Newborn Screening for Congenital Hypothyroidism in Michigan: Past, Present,

1
Newborn Screening for Congenital Hypothyroidism
in MichiganPast, Present, Future

• Steven J. Korzeniewski, MA, MSc,
• Maternal Child Health Epidemiology Section
  Manager
• Bureau of Epidemiology, Division of Genomics,
  Perinatal Health and Chronic Disease Epidemiology

2
Objective

• Provide a historical account of CH screening in
  Michigan and detail the application of
  epidemiology to improve outcomes.
• Outline
• Case definition
• Screening methods over time
• Regional efforts to improve screening outcomes
• CH detection among LBW/Premature/NICU infants
• Three year follow-up study preliminary findings
• Future directions

3
Congenital Hypothyroidism Background

• No standardized case definition
• Defined as inadequate thyroid hormone production
• Usually characterized by increased thyroid
  stimulating hormone (TSH) concentration and
  decreased thyroxin (T4) concentration.
• Insufficient thyroid hormone impacts brain
  development specifically, myelination and
  neuronal connections are impacted

4
Hypothalamic-pituitary-thyroid axis

• T4 is stimulated by secretion of TSH by the
  pituitary.
• Low T4 (negative feedback) causes hypothalamus
  to secrete thyrotropin releasing hormone (TRH)
  which stimulates release of TSH thereby
  stimulating the thyroid gland to increase
  secretion of T4.
• When the thyroid does not respond to TSH
  stimulation (primary CH), due to dysgenesis or
  dyshormonogenesis, the result is low T4 and high
  TSH.

Hypothalamus
TRH
T3, T4 -
Pituitary Gland
T3, T4 -
TSH
Thyroid
(-) inhibition () stimulation
5
Etiology

• Dysgenesis (thyroid gland developmental defect)
• Dyshormongenesis (defects in thyroid hormone
  synthesis)
• Mutations in thyroid development genes or TSH
  receptor
• Defects in hypothalamus or pituitary (central or
  secondary/tertiary hypothyroidism) rare
  (1/25,000 1/100,000)
• Iodine deficiency

Refetoff S, Dumont JE, Vassart G 2001 Thyroid
disorders. In Scriver CR, ed. The metabolic and
molecular bases of inherited diseases. Vol 3. New
York McGraw-Hill 40294076 de Felice M, Di
Lauro R 2004 Thyroid development and its
disorders genetics and molecular mechanisms.
Endocr Rev 25722746 LaFranchi S. Congenital
hypothyroidism etiologies diagnosis and
management. Thyroid. 1999, 7735-40
6
Newborn Screening

• NBS for CH
• Quebec, Canada, and Pittsburgh, Pennsylvania in
  1974
• Michigan, 1977

7
Michigan CH Screening Algorithms

• 1977-1998 Primary T4 Secondary TSH
• lt10th T4 had TSH evaluated
• 1998-2003 Both T4 TSH Initially screened
• Increased sensitivity/specificity for primary CH
• Analysis of these data is underway
• 2003-Present Primary TSH screen
• T4 removed from algorithm due to extreme FPR
  (3-5) and few cases of secondary CH detected
• Fixed cut-off used (age in hours, uIU/mL)
  (24-36h, lt33), (37h-6 days, lt 25), (7-31 days, lt
  13), (gt31 days, lt10)

