ESPE Abstracts

Background: There is a link between congenital hypothyroidism (CH) and neurodevelopmental abnormalities, but no longitudinal studies have yet identified reliable quantifiable measures.

Purpose: To evaluate associations between CH and abnormalities in neurodevelopment and growth, and identify the timing of these abnormalities and the best time for intervention.

Methods: Data from the National Health Insurance System (NHIS) were used to analyze 919,707 children born in Korea from 2008 to 2009. Eligible children were tracked annually until the age of 72 months. Claims-based definitions for CH were developed using a combination of ICD-10 diagnosis codes (E03.0 or E03.1) and use of thyroid hormone medication for 5 or more years. The primary outcome was suspected neurodevelopmental disorder in five domains of the Korean Ages and Stages Questionnaires (K-ASQ): gross motor skills, fine motor skills, communication, problem-solving, personal/social ability. The secondary outcomes were anthropometric measures. A generalized estimating equation (GEE) model was used to investigate the relationship of neurodevelopmental outcomes and anthropometric growth.

Results: The prevalence of CH in our population was 0.05% (n= 408). Relative to the control group, the CH group had a higher risk of suspected neurodevelopmental disorders based on the total K-ASQ score (propensity score [PS] –weighted, odds ratio: 4.52, 95% CI: 2.91, 7.02), and significantly increased risk in each of the five K-ASQ domains. There was no interaction of time with outcome based on analysis of the date of the developmental examination (all P for interaction > 0.05). The association of CH with neurodevelopmental disorders was greater when thyroid medication began after the age of 6 weeks than at or before the age of 6 weeks. Moreover, neurodevelopmental outcome during each screening round was worse when initiation of treatment was delayed (all P < 0.001). The CH group also had a higher risk for low height-for-age z-score (PS-weighted aOR: 2.97, 95% CI: 1.83, 4.80), but elevated BMI-for-age z-score had no effect.

Conclusions: The CH group had worse neurodevelopmental outcome overall and in each of the five K-ASQ domains, and also had reduced height-for-age z-score, but there was no interaction of time with the outcome measurement. Outcomes were worse when the onset of treatment was increasingly delayed.