- Email: [email protected]
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
F
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 Maya K. Hunter, MD, Scott H. Mandel, MD, David E. Sesser, BS, Richard S. Miyahira, BS, Leanne Rien, RN, Michael R. Skeels, PhD, and Stephen H. LaFranchi, MD Objectives: To determine the type and frequency of thyroid disorders detected in infants with low thyroxine (T4) and nonelevated thyroid-stimulating hormone (TSH) screening test results in the Northwest Regional Newborn Screening Program (NWRNSP) over the 20-year period from May 1975 to May 1995 and to determine the effect of follow-up of these infants on the overall recall rate. Study design: The NWRNSP requests a serum specimen in infants with an absolute T4 level < 38.6 nmol/L (<3 mg/dl) and in infants with two filter paper T4 concentrations less than the 3%, regardless of the TSH concentration. We conducted a retrospective analysis of infants who were followed up because of low T4 and nonelevated TSH concentrations on newborn screening. To determine the effect of follow-up of infants with low T4 levels, nonelevated TSH concentrations on the recall rate, we selected 1 year (1994) for review. Serum sample requests were evaluated to determine the reason for the request. Results: Over this 20-year period, the NWRNSP detected 450 infants with primary hypothyroidism among 1,747,805 infants screened (1:3,884). Of these, 416 were detected on the basis of low T4 levels and nonelevated TSH screening test results, whereas an additional 34 infants with primary hypothyroidism and 29 infants with hypopituitary hypothyroidism were detected as a result of follow-up of low T4 levels and nonelevated TSH screening test results. This included 25 infants with delayed TSH rise (1:67,226), 9 infants with mild hypothyroidism (TSH levels <25 mU/L) (1:194,212), 29 infants with hypopituitary hypothyroidism (1:60,269), and 434 infants with T4-binding globulin deficiency (1:4,027). Excluding those with T4-binding globulin deficiency, the false-positive rate was 43.5:1. This compares with an overall false-positive rate of 12:1 for our screening program. Conclusion: Follow-up of infants with low T4 and nonelevated TSH concentration on screening led to the detection of 63 additional infants with hypothyroidism, for an overall frequency of 1:27,743. We believe this yield justifies continued follow-up of infants with low T4 levels, nonelevated (TSH) screening test results in our program. (J Pediatr 1998;132:70-4)
From the Department of Pediatrics, Oregon Health Sciences University and the Oregon State Public Health Laboratory, Oregon Health Division, Portland, Oregon. Submitted for publication Sep. 18, 1996; accepted April 29, 1997. Reprint requests: Stephen H. LaFranchi, MD, Department of Pediatrics, (NRC-5), Oregon Health Sciences University, 3181 SW Sam Jackson Park Rd., Portland, Oregon 97201. Copyright © 1998 by Mosby, Inc. 0022-3476/98/$5.00 + 0 9/21/83077
70
Most North American screening programs for congenital hypothyroidism use a primary thyroxine measurement, with thyroidstimulating hormone measurement in infants with a T4 less than a selected cutoff (primary T4-backup TSH approach). The primary goal of these screening programs is to detect infants with primary hypothyroidism, which has an overall incidence of approximately 1 in 4000 newborns.1 All screening programs using the primary T4TSH backup approach will follow up infants with low T4 and elevated TSH concentrations. The recall rate using this method is approximately 0.05%, resulting in two infants recalled for confirmatory serum NWRNSP T4 TBG TSH
Northwest Regional Newborn Screening Program Thyroxine Thyroxine-binding globulin Thyroid-stimulating hormone
testing for every one case of hypothyroidism. In some primary T4 screening programs, including the Northwest Regional Newborn Screening Program (NWRNSP), infants with a filter paper T4 concentration less than an absolute cutoff (e.g., 38.6 nmol/L or 3.0 µg/dl) or with repeated low filter paper T4 measurements (in those programs using two separate specimen collections at two different time periods) will also be followed up regardless of TSH concentrations. The recall rate with this method is approximately 0.30%,2 leading to a higher false-positive rate. This will result in 12 normal infants being recalled for confirmatory serum testing for every one case of hypothyroidism. Most of these infants are normal, and therefore the psychologic harm created
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
HUNTER
ET AL.
Table I. Thyroid function tests in infants with delayed TSH rise
First filter paper specimen Patient No.
Age (days)
T4 (nmol/L)*
TSH† (mU/L)
Second filter paper specimen Age (days)
T4 (nmol/L)*
Serum sample
TSH (mU/L)
Age (days)
T4 (nmol/L)*
Free T4 (pmol/L)‡ —
43
— — — — — — — <3.9 — — 5.1 — 7.7 — — — — — — — <1.3 — —
>100 46 >100 >100 152 72 50.0 >50 >50 98 99 >50 >100 274 Elevated 50 >37 104 >50 47 186 116 >100 >100
1
7
25.7
<25
—
—
—
62
55.3
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
2 17 5 8 15 1 2 3 5 7 1 2 5 5 11 3 5 6 2 14 7
70.8 38.6 65.6 77.2 68.2 96.5 75.9 124.8 32.2 54.1 169.9 18.0 77.2 61.8 81.1 137.7 123.6 74.6 95.2 104.2 48.9 42.5 132.6 122.3
<25 <15 <25 <30 <25 <25 <25 <25 11.8 <25 <25 <25 18 19 24 17 22 22 <25 <25 <25 13 <20 <20
— 26 77 16 33 8 18 — 34 14 15 — 48 27 14 15 26 14 37 42 11
— <15 — <30 <25 <25 <25 — >220 40 61 — >200 278 38 117 >200 126 258 90 56 >220 175 45
? 46 132 29 67 27 100 4 39 16 34 28 57 30 49 25 37 22 48 47 36
31 21
— 56.6 68.2 77.2 <12.9 nl 51.5 — 3.9 16.7 48.9 — 37 74.6 106.8 55.3 11.6 77.6 24.5 30.9 63.1 <25.7 72.1 96.5
18.0 79.8 160.9 137.7 19.3 30.9 96.5 163.4 — 46.3 63.1 — 78.5 — 51.5 112.0 15.4 37.3 61.8 36 86.2 — 39.9 12.9
<25
>126
<25
95-185
12-28
1.7-9.1
Normal range
§ §
3 >126
§
§
49 §
TSH (mU/L)
nl, Normal. *Multiply by 0.0777 to convert to µg/dl. The T4 cutoff of >126 nmol/L represents an average for the 10%. T4 concentrations reported as >126 nmol/L were < 10% for that particular assay and so a TSH was determined. †Where TSH assay sensitivity allowed, an actual TSH concentration is reported; otherwise the concentration is reported as < 25 mU/L. ‡Multiply by 0.0777 to convert to ng/dl. §Data not available.
