The impact of global warming on health and mortality.

Abstract: Initial concern about the possible effects of global warming on infections has declined with the realization that the spread of tropical diseases is likely to be limited and controllable. However, the direct effects of heat already cause substantial numbers of deaths among vulnerable people in the summer. Action to prevent these deaths from rising is the most obvious medical challenge presented by a global rise in temperature. Strategies to prevent such deaths are in place to some extent, and they differ between the United States and Europe. Air conditioning has reduced them in the United States, and older technologies such as fans, shade, and buildings designed to keep cool on hot days have generally done so in Europe. Since the energy requirements of air conditioning accelerate global warming, a combination of the older methods, backed up by use of air conditioning when necessary, can provide the ideal solution. Despite the availability of these technologies, occasional record high temperatures still cause sharp rises in heat-related deaths as the climate warms. The most important single piece of advice at the time a heat wave strikes is that people having dangerous heat stress need immediate cooling, eg, by a cool bath. Such action at home can be more effective than transporting the patient to hospital. Meanwhile, it must not be forgotten that cold weather in winter causes many more deaths than heat in summer, even in most subtropical regions, and measures to control cold-related deaths need to continue.

Key Words: cold-related deaths, global warming, heat-related deaths, heat stress

**********

Infections in the summer months were largely brought under control by the beginning of the last century. Since then, winter cold has been the major seasonal factor causing death in all but tropical regions of the world. However, more direct effects of heat in summer cause a smaller but significant number of deaths. These sometimes increase sharply during record high temperatures in particular regions, and recent concerns about the effects of global warming on health have concentrated on the need to prevent these deaths from increasing. Both the cold-related and the heat-related deaths occur almost entirely in elderly people. Both are largely avoidable, and both need to be taken into account when considering the effects of rising temperatures.

Global warming has now been under way for long enough to allow some direct assessments to be made of its effects on seasonal deaths. These effects are often very different from earlier expectations, and depend a great deal on the effectiveness of measures taken to limit heat stress on people. These measures in turn vary greatly between Europe and North America. The purpose of this article is to summarize current information on the ways in which hot and cold weather cause deaths and on measures to prevent these.

Mortality rates are generally lowest at a mean daily temperature around 66[degrees]F, (1-5) the precise level varying with region. The mortality rate in each region rises as temperature either falls or rises from that level. The numbers of these excess deaths per year give the best measures of both heat-related and cold-related mortality rates. An older measure that had been used for winter mortality rate alone was the number of deaths per million during December to March, as an excess over mean mortality rate during the rest of the year. That is simple to calculate and can often provide a rough comparison of winter deaths at different times and in different regions, but it has major drawbacks. The worst of these is that it is determined by death caused by hot weather in summer as well as cold weather in winter. If there are more deaths caused by hot weather than cold weather, that comparison will suggest no excess winter deaths in that region, although cold is in fact producing high mortality rates in winter, balanced by high mortality rates caused by heat in summer.

Causes of Cold-related and Heat-related Deaths

Cold-related deaths are far more numerous than heat-related deaths in the United States, Europe, and almost all countries outside the tropics, and almost all of them are due to common illnesses that are increased by cold. Coronary and cerebral thrombosis account for about half of the cold-related deaths and respiratory disease for about half the rest. (6) Cold stress causes the increase in arterial thrombosis because the blood becomes more concentrated, and so more liable to clot, during exposure to cold. The sequence of events is that the body's first adjustment to cold stress is to shut down blood flow to the skin, to conserve body heat. (7-9) The shift of blood from the skin produces an excess of blood in central parts of the body. To correct that, salt and water are moved out from the blood into tissue spaces and are eventually excreted. This leaves behind increased levels of red cells, white cells, platelets, and fibrinogen, and causes increased viscosity of the blood (Fig. 1). All of these changes promote clotting.

