience in the Northwest Regional Newborn Screening Program. Journal of
Pediatrics, 1998. 132(1): p. 70-74.

21.

Larson, C., et al., Risk factors associated with delayed thyrotropin elevations in
congenital hypothyroidism. Journal of Pediatrics, 2003. 143(5): p. 587-591.

22.

Mandel, S.J., et al., Atypical hypothyroidism and the very low birthweight infant.
Thyroid., 2000. 10(8): p. 693-5.

23.

Brown, R.S. and L.A. Demmer, The Etiology of Thyroid Dysgenesis--Still an
Enigma after All These Years. J Clin Endocrinol Metab, 2002. 87(9): p. 40694071.

24.

Rastogi, M.V. and S.H. LaFranchi, Congenital hypothyroidism. Orphanet Journal
of Rare Diseases, 2010. 5.

25.

Zung, A., et al., Neonatal hyperthyrotropinemia: population characteristics,
diagnosis, management and outcome after cessation of therapy. Clinical
Endocrinology, 2010. 72(2): p. 264-271.

26.

Parks, J.S., et al., The Impact of Transient Hypothyroidism on the Increasing
Rate of Congenital Hypothyroidism in the United States. Pediatrics, 2010. 125: p.
S54-S63.

27.

Rose, S.R., et al., Update of newborn screening and therapy for congenital
hypothyroidism. Pediatrics, 2006. 117(6): p. 2290-2303.
85

28.

Forghani, N. and T. Aye, Hypothyroxinemia and Prematurity. Neoreviews, 2008.
9(2): p. e66-71.

29.

Berbel, P., et al., Iodine supplementation during pregnancy: a public health
challenge. Trends in Endocrinology & Metabolism, 2007. 18(9): p. 338-343.

30.

Zimmermann, M.B., Iodine Deficiency. Endocrine Reviews, 2009. 30(4): p. 376408.

31.

Pass, K.A. and E.C. Neto, Update: Newborn Screening for Endocrinopathies.
Endocrinology and Metabolism Clinics of North America, 2009. 38(4): p. 827-+.

32.

Moreno, J.C. and T.J. Visser, New phenotypes in thyroid dyshormonogenesis:
hypothyroidism due to DUOX2 mutations. Endocr Dev., 2007. 10: p. 99-117.

33.

Brown, R.S., et al., Incidence of transient congenital hypothyroidism due to
maternal thyrotropin receptor-blocking antibodies in over one million babies.
Journal of Clinical Endocrinology & Metabolism, 1996. 81(3): p. 1147-1151.

34.

Calaciura, F., et al., Subclinical hypothyroidism in early childhood: A frequent
outcome of transient neonatal hyperthyrotropinemia. Journal of Clinical
Endocrinology & Metabolism, 2002. 87(7): p. 3209-3214.

35.

Filippi, L., et al., Dopamine infusion: A possible cause of undiagnosed congenital
hypothyroidism in preterm infants. Pediatric Critical Care Medicine, 2006. 7(3): p.
249-251.

36.

Filippi, L., et al., Dopamine infusion and hypothyroxinaemia in very low birth
weight preterm infants. European Journal of Pediatrics, 2004. 163(1): p. 7-13.

37.

Porterfield, S.P., Thyroidal dysfunction and environmental chemicals - Potential
impact on brain development. Environmental Health Perspectives, 2000. 108: p.
433-438.

38.

Dezegher, F., et al., DOPAMINE SUPPRESSES THYROID-STIMULATING
HORMONE-SECRETION IN NEONATAL-HYPOTHYROIDISM. Acta Paediatrica,
1995. 84(2): p. 213-214.

39.

Miller, M.D., et al., Thyroid-Disrupting Chemicals: Interpreting Upstream
Biomarkers of Adverse Outcomes. Environmental Health Perspectives, 2009.
117(7): p. 1033-1041.

40.

Kirk, A.B., Environmental perchlorate: Why it matters. Analytica Chimica Acta,
2006. 567(1): p. 4-12.

41.

Buffler, P.A., et al., Thyroid function and perchlorate in drinking water: An
evaluation among California newborns, 1998. Environmental Health
Perspectives, 2006. 114(5): p. 798-804.
86

42.

Nedveckaite, T., et al., Environmental releases of radioactivity and the incidence
of thyroid disease at the Ignalina Nuclear Power Plant. Health Physics, 2000.
79(6): p. 666-674.

43.

Porterfield, S.P., VULNERABILITY OF THE DEVELOPING BRAIN TO THYROID
ABNORMALITIES - ENVIRONMENTAL INSULTS TO THE THYROID SYSTEM.
Environmental Health Perspectives, 1994. 102: p. 125-130.

