Using record linkage to investigate perinatal factors and neonatal thyroid-stimulating hormone.

Abstract

Studies examining the relationship between maternal and infant thyroid parameters have shown conflicting results. Record linkage provides an opportunity to examine the association between maternal and infant thyroid-stimulating hormone (TSH) levels. Our aim was to demonstrate the feasibility of record linkage of newborn screening (NBS), laboratory and birth databases for research by investigating the association between maternal and newborn TSH levels. The records of 2802 women with first trimester serum TSH concentrations were linked with population-based birth data and NBS data containing infant TSH levels. Association between moderately high neonatal TSH levels (>5 mIU/L) and maternal and infant characteristics was evaluated. The correlation and association between maternal and infant TSH levels were assessed using Pearson's correlation coefficient and multivariable linear regression, respectively. Of maternal and birth records, 99.3% linked with an NBS record. Mother's country of birth, gestational age (>41 weeks) and lower birthweight were associated with neonatal TSH levels >5 mIU/L. Neonatal and maternal first trimester TSH levels were not correlated, although statistically significant (r = 0.05, P = 0.008). The association between neonatal TSH and maternal TSH, after adjusting for maternal age, gestational age and age at NBS testing, was also small (b = 0.039, P = 0.009). Record linkage is a feasible and cost-efficient way to investigate the association between maternal factors and neonatal hormone levels. First trimester maternal thyroid levels are not correlated with neonatal TSH levels. This method of outcome assessment can be used for future research examining long-term outcomes for infants with different NBS results.