Detection of Congenital Hypothyroidism in Premature Infants

Source: Shah AN , Li W , Zheng D , et al . A practical gestational age-based algorithm for timely detection of hypothyroidism in premature infants (published online ahead of print May, 11, 2024). J Perinatol. 2024; doi: 10.1038/s41372-024-01985-5.Investigators from the University of Texas Health Sciences Center at Houston, Houston, TX, and the University of Wisconsin, Madison, WI, conducted a retrospective study to evaluate the use of an algorithm for repeated screening for congenital hypothyroidism in premature infants. The algorithm was developed for use in newborns born at <35 weeks’ gestation. In addition to screening for congenital hypothyroidism as part of routine statewide neonatal screening programs, the algorithm recommends obtaining thyroid-stimulating hormone (TSH) and free thyroxine (fT4) levels at approximately 1 month of age (21–35 days of life). Recommended follow-up, including repeated testing in 1–2 weeks or at discharge or endocrinology consult for possible treatment with levothyroxine (L4), is based on the TSH level category (low, <0.5 uIU/mL; normal, 0.5–6.0 uIU/mL; mildly elevated, 6.1–10 uIU/mL; or elevated, >10 uIU/mL) and fT4 level.To evaluate the outcomes of the screening algorithm, medical record data on infants born at <35 weeks’ gestation between June 2018 and April 2021 and treated at a single NICU where the algorithm was used were reviewed. In addition to state newborn screening results and fT4 and TSH levels, data abstracted on study participants included gestational age at birth, birth weight, and treatment with LT4. For those receiving LT4, length of treatment with LT4 was collected from the medical record to assess whether the patient had transient or permanent hypothyroidism. Differences between infants treated for hypothyroidism and those without hypothyroidism were assessed with Wilcoxon rank sum tests. In addition, rates of treatment in the 10 years prior to implementation of the algorithm and following implementation were compared.Data were analyzed on 938 preterm infants. Of these, 29% were born at <28 weeks’ gestation, and 39% were born at 28 to <32 weeks’ gestation. A total of 16 infants were treated with LT4, yielding an incidence of treated hypothyroidism of 1.7%. Three of the 16 patients were identified by the state newborn screen. Mean gestational ages were similar among those treated for hypothyroidism and infants without hypothyroidism (30.6 and 30.4 weeks’ gestation, respectively; P = 0.8), but those in the treatment group had significantly lower birth weights (mean, 815g and 1,350g, respectively; P = 0.02). Four infants treated for hypothyroidism died prior to NICU discharge. Of the remaining 12, 2 were lost to follow-up, 2 were classified with transient hypothyroidism, and 8 with permanent hypothyroidism. Overall, the rate of treatment for hypothyroidism was significantly higher following implementation of the algorithm compared to the previous 10 years (1.7% vs 1.0%; P = 0.03). This increase was because of more treatment of premature infants with birth weights ≥2,500g. There was no significant difference in rate of treatment for those with birth weights <2,500g.The authors conclude that repeated fT4 and TSH testing led to early detection of hypothyroidism in premature infants.Dr Beam has disclosed no financial relationship relevant to this commentary. This commentary does not contain a discussion of an unapproved/investigative use of a commercial product/device.Early detection of congenital hypothyroidism (CH) on newborn screens and prompt treatment have resulted in a reduction of the long-term neurodevelopmental consequences of the untreated disorder.1 However, questions persist regarding treatment of CH, particularly in preterm infants. There are no clear guidelines on the treatment threshold, what dose of LT4 to initiate, or duration of treatment in infants <35 weeks’ gestation.2 Consultation with a pediatric endocrinologist is necessary for treatment and follow-up but can be challenging in some areas of the country where specialists are scarce.3Although the results of the current study demonstrate improved detection of CH at a single institution, questions remain regarding the method of detection used on newborn screens and whether utilizing a more sensitive approach across all states may improve detection rates. The primary TSH method is the least expensive and most utilized.4 Ultimately, the gestational age-based algorithm put forth by the authors provides a streamlined approach to follow-up on newborn screens despite the testing method implemented on the newborn screen.Finally, understanding populations that are at increased risk of CH, whether transient, mild, or severe, is critical in identifying infants with this disorder. Of the infants identified in the current study, extremely low birth weight infants, infants with trisomy 21, and infants with critical congenital heart disease were over-represented in comparison to the general population.CH has increased prevalence in premature infants; using an expanded algorithm examining TSH and fT4 levels can improve detection in this vulnerable population.The serial screening algorithm employed by the current authors closely aligns with the 2023 AAP recommendations for CH screening and management.4