• Energy
    • Oil and gas
    • Coal
    • Emissions
    • Renewables
    • Technology
  • Policy
    • International
    • Africa
    • Asia & Pacific
    • Europe
    • North America
    • South America
  • Nature
    • Food and farming
    • Plants and forests
  • Finance
    • Public finance
    • Private finance
  • Science
    • IPCC
    • Oceans
    • Temperature
    • Extreme weather
  • Insights
    • Briefings
    • Series
    • Newsletters
    • Unlocking key terms
  • ZCA In The Media
Policy

Posted on: Nov 2026

Reading time: 23 min

Policy

Posted on: 02 Nov 2026

Reading time: 23 min

Share:

Climate change is already driving preterm births in Asia – and the toll is set to rise

Climate change is already driving preterm births in Asia – and the toll is set to rise
Visualss, Unsplash
Visualss, Unsplash
Asia & Pacific Briefings Extreme weather

With thanks to Dr Rashmi Agarwalla – Nodal Officer: Climate Change, Associate Professor, Dept of Community and Family Medicine, AIIMS Guwahati – for her insightful review of this briefing.

Key points

  • Human-caused climate change is causing more frequent and prolonged extreme heat events, which disproportionately affect vulnerable populations, such as pregnant women and newborns. South and Southeast Asia suffered such an event in April 2024, experiencing exceptionally high temperatures.  
  • Exposure to extreme heat increases the risk of preterm birth, which is the leading cause of neonatal mortality globally and is associated with long-term physical, neurodevelopmental and socioeconomic effects for babies.
  • This is of particular concern in countries disproportionately affected by extreme heat and with some of the highest preterm birth rates, such as India, Bangladesh and the Philippines.
  • New ZCA analysis estimates that the April 2024 heat event – made at least 45 times more likely by human-caused climate change – was associated with a 3.6% increase in preterm births in India (over those expected in a typical April), an 11.9% increase in Bangladesh, and a 10% increase in the Philippines. This equates to over 16,000 excess preterm births.
  • In a separate estimate, ZCA finds that around 10,600 preterm births across the three countries are attributable specifically to the additional heat that human-caused climate change contributed to the April 2024 event.  
  • Under future warming scenarios (SSP2-4.5 and SSP3-7.0), India alone is projected to experience over 15,000 excess preterm births in a typical April by mid-century. This suggests that an ordinary April in 2050 could carry an excess preterm birth burden higher than that of the exceptionally hot April of 2024.
  • Maternal health programmes in all three countries have not been designed with extreme heat in mind. India and Bangladesh recently issued pregnancy-specific heat guidance, but heat protection is not yet a routine aspect of maternal care. 
  • The challenge is in embedding heat protection into antenatal care, community maternal health services and health-system preparedness. ZCA’s sub-national estimates can inform this by highlighting vulnerable communities where greater health services capacity may be needed.  
  • Better antenatal care, cooling and early warning could prevent some of the preterm birth burden imposed by climate change, but not all of it.

Pregnant women are particularly vulnerable to extreme heat fuelled by human-caused climate breakdown

Human-caused climate change is increasing the frequency and intensity of extreme weather events worldwide, with extreme heat among the most deadly. For example, heatwaves in 2023 are estimated to have caused more than 178,000 deaths globally, with almost half occurring in Asia. Around 54% of the total death toll can be attributed to human-caused climate change. 

While ‘heat events’ – periods of abnormally hot weather lasting days to weeks – affect entire populations, the health and socioeconomic burdens of climate impacts fall disproportionately on women. Pregnant women and newborns are particularly at risk, for both physiological and social reasons. 

Women and newborns living in poverty are especially vulnerable. Poverty removes the buffers that would otherwise blunt climate shocks by reducing access to healthcare and antenatal care, limiting resources that support traditional livelihoods, lowering mobility to adapt or relocate, and causing higher baseline rates of undernutrition. 

Heat exposure often coincides with other vulnerabilities that may reduce a woman’s ability to cope with the physiological stress of pregnancy, including nutritional anaemia, strenuous physical work, and gaps in antenatal care. Heat is not only an outdoor risk: many women spend long hours in poorly ventilated homes, often under roofs made of asbestos, tin or other materials that trap heat, without reliable, affordable electricity to run cooling devices in hot weather.

In remote areas, access to healthcare can also be disrupted during floods or other extreme weather, which especially affects vulnerable groups, including pregnant women and children.

High temperatures increase the risk of preterm births

Several biological mechanisms may explain the link between heat exposure during pregnancy and adverse outcomes. These processes include dehydration, reduced blood flow to the placenta, inflammation, and the activation of pathways that initiate labour.1Physiologically, women generally also have higher working metabolic rates, lower heat dissipation from sweating, and lower effective radiative cooling than men.

However, the relative contributions of these mechanisms to adverse outcomes are not yet fully understood, and effects may vary depending on the mother’s stage of pregnancy, length of exposure, and socioeconomic circumstances.  

Exposure to high temperatures – even for just a few days – can increase the odds of a preterm birth (defined by the World Health Organization as the birth of a live infant before 37 full weeks of pregnancy). A synthesis of global studies finds that a 1°C increase in heat exposure is associated with 4% higher odds of preterm birth. This climbs to 26% during heatwaves. 

