The incidence of most, if not all, infectious diseases depends on climate. For any given infection, however, climate change is but one of many factors that determine disease epidemiology, and often it is not the most influential factor. Even in instances in which climate change creates conditions favorable to the spread of infections, diseases may be kept in check through interventions such as vector control or antibiotic treatment.
Detecting climate-change influence on an emerging human disease can be challenging. Research with animal pathogens, which in most instances are less well monitored and intervened upon than that with their human counterparts, has suggested how climate change may influence disease spread. For example, the life cycle of nematode parasites of caribou and musk oxen shortens as temperatures rise. As the Arctic has warmed, higher nematode burdens and consequently higher rates of morbidity and mortality have been observed. Other examples from animals, such as the spread of the protozoan parasite Perkinsus marinus in oysters, demonstrate how warming can enable range expansion of pathogens previously held in check by colder temperatures.
As these and other examples from studies of animals make clear, the influence of climate change on infectious diseases can be pronounced. The following sections deal with the infectious diseases for which research has explored the influence of climate change.
Because insects are cold-blooded, ambient temperature dictates their geographic distribution. With increases in temperatures (in particular, nighttime minimum temperatures), insects are freed to move poleward and up mountainsides. At the same time, as new areas become climatically suitable, current mosquito habitats may become unsuitable as a result of heat extremes.
In addition, insects tend to be sensitive to water availability. Mosquitoes that transmit malaria, dengue, and other infections may breed in pools of water created by heavy downpours. As has been observed in the Amazon, breeding pools can also appear during periods of drought when rivers recede and leave behind stagnant pools of water for Anopheles mosquitoes. These circumstances have raised interest in the potentially favorable impact of water-cycle intensification on the spread of mosquito-borne disease.
Higher temperatures promote higher mosquito-biting rates, shorter parasite reproductive cycles, and the potential for the survival of mosquito vectors of Plasmodium infection in locations previously too cold to sustain them. Modeling experiments have identified highland areas of East Africa and South America as perhaps most vulnerable to increased malarial incidence as a result of rising temperatures. In addition, an analysis of interannual malaria in Ecuador and Colombia has documented a greater incidence of malaria at higher altitudes in warmer years. Highland populations may be more vulnerable to malaria epidemics because they lack immunity.
Although rising temperature has the potential to expand the viable range of disease, malaria incidence is not associated with temperature in a strictly linear fashion. While mosquitoes and parasites may adapt to a warming climate, the present optimal temperature for malaria transmission is ~25°C, with a range of transmission temperatures between 16°C and 34°C. Rising temperatures also can have differential effects on parasite development during external incubation and on the mosquitoes’ gonotrophic cycle. Asynchrony between these two temperature-sensitive processes has been shown to decrease the vectorial capacity of mosquitoes.1
The abundance of Anopheles mosquitoes is strongly correlated with the availability of surface-water pools for mosquito breeding, and biting rates have been linked to soil moisture (a surrogate for breeding pools). Research in the East African highlands has documented that increased variance in rainfall over time has strengthened the association between precipitation and disease incidence. These disease-promoting effects of precipitation may be countered by the potential for extreme rainfall to flush mosquito larvae from breeding sites.
Climate models have begun to deliver output on regional scales, permitting projections of climate-suitable regions to assist national and local health authorities. Climate models speak to the temperature and precipitation ranges necessary for malaria transmission but do not—and cannot—account for the capacity of malaria control programs to halt the spread of disease. The global reduction in malaria distribution over the past century makes it clear that, even with climate change, malaria occurs in far fewer places today because of public health interventions.
Despite intensive efforts, malaria remains the single greatest vector-borne disease cause of morbidity and death in the world. Particularly in regions that are most affected by malaria and where the public health infrastructure is inadequate to contain it, climate modeling may provide a useful tool in determining where the disease may spread. Modeling studies in sub-Saharan Africa have suggested that, although East African nations may encompass regions that will become more climatically suitable for malaria over this century, West African nations may not. By 2100, temperatures in West Africa may largely exceed those optimal for malaria transmission, and the climate may become drier; in contrast, higher temperatures and changes in precipitation may allow malaria to move up the mountainsides of East African countries. Climate change may create conditions favorable to malaria in subtropical and temperate regions of the Americas, Europe, and Asia as well.
Like malaria epidemics, dengue fever epidemics depend on temperature (Fig. 120-5). Higher temperatures increase the rate of larval development and accelerate the emergence of adult Aedes mosquitoes. The daily temperature range may also influence dengue virus transmission, with a smaller range corresponding to a higher transmission potential. Temperatures <15°C or >36°C substantially reduce mosquito feeding. In a Rhesus model of dengue, viral replication can occur in as little as 7 days with temperatures of >32–35°C; at 30°C, replication takes ≥12 days; and replication does not reliably occur at 26°C. Research on dengue in New Caledonia has shown peak transmission at ~32°C, reflecting combined effects of a shorter extrinsic incubation period, a higher feeding frequency, and more rapid development of mosquitoes. Along with temperature, peak relative humidity is a strong predictor of dengue outbreaks.