in follow-up of infants with false-positive screening test results must be considered when choosing the screening method. The main purpose of this study was to determine the type and frequency of thyroid disorders detected in infants with low T4 and nonelevated TSH screening concentrations in the NWRNSP over the 20year period from 1975 to 1995. We also selected 1 year (1994) to analyze the effect of follow-up of these infants on the overall recall rate.
METHODS The NWRNSP, consisting of Oregon, Idaho, Montana (withdrew October 1988), Alaska, and Nevada, began screen-
ing infants for congenital hypothyroidism in May 1975.3 The NWRNSP uses a primary T4-backup TSH approach. Both T4 and TSH measurements are obtained from dried blood on filter paper specimens that are collected from newborn infants during the first few days of life. Most of the infants born in Oregon (95%) and approximately 60% of infants born in the other four states undergo routine second testing between 2 and 6 weeks of age. If one of the following criteria for recall testing is met, a serum specimen collected by venipuncture is requested: (1) filter paper T4 concentration < the 10th percentile as determined for each day’s T4 assays and elevated TSH concentration, typically > 25 mU/L; (2) filter paper T4
concentration less than an absolute cutoff of 38.6 nmol/ L (3 µg/dl ) regardless of the TSH result, which is still pending at the time of the request; or (3) filter paper T4 concentrations < the 3rd percentile in both first and second routine specimens, regardless of the TSH concentration. In premature infants weighing less than 2.5 kg, a third filter paper specimen is requested before confirmation by serum venipuncture. The primary goal of the newborn thyroid screening program is to detect primary hypothyroidism; however, obtaining serum from infants with low T4 concentrations and nonelevated TSH concentrations (criteria 2 and 3) enabled us to retrospectively review the outcome of these infants to de71
HUNTER
ET AL.
THE JOURNAL
OF PEDIATRICS JANUARY 1998
Table II. Thyroid function tests in infants with mild hypothyroidism (TSH < 25 mU/L)
First filter paper specimen Patient No.
Age (days)
T4 (nmol/L)*
TSH† (mU/L)
Second filter paper specimen Age (days)
T4 (nmol/L)*
TSH (mU/L)
Serum sample Age (days)
T4 (nmol/L)*
1
2
64.4
<25
42
74.6
<25
56
118.4
2 3 4 5 6 7 8 9
1 3 1 1 3 2 3 11
77.2 87.5 101.7 >10% 61.8 110.7 59.2 83.7
<25 <25 <25
29 26 22 9 28 7 45 24
63.1 29.6 32.2 38.6 42.5 47.6 73.4 100.4
<25 <25 <25 <25 <25 <25 <25 14
80 53
51.5
>126
<25
>126
<25
Normal range
—
<25 <25 <25 12
§
62 §
23 70 32
Free T4 (pmol/L)‡ — —
52.8 60.5 83.7 60.5
<3.9 3.9 24.5 10.3 — — —
95-185
12-28
— — —
TSH (mU/L) 21
14 16 14 18 19 24 21 13 1.7-9.1
*Multiply by 0.0777 to convert to µg/dl. †Where TSH assay sensitivity allowed, an actual TSH concentration is reported; otherwise the concentration is reported as <25 mU/L. ‡Multiply by 0.0777 to convert to ng/dl. §Data not available.
termine the type and frequency of thyroid disorders detected. Serum specimens collected by venipuncture were routinely assayed for T4,4 TSH,5 triiodothyronine resin uptake,6 and recently free T47 in a centralized laboratory. Thyroxine-binding globulin levels by radioimmunoassay,8 if obtained, were performed at several different laboratories based on the decision of the infant’s private physician. In the analysis of newborns with initial or second routine filter paper T4 < 38.6 nmol/L (<3 µg/dl) or T4 < 10% and TSH “not elevated” (<25 mU/L), four categories of thyroid disorders were discovered, defined as follows: (1) primary hypothyroidism with delayed TSH rise: TSH elevated to > 25 mU/L on either second routine filter paper specimen or confirmatory serum specimen; (2) mild primary hypothyroidism: TSH between 7 and 25 mU/L on confirmatory serum specimen; (3) hypopituitary hypothyroidism: subnormal T4 (<96.5 nmol/L or 7.5 mg/dl ) or free T4 (<10.3 pmol/L or 0.8 ng/dl) and TSH low or “normal” (<7 mU/L) on the confirmatory serum specimen, and (4) TBG deficiency: low T4 and either elevated triiodothyronine resin uptake, normal free T4, and/or low TBG concentration on the confirmatory serum
72
specimen. In infants in the first two categories, the TSH screening concentration appeared “nonelevated,” but in fact the serum TSH concentration was abnormally elevated, thus the reason for using quotation marks in subsequent references to nonelevated TSH concentrations. To analyze the effect of follow-up of infants with low T4-nonelevated TSH concentrations on the infant recall rate, we selected 1 year (1994) for review. Serum sample requests were evaluated to determine the reason for the request. Those that were requested as a result of (1) two filter paper T4 concentrations < 3% with nonelevated TSH concentration or (2) filter paper T4 concentration < 3 mg/dl with nonelevated TSH concentration were selected for review to determine the false-positive rate by using these follow-up criteria.