The anticoagulant protein C, which normally hinders intravascular clotting, might be expected to increase in line with the increase in concentration of the thrombogenic elements of blood during exposure to cold. If it did, it would counter the effects of those elements, at least in part. However, it has a small enough molecule to allow it to diffuse through the walls of blood vessels. As a result, it redistributes from the blood into the tissue spaces of the body, and its plasma concentration hardly changes during general hemoconcentration in the cold.

Since cold exposure causes this increase in concentration of thrombogenic factors in the blood of both young and elderly people, it might be expected that it would increase mortality rates in both age groups. In practice, the increase in mortality rate is virtually confined to the elderly. Part of the explanation for this is that baseline fibrinogen levels are much higher in the elderly, so that even in the cold, fibrinogen levels in young adults do not reach the baseline level of the elderly (Fig. 2). A more important reason is that young people are protected from intra-arterial thrombosis by the fact that their arteries have a healthy endothelial lining. The inner surfaces of arteries of elderly people are commonly affected by atheroma and so are much more prone to thrombose.

The increase in respiratory infections and respiratory deaths in winter is due partly to respiratory infections spreading more readily in cold weather. People crowd together in poorly ventilated spaces when it is cold. Apart from that, breathing of cold air stimulates coughing and running of the nose, and this helps to spread respiratory viruses and bacteria. Cold stress also tends to suppress immune responses to infections. Respiratory infections also increase the plasma level of fibrinogen, and this contributes to the rise in arterial thrombosis in winter. (10)

[FIGURE 1 OMITTED]

The train of events leading to respiratory deaths in winter often starts with a cold or some other minor infection of the upper airways. This spreads to the bronchi and to the lungs. Secondary infection often follows and can lead to pneumonia. Anything that ameliorates a cold will therefore reduce more serious respiratory infections. The simplest treatment for a cold is the old remedy of inhaling steam for 30 minutes or so. This not only reduces the symptoms of a cold at once, but moderates the entire subsequent course of the illness. (11,12) The reasons for this are not fully known, but one is that a rise in temperature to around 113[degrees]F causes heat shock damage to the rhinoviruses that cause many colds. (13)

Epidemics of influenza used to kill hundreds of thousands of people every 2 or 3 years, often through secondary bacterial pneumonia. Since the 1970s, these epidemics have been much less common and less severe worldwide. The decline mostly occurred before the start of immunization against influenza. (14,15) The reasons for it are not fully known, but an important factor appears to be reduced opportunities for exchanges of genetic material between human, avian, and porcine strains of influenza virus. In the past, particularly in times of war and civil disturbance, there was frequent contact of people with ducks, chickens, and pigs in living spaces. Settled times allow better hygiene, which allows less contact between the different viruses. A potential new risk has appeared with the mass rearing of chickens, which allows epidemics of avian influenza to infect huge flocks of birds. Spread to human beings of these viruses and of the frequently lethal SARS virus has so far been limited by mass slaughter of infected birds. However, there is a risk that one of the new strains of influenza will develop, through contact with human influenza virus, the capacity for rapid transmission from person to person. If this happened, it could trigger a human epidemic on the scale of the influenza pandemic of 1918.

[FIGURE 2 OMITTED]

Surprisingly, few of the excess deaths in winter are due to the body simply cooling until vital organs such as the heart and brain cease to function. When hypothermia does occur it is usually a consequence of other illness and it is not a common cause of death in North America (16,17) and is a rare one in Britain. (18,19)

In the past, deficiency of vitamin C in the winter diet may have contributed to winter mortality rates, but winter death is now due to the effects of cold on people. Vitamin C does have a protective action against arterial thrombosis, but fruits and vegetables are now freely available even in winter. Statistical analysis shows that the winter mortality rate is closely associated with cold weather. Time series analysis on deseasonalized data, using multiple single regressions, (6) shows that cold spells are closely associated with sharp increases in mortality rates. The deaths continue for many days after a cold spell ends and account for all of the excess mortality rate in winter.