Effects of temperature on variables associated with dengue transmission. Shown are the number of days required for development of immature Aedes aegypti mosquitoes to adults, the length of the dengue virus type 2 extrinsic incubation period (EIP), the percentage of A. aegypti mosquitoes that complete a blood meal within 30 min after a blood source is made available, and the percentage of hatched A. aegypti larvae surviving to adulthood. (Reproduced from CW Morin et al: Climate and dengue transmission: Evidence and implications. Environ Health Perspect 121:1264, 2013.)
The association between dengue epidemics and precipitation is less consistent in the peer-reviewed literature, possibly because of the mosquito vector’s greater reliance on domestic breeding sites than on natural pools of water. For instance, in some studies, increased access to a piped water supply has been linked to dengue epidemics, presumably because of associated increased domestic water storage. Nonetheless, several studies have established rainfall as a predictor of the seasonal timing of dengue epidemics.
The current global distribution of dengue largely overlaps the geographic spread of Aedes mosquitoes (Fig. 120-6). The presence of Aedes without dengue endemicity in large regions of North and South America and Africa illustrates the relevance of variables other than climate to disease incidence. Nevertheless, coupled climatic–epidemiologic modeling suggests dramatic shifts in the relative vectorial capacity for dengue by the end of this century should little or no mitigation of greenhouse gas emissions occur (Fig. 120-7). Given the joint effects of climate change and population growth, the number of people exposed to A. aegypti globally may nearly double by 2100 from roughly 4 billion to 8 billion or more.
Distribution of Aedes aegypti mosquitoes (turquoise) and dengue fever epidemics (red). (Map produced by the Agricultural Research Service of the U.S. Department of Agriculture.)
Trend of annually averaged global dengue epidemic potential (rVc). Differences in rVc are based on 30-year averages of temperature and daily temperature range. A. Differences between 1980–2009 and 1901–1930. B. Differences between 2070–2099 and 1980–2009. The mean value of rVc was averaged from five global climate models under RCP8.5, a scenario of high greenhouse-gas emission. The color bar describes the values of the rVc. (From J Liu-Helmersson et al: Vectorial capacity of Aedes aegypti: Effects of temperature and implications for global dengue epidemic potential. PLoS ONE 9:e89783, 2014 [doi:10.1371/journal.pone.0089783].)
Other Arbovirus Infections
Climate change may favor increased geographic spread of other arboviral diseases, including Zika virus disease, chikungunya virus disease, West Nile virus disease, and eastern equine encephalitis. Zika virus moved to the Western Hemisphere from French Polynesia around 2013 and rapidly spread in Brazil in 2016. Although air travel was essential for the delivery of the virus to the Americas, the available evidence suggests that the 2015 El Niño event provided an optimal climate for the infection to take root and spread. A. aegypti is the primary vector for Zika virus. Chikungunya virus disease emerged in Italy in 2007, having previously been mostly a disease of African nations. Climate models predict that, should competent vectors be present, conditions will be suitable for chikungunya virus to gain a foothold in Western Europe, especially France, in the first half of the twenty-first century. In North America, areas favorable to West Nile virus outbreaks are expected to shift northward in this century. Current hotspots in North America are the California Central Valley, southwestern Arizona, southern Texas, and Louisiana, which have both compatible climates and avian reservoirs for the disease. By mid-century, the upper Midwest and New England will be more climatically suited to West Nile virus; by the end of the century, the area of risk may shift even further north to southern Canada. Whether the disease will ultimately move northward will depend on reservoir availability and mosquito control programs, among other factors.
In the past few decades, Ixodes scapularis, the primary tick vector for Lyme disease as well as for anaplasmosis and babesiosis in New England, has become established in Canada because of warming temperatures. With climate change, the range of the tick is expected to expand further (Fig. 120-8).
Present and projected probability of establishment of Ixodes scapularis. (From U.S. National Climate Assessment 2014, adapted from JS Brownstein et al: Effect of climate change on Lyme disease risk in North America. Ecohealth 2:38, 2005.)
Lyme disease, caused by the spirochete Borrelia burgdorferi, is the most commonly reported vector-borne disease in North America, with ~30,000 cases per year. The model used in Fig. 120-8 showed 95% accuracy in predicting current I. scapularis distribution and suggests substantial expansion of tick habitat and consequently of populations at risk for the diseases this tick transmits, particularly in Quebec, Iowa, and Arkansas, by 2080. Of note, some areas on the Gulf Coast may become less suitable for ticks by the end of the century.
