Picture the public image of a heat death. A marathon runner collapsing at mile twenty. A labourer on a roof. Someone with a core temperature of 41 degrees, confused and flushed, rushed to an emergency department packed with ice.
Those deaths are real. They are also the minority.
Most people killed by a heat wave never get a diagnosis of heatstroke. They die at home, often alone, often at night, of what the death certificate records as a heart attack, a stroke, heart failure or “sudden cardiac death.” The heat is the trigger. The heart is the route. And because the paperwork names the route rather than the trigger, the health system consistently underestimates what hot weather is doing to its patients.
The numbers hide in plain sight
The scale only becomes visible when epidemiologists compare deaths during hot periods with what would normally be expected. When a team led by Joan Ballester did exactly that for Europe, publishing in Nature Medicine in 2023, it estimated 61,672 heat-related deaths between late May and early September 2022. Italy alone accounted for about 18,000. Very few of those people were counted as heat casualties at the time.
Globally, the 2025 report of the Lancet Countdown on health and climate change put heat-related deaths at around 546,000 a year, a rise of 63 per cent since the 1990s.
In the United States, a study in Circulation supported by the National Heart, Lung, and Blood Institute looked specifically at cardiovascular deaths. The study, led by Sameed Khatana and published in 2023, found that extreme heat currently accounts for less than 1 per cent of cardiovascular deaths, which sounds small. But it estimated about 1,651 excess cardiovascular deaths a year linked to extreme heat over 2008 to 2019, and projected that figure to rise to between 4,320 and 5,491 a year by mid-century, depending on the emissions path. Older adults and Black adults were projected to see much larger increases than younger and white adults.
That last detail matters. Heat does not create cardiovascular risk from nothing. It finds the risk that is already there and pulls the date forward.
Why the heart takes the hit
To shed heat, the body pushes blood to the skin. The heart has to pump harder and faster to keep pressure up while fluid is lost through sweat. Blood becomes more concentrated and more prone to clotting. For a healthy 30-year-old this is an inconvenience. For an 80-year-old with heart failure, coronary disease or poorly controlled hypertension, it is a stress test they did not sign up for, repeated every day and every night the heat persists.
Night matters more than we admit. Hot nights deny the body its recovery period. Someone who could tolerate one hot afternoon may not tolerate five in a row without relief.
This physiology also explains the timing that confuses people. Deaths do not always peak on the hottest day. They often cluster on the day of and the few days after, as strain accumulates, fluid balance drifts and a vulnerable heart finally decompensates. By the time the forecast has cooled and the news has moved on, the mortality curve is still climbing. Anyone measuring the success of a heat response by what happens on the peak afternoon is looking at the wrong window.
Why the paperwork keeps getting it wrong
The coding problem is not carelessness. A death certificate asks the certifier for a chain of causes leading to death. An elderly man with known heart failure who dies at home in a hot flat has a perfectly plausible cardiac chain. Nothing in the process asks what the temperature was. The certifier writes what is clinically defensible, and the heat disappears.
That creates a quiet feedback loop. Health systems budget and plan around counted deaths. Heat deaths are mostly uncounted, so heat gets little planning, so the next summer’s deaths are also mostly uncounted. The only reason we see the true scale at all is the statistical work of excess-mortality studies, which arrive months or years later, long after budgets have been set.
The loop matters for money as much as for mortality. Hospitals see a surge of cardiac and renal admissions during hot spells and record them under cardiology and nephrology. The cost of heat is real and already on the books; it is just filed under other names.
The uncomfortable part: some of this risk sits in our prescriptions
Here is the part clinicians rarely discuss at the start of summer. Many of the medicines that keep cardiovascular patients alive also blunt their ability to cope with heat. Diuretics deplete fluid and salt. Some blood pressure drugs impair the body’s ability to raise cardiac output or to feel thirst properly. Medicines with anticholinergic effects can reduce sweating.
None of this means those drugs should be stopped. It means the patients most at risk from heat are frequently the ones we have most actively treated, and very few of them have ever had a conversation about what to do when the forecast turns. A heat plan is rarely part of a heart failure care plan. It should be as routine as a sick-day plan for someone with diabetes.
A back-of-envelope for one practice
Take an illustrative primary care practice of 10,000 patients in a temperate city. Suppose a few hundred have diagnosed heart failure or a recent cardiac event, and a larger group are over 75 and on diuretics or several blood pressure drugs. The exact numbers will vary, but most practices could pull such a list in an afternoon.
