ESPE Abstracts

Background: Transient hypothyroxinaemia of prematurity (THOP), defined as low levels of FT4 without the expected TSH surge, may concern up to 50% of infants born <30 weeks’ gestational age. Most studies showed a link between THOP and impaired neurodevelopment, as shown in the review of Eerdekens. Data about the benefit of supplementation are scarce, with few randomised trials and inconclusive results. To date, there are no clear recommendations regarding screening, diagnosis criteria and treatment of THOP.

Aim: To compare neurodevelopment at two years of corrected age between newborns with treated and untreated THOP.

Materials and methods: This study was a multicentered, retrospective cohort study, based on prospectively collected data in 4 neonatal ICU, during years 2009 to 2020. Newborns were included if they were born before 32 weeks of gestation (GW) and underwent a thyroid function test. They were divided into 3 groups: noTHOP, treated THOP and untreated THOP. Primary outcome was neurodevelopment at two years of corrected age, evaluated by the child’s paediatrician as normal or not. Secondary outcomes were revised Brunet Lezine score at 2 years, ASQ-3 scores.

Results: 601 newborns met inclusion criteria, 373 were included. Compared to noTHOP group, subjects with THOP had lower gestational age, lower birth weight and younger age at first thyroid evaluation (20 days versus 25 days). Subjects from the treated group had lower FT4 than untreated ones (8.4 pmol/L versus 9.4 pmol/L). In the treated group, L-thyroxine was started at median age of 22 days at mean dose of 7.5 µg/kg/j and continued until 39 GW. 266 subjects were analysed. There was no statistically significant difference in neurodevelopment at two years of corrected age between the 3 groups. We showed a tendency for better neurodevelopment in noTHOP group as compared with THOP group (OR 1.37 [0.8-2.3]) and worse neurodevelopment in untreated group as compared with treated group (OR 0.8 [0.3-1.9]). Results remained the same after adjusting for confounding factors. There were no statistically significant differences on secondary outcomes.

Conclusion: Premature newborns who received LT4 for THOP may have a better neurodevelopment than untreated infants. Our results are not statistically significant but relied on an observational study with inherent bias (the biggest being the absence of randomisation for treatment allocation). The absence of statistically significant difference does not mean the absence of effect. Randomised controlled trials should follow this study to improve our knowledge on this subject.