Objective: To summarize the evidence from epidemiological studies examining the association between Helicobacter pylori infection and hyperemesis gravidarum.
Design: Systematic review and meta-analysis of case-control studies.
Material and methods: We searched the Medline/PubMed, Embase, Cinahl, ISI Web of Science and Biological Abstracts databases from 1966 to 17 June 2008 and finally Google Scholar. A total of 25 case-control studies were identified. Both fixed-effect and random-effect models were used to synthesize the results of individual studies. The Cochran Q, tau(2) of between-study variance and index of heterogeneity (I(2)) were used to evaluate heterogeneity. Heterogeneity between studies was examined by subgroup and random effect meta-regression analyses. Publication bias was evaluated.
Results: Publication bias was not observed. The random model pooled estimate was odds ratio = 3.32, 95 % confidence interval (CI): 2.25-4.90. A high heterogeneity was pinpointed (I(2) = 80 %, 95 % CI: 65-89). Subgroup analysis and meta-regression showed a weaker association in studies with a clear definition of hyperemesis gravidarum compared to studies without this condition, and weaker association in recent studies compared to earlier studies. Meta-regression showed that these two study characteristics explained 40% of heterogeneity between studies.
Conclusions: Exposure to H. pylori appears to be associated with an increased risk of hyperemesis gravidarum. The residual heterogeneity might have different reasons. Given the high prevalence of H. pylori, the public health consequence of H. pylori with regard to hyperemesis gravidarum may be important.