Skip to main content

26-05-2022 | Diabetes | News | Article

Meta-analysis highlights adverse pregnancy outcomes in gestational diabetes

Author: Laura Cowen

medwireNews: A systematic review and meta-analysis, which adjusted for confounding factors, has shown that gestational diabetes is associated with a range of adverse pregnancy outcomes that vary according to insulin use.

Writing in The BMJ, Fangkun Liu and colleagues from Central South University in Changsha, China, say that, up to now, “[n]o comprehensive study has assessed the relation between gestational diabetes mellitus and various maternal and fetal adverse outcomes after adjustment for confounders.”

They also note that previous studies have often been restricted to single clinical centers or regions, which limits their generalizability to more diverse populations.

The meta-analysis findings therefore “contribute to a more comprehensive understanding of the adverse outcomes of pregnancy related to gestational diabetes mellitus,” the authors remark.

Their analysis included 156 studies with data for 7,506,061 pregnancies, of which 68% had a high risk for bias.

The researchers report that, in studies with no insulin use (n=35), women with gestational diabetes were significantly more likely than those without diabetes to require cesarean section (odds ratio [OR]=1.16), after adjustment for at least minimal confounders.

This means that the studies were adjusted for at least one of seven confounding factors, namely maternal age, pregestational BMI, gestational weight gain, gravidity, parity, smoking history, and chronic hypertension.

The neonates of mothers with gestational diabetes but no insulin use had an increased odds for macrosomia (OR=1.70), being born large for gestational age (OR=1.57), having a low 1-minute Apgar score (OR=1.43), and being born preterm (odds ratio [OR]=1.51) relative to those of mothers without diabetes.

In studies in which the women were given insulin (n=63), there were no significant associations between gestational diabetes and adverse maternal outcomes, but neonates were at increased risk for requiring neonatal intensive care unit admission (OR=2.29), being large for gestational age (OR=1.61), and having respiratory distress syndrome (OR=1.57) or neonatal jaundice (OR=1.28) if their mothers had gestational diabetes.

Finally, in the studies that did not report details of insulin use, women with gestational diabetes had increased odds for induction of labor (OR=1.88), pre-eclampsia (OR=1.46), cesarean section (OR=1.38), and premature rupture of membrane (OR=1.13), while their neonates were at increased risk for hypoglycemia (OR=11.71), admission to the intensive care unit (OR=2.28), preterm delivery (OR=1.51), macrosomia (OR=1.48), and congenital malformation (OR=1.18).

There were no significant differences between women with and without diabetes in any of the analyses in the risks for instrumental delivery, shoulder dystocia, postpartum hemorrhage, stillbirth, neonatal death, low 5-minute Apgar score, low birthweight, and small for gestational age, after adjusting for confounders.

Subgroup analyses revealed that country status (developed vs developing), adjustment for BMI, and the screening methods used significantly contributed to heterogeneity among the studies for several adverse outcomes. For example, estimated ORs were generally lower in studies that used universal one-step diabetes screening than those that used the universal glucose challenge test or selective screening methods.

Liu et al say their findings “support the need for an improved understanding of the pathophysiology of gestational diabetes mellitus to inform the prediction of risk and for precautions to be taken to reduce adverse outcomes of pregnancy.”

However, they caution that “adjustment for at least one confounder had limited power to deal with potential confounding effects” but “defining a broader set of multiple adjustment variables was difficult” because adjustment factors differed across the studies.

“This major concern should be looked at in future well designed prospective cohort studies, where important prognostic factors are controlled,” they write.

medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2022 Springer Healthcare Ltd, part of the Springer Nature Group

BMJ 2022; 377: e067946

See the research in context now

with trial summaries, expert opinion and congress coverage