Objective: Thyroid dysfunction, a common complication of pregnancy, is associated with adverse obstetric and neonatal consequences. This study aimed to determine the effect of TSH levels on early pregnancy outcome in a prospective population-based cohort study.

Design and methods: The serum TSH, free thyroxine, free triiodothyronine, thyroid peroxidase antibody levels and urinary iodine concentration of 418 pregnant women in their first trimester of pregnancy were measured. According to the American Thyroid Association (ATA) and the local reference ranges for TSH, women were divided into two groups of 0.1-2.5, >2.5 mIU/L and 0.2-4.6, >4.6 mIU/L. The risk of spontaneous abortion (SA) was calculated for each group.

Results: Spontaneous abortion was detected in 7.2% (n = 30) of total 418 pregnancies. Women with TSH levels > 2.5 mIU/L had an increased risk of SA, compared to women with TSH levels of 0.1-2.5 mIU/L (relative risk [RR] 3.719, 95% confidence interval [CI]:1.713-8.074). The risk of SA was increased in women with TSH levels > 4.6 mIU/L (RR 5.939, 95% CI: 1.711-20.620). The rate of SA was increased by 78% for every unit increase in standard deviation of TSH concentration (RR 1.35, 95% CI: 1.09-1.70). The rate of miscarriages in the treated group by levothyroxine was 9.8% (n = 6) compared to 28.6% (n = 8) in the untreated group (P = 0.024).

Conclusions: Our finding suggests that the upper limit for the TSH normal range should be redefined to <2.5 mIU/L during the first trimester of gestation. The local upper limit was 4.6 mIU/L, consistent with 4.0 mIU/L cut-off value recommended by the ATA.