Spatial analysis of human granulocytic ehrlichiosis near Lyme, Connecticut - Research

Emerging Infectious Diseases, Sept, 2002 by Emma K. Chaput, James I. Meek, Robert Heimer

Geographic information systems combined with methods of spatial analysis provide powerful new tools for understanding the epidemiology of diseases and for improving disease prevention and control. In this study, the spatial distribution of a newly recognized tick-borne disease, human granulocytic ehrlichiosis (HGE), was investigated for nonrandom patterns and clusters in an area known to be endemic for tick-borne diseases. Analysis of confirmed cases of HGE identified in 1997-2000 in a 12-town area around Lyme, Connecticut, showed that HGE infections are not distributed randomly. Smoothed HGE incidence was higher around the mouth of the Connecticut River and lower to the north and west. Cluster analysis identified one area of increased HGE risk (relative risk=1.8, p=0.001). This study demonstrates the utility of geographic information systems and spatial analysis to clarify the epidemiology of HGE.

**********

Historically, the study of the spread of diseases within populations has included a spatial component. New tools, including geographic information systems (GIS) and spatial statistics methods, enable epidemiologists to address the spatial aspects of disease rates and transmission more thoroughly and less subjectively. The emergence of tick-borne infections in the United States has been attributed to reforestation and second-growth forests, with the associated increases in reservoir and vector populations, as well as to human behavior changes including residential preferences and the increased popularity of outdoor recreational activities (1-3).

Our study used a GIS and spatial statistics to analyze the spatial distribution of a newly recognized tick-borne disease, human granulocytic ehrlichiosis (HGE). This disease was first described in a series of patients from northern Minnesota and Wisconsin in 1994 (4). The agent of HGE (1) "is most closely related to Ehrlichia phagocytophila, which infects sheep and cattle, and E. equi, which causes disease in horses. Recent research has suggested that rather than three separate species, these organisms are three variants of the same species (5-7). In the eastern and midwestern United States, the agent of HGE is transmitted to humans by the tick vector, Ixodes scapularis. This tick is also the vector of Borrelia burgdorferi and Babesia microti, the agents of Lyme disease and human babesiosis, respectively (8). The HGE agent is well established in vector populations in the Northeast (9-11), and infection with multiple I. scapularis--borne pathogens has been documented in both humans and wild mammal reservoirs (9,11-18). Since 1995, HGE has been a physician- and laboratory-reportable condition in Connecticut. In addition, an active surveillance system for HGE was established in 1997 in a 12-town area around Lyme, Connecticut (Figure 1), where Lyme disease was first described and remains highly endemic (19). This region has a total population of 83,600 and encompasses 330.7 square miles. During the 4 years of surveillance (1997-2000), the average annual incidence of confirmed cases of HGE in the 12-town area was 42 cases per 100,000 persons. For the same 4-year period, the average annual incidence of Lyme disease in the 12-town area was 234 cases per 100,000 persons.

[FIGURE 1 OMITTED]

The use of a GIS with spatial statistics, including spatial filtering (smoothing) and cluster analysis, has been applied to other diseases, in which it is often used to analyze and more clearly display the spatial patterns of disease (20-25). Smoothing decreases the random variation associated with small case numbers and small populations, enabling disease gradients or holes to be observed that may not be apparent with raw data (20,26,27). Cluster analysis identifies whether geographically grouped cases of disease can be explained by chance or are statistically significant (23,28); it detects true clusters of disease from cases grouped around population centers (29). While many risk factors and environmental cues may be similar for Lyme disease and HGE, investigating the spatial nature of the latter in an area known to be endemic for both may increase our understanding of the epidemiology of HGE and enhance our ability to focus education and control efforts to reduce human disease risk. The goal of our study was to describe the spatial distribution of HGE within a highly endemic area and to provide the groundwork for further study to identify the environmental and landscape characteristics associated with increased risk for HGE infections.

Materials and Methods

Cases

The confirmed cases of HGE analyzed in this study were identified through active and passive surveillance systems described elsewhere (19). Informed consent for participation in the active tick-borne disease surveillance study was obtained from all participants or their parents or guardians, according to a protocol approved by the Yale School of Medicine Human Investigation Committee. That committee approved a waiver of consent for this analysis. Only cases detected in 1997-2000 in residents of the 12-town area around Lyme, Connecticut, were included in the analysis. A confirmed case was defined as illness in a patient who had a seroconversion or [greater than or equal to] 4-fold change in antibody titer between acute-and convalescent-phase serum specimens (by indirect fluorescent antibody or enzyme-linked immunosorbent assay), a positive polymerase chain reaction assay with primer pairs directed to genomic sequences specific to HGE, or detection of the specific 44-kDa protein band by Western blot analysis. A probable case was defined as an illness in a patient with a positive antibody titer from only a single serum sample or a <4-fold change in antibody titer between acute- and convalescent-phase serum samples.