Heat-related death, like cold-related death, is often due to thermal stress exacerbating conditions that commonly cause deaths among elderly people. This is particularly true of temperate regions with moderate summer temperatures. However, in regions where people are exposed to air close to or above body temperature for long periods, a substantial amount of the heat-related deaths are due to simple hyperthermia, overheating of the body until the body proteins are denatured. (20) A wide range of psychoactive drugs, particularly those with anticholinergic or narcotic actions, increase the risk of this by impairing sweating and other responses to heat. Psychiatric patients are accordingly particularly vulnerable to heat stress.

In regions with cool summers, such as Britain, almost all heat-related deaths are due to other factors. Unlike cold-related deaths, infections play little part in these, but as with cold-related deaths, coronary and cerebral thromboses account for many heat-related deaths. These thromboses again result from hemoconcentration, which in this case has a simple cause, loss of salt and water in sweat. (21) Other heat-related deaths result from a range of other factors that are not well understood but which probably include strain on failing hearts unable to provide the additional blood flow to the skin needed to increase loss of heat from the body. One mitigating factor in relation to heat-related death in relatively cool countries such as Britain is that increased deaths during a few days of hot weather are followed by a lower than normal mortality rate. The likely reason is that many of those dying in the heat are already seriously ill and even without heat stress would have died within the next 2 or 3 weeks.

How Is Global Warming Changing Mortality in Practice?

Global warming has been under way for at least 25 years, and there is strong evidence that it is largely man-made and is continuing. (22) A large part of the warming is due to the burning of fossil fuels and consequent increase of carbon dioxide in the atmosphere. Since heat-related deaths are generally much fewer than cold-related deaths, the overall effect of global warming on health can be expected to be a beneficial one. Inevitably, though, when it was recognized in the 1990s that global warming was under way, attention shifted from the hazards of cold to those of hot weather.

The main concern at first was that diseases transmitted by insects, such as malaria, would spread to cooler regions of the world and would become a problem there. Closer examination showed that this was unlikely to happen to a serious extent. Malaria, for example, was once prevalent in most of Europe and even in Russia but had already been eliminated. The main reason was that modern farming methods and changes in human living conditions had reduced the number of the mosquitoes that spread the disease and had reduced their access to people. (23) From time to time, global warming, together with rapid air travel, is likely to cause new health problems from insect borne illnesses, such as the recent outbreak of West Nile Fever in New York. These should remain relatively easy to contain by measures such as spraying to kill mosquito larvae and preventing access of mosquitoes to infected patients.

A simple assessment of the immediate effect of rising temperature can be made on the assumption that particular degrees and patterns of heat or cold will continue to produce the same mortality rates as they did previously. Lack of daily statistics has prevented accurate assessment of this kind for some regions, but outside the tropics, it indicates that rises in temperature over the next few years would increase heat-related deaths less than they decrease cold-related deaths. For example, on this assumption, the rise in temperature of 3.6[degrees]F expected over the next 50 years would increase heat-related deaths in Britain by about 2,000, but reduce cold-related deaths by about 20,000. (5)

Of course, people dying as a result of heat will not be reassured by being told that fewer people will die next winter as a result of cold. It is important to minimize any increase in heat-related mortality rates, regardless of falls in cold-related mortality rates. Studies of populations living in widely different climates show that they have in fact adjusted to their own climates remarkably effectively over time. The Eurowinter study, which made active surveys of 8,000 people in 8 regions of Europe, and related mortality rates to daily temperature in each region, showed that people in cold regions such as the north of Finland had no more winter deaths than people in regions with much milder winters such as London and Athens. (1) Studies in Siberia provided even more striking examples. There was little excess winter death in the big industrial town of Yekaterinburg (24) in western Siberia. There was none at all in Yakutsk, in eastern Siberia, the coldest city in the world with temperatures averaging -33[degrees]F in winter. (25) The heat-related mortality rate showed evidence of corresponding adjustment. Such mortality rates were not significantly greater in the hot summers of Athens than in the cooler summers of the north of Finland. (3)