RESULTS From May 1975 to May 1995, a total of 1,747,805 infants was screened in the NWRNSP. Follow-up of infants with low T4 and nonelevated TSH concentrations led to the detection of 34 infants with primary hypothyroidism, including 25 infants with delayed TSH rise
(1:67,226), 9 infants with mild hypothyroidism (serum TSH > 25 mU/L) (1:194,212), and 29 infants with hypopituitary hypothyroidism (1:60,269). Also detected were 434 infants with TBG deficiency (1:4,027).9 The individual filter paper and serum T4, or free T4, and TSH concentrations in the 25 infants with delayed TSH rise are presented in Table I. The mean thyroid concentrations for the 25 infants identified with delayed TSH rise were T4 82.4 nmol/L, TSH “18.4” mU/L (first filter paper specimen), T4 54.1 nmol/L, TSH “117 mU/L” (second filter paper specimen; four infants did not have second filter paper specimens obtained), and T4 66.9 nmol/L, TSH 102.8 mU/L (confirmatory serum specimen). Of the 15 infants with clinical information available, 8 were term newborns. Of these, five had a history of medical problems that could potentially alter thyroid function, including one with a protein-losing enteropathy, one with necrotizing enterocolitis, one with renal failure, one with trisomy 21, and one with two other male siblings receiving thyroid hormone replacement. Of these 15 infants, 7 were preterm (<36 weeks’ gestation), with 5 of these 7 having birth weights <1.35 kg. One preterm infant
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
HUNTER
ET AL.
Table III. Mean thyroid function test results in the three hypothyroid groups
First filter paper specimen
Second filter paper specimen
Serum specimen
n
T4 (nmol/L)*
TSH† (mU/L)
T4 (nmol/L)*
TSH† (mU/L)
T4 (nmol/L)*
TSH (mU/L)
Delayed TSH rise
25
82.4
“<18.4”
54.1
117
66.9
102.8
Mild hypothyroidism
9
81.1
“<23.4”
55.3
“<23.8”
70.8
17.8
29 hypothyroidism
70.8
“12.9”
74.6
“10.6”
56.6
4.1
Hypopituitary
*Multiply by 0.0777 to convert to µg/dl. †The TSH concentrations in quotes reflect approximations; several values were reported as <25 mU/L and were not further quantified.
born at 36 weeks’ gestation was diagnosed with chronic renal failure, which was managed by peritoneal dialysis. No obvious clinical signs of hypothyroidism were found in these 15 infants diagnosed with delayed TSH rise. Thyroid results in the nine infants with mild hypothyroidism (TSH < 25 mU/L) are presented in Table II. Mean thyroid concentrations for this group of infants were T4, 81.1 nmol/L; TSH, “<23.4” mU/L (first filter paper specimen); T4, 55.3 nmol/L; TSH, “< 23.8” mU/L (second filter paper specimen); and T4, 70.8 nmol/L; TSH, 17.8 mU/L (serum specimen) (Table III). Seven of the nine infants were term newborns and one was preterm (no further information was available for one infant). Of the seven, four were identified with one of the following: ectopic thyroid gland (two), trisomy 21, and congenital nephrotic syndrome. The detection of congenital hypopituitary hypothyroidism by the NWRNSP has been previously reported.10 In that study carried out between 1975 and 1985, 8 infants were identified by screening, and an additional 11 infants were diagnosed on the basis of clinical features, typically before newborn screening results were available. On the basis of thyroid function tests obtained in the neonatal period, we believe that 7 of these 11 infants would have been detected by newborn screening tests had they not been first diagnosed by clinical manifestations. In this study, we identified an additional 21 hypopituitary hypothyroid infants by screening for a total of 29 infants identified through new-
born screening (1:60,269). Mean thyroid values for this group of infants were T4, 70.8 nmol/L; TSH, “12.9” mU/L (first filter paper specimen); T4, 74.6 nmol/L; TSH, “10.6” mU/L (second filter paper specimen); and T4, 56.6 nmol/L; TSH, 4.1 mU/L (serum specimen) (Table II). Mandel et al.9 reported the incidence of inherited TBG deficiency in the NWRNSP in a previous study. Extending the results of this study, we have now identified a total of 434 (1:4027) infants in the 20year period studied. We selected 1 year to analyze the effect of follow-up of low T4-nonelevated TSH on the infant recall rate. In 1994, of 116 infants recalled because of a low T4nonelevated TSH screening test result, 29 were diagnosed with TBG deficiency and 2 with hypopituitary hypothyroidism. Excluding those with TBG deficiency, the false-positive rate was 43.5:1. This compares with an overall false-positive rate of 12:1 for our screening program.
DISCUSSION Congenital hypothyroidism is one of the most common preventable causes of mental retardation. The optimal screening approach is still debated; most North American programs use the primary T4backup TSH approach, whereas most European and Japanese programs screen by primary TSH measurement. The NWRNSP uses a primary T4-backup TSH screening method; in nearly all infants born in Oregon, and most infants
born in Idaho, Alaska, and Nevada, a second filter paper specimen is routinely obtained between 2 and 6 weeks of age.11 All screening programs using the primary T4-backup TSH method will follow up infants with low T4 and elevated TSH concentrations. The NWRNSP also follows up infants with a filter paper T4 concentration less than an absolute cutoff (38.6 nmol/L) and infants with two low filter paper T4 concentrations (<10%) regardless of TSH concentration. The low T4, nonelevated TSH profile is seen in 3% to 5% of neonates on newborn screening, but these infants rarely have thyroid insufficiency. These results often can be explained by immaturity of the hypothalamic-pituitary axis, which is seen more frequently in premature infants. Less common thyroid problems include protein-binding abnormalities, such as TBG deficiency, hypopituitary hypothyroidism, and primary hypothyroidism associated with delayed TSH rise or mild TSH elevation. Routine follow-up of infants with low T4, nonelevated TSH screening tests allowed the NWRNSP to determine the type and frequency of thyroid disorders present in these infants. Over the 20-year period between 1975 and 1995, such follow-up detected 63 infants with hypothyroidism, including 29 infants with hypopituitary hypothyroidism, 25 infants with delayed TSH rise, and 9 infants with mild hypothyroidism (TSH < 25). Perhaps the most serious condition causing a low T4 without elevation of TSH is hypopituitary hypothyroidism.
73
HUNTER
ET AL.
