Summary: In 200 women with gestational diabetes mellitus, a single-centre randomised trial found oral metformin no different from insulin for glycaemic control (fasting glucose, HbA1c and 2-hour postprandial glucose all P>0.05). Several secondary outcomes (neonatal birth weight, neonatal and maternal hypoglycaemia, NICU admission, compliance) favoured metformin (P<0.05), though no effect sizes or confidence intervals were reported.
PICO Summary
| Element | Detail |
|---|---|
| Population | 200 women with gestational diabetes mellitus; single-centre randomised controlled trial, university hospital outpatient clinics, Egypt, March to August 2024. |
| Intervention | Oral metformin (arm n not separately reported; presumed approximately 100). |
| Comparison | Insulin injections (arm n not separately reported; presumed approximately 100). |
| Outcome | No significant between-group difference in fasting blood glucose, HbA1c or 2-hour postprandial glucose (all P>0.05). Metformin associated with lower neonatal birth weight, fewer cases of neonatal hypoglycaemia and NICU admission, and lower maternal hypoglycaemia, with higher compliance (P<0.05). Overall maternal and neonatal complications reported as similar. No effect sizes, 95% CIs, exact p-values, event counts or ARR/NNT were reported in the abstract. |
Expert Commentary
The verdict on this trial is modest, not paradigm-shifting despite the title. For the primary aim of glycaemic control, metformin was statistically no different from insulin, so the result is best read as compatibility rather than superiority for glucose lowering. The secondary signals favouring metformin, lower neonatal birth weight and fewer episodes of neonatal and maternal hypoglycaemia, are biologically plausible and align with the broader literature, but they must be treated cautiously here. The principal limitation is that the report provides no effect sizes, no confidence intervals and no event counts, only bare P-value thresholds, which makes the magnitude and precision of any benefit impossible to judge and raises the risk of selective emphasis on positive secondary endpoints. The single-centre design, modest sample, short recruitment window and the inherently unblinded nature of comparing an oral tablet against injected insulin all temper confidence further; no industry or manufacturer sponsorship was declared. Can I use this with my patients? Cautiously, yes, for the woman with mild gestational diabetes who declines or cannot manage insulin, metformin remains a reasonable first-line option, but this trial does not change that calculus. I would like to see the full paper with absolute numbers, randomisation detail and longer-term neonatal follow-up before weighting these secondary outcomes heavily.
References
Abd El Tawwab AE, Mahmoud MS, El-Kady HA, Elashiry AA, Elsayed HG, Seifelnasr MF. Changing Paradigms: Neonatal Outcomes after Oral Metformin versus Insulin in the Control of Gestational Diabetes Mellitus (Randomized controlled trial). Clin Ter. 2025;176(2):148-149. doi:10.7417/CT.2025.5197