Meta-analysis of risk factors for canine leptospirosis

Abstract

Leptospirosis is one of the most widespread zoonotic diseases, with more than one million human cases reported worldwide every year. Dogs could develop infections that range from asymptomatic to severe, and shed leptospires with their urine. Given their close contact with humans, dogs may act both as epidemiological links or as sentinels of pathogenic leptospires in the environment. The aims of our study were to quantitatively summarize the overall prevalence of leptospiral antibodies and to identify factors associated with the probabilities of infection. We searched the electronic databases Scopus, PubMed, PMC and ScienceDirect for observational studies on canine leptospirosis published between 1989 and December 2019 and written in English, Spanish or Portuguese. We fitted a series of multilevel random effects meta-analysis models to estimate the prevalence of antibodies against Leptospira for different types of dogs, health statuses, diagnostic tests, geographic regions and income categories of the countries. We also fitted a number of random effects meta-analysis models to estimate the pooled odds-ratio of factors associated with canine leptospirosis. After removing duplicates and articles not meeting selection criteria, a total of 130 studies in 91 articles were included in this work. We found lower seroprevalence estimates in North America countries (P<0.001) and other high income countries (P<0.001). We also found higher probabilities of leptospiral infection in adult (P=0.017), male dogs with access to the streets (P<0.001). Identifying the profile of dogs that are more exposed to leptospirosis could be useful in the design of public health strategies for the prevention and control of leptospirosis.