44.

La Franchi, S., Hypothyroidism. Pediatric Clinics of North America, 1979. 26(1):
p. 33-51.

45.

Grant, D.B., et al., Congenital hypothyroidism detected by neonatal screening:
relationship between biochemical severity and early clinical features. Arch Dis
Child., 1992. 67(1): p. 87-90.

46.

Medda, E., et al., Risk factors for congenital hypothyroidism: results of a
population case-control study (1997-2003). Eur J Endocrinol, 2005. 153(6): p.
765-773.

47.

Fernboff, P.M., Congenital hypothyroidism and associated birth defects:
Implications for investigators and clinicians. Journal of Pediatrics, 1998. 132(4):
p. 573-574.

48.

Roberts, H.E., et al., Population study of congenital hypothyroidism and
associated birth defects, Atlanta, 1979-1992. American Journal of Medical
Genetics, 1997. 71(1): p. 29-32.

49.

AlJurayyan, N.A.M., et al., Congenital anomalies in infants with congenital
hypothyroidism: Is it a coincidental or an associated finding? Human Heredity,
1997. 47(1): p. 33-37.

50.

de Papendieck, L.G., et al., Thyroid dysfunction and high thyroid stimulating
hormone levels in children with Down's syndrome. Journal of Pediatric
Endocrinology & Metabolism, 2002. 15(9): p. 1543-1548.

51.

Tuysuz, B. and D.B. Beker, Thyroid dysfunction in children with Down's
syndrome. Acta Paediatrica, 2001. 90(12): p. 1389-1393.

52.

Jaruratanasirikul, S., N. Patarakijvanich, and C. Patanapisarnsak, The
association of congenital hypothyroidism and congenital gastrointestinal
anomalies in Down's syndrome infants. Journal of Pediatric Endocrinology &
Metabolism, 1998. 11(2): p. 241-246.

53.

Raiti, S. and G.H. Newns, Cretinism. Archives of Disease in Childhood, 1971.
46(249): p. 692-694.

87

54.

Klein, A.H., S. Meltzer, and F.M. Kenny, Improved prognosis in congenital
bypothyroidism treated before age three months. The Journal of pediatrics, 1972.
81(5): p. 912-915.

55.

Jacobsen, B.B. and N.J. Brandt, Congenital hypothyroidism in Denmark.
Archives of Disease in Childhood, 1981. 56(2): p. 134-136.

56.

Dussault, J.H., The Anecdotal History of Screening for Congenital
Hypothyroidism. J Clin Endocrinol Metab, 1999. 84(12): p. 4332-4334.

57.

Dussault, J.H. and C. Laberge, [Thyroxine (T4) determination by
radioimmunological method in dried blood eluate: new diagnostic method of
neonatal hypothyroidism?]. Union Med Can., 1973. 102(10): p. 2062-4.

58.

Dussault, J.H., et al., TSH measurements from blood spots on filter paper: A
confirmatory screening test for neonatal hypothyroidism. The Journal of
pediatrics, 1976. 89(4): p. 550-552.

59.

Dussault, J., et al., Thyroxine-binding globulin capacity and concentration
evaluated from blood spots on filter-paper in a screening program for neonatal
hypothyroidism. Clin Chem, 1980. 26(3): p. 463-465.

60.

Fisher, D.A., et al., Screening for congenital hypothyroidism: Results of screening
one million North American infants. The Journal of pediatrics, 1979. 94(5): p.
700-705.

61.

Illig, R., T. Torresani, and B. Sobradillo, Early detection of neonatal
hypothyroidism by serial TSH determination in dried blood. Six months
experience with a reliable, efficient and inexpensive method. Helv Paediatr Acta.,
1977. 32(4-5): p. 289-97.

62.

Delange, F., et al., Serum thyrotrophin determination on day 5 of life as
screening procedure for congenital hypothyroidism. Arch Dis Child., 1977. 52(2):
p. 89-96.

63.

MIYAI, K., et al., A New Method of Paired Thyrotropin Assay as a Screening Test
for Neonatal Hypothyroidism. J Clin Endocrinol Metab, 1978. 47(5): p. 10281033.

64.

The Endocrine Problems of Infancy. Handbook of Clinical Pediatric
Endocrinology. 2008: Blackwell Publishing Ltd. 14-32.

65.

Pass, K.A., et al., US newborn screening system guidelines II: follow-up of
children, diagnosis, management, and evaluation. Statement of the Council of
Regional Networks for Genetic Services (CORN). J Pediatr., 2000. 137(4 Suppl):
p. S1-46.

88

66.

Jones, J.H., et al., Improvement in screening performance and diagnosis of
congenital hypothyroidism in Scotland 1979-2003. Archives of Disease in
Childhood, 2006. 91(8): p. 680-685.

