Cover: 4.6 million deaths a year. Cold weather deaths, what the evidence shows.
|

Cold Weather Deaths: What the Evidence Shows

Key takeaways · 11 min read

  • In a 13-country study, cold was linked to 7.29% of deaths and heat to 0.42%.
  • Worldwide, about 4.6 million deaths a year are associated with cold and about 489,000 with heat.
  • Most cold-related deaths happen on moderately cold days rather than in extreme freezes.
  • Researchers disagree on how much winter mortality is caused by cold, and how much by infections and season.

Heatwaves get the headlines. They are sudden, visible and increasingly common, and their death tolls are reported in real time. Cold weather draws attention mainly when something dramatic happens, such as the Texas power grid failing in February 2021.

In most of the world, though, cold kills far more people than heat. The Lancet published a study of 74 million deaths in 13 countries in 2015 and found cold linked to about 17 times as many deaths as heat. A global analysis in 2021 put cold-related deaths at about 4.6 million a year.

Those numbers are less simple than they look. Researchers disagree about how much of the winter rise in deaths is caused by cold itself, and about whether climate warming will reduce it. The sections below cover what the evidence shows, where the uncertainty lies, and who is most at risk.

How many deaths are linked to cold?

Antonio Gasparrini and colleagues in the Multi-Country Multi-City Collaborative Research Network produced the most influential study. They analysed 74,225,200 deaths in 384 locations across 13 countries, including Australia, Canada, China, Japan, the UK and the USA, and modelled how the risk of death changed with daily temperature over the following three weeks.

Of all deaths, 7.71% were attributable to temperatures above or below the level at which mortality was lowest. Cold accounted for 7.29% of all deaths and heat for 0.42%.

In The Lancet Planetary Health in 2021, Qi Zhao and colleagues extended the approach to the whole world, using data from 750 locations in 43 countries. They estimated that about 5.08 million deaths a year between 2000 and 2019 were associated with non-optimal temperatures, 9.43% of all deaths. About 4.59 million were linked to cold and about 489,000 to heat, a ratio of roughly nine to one.

Share of all deaths attributable to cold and heat

384 locations in 13 countries, 1985 to 2012.

Cold7.29%
Heat0.42%

Gasparrini A and colleagues, The Lancet, 2015. Bars scaled to the cold figure.

Deaths linked to non-optimal temperatures worldwide

Estimated deaths per year, 2000 to 2019.

4.59 milliondeaths a year linked to cold
489,000deaths a year linked to heat

Zhao Q and colleagues, The Lancet Planetary Health, 2021. Data from 750 locations in 43 countries.

A pointillist illustration: a terrace of dark houses under snow at night, snow falling, a single window glowing amber.
A snowy terrace at night. Most cold-related deaths happen on ordinary winter days, not in the worst freezes.

It is not mostly the extreme cold

One finding in Gasparrini’s study is easy to miss. Extreme temperatures, both hot and cold, accounted for only 0.86% of all deaths. Most temperature-related deaths happened on moderately cold days, and few on the rare days of severe frost.

Frequency explains it. Each moderately cold day adds only a small amount of risk, but a winter has many such days. A cold snap raises risk more sharply but lasts a few days, so over a year the ordinary cold days add up to more deaths.

The practical consequence is for warnings. Public warnings and emergency responses are usually triggered by extreme events, and if most cold-related deaths happen outside those events, warnings alone will miss most of the problem.

Heat and cold also behave differently. The risk from heat rises steeply once temperatures pass a threshold, so a heatwave concentrates deaths into a few days. Cold risk rises more gradually and lasts longer, with deaths from heart disease and respiratory illness often following a cold spell by days or weeks.

Most temperature-related deaths are not on extreme days

Share of all deaths attributable to non-optimal temperatures.

All non-optimal temperatures7.71%
Extreme heat and cold only0.86%

Gasparrini A and colleagues, The Lancet, 2015. Extreme means below the 2.5th or above the 97.5th percentile of local temperature.

How cold harms the body

Hypothermia, a fall in core body temperature, is the most direct way cold kills, yet it accounts for only a small share of cold-related deaths in most countries.

Heart attacks, strokes and respiratory disease supply most of the excess. Cold makes blood vessels near the skin constrict, which raises blood pressure, and blood becomes thicker and more likely to clot. Breathing cold air can irritate the airways and worsen chronic lung disease. Respiratory infections also spread more easily in winter, when people spend more time indoors.

A pointillist illustration: a frosted window above a cold grey radiator, a single amber valve glowing at its side.
A frosted window over a radiator. Indoor temperature is one of the few parts of cold risk people can change.

