Risk of pre‐eclampsia in women taking metformin: a systematic review and meta‐analysis

Aims To perform meta‐analyses of studies evaluating the risk of pre‐eclampsia in high‐risk insulin‐resistant women taking metformin prior to, or during pregnancy. Methods A search was conducted of the Medline, EMBASE, Web of Science and Scopus databases. Both randomized controlled trials and prospective observational cohort studies of metformin treatment vs. placebo/control or insulin either prior to or during pregnancy were selected. The main outcome measure was the incidence of pre‐eclampsia in each treatment group. Results Overall, in five randomized controlled trials comparing metformin treatment (n = 611) with placebo/control (n = 609), no difference in the risk of pre‐eclampsia was found [combined/pooled risk ratio (RR), 0.86 (95% CI 0.33–2.26); P = 0.76; I2 = 66%]. Meta‐analysis of four cohort studies again showed no significant effect [RR, 1.21 (95% CI 0.56–2.61); P = 0.62; I2 = 30%]. A meta‐analysis of eight randomized controlled trials comparing metformin (n = 838) with insulin (n = 836), however, showed a reduced risk of pre‐eclampsia with metformin [RR, 0.68 (95% CI 0.48–0.95); P = 0.02; I2 = 0%]. No heterogeneity was present in the metformin vs. insulin analysis of randomized controlled trials, whereas high levels of heterogeneity were present in studies comparing metformin with placebo/control. Pre‐eclampsia was a secondary outcome in most of the studies. The mean weight gain from time of enrolment to delivery was lower in the metformin group (P = 0.05, metformin vs. placebo; P = 0.004, metformin vs. insulin). Conclusions In studies randomizing pregnant women to glucose‐lowering therapy, metformin was associated with lower gestational weight gain and a lower risk of pre‐eclampsia compared with insulin. What's new? This study shows that metformin with and without insulin treatment is associated with a lower incidence of pre‐eclampsia than insulin treatment alone in women with gestational diabetes mellitus or Type 2 diabetes; this is probably linked to reduced weight gain during pregnancy in those taking metformin. In other high‐risk pregnancies where hypoglycaemic agents are not essential, metformin does not appear to be beneficial. Clinical guidelines, which recommend insulin as the first‐line treatment in pregnancies complicated by diabetes, should be reviewed in light of these findings, and adequately designed randomized controlled trials with pre‐eclampsia as a primary outcome carried out.