67.

Perry, R.J., et al., Combined ultrasound and isotope scanning is more informative
in the diagnosis of congenital hypothyroidism than single scanning. Archives of
Disease in Childhood, 2006. 91(12): p. 972-976.

68.

Eugster, E.A., et al., Definitive diagnosis in children with congenital
hypothyroidism. Journal of Pediatrics, 2004. 144(5): p. 643-647.

69.

Smith, L., Updated AAP guidelines on newborn screening and therapy for
congenital hypothyroidism. American Family Physician, 2007. 76(3): p. 439-+.

70.

La Gamma, E.F., et al., Phase 1 Trial of 4 Thyroid Hormone Regimens for
Transient Hypothyroxinemia in Neonates of < 28 Weeks' Gestation. Pediatrics,
2009. 124(2): p. E258-E268.

71.

Green, N.S. and K.A. Pass, Neonatal screening by DNA microarray: spots and
chips. Nat Rev Genet, 2005. 6(2): p. 147-151.

72.

Lafranchi, S., et al., NEWBORN SCREENING FOR CONGENITAL
HYPOTHYROIDISM - RECOMMENDED GUIDELINES. Thyroid, 1993. 3(3): p.
257-263.

73.

DeLong, G.R., J.B. Stanbury, and R. Fierro-Benitez, Neurological signs in
congenital iodine-deficiency disorder (endemic cretinism). Dev Med Child
Neurol., 1985. 27(3): p. 317-24.

74.

Gruters, A., A. Jenner, and H. Krude, Long-term consequences of congenital
hypothyroidism in the era of screening programmes. Best Practice & Research
Clinical Endocrinology & Metabolism, 2002. 16(2): p. 369-382.

75.

Huffmeier, U., H.U. Tietze, and A. Rauch, Severe skeletal dysplasia caused by
undiagnosed hypothyroidism. European Journal of Medical Genetics, 2007.
50(3): p. 209-215.

76.

Delvecchio, M., et al., Longitudinal assessment of levo-thyroxine therapy for
congenital hypothyroidism: Relationship with aetiology, bone maturation and
biochemical features. Hormone Research, 2007. 68(3): p. 105-112.

77.

Delvecchio, M., et al., Factors predicting final height in early treated congenital
hypothyroid patients. Clinical Endocrinology, 2006. 65(5): p. 693-697.

78.

Niu, D.M., et al., Contributions of bone maturation measurements to the
differential diagnosis of neonatal transient hypothyroidism versus
dyshormonogenetic congenital hypothyroidism. Acta Paediatrica, 2004. 93(10):
p. 1301-1306.
89

79.

Van Vliet, G., Neonatal hypothyroidism: Treatment and outcome. Thyroid, 1999.
9(1): p. 79-84.

80.

Verrotti, A., et al., Bone metabolism in children with congenital hypothyroidism A longitudinal study. Journal of Pediatric Endocrinology & Metabolism, 1998.
11(6): p. 699-705.

81.

Grosse, S.D. and G.V. Vliet, Prevention of intellectual disability through
screening for congenital hypothyroidism: how much and at what level? Archives
of Disease in Childhood, 2011.

82.

Frost, G.J. and J.M. Parkin, MANAGEMENT OF PATIENTS WITH
CONGENITAL HYPOTHYROIDISM. British Medical Journal, 1985. 290(6480): p.
1485-1489.

83.

Hulse, A., CONGENITAL HYPOTHYROIDISM AND NEUROLOGICAL
DEVELOPMENT. Journal of Child Psychology and Psychiatry and Allied
Disciplines, 1983. 24(4): p. 629-635.

84.

Grant, D.B., et al., PSYCHOMOTOR DEVELOPMENT IN INFANTS WITH
CONGENITAL HYPOTHYROIDISM DIAGNOSED BY NEONATAL SCREENING.
Acta Medica Austriaca, 1992. 19: p. 54-56.

85.

Herbstman, J., et al., Maternal, infant, and delivery factors associated with
neonatal thyroid hormone status. Thyroid, 2008. 18(1): p. 67-76.

86.

Peter, F., Thyroid dysfunction in the offspring of mothers with autoimmune
thyroid diseases. Acta Paediatrica, 2005. 94(8): p. 1008-1010.

87.

Evans, C., et al., Potent thyrotrophin receptor-blocking antibodies: a cause of
transient congenital hypothyroidism and delayed thyroid development. European
Journal of Endocrinology, 2004. 150(3): p. 265-268.

