Potential Influence of Climate Change on Vector-borne and Zoonotic Diseases: A Review and Proposed Research Plan

James N. Mills; Kenneth L. Gage; Ali S. Khan


Environ Health Perspect. 2010;118(11):1507-1514. 

In This Article

What Kinds of Studies will be Useful?

The CDC policy on climate change and public health (CDC 2010) sets forth 11 priority public health actions that address research, prediction, training, communication, preparedness, and prevention to prepare the nation and the world for confronting potential health problems associated with climate change. Using these general priorities, we propose six specific goals that we believe would lead to a better understanding of the interaction between climate and VBZDs and an improved capability for predicting how climate change will affect the incidence and distribution of VBZDs.

Pathogen Identification and Characterization

Our knowledge of tropical biota and the pathogens associated with them is inadequate [see Supplemental Material (doi:10.1289/ehp.0901389)]. Calls for the study of worldwide (and especially tropical) biodiversity have come from a variety of sources [e.g., the United Nations, the U.S. National Science Foundation (NSF), and the Ecological Society of America]. Understanding biodiversity of plants and animals has scientific, economic, and esthetic appeal and has attracted much support (e.g., NSF's Planetary Biodiversity Inventories program; NSF 2003). Understanding the diversity of tropical pathogens has obvious practical importance yet has received little attention. Although a complete survey of potential pathogens in nature is an overwhelming task, a useful beginning would be to select the most likely potential hosts and vectors (e.g., bats and rodents, ticks and mosquitoes) and catalog those groups of pathogens with known propensities for causing disease in humans (e.g., hemorrhagic fever viruses, arboviruses, rickettsiae, and some bacteria).

Establish Baseline Data on the Geographic and Habitat Distribution of Recognized Zoonotic and Vector-borne Pathogens and their Hosts and Vectors

Accurate information on the current distribution of these agents and their hosts and vectors defines current potential disease-endemic areas and helps to quantify relative risk among habitat types. These data also are prerequisite to documenting spatial changes in the distribution of pathogens associated with climate change. Because pathogen distributions do not always coincide with host and vector distributions (Mills and Childs 1998), studies cannot be limited to documenting host and vector presence or absence but must include sampling hosts and vectors for pathogen presence. Finally, because distinct pathogens may be associated with genetically distinct populations of hosts or vectors (e.g., at the subspecies level) (Mills and Childs 1998), accompanying taxonomic studies will be required. Collection of accurate baseline data is the first step in the establishment of long-term monitoring programs.

Establish Longitudinal Monitoring Programs

A major impediment to accurate prediction has been the lack of long-term monitoring programs (Harvell et al. 2002; Jannett et al. 2007; Pascual and Bouma 2009; Sagarin 2002). These studies would track changes in zoonotic disease risk to humans by following changes in population dynamics and infection prevalence of a few zoonotic pathogens in their host and vector populations. These studies, whose length should be measured in decades, would monitor changes in environmental variables using field-based and remote-sensing platforms and identify factors associated with changes in host, vector, and pathogen dynamics at selected sites. Long-term sampling should also monitor changes in the distribution of hosts and vector populations and their distribution across latitudinal, elevational, and environmental gradients. Although only a few agents, hosts, and vectors can be addressed this way, these studies might be applied to other pathogens, even those yet to be discovered, that have the same or similar host or vector associations.

A few long-term databases describing temporal dynamics of animal populations or disease incidence already exist (e.g., data from long-term ecological research sites and records of reportable diseases), and a few more are proposed (e.g., NSF's National Ecological Observatory Network), but such ecological data were rarely collected for the purposes of infectious disease studies. Conversely, surveillance records for human diseases rarely contain significant amounts of useful ecological data. Nevertheless, some of these data sets might be paired with long-term weather data to provide insight into climate effects (Pascual et al. 2002; Stenseth et al. 2006).

