Abstract
Leptospirosis is an underappreciated cause of human illness and livestock production loss in the tropics. Our understanding of the burden of disease, reservoirs and sources of human infection, and the performance of point-of-care diagnostic tests is limited in African countries. To fill data gaps, we recruited patients presenting with fever to two hospitals in Moshi, northern Tanzania, during 2012-2014 and at Endulen Hospital during 2016-17. We tested participants for leptospirosis using microscopic agglutination testing (MAT), Leptospira culture, and in a subset, polymerase chain reaction (PCR) and IgM point-of-care tests and evaluated test accuracy. We conducted a case-control study to identify risk factors for acute leptospirosis and conducted a survey of community livestock owners across northern Tanzania. We estimated leptospirosis incidence from the prevalence in Moshi Rural and Moshi Urban Districts by using multipliers derived from a health-care utilisation survey and calculated disability adjusted life years (DALYs). We compared MAT serogroup reactivity and Leptospira species detected in people with those detected in animals. To assess potential animals hosts of Leptospira serogroups and species we conducted a systematic review of all typed isolations and PCR detections of Leptospira from animals. We estimated the incidence of leptospirosis in several districts in northern Tanzania using a model that predicted an individual’s probability of leptospirosis based on their risk factors and the presence of recent fever. We estimated the annual leptospirosis incidence in Moshi Rural and Moshi Urban Districts as 11-18 cases per 100,000 people and 169 DALYs lost annually. Annual incidence in surrounding Districts was up to 85 per 100,000 persons. At our Moshi and Endulen sites the most commonly reactive serogroups were Australis and Djasiman. The systematic review found that serovars from the Australis serogroup have been isolated in Africa from field rats, multi-mammate rats, and cattle. Serovars of the Djasiman serogroup have not been isolated in Africa but have been isolated from dogs and small mammals in South America, Asia, and Europe. Using PCR, we identified L. borgpetersenii, L. interrogans, and L. kirschneri in the serum of patients with fever. Working in rice fields (odds ratio [OR] 14.6, 95% confidence intervals [CI] 2.9-59.5), cleaning cattle waste (OR 4.3, CI 1.2-12.9), feeding cattle (OR 3.9, CI 1.3-10.3) and being a farmer (OR 3.3, CI 1.3-8.2) were risk factors for acute leptospirosis. Increasing cumulative cattle urine exposure (OR 2.3, CI 1.1-4.7) and rodent urine exposure (OR 1.7, CI 1.1-2.8) were associated with leptospirosis on bivariable but not multivariable logistic regression. Working in rice fields (OR 3.6, 95% CI 1.5-9.0), slaughtering goats (OR 2.3, 95% CI 1.0-4.8), working as a farmer (OR 1.8, 95% CI 1.3-2.5), and frequently seeing rodents in the kitchen (OR 1.5, 95% CI 1.1-2.1) were risk factors for Leptospira seropositivity. All IgM point-of-care assays had low sensitivity compared with MAT. Leptospirosis causes substantial morbidity in northern Tanzania. Prevention of Leptospira infection in livestock is likely to reduce the burden of human leptospirosis. Clinicians should suspect leptospirosis in patients with fever who are rice workers or exposed to cattle urine, however IgM point-of-care tests are insufficiently accurate for clinical use.