Dogs as sentinels for Lyme disease in Massachusetts

Background. An investigation of the relationship between incident human cases of Lyme disease and seroprevalence of antibodies to B. burgdorferi in dogs was undertaken in order to determine whether dogs might serve as sentinels for Lyme disease. Methods. 3011 canine serum samples were analyzed by ELISA for antibody to B. burgdorferi. Records of incident human cases of Lyme disease were obtained from the Massachusetts Department of Public Health. Results. Regression analyses of the relationship between the log 10 (mean incidence in people 1985-1989) and canine seroprevalence from July 1988-August 1989 revealed that canine seroprevalence was highly predictive of incidence (R 2 = 0.86, p < .0001). A logistic regression model that incorporates the altitude of the town where each dog was resident, the date of sampling, and information on each dog's age, sex, and breed adequately explained the risk of canine seropositivity. Dogs resident at altitudes less than 200 feet, of sporting or large mixed breeds, and greater than two years of age were five times, four times, and almost three times more likely, respectively, to exhibit seropositivity than were other dogs. Conclusions. Estimates of the prevalence of antibody to B. burgdorferi in dog populations offers a sensitive, reliable, and convenient measure of the potential risk to people of B. burgdorferi in the environment. Risk factors for canine seropositivity may directly or indirectly illuminate certain aspects of the epidemiology of human Lyme disease.