The surveys, covering a thousand homes in each region, showed how people had adjusted. Those in the colder regions not only kept their houses warmer than people in warm regions, even at the same level of outdoor temperature, they also dressed more warmly and were more likely to keep moving when outside. These comparisons were made for the same outdoor temperature and for people of the same ages. People in cold countries were simply much more effective in keeping warm. People in Yakutsk wore massive fur clothing outdoors, and on days of extreme cold reduced the time they spent outdoors. The ways in which people in hot regions adjusted to heat are less well documented, but people in European countries with hot summers have long-standing and well known strategies for this. The siesta, ceiling fans, and outside shutters that prevent sunlight from entering windows to cause greenhouse warming are obvious examples of these in southern Europe.

North American data showed a discrepancy from the European data with regard to winter as well as summer mortality, which appears to be due to better heating in winter and more air conditioning in summer. In North Carolina, the mortality rate in cold weather rose no more steeply with a fall in outside temperature than it did in the much colder European region of north Finland (1) (see Figure 3). It rose much more steeply in parts of Europe with winter climates similar to North Carolina, but with poorer defenses against cold, such as London or Athens. (3,4) Until recently, central heating was widely installed in Europe only in the coldest countries, whereas it was usual for most homes to be kept fully warm in even moderately cold weather in the United States.

In recent years, temperature and mortality data from several countries shows that cold-related deaths in each age group are falling in most countries. Much of that was due to rising climatic temperature and better home heating. The reduced frequency and severity of new epidemics of winter influenza has contributed, (26) but since 1976, campaigns for warmer housing and advice on clothing and exercise to keep warm outdoors can take most of the credit.

A surprising finding is that the heat-related mortality rate has stabilized or fallen, despite rising temperatures. Air conditioning has been a major factor in the United States. Heat-related deaths there are lower among people with air conditioning. (20,27,28) An extension of air conditioning was accompanied by the virtual disappearance of heat-related death in North Carolina, despite summers becoming hotter. (4) In 1971, such a mortality rate in North Carolina was similar to that in Athens. In North Carolina, summer temperatures then rose around 1.8[degrees]F, humidity increased and wind decreased, but heat-related death virtually disappeared. The spread of air conditioning in the South Atlantic region of the United States, which includes North Carolina, from 56 to 72% of homes from 1976 to 1997, provides an explanation. A recent study confirms falling heat-related mortality rates in many cities of the United States, (29) although the method used to estimate it makes quantitative comparison difficult.

[FIGURE 3 OMITTED]

Britain and the rest of northern Europe still have little air conditioning, and the heat-related mortality rate in London has not fallen. Nor has it risen, however, despite a 3.6[degrees]F rise in summer temperature since 1971. Such factors as more relaxed lifestyle, more informal clothing, and purchase of electric fans as prosperity increased have apparently countered heat stress there. (4)

What Can Be Done to Control Seasonal Mortality During Global Warming?

It would be easy to look at these facts and say that we need do nothing. Global warming is not likely to increase overall mortality rates over time. People make their own adjustments to hotter summers, and in time this will prevent much increase in summer mortality rates.

Despite this, much needs to be done. Sudden heat waves can be expected to produce record high temperatures every few years as the climate warms. These will expose populations to higher environmental temperatures than they have ever encountered before. The record heat is accordingly liable to cause high mortality rates for a few days among people who are not prepared for it. This happened in France in the summer of 2003, with around 15,000 excess deaths in 2 weeks. Record temperatures can also cause unacceptable working conditions. One obvious solution is to install widespread air conditioning more widely in Europe, as in the United States. Unfortunately, this solution has penalties in the long term. The most important is high energy consumption, which accelerates global warming, as it requires increased generation of electricity and this in turn involves burning more fossil fuel.