A ten-year analysis published by The Lancet shows preterm birth is already the leading cause of neonatal mortality globally, and is ‘associated with long-term physical, neurodevelopmental, and socioeconomic effects.’

Extreme heat is linked to a range of adverse birth outcomes

Beyond preterm birth, heat events, particularly in low and middle-income settings, have been linked to the heightened risk of a variety of adverse birth outcomes. These include congenital defects in babies and pre-eclampsia in mothers, which causes high blood pressure and organ damage.

Based on a global synthesis of studies, high heat is associated with 48% higher odds of congenital anomalies and 28% higher odds of gestational diabetes, while the odds of any obstetric complication grow by 5% per 1°C increase and 25% during a heatwave. However, the evidence for these outcomes is less extensive than for preterm birth, with studies varying considerably in their design, how they define each outcome, and how they measure heat.

In South Asia specifically, attribution studies have made direct links between heat events triggered by climate change and both low birthweight and neonatal deaths:

  • Heat and low birthweight: Human-caused climate change has contributed to approximately 59% of temperature-related low birthweight cases in South Asia. In India, 67% of cases, resulting in 3,468 cases between 1990 and 2018, were attributable to climate change. 
  • Heat and neonatal mortality: In the Philippines, almost four in five (79%) of heat-related neonatal deaths were attributed to climate change, with a rate of 16 per 100,000 births and total deaths of 7,043 between 2001 and 2019. Over the same period, climate change accounted for 34% of Bangladesh’s heat-related neonatal deaths (8,064) and 20% of India’s (47,046).

Factors exacerbating how heat affects maternal and prenatal health and neonatal vulnerability to heat

Socio-cultural factors can render pregnant women even more vulnerable to the effects of extreme heat, especially in developing countries. These include culturally-prescribed heavy clothing that traps heat, as well as restricted access to cooling centres and healthcare due to personal safety concerns and limited transportation.

The same is true of economic factors, such as the need to keep working in challenging conditions. In Tamil Nadu, India, pregnant women engaging in moderate to heavy physical work during heat stress episodes had roughly double the risk of miscarriage and face a higher risk of adverse pregnancy or birth outcomes. 

Despite the evidence that high temperatures during pregnancy are linked to adverse outcomes for mothers and babies, targeted strategies remain scarce worldwide. Yet this is an issue that will become of increasing concern as heat events due to human-caused climate change become more frequent.

Human-caused climate change is accelerating pregnancy heat risk in South and Southeast Asia

This briefing focuses on India, Bangladesh and the Philippines: three climate-vulnerable countries2The Economic and Social Commission for Asia and the Pacific (ESCAP) estimates that, by the end of the century, India and Bangladesh could experience over 300 days each year above 35°C and, in many areas, more than 200 days above 41°C. The Lancet Countdown on Health and Climate Change reports a significant increase in heat exposure in the Philippines, with six times as many heatwave days in 2020-2024 as in a scenario without climate change. which ranked 15th, 13th and 7th respectively in Germanwatch’s Climate Risk Index 2026. The Index highlights that ‘Human-induced climate change affects the frequency and intensity of extreme weather events’. 

These are countries where pregnancy-specific heat exposure is rising faster than almost anywhere in the world, and which were directly affected by a series of extreme, prolonged heat events in April 2023, 2024 and 2026.

In May 2025, Climate Central analysis revealed that Southeast Asia is experiencing some of the world’s most dramatic climate-driven increases in ‘pregnancy heat-risk days’ – days with maximum temperatures above a threshold linked to increased preterm birth risk. 

It finds that, between 2020 and 2024, climate change was responsible for 30 of the 36 pregnancy heat-risk days the Philippines experienced per year over 2020-2024 (a striking 83% of the total), compared with six of the 19 days in India (32%) and seven of the 34 days (21%) in Bangladesh.

Put another way, the impact of climate change means the Philippines experiences, on average, an additional full month of pregnancy heat-risk days per year.

High preterm birth rates in climate-vulnerable countries are exacerbated by rising temperatures\

South Asia has the highest rates of preterm birth of any region, at 13.2% in 2020, and Bangladesh the highest average rate of any country, at 16.2% in the same year, with limited improvement over the last decade.

India’s preterm birth rate is also one of the highest, at 13.0%, while the rate is lower in the Philippines, at 7.7%, even as its heat exposure has risen the fastest of the three.  

India accounted for 20% of all global preterm births in 2020, according to The Lancet. While global preterm birth rates fell by 6.3% between 1990 and 2021, there has been a concerning rise in preterm births in India, with wide state-level variation in rates. 

The link between exposure to extreme heat and preterm birth in India (as well as other maternal and neonatal conditions) is well established, but as yet, the risk of preterm birth from exposure to extreme heat in Bangladesh and the Philippines has not been quantified, despite growing concern.

Consecutive and prolonged pre-monsoon heat events are compounding heat stress in South and Southeast Asia

The fingerprint of climate change on South and Southeast Asian pre-monsoon heat – typically experienced from mid-March to May – has become more pronounced in recent years, exposing hundreds of millions of people to deadlier temperatures for longer periods. 

Studies by World Weather Attribution (WWA) found that human-caused climate change increased the likelihood of:

  • The April 2023 India-Bangladesh heat event by at least 30 times 
  • The South Asia April 2024 heat event by around 45 times – and noted that the temperatures experienced in the Philippines would be  ‘impossible’ without human-caused warming, 
  • The April 2026 northwestern India and Pakistan heatwave by roughly 3 times.