Hypopituitary hypothyroidism is a relatively rare condition (1:60,269) and is often associated with other hormonal deficiencies. Recognition of clinical signs and symptoms of hypopituitarism, such as hypoglycemia, microphallus, or prolonged jaundice, will aid in the diagnosis of these infants. An advantage of follow-up of low T4, nonelevated TSH test results is the early recognition of affected infants, leading to earlier identification of other associated hormone deficiencies, or congenital ophthalmologic or neurologic abnormalities, and prevention of their complications. A disproportionate number of infants in the delayed TSH rise group were premature. Premature infants represent 5% of all newborns, but in this group 7 of the 15 infants with clinical information available were premature. Immaturity of the hypothalamic-pituitary axis may potentially explain their delayed TSH rise. Overall, premature infants are disproportionately represented in infants with a screening T4 < 10%, as illustrated in a recent report by Frank et al.12 Nine infants with “mild” hypothyroidism, as defined by a serum TSH > 25 mU/L, were detected. Most of these infants were term newborns. It is generally recommended that these infants be treated as others with congenital hypothyroidism until the infant has reached the age of 3 years, when myelinization of the nervous system is complete. At this age it would be safe to discontinue thyroid hormone replacement to reevaluate whether the thyroid abnormality is permanent or transient. Follow-up of infants with low T4, nonelevated TSH concentrations will also identify infants with congenital TBG deficiency (1:4,027). This is an Xlinked recessive condition and is more common in males (1:2,400) than congenital hypothyroidism.9 Although this condition has no detrimental effects on a child’s health, it is important to recognize so that inappropriate treatment can be avoided. Several groups have reported on the frequency of missed cases of congenital hypothyroidism with newborn screening.13,14 In addition to errors in sample collection and processing, thyroid disor74
THE JOURNAL
ders such as hypopituitary hypothyroidism, delayed TSH rise, and mild disease accounted for most missed cases. By following up infants with low T4, nonelevated TSH screening results, many of these conditions can be diagnosed early; the price that is paid, however, is a higher recall rate.2 During 1994 the NWRNSP recalled an additional 116 infants as a result of our follow-up of infants with low T4, nonelevated TSH screening test results. Of these 116 infants, 29 were TBG deficient and 2 had hypopituitary hypothyroidism. With the exclusion of those with TBG deficiency, the false-positive rate was 43.5:1, compared with the overall false-positive rate of 12:1 for our screening program. No consistent follow-up approach exists among programs with low T4, nonelevated TSH screening results. Most of these infants are normal or have TBG deficiency; therefore the psychologic distress of a false-positive case versus identifying the child with delayed TSH rise, mild hypothyroidism, or hypopituitary hypothyroidism (overall detection rate 1:27,743) must be evaluated carefully. Despite the increased detection rate, many programs have chosen not to follow up infants with these screening results because of the increased cost of follow-up testing, relatively low number of cases, and uncertain long-term prognosis for some children in these categories. However, the large health care costs an untreated hypothyroid infant will accrue over a lifetime may justify confirmatory thyroid function tests in infants falling in this category. The NWRNSP chooses to continue to follow up infants with low T4, nonelevated TSH concentrations. Clearly, some infants with mild or unusual forms of hypothyroidism will go undetected regardless of the screening strategy and follow-up approach. Physicians caring for infants must not depend solely on screening to detect all cases, rather they must continue to be vigilant. The presence of clinical symptoms or signs of hypothyroidism should lead to prompt assessment of thyroid function.
OF PEDIATRICS JANUARY 1998
REFERENCES 1. Fisher DA. Second international conference on neonatal thyroid screening: progress report. J Pediatr 1983; 102:653-4. 2. The Council of Regional Networks for Genetics Services Newborn Screening Report for 1990. Meaney FJ, Chair, Riggle SM, Data coordinator. New York: The Council of Regional Networks for Genetics Services; 1992. 3. LaFranchi SH, Murphey WH, Foley TP Jr, Larsen PR, Buist NRM. Neonatal hypothyroidism detected by the Northwest Regional Screening Program. Pediatrics 1979;63:180-91. 4. Murphy BEP, Patter CJ. The determination of thyroxine utilizing the property of protein binding. J Clin Endocrinol Metab 1964;24:187-96. 5. Odell WE, Wilber JP, Paul WE. Radioimmunoassay of thyrotropin in human serum. J Clin Endocrinol Metab 1965;25:1179-88. 6. Nara M, DeGroot LJ. Resin uptake of 131I labeled triiodothyronine as a test of thyroid function. N Engl J Med 1962;266:1307-10. 7. Welby ML, Guthrie L, Reiley CP. Evaluation of a new free-thyroxine assay. Clin Chem 1981;27:2022-4. 8. Hesch RD, Gatz J, McIntosh CHS, Janzen J, Hehrmann R. Radioimmunoassay of thyroxine-binding globulin in human plasma. Clin Chem Acta 1976; 70:33-42. 9. Mandel S, Hanna C, Boston B, Sesser D, LaFranchi SH. Thyroxine-binding globulin deficiency detected by newborn screening. J Pediatr 1993;122:227-30. 10. Hanna CE, Krainz PL, Skeels MR, Miyahira RS, Sesser DE, LaFranchi SH. Detection of congenital hypopituitary hypothyroidism: ten-year experience in the Northwest Regional Screening Program. Pediatrics 1986;109:959-64. 11. LaFranchi SH, Hanna CE, Krainz PL, Skeels MR, Miyahira RS, Sesser DE. Screening for congenital hypothyroidism with specimen collection at two time periods: results of the Northwest regional screening program. Pediatrics 1985; 76:734-9. 12. Frank JE, Faix JE, Hermos RJ, Mullaney DM, Rojan DA, Mitchell ML, et al. Thyroid function in very low birth weight infants: effects on neonatal hypothyroidism screening. J Pediatr 1996;128:548-54. 13. Leger J. Screening for congenital hypothyroidism in France. Misdiagnosed cases: collaborative study of screening centers in France. Eur J Pediatr 1990;149:605-7. 14. Fisher DA. Effectiveness of newborn screening programs for congenital hypothyroidism: prevalence of missed cases. Pediatr Clin North Am 1987;34:881-9.