Alternative strategies for keeping cool in hot weather involve the design and management of buildings. One important element in this is high thermal insulation in their outer walls and high thermal mass internally, to provide a more even indoor temperature throughout the 24-hour cycle. It is particularly important to prevent sunlight coming in through windows, causing greenhouse heating. External slatted shutters are a traditional and effective way of doing this in southern Europe. Windows can be opened at night, and then closed after dawn for as long as the interior of a building remains cooler than the outside air. Cooking should be kept to a minimum in places where it can warm up living space and increase humidity. Crowding of people indoors will also increase temperature and humidity. Once the interior of a building does become uncomfortably hot, a combination of light clothing, air movement from a fan, open windows, and a sprinkling of water on the clothing can normally control heat stress. People should continue to eat regular meals with moderate salt content and to drink water in hot weather, even if they do not feel hungry or thirsty.

These measures will normally be effective even for people who are elderly or vulnerable because of illnesses such as diabetes, or who are taking drugs that suppress sweating. Sprinkling water on clothing can substitute for sweating and allow evaporative cooling even in these people. Evaporative cooling requires that the air should not be saturated; when heat stress develops, ventilation by air from an open window will provide the unsaturated moving air needed for evaporative cooling. Outside air is not saturated at hot times of the day during heat waves in temperate regions.

Anyone who becomes seriously overheated, with a mouth temperature around or above 104.9[degrees]F, needs to be cooled immediately. This should be started at once rather than waiting for help from the emergency services. If other measures are difficult to implement, immersion in a cool bath can used, but it must be remembered that very cold water will cause vasoconstriction. This can retard cooling, so that cool rather than cold water should be used. Immediate measures of that kind are far more effective than transporting a patient to hospital. A high proportion of those who died in France in the heat wave of 2003 appear to have done so in hospitals. The objective should be to keep people cool at home, and to cool them there. However, hospitals and other institutions containing numbers of elderly and other vulnerable people have a particular need for air conditioning, unless the design of the building can keep those within at safe temperatures with any level of outdoor temperature that can reasonably be expected. Finally, measures to control and treat tropical diseases need to be available if these diseases spread outside their present ranges.

Some of these measures require long-term planning, but warnings with advice on preventing and treating heat stress are important when heat waves are forecast. Broadcasting advice together with such forecasts can also allow elderly people and their neighbors to check, for example, that fans and water are available and that windows can be opened. Air conditioning will be increasingly needed in heat waves, but a possible strategy is to have it available in most regions but to use it only as a second line of defense when other methods fail or become excessively burdensome.

Important as such measures against heat stress are, we should not lose sight of the fact that even in climates as warm as southern Europe or North Carolina, cold weather causes more deaths than hot weather. The importance of warm housing in preventing winter death is well recognized, (30) but the large contribution made by outdoor cold stress to winter death (1,31) is generally not. Global warming will reduce this at first, but the improvement is not likely to continue without action to promote defenses against cold. People in regions with mild winters become careless about cold stress, protect themselves less effectively against cold, and generally have more winter deaths than people in colder regions. (1,24,25) Climatic warming therefore calls for action to control cold stress as well as heat stress. If this is taken, rising temperatures could reduce overall mortality rates.

Wider Implications of Global Warming

It might be supposed that the wider hazards presented by global warming would bring action to halt it, and so avoid the need to deal with its effects on people. In practice, this is unlikely to happen. The most serious long-term hazard of global warming is that melting of the ice caps and warming of the oceans may cause flooding. It has been estimated that the sea level could rise by 34 cm by the year 2100. (32) This would have serious consequences in some coastal cities. Further warming over several centuries could cause much larger rises, with massive flooding of heavily populated regions. It would also cause climate changes affecting the habitability of many parts of the world and might increase net mortality rates in some tropical countries.

