Non-genetic risk indicators for preterm birth: An umbrella review of systematic reviews and meta-analyses

This umbrella review, aimed to summarize systematic evidence on non-genetic risk indicators that are present at the start of pregnancy and their potential association with preterm birth.

We searched Medline and Embase up to June 30, 2021, for systematic reviews and meta-analyses of observational studies. We included systematic reviews and meta-analyses on observational studies evaluating putative risk indicators that are present at conception. Studies that assessed indicators that arise after conception, and were unlikely to be present before pregnancy were excluded. Preterm birth was analyzed as an overall outcome, as subtype-specific data were not consistently reported. For each meta-analysis, we calculated summary effect sizes, 95% confidence and prediction intervals, heterogeneity (I²), small-study effects, and excess significance bias. Evidence was graded from convincing to non-significant. Methodological quality was assessed using AMSTAR 2.

In total, 134 systematic reviews were included, covering 163 meta-analyses (total sample size = 170,564,182 women). Eleven showed convincing evidence and 34 highly suggestive evidence. Ten associations had convincing or highly suggestive evidence with an odds ratio or relative risk > 2: advanced maternal age, previous singleton preterm birth, previous stillbirth, amphetamine use, cocaine use, maternal personality disorder, anti-TNF-α agents, systemic lupus erythematosus, chronic hypertension, and cold knife conization. No indicators showed strong protective effects. These risk indicators should be considered a core set of baseline characteristics in future research and risk stratification efforts related to preterm birth.

This umbrella review maps many indicators that have strong associations with preterm birth, these indicators should be considered as a core set of baseline characteristics background knowledge in future work on overall preterm birth.