88.

de Escobar, G.M., M.J. Obregon, and F.E. del Rey, Maternal thyroid hormones
early in pregnancy and fetal brain development. Best Practice & Research
Clinical Endocrinology & Metabolism, 2004. 18(2): p. 225-248.

89.

Smit, B.J., et al., Neurologic development of the newborn and young child in
relation to maternal thyroid function. Acta Paediatrica, 2000. 89(3): p. 291-295.

90.

Dussault, J.H. and D.A. Fisher, Thyroid function in mothers of hypothyroid
newborns. Obstetrics and Gynecology, 1999. 93(1): p. 15-20.

91.

New England Congenital Hypothyroidism Collaborative, Elementary school
performance of children with congenital hypothyroidism. The Journal of
pediatrics, 1990. 116(1): p. 27-32.

90

92.

Simons, W.F., et al., Educational progress, behaviour, and motor skills at 10
years in early treated congenital hypothyroidism. Archives of Disease in
Childhood, 1997. 77(3): p. 219-222.

93.

Simoneau-Roy, J., et al., Cognition and behavior at school entry in children with
congenital hypothyroidism treated early with high-dose levothyroxine. The
Journal of pediatrics, 2004. 144(6): p. 747-752.

94.

Bargagna, S., et al., School attainments in children with congenital
hypothyroidism detected by neonatal screening and treated early in life.
European Journal of Endocrinology, 1999. 140(5): p. 407-413.

95.

Bongers-Schokking, J.J., et al., Influence of timing and dose of thyroid hormone
replacement on development in infants with congenital hypothyroidism. The
Journal of pediatrics, 2000. 136(3): p. 292-297.

96.

Dimitropoulos, A., et al., Children With Congenital Hypothyroidism: Long-Term
Intellectual Outcome After Early High-Dose Treatment. Pediatric Research, 2009.
65(2): p. 242-248.

97.

Rovet, J.F., Congenital hypothyroidism: Long-term outcome. Thyroid, 1999. 9(7):
p. 741-748.

98.

Song, S.I., D. Daneman, and J. Rovet, The influence of etiology and treatment
factors on intellectual outcome in congenital hypothyroidism. Journal of
Developmental and Behavioral Pediatrics, 2001. 22(6): p. 376-384.

99.

Toublanc, J.E., Guidelines for neonatal screening programs for congenital
hypothyroidism. Working Group for Neonatal Screening in Paediatric
Endocrinology of the European Society for Paediatric Endocrinology. Acta
Paediatr Suppl., 1999. 88(432): p. 13-4.

100.

Gunn, A.J., M. Wake, and W.S. Cutfield, High and low dose initial thyroxine
therapy for congenital hypothyroidism. Journal of Paediatrics and Child Health,
1996. 32(3): p. 242-5.

101.

Salerno, M., et al., Effect of different starting doses of levothyroxine on growth
and intellectual outcome at four years of age in congenital hypothyroidism.
Thyroid., 2002. 12(1): p. 45-52.

102.

Fisher, D.A., Second International Conference on Neonatal Thyroid Screening:
progress report. J Pediatr., 1983. 102(5): p. 653-4.

103.

Lafranchi, S., CONGENITAL HYPOTHYROIDISM - A NEWBORN SCREENING
SUCCESS STORY. Endocrinologist, 1994. 4(6): p. 477-486.

104.

Klett, M., Epidemiology of congenital hypothyroidism. Experimental and Clinical
Endocrinology & Diabetes, 1997. 105: p. 19-23.
91

105.

LaFranchi, S., Congenital hypothyroidism: Etiologies, diagnosis, and
management. Thyroid, 1999. 9(7): p. 735-740.

106.

Harris, K.B. and K.A. Pass, Increase in congenital hypothyroidism in New York
State and in the United States. Molecular Genetics and Metabolism, 2007. 91(3):
p. 268-277.

107.

Rendon-Macias, M.E., et al., Birth prevalence of congenital hypothyroidism in
Mexico. Paediatric and Perinatal Epidemiology, 2008. 22(5): p. 478-485.

108.

Deladoey, J., N. Belanger, and G. Van Vliet, Random variability in congenital
hypothyroidism from thyroid dysgenesis over 16 years in Quebec. Journal of
Clinical Endocrinology & Metabolism, 2007. 92(8): p. 3158-3161.

109.

Mikelsaar, R.V., et al., Neonatal screening for congenital hypothyroidism in
Estonia. Journal of Medical Screening, 1998. 5(1): p. 20-21.

110.

Ray, M., et al., Audit of screening programme for congenital hypothyroidism in
Scotland 1979-93. Archives of Disease in Childhood, 1997. 76(5): p. 411-415.

111.

AlJurayyan, N.A.M., et al., Neonatal screening for congenital hypothyroidism in
Saudi Arabia: Results of screening the first 1 million newborns. Screening, 1996.
4(4): p. 213-220.