The Eurowinter study, by William Keatinge, Gavin Donaldson and colleagues in The Lancet in 1997, compared regions across Europe from Finland to Greece. Mortality rose more for each degree of cold in regions with mild winters than in regions with harsh ones. People in colder places tended to heat their homes more, wear more clothing outdoors and spend less time outside without protection.

That finding has shaped much of the thinking since. It suggests that cold-related deaths depend on how well people are protected from the cold they experience, more than on how cold a place gets. It is also one reason why countries with mild winters and poorly insulated housing, such as the UK, Ireland and Portugal, have often recorded larger seasonal rises in deaths than Scandinavian countries with much harsher winters.

Homes, heating and who is at risk

Most of the risk falls on older people. The body’s ability to sense and respond to cold declines with age, and older people are more likely to have the heart and lung conditions that cold makes worse.

Housing has been a major focus, especially in the UK. In a 2011 report for Friends of the Earth, the Marmot Review Team estimated that 21.5% of excess winter deaths in England and Wales could be attributed to the coldest quarter of homes, almost three times as many as if deaths were spread evenly across housing.

Evidence at the individual level is less clear-cut than that estimate suggests. In the British Journal of General Practice in 2018, Peter Tammes and colleagues used primary care records to study who died on colder days in England. Each 1°C fall in the three-day average temperature raised the risk of death by about 1.1%. They found no clear modifying effect of sociodemographic characteristics or the energy efficiency of people’s homes, although risk was higher in northern England.

Cold homes plausibly contribute to winter deaths, then, but the effect of housing is harder to measure than the broad population estimates imply.

The debate about winter deaths

Not everyone accepts that the large cold-related death figures reflect cold itself.

Patrick Kinney and colleagues looked at 36 US and 3 French cities in Environmental Research Letters in 2015. Winter mortality was about 16% higher than summer mortality across all of them, whether their winters averaged below freezing or above 20°C, and how cold a city’s winter was had no relationship to how many extra deaths it had. They also showed that if a model does not control well for season, cold effects look much larger than they are.

Philip Staddon and colleagues found, in Nature Climate Change in 2014, that excess winter deaths in England and Wales had stopped tracking the number of cold days, and that influenza better explained year-to-year variation. They argued that climate warming would not reduce winter mortality.

Others disagree. Donaldson and Keatinge, in the BMJ in 2002, estimated that only 2.4% of excess winter deaths in south-east England over a decade were due to influenza, and argued that cold stress was the main driver. Shakoor Hajat and Gasparrini, in Epidemiology in 2016, argued that the common measure of excess winter deaths, which compares December to March with other months, is itself biased: it misses cold-related deaths that happen outside those months and counts deaths that are seasonal for other reasons.

The disagreement matters for climate policy. If cold causes most winter deaths, warming will save many lives in cold regions and partly offset the rising toll from heat. If infections and other seasonal factors mainly drive winter deaths, the savings will be smaller.

Is winter mortality really about cold?

Studies that reach different conclusions.

StudyFindingImplication
Kinney and colleagues, 2015Similar winter excess in cold and warm citiesCold effect may be overstated
Staddon and colleagues, 2014Winter deaths no longer track cold daysWarming may not cut winter deaths
Donaldson and Keatinge, 2002Influenza only 2.4% of excess winter deathsCold stress is the main driver
Hajat and Gasparrini, 2016Winter death index is biasedNeed temperature-based methods

Environmental Research Letters; Nature Climate Change; BMJ; Epidemiology.

When the power goes out: Texas, 2021

In February 2021, Winter Storm Uri brought record cold to Texas, the most dramatic recent example of cold-related death in a wealthy country. The state’s power grid failed, leaving millions of people without electricity or heat for days.

The official state death toll, revised in January 2022, was 246 people in 77 counties, ranging in age from infants to a 102-year-old. Most died of hypothermia. Others died in road accidents, falls and house fires, from carbon monoxide poisoning while trying to heat their homes (at least 19 such deaths), or after losing power to oxygen machines and other medical equipment. Independent analyses of excess deaths suggested the true toll was higher.

Texas was an extreme event in a place unused to extreme cold, which fits the Eurowinter finding that people in milder climates are less protected. It also showed that in a modern cold disaster the main danger is often the loss of heat and power, as well as the outdoor temperature.

Winter Storm Uri, Texas, February 2021

Official state count of storm-related deaths.

246deaths in 77 counties, most from hypothermia
19+deaths from carbon monoxide while trying to keep warm

Texas Department of State Health Services, revised January 2022, as reported by NPR.