Outbreaks of waterborne disease are associated with heavy rainfall events. A review of 548 waterborne disease outbreaks in the United States found that 51% were preceded by precipitation levels above the 90th percentile. Since 1900, most regions of the United States except the Southwest and Hawaii have experienced an increase in heavy downpours (Fig. 120-9), with the greatest intensification of the water cycle in New England and Alaska. Climate models suggest that by 2100 daily heavy-precipitation events, which are defined as a cumulative daily amount that now occurs once every 20 years, will increase nationwide (Fig. 120-10). This scenario may be from two to as much as five times more likely, depending on the extent of greenhouse gas emission reductions achieved early in the twenty-first century.
Percentage changes in the annual amount of precipitation falling in very heavy events, defined as the heaviest 1% of all daily events from 1901 to 2012 for each region. Changes are relative to a 1901–1960 average for all regions except values for Alaska and Hawaii, which are relative to the 1951–1980 average. (From U.S. National Climate Assessment 2014, NOAA National Climate Data Center/Cooperative Institute for Climate and Satellites, North Carolina.)
Increased frequency of extreme daily precipitation events (defined as a daily amount that now occurs once in 20 years) by the latter part of the twenty-first century (2081–2100) compared to the frequency in the latter part of the twentieth century (1981–2000). A representative concentration pathway (RCP) describes a plausible climate future based on a net radiative forcing (e.g., 2.6 or 8.5) in 2100. (From U.S. National Climate Assessment 2014, NOAA National Climate Data Center/Cooperative Institute for Climate and Satellites, North Carolina.)
Most disease outbreaks after heavy precipitation occur through contamination of drinking-water supplies. While outbreaks related to surface-water contamination generally occur within a month of the precipitation event, disease outbreaks from groundwater contamination tend to occur ≥2 months later. According to a review of published reports of waterborne disease outbreaks, Vibrio and Leptospira species are the pathogens most commonly involved in the wake of heavy precipitation.
Roughly 40 million people in the United States and millions more around the world rely on combined sewer systems in which storm water and sanitary wastewater are conveyed in the same pipe to treatment facilities. These systems were designed on the basis of the nineteenth-century climate, in which heavy downpours were less frequent than they are today. The frequency of combined sewer overflows resulting in untreated sewage discharge, usually into freshwater bodies, has been increasing in cities worldwide. Overflows are associated with discharges of heavy metals and other chemical pollutants as well as a variety of pathogens. Outbreaks of hepatitis A, Escherichia coli O157:H7 infection, and cryptosporidial disease have been associated with sewer overflows in the United States.
Rising Temperatures and Vibrio Species
Warmer temperatures favor proliferation of Vibrio species and disease outbreaks, as has been demonstrated in countries surrounding the Baltic Sea, Chile, Israel, northwestern Spain, and the U.S. Pacific Northwest. Around the Baltic Sea, outbreaks of Vibrio infection may be particularly likely because of faster warming near the poles and the sea’s relatively low salt content. In 2004, a Vibrio parahaemolyticus outbreak arising from consumption of Alaskan oysters occurred. This pathogen was unknown in Alaskan oysters prior to this event and extended the known geographic range of the disease 1000 km northward.
In the past, El Niño events were used as a model to investigate the potential for extreme weather–related infectious disease epidemics occurring in association with climate change. Recent evidence indicates that climate change itself may be strengthening El Niño events. These events tend to promote epidemic infections in certain regions (Fig. 120-11).
Characteristic patterns of disease outbreaks associated with El Niño events, determined on the basis of 2006–2007 conditions. (From A Anyamba et al: Developing global climate anomalies suggest potential disease risks for 2006–2007. Int J Health Geogr 5:60, 2006.)
Associations of El Niño with outbreaks of Rift Valley fever in eastern and southern Africa have been known since the 1950s. El Niño favors wet conditions suitable for the insect vectors of the disease in these regions. Given the strong association between El Niño conditions and disease incidence, models have successfully predicted Rift Valley fever epidemics in humans and animals. In the 2006–2007 El Niño season, for example, outbreaks of Rift Valley fever were accurately predicted 2–6 weeks prior to epidemics in Somalia, Kenya, and Tanzania.
El Niño has had inconsistent associations with malaria incidence in African countries. Some of the strongest associations between El Niño and malaria have been identified in South Africa and Swaziland, where available data on incidence are relatively robust; however, even in these instances, the observed increased risk did not reach statistical significance. A stronger link to El Niño has been found in several studies done in South America. Research on malaria incidence in Colombia between 1960 and 2006 found that a 1°C temperature rise contributed to a 20% increase in incidence.
El Niño years are often associated with an increased incidence of dengue. Research on dengue outbreaks in Thailand from 1996 to 2005 revealed that 15–22% of the variance in monthly dengue disease incidence was attributable to El Niño. In South America, data on dengue outbreaks between 1995 and 2010 showed an increased incidence during the El Niño events of 1997–1998 and 2006–2007.