However, the extent and rate of such effects is still very controversial, and stopping them by halting the burning of fossil fuels would carry a high cost to the standard of living of a growing world population. As a result, even acceptance of the Kyoto Protocol, which would only slightly reduce global warming, is still uncertain. Wind, wave, and solar power could only be a partial substitute for fossil fuels. Nuclear power could make a larger contribution, but its risks would have to be accepted. In the medium term, the likelihood is that despite some measures of this kind, substantial global warming will continue and will require action to deal with its effects on health. Fortunately, effective action is available.

All are lunatics, but he who can analyze his delusion is called a
philosopher.
--Ambrose Bierce

Accepted June 9, 2004.

References

1. Eurowinter Group, Keatinge WR, Donaldson GC, Bucher K, et al. Cold exposure and winter mortality from ischaemic heart disease, cerebro-vascular disease, respiratory disease, and all causes in warm and cold regions of Europe. Lancet 1997;349:1341-1346.

2. Shen T, Howe HL, Alo C, Moolenaar RL. Towards a broader definition of heat related death: comparison of mortality estimates from total death differentials during the July 1995 heat wave in Chicago, Illinois. Am J Forensic Med Pathol 1998;19:113-118.

3. Keatinge WR, Donaldson GC, Cordioli E, et al. Heat related mortality in warm and cold regions of Europe, observational study. Br Med J 2000;321:670-673.

4. Donaldson GC, Keatinge WR, Nayha S. Changes in summer temperature and heat related mortality since 1971 in North Carolina, South Finland and Southeast England. Environ Res 2003;91:1-7.

5. Donaldson GC, Kovats RS, Keatinge WR, McMichael RJ. Heat- and cold-related mortality and morbidity and climate change. Chapter 4.1 in Health effects of climate change in the UK, p70-80 in: Report to the Department of Health (UK), 2001, Ed Maynard RL.

6. Donaldson GC, Keatinge WR. Early increases in ischaemic heart disease mortality dissociated from, and later changes associated with, respiratory mortality, after cold weather in south east England. J Epidemiol Comm Health 1997;51:643-648.

7. Keatinge WR, Coleshaw SRK, Cotter F, et al. Increases in platelet and red cell counts, blood viscosity, and arterial pressure during mild surface cooling: factors in mortality from coronary and cerebral thrombosis in winter. Br Med J 1984;289:1405-1408.

8. Neild PJ, Syndercombe-Court D, Keatinge WR, et al. Role of cold diuresis and natriuresis in the haemo-concentration caused by exposure to cold in man, in Milton AS (ed): Thermal Physiology. Aberdeen, International Union of Physiological Sciences Thermal Physiology Commission, 1993.

9. Neild PJ, Syndercombe-Court D, et al. Cold-induced increases in erythrocyte count, plasma cholesterol and plasma fibrinogen of elderly people without comparable rise in protein C or factor X. Clin Sci 1994;86:43-48.

10. Woodhouse PR, Khaw K-T, Plummer M, et al. Seasonal variations of plasma fibrinogen and factor VII in the elderly: winter infections and seath from cardiovascular disease. Lancet 1994;343:435-439.

11. Ophir D, Elad Y. Effects of steam inhalation on nasal patency and nasal symptoms in patients with the common cold. Am J Otolaryngol 1987;3:149-153.

12. Tyrrell D, Barrow I, Arthur J. Local hypothermia benefits natural and experimental common colds. Br Med J 1989;298:1280-1283.

13. Conti C, deMarco A, Mastromarino P, et al. Antiviral effect of hyperthermic treatment in rhinovirus infection. Antimicrob Agents Chemother 1999;43:822-829.

14. Keatinge WR, Coleshaw SRK, Holmes J. Changes in seasonal mortality with improvement in home heating in England and Wales 1964-1984. Int J Biometeorology 1989;33:71-76.

15. Donaldson GC, Keatinge WR. Excess winter mortality; influenza or cold stress? Observational study. Br Med J 2002;324:89-90.

