Meta-analysis of ART outcomes in women with different preconception TSH levels

To assess whether elevated thyroid-stimulating hormone (TSH) levels before conception can predict poor outcomes of assisted reproductive technology (ART).

Prior to July 2018, we searched the PubMed, EMBASE, COCHRANE, Google Scholar, and CNKI databases for studies. Retrospective or prospective reports that compared ART results in patients with subclinical hypothyroidism (SCH) with normal thyroid function were selected. Two reviewers separately reviewed each potential article for qualification, analyzed the quality of the studies according to the Newcastle-Ottawa scale, and extracted the data. The PRISMA guidelines were adopted.

We selected a total of 18 publications that included 14,846 participants for this meta-analysis. When the TSH cut-off value for SCH was set at 2.5 mIU/L, no significant differences were observed in ART-related outcomes between SCH patients and normal women. The evaluated outcomes included the live birth rate (LBR) (OR: 0.93; 95% CI (0.77,1.12), P = 0.43), clinical pregnancy rate (CPR) (OR:1.02; 95% CI (0.90,1.17); P = 0.74), pregnancy rate (PR) (OR: 1.00; 95% CI (0.89,1.12); P = 0.99), and miscarriage rate (MR) (OR:1.24; 95% CI (0.85, 1.80); P = 0.26). Furthermore, when a higher TSH level was used as the cut-off value to diagnose SCH (i.e., 3.5-5 mIU/L), a significant difference was found in the MR (OR: 1.91; 95% CI (1.09, 3.35); P = 0.02) between the two groups of ART-treated women. However, when a broader cut-off value was used to define SCH, no significant differences were observed in the LBR (OR: 0.72; 95% CI (0.47,1.11); P = 0.14), CPR (OR: 0.82; 95% CI (0.66,1.00); P = 0.052), or PR (OR: 1.07; 95% CI (0.72,1.60); P = 0.74) between the two groups of ART-treated women.

No difference was observed in ART outcomes when a TSH cut-off value of 2.5 mIU/L was used. However, when a broader TSH cut-off value was used, preconception SCH resulted in a higher miscarriage rate than in normal women.