Population-based TSH Screening of Newborns for Hyperthyroidism: It May Be Feasible, but Is It Justified?

Neonatal hyperthyroidism (NH) is a serious condition that occurs mostly in the offspring of women with past or active Graves’ disease. Active Graves’ disease is estimated to affect about 1 in 500 women of childbearing age (1), but only those with a high titer of thyrotropin [thyroid-stimulating hormone (TSH)]-receptor activating immunoglobulins [thyrotropin receptor antibodies (TRAbs)], which cross the placenta, are at risk of passing the disease to their fetus in approximately 1% to 2% of cases (2). Thus, NH is expected in 1 newborn in 25 000 (3). This prevalence estimate has been widely quoted in the literature but is likely imprecise.

Banigé et al (4) report a proof-of-concept study of population-based screening of newborns for NH based on TSH on a dried blood spot collected on day 3. Although this concept appears attractive and relevant, given that TSH-based newborn screening (NBS) for congenital hypothyroidism is already established in high-income countries, we do not think NBS for NH is necessarily justified.

First, in their case-control study, Banigé et al report a prevalence of maternal Graves’ disease of 1 in 670, similar to that previously quoted (1 in 500), but a prevalence of NH of 1/5600 in a cohort of 280 000 births, 4 to 5 times higher than expected (4). Even when considering only cases with severe NH (17 of the 48 NH cases), the prevalence would be 1/16 470, almost double than expected. This prevalence is similar to that of congenital adrenal hyperplasia (CAH), for which population-based biochemical NBS is not universal because of questions about its rationale and the rarity of severe outcomes (5). Similar to that of CAH, the death rate in newborns with NH is based on historical series, which may be outdated: the death rate of 25% quoted by Banigé et al does not correspond to the 1 death among 11 affected fetuses reported by Luton et al (6). In addition, even if rarer than in the past, this tragic outcome should be preventable by early expert multidisciplinary care of women with Graves’ disease, and NBS is irrelevant to its prevention.

Second, the generally accepted estimate of the prevalence of NH (1 in 25 000) applies only to clinically overt disease and may be higher with active case finding. However, in the 31 cases of moderate NH, Banigé et al do not report how many of the clinical diagnostic criteria were actually met (4). Even more important, the management of moderate NH differs significantly across jurisdictions. It can be argued that NH not severe enough to require treatment with antithyroid medications may not warrant hospitalization and that ambulatory follow-up with a pediatric endocrinologist would suffice. This outpatient follow-up is short, since NH is a self-limiting condition, disappearing as the TRAbs are cleared from the infant’s circulation. By contrast, we concur that during the early course of severe NH, the cardiovascular manifestations may be life-threatening and require intensive care. The age at which this occurred in the 12 neonates with severe NH is not provided, neither do we know whether the screening TSH was already known to be <0.18 mU/L at the time. As for CAH, the turnaround time of NBS is crucial.

Third, Banigé et al conclude that the measurement of TSH on dried blood spots, which is routinely carried out to detect hypothyroidism, can be adapted to population-based detection of NH (4). The cutoffs tested were as low as 0.18 mU/L. Not all NBS laboratories express low TSH results with such precision. Indeed, the manufacturer of the kit used by Banigé et al reports an analytical sensitivity of 2 mU/L⁶, and the investigators do not explain how they were able to measure a TSH as low as 0.18 mU/L on dried blood spots.

Fourth, an important criterion for implementing NBS is that it detects treatable disease at a presymptomatic stage. Indeed, the vast majority of severe cases were detected before birth and/or due to a maternal history of Graves’ disease, which represent the true positives in the study of Banigé et al (4). In addition, the 5 newborns with negative NBS were also known to have the disease before birth; their normal TSH was due to the fact that their mothers were taking antithyroid drugs at the time their baby had the NBS test. This illustrates the complexity of NH, which may be initially masked by maternal antithyroid drug treatment only to become apparent once the effect of the drugs wane [their half-life (days) being much shorter than that of immunoglobulins (weeks)] (1).

Fifth, NBS is a public health program and, as such, should be evaluated from the economic standpoint (7). The authors appropriately attempted to estimate potential costs and benefits. However, the costs were calculated based on performing confirmatory serum measurements of TSH, free thyroxine, and total triiodothyronine in only 32 babies (the estimated number of babies with NH), but clinicians would likely perform these tests in all 540 babies with positive screening results. This, in addition to the arguable need for admitting newborns not needing treatment, introduces bias to the cost estimate. On the other hand, the 5 newborns requiring costly intensive care were the 5 false negatives described in the previous discussion and therefore had not been identified by NBS.

In conclusion, the study of Banigé et al indeed shows that NBS for NH based solely on TSH levels may be feasible (4). However, given the questions about rationale, assay sensitivity, and costs as previously outlined, it is likely that replication in other populations and in larger cohorts will be required by the government authorities that fund NBS. In the meantime, measuring TRAbs at midgestation in women with active Graves disease is essential and yet often forgotten. Even more important, clinicians should realize that among the thousands of pregnant women taking levothyroxine, some may have a history of thyroid ablation for Graves’ disease and, while not being at risk for developing hyperthyroidism themselves, may still have high titer TRAbs with potentially serious consequences for their offspring. NBS should not replace awareness of the risk of NH in young women with past or active Graves’ disease and appropriate education of patients and clinicians.

Abbreviations

Abbreviations
 
• CAH

  congenital adrenal hyperplasia

 
• NBS

  newborn screening

 
• NH

  neonatal hypothyroidism

 
• TRAbs

  thyrotropin receptor antibodies

 
• TSH

  thyroid-stimulating hormone (thyrotropin)

Acknowledgments

We thank Drs Marie-Thérèse Berthier (clinical biochemist, Programme Québécois de Dépistage Néonatal Sanguin), François Audibert (Department of Obstetrics and Gynecology, Centre Hospitalier Universitaire Sainte-Justine and Université de Montréal), and Scott D. Grosse (senior health economist, Centers for Disease Control and Prevention, Atlanta, GA, USA) for critical reading of this commentary.

Conflict of Interest

The authors have no conflicts of interest to declare.

Data Availability

Data sharing is not applicable to this article because no data sets were generated or analyzed during the present study.

References