Greenhouse gas emissions from human activity made all of these events hotter by 0.85-2°C. 

Globally, April 2024 is the warmest April on record, and the heat event in South and Southeast Asia was prolonged and extreme, lasting a month in India and Bangladesh and 15 days in the Philippines, with temperatures pushing well above historical norms.3Temperatures reached 0.9°C above the 1991-2020 average in India, 2.9°C in Bangladesh and 2.4°C in the Philippines; our own calculations.

The India Meteorological Department issued a code red for parts of the country; Kolkata, the largest city in the state of West Bengal, experienced its longest heatwave on record, with temperatures 5-10°C above the seasonal average.

New analysis: How the April 2024 heat event affected preterm births in India, Bangladesh and the Philippines, and the contribution of human-caused climate change

We set out to estimate the number of preterm births associated with the 2024 heat event in India, Bangladesh and the Philippines. Throughout, we use ‘excess preterm births’ to mean the preterm births that would not have occurred if temperatures had been at their baseline level rather than the level being assessed.4We used ERA5-Land reanalysis data at 0.1° resolution to calculate temperature anomalies for the event window (full April for South Asia; April 15-29 for the Philippines). Each grid cell’s anomaly was calculated as its average deviation from a 1991-2020 baseline for the same calendar window, and state/division/region-level values were obtained as the population-weighted mean across the grid cells. Population weighting ensures the figure reflects the heat experienced by the average person, rather than a simple geographic land average. We used state-specific preterm birth rates for India, and national rates for Bangladesh and the Philippines as we were unable to obtain regional data for these countries. We then applied a pooled epidemiological odds ratio from Lakhoo et al. 2025 (OR 1.04 per 1°C increase) to the expected births to calculate excess preterm cases. This follows previous work estimating the heat-related health burden from published exposure-response relationships and baseline birth and preterm birth rates (Zhang et al., 2022). The 1.04 estimate from Lakhoo et al. 2025 pools odds ratios, risk ratios and hazard ratios, which the authors treat as comparable following the Cochrane Handbook. Final estimates and confidence intervals were derived using a 10,000-iteration Monte Carlo simulation. The range of the published risk estimate is 1.03-1.06 per 1°C, and so we ran the calculation 10,000 times using different values from across that range. We report the median (central) value, and the confidence interval (CI) shows where 95% of the results fell. The intervals reflect uncertainty in the risk estimate (and, for the climate-attributable figure, in WWA’s temperature estimate), which is the largest source of uncertainty. They do not include uncertainty in birth numbers or preterm birth rates, which would only shift our estimates slightly. We applied the same pooled risk estimate to all three countries. This estimate comes from the most comprehensive recent synthesis, covering 198 studies across 66 countries, and using one estimate keeps results comparable across countries. While differences in housing, humidity, healthcare access and physiological adaptation mean the true effect may vary by country and region, the evidence base suggests it is more likely to be higher than lower in these settings. Some of the preterm births in our analysis may not be additional overall: heat can bring forward a birth that would have been preterm anyway; just days or weeks later. Our method counts births during the heat window and cannot separate these from births that would otherwise have reached full term.

We asked: 

  1. How unusual was the heat in April 2024, and what did that mean for preterm births?
    We compare April 2024 in each state, division or region with its own average April over 1991-2020 – this is the current climatological baseline defined by the World Meteorological Organisation, and the standard reference period for describing how far conditions have departed from normal. As we calculate the temperature departure separately for each region, this allows us to assess at the subnational level where the heat in 2024 was most unusual and where the preterm birth burden fell.

The figure we report for this question does not, however, tell us how much of that unusual heat was caused specifically by human-caused climate change, as the 1991-2020 baseline already contains a substantial amount of the warming to date. So to isolate the effect of human-caused warming on preterm births, we asked:

  1. How many of April 2024’s preterm births would not have occurred without human-caused climate change?
    We use WWA’s finding that climate change made the event around 0.85°C hotter in South Asia and 1.2°C hotter in the Philippines, and ask how many preterm births would not have occurred had the heat been that much less intense. The comparison here is to a world without human-caused warming, rather than with a typical April. WWA gives a single figure for each region, rather than for individual states, so this estimate is only reported nationally. 

We report the values along with the 95% confidence interval (CI), which is the range our estimate is likely to fall within, given uncertainty in how strongly heat affects preterm birth. 