What reduces the risk

Most cold-related deaths are spread across ordinary winter days, and the measures with the best supporting evidence are ordinary too.

Public health guidance in England, including the National Institute for Health and Care Excellence guideline on cold homes published in 2015, focuses on identifying people at risk, particularly older people and those with heart or lung disease, and helping them keep their homes warm. That includes referral to heating and insulation schemes and checking that people can afford to heat their homes.

The UK Health Security Agency and the Met Office have issued cold-health alerts in England since 2023, graded by the expected impact on health services rather than by temperature alone. This follows directly from the research: the alerts aim to prompt action by health and care services before and during cold spells, including outside the most extreme weather.

Seasonal flu and COVID-19 vaccination also matter, whichever side of the cold-or-infection debate one takes, because respiratory infections contribute a share of winter deaths. Where cold is rare, as in Texas, the evidence from 2021 points to resilient power and heating systems as the first line of protection.

A pointillist illustration: power lines and poles stretching across a frozen field under a grey sky, a single farmhouse window glowing amber in the distance.
Power lines across a frozen field. In Texas in 2021, the loss of heat and power turned cold into a disaster.

Staying safe in the cold: what the CDC, NWS and UKHSA say

This section collects practical advice from US and UK public-health pages. It is general information, not medical advice.

Practical advice
TopicWhat the pages say
Hypothermia signsAdults: shivering, exhaustion, confusion, fumbling hands, memory loss, slurred speech, drowsiness. Babies: bright red, cold skin and very low energy. The CDC says that if body temperature is below 95°F, get medical attention immediately, and notes hypothermia can occur above 40°F if you are chilled while wet.
What to do (CDC)Get medical attention as soon as possible. Move the person to a warm room, remove wet clothing, and warm the centre of the body (chest, neck, head, groin). Warm drinks can help but avoid alcohol, and give no drinks to an unconscious person. If someone is severely hypothermic and unresponsive, handle gently, call emergency help and begin CPR.
Heating and carbon monoxide (CDC)Install a battery-operated CO detector. Never use generators, gas or charcoal grills or camp stoves inside the home; keep generators at least 20 feet from windows, doors and vents. Do not use the stove for heat. Keep space heaters at least 3 feet from drapes, furniture and bedding. Service the heating system yearly.
Wind chill (NWS)The index applies at or below 50°F with wind above 3 mph. NWS example: 0°F with a 15 mph wind feels like −19°F, and bare skin can freeze in about half an hour.
Indoor warmth (UKHSA, England)Heat the rooms you spend most time in to at least 18°C. Have an EN 50291 carbon monoxide alarm fitted and appliances checked by a Gas Safe engineer.
NeighboursThe CDC says to check on older friends and neighbours often to make sure their homes are heated. UKHSA says to keep in touch and ask whether they feel unwell or need practical help.

A 2014 Public Health England review proposed 18°C as a winter threshold but said the evidence was weak and should not be framed as strong.

What we could not confirm: a numeric WHO indoor temperature from a WHO page, any NHS.uk page (none fetched), and the CDC’s separate hypothermia page, which failed to load.

What the evidence does not show

Not every death attributed to cold was necessarily caused by cold. The estimates come from statistical models, and results depend on how they handle season and other factors.

A large net fall in temperature-related deaths from climate warming is not shown either. Some studies expect savings in cold regions; others argue winter deaths will change little, while heat deaths will rise.

How much cold homes matter is also unsettled. Population estimates are large, but individual-level studies have found weaker or inconsistent effects of housing quality.

Heat is still important. Heat deaths are concentrated in time, rising, and more sensitive to climate change.

What to do

Steps sorted by when they matter. Each one says how strong the evidence is: Measured Official advice Untested

Before

  • Know who is at risk around you, especially older people and those with heart or lung disease, and check they can keep their home warm. Official advice see Sources
  • Ask about heating and insulation schemes, and whether heating is affordable. Official advice see Sources
  • Get seasonal flu and COVID-19 vaccinations. Respiratory infections contribute a share of winter deaths. Official advice see Sources

If it happens

  • Act on cold-health alerts before and during cold spells, not only in the most extreme weather. Most cold-related deaths fall on ordinary winter days. Measured see Sources

Questions people ask

Does cold kill more people than heat?

In most places, yes. A study of 74 million deaths found cold linked to 7.29% of deaths and heat to 0.42%.

Do most cold deaths happen in severe cold snaps?