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 Maya K. Hunter, MD, Scott H. Mandel, MD, David E. Sesser, BS, Richard S. Miyahira, BS, Leanne Rien, RN, Michael R. Skeels, PhD, and Stephen H. LaFranchi, MD Objectives: To determine the type and frequency of thyroid disorders detected in infants with low thyroxine (T4) and nonelevated thyroid-stimulating hormone (TSH) screening test results in the Northwest Regional Newborn Screening Program (NWRNSP) over the 20-year period from May 1975 to May 1995 and to determine the effect of follow-up of these infants on the overall recall rate. Study design: The NWRNSP requests a serum specimen in infants with an absolute T4 level < 38.6 nmol/L (<3 mg/dl) and in infants with two filter paper T4 concentrations less than the 3%, regardless of the TSH concentration. We conducted a retrospective analysis of infants who were followed up because of low T4 and nonelevated TSH concentrations on newborn screening. To determine the effect of follow-up of infants with low T4 levels, nonelevated TSH concentrations on the recall rate, we selected 1 year (1994) for review. Serum sample requests were evaluated to determine the reason for the request. Results: Over this 20-year period, the NWRNSP detected 450 infants with primary hypothyroidism among 1,747,805 infants screened (1:3,884). Of these, 416 were detected on the basis of low T4 levels and nonelevated TSH screening test results, whereas an additional 34 infants with primary hypothyroidism and 29 infants with hypopituitary hypothyroidism were detected as a result of follow-up of low T4 levels and nonelevated TSH screening test results. This included 25 infants with delayed TSH rise (1:67,226), 9 infants with mild hypothyroidism (TSH levels <25 mU/L) (1:194,212), 29 infants with hypopituitary hypothyroidism (1:60,269), and 434 infants with T4-binding globulin deficiency (1:4,027). Excluding those with T4-binding globulin deficiency, the false-positive rate was 43.5:1. This compares with an overall false-positive rate of 12:1 for our screening program. Conclusion: Follow-up of infants with low T4 and nonelevated TSH concentration on screening led to the detection of 63 additional infants with hypothyroidism, for an overall frequency of 1:27,743. We believe this yield justifies continued follow-up of infants with low T4 levels, nonelevated (TSH) screening test results in our program. (J Pediatr 1998;132:70-4)
From the Department of Pediatrics, Oregon Health Sciences University and the Oregon State Public Health Laboratory, Oregon Health Division, Portland, Oregon. Submitted for publication Sep. 18, 1996; accepted April 29, 1997. Reprint requests: Stephen H. LaFranchi, MD, Department of Pediatrics, (NRC-5), Oregon Health Sciences University, 3181 SW Sam Jackson Park Rd., Portland, Oregon 97201. Copyright © 1998 by Mosby, Inc. 0022-3476/98/$5.00 + 0 9/21/83077
70
Most North American screening programs for congenital hypothyroidism use a primary thyroxine measurement, with thyroidstimulating hormone measurement in infants with a T4 less than a selected cutoff (primary T4-backup TSH approach). The primary goal of these screening programs is to detect infants with primary hypothyroidism, which has an overall incidence of approximately 1 in 4000 newborns.1 All screening programs using the primary T4TSH backup approach will follow up infants with low T4 and elevated TSH concentrations. The recall rate using this method is approximately 0.05%, resulting in two infants recalled for confirmatory serum NWRNSP T4 TBG TSH
Northwest Regional Newborn Screening Program Thyroxine Thyroxine-binding globulin Thyroid-stimulating hormone
testing for every one case of hypothyroidism. In some primary T4 screening programs, including the Northwest Regional Newborn Screening Program (NWRNSP), infants with a filter paper T4 concentration less than an absolute cutoff (e.g., 38.6 nmol/L or 3.0 µg/dl) or with repeated low filter paper T4 measurements (in those programs using two separate specimen collections at two different time periods) will also be followed up regardless of TSH concentrations. The recall rate with this method is approximately 0.30%,2 leading to a higher false-positive rate. This will result in 12 normal infants being recalled for confirmatory serum testing for every one case of hypothyroidism. Most of these infants are normal, and therefore the psychologic harm created
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
HUNTER
ET AL.
Table I. Thyroid function tests in infants with delayed TSH rise
First filter paper specimen Patient No.
Age (days)
T4 (nmol/L)*
TSH† (mU/L)
Second filter paper specimen Age (days)
T4 (nmol/L)*
Serum sample
TSH (mU/L)
Age (days)
T4 (nmol/L)*
Free T4 (pmol/L)‡ —
43
— — — — — — — <3.9 — — 5.1 — 7.7 — — — — — — — <1.3 — —
>100 46 >100 >100 152 72 50.0 >50 >50 98 99 >50 >100 274 Elevated 50 >37 104 >50 47 186 116 >100 >100
1
7
25.7
<25
—
—
—
62
55.3
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
2 17 5 8 15 1 2 3 5 7 1 2 5 5 11 3 5 6 2 14 7
70.8 38.6 65.6 77.2 68.2 96.5 75.9 124.8 32.2 54.1 169.9 18.0 77.2 61.8 81.1 137.7 123.6 74.6 95.2 104.2 48.9 42.5 132.6 122.3
<25 <15 <25 <30 <25 <25 <25 <25 11.8 <25 <25 <25 18 19 24 17 22 22 <25 <25 <25 13 <20 <20
— 26 77 16 33 8 18 — 34 14 15 — 48 27 14 15 26 14 37 42 11
— <15 — <30 <25 <25 <25 — >220 40 61 — >200 278 38 117 >200 126 258 90 56 >220 175 45
? 46 132 29 67 27 100 4 39 16 34 28 57 30 49 25 37 22 48 47 36
31 21
— 56.6 68.2 77.2 <12.9 nl 51.5 — 3.9 16.7 48.9 — 37 74.6 106.8 55.3 11.6 77.6 24.5 30.9 63.1 <25.7 72.1 96.5
18.0 79.8 160.9 137.7 19.3 30.9 96.5 163.4 — 46.3 63.1 — 78.5 — 51.5 112.0 15.4 37.3 61.8 36 86.2 — 39.9 12.9
<25
>126
<25
95-185
12-28
1.7-9.1
Normal range
§ §
3 >126
§
§
49 §
TSH (mU/L)
nl, Normal. *Multiply by 0.0777 to convert to µg/dl. The T4 cutoff of >126 nmol/L represents an average for the 10%. T4 concentrations reported as >126 nmol/L were < 10% for that particular assay and so a TSH was determined. †Where TSH assay sensitivity allowed, an actual TSH concentration is reported; otherwise the concentration is reported as < 25 mU/L. ‡Multiply by 0.0777 to convert to ng/dl. §Data not available.