Our findings 

  1. Using three decades of temperature data,5The physiological mechanisms of heat stress differ between dry and humid environments. In dry environments, severe thermal strain occurs directly from extreme ambient heat. Conversely, in highly humid regions, atmospheric moisture physically prevents sweat evaporation – the human body’s primary cooling mechanism. This inability to evaporate sweat causes internal core temperatures to rise rapidly, inflicting severe physiological stress even when the thermometer reading is lower. In this analysis, we report temperature rather than a heat index (which accounts for humidity) to match the event definitions used by World Weather Attribution. Our results may therefore underestimate the burden of heat risks from humid heat. we estimate that the April 2024 heat event was associated with roughly 16,150 excess preterm births (95% CI 10,100-22,400) across the three countries when compared with a typical April over 1991-2020 (Table 1).6We follow WWA’s event definitions for each region, so that our estimates describe the same events their attribution findings refer to. For South Asia, WWA defined the event as the April mean temperature, and we therefore use average daily temperature across the full month for India and Bangladesh. For the Philippines, they defined it as the 15-day average of the maximum daily temperature, and so we use the same measure for 15-29 April. This means that the temperature figures for the Philippines are not directly comparable with those from Bangladesh and India. The risk estimate we use – OR 1.04 per 1°C increase – is drawn from a range of studies using a mix of exposure measures, and therefore the use of either daily maximum or daily average is justified.
Table 1
  1. Separately, we estimate that around 10,600 preterm births (95% CI 5,100-17,700) across the three countries would not have occurred in April 2024 without human-caused climate change (Table 2). This compares April 2024 as it happened with the same event without the warming climate change added – about 0.85°C in South Asia and 1.2°C in the Philippines – rather than with a typical April.7To estimate how much of this burden is attributable to climate change, we used WWA’s finding that the April 2024 heat was around 0.85°C hotter in South Asia and 1.2°C hotter in the Philippines than it would have been without human-caused warming. We then asked how many of the preterm births we estimate for the event would not have occurred had the heat been that much less intense (i.e. subtracting the temperature increase attributed to climate change), applying the same odds ratio (OR) of 1.04 per 1°C increase. This approach follows Stuart-Smith et al. (2025), who construct a counterfactual by subtracting the temperature change attributable to human influence from observed temperatures; an approach they recommend for single events. We use a single published estimate of the temperature increase for each region, as WWA give only a single figure for the whole of South Asia and a single figure for the Philippines. The true increase would have been larger in some places and smaller in others, which is why we report this figure nationally only. We also acknowledge that the underlying evidence base for the 1.04 OR is heavily drawn from high-income and temperate regions and therefore might be conservative for lower-income settings.
Table 2

This is not a share of the 16,150, as the two figures use different comparisons: the first measures the April 2024 event against the climate of the past three decades – which already includes a lot of the warming to date – while the second figure measures the event against a world without human-caused climate change.8For example, states that were cooler than their 1991-2020 April would contribute nothing to the first figure, but as climate change still made them hotter than they would otherwise have been, they would contribute to the second.

The excess preterm birth burden of April 2024 compared to a typical April: subnational insights

Our estimates suggest that the April 2024 heat event was associated with more preterm births than would be expected in a typical April in all three countries: by 3.6% in India, by 11.9% in Bangladesh, and by 10.0% in the Philippines.9For India, this counts only the states that were warmer than normal. If the slightly cooler April in the nine states that were below their usual temperature – mainly in the north and centre of the country – is netted out, the figure is 3.2%. 10As an additional check, we also estimated the burden by counting only heatwave days – those (two or more days) above the 90th percentile of the 1991-2020 April distribution – and applying the heatwave-specific risk estimate of OR=1.26. This gives 9,105 excess preterm births for India and 5,011 for Bangladesh, which are close to our main estimates. For the Philippines, the value is 948, which is higher than our estimate of 585.

To date, estimates of the preterm births resulting from extreme heat have been made at the global or national level. which conceals the fact that the burden is not evenly spread. A better understanding of the heat-attributable burden of preterm births at the sub-national level in India, Bangladesh and the Philippines could help inform targeted interventions and highlight vulnerable populations. 

These subnational estimates, shown in Figures 1-3 and Table 3 below, are reported against a typical April, not as a share attributable to human-caused climate change.

We use three measures:

  1. Absolute excess: where the most extra preterm births occurred. This is driven by population size as well as heat, so it shows where health services face the greatest load
  2. Rate per 1,000 births: the extra preterm births relative to the number of births, so large and small regions can be compared. This is driven by heat and how common preterm birth already is. 
  3. Risk increase %: how much the heat raised the chance of any given birth being preterm. This is driven only by how unusual the heat was. 

The excess preterm birth burden of April 2024 is concentrated and uneven

The subnational estimates show that the burden was concentrated and uneven, which national figures alone would miss (Figures 1-3). Some places carry the largest numbers of excess preterm births and need the most health-service capacity, while others saw the sharpest rises in risk despite their small populations, and could be overlooked if only national or total figures are used. 

These findings matter for policy: as a small number of states and divisions account for most of the excess preterm births, heat-health measures such as early warnings, antenatal outreach and cooling in health facilities could be targeted where they would protect the most pregnant women – while ensuring that places where risk is rising sharply are not missed.

India

For India, we estimate 10,050 excess preterm births occurred during April 2024, with the Indo-Gangetic plain states of Bihar, Uttar Pradesh, West Bengal and Jharkhand carrying 65% of the national excess (Figure 1). 

Nationally, the average person in India experienced an April that was 0.94°C hotter than normal, though this conceals wide variation: nine states were cooler than usual, while Tripura and Mizoram were around 2.7°C warmer, implying rises in ‘preterm birth risk’ of around 11% in these two states (Table 3).

The preterm birth risk % is the change in the chance that any given birth is preterm: where risk rose by 11%, a woman who would have faced a 13% chance of delivering early faced roughly 14.4%11i.e. 11% of 13% is 1.43%; 13%+1.43%=14.43%. instead.12Because the underlying estimate pools odds ratios with risk ratios, we treat it as a change in risk. Where the underlying studies report odds ratios, the true change in risk will be slightly smaller, particularly in India and Bangladesh where preterm birth is more common. The preterm birth risk increase % depends only on how much hotter the region was, not on how many babies were born there or how common preterm birth already is.