Now suppose that list is 300 people. A structured pre-summer call of ten minutes each is 50 hours of a nurse’s time, spread across a spring. During a declared heat wave, a shorter check-in to the highest-risk third, 100 people living alone, is perhaps two days of work for one person. Against that, a single avoided emergency admission for decompensated heart failure typically saves several days of inpatient care.
This is not proof of benefit; there are few trials of exactly this intervention. But it shows the scale of effort involved is small enough that the question becomes why it is not routine, rather than whether it is affordable.
There is an incentive problem here too. Heat mortality falls across cardiology, primary care, emergency medicine, public health and city government. Because it belongs to everyone, it belongs to no one’s performance targets.
The case against medicalising heat
There is a reasonable objection here. Framing heat as a cardiology problem risks turning a structural issue into an individual one. The real drivers of heat mortality include housing that traps heat, urban design without shade, poverty that makes air conditioning unaffordable and, underneath, a warming climate. A phone call does not fix a top-floor flat with windows that do not open.
That critique partly holds. Clinical outreach is triage, not a solution, and it would be a mistake for health systems to treat it as a substitute for building standards, urban greening and energy support. There is also a risk that air conditioning, the most effective personal protection, raises energy demand and emissions if it becomes the default answer.
But the objection is strongest over a decade and weakest over a single summer. Housing stock changes slowly. The patients at risk next July are already known, already prescribed, already on registers. Both things can be true: heat is a structural problem, and the health system is sitting on the one list that can reduce deaths this year.
What works is boring and local
The best evidence for prevention does not come from high technology. It comes from cities that organised themselves.
Ahmedabad in India launched South Asia’s first heat action plan in 2013 after a deadly heat wave. The plan combined early warnings, training for health workers, public messaging and coordination across agencies. An evaluation by Jeremy Hess and colleagues, published in 2018, estimated about 2,380 deaths avoided across the 2014 and 2015 heat seasons compared with a 2007 to 2010 baseline, roughly 1,190 a year. The confidence interval was wide, and the authors called it a pilot evaluation. But the direction was clear, and the ingredients were ordinary.
For cardiovascular risk specifically, a serious heat response would look like this:
- Registers, not just warnings. Clinics know which patients have heart failure, recent heart attacks or complex medication regimens. Heat alerts should trigger outreach to those people, not only a press release.
- Medication reviews before summer. A short, planned conversation about fluids, symptoms and when to call, especially for patients on diuretics.
- Checking on people who live alone. Social isolation is a heat risk factor because nobody notices early decline.
- Cool places that people can actually reach. A cooling centre across a highway is not a cooling centre for an 85-year-old without a car.
- Better cause-of-death coding. If a heat wave is not recorded as a contributor, it will not be budgeted for.
Who needs to move
For health ministries, the cheapest high-value step is to make heat-health action plans require clinical components, not only public messaging, and to fund excess-mortality monitoring in near real time, so heat deaths are counted during the season rather than estimated years later.
For hospital leaders, it means treating hot spells like predictable surges: planning staffing for cardiac and renal admissions and making sure discharge planning asks whether a frail patient is going home to a dwelling that will be dangerous that week.
For payers and insurers, it means recognising pre-summer medication reviews and heat check-ins as reimbursable care rather than goodwill. What is not paid for is rarely done at scale.
For clinicians, the change is small and concrete: add heat to the heart failure and hypertension conversation, agree what the patient should do with diuretic doses or fluid limits during hot weather, and write it down. Those adjustments are individual clinical decisions, which is exactly why they need a planned conversation rather than a leaflet.
For founders building remote monitoring or care-coordination tools, the opportunity is not another consumer heat app. It is plugging weather alerts into clinical registers so that the right list lights up on the right morning.
What to watch
Over the next year or two, watch whether more countries publish timely heat excess-mortality figures during the season rather than after it, and whether cardiology and heart failure guidance begins to include explicit heat advice. Watch cause-of-death coding practice, since even a prompt to record heat as a contributing factor would change what gets counted. And watch cities with heat action plans for evaluations that report cardiovascular outcomes, not only total deaths. That is the evidence that would move this from sensible idea to standard of care.
The takeaway
We have framed heat as an environmental emergency and treated it with environmental tools: warnings, maps and shade. Those help. But the body count is being produced inside cardiology and primary care caseloads, by patients whose risk we already know and have already documented.
The most effective heat intervention available to most health systems this year is not a new technology. It is a phone call, made before the heat wave, to the patient on three cardiovascular drugs who lives alone on the top floor.
Does your service have a heat protocol for cardiovascular patients, or does it stop at the public warning? Tell us what you have seen work.
