Europe’s Heat Killed 50,391 in Cities Official Counts Still Miss

Heat killed an estimated 50,391 people in 854 European cities this summer, according to peer-reviewed models released on Oct. 23, 2026. About two-thirds of those deaths, 32,619, were tied to extra heat from burning coal, oil and gas.

National heat ledgers still track a smaller, differently coded total. That gap is now large enough to leave hospitals and heat plans sized for 2003 and 2022 while the temperature itself has moved.

The 50,391 Figure Stops at the City Limits

Clair Barnes of Imperial College London and colleagues at the London School of Hygiene & Tropical Medicine published two papers in Nature Health. The 2026 paper estimates modelled heat deaths in 854 urban areas across the 27 EU states plus Britain, Norway and Switzerland, from June 1 to August 31. The 95% empirical confidence interval runs from 46,026 to 53,878.

Those towns and cities hold about 40% of the 30 countries’ people. Rural deaths sit outside the total, and so does a late-May heatwave that Copernicus recorded as the season’s opening blow. The authors say the 50,391 figure is an underestimate on both counts.

The same team’s method put 29,480 urban heat deaths in 2022 (26,643 to 32,006) and 28,462 in 2003 (25,709 to 31,199). Against 2022, the new total is 71% higher. Against 2003, it is 77% higher. Garyfallos Konstantinoudis, an assistant professor at Imperial’s Grantham Institute, said the 2026 jump held after the team controlled for larger city populations, which points to the heat itself rather than aging cities alone.

The 2026 analysis assigns around two-thirds to fossil-fuel warming. In a climate without that extra heat, the models imply about 17,772 of these urban deaths would still have occurred. Barnes, lead author on both papers, said some of these waves would have killed at a high rate even in a preindustrial climate, and that fossil-fuel warming pushed that burden substantially higher.

Italy Led, and 6,000 Were Younger Than 65

Southern and western Europe had the hottest readings and the highest modeled death rates. Northern cities that used to treat summer as the quiet season were not spared. Pierre Masselot, an assistant professor at LSHTM and a co-author, said parts of western and northern Europe are now seeing extreme heat and a historically high health burden, including among younger adults.

Country estimates from the 2026 study put Italy first. France, Spain, Germany and the United Kingdom follow. Around 6,000 of the 50,391 dead were younger than 65.

HEAT DEATHS IN THE FIVE HARDEST-HIT COUNTRIES

Those five countries account for 38,089 of the 50,391 urban deaths, leaving 12,302 in the other 25 countries in the study. Age-adjusted, the UK rate was 158 deaths per million people, against a study-wide average of 276. The UK still had its heaviest modeled urban summer in the series.

A Third of the Dead Fell in Two Weeks in June

Nearly one-third of the modeled 2026 deaths occurred between June 17 and July 1, when average daily temperatures across the study region sat more than 5C above the 1991-2020 average. On some days, city readings ran as much as 12C above the seasonal norm. Western Europe, in the urban series, averaged about 2.9C above 1991-2020.

Barnes said much of western and northern Europe has buildings and services that were never designed for the temperatures now arriving. That is the practical half of the death count: the heat was worse, and the rooms people slept in, worked in and were nursed in were still built for a cooler climate.

FROM THE 2003 SHOCK TO THIS SUMMER’S PAPERS

1. August 1-20, 2003: France records 14,802 excess deaths in a single heatwave, about 82% of them in people aged 75 or older, and issues a national heat plan the next year.
2. Summer 2022: The same 854-city model later puts urban heat deaths at 29,480, then the series peak.
3. June 11, 2026: WHO/Europe launches the second edition of its heat-health action plan guidance as the first major western heatwave is already underway.
4. June 17-July 1, 2026: About one-third of the eventual modeled urban deaths occur in a fortnight more than 5C above the 1991-2020 average.
5. October 2, 2026: The Copernicus Climate Change Service confirms western Europe’s warmest summer on record and a new mark for strong heat-stress days.
6. October 23, 2026: The 2026 mortality paper is released, putting the urban total at 50,391.

The 2003 French inquiry found hospitals and nursing homes understaffed in August, with no heat watch in place. The 2026 season arrived with warnings, in-season death estimates in Britain, France, Germany and Spain, and still a modeled urban toll 71% above 2022.