8
Diagnosed CH Cases () by Serum TSH FT4,
Michigan, 9/2003-9/2007
Normal Range Serum FT4 with Significant
Elevations of Serum TSH
Serum FT4 Serum TSH Serum TSH Serum TSH Serum TSH
Serum FT4 gt40 N48 (30.0) gt40 N48 (30.0) gt100 N16 (14.5) gt100 N16 (14.5)
Serum FT4 Median TSH Median FT4 Median TSH Median FT4
gt.9 73.1 1.2 125.5 1.1
with Serum FT4 gt .9 / with Serum TSH gt
40, or 100
9
Disorders Identified via Newborn Screening,
Michigan Residents, 1965-2007
Type of Disorder Classification (Year Screening Began) Cases in 2007 (N) Cases Through 2007 (N) Cumulative Detection Rate
Galactosemia (1985) 8 124 125,300
Biotinidase Deficiencies (1987) 12 160 117,885
Amino Acid Disorders (1965) 9 599 110,177
Organic Acid Disorders (2005) 3 13 129,391
Fatty Acid Oxidation Disorders (2003) 16 55 111,684
Congenital Hypothyroidism (1987) 88 1,482 11,931
Congenital Adrenal Hyperplasias (1993) 3 106 118,744
Hemoglobinopathies (1987) 53 1,389 12,060
Cystic Fibrosis (since Oct. 2007) 7 7 14,436
38 of all cases
10
CH Screening Results and Performance Metrics,
Michigan, 2007
Total N Total N ( NICU) Strong N Confirmed CH N Positive Detection Rate Among All Positive Screens Among All Positive Screens Among Strong Positive Screens Among Strong Positive Screens
Total N Total N ( NICU) Strong N Confirmed CH N Positive Detection Rate FPR PPV FPR PPV
123,146 697 (34.4) 183 88 11,399 0.49 12.63 0.10 34.43
11
Program Evaluation
12
Evaluation of TSH Distributions

• Charles Mundt evaluated variation
• in initial TSH distributions significant
  variation by
• -gestational age
• -birthweight
• -race
• -month
• -state

13
Michigan TSH Variation by Selected Demographic
Factors, Jan 2005-Dec 2006
Factor Factor N Median 99.5 99.8
Sex Male 126,286 11 36 44
Sex Female 120,756 10 36 45
NICU Yes 26,076 9 50 76
NICU No 223,361 11 34 41
Race White 162,589 11 36 43
Race Black 42,511 11 40 47
Birthweight lt2,500g 23,098 9 45 59
Birthweight gt2,500g 226,339 11 35 43
Gestational Age lt 37weeks 35,181 9 42 57
Gestational Age gt37 weeks 214,256 11 35 43
Age at Collection 1-23 Hours 3,455 12 100 138
Age at Collection 24-36 Hours 212,896 11 35 42
Age at Collection gt37 Hours 33,086 8 31 38
14
(No Transcript)
15
Michigan Mean TSH by Month, Jan 2005-Dec 2006
All Selected (Full Term, White, Female,
Non-NICU, NBW, Screened 24-36 hrs
16
TSH values, Michigan Resident Births, Jan
2005-Dec 2006, Screened lt 14 Days of Life
month
17
Implications of TSH Variation

• Given the variation observed, a single cut-off
  approach is unlikely to yield improvements in
  PPV/FPR.
• Percentile based approaches to cut-off
  calculations appear warranted
• However, stratified percentile based cut-offs
  would require further study and implementation
  would require significant changes in the NBS
  laboratory

18
NICU Screening Protocol
19
CH Detection in Premature Infants

• Maturity of the endocrine system influences
  initial dried blood spot TSH values.
• Some premature infants with CH have late rising
  TSH and are therefore not detectable by an
  initial screen at 24-36hrs of life
• NICU/LBW screening protocol implemented (2007) to
  account for unreliable screens in premature
  newborns

20
NICU/LBW Screening Algorithm
21
CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008 CH Cases by Birth Weight Birth Year, Michigan, 1/1/2005-5/21/2008
Birth Year Birth Weight Birth Weight Birth Weight Birth Weight Birth Weight Birth Weight Birth Weight Birth Weight Birth Weight
Birth Year lt 1800g lt 1800g lt 1800g 1800-2499g 1800-2499g 1800-2499g gt 2500g gt 2500g gt 2500g
Birth Year Cases Screened Rate Cases Screened Rate Cases Screened Rate
2005 5 2587 1517 9 6873 1764 58 108052 11863
2006 4 2793 1698 5 7300 11460 50 114136 12283
2007 14 2961 1211 5 7050 11410 68 111893 11645
1/1/08- 5/21/08 5 996 1199 2 2757 11379 23 42961 11868
22
Number of Infants Screened by Birth Weight,
Michigan, 2007
Birth Weight Frequency Cumulative Frequency Cumulative Estimated Number of Infants Added to the NICU Protocol (Cumulative)
lt 1800g 2,794 2.31 2,794 2.31  -
1800-1999g 1,019 0.84 3,813 3.15 1,019
2000-2199g 1,554 1.28 5,367 4.43 2,573
2200-2499g 4,422 3.65 9,789 8.08 6,995
gt2500g 111,291 91.92 121,080 100  -
23
NICU Protocol Evaluation Conclusions