16. Danzl DF, Pozos RS, Auerbach PS, et al. Multicenter hypothermia survey. Ann Emerg Med 1987;16:1042-1055.

17. Thomas DR. Accidental hypothermia in the sunbelt. J Gen Intern Med 1988;3:552-554.

18. Coleshaw SRK, Easton JC, Keatinge WR, et al. Hypothermia in emergency admissions in cold weather. Clin Sci 1986;70(Suppl 13):93P-94P.

19. Woodhouse P, Coleshaw SRK, Keatinge WR. Factors associated with hypothermia in patients admitted to a group of inner city hospitals. Lancet 1989;2:1201-1203.

20. Rogot E, Sorlie PD, Backlund E. Air-conditioning and mortality in hot weather. Am J Epidemiol 1992;136:106-116.

21. Keatinge WR, Coleshaw SRK, Easton JC, et al. Increased platelet and red cell counts, blood viscosity and plasma cholesterol level during heat stress, and mortality from coronary and cerebral thromboses. Am J Med 1986;81:795-800.

22. Hulme M, Jenkins G, Brooks D, et al. What is happening to global climate and why? in Maynard RL (ed): Health Effects of Climate Change in the UK. Report to the Department of Health (UK), 2001. pp 18-49.

23. Rogers DJ, Randolph S, Lindsay S, Thomas C. Vector-borne diseases and climate change, in Maynard RL (ed): Health Effects of Climate Change in the UK. Report to the Department of Health (UK), 2001. Pp 85-98.

24. Donaldson GC, Tchernavskii VF, Ermakov SP, et al. Winter mortality and cold stress in Yekaterinburg, Russia: interview survey. Br Med J 1998;316:514-518.

25. Donaldson GC, Ermakov SP, Komarov YM, et al. Cold-related mortalities and protection against cold in Yakutsk, eastern Siberia: observation and interview study. Br Med J 1998;317:978-982.

26. Donaldson GC, Keatinge WR. Mortality related to cold weather in southeast England, 1979-1994. Br Med J 1997;315:1055-1056.

27. Taylor AJ, McGwin G. Temperature-related deaths in Alabama. South Med J 2000;93:787-792.

28. Greenberg JH, Bromberg J, Reed CM, et al. The epidemiology of heat-related deaths, Texas--1950, 1970-79, and 1980. Am J Public Health 1983;73:805-807.

29. Davis RE, Knappenberger PC, Michaels PJ, Novicoff WM. Changing heat-related mortality in the United States. Environ Health Perspect 2003;111:1712-1718.

30. Healy JD. Excess winter mortality in Europe: a cross country analysis identifying key risk factors. J Epidemiol Comm Health 2003;57:784-789.

31. Keatinge WR. Seasonal mortality among elderly people with unrestricted home heating. Br Med J 1986;293:732-733.

32. Titus JG, Narayanan V. The risk of sea level rise: a Delphic Monte Carlo analysis in which twenty researchers specify subjective probability distributions for model coefficients within their respective areas of expertise. Climatic Change 1996;33:151-212.

RELATED ARTICLE: Key Points

* Heat stress causes deaths among elderly people in summer.

* Cold causes larger numbers of deaths in winter, from arterial thrombosis and respiratory disease, and efforts to improve protection against cold should not be relaxed during global warming.

* Populations accustomed to heat or cold adjust to them, but record high temperatures for a locality during global warming have caused high mortality rates.

* Measures to prevent heat stress in vulnerable people need to be in place before unprecedented heat waves occur.

W. R. Keatinge, MA, MB, BCHIR, PHD, FRCP and G. C. Donaldson, BA, PHD

From Bart's and the London, Queen Mary's School of Medicine and Dentistry, University of London, London, United Kingdom.

Reprint requests to Prof. W.R. Keatinge, Medical Sciences Building, Queen Mary College, Mile End Road, London E1 4NS, UK. Email: [e-mail]