Challenges in screening for congenital hypothyroidism: Optimization of thyrotropin cut-off values.

Abstract Background aim Different protocols exist for newborn screening of congenital hypothyroidism (CH) worldwide, with different thyrotropin cut-off values for repetition and confirmation tests. This study aimed to assess local protocol in terms of specificity and improve our screening process by optimizing thyrotropin cut-off values. Subsequently, the cut-off values obtained were retrospectively applied to evaluate the number of tests avoided. Methods Retrospective observational study between 2013-2019. All newborn children with a confirmation test for CH were considered for the study. ROC curve analysis was performed for thyrotropin cut-off value optimization in DBS which triggers a confirmatory test, and odds ratios were calculated. For individuals affected by the cut-off value modification, serum thyrotropin and free thyroxine in the confirmation test were analyzed for consideration of clinical outcomes. Results A total of 72,133 newborn children were screened for CH, and 208 individuals were included in the study. Incidence in our population was 1:2,000 live births. The area under the ROC curve was 0.819 (CI 95%: 0.748–0.897). While the current cut-off value (thyrotropin≥10mIU/L) had a specificity of 31.8% [ORs: 3.5 (CI 95%: 1.4–8.8)], the optimal cut-off value (thyrotropin≥15mIU/L) yielded a specificity of 92.4% for the detection of CH and transient hypothyroidism [ORs: 15.9 (CI 95%: 7.1–35.8)], with no loss of sensitivity. Discussion While keeping a maximum sensitivity, optimization of cut-off values may be of great use not only in management, but also in reducing family stress, which is of special relevance for the newborn.