Background: Metformin is increasingly accepted as an alternative to insulin therapy in gestational diabetes mellitus (GDM). The Metformin in Gestational Diabetes (MiG) trial reported similar pregnancy outcomes for metformin versus insulin; however, supplemental insulin was required in 46% of women on metformin.

Aims: We aimed to identify predictors of response to metformin monotherapy in women with GDM attending a general hospital antenatal clinic. Methods We offered extended-release metformin to women diagnosed with GDM (ADIPS 1998 criteria) at ≥24 weeks of gestation. If glucose targets were not achieved (≤5.0 mmol/L fasting, ≤6.7 mmol/L two-h post-meal), women were changed to insulin. We carried out an audit to determine characteristics of metformin responders versus nonresponders.

Results: Twenty-five women chose initial metformin therapy; 16 (64%) achieved satisfactory glycaemic control (responders). Nine women (36%) were changed to insulin: seven due to inadequate control (nonresponders) and two had metformin intolerance. Fasting glucose at oral glucose tolerance test (OGTT) was significantly lower in metformin responders versus nonresponders; two-h glucose and BMI did not differ. Ninety-three percent of women with fasting glucose ≤5.2 mmol/L responded to metformin: conversely, at fasting glucose >5.2 mmol/L, 33% responded (P = 0.005). Neonatal outcomes were similar in metformin responders and nonresponders, women who chose initial insulin therapy (n = 25), or were diet-controlled (n = 21).

Conclusions: In women with GDM, fasting glucose on OGTT predicted response to metformin: at fasting glucose ≤5.2 mmol/L, the probability of response was 93%. Antenatal clinics should determine locally relevant predictors of response to metformin in women with GDM.