in follow-up of infants with false-positive screening test results must be considered when choosing the screening method. The main purpose of this study was to determine the type and frequency of thyroid disorders detected in infants with low T4 and nonelevated TSH screening concentrations in the NWRNSP over the 20year period from 1975 to 1995. We also selected 1 year (1994) to analyze the effect of follow-up of these infants on the overall recall rate.
METHODS The NWRNSP, consisting of Oregon, Idaho, Montana (withdrew October 1988), Alaska, and Nevada, began screen-
ing infants for congenital hypothyroidism in May 1975.3 The NWRNSP uses a primary T4-backup TSH approach. Both T4 and TSH measurements are obtained from dried blood on filter paper specimens that are collected from newborn infants during the first few days of life. Most of the infants born in Oregon (95%) and approximately 60% of infants born in the other four states undergo routine second testing between 2 and 6 weeks of age. If one of the following criteria for recall testing is met, a serum specimen collected by venipuncture is requested: (1) filter paper T4 concentration < the 10th percentile as determined for each day’s T4 assays and elevated TSH concentration, typically > 25 mU/L; (2) filter paper T4
concentration less than an absolute cutoff of 38.6 nmol/ L (3 µg/dl ) regardless of the TSH result, which is still pending at the time of the request; or (3) filter paper T4 concentrations < the 3rd percentile in both first and second routine specimens, regardless of the TSH concentration. In premature infants weighing less than 2.5 kg, a third filter paper specimen is requested before confirmation by serum venipuncture. The primary goal of the newborn thyroid screening program is to detect primary hypothyroidism; however, obtaining serum from infants with low T4 concentrations and nonelevated TSH concentrations (criteria 2 and 3) enabled us to retrospectively review the outcome of these infants to de71
HUNTER
ET AL.
THE JOURNAL
OF PEDIATRICS JANUARY 1998
Table II. Thyroid function tests in infants with mild hypothyroidism (TSH < 25 mU/L)
First filter paper specimen Patient No.
Age (days)
T4 (nmol/L)*
TSH† (mU/L)
Second filter paper specimen Age (days)
T4 (nmol/L)*
TSH (mU/L)
Serum sample Age (days)
T4 (nmol/L)*
1
2
64.4
<25
42
74.6
<25
56
118.4
2 3 4 5 6 7 8 9
1 3 1 1 3 2 3 11
77.2 87.5 101.7 >10% 61.8 110.7 59.2 83.7
<25 <25 <25
29 26 22 9 28 7 45 24
63.1 29.6 32.2 38.6 42.5 47.6 73.4 100.4
<25 <25 <25 <25 <25 <25 <25 14
80 53
51.5
>126
<25
>126
<25
Normal range
—
<25 <25 <25 12
§
62 §
23 70 32
Free T4 (pmol/L)‡ — —
52.8 60.5 83.7 60.5
<3.9 3.9 24.5 10.3 — — —
95-185
12-28
— — —
TSH (mU/L) 21
14 16 14 18 19 24 21 13 1.7-9.1
*Multiply by 0.0777 to convert to µg/dl. †Where TSH assay sensitivity allowed, an actual TSH concentration is reported; otherwise the concentration is reported as <25 mU/L. ‡Multiply by 0.0777 to convert to ng/dl. §Data not available.
termine the type and frequency of thyroid disorders detected. Serum specimens collected by venipuncture were routinely assayed for T4,4 TSH,5 triiodothyronine resin uptake,6 and recently free T47 in a centralized laboratory. Thyroxine-binding globulin levels by radioimmunoassay,8 if obtained, were performed at several different laboratories based on the decision of the infant’s private physician. In the analysis of newborns with initial or second routine filter paper T4 < 38.6 nmol/L (<3 µg/dl) or T4 < 10% and TSH “not elevated” (<25 mU/L), four categories of thyroid disorders were discovered, defined as follows: (1) primary hypothyroidism with delayed TSH rise: TSH elevated to > 25 mU/L on either second routine filter paper specimen or confirmatory serum specimen; (2) mild primary hypothyroidism: TSH between 7 and 25 mU/L on confirmatory serum specimen; (3) hypopituitary hypothyroidism: subnormal T4 (<96.5 nmol/L or 7.5 mg/dl ) or free T4 (<10.3 pmol/L or 0.8 ng/dl) and TSH low or “normal” (<7 mU/L) on the confirmatory serum specimen, and (4) TBG deficiency: low T4 and either elevated triiodothyronine resin uptake, normal free T4, and/or low TBG concentration on the confirmatory serum
72
specimen. In infants in the first two categories, the TSH screening concentration appeared “nonelevated,” but in fact the serum TSH concentration was abnormally elevated, thus the reason for using quotation marks in subsequent references to nonelevated TSH concentrations. To analyze the effect of follow-up of infants with low T4-nonelevated TSH concentrations on the infant recall rate, we selected 1 year (1994) for review. Serum sample requests were evaluated to determine the reason for the request. Those that were requested as a result of (1) two filter paper T4 concentrations < 3% with nonelevated TSH concentration or (2) filter paper T4 concentration < 3 mg/dl with nonelevated TSH concentration were selected for review to determine the false-positive rate by using these follow-up criteria.
