One number has been doing a lot of traveling lately: 546,000 heat deaths a year. It comes from a serious report by serious people. It also tells you almost nothing about your own risk.
The figure is global. It is an annual average, not a count from any single year. And it has a date on it, from the 2025 report of the Lancet Countdown, which the World Health Organization publicized on October 29, 2025. Moving a worldwide annual average into a sentence about a reader in Ohio is a trick, and it is worth naming as one.
Climate reaches a human body through a small number of specific routes. Heat. Air quality. Things that bite. Things that contaminate water and food. Weather events violent enough to injure people and wreck the places they live. The mental aftermath of all of it.
Those routes are not equally well understood. Some are measured directly in American data. Some are projections under assumptions most readers never see. Sorting them is the whole job, and this article is about how to do that sorting yourself.
The pathways the CDC names
The Centers for Disease Control and Prevention runs a Climate and Health Program, and its page on health effects lists ten topic areas: air pollution; allergens and pollen; food security; food and waterborne diarrheal disease; mental health and stress-related disorders; pollen and your health; precipitation extremes; temperature extremes; vectors; and wildfires. Two of the ten cover pollen, so the real count of distinct subjects is nine.
That page carries no numbers at all. Its umbrella statement is a list of categories: more respiratory and cardiovascular disease, injuries and premature deaths from extreme weather, shifts in where food-borne and water-borne illnesses occur, and threats to mental health.
CDC’s regional page, last reviewed August 12, 2026, is more specific and more hedged. CDC says temperature and precipitation changes are increasing health risks tied to wildfire and ground-level ozone pollution. It says warmer air and water with more intense extreme events are expected to increase exposure to waterborne and foodborne disease. It projects that cold-related deaths will fall and heat-related deaths will rise, with the increases outpacing the reductions in most regions. It expects allergic illnesses including hay fever to increase. It says warming may expose more people to the ticks and mosquitoes that carry Lyme disease, and names Zika, West Nile and dengue as other examples. And it says extreme weather can have lasting mental health consequences in affected communities.
Notice the verbs. Is increasing is one kind of claim. Are expected to increase is another. CDC switches between them deliberately, and the switch is the information.
One oddity worth flagging: the pathway list itself was last reviewed on February 29, 2024, while the regional page describing those pathways was reviewed in August 2026. The summary page a reader is most likely to land on is the older one.
Four kinds of number, and why mixing them up ruins everything
Almost every confusing claim in this subject comes from blending four different kinds of statement. Keep them separate and most of the confusion evaporates.
A measured change in exposure. Something about the environment got measurably different. Pollen seasons got longer. Hot days got more frequent. This is often the best-evidenced category, because the thing being measured is physical and instruments have been pointed at it for decades.
A measured change in a health outcome. Actual people showed up sick or died, and somebody counted. This is harder than it sounds, because what gets counted depends on how doctors code, how surveillance systems define a case, and whether anyone looked.
A projected outcome under a scenario. Cases are modeled to rise by some percentage by some year, assuming a specific path for future emissions. The scenario label is load-bearing and usually missing from the summary a reader sees.
An attributed outcome. A share of observed deaths or illnesses is assigned to human-caused warming by comparing the world that happened against a modeled world that did not. This is legitimate research and the most easily misread number of the four.
Sorting the pathways by how well they hold up
Heat is the strongest. The physiological route is direct, the exposure is easy to measure, and the outcome shows up in emergency departments within hours. Even here, American counting is openly contested. EPA’s heat-related deaths indicator reports that more than 14,000 Americans have died directly from heat-related causes since 1979 according to death certificates. The same page offers two different annual figures: more than 1,300 deaths a year from other estimation approaches, and about 750 a year in the underlying-and-contributing-causes dataset the indicator charts. EPA attaches a warning to its own indicator: the count runs low, because the certificate often names a heart or lung cause rather than the heat behind it. Heat as an outdoor and workplace exposure, and the physiology of heat and cold, are covered separately in this library.
