Climate factors influencing coccidioidomycosis seasonality and outbreaks

Environmental Health Perspectives, June, 2005 by Andrew C. Comrie

Although broad links between climatic factors and coccidioidomycosis have been established, the identification of simple and robust relationships linking climatic controls to seasonal timing and outbreaks of the disease has remained elusive. Using an adaptive data-oriented method for estimating date of exposure, in this article I analyze hypotheses linking climate and dust to fungal growth and dispersion, and evaluate their respective roles for Pima County, Arizona. Results confirm a strong bimodal disease seasonality that was suspected but not previously seen in reported data. Dispersion-related conditions are important predictors of coccidioidomycosis incidence during fall, winter, and the arid foresummer. However, precipitation during the normally arid foresummer 1.5-2 years before the season of exposure is the dominant predictor of the disease in all seasons, accounting for half of the overall variance. Cross-validated models combining antecedent and concurrent conditions explain 80% of the variance in coccidioidomycosis incidence. Key words: climate, Coccidioides, coccidioidomycosis, environment, meteorologic factors, rain, seasonal variation, southwestern United States, weather. Environ Health Perspect 113:688-692 (2005). doi:10.1289/ehp.7786 available via http://dx.doi.org/[Online 3 March 2005]

**********

Coccidioidomycosis, or valley fever, is caused by inhalation of spores from Caccidioides immitis and Coccidioides pasadasii. These dimorphic soil fungi are endemic to the deserts of the southwestern United States, Mexico, and elsewhere in Central and South America (Fisher et al. 2002; Kolivras et al. 2001). Although approximately 60% of people infected with the disease are asymptomatic, others experience mild influenza-like symptoms, and a small percentage experience severe effects and sometimes death resulting from dissemination of the disease to other parts of the body (Kolivras et al. 2001). Those at greatest risk for coccidioidomycosis infection include immunocompromised patients, young children, the elderly, and members of several ethnic minorities in the United States (Kolivras et al. 2001; Pappagianis 1988). In Arizona alone, > 2,000 cases per year have been reported (Komatsu et al. 2003), and the incidence of coccidioidomycosis is greater than that for other emerging infectious diseases in the region such as West Nile virus [Centers for Disease Control and Prevention (CDC) 2004a]. The number of Arizona cases is likely to exceed 3,000 by the end of 2004 (CDC 2004b).

Environmental conditions appear to have an important impact on coccidioidomycosis incidence. The current Arizona coccidioidomycosis epidemic has been linked to climate conditions (Kolivras and Comrie 2003; Komatsu et al. 2003; Park et al. 2005), whereas California experienced an epidemic in the 1990s that was possibly linked to drought conditions (Jinadu 1995). Initial links between climate conditions and the disease were identified several decades ago (Hugenhohz 1957; Maddy 1965). It is only recently that further details on climate and coccidioidomycosis have been published (Kolivras and Comrie 2003; Komatsu et al. 2003). These studies identified associations linking climate and other factors to seasonal patterns of coccidioidomycosis and to interannual variability and trends in the disease. Significant variables included drought indices, lagged precipitation, temperature, wind speed, and dust during the preceding 1 or more years. The relationships to coccidioidomycosis were quite complex, however, perhaps because of disease data issues outlined below. In this article I aim to identify simple and robust relationships linking climatic controls to seasonal timing and outbreaks of the disease, which until now have remained elusive and poorly understood. Important public health opportunities exist if environmental factors controlling coccidioidomycosis outbreaks and trends can be better comprehended, including the timing and degree of mitigation efforts such as soil treatment and the development of an advance warning system for public health management.

Part of the reason for the current state of knowledge has been the lack of high-quality disease data series. In fact, a major challenge to understanding more about the links between climate and infectious disease continues to be the difficulty in obtaining regular time series of disease data (National Research Council 2001). This is especially true for coccidioidomycosis with respect to data on Coccidioides in the soil or atmosphere. The current environmental detection method using laboratory mice is expensive and time-consuming, and although there is ongoing research into more rapid detection techniques (e.g., using polymerase chain reaction analysis to detect the fungus in soil samples), it will be several years before time series of such data become available. In the absence of suitable data on the environmental variability of the fungus itself, there is a need to exploit epidemiologic data in different ways to better identify the role of environmental controlling factors such as climate. Thus, for now, disease incidence data offer the best (and only) available multiyear time series for comparison with climatic conditions.

The use of human disease data to study potential relationships to climate conditions introduces numerous methodologic and analytical issues related to collection and reporting. Incidence data do not provide a homogeneous time series because of changes in reporting requirements, changes in population demographics, and the introduction of new diagnostic tests. In addition, the reported data necessarily contain imprecise estimations of disease onset dates because of various factors including patient recall, incorrect or delayed diagnoses caused by displacement of diagnoses during the respiratory disease season, and the variability in disease incubation and onset of symptoms from case to case.