112.

AlJurayyan, N.A.M., et al., Congenital hypothyroidism: Increased incidence in
Najran Province, Saudi Arabia. Journal of Tropical Pediatrics, 1996. 42(6): p.
348-351.

113.

Fernandeziglesias, C., et al., NEONATAL SCREENING FOR
PHENYLKETONURIA AND CONGENITAL HYPOTHYROIDISM IN
PRINCIPADO DE ASTURIAS (SPAIN) USING 2 TYPES OF BLOOD-SAMPLES.
Screening, 1995. 4(3): p. 131-138.

114.

Inoue, B., et al., RESULTS AND EVALUATION OF 12 YEARS OF NEWBORN
SCREENING FOR CONGENITAL HYPOTHYROIDISM IN JAPAN OKAYAMA
DISTRICT. Screening, 1992. 1(3): p. 203-209.

115.

Connelly, J.F., et al., Newborn screening for congenital hypothyroidism, Victoria,
Australia, 1977-1997. Part 1: The screening programme, demography, baseline
perinatal data and diagnostic classification. Journal of Pediatric Endocrinology &
Metabolism, 2001. 14(9): p. 1597-1610.

116.

Skordis, N., et al., High prevalence of congenital hypothyroidism in the Greek
Cypriot population: Results of the neonatal screening program 1990-2000.
Journal of Pediatric Endocrinology & Metabolism, 2005. 18(5): p. 453-461.

117.

Stranieri, I. and O.A. Takano, Evaluation of the Neonatal Screening Program for
congenital hypothyroidism and phenylketonuria in the State of Mato Grosso,
92

Brazil. Arquivos Brasileiros De Endocrinologia E Metabologia, 2009. 53(4): p.
446-452.
118.

Ordookhani, A., et al., A high prevalence of consanguineous and severe
congenital hypothyroidism in an Iranian population. Journal of Pediatric
Endocrinology & Metabolism, 2004. 17(9): p. 1201-1209.

119.

Shapira, S.K., M.A. Lloyd-Puryear, and C. Boyle, Future Research Directions to
Identify Causes of the Increasing Incidence Rate of Congenital Hypothyroidism in
the United States. Pediatrics, 2010. 125: p. S64-S68.

120.

Olney, R.S., S.D. Grosse, and R.F. Vogt, Prevalence of Congenital
Hypothyroidism-Current Trends and Future Directions: Workshop Summary.
Pediatrics, 2010. 125: p. S31-S36.

121.

Hertzberg, V., J. Mei, and B.L. Therrell, Effect of Laboratory Practices on the
Incidence Rate of Congenital Hypothyroidism. Pediatrics, 2010. 125: p. S48-S53.

122.

Hinton, C.F., et al., Trends in Incidence Rates of Congenital Hypothyroidism
Related to Select Demographic Factors: Data From the United States, California,
Massachusetts, New York, and Texas. Pediatrics, 2010. 125: p. S37-S47.

123.

van Tijn, D.A., et al., Neonatal detection of congenital hypothyroidism of central
origin. Journal of Clinical Endocrinology & Metabolism, 2005. 90(6): p. 33503359.

124.

Kempers, M.J.E., et al., Neonatal screening for congenital hypothyroidism based
on thyroxine, thyrotropin, and thyroxine-binding globulin measurement: Potentials
and pitfalls. Journal of Clinical Endocrinology & Metabolism, 2006. 91(9): p.
3370-3376.

125.

National Newborn Screening Information System. 2006 [cited 2011 1/16/2011];
Available from:
http://nnsis.uthscsa.edu/xreports.aspx?XREPORTID=103&FORMID=71&FCLR=
1.

126.

Kempers, M.J.E., et al., Neonatal screening for congenital hypothyroidism in The
Netherlands: Cognitive and motor outcome at 10 years of age. Journal of Clinical
Endocrinology & Metabolism, 2007. 92(3): p. 919-924.

127.

Coakley, J.C., et al., Transient primary hypothyroidism in the newborn:
Experience of the Victorian Neonatal Thyroid Screening Programme. Journal of
Paediatrics and Child Health, 1989. 25(1): p. 25-30.

128.

Davy, T., et al., Congenital Hypothyroidism: The Effect of Stopping Treatment at
3 Years of Age. Am J Dis Child, 1985. 139(10): p. 1028-1030.

93

129.

Hashemipour, M., et al., Permanent and transient congenital hypothyroidism in
Isfahan-Iran. Journal of Medical Screening, 2009. 16(1): p. 11-16.

130.

Ordookhani, A., et al., Transient neonatal hypothyroidism is associated with
elevated serum anti-thyroglobulin antibody levels in newborns and their mothers.
Journal of Pediatrics, 2007. 150(3): p. 315-317.

131.

Gaudino, R., et al., Proportion of various types of thyroid disorders among
newborns with congenital hypothyroidism and normally located gland: a regional
cohort study. Clinical Endocrinology, 2005. 62(4): p. 444-448.

132.

Kemper, A.R., L.J. Ouyang, and S.D. Grosse, Discontinuation of thyroid hormone
treatment among children in the United States with congenital hypothyroidism:
findings from health insurance claims data. Bmc Pediatrics, 2010. 10.

133.

Corbetta, C., et al., A 7-year experience with low blood TSH cutoff levels for
neonatal screening reveals an unsuspected frequency of congenital
hypothyroidism (CH). Clinical Endocrinology, 2009. 71(5): p. 739-745.

134.

Costa, P., et al., [Re-evaluation of the diagnosis in congenital hypothyroidism].
Ann Ist Super Sanita., 1998. 34(3): p. 343-7.

135.

Weber, G., et al., Congenital hypothyroidism with gland in situ: Diagnostic reevaluation. Journal of Endocrinological Investigation, 2005. 28(6): p. 516-522.

136.

Nair, P., S. Sobhakumar, and L. Kailas, Diagnostic Re-evaluation of children with
congenital hypothyroidism. Indian Pediatrics, 2010. 47(9): p. 757-760.

137.

Tamam, M., et al., Diagnostic spectrum of congenital hypothyroidism in Turkish
children. Pediatrics International, 2009. 51(4): p. 464-468.

138.

Panoutsopoulos, G., et al., Scintigraphic evaluation of primary congenital
hypothyroidism: results of the Greek screening program. European Journal of
Nuclear Medicine, 2001. 28(4): p. 529-533.

139.

Yang, R.L., et al., Obervation time for drug administration and withdrawal in the
treatment of children with congenital hypothyroidism. Zhejiang Da Xue Xue Bao
Yi Xue Ban, 2007. 36(5): p. 493-497.

140.

Oakley, G.A., et al., Increased incidence of congenital malformations in children
with transient thyroid-stimulating hormone elevation on neonatal screening. The
Journal of pediatrics, 1998. 132(4): p. 726-730.

141.

Kurinczuk, J.J., et al., Congenital hypothyroidism in Western Australia 19811998. Journal of Paediatrics and Child Health, 2002. 38(2): p. 187-191.

142.

Therrell, B.L., U.S. Newborn Screening Policy Dilemmas for the Twenty-First
Century. Molecular Genetics and Metabolism, 2001. 74(1-2): p. 64-74.
94

143.

Kemper, A.R., et al., Long-term follow-up after diagnosis resulting from newborn
screening: Statement of the US Secretary of Health and Human Services'
Advisory Committee on Heritable Disorders and Genetic Diseases in Newborns
and Children. Genetics in Medicine, 2008. 10(4): p. 259-261
10.1097/GIM.0b013e31816b64f9.

144.

Slazyk, W.E. and W.H. Hannon, Quality assurance in the newborn screening
laboratory, in Laboratory Methods for Neonatal Screening, B. Therrell, Editor.
1993, American Public Health Association: Washington, DC. p. 23-46.

145.

Loeber, G., D. Webster, and A. Aznarez, Quality evaluation of newborn
screening programs. Acta Paediatrica, 1999. 88: p. 3-6.

146.

Buyukgebiz, A., Newborn screening for congenital hypothyroidism. Journal of
Pediatric Endocrinology & Metabolism, 2006. 19(11): p. 1291-1298.

147.

Dussault, J.H. and J. Morissette, Higher Sensitivity of Primary Thyrotropin in
Screening for Congenital Hypothyroidism: A Myth? J Clin Endocrinol Metab,
1983. 56(4): p. 849-852.

148.

Zamboni, G., et al., Diagnostic effectiveness of simultaneous thyroxine and
thyroid-stimulating hormone screening measurements. Thirteen years'
experience in the Northeast Italian Screening Programme. Journal of Medical
Screening, 2004. 11(1): p. 8-10.

149.

Wang, S.T., S. Pizzolato, and H.P. Demshar, Diagnostic effectiveness of TSH
screening and of T4 with secondary TSH screening for newborn congenital
hypothyroidism. Clinica Chimica Acta, 1998. 274(2): p. 151-158.

150.

Tuerck, J.M., et al., Newborn screening strategies: routine second tests —
Oregon, in Early hospital discharge: impact on newborn screening., K.A. Pass
and H.L. Levy, Editors. 1995, CORN: Atlanta, GA. p. 201-212.