Heat Plans Built After 2003 Faced a New Ceiling

France’s 2004 plan added a four-level watch, cooling rooms in senior centers, and daily alerts from June to September. Sister systems spread across Europe. They cut later tolls relative to 2003 even as summers warmed. They were not built for a western summer that Copernicus now puts at 21.7C, 2.5C above the 1991-2020 average, and above the 2003 record.

On June 11, 2026, while western Europe was already in its first major heatwave, Dr Hans Henri P. Kluge, WHO regional director for Europe, used the launch of updated guidance to cite more than 200,000 heat deaths in four years across the EU and associated countries. He said most of those deaths were entirely preventable. The new playbook landed in the same weeks the 854-city model now treats as the deadliest stretch of the century.

Housing, care homes and outdoor work rules still assume nights that cool off and summers that pause. That is why a season of this length produced a death count the post-2003 plans never budgeted, even with alerts on. The people who die are mostly old and already ill, which is also why a death certificate so often names the heart or the lungs and never the heat.

Here in the UK the summer is historically the quietest time for our health service. This summer, however, saw demand and strain that was akin to a really bad winter virus season.

Andrew Hartle, intensive care consultant, Imperial College Healthcare NHS Trust

The UK Health Security Agency estimated almost 2,900 heat-related deaths in England in May and June alone, against 1,504 for the whole of summer 2025. Hartle was not on the study team. His ward comparison is what a heat plan looks like when the quiet season disappears.

How Scientists Turned Temperature Into a Death Toll

The 2026 paper does not count death certificates marked heatstroke. It takes daily temperatures in the 854 cities, then applies city- and age-specific curves that link heat to extra deaths from all causes. Those curves come from observed relationships in 232 of the cities between 2000 and 2019.

To isolate warming, the team compared the temperatures that occurred with a modeled world that had not been heated by fossil fuels, a present climate the authors treat as about 1.4C above preindustrial. The gap between the two death estimates is the climate share. Exposure-response functions, baseline death rates and population were held constant, so the only moving piece is the extra heat.

A companion paper in the same journal, covering the same 854 cities, estimates 430,365 heat-related deaths from 2000 to 2025 (376,535 to 481,458). The climate share of that burden rose from 39% in 2000 to 68% in 2025. Annual modeled urban heat deaths went from 9,720 in 2000 to 24,572 in 2025. Combined population in that paper is almost 160 million.

Marcos Quijal Zamorano, an environmental epidemiologist at the University of Bern who was not an author, has called this style of estimate robust, with the caveat that a city’s 2026 conditions can drift from the 2000-2019 baseline. The authors note that earlier runs of the method have lined up with figures national health agencies later published.

Heart and Lung Failure Carry the Real Toll

Governments usually log heatstroke and similar diagnoses. They often miss deaths from heart and lung disease that heat makes worse. That is the oldest finding in European heat epidemiology, restated after 2003, and it is why a modeled all-cause total and an official heat column refuse to match.

Spain makes the mismatch concrete. The Carlos III Health Institute’s MoMo system, which underpins the national heat plan, estimated 5,807 heat-attributable deaths nationwide from May 15 to September 30, the highest in its series, which starts in 2015. That is 51.5% above MoMo’s 3,832 in 2025 and above 4,789 in 2022. The Imperial-LSHTM model, looking only at 90 Spanish cities in June, July and August, puts the urban toll at 6,824. A nationwide official figure over 4.5 months still sits below a three-month city-only model.

WHO DIES WHEN THE HEAT SETTLES IN

• Older adults: Spain’s MoMo file for 2026 puts 97% of estimated heat deaths in people aged 65 or older and 62% in people over 85, with women 3,486 of the 5,807 (60%) and men 2,321.
• People under 65: The 854-city model still finds around 6,000 deaths under 65, which is the slice heat plans written around care homes never fully cover.
• Heart and lung patients: Heatstroke is the label that makes a heatwave official; most of the extra deaths arrive as heart, lung and other chronic disease made worse by high temperature, including at night.
• Outdoor and poorly housed workers: WHO’s 2026 guidance lists outdoor workers, people in poorly ventilated buildings, pregnant women, infants and people with limited mobility among those at raised risk, alongside older adults.

