Infectious Diseases, Weather, and Climate

Over two and a half thousand years ago, Hippocrates and his followers used the term “epidemics” to describe diseases that propagated through populations in a seasonal fashion [1, 2]. Another Hippocratic treatise, “Airs, Waters, Places,” described the influence of the environment and seasons on constitution and instructed physicians to observe the health of a community with regard to sun exposure, soil, elevation, climate, and geography [1–3]. During the 16th through the 18th centuries, interest in the effects of climate and health was facilitated by the ability to measure environmental conditions with new instruments. For example, in the United States, both Thomas Jefferson and Noah Webster collected information on weather and disease [4]. The formation of what became the US National Weather Bureau resulted, in part, from an interest in the connection between weather and health [5]. In the mid- to late 19th and through much of the 20th century, interest in the role of seasons and climate shifted to elucidating host- and pathogen-related risk factors for infectious diseases. This change coincided with the loss of interest in the erroneous miasmic theory, the advent of germ theory, and the expanding field of microbiology. Epidemiological investigations also helped deemphasize the importance of climate with regard to disease, and John Snow’s work on the cholera epidemic in the 1850s provides a specific example [1]. Concern about climate change from the late 20th century to the present renewed interest in understanding the effects of the environment, climate, and weather on the incidence of infectious diseases and other health-related conditions [6]. Climate and weather differ in terms of scale: weather describes atmospheric conditions in the short-term (as in minutes, hours, days, weeks, months, or seasons), whereas climate, which is averaged over space and time, describes those in the longer term (years, decades, millennia, etc.). Most investigations regarding changes in climate and the risk of infectious diseases involve environmental changes that facilitate the spread of disease vectors into new regions, placing new populations at risk. Weather changes on time scales close to the incubation period for most infectious diseases. Accordingly, changes in weather provide opportunities for “natural experiments,” which can demonstrate the possible effects of atmospheric variables on infectious diseases. The effects of weather may be especially important to study because even slight increases in global temperatures may lead to extreme events, such as hurricanes: warming sea-surface temperatures (SSTs) can increase hurricane intensity, and warmer temperatures encourage evaporation and therefore precipitation in certain latitudes [7]. Weather conditions, such as rainfall, flooding, humidity, and heat waves have well-documented effects on infectious diseases. Prolonged periods of heavy rain increase the opportunities for vector-borne diseases to spread [8]. Longer seasons of mild temperatures may increase the transmission likelihood of vector-borne diseases. In warmer temperatures, vectors become infectious more quickly and can transmit virus earlier in their lives [9]. In addition, the temperature must be “just right”; if too cold, the virus takes too long to replicate, and if too hot, the vectors’ lifespans are decreased [10]. Standing water after excessive rains facilitates growth of mosquito populations [11]. The effects of heavy rainfall in areas with heavy clays, hard-packed soil, impermeable igneous or metamorphic rock close to the surface, or high water tables, are especially likely to provide the geological conditions necessary to collect surface standing water. Interestingly, droughts can also provide new breading grounds for mosquitoes because desiccated riverbeds offer an impermeable surface to trap pooling water. The Zika Pan-American virus spread in Brazil during a period of prolonged drought and warm temperatures, possibly caused by the 2015 El Niño event [12]. In addition, coccidioidomycosis outbreaks are associated with “grow and blow” conditions, in which wet weather is favorable to growth, and subsequent dry weather promotes transmission [13, 14]. Dramatic changes in levels of precipitation, even in normally dry areas, can increase rodent populations and thus increase human–rodent interactions, as in the Hanta virus outbreak in the Southwest United States in 1993 [15]. Excessive rainfall can cause sewage system outflow, leading to the spread of infectious diseases, both viral (e.g., norovirus, rotavirus) and bacterial (e.g., Vibrio spp. and Leptospira spp.) [8]. The outbreak of leptospirosis in Puerto Rico after hurricane Maria is yet another example of how flooding, when associated with extreme weather events, can increase the spread of infectious diseases [16]. Changing temperatures can also influence the risk of spreading infectious diseases. Warming SSTs may directly increase populations of V. cholera, a disease with a long association with seaside communities [17]. Longer periods of warm weather can increase the population of new insect vectors. During warmer periods of weather, tick populations may migrate to higher latitudes and/or altitudes, expanding the geographical distribution of tick-borne infections [18, 19]. In addition, even shorter periods of warmer weather appear to increase a range of infections not historically associated with environmental risk factors. The incidence of cellulitis is seasonal [20], and a substantial portion of this seasonality seems to be related to warmer temperatures [21]. Similarly, surgical-site infections are also seasonal and associated with warmer weather [22]. Urinary-tract infections demonstrate a seasonal increase in warmer weather [23]. Some gram-negative infections also exhibit a seasonal pattern [24]. Thus, warmer temperatures may increase a broad range of infections not typically associated with weather or climate. Although reports of different weather-related risk factors for infectious diseases exist, further work is needed in this area. Also, in order to understand how weather-related risk factors translate to population effects as the climate changes, we need to understand if and how populations can adapt. At least three major barriers obstruct efforts to improve understanding of how weather and ultimately changes in climate affect the incidence and spread of infectious diseases. Important privacy protections often require the aggregation of patient-related health data over larger geographic regions (e.g., state or census region) and longer periods of time (e.g., quarterly). Using less granular health-outcome data makes it difficult to investigate the effects of differing weather patterns on the incidence of an infectious disease. Quarterly data by state, for example, are not sufficient to make statistical inferences regarding the effects of weather on the risk for disease incidence: weather changes too quickly and affects local regions diversely. Furthermore, although it may be possible to obtain time-stamped and geolocated cases for specific diseases (especially those that are reportable), one needs to know not only about the environmental conditions experienced by cases but also conditions experienced by the controls: the patients who did not contract the disease. Thus, case-related data typically collected for public-health reporting may not be sufficient to be used to determine the role of weather on the incidence of a specific infectious disease. Finally, investigations including large geographic regions with variable weather conditions are most useful. However, the policies and procedures for collecting and reporting health data often vary dramatically from country-to-country and health system-to-health system. Ideally, researchers in different countries could collaborate and easily share data. If granular disease data could be collected, while still protecting the privacy of individuals, the pace of weather-and-health-related research would accelerate. The methods required to understand both the auto-correlated and spatially correlated data used for weather/disease-related projects are more commonly found outside the field of infectious diseases. Thus, there is a critical need for researchers within the infectious disease community to collaborate directly with investigators from statistics, engineering, computer science, and climate science. The necessity for such collaborations will increase with the scale and complexity of the data. Collaborations will be needed not only for analysis but also for efforts to assemble, organize, and fuse data from multiple sources. Many important health problems that require resources, and recent extreme weather events and their potential impact on the spread of infectious diseases highlight the need for funding of environmental studies related to infectious diseases. Although it is important to characterize disease risk factors, solutions are also necessary. Future work should focus on discovering the mechanisms by which weather affects infectious diseases, in order to help anticipate and even predict outbreaks, engineer solutions, and focus limited resources. In some cases, the mechanisms behind the links between weather and infectious diseases are clear (e.g., flooding increases mosquito breeding grounds). However, it is not obvious, for example, why warmer weather would increase the risk for skin and soft-tissue infections. Knowledge of how weather affects the incidence of infectious diseases may ultimately help increase the understanding of host and pathogen interactions. In addition, an appreciation of the role of weather may illuminate the effect of climate change on human health and ultimately convince society to invest the necessary resources to respond to changes in infectious-disease emergence and allow us to adapt. Potential conflicts of interest. The authors report no external funding and no conflicts of interest relating to this manuscript. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.