JanVayu Blog · Fact check

Delhi's recorded preterm births rose 51% in five years. Studies link PM2.5 to early birth, but the Delhi count cannot show it

Published: 4 November 2026 | Author: Team JanVayu | Reading time: 6 min

On 22 May 2026 Newslaundry published an investigation by Astha Savyasachi and Akankhya Rout, "5 years, over 21% rise in preterm birth rate: Delhi's air is cutting pregnancies short". A card summarising it has since circulated on Instagram, and it reached us as a screenshot. It makes four claims that readers will repeat: that fine particles can inflame and starve the foetus, that India carries the world's largest preterm birth burden, that China has a large body of research on pollution and pregnancy, and that India has no comparable study. We checked each against the sources. Most of it holds. One claim is too strong, and the Delhi count at the centre of the report cannot do the job the headline gives it.

What the report found

The authors obtained month-wise childbirth data from Delhi's Directorate of Family Welfare under the Right to Information Act. They report that preterm births rose from 21,134 in 2021 to 31,963 in 2025, an increase of 51%, and that the preterm share of live births rose from 9.8% to 11.9%, a relative rise of 21.4%. Preterm births peak in August or September every year. A baby born preterm then would have spent early pregnancy in the winter months when Delhi's air is at its worst. We have not seen the RTI data, so every Delhi figure here is as Newslaundry reports it.

What holds up

The burden. The Lancet's 2023 estimates (Ohuma and colleagues) put global preterm births in 2020 at 13.4 million, with India having the most, at 3.02 million. The interval around India's figure is wide, 2.2 to 4.0 million. The report and the card say India had "more than 23%" of the world's total. The Lancet paper itself says "over 20%", and 3.02 million of 13.38 million is 22.6%. The "more than 23%" appears in a 2025 review that quotes the same 3.02 million, and it does not follow from those numbers. "About a fifth to a quarter" is what the data supports.

The mechanism. Inflammation, oxidative stress and reduced oxygen and nutrient supply are plausible routes, and one Indian study points the same way. A hospital cohort of 406 pregnant women (Parikh and colleagues, 2026) found a significant association between each 10 µg/m³ of PM2.5 and cord-blood levels of four inflammatory markers (CRP, IL-6, TNF-α and IL-1β), and black carbon was found on the foetal side of the placenta only in the exposed group. That is one small cohort. It supports plausibility and does not explain a city-wide rise.

The association. A 2021 meta-regression in PLOS Medicine (Ghosh and colleagues) pooled studies and found that each 10 µg/m³ of ambient PM2.5 went with a 12% higher risk of preterm birth (risk ratio 1.12, 95% interval 1.06 to 1.19), an 11% higher risk of low birth weight, and a birth weight about 22 grams lower. Most of those studies came from North America, Europe and Australia. The authors note that a few recent ones came from China, India, sub-Saharan Africa and South America.

Where "India has no comparable study" goes too far

India has several studies, and most of them use national surveys linked to satellite estimates of PM2.5:

What is true is narrower. These studies rely on survey records, satellite exposure and, in India's case, mostly birth weight and mortality. The 406-woman hospital cohort is the only clinical cohort we found with measured outcomes and biomarkers, and its authors write that the link between PM2.5 and adverse pregnancy outcomes in India "remains largely unexplored". We did not check the Chinese multi-city studies the report cites, so we cannot say how closely India's evidence compares with them. "India has less and weaker evidence than it should" is defensible. "India has none" is not.

What the Delhi count cannot show

It holds no exposure data. The RTI figures are counts of births by month. No mother's exposure is attached to any of them, so the count shows that preterm births peak in August and September and cannot show why.

The rise may partly be in the denominator. If 21,134 preterm births were 9.8% of live births in 2021 and 31,963 were 11.9% in 2025, the implied number of recorded live births rose from about 216,000 to about 269,000, an increase of roughly a quarter in four years. That is our arithmetic from rounded rates, so the range is about 23% to 26%. The report does not explain it. A change in which facilities report, in how gestational age is recorded, or in migration into the city could move the count and the rate without any change in biology. We do not know whether any of these happened.

Heat is a competing explanation for the August peak. A baby born at 34 to 36 weeks in August spent the last trimester in the Delhi summer. A 2020 BMJ systematic review (Chersich and colleagues, 70 studies in 27 countries) found that the odds of preterm birth rose 1.05-fold per 1 °C increase in temperature and 1.16-fold during heatwaves. That does not rule out pollution, and the two can act together. It means the August peak cannot separate them without exposure data.

Clinical impressions are not a count. The doctors quoted describe pregnancies in which they found no explanation other than the air. One of them, Dr Archana Dhawan Bajaj, says air pollution "cannot be identified as the sole cause" and that the evidence to settle it "does not exist yet". We agree with her. Two further claims in the report, that pollution reduces the body's absorption of folic acid and that a 1,400-pregnancy Varanasi study showed that "the evidence is sufficient to infer a causal relationship", we have not verified, and we have not repeated them.

What would settle it

The question can be answered with the right records. If facility-level birth records from Delhi, with dated gestational age and the mother's area of residence, were released with identifying details removed, they could be paired with daily CPCB readings. That would let researchers compare mothers in the same hospital and the same season by their exposure, and test heat and pollution separately. We suggest researchers and journalists ask for such a file, and we will add any such study to the Reading List.

JanVayu has added the Indian and South Asian studies above that were missing to its Reading List, with downloadable citation files.

Sources

Share WhatsApp X LinkedIn Telegram Email