Climate change’s most dangerous health effect might not be the heatstroke deaths that make headlines during a heat wave. It could be happening quietly, weeks before a baby is even born.
A pregnancy exposed to one extra degree of heat carries a measurably higher risk of ending early. An international team led by researchers at the University of the Witwatersrand pulled together 198 studies from 66 countries, covering millions of pregnancies, to answer a question that had never been quantified across the full picture of maternal, fetal and neonatal health at once: exactly how much does heat exposure raise the odds of things going wrong.
The answer, published in Nature Medicine, is unambiguous. For every 1°C increase in heat exposure during pregnancy, the odds of a premature birth rose by 4%. During an actual heat wave, that risk jumped by 26%. Stillbirths, congenital anomalies and gestational diabetes all climbed too, with the odds of any obstetric complication rising by a quarter during heat waves.
The reason comes down to basic physiology. Pregnancy already pushes the cardiovascular system harder, with more blood volume, more heat generated by the fetus, and a body working overtime to dissipate it all through sweating and dilated blood vessels. When conditions like hypertension or dehydration are already straining that system, extreme heat can tip it over. Researchers believe elevated body temperature, hormonal disruption and inflammation from heat shock proteins may all play a role in triggering labor too early or damaging a developing fetus.
Timing matters just as much as intensity. The study found that different complications have different windows of vulnerability. Gestational diabetes risk climbed mainly when heat hit during the second trimester, right around the time doctors screen for it. Hypertensive disorders were more closely tied to heat exposure in the very first weeks after conception, when the placenta is still forming. Preterm birth, the outcome with the strongest and most consistent evidence, showed risk building from both short bursts of heat in the final weeks before delivery and prolonged heat exposure earlier in the pregnancy.
The burden falls hardest on the countries least responsible for warming the planet. When researchers split the data by national income, the odds of a heat-related preterm birth were 1.61 in low-income countries, compared with 1.11 in high-income ones. Yet only 3% of the 198 studies came from low-income countries at all, meaning the places likely suffering the most are also the ones science has studied the least.
One American study included in the review followed 56 million births and estimated that extreme heat days were already shortening pregnancies enough to account for roughly 25,000 babies born earlier than they should have been in a single year, and 150,000 lost days of gestation overall. A separate analysis attributed more than 4,600 preterm births a year in China directly to heat waves linked to human-caused climate change, with an economic cost projected to exceed a billion dollars annually.
The authors are careful about the limits of the evidence. Almost all of the underlying studies are observational, meaning they can show a strong statistical association between heat and these outcomes without proving the exact biological chain of cause and effect in every case, and the ways different studies measured heat exposure varied enough to complicate direct comparisons. Even so, with effect estimates from 221 separate analyses almost universally pointing in the same harmful direction, the researchers describe the pattern as one of the most consistent signals in climate health research to date, and they are now calling for pregnant women and newborns to be explicitly written into national climate adaptation plans, from early heat warnings to cooling in maternity wards.
Source: Lakhoo, D.P., Brink, N., Radebe, L. et al. “A systematic review and meta-analysis of heat exposure impacts on maternal, fetal and neonatal health.” Nature Medicine, 2025. DOI: 10.1038/s41591-024-03395-8.