RESULTS From May 1975 to May 1995, a total of 1,747,805 infants was screened in the NWRNSP. Follow-up of infants with low T4 and nonelevated TSH concentrations led to the detection of 34 infants with primary hypothyroidism, including 25 infants with delayed TSH rise
(1:67,226), 9 infants with mild hypothyroidism (serum TSH > 25 mU/L) (1:194,212), and 29 infants with hypopituitary hypothyroidism (1:60,269). Also detected were 434 infants with TBG deficiency (1:4,027).9 The individual filter paper and serum T4, or free T4, and TSH concentrations in the 25 infants with delayed TSH rise are presented in Table I. The mean thyroid concentrations for the 25 infants identified with delayed TSH rise were T4 82.4 nmol/L, TSH “18.4” mU/L (first filter paper specimen), T4 54.1 nmol/L, TSH “117 mU/L” (second filter paper specimen; four infants did not have second filter paper specimens obtained), and T4 66.9 nmol/L, TSH 102.8 mU/L (confirmatory serum specimen). Of the 15 infants with clinical information available, 8 were term newborns. Of these, five had a history of medical problems that could potentially alter thyroid function, including one with a protein-losing enteropathy, one with necrotizing enterocolitis, one with renal failure, one with trisomy 21, and one with two other male siblings receiving thyroid hormone replacement. Of these 15 infants, 7 were preterm (<36 weeks’ gestation), with 5 of these 7 having birth weights <1.35 kg. One preterm infant
THE JOURNAL OF PEDIATRICS Volume 132, Number 1
HUNTER
ET AL.
Table III. Mean thyroid function test results in the three hypothyroid groups
First filter paper specimen
Second filter paper specimen
Serum specimen
n
T4 (nmol/L)*
TSH† (mU/L)
T4 (nmol/L)*
TSH† (mU/L)
T4 (nmol/L)*
TSH (mU/L)
Delayed TSH rise
25
82.4
“<18.4”
54.1
117
66.9
102.8
Mild hypothyroidism
9
81.1
“<23.4”
55.3
“<23.8”
70.8
17.8
29 hypothyroidism
70.8
“12.9”
74.6
“10.6”
56.6
4.1
Hypopituitary
*Multiply by 0.0777 to convert to µg/dl. †The TSH concentrations in quotes reflect approximations; several values were reported as <25 mU/L and were not further quantified.
born at 36 weeks’ gestation was diagnosed with chronic renal failure, which was managed by peritoneal dialysis. No obvious clinical signs of hypothyroidism were found in these 15 infants diagnosed with delayed TSH rise. Thyroid results in the nine infants with mild hypothyroidism (TSH < 25 mU/L) are presented in Table II. Mean thyroid concentrations for this group of infants were T4, 81.1 nmol/L; TSH, “<23.4” mU/L (first filter paper specimen); T4, 55.3 nmol/L; TSH, “< 23.8” mU/L (second filter paper specimen); and T4, 70.8 nmol/L; TSH, 17.8 mU/L (serum specimen) (Table III). Seven of the nine infants were term newborns and one was preterm (no further information was available for one infant). Of the seven, four were identified with one of the following: ectopic thyroid gland (two), trisomy 21, and congenital nephrotic syndrome. The detection of congenital hypopituitary hypothyroidism by the NWRNSP has been previously reported.10 In that study carried out between 1975 and 1985, 8 infants were identified by screening, and an additional 11 infants were diagnosed on the basis of clinical features, typically before newborn screening results were available. On the basis of thyroid function tests obtained in the neonatal period, we believe that 7 of these 11 infants would have been detected by newborn screening tests had they not been first diagnosed by clinical manifestations. In this study, we identified an additional 21 hypopituitary hypothyroid infants by screening for a total of 29 infants identified through new-
born screening (1:60,269). Mean thyroid values for this group of infants were T4, 70.8 nmol/L; TSH, “12.9” mU/L (first filter paper specimen); T4, 74.6 nmol/L; TSH, “10.6” mU/L (second filter paper specimen); and T4, 56.6 nmol/L; TSH, 4.1 mU/L (serum specimen) (Table II). Mandel et al.9 reported the incidence of inherited TBG deficiency in the NWRNSP in a previous study. Extending the results of this study, we have now identified a total of 434 (1:4027) infants in the 20year period studied. We selected 1 year to analyze the effect of follow-up of low T4-nonelevated TSH on the infant recall rate. In 1994, of 116 infants recalled because of a low T4nonelevated TSH screening test result, 29 were diagnosed with TBG deficiency and 2 with hypopituitary hypothyroidism. Excluding those with TBG deficiency, the false-positive rate was 43.5:1. This compares with an overall false-positive rate of 12:1 for our screening program.
DISCUSSION Congenital hypothyroidism is one of the most common preventable causes of mental retardation. The optimal screening approach is still debated; most North American programs use the primary T4backup TSH approach, whereas most European and Japanese programs screen by primary TSH measurement. The NWRNSP uses a primary T4-backup TSH screening method; in nearly all infants born in Oregon, and most infants
born in Idaho, Alaska, and Nevada, a second filter paper specimen is routinely obtained between 2 and 6 weeks of age.11 All screening programs using the primary T4-backup TSH method will follow up infants with low T4 and elevated TSH concentrations. The NWRNSP also follows up infants with a filter paper T4 concentration less than an absolute cutoff (38.6 nmol/L) and infants with two low filter paper T4 concentrations (<10%) regardless of TSH concentration. The low T4, nonelevated TSH profile is seen in 3% to 5% of neonates on newborn screening, but these infants rarely have thyroid insufficiency. These results often can be explained by immaturity of the hypothalamic-pituitary axis, which is seen more frequently in premature infants. Less common thyroid problems include protein-binding abnormalities, such as TBG deficiency, hypopituitary hypothyroidism, and primary hypothyroidism associated with delayed TSH rise or mild TSH elevation. Routine follow-up of infants with low T4, nonelevated TSH screening tests allowed the NWRNSP to determine the type and frequency of thyroid disorders present in these infants. Over the 20-year period between 1975 and 1995, such follow-up detected 63 infants with hypothyroidism, including 29 infants with hypopituitary hypothyroidism, 25 infants with delayed TSH rise, and 9 infants with mild hypothyroidism (TSH < 25). Perhaps the most serious condition causing a low T4 without elevation of TSH is hypopituitary hypothyroidism.
73
HUNTER
ET AL.