Masselot said cities remain poorly adapted, and that cooling them means more green space, less pollution and health care aimed at the people the heat actually kills. The 6,000 under 65 are the clearest sign the burden is no longer confined to August in a nursing home.

Record Heat Stress, Low Rivers and Fires Arrived Together

The Copernicus Climate Change Service, in its summer review issued on Oct. 2, 2026, called 2026 Europe’s third-warmest summer on record and western Europe’s warmest, at 21.7C. That western figure is 2.5C above 1991-2020 and above 2003. It is a land-average for a mapped box, not the 2.9C urban-study anomaly, and both can stand.

Heat stress is how Copernicus translates air, wind, humidity and radiation into what a body feels. By the end of the season, 52% of the continent had been exposed to “very strong” heat stress, a feels-like temperature of 38C or more. Averaged across western Europe, strong heat stress landed on a record 41 days of strong heat stress, five more than in 2003. Very strong heat stress reached 13 days, also a record.

WESTERN EUROPE, SUMMER 2026, IN COPERNICUS FIGURES

• Days above 30C: 17 days with at least 60% of western Europe above 30C, against 11 such days in 2003 and 7 in 2025.
• Strong heat stress: 41 days, the series high, five above 2003.
• Very strong heat stress: 13 days, a new mark, with 52% of the continent reaching feels-like 38C or more at some point.
• Rivers and fire: Summer flows were the lowest in the EFAS record back to 1992 on the Rhine, Danube, Po, Loire, Rhône and Seine; by Sept. 2, EU burned area was almost twice the 2006-2025 average, with the largest modern fires in France and Spain displacing 300,000 people.

Samantha Burgess, strategic lead for climate at ECMWF, said the summer stacked record heat stress, dry ground, low rivers and warm seas at once, with effects on water, energy, transport, ecosystems and health. Marine heat in the western and central Mediterranean limited night cooling on the coast and added humidity, which is how a sea record becomes a body-stress record.

Every western European summer since 2015 has been warmer than the 1991-2020 average. The 2026 season also set daily records in May and September, which is why a June-August death model misses part of the year people actually lived through.

Frequently Asked Questions

How Much of Europe’s 2000-2025 Urban Heat Toll Fell on People Aged 85 or Older?

Of the 430,365 modeled heat deaths in the 854 cities from 2000 to 2025, 165,992 (147,573 to 185,609) were in people aged 85 or older, the largest age group in that paper. That long-run concentration is why a heat plan that only opens cooling rooms in care homes still misses outdoor workers and the around 6,000 people under 65 in the 2026 file, but it is also why most certificates never mention heat.

What Is a Heat-Health Action Plan Under WHO’s 2026 Guidance?

WHO/Europe’s second edition, issued on June 11, 2026, frames a plan around eight elements: governance, a heat-health warning system, populations at increased risk, communication, health-system resilience, cutting heat exposure, heat-health surveillance, and monitoring and learning. The first edition dated to 2008. Kluge said the plans save lives when they are funded and staffed, which is a separate question from whether a warning was issued.

Why Does the 50,391 Total Omit May and Rural Areas?

The 2026 paper is limited to June 1 through August 31 in 854 cities that hold about 40% of the 30 countries’ population. Copernicus recorded the first major western heatwave in late May, with further records in September, so early and late deaths are outside the modeled total. Authors have said both the May omission and the city-only frame make 50,391 a floor.

What Does the 1.4C Figure Mean in the 2026 Analysis?

The team compared the temperatures cities actually had with a counterfactual climate without human-caused warming, treating the present world as about 1.4C warmer than preindustrial. Only daily temperature changes between those two climates; death curves, baseline rates and population stay fixed. The 32,619 figure is that temperature gap expressed as extra urban deaths, not a count of certificates that name climate change.

The 50,391 count still leaves out May, September and everyone who died outside those 854 cities. Around 6,000 of the people inside it were younger than 65.

Disclaimer: This article is news reporting and analysis of published mortality models and official heat surveillance. It is informational only and is not medical advice, a diagnosis, or a guide to treating heat illness, heart disease or lung disease. Readers who need care during hot weather, or who are changing medicines or work practices because of heat, should speak with a licensed physician or occupational health clinician. Modeled totals, confidence intervals and official MoMo and agency figures reflect the papers and agencies cited as of Oct. 23, 2026, and later revisions can change them.