Short title: Genetic and non-genetic risk factors for preeclampsia
Keywords: Preeclampsia, Risk factors, Epidemiology, Meta-analysis, Umbrella reviewThis article is protected by copyright. All rights reserved.This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/uog.18959
Accepted Article
AbstractObjective: To summarize evidence from the literature on the risk factors associated with preeclampsia, assess the presence of statistical biases and identify associations with robust evidence.
Methods:We searched PubMed and ISI Web of Science from inception to October, 2016, to identify systematic reviews and meta-analyses of observational studies examining associations between genetic and non-genetic risk factors for preeclampsia.For each meta-analysis we estimated the summary effect size by random-effects and fixed-effects models, the 95% confidence interval and the 95% prediction interval. We estimated the between-study heterogeneity expressed by I 2 (considering above 75% as very large), evidence of small-study effects (large studies had significantly more conservative results than smaller studies and evidence of excess significance bias (too many studies with statistically significant results).Results: Fifty-seven eligible papers were identified providing data on 130 associations including 1466 primary studies, covering a very wide range of risk factors: co-morbid diseases, genetic factors, exposure to environmental agents and a range of biomarkers.Sixty-five (50%) associations had nominally statistically significant findings at P<0.05, while sixteen (12%) were significant at P<10 -6 . Sixty-four (49%) associations had large or very large heterogeneity. Evidence for small-study effects and excess significance bias was found in ten (8%) and twenty-six (20%) associations, respectively. Oocyte donation vs normal conception (OR 4.33, 95% CI: 3.11-6.03) had >1000 cases, 95% prediction intervals excluding the null, not suggestive of large heterogeneity (I 2 <50%), small-study effects (P for Egger's test>0.10), or excess of significance (P>0.05