151.

Therrell, B., Second testing in newborn screening program in the US, in Early
hospital discharge: impact on newborn screening., K.A. Pass and H.L. Levy,
Editors. 1995, CORN: Atlanta, GA. p. 75-86.

152.

Baumgartner, J.H., Screening for primary hypothyroidism in Missouri, using a
thyroxine and TSH assay for all specimens: a five year experience., in
Proceedings of the 11th National Neonatal Screening Symposium, B. Therrell
and B.G. Aldis, Editors. 1995: Corpus Christi, TX. p. 39-42.

153.

LaFranchi, S., Newborn screening strategies for congenital hypothyroidism: an
update. Journal of Inherited Metabolic Disease, 2010. 33(0): p. 225-233.

154.

CLSI, Newborn Screening for Preterm, Low Birth Weight, and Sick Newborns;
Approved Guideline., in CLSI document I/LA31-A. 2009, Clinical And Laboratory
Standards Institute: Wayne, PA.
95

155.

Silva, S.A.B., et al., Screening for Congenital Hypothyroidism in Extreme
Premature and/or Very Low Birth Weight Newborns: The Importance of a
Specific Protocol. Journal of Pediatric Endocrinology & Metabolism, 2010. 23(12): p. 45-52.

156.

Madison, L.D. and S. LaFranchi, Screening for congenital hypothyroidism:
current controversies. Current Opinion in Endocrinology, Diabetes and Obesity,
2005. 12(1): p. 36-41.

157.

Korada, S.M., et al., Difficulties in selecting an appropriate neonatal thyroid
stimulating hormone (TSH) screening threshold. Archives of Disease in
Childhood, 2010. 95(3): p. 169-U24.

158.

Mengreli, C., et al., Screening for Congenital Hypothyroidism: The Significance of
Threshold Limit in False-Negative Results. Journal of Clinical Endocrinology &
Metabolism, 2010. 95(9): p. 4283-4290.

159.

Harada, S., et al., LATER MANIFESTATIONS OF CONGENITAL
HYPOTHYROIDISM PREDICTED BY SLIGHTLY ELEVATED THYROTROPIN
LEVELS IN NEONATAL SCREENING. Screening, 1995. 3(4): p. 181-192.

160.

Chace, D.H., et al., Rapid metabolic and newborn screening of thyroxine (T-4)
from dried blood spots by MS/MS. Clinica Chimica Acta, 2009. 403(1-2): p. 178183.

161.

Miyai, K., et al., Evaluation of highly sensitive thyrotropin assay for detecting
thyroid diseases in neonatal screening: Preliminary studies. Endocrine Journal,
1998. 45(6): p. 761-766.

162.

Rapaport, R., et al., Time for Thyrotropin Releasing Hormone to Return to the
United States of America. Thyroid, 2010. 20(9): p. 947-948.

163.

Van Tijn, D.A., J.J.M. de VijIder, and T. Vulsma, Role of the thyrotropin-releasing
hormone stimulation test in diagnosis of congenital central hypothyroidism in
infants. Journal of Clinical Endocrinology & Metabolism, 2008. 93(2): p. 410-419.

164.

Krude, H. and O. Blankenstein, Treating patients not numbers: the benefit and
burden of lowering TSH newborn screening cut-offs. Archives of Disease in
Childhood, 2011. 96(2): p. 121-122.

165.

Robert, R., Thyroid function in the very low birth weight newborn: Rescreen or
reevaluate? The Journal of pediatrics, 2002. 140(3): p. 287-289.

166.

Alm, J., et al., INCIDENCE OF CONGENITAL HYPOTHYROIDISM RETROSPECTIVE STUDY OF NEONATAL LABORATORY SCREENING
VERSUS CLINICAL SYMPTOMS AS INDICATORS LEADING TO DIAGNOSIS.
British Medical Journal, 1984. 289(6453): p. 1171-1175.
96

167.

Azizi, F., et al., Effects of transient neonatal hyperthyrotropinemia on intellectual
quotient and psychomotor performance. International Journal for Vitamin and
Nutrition Research, 2001. 71(1): p. 70-73.

168.

Leonardi, D., et al., Longitudinal study of thyroid function in children with mild
hyperthyrotropinemia at neonatal screening for congenital hypothyroidism.
Journal of Clinical Endocrinology & Metabolism, 2008. 93(7): p. 2679-2685.

169.

McDermott, M.T. and E.C. Ridgway, Subclinical Hypothyroidism Is Mild Thyroid
Failure and Should be Treated. J Clin Endocrinol Metab, 2001. 86(10): p. 45854590.

170.