Because preterm birth is more common in some states than others, the rate per 1,000 births does not always follow the heat. Karnataka had the highest rate, at 12 per 1,000 births, reflecting both a 2°C anomaly and a higher-than-average baseline preterm birth rate. Mizoram, despite one of the largest rises in preterm birth risk, had one of the lowest rates per 1,000 births, because preterm birth there is less common. 

Among the states with the largest absolute excess preterm birth burden, West Bengal saw a rise in preterm birth risk of 10% and Bihar 7%, while in Uttar Pradesh, where April was only 0.74°C warmer than normal, risk rose by around 3% – its large burden reflects its population rather than the intensity of the heat. By contrast, Bihar and West Bengal combined large populations with some of the larger rises in preterm birth risk, so their burden reflects the intensity of the heat as well as their size.

Bangladesh

For Bangladesh, we estimate 5,515 excess preterm births during the April 2024 heat event, around 19 per 1,000 births – the highest rate of the three countries. The two most populous divisions – Dhaka and Chittagong – carried 47% of the national burden (Figure 2).

Bangladesh experienced the most unusual heat of the three countries relative to its own climate, with temperatures averaging around 2.88°C higher than its 1991-2020 April baseline.

Preterm birth risk rose by roughly 10% to 15% in every division, with the largest increase in Mymensingh, which was 3.59°C warmer in 2024 than a typical April. Because we use a single national preterm birth rate for Bangladesh, the rate per 1,000 births follows the heat directly: Mymensingh also had the highest rate of excess preterm births of any division, at 24 per 1,000, despite ranking sixth of eight on total excess preterm numbers, while Sylhet, the least affected, had the lowest at 16 per 1,000.

The Philippines

For the Philippines, we estimate 585 excess preterm births over the 15-day heat event of April 2024, around 8 per 1,000 births. Calabarzon had the most (around 100), followed by Metropolitan Manila and Central Luzon (around 80 each) – Metropolitan Manila accounted for 14% of the national total. The Philippine total is lower than India’s or Bangladesh’s because the event window is half as long, the preterm birth rate is lower, and the population is smaller.  

The average person experienced temperatures 2.44°C above normal. As in Bangladesh, a single national preterm birth rate means the rate per 1,000 births broadly follows the heat: Cagayan Valley in northern Luzon had both the largest rise in preterm birth risk, around 15%, and the highest rate, at 12 per 1,000. Within it, Cagayan province was 4.01°C hotter than normal, with a rise in risk of 17%. Caraga, in Mindanao, had the lowest rate at 3 per 1,000.

Figure 1
Figure 2
Figure 3
Table 3

New projection: By mid-century, climate change could drive the preterm birth burden of an ordinary April in India to exceed that of the exceptionally hot April of 2024

To understand how ongoing planetary warming caused by climate change might affect the heat-driven preterm birth burden, we separately projected the relationship forward using the INDRA-CMIP6 dataset13A high-resolution, bias-corrected multi-model ensemble covering the Indian subcontinent. Four models (CanESM5, ACCESS-CM2, NorESM2 and MIROC6) were chosen because they span the Equilibrium Climate Sensitivity range of CMIP6. We applied the same dose-response relationship used throughout this analysis, adjusted for projected population change, and generated uncertainty ranges (confidence intervals) using Monte Carlo simulation. for India and Bangladesh.14Given India and Bangladesh’s substantial contribution to the preterm burden for the April 2024 heat event, together accounting for about 96% of the three-country total, as well as their comparatively high preterm birth rates, we focused on these two countries for our future projections.

We compared projected temperatures for mid-century (2041-2060) under two climate emissions pathways: SSP2-4.5, a moderate emissions scenario, and SPP3-7.0, a higher emissions scenario with weaker near-term emissions controls. Our estimates assume no measures are taken towards climate adaptation.  

By mid-century (2041-2060), combining both emissions scenarios, April temperatures in India are projected to be around 1.6°C above the 1985-2014 baseline, implying roughly 15,300 excess preterm births (95% range 6,800-28,100) in a typical April. For Bangladesh, the projected warming is around 1.1°C, implying roughly 1,500 excess preterm births (630-2,900).15These projections assume the temperature–preterm birth relationship observed today holds unchanged in a warmer climate, with no physiological, behavioural or health-system adaptation, though births are scaled to projected levels using UN World Population Prospects 2024. As fertility is falling in both countries, projected annual births by mid-century are around 18% lower in India and 22% lower in Bangladesh compared to 2024.

This means that, by mid-century, the impact of climate change in India could see an ordinary April carry an excess preterm birth burden exceeding that of the exceptional heat event of April 2024.16Projections use daily maximum temperature while the April 2024 analysis uses daily average. In India, the two metrics produced anomalies within 7% of each other for the April 2024 event, so the mid-century projection is broadly comparable to the 2024 estimate in magnitude. The projections also measure warming from a slightly earlier baseline (1985–2014, rather than 1991-2020), which makes projected warming appear a little larger. Neither difference changes the conclusion: India’s projected burden is around 50% higher than in April 2024, so the mid-century projection is broadly comparable to the 2024 estimate in magnitude. This is an average burden – a particularly hot April would exceed it.