Track data on the Geographic Distribution, Severity, and Frequency of Outbreaks of Wildlife Diseases and VBZDs in Humans

This is another type of longitudinal monitoring program, but the list of pathogens and hosts would be broader and the data collected more restricted (e.g., geographic coordinates, dates, number of cases). These data can be used to track geographic and temporal trends in VBZD incidence, identify vulnerable populations, and test predictions based on forecasting models. Several existing databases provide useful formats (e.g., U.S. Geological Service's Global Wildlife Disease News Map; National Biological Information Infrastructure 2010), the CDC's ArboNet, and the Global Early Warning System for Major Animal Diseases including Zoonoses [World Health Organization (WHO) 2010]. Integration of such formats with the monitoring, data gathering, and data dissemination capacity of CDC and WHO networks would provide a powerful tool, making geographic and incidence data quickly available to the broad community of public health professionals, researchers, and modelers.

Conduct Experimental Laboratory and Field Studies of Effects of Climate change on Hosts and Vectors and their Abilities to Maintain and Transmit Pathogens

Laboratory studies and manipulative field studies (Post et al. 2008) can be used to test mechanisms of climate change effects on hosts, vectors, and pathogens or to provide hypotheses for field testing. Because specific environmental factors cannot be isolated and controlled in the field, laboratory studies will be the best way to test the effects of specific changes in temperature, humidity, or physiological stress on host, vector, or pathogen populations. For example, stress-related immunosuppression has been suggested as an explanation for increased rates of transmission and for viral recrudescence including increases in pathogen load in hantavirus hosts observed in field studies (Botten et al. 2003; Kuenzi et al. 2005). Laboratory experiments provided preliminary evidence in support of this hypothesis in the case of Seoul hantavirus infection in Norway rats (Rattus norvegicus) (Klein et al. 2004). Additional controlled laboratory experiments will be necessary to confirm this hypothesis for other VBZD pathogens and to help quantify the environmental and social factors (temperature, crowding, aggression, breeding, etc.) associated with transmission and recrudescence.

Neither field nor laboratory studies alone can provide complete answers to these questions. Physiological patterns observed in the laboratory must be tested under natural field conditions and the physiological mechanisms for patterns observed in the field must be tested in the laboratory.

Use Data from Laboratory and Field Studies, Epidemiological Studies, and Remote Sensing to Develop Predictive Models of changes in Zoonotic Disease Risk and the Projected Distribution and Abundance of Major Hosts and Vectors

The utility of such predictive models has been proven. Long-term direct monitoring of host population density and prevalence of hantavirus infection in North American deer mice provided early warning of elevated risk for HPS in the southwestern United States (CDC 1998, 1999; Yates et al. 2002). Models have used rainfall and temperature data to identify high-risk areas for plague and HPS in the same geographic area (Eisen et al. 2007; Enscore et al. 2002; Parmenter et al. 1999). Remote sensors have demonstrated climate-associated changes in vegetation or other environmental parameters that foretold increased risk for Rift Valley fever (Anyamba et al. 2001) or HPS (Glass et al. 2002). Ecological niche modeling has been used to provide evidence for climate change-mediated range shifts for plague and tularemia and to provide predictions of continued future poleward expansion (Nakazawa et al. 2007). Ecological niche models also used climatic data and other environmental datasets to predict the geographic distribution of the hosts and vectors for Chagas disease and the filovirus hemorrhagic fevers (Peterson et al. 2002, 2004). Another model predicted the northward spread of the Lyme disease and anaplasmosis vector I. scapularis under proposed climate change scenarios (Brownstein et al. 2005). Recent surveillance has confirmed that expansion (Ogden et al. 2010). Continued data collection from long-term field studies combined with laboratory data providing measured physiological tolerances for specific hosts and vectors and effects of stressors on pathogen transmission can be used to develop increasingly accurate models of range expansions and effects of increased population density on heat stress and pathogen transmission.

Although predictive models may accurately identify the changes in suitable physical habitat for host and vector species, it is important to remember that the realized niche (actual range of conditions occupied by a species) is always smaller than the fundamental niche (potential range in the absence of predators, competitors, and other limiters); the many nonclimatic confounders discussed above (barriers to dispersal, predation, competition) will need to be included in accurate models (Lafferty 2009).


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: