Exercise in Pregnancy Cuts Risk of Excess Weight Gain and Oversized Babies, Major Analysis Finds

For decades, pregnant women have been told to stay active, yet the precise payoff of that advice has remained frustratingly vague. How much movement is enough? Does timing matter? And does exercise actually protect against the two of the most common pregnancy complications linked to weight: gaining too many pounds and delivering an unusually large baby? A new dose-response meta-analysis published in the International Journal of Obesity offers some of the clearest answers yet, and its headline finding is striking: the most physically active pregnant women had a 30 percent lower risk of excessive gestational weight gain and a 16 percent lower risk of delivering a large-for-gestational-age infant compared with the least active.

The study, led by Ziqin Lin and colleagues at Beijing Sport University, is not a single clinical trial but a synthesis of 34 independent observational studies pulled from eight databases. That scale matters. Individual studies of pregnancy and exercise often disagree, partly because they measure activity in incompatible ways—some count brisk walks, others tally structured workouts, and still others rely on questionnaires of everyday movement. By converting every study’s activity data into a common currency, metabolic equivalent hours per week (MET-h/week), the team could finally stack the evidence on a single ruler and ask how risk changes with each increment of activity.

The technical machinery behind the analysis is worth appreciating. The researchers scored each study’s methodological rigor with the Newcastle-Ottawa Scale, tested for publication bias using Begg’s and Egger’s tests, and—crucially—modeled the dose-response relationship with restricted cubic splines, a flexible statistical technique that lets the data reveal whether risk falls in a straight line or curves at higher activity levels. All computations were performed in Stata 16.0. The verdict: the relationship between physical activity and excessive gestational weight gain is essentially linear, meaning more activity is consistently associated with lower risk, with no obvious plateau or threshold within the observed range.

The numbers translate into something every expectant mother can grasp. Women who accumulated 10 MET-h/week of physical activity—roughly equivalent to about three hours of brisk walking spread across a week—had a 4 percent lower risk of excessive gestational weight gain than the least active women. At 20 MET-h/week, the risk reduction grew to 9 percent, and at 50 MET-h/week it reached 20 percent. These are modest-sounding percentages, but applied across the millions of pregnancies each year in which women exceed recommended weight-gain limits, the population-level impact could be substantial.

Perhaps the most clinically provocative finding concerns timing. When the researchers split the analysis by trimester, the second trimester showed no statistically significant dose-response association between activity and weight-gain risk. The third trimester, by contrast, was where the action was: 10, 20, and 50 MET-h/week of activity in those final months corresponded to 8, 16, and 35 percent lower risks of excessive weight gain, respectively. The dose-response curve for the third trimester was also the steepest, with relative risk falling to roughly 0.45 at 120 MET-h/week and confidence intervals remaining consistently below 1.0—a sign of statistical robustness rather than a fleeting signal.

That pattern makes physiological sense. The third trimester is when fetal growth accelerates and maternal insulin resistance rises sharply, driven by placental hormones that deliberately shift the mother’s metabolism toward glucose sparing for the baby. When that shift overshoots, excess glucose fuels both maternal fat deposition and fetal overgrowth. Physical activity is one of the most powerful non-pharmacological levers on insulin sensitivity, enhancing glucose uptake in muscle through both insulin-dependent and contraction-mediated pathways. In late pregnancy, when insulin resistance peaks, the metabolic counterweight provided by regular movement may be at its most valuable.

The implications extend beyond the mother’s scale. Large-for-gestational-age infants—those weighing more than the 90th percentile for their gestational age—face elevated risks of birth trauma, shoulder dystocia during delivery, cesarean birth, and later childhood obesity. Excessive gestational weight gain is itself a well-documented predictor of delivering an LGA baby, independent of pre-pregnancy body mass index, and it also predicts postpartum weight retention, setting up a trajectory toward long-term cardiometabolic risk for the mother. By trimming both outcomes, physical activity appears to act on a shared upstream mechanism rather than two unrelated problems.

The LGA findings, while less dramatic than the weight-gain results, still point in a protective direction. Across trimester subgroups, the pooled relative risks for LGA were 0.82 for first-trimester activity, 0.78 for second-trimester activity, and 0.61 for third-trimester activity, with an overall pooled estimate of 0.84. Not every subgroup was statistically clean—the combination of first and second trimester activity yielded a relative risk of 1.07, slightly above the null—but the overall pattern, particularly the strong third-trimester signal, reinforces the same biological story: late-pregnancy movement helps keep fetal growth on track.

The authors are careful about what their design can and cannot prove. As a meta-analysis of observational studies, it establishes associations, not causation; women who exercise during pregnancy may differ in diet, socioeconomic status, baseline health, or pregnancy complications in ways that partially explain the results. The team also confronted publication bias directly, using funnel-plot asymmetry and trim-and-fill imputation to probe whether smaller studies with null findings were missing from the literature. And heterogeneity across the included studies—inevitable when pooling research from different countries, populations, and measurement tools—was quantified with the I-squared statistic throughout. These caveats temper, but do not erase, a dose-response gradient that is difficult to explain entirely by confounding.

For clinicians and guideline writers, the study arrives at a moment when most major bodies—including the American College of Obstetricians and Gynecologists and the World Health Organization—already recommend at least 150 minutes of moderate-intensity activity per week during uncomplicated pregnancies. What this analysis adds is the dose-response architecture beneath that recommendation: a linear relationship with weight-gain risk, a quantified benefit per increment of MET-hours, and a pointed reminder that the third trimester is not a time to retreat to the couch. Maintaining adequate activity throughout pregnancy, and especially in its final stretch, may be one of the simplest, safest interventions available to bend the curve on two of obstetrics’ most common weight-related complications.

Subject of Research: Dose-response association between physical activity during pregnancy and risks of excessive gestational weight gain and large-for-gestational-age infants

Article Title: Physical activity during pregnancy and the risks of excessive gestational weight gain and large-for-gestational-age: a dose-response meta-analysis of observational studies

Article References: Lin, Z., Zhang, L., Li, H., Ren, Y., Dai, X., Zhou, H., Zhou, C., & Li, R. (2026). Physical activity during pregnancy and the risks of excessive gestational weight gain and large-for-gestational-age: a dose-response meta-analysis of observational studies. International Journal of Obesity. https://doi.org/10.1038/s41366-026-02236-7

Image Credits: AI Generated

DOI: 10.1038/s41366-026-02236-7

Keywords: pregnancy, physical activity, gestational weight gain, large-for-gestational-age, meta-analysis, dose-response, third trimester, MET-h/week, obstetrics, maternal health, fetal growth, insulin resistance