The mid-century excess preterm burden: subnational insights

In India, the state of Uttar Pradesh is projected to have the largest excess burden in absolute terms, followed by Bihar and Rajasthan, while in Bangladesh, Dhaka Division is projected to have the highest burden, followed by Chittagong and Rajshahi (Table 4).

Table 4

While these projections focus on the future toll of rising temperatures in South Asia, it is worth remembering that many other climate impacts present a concurrent – and growing – threat to maternal and neonatal health in the region (Box 1).

Box 1. Beyond heat: other climate impacts putting mothers and babies at risk

In addition to heat, various climate impacts are affecting the health of mothers and babies. The empirical evidence below quantifies some of these risks.

In Bangladesh:

  • The infant death rate is 8% higher in flood-prone areas, resulting in an estimated excess of 152,753 infant deaths between 1988-2017.
  • Saline contamination of drinking water is increasing with climate change and sea-level rise. Consumption of severely saline water resulted in a 3% higher probability of infant mortality, resulting in an estimated 3,927.5 infant deaths across six coastal Bangladeshi subdivisions in 2011 and 2014, based on national survey data. 
  • Exposure to tropical cyclones, either in utero or in the first year of life, led to 10.4 additional infant deaths per 1,000 live births – a 27% increase from the baseline average infant mortality rate. 

In India:

  • Flood-affected regions have a higher prevalence of low and very low birthweight infants (8% and 27% higher, respectively), and among the poorest households, flood exposure raised the likelihood of low birth weight by 5% compared to those not exposed to floods. 
  • For agricultural households, climate-induced low temperatures and low rainfall before conception increase infant mortality rates by 1.09 and 1.53 percentage points, respectively. 
  • Exposure to excessive monsoon rainfall in utero and during the first year after birth is linked to undernutrition in children under five years of age, resulting in a higher likelihood of stunted growth. 

At the same time, air pollution – including from the same burning of fossil fuels that is driving climate change – was responsible for 4.4-4.9% of Bangladesh’s preterm births between 2015 and 2019, while 17,800 of the country’s newborn deaths could be attributed to air pollution in 2023.

Indian mothers exposed to higher levels of ambient PM2.5 air pollution during pregnancy had 67% higher odds of delivering preterm and a 37% higher odds of having a low-birthweight baby, even after accounting for temperature.

Maternal health policies and heat action plans do not yet systematically protect pregnant women and babies in India, Bangladesh or the Philippines against heat risks

India has launched various large-scale public health programmes aimed at predicting, preventing and monitoring preterm births. These include the India Newborn Action Plan (INAP), which aims to reduce neonatal mortality and stillbirth rates to fewer than 10 per 1,000 live births by 2030, and the Pradhan Mantri Surakshit Matritva Abhiyan (PMSMA), which has the goal of improving maternal and fetal health through early detection and management of maternal complications. 

Similarly, the Bangladesh Every Newborn Action Plan (BENAP) targets reducing neonatal mortality to no more than 12 per 1,000 live births, and stillbirths to no more than 12 per 1,000 total births, by 2030. The Philippines has implemented the Maternal, Newborn, Child Health and Nutrition (MNCHN) strategy, which aims to reduce maternal and neonatal deaths through the provision of health services.

However, none of these programmes have been designed with heat specifically in mind: they address newborn and maternal survival more broadly, and heat risk is not yet systematically integrated into routine maternal care in any of the three countries.

Baby steps toward embedding heat-risk in maternal care

In India, heat risk is just beginning to be integrated into maternal care pathways. The Ministry of Health and Family Welfare has published a guideline on heat and pregnancy for healthcare professionals and community health workers. It covers heat-risk assessment, high-risk pregnancies, warning signs, cooling, hydration, referral and family preparedness.

The country’s National Programme on Climate Change and Human Health (NPCCHH) and State Action Plans on Climate Change and Human Health also provide a route for building maternal heat protection into state and district health systems. 

India also has more Heat Action Plans (HAPs) – strategies for adapting to extreme heat, including measures, technologies and interventions – than any other low or middle-income country, and many recognise pregnant women as a vulnerable group:  (60%) of Indian state HAPs do so, compared to 46% elsewhere in the world. 

But recognition has not consistently translated into protective measures. A review of plans from every state and union territory – 19 HAPs and, where none existed, 16 State Action Plans on Climate Change – found that one third (12) of the 35 contained no recommendations for pregnant or lactating women.

The remaining 23 listed 214 strategies that referred to this group, either directly or as part of a wider vulnerable population, but only 40 were aimed at lactating or pregnant women specifically, and the largest share focused on behaviour change rather than structural measures such as cooling or health-system support.

The fact that a plan contains pregnancy-specific recommendations is therefore not a guarantee that it will be effectively implemented or delivered. The evidence points to a gap in policy design: India’s HAPs have also been criticised for being poorly designed, including a lack of local context and poor targeting of vulnerable groups.17This appears to be a limitation of HAPs more broadly, with a 2025 review finding that none comprehensively address maternal, newborn, and child health risks during extreme heat.

Efforts are being made to close the evidence gap on the ineffective implementation of HAPs in India. The Heat in Pregnancy-India (HiP-India) project is running a study across three climate zones in India to understand the mechanisms linking heat to birth outcomes, and includes a review of existing HAPs to identify where they can be strengthened. 