• NICU protocol inclusion criteria currently does
  not consider gestational age.
• GA is thought to be a better indicator of
  maturity than BW however, concerns about data
  quality have led the latter to be used.
• Adding very preterm and below (GA lt 32 wks)
  infants would add 215 infants based on 2007 data.
• We are evaluating the impact of
• increasing the birthweight inclusion criterion
  to 2300g,
• adding very preterm and below (GA lt 32 wks)
  infants, and
• eliminating 2nd screens for infants born weighing
  less than 1,000g-1,500g.
• Clinical Laboratory Standards Institute (CLSI)
  has a subcommittee working on a proposed standard
  for NBS of SCBU/NICU infants

24
CH Three-Year Follow-up Study
25
CH Three Year Follow-up

• Standard of care is to follow-up CH cases until
  at least age three years (AAP Guidelines)
• Diagnostic verification recommended (permanent
  vs. transient CH)
• Thyroid function trial
• Rate and outcome of diagnostic verification
  unknown

26
CH Three Year Follow-up of Borderline Cases

• Borderline was initially considered as having
  pre-treatment serum TSH values below the 15Th
  percentile.
• We recently expanded our criteria to include
  cases below the 25th percentile.
• Survey developed and pilot tested in
  collaboration with PEAC endocrinologists
• Medical management data maintained by NBS
  Follow-up program used to locate cases
  physicians.
• LTFU mitigated by use of MCIR and phone based
  survey

27
Demographics of Michigan CH Cases Detected by NBS
Classified by Pre-Treatment Serum TSH/FT4 Values,
9/2003-9/2007
Classification N Mean GA Mean BW White () Male () NICU ()
FT4 lt .9, TSH gt40 103 44.78 37.9 3156 79.1 39.2 19.6
FT4 gt.9, TSH gt40 48 20.87 38 3254 61.5 33.3 18.2
FT4 lt .9, TSH lt 40 9 3.91 34.1 2461 50 66.7 55.6
FT4 gt.9, TSH lt40 70 30.43 38.5 3001 53.1 58 25.8
28
Preliminary Findings of the CHThree year
follow-up study

• Preliminary findings indicate
• 44 (8/18) of cases with completed diagnostic
  re-evaluation to date appear not to have
  permanent CH.
• A surprising number of patients (families) have
  stopped treatment of their own accord (7/22 with
  completed follow-up) this phenomenon requires
  further study.
• One case (family) described miscommunication with
  health care provider and lack of follow-up as
  impetus to treatment cessation.
• While all patients who stopped treatment on their
  own are thought to be transient CH (confirmation
  of transient CH is pending TSH value receipt- two
  patients have provided such confirmation to
  date), only 1/11 cases evaluated by an
  endocrinologist is considered transient.
• This finding may have implications for the method
  of diagnostic re-evaluation.
• Some cases were confirmed based on increasing
  treatment dosage at approximately 6 months of
  age should this preclude three year thyroid
  challenge?
• Questions remain about process and outcome of
  three year CH re-evaluation.

29
Conclusion

• Epidemiology is being used to improve CH
  screening outcomes by evaluating
• variations in TSH distributions and potential
  implications for initial CH screens
• detection among premature/LBW infants and
  potential algorithm changes
• The role of follow-up in differentiation of
  transient from permanent CH

30
Future Directions

• Analyze data from the tandem primary T4 TSH
  screening period (1998-2003)
• Evaluate impact of percentile vs. fixed initial
  TSH cutoffs
• Further analyze impact of NICU/LBW screening
  protocol inclusion criteria changes
• Follow-up with families of cases taken off of
  treatment
• Continue to collect informaiton on diagnostic
  re-evaluation in hopes of initiating discussion
  about standardization.

31
Acknowledgements

• Bill Young, Manger, Newborn Screening Follow-up
  Program
• Violanda Grigorescu, Director, Division of
  Genomics, Perinatal Health Chronic Disease
  Epidemiology
• Charles Mundt, Outreach Specialist, Institute for
  Healthcare Studies, MSU
• Newborn Screening Follow-up Staff

32
Your Thoughts, Questions?

• KorzeniewskiS_at_Michigan.gov
• http//www.michigan.gov/mchepi