No. Most occur on moderately cold days, which are far more common. Extreme temperatures accounted for 0.86% of all deaths.

What do people actually die of in cold weather?

Mostly heart attacks, strokes and respiratory disease. Hypothermia is a small share of cold-related deaths.

How many people died in the 2021 Texas winter storm?

The official toll is 246, most from hypothermia. Independent analyses suggest the real figure was higher.

Will climate change reduce winter deaths?

Researchers disagree. Some expect fewer cold deaths as winters warm; others find winter deaths no longer track cold and may not fall much.

The short version

  • In a 13-country study, cold was linked to 7.29% of deaths and heat to 0.42%.
  • Worldwide, about 4.6 million deaths a year are associated with cold and about 489,000 with heat.
  • Most cold-related deaths happen on moderately cold days rather than in extreme freezes.
  • Researchers disagree on how much winter mortality is caused by cold, and how much by infections and season.
  • The 2021 Texas storm killed at least 246 people, most from hypothermia after the power grid failed.

This article summarises published research on cold-related mortality. It is not medical advice. If you are worried about an older relative or someone with heart or lung disease during cold weather, their doctor or a local public health service can advise on keeping warm safely.

Further reading. Gasparrini and colleagues, ‘Mortality risk attributable to high and low ambient temperature’, The Lancet, 2015, is the foundation of most current estimates and is open access. Kinney and colleagues, ‘Winter season mortality: will climate warming bring benefits?’, Environmental Research Letters, 2015, is the clearest statement of the sceptical view.

Sources

  1. Gasparrini A, Guo Y, Hashizume M, and colleagues. Mortality risk attributable to high and low ambient temperature: a multicountry observational study. The Lancet, 2015;386(9991):369–375. doi:10.1016/S0140-6736(14)62114-0.
  2. Zhao Q, Guo Y, Ye T, and colleagues. Global, regional, and national burden of mortality associated with non-optimal ambient temperatures from 2000 to 2019. The Lancet Planetary Health, 2021;5(7):e415–e425. doi:10.1016/S2542-5196(21)00081-4.
  3. The Eurowinter Group. Cold exposure and winter mortality from ischaemic heart disease, cerebrovascular disease, respiratory disease, and all causes in warm and cold regions of Europe. The Lancet, 1997;349(9062):1341–1346. doi:10.1016/S0140-6736(96)12338-2.
  4. Kinney PL, Schwartz J, Pascal M, and colleagues. Winter season mortality: will climate warming bring benefits? Environmental Research Letters, 2015;10(6):064016. doi:10.1088/1748-9326/10/6/064016.
  5. Staddon PL, Montgomery HE, Depledge MH. Climate warming will not decrease winter mortality. Nature Climate Change, 2014;4(3):190–194. doi:10.1038/nclimate2121.
  6. Donaldson GC, Keatinge WR. Excess winter mortality: influenza or cold stress? Observational study. BMJ, 2002;324(7329):89–90. doi:10.1136/bmj.324.7329.89.
  7. Hajat S, Gasparrini A. The excess winter deaths measure: why its use is misleading for public health understanding of cold-related health impacts. Epidemiology, 2016;27(4):486–491. doi:10.1097/EDE.0000000000000479.
  8. Tammes P, Sartini C, Preston I, Hay AD, Lasserson D, Morris RW. Use of primary care data to predict those most vulnerable to cold weather: a case-crossover analysis. British Journal of General Practice, 2018;68(668):e146–e156. doi:10.3399/bjgp18X694829.
  9. Marmot Review Team. The Health Impacts of Cold Homes and Fuel Poverty. Friends of the Earth and the Marmot Review Team, 2011.
  10. National Institute for Health and Care Excellence. Excess winter deaths and illness and the health risks associated with cold homes. NICE guideline NG6, 2015.
  11. UK Health Security Agency. Weather-health alerting system: cold-health alerts. 2023 onwards.
  12. Texas Department of State Health Services. February 2021 winter storm-related deaths, Texas. Revised January 2022; as reported by NPR, 3 January 2022.
  13. CDC. Preventing Hypothermia (Aug. 25, 2026), cdc.gov; Recognizing Hypothermia (Aug. 26, 2026), cdc.gov; Safety Guidelines: During & After a Winter Storm, cdc.gov.
  14. National Weather Service. Wind Chill Temperature Index brochure. weather.gov.
  15. UKHSA. Keeping warm and well: staying safe in cold weather, updated Apr. 28, 2026. gov.uk.
  16. Public Health England. Minimum home temperature thresholds for health in winter, Oct. 2014. gov.uk.

Similar Posts