Meanwhile, healthcare workers say they are already seeing the effects of heat on pregnant women, and feel underprepared to respond. A September 2026 survey by the Wellcome Trust of 1,000 maternal healthcare professionals across five countries, including India, found that 73% had seen an increase in heat-related cases or complications among pregnant women over the past five years, and 92% wanted more training and resources to protect them.

Bangladesh has also recently begun to take a more targeted approach: in 2024, the Directorate General of Health Services, working with UNICEF, launched national guidelines specifically to protect children and pregnant women from heat stress, citing preterm birth risk as a rationale. 

We found no heat-specific maternal health guideline or HAP for the Philippines.

The challenge, therefore, is no longer in recognising that pregnant women are vulnerable, but in making heat protection a routine part of antenatal care, community maternal health services and health-system preparedness. Our estimates show the burden that climate change added, not how much of it could be avoided: better antenatal care, cooling and early warning could prevent some of it, but not all.

  • 1
    Physiologically, women generally also have higher working metabolic rates, lower heat dissipation from sweating, and lower effective radiative cooling than men. ↩︎
  • 2
    The Economic and Social Commission for Asia and the Pacific (ESCAP) estimates that, by the end of the century, India and Bangladesh could experience over 300 days each year above 35°C and, in many areas, more than 200 days above 41°C. The Lancet Countdown on Health and Climate Change reports a significant increase in heat exposure in the Philippines, with six times as many heatwave days in 2020-2024 as in a scenario without climate change. ↩︎
  • 3
    Temperatures reached 0.9°C above the 1991-2020 average in India, 2.9°C in Bangladesh and 2.4°C in the Philippines; our own calculations. ↩︎
  • 4
    We used ERA5-Land reanalysis data at 0.1° resolution to calculate temperature anomalies for the event window (full April for South Asia; April 15-29 for the Philippines). Each grid cell’s anomaly was calculated as its average deviation from a 1991-2020 baseline for the same calendar window, and state/division/region-level values were obtained as the population-weighted mean across the grid cells. Population weighting ensures the figure reflects the heat experienced by the average person, rather than a simple geographic land average. We used state-specific preterm birth rates for India, and national rates for Bangladesh and the Philippines as we were unable to obtain regional data for these countries. We then applied a pooled epidemiological odds ratio from Lakhoo et al. 2025 (OR 1.04 per 1°C increase) to the expected births to calculate excess preterm cases. This follows previous work estimating the heat-related health burden from published exposure-response relationships and baseline birth and preterm birth rates (Zhang et al., 2022). The 1.04 estimate from Lakhoo et al. 2025 pools odds ratios, risk ratios and hazard ratios, which the authors treat as comparable following the Cochrane Handbook. Final estimates and confidence intervals were derived using a 10,000-iteration Monte Carlo simulation. The range of the published risk estimate is 1.03-1.06 per 1°C, and so we ran the calculation 10,000 times using different values from across that range. We report the median (central) value, and the confidence interval (CI) shows where 95% of the results fell. The intervals reflect uncertainty in the risk estimate (and, for the climate-attributable figure, in WWA’s temperature estimate), which is the largest source of uncertainty. They do not include uncertainty in birth numbers or preterm birth rates, which would only shift our estimates slightly. We applied the same pooled risk estimate to all three countries. This estimate comes from the most comprehensive recent synthesis, covering 198 studies across 66 countries, and using one estimate keeps results comparable across countries. While differences in housing, humidity, healthcare access and physiological adaptation mean the true effect may vary by country and region, the evidence base suggests it is more likely to be higher than lower in these settings. Some of the preterm births in our analysis may not be additional overall: heat can bring forward a birth that would have been preterm anyway; just days or weeks later. Our method counts births during the heat window and cannot separate these from births that would otherwise have reached full term. ↩︎
  • 5
    The physiological mechanisms of heat stress differ between dry and humid environments. In dry environments, severe thermal strain occurs directly from extreme ambient heat. Conversely, in highly humid regions, atmospheric moisture physically prevents sweat evaporation – the human body’s primary cooling mechanism. This inability to evaporate sweat causes internal core temperatures to rise rapidly, inflicting severe physiological stress even when the thermometer reading is lower. In this analysis, we report temperature rather than a heat index (which accounts for humidity) to match the event definitions used by World Weather Attribution. Our results may therefore underestimate the burden of heat risks from humid heat. ↩︎
  • 6
    We follow WWA’s event definitions for each region, so that our estimates describe the same events their attribution findings refer to. For South Asia, WWA defined the event as the April mean temperature, and we therefore use average daily temperature across the full month for India and Bangladesh. For the Philippines, they defined it as the 15-day average of the maximum daily temperature, and so we use the same measure for 15-29 April. This means that the temperature figures for the Philippines are not directly comparable with those from Bangladesh and India. The risk estimate we use – OR 1.04 per 1°C increase – is drawn from a range of studies using a mix of exposure measures, and therefore the use of either daily maximum or daily average is justified. ↩︎
  • 7
    To estimate how much of this burden is attributable to climate change, we used WWA’s finding that the April 2024 heat was around 0.85°C hotter in South Asia and 1.2°C hotter in the Philippines than it would have been without human-caused warming. We then asked how many of the preterm births we estimate for the event would not have occurred had the heat been that much less intense (i.e. subtracting the temperature increase attributed to climate change), applying the same odds ratio (OR) of 1.04 per 1°C increase. This approach follows Stuart-Smith et al. (2025), who construct a counterfactual by subtracting the temperature change attributable to human influence from observed temperatures; an approach they recommend for single events. We use a single published estimate of the temperature increase for each region, as WWA give only a single figure for the whole of South Asia and a single figure for the Philippines. The true increase would have been larger in some places and smaller in others, which is why we report this figure nationally only. We also acknowledge that the underlying evidence base for the 1.04 OR is heavily drawn from high-income and temperate regions and therefore might be conservative for lower-income settings. ↩︎
  • 8
    For example, states that were cooler than their 1991-2020 April would contribute nothing to the first figure, but as climate change still made them hotter than they would otherwise have been, they would contribute to the second. ↩︎
  • 9
    For India, this counts only the states that were warmer than normal. If the slightly cooler April in the nine states that were below their usual temperature – mainly in the north and centre of the country – is netted out, the figure is 3.2%. ↩︎
  • 10
    As an additional check, we also estimated the burden by counting only heatwave days – those (two or more days) above the 90th percentile of the 1991-2020 April distribution – and applying the heatwave-specific risk estimate of OR=1.26. This gives 9,105 excess preterm births for India and 5,011 for Bangladesh, which are close to our main estimates. For the Philippines, the value is 948, which is higher than our estimate of 585. ↩︎
  • 11
    i.e. 11% of 13% is 1.43%; 13%+1.43%=14.43%. ↩︎
  • 12
    Because the underlying estimate pools odds ratios with risk ratios, we treat it as a change in risk. Where the underlying studies report odds ratios, the true change in risk will be slightly smaller, particularly in India and Bangladesh where preterm birth is more common. The preterm birth risk increase % depends only on how much hotter the region was, not on how many babies were born there or how common preterm birth already is. ↩︎
  • 13
    A high-resolution, bias-corrected multi-model ensemble covering the Indian subcontinent. Four models (CanESM5, ACCESS-CM2, NorESM2 and MIROC6) were chosen because they span the Equilibrium Climate Sensitivity range of CMIP6. We applied the same dose-response relationship used throughout this analysis, adjusted for projected population change, and generated uncertainty ranges (confidence intervals) using Monte Carlo simulation. ↩︎
  • 14
    Given India and Bangladesh’s substantial contribution to the preterm burden for the April 2024 heat event, together accounting for about 96% of the three-country total, as well as their comparatively high preterm birth rates, we focused on these two countries for our future projections. ↩︎
  • 15
    These projections assume the temperature–preterm birth relationship observed today holds unchanged in a warmer climate, with no physiological, behavioural or health-system adaptation, though births are scaled to projected levels using UN World Population Prospects 2024. As fertility is falling in both countries, projected annual births by mid-century are around 18% lower in India and 22% lower in Bangladesh compared to 2024. ↩︎
  • 16
    Projections use daily maximum temperature while the April 2024 analysis uses daily average. In India, the two metrics produced anomalies within 7% of each other for the April 2024 event, so the mid-century projection is broadly comparable to the 2024 estimate in magnitude. The projections also measure warming from a slightly earlier baseline (1985–2014, rather than 1991-2020), which makes projected warming appear a little larger. Neither difference changes the conclusion: India’s projected burden is around 50% higher than in April 2024, so the mid-century projection is broadly comparable to the 2024 estimate in magnitude. ↩︎
  • 17
    This appears to be a limitation of HAPs more broadly, with a 2025 review finding that none comprehensively address maternal, newborn, and child health risks during extreme heat. ↩︎
Joanne Bentley-McKune