Air quality is strong on one joint and weaker on the other. That breathing fine particles raises mortality risk is among the best-supported findings in environmental epidemiology, built on cohorts of tens of millions of Americans. Our article on air pollution standards covers that evidence and the rules built on it. The weaker joint is the climate-to-exposure link: how much of a given bad air day is attributable to warming rather than to traffic, industry or wind. Wildfire smoke is covered separately as well.
Pollen is unusually well measured, on the exposure side only. Ziska and colleagues reported in the Proceedings of the National Academy of Sciences in 2011 that the ragweed pollen season lengthened by as much as 13 to 27 days at latitudes above roughly 44 degrees north since 1995, driven primarily by a delay in the first fall frost. The lead author worked at the USDA Agricultural Research Service. That is a clean measured change in exposure. There is no comparable American measurement of how many additional people got sick because of it.
Vector-borne disease is the messiest. CDC reports that more than 80,000 Lyme disease cases are reported each year by states and jurisdictions, and that an estimate built from insurance claims suggests around 476,000 people may be diagnosed and treated annually. CDC attaches its own caveat to the larger figure: it may be an overestimate, because billing codes do not always reflect a confirmed diagnosis. EPA’s Lyme indicator, last updated June 13, 2025, shows reported incidence ranging from a minimum of 3.2 cases per 100,000 people in 1993 to a maximum of 18.8 per 100,000 in 2022, with 15 states plus the District of Columbia accounting for about 95 percent of reported cases in 2022. EPA cautions on the same page that changes in how Lyme disease has been defined and reported over time make long-term trends challenging to identify. Rising case counts here reflect some mixture of real range expansion, better awareness, more testing, and three decades of surveillance rule changes, and nobody has cleanly separated them.
Water-borne and food-borne illness sits mostly in the projection category. The Fifth National Climate Assessment projects vibriosis cases rising 51 percent by 2090 under an intermediate scenario, and Valley fever rising 220 percent by the end of the century under a very high scenario. Those are modeled futures, and the scenario attached to each changes what it means.
Mental health is real and barely quantified at national scale. CDC’s statement is qualitative. There is no American equivalent of a heat death count for the psychological aftermath of a flood.
Three American numbers, and exactly what each counts
A measured outcome, dated and bounded: CDC’s MMWR reported on September 17, 2026 that the ESSENCE syndromic surveillance system recorded 145,321 heat-related illness emergency department visits in 2026 as of September 10. About 15 percent of those, 21,775 visits, occurred in federal health regions 1 through 7 during a heat dome from June 28 to July 7, 2026. During that stretch, roughly 65 percent of the US population, 221 million people across 42 states and the District of Columbia, had at least one day of major or extreme heat risk.
A measured outcome with a known floor problem: the Fifth National Climate Assessment reports that from 2011 to 2019, an average of 3,500 heat-related injuries and 38 fatalities a year were reported to OSHA. That is what employers reported, which is not the same as what happened.
A measured change in the physical setting: the assessment also reports that temperatures in the contiguous United States have risen 2.5 degrees Fahrenheit and Alaska by 4.2 degrees since 1970, and that across 50 large US cities, the average number of heatwaves has doubled since the 1980s.
Two more from the same assessment are worth knowing because they describe the system meant to treat you: the healthcare sector is responsible for 8.5 percent of US greenhouse gas emissions, and 9.3 percent of hospitals, 10.2 percent of nursing homes and 12.1 percent of pharmacies sit at risk of flooding.
The scale problem
Back to 546,000.
The Lancet Countdown’s 2025 report puts the rate of heat-related deaths up 23 percent since the 1990s, reaching that annual average. WHO’s release on the same report adds that heat exposure cost 640 billion potential labor hours in 2024. Both figures are global.
WHO’s own climate and health fact sheet, dated October 12, 2023, projects approximately 250,000 additional deaths per year between 2030 and 2050 from undernutrition, malaria, diarrhoea and heat stress alone, and says 3.6 billion people already live in areas highly susceptible to climate change. Also global.
The attribution research behind these figures is real. Vicedo-Cabrera and colleagues, writing in Nature Climate Change in 2021, used data from 732 locations in 43 countries between 1991 and 2018 and estimated that 37.0 percent of warm-season heat-related deaths can be attributed to human-caused climate change, with a range across locations of 20.5 to 76.3 percent. That range is as important as the headline. The burden is not evenly distributed, and a 43-country average belongs to no individual country.
None of these numbers has a denominator a reader belongs to. Air conditioning prevalence, housing stock, age structure, baseline health and the capacity of the local health system all vary enormously, and the global figure averages across all of it. Use these numbers as evidence the problem is large. Do not use them as a personal risk estimate, and be skeptical of any article that slides from one to the other in a single paragraph.
There is also a genuine disagreement here that nobody has resolved. The Lancet Countdown’s own page on the 2025 report says: “Health risks and impacts of climate change are worse than ever before across 13 of 20 impact indicators.” WHO’s release about that same report says: “12 of 20 key indicators tracking health threats have reached record levels.” Thirteen and twelve. These may describe different indicator sets, or one may simply be wrong. Both statements are published by the organizations that made them, and this article is not going to pick.
One housekeeping note that will matter soon. The Lancet Countdown’s 2026 global report, its tenth, is scheduled to launch on October 28, 2026. Every Countdown figure above comes from the 2025 report. Any source citing the most recent Lancet Countdown without naming a year should be treated as undated.
When the official record itself moves
This is the part most coverage leaves out, and a reader checking sources will run straight into it.
On October 20, 2026, globalchange.gov, the domain of the US Global Change Research Program, served a page with five items, no navigation, no footer and no archive banner. The items included a NASA soil moisture satellite described as new, confirmation that 2014 was the hottest year on record, and the release of the Third National Climate Assessment in May 2014. Meanwhile nca2023.globalchange.gov serves the complete Fifth National Climate Assessment. The current assessment exists and is reachable. The front door to it serves content from a decade ago.
The Fifth Assessment is also under proposed amendment. The Office of Science and Technology Policy, on behalf of the Global Change Research Program, published a notice of availability for public comment on August 25, 2026, at 91 FR 54893. The proposed amendment addresses the interpretation and federal use of assessment results that depend materially on RCP8.5, SSP5-8.5, SRES A2 or analogous high-emissions pathways. Comments closed at 11:59 p.m. Eastern on September 9, 2026, and no final action had appeared as of October 20, 2026.
That is directly relevant to the numbers above. The Valley fever projection of 220 percent is a very high scenario result. The vibriosis projection of 51 percent is an intermediate scenario result. Readers almost never see the scenario label, and a federal process is now arguing about how much weight the high-scenario results should carry.
EPA’s climate indicator pages tell a similar story in miniature. The index page loads and still lists the Health and Society indicators, including heat-related deaths, Lyme disease, West Nile virus and ragweed pollen season. The index carries a last-updated date of March 31, 2022, while the heat-related deaths page it points to is dated February 2025 and the Lyme page June 13, 2025. The ragweed pollen season page would not open for us at all.
Then there is a quieter break in the chain. NOAA announced on May 8, 2025 that its Billion Dollar Weather and Climate Disasters product would be retired, with no updates beyond calendar year 2024, and that the 1980 to 2024 reports and data remain archived and available. The Fifth Assessment’s drought figure, roughly 99 climate-related deaths a year, is drawn from that database. A retired dataset is now embedded inside a federal assessment that is itself under proposed amendment.
Describe what loads. That is all anyone can honestly do here.
Four questions to ask of any climate-and-health number
- Global or local? If the answer is global, it does not describe your risk. Ask whether a US-specific version exists.
- Single year or annual average? An average smooths out the extreme years that make a stretch of weather memorable, and a single year does the opposite.
- Measured, projected or attributed? And if projected, under which emissions scenario? A high-scenario result is a bounding exercise, not a forecast.
- What date is on it, and has the source moved? Check the page’s own last-reviewed date, not the date of the article quoting it.
When this becomes a medical question rather than a data question
A few specifics are worth separating from the general subject.
If you have asthma, chronic obstructive pulmonary disease or heart failure, and your symptoms reliably worsen during hot or smoky stretches, that pattern belongs in a conversation with your clinician before the next one arrives, not during it.
If you find an attached tick, or develop an expanding rash, fever, headache or joint pain in the weeks after a tick bite, that is a same-week call. Lyme disease is far more straightforward to treat early, and the diagnosis relies heavily on exposure history that only you can supply.
Confusion, hot and dry skin, or collapse during hot weather is a 911 call, not a wait-and-see.
If a flood, fire or storm has left you or someone in your household unable to sleep, unable to function, or thinking about self-harm weeks afterward, that is a reason to contact a clinician. The 988 Suicide and Crisis Lifeline takes calls and texts around the clock.
None of this is medical advice, and nothing here should be used to delay care you would otherwise seek.
What to take away
Before accepting any claim in this subject, find out which of the four kinds of number it is: a measured exposure, a measured outcome, a projection under a named scenario, or an attributed share. Most of the confusion in climate and health is not scientists disagreeing. It is four very different claims wearing the same clothes.
Then narrow to what you can act on. For nearly everyone in the United States, three of the pathways are local, checkable and worth attention this year: heat where you live, air quality where you live, and the ticks and mosquitoes in your county. The global totals are evidence the problem is large. They were never going to tell you what to do on Tuesday.
Sources
- U.S. Global Change Research Program. “Fifth National Climate Assessment, Chapter 15: Human Health.”
- U.S. Global Change Research Program. “Fifth National Climate Assessment, Chapter 2: Climate Trends.”
- U.S. Global Change Research Program. “GlobalChange.gov home page (as served 20 October 2026).”
- Office of Science and Technology Policy. “United States Global Change Research Program; Availability for Public Comment on the Proposed Amendment to the Fifth National Climate Assessment (NCA5), 91 FR 54893, 25 August 2026.” federalregister.gov
- Centers for Disease Control and Prevention. “Effects of Climate Change on Health (last reviewed 29 February 2024).” cdc.gov
- Centers for Disease Control and Prevention. “Regional Climate Change Effects on Health (last reviewed 12 August 2026).” cdc.gov
- Centers for Disease Control and Prevention. “Lyme Disease Data and Surveillance: Facts and Statistics (last reviewed 9 September 2026).” cdc.gov
- Centers for Disease Control and Prevention. “Notes from the Field: Heat-Related Emergency Department Visits, June 28-July 7, 2026. MMWR 2026;75(36):563-565, 17 September 2026. DOI: 10.15585/mmwr.mm7536a2.” cdc.gov
- U.S. Environmental Protection Agency. “Climate Change Indicators: Heat-Related Deaths.”
- U.S. Environmental Protection Agency. “Climate Change Indicators: Lyme Disease (last updated 13 June 2025).”
- U.S. Environmental Protection Agency. “Climate Change Indicators in the United States (index page, last updated 31 March 2022).”
- NOAA National Environmental Satellite, Data, and Information Service. “Billion-Dollar Weather and Climate Disasters: Notice of Changes, 8 May 2025.” nesdis.noaa.gov
- Lancet Countdown on Health and Climate Change. “The 2025 Report of the Lancet Countdown on Health and Climate Change: Climate change action offers a lifeline.” lancetcountdown.org
- World Health Organization. “Climate inaction is claiming millions of lives every year, warns new Lancet Countdown report (29 October 2025).” who.int
- World Health Organization. “Climate change and health (fact sheet, 12 October 2023).” who.int
- PubMed. “Ziska L, Knowlton K, Rogers C, et al. "Recent warming by latitude associated with increased length of ragweed pollen season in central North America." Proc Natl Acad Sci U S A, 2011;108(10):4248-51. PubMed. DOI: 10.1073/pnas.1014107108.” doi.org
- PubMed. “Vicedo-Cabrera AM, Scovronick N, Sera F, et al. "The burden of heat-related mortality attributable to recent human-induced climate change." Nature Climate Change, 2021;11(6):492-500. PubMed. DOI: 10.1038/s41558-021-01058-x.” doi.org