Tyfield, L.A., et al., Persistent hyperthyrotropinaemia since the neonatal period in
clinically euthyroid children. European Journal of Pediatrics, 1991. 150(5): p.
308-309.

171.

Miki, K., et al., Transient infantile hyperthyrotrophinaemia. Archives of Disease in
Childhood, 1989. 64(8): p. 1177-1182.

172.

Cody, D., et al., The differing outcomes of hyperthyrotropinaemia. Journal of
Pediatric Endocrinology & Metabolism, 2003. 16(3): p. 375-378.

173.

Kohler, B., et al., Transient congenital hypothyroidism and hyperthyrotropinemia:
Normal thyroid function and physical development at the ages of 6-14 years.
Journal of Clinical Endocrinology & Metabolism, 1996. 81(4): p. 1563-1567.

174.

Demirel, F., et al., l-thyroxin treatment in infants with hyperthyrotropinaemia: 4year experience. International Journal of Clinical Practice, 2007. 61(8): p. 13331336.

175.

Salerno, M., et al., Long-term cardiovascular effects of levothyroxine therapy in
young adults with congenital hypothyroidism. Journal of Clinical Endocrinology &
Metabolism, 2008. 93(7): p. 2486-2491.

176.

Reuss, M.L., et al., The Relation of Transient Hypothyroxinemia in Preterm
Infants to Neurologic Development at Two Years of Age. New England Journal of
Medicine, 1996. 334(13): p. 821-827.

177.

Ng, S.M., et al., TIPIT: A randomised controlled trial of thyroxine in preterm
infants under 28 weeks' gestation. Trials, 2008. 9.

178.

Bakker, B., et al., Dynamics of the plasma concentrations of TSH, FT4 and T3
following thyroxine supplementation in congenital hypothyroidism. Clinical
Endocrinology, 2002. 57(4): p. 529-537.

179.

FISHER, D.A. and B.L. FOLEY, Early Treatment of Congenital Hypothyroidism.
Pediatrics, 1989. 83(5): p. 785-789.

97

180.

Rovet, J. and R. Ehrlich, Long-term effects of L-thyroxine therapy for congenital
hypothyroidism. The Journal of pediatrics, 1995. 126(3): p. 380-386.

181.

Rovet, J. and M. Alvarez, Thyroid hormone and attention in congenital
hypothyroidism. Journal of Pediatric Endocrinology & Metabolism, 1996. 9(1): p.
63-66.

182.

Ng, S.M., D. Anand, and A.M. Weindling, High versus low dose of initial thyroid
hormone replacement for congenital hypothyroidism. Cochrane Database of
Systematic Reviews, 2009(1).

183.

Selva, K.A., et al., Initial treatment dose of L-thyroxine in congenital
hypothyroidism. Journal of Pediatrics, 2002. 141(6): p. 786-792.

184.

Selva, K.A., et al., Neurodevelopmental Outcomes in Congenital Hypothyroidism:
Comparison of Initial T4 Dose and Time to Reach Target T4 and TSH. The
Journal of pediatrics, 2005. 147(6): p. 775-780.

185.

Telfair, J., Long-term follow-up after diagnosis resulting from newborn screening:
Statement of the US secretary of Health and Human Services' Advisory
Committee on Heritable Disorders and Genetic Diseases in Newborns and
Children (vol 10, pg 259, 2008). Genetics in Medicine, 2008. 10(5): p. 368-368.

186.

Eugster, E.A., et al., Development of a congenital hypothyroidism follow-up
program - The Indiana experience. Endocrinologist, 2000. 10(3): p. 185-193.

187.

Kemper, A.R., et al., Primary care physicians' attitudes regarding follow-up care
for children with positive newborn screening results. Pediatrics, 2006. 118(5): p.
1836-1841.

188.

Korzeniewski, S.J., et al., Methodological innovations in data gathering: newborn
screening linkage with live births records, Michigan, 1/2007-3/2008. Matern Child
Health J, 2010. 14(3): p. 360-4.

189.

Alexander, G.R., M.D. Kogan, and J.H. Himes, 1994-1996 U.S. singleton birth
weight percentiles for gestational age by race, Hispanic origin, and gender.
Matern Child Health J., 1999. 3(4): p. 225-31.

190.

Korada, M., et al., Repeat testing for congenital hypothyroidism in preterm infants
is unnecessary with an appropriate thyroid stimulating hormone threshold.
Archives of Disease in Childhood-Fetal and Neonatal Edition, 2008. 93(4): p.
F286-F288.

191.

Kemper, A.R., L.J. Ouyang, and S.D. Grosse, Discontinuation of thyroid hormone
treatment among children in the United States with congenital hypothyroidism:
findings from health insurance claims data. Hormone Research, 2009. 72: p.
478-478.
98