Neonatal Screening for Hyperthyroidism Proof of Concept

Abstract

Abstract Context Early treatment is essential to avoid the cardiac complication of neonatal hyperthyroidism (NH). Our results have direct implications for clinical care. Objective NH can cause potentially fatal neonatal thyrotoxicosis. Here, we have evaluated the feasibility of neonatal hyperthyroidism screening using the thyroid-stimulating hormone value in dried blood collected routinely on filter paper on the third postnatal day of life for congenital hypothyroidism screening. Methods Retrospective case–control study. Cases were identified using data from our previously published study of 280 000 infants born in 10 maternity units in France in 2007-2014. Controls were identified among the 1 362 564 infants born in the Ile-de-France region during the same period. Results A screening thyroid-stimulating hormone level below 0.18 mIU/L on the third postnatal day had 71% (95% CI 44-90%) sensitivity, 99% (95% CI 99-100%) specificity, 81% (95% CI 74-86%) positive predictive value, and 98% (95% CI 97-99%) negative predictive value for detecting severe NH. By univariate regression analysis, the screening thyroid-stimulating hormone value was the strongest predictor of NH (P < .00001), with an area under the receiver-operating characteristics curve of 0.98 (95% CI 0.95-1.0). Expected frequencies were not significantly different from observed frequencies (Hosmer–Lemeshow test, P = .99). Conclusion The screening thyroid-stimulating hormone test can be used to detect severe NH, the optimal cut-off being 0.18 mIU/L. The additional cost compared with screening for congenital hypothyroidism would be small. Infants with neonatal hyperthyroidism would benefit from an earlier diagnosis with treatment initiation at the presymptomatic stage in many cases, ensuring optimal outcomes.