Joanne Bentley-McKune

Jo is a researcher in climate science and nature, looking at climate risks to natural ecosystems.

Joanne Bentley-McKune

Joanne Bentley-McKune

Jo is a researcher in climate science and nature, looking at climate risks to natural ecosystems.

Victoria Kalyvas

Victoria Kalyvas

Victoria is the team’s climate science and policy researcher, specialising in carbon removal and international climate policy.

Victoria Kalyvas

Victoria Kalyvas

Victoria is the team’s climate science and policy researcher, specialising in carbon removal and international climate policy.

Keep reading

Briefings Public finance Series

Oct 2024

Reforming climate finance: Unlocking funds from multilateral development banks

Multilateral development banks play a key role in climate finance, but reforms could allow them to lend hundreds of billions of dollars more.

Read More
Briefings Oil and gas Renewables

Apr 2024

Bullish Asian gas demand forecasts eroded by renewable surge

Increased momentum for the energy transition has led the IEA to lower its 2030 gas demand forecast by over 421 bcm.

Read More
Briefings Public finance

Apr 2023

Climate change requires a new approach from international financial institutions

Addressing climate change effectively will not be possible without reforming key institutions such as the IMF and World Bank.

Read More

About

  • About Us
  • Cookie Policy
  • Privacy Policy
  • Terms of Use

Follow Us

Get In Touch:

216
Join Our Newsletters!
Manage Consent
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behavior or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes. The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
  • Manage options
  • Manage services
  • Manage {vendor_count} vendors
  • Read more about these purposes
View preferences
  • {title}
  • {title}
  • {title}
Manage Consent
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behavior or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes. The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
  • Manage options
  • Manage services
  • Manage {vendor_count} vendors
  • Read more about these purposes
View preferences
  • {title}
  • {title}
  • {title}