Hypopituitary hypothyroidism is a relatively rare condition (1:60,269) and is often associated with other hormonal deficiencies. Recognition of clinical signs and symptoms of hypopituitarism, such as hypoglycemia, microphallus, or prolonged jaundice, will aid in the diagnosis of these infants. An advantage of follow-up of low T4, nonelevated TSH test results is the early recognition of affected infants, leading to earlier identification of other associated hormone deficiencies, or congenital ophthalmologic or neurologic abnormalities, and prevention of their complications. A disproportionate number of infants in the delayed TSH rise group were premature. Premature infants represent 5% of all newborns, but in this group 7 of the 15 infants with clinical information available were premature. Immaturity of the hypothalamic-pituitary axis may potentially explain their delayed TSH rise. Overall, premature infants are disproportionately represented in infants with a screening T4 < 10%, as illustrated in a recent report by Frank et al.12 Nine infants with “mild” hypothyroidism, as defined by a serum TSH > 25 mU/L, were detected. Most of these infants were term newborns. It is generally recommended that these infants be treated as others with congenital hypothyroidism until the infant has reached the age of 3 years, when myelinization of the nervous system is complete. At this age it would be safe to discontinue thyroid hormone replacement to reevaluate whether the thyroid abnormality is permanent or transient. Follow-up of infants with low T4, nonelevated TSH concentrations will also identify infants with congenital TBG deficiency (1:4,027). This is an Xlinked recessive condition and is more common in males (1:2,400) than congenital hypothyroidism.9 Although this condition has no detrimental effects on a child’s health, it is important to recognize so that inappropriate treatment can be avoided. Several groups have reported on the frequency of missed cases of congenital hypothyroidism with newborn screening.13,14 In addition to errors in sample collection and processing, thyroid disor74
THE JOURNAL
ders such as hypopituitary hypothyroidism, delayed TSH rise, and mild disease accounted for most missed cases. By following up infants with low T4, nonelevated TSH screening results, many of these conditions can be diagnosed early; the price that is paid, however, is a higher recall rate.2 During 1994 the NWRNSP recalled an additional 116 infants as a result of our follow-up of infants with low T4, nonelevated TSH screening test results. Of these 116 infants, 29 were TBG deficient and 2 had hypopituitary hypothyroidism. With the exclusion of those with TBG deficiency, the false-positive rate was 43.5:1, compared with the overall false-positive rate of 12:1 for our screening program. No consistent follow-up approach exists among programs with low T4, nonelevated TSH screening results. Most of these infants are normal or have TBG deficiency; therefore the psychologic distress of a false-positive case versus identifying the child with delayed TSH rise, mild hypothyroidism, or hypopituitary hypothyroidism (overall detection rate 1:27,743) must be evaluated carefully. Despite the increased detection rate, many programs have chosen not to follow up infants with these screening results because of the increased cost of follow-up testing, relatively low number of cases, and uncertain long-term prognosis for some children in these categories. However, the large health care costs an untreated hypothyroid infant will accrue over a lifetime may justify confirmatory thyroid function tests in infants falling in this category. The NWRNSP chooses to continue to follow up infants with low T4, nonelevated TSH concentrations. Clearly, some infants with mild or unusual forms of hypothyroidism will go undetected regardless of the screening strategy and follow-up approach. Physicians caring for infants must not depend solely on screening to detect all cases, rather they must continue to be vigilant. The presence of clinical symptoms or signs of hypothyroidism should lead to prompt assessment of thyroid function.
OF PEDIATRICS JANUARY 1998
REFERENCES 1. Fisher DA. Second international conference on neonatal thyroid screening: progress report. J Pediatr 1983; 102:653-4. 2. The Council of Regional Networks for Genetics Services Newborn Screening Report for 1990. Meaney FJ, Chair, Riggle SM, Data coordinator. New York: The Council of Regional Networks for Genetics Services; 1992. 3. LaFranchi SH, Murphey WH, Foley TP Jr, Larsen PR, Buist NRM. Neonatal hypothyroidism detected by the Northwest Regional Screening Program. Pediatrics 1979;63:180-91. 4. Murphy BEP, Patter CJ. The determination of thyroxine utilizing the property of protein binding. J Clin Endocrinol Metab 1964;24:187-96. 5. Odell WE, Wilber JP, Paul WE. Radioimmunoassay of thyrotropin in human serum. J Clin Endocrinol Metab 1965;25:1179-88. 6. Nara M, DeGroot LJ. Resin uptake of 131I labeled triiodothyronine as a test of thyroid function. N Engl J Med 1962;266:1307-10. 7. Welby ML, Guthrie L, Reiley CP. Evaluation of a new free-thyroxine assay. Clin Chem 1981;27:2022-4. 8. Hesch RD, Gatz J, McIntosh CHS, Janzen J, Hehrmann R. Radioimmunoassay of thyroxine-binding globulin in human plasma. Clin Chem Acta 1976; 70:33-42. 9. Mandel S, Hanna C, Boston B, Sesser D, LaFranchi SH. Thyroxine-binding globulin deficiency detected by newborn screening. J Pediatr 1993;122:227-30. 10. Hanna CE, Krainz PL, Skeels MR, Miyahira RS, Sesser DE, LaFranchi SH. Detection of congenital hypopituitary hypothyroidism: ten-year experience in the Northwest Regional Screening Program. Pediatrics 1986;109:959-64. 11. LaFranchi SH, Hanna CE, Krainz PL, Skeels MR, Miyahira RS, Sesser DE. Screening for congenital hypothyroidism with specimen collection at two time periods: results of the Northwest regional screening program. Pediatrics 1985; 76:734-9. 12. Frank JE, Faix JE, Hermos RJ, Mullaney DM, Rojan DA, Mitchell ML, et al. Thyroid function in very low birth weight infants: effects on neonatal hypothyroidism screening. J Pediatr 1996;128:548-54. 13. Leger J. Screening for congenital hypothyroidism in France. Misdiagnosed cases: collaborative study of screening centers in France. Eur J Pediatr 1990;149:605-7. 14. Fisher DA. Effectiveness of newborn screening programs for congenital hypothyroidism: prevalence of missed cases. Pediatr Clin North Am 1987;34:881-9.