Impact of environment and social gradient on Leptospira infection in urban slums

Abstract

Background: Leptospirosis has become an urban health problem as slum settlements have expanded worldwide. Efforts to identify interventions for urban leptospirosis have been hampered by the lack of population-based information on Leptospira transmission determinants. The aim of the study was to estimate the prevalence of Leptospira infection and identify risk factors for infection in the urban slum setting.

Methods and findings: We performed a community-based survey of 3,171 slum residents from Salvador, Brazil. Leptospira agglutinating antibodies were measured as a marker for prior infection. Poisson regression models evaluated the association between the presence of Leptospira antibodies and environmental attributes obtained from Geographical Information System surveys and indicators of socioeconomic status and exposures for individuals. Overall prevalence of Leptospira antibodies was 15.4% (95% confidence interval [CI], 14.0-16.8). Households of subjects with Leptospira antibodies clustered in squatter areas at the bottom of valleys. The risk of acquiring Leptospira antibodies was associated with household environmental factors such as residence in flood-risk regions with open sewers (prevalence ratio [PR] 1.42, 95% CI 1.14-1.75) and proximity to accumulated refuse (1.43, 1.04-1.88), sighting rats (1.32, 1.10-1.58), and the presence of chickens (1.26, 1.05-1.51). Furthermore, low income and black race (1.25, 1.03-1.50) were independent risk factors. An increase of US$1 per day in per capita household income was associated with an 11% (95% CI 5%-18%) decrease in infection risk.

Conclusions: Deficiencies in the sanitation infrastructure where slum inhabitants reside were found to be environmental sources of Leptospira transmission. Even after controlling for environmental factors, differences in socioeconomic status contributed to the risk of Leptospira infection, indicating that effective prevention of leptospirosis may need to address the social factors that produce unequal health outcomes among slum residents, in addition to improving sanitation.

Figure 1. Slum community site in the city of Salvador, Brazil.

(A) The yellow line in the aerial photograph is the boundary of the study site in the Pau da Lima community. The map in the bottom left corner shows the location of Salvador in Brazil and the study site (red) within the city. (B) Photograph of the typical environment at the community study site, which shows a valley in which households is situated and the proximity of households to open sewers and refuse. (C) Close-up view of the orthomap used to georeference households (red and black dots) and environmental attributes, such as open sewers (blue line) and refuse deposits, for the region marked as a yellow box in Panel A. The red arrow represents the direction from which the photograph in Panel B was taken.

Figure 2. Distribution of reciprocal microscopic agglutination test titers for 3,171 subjects from the slum community site.

Labels above the bars indicate the number of subjects (% of total), according to their highest reciprocal titer. The open bar represents seronegative subjects. Subjects with highest reciprocal titres against L. interrogans serovar Copenhageni, multiple serovars and serovars other than Copenhageni are shown as black bars, grey bars and crosshatched bars, respectively.

Figure 3. Spatial distribution of subjects with Leptospira antibodies and all enrolled subjects, according to place of residence, and environmental attributes of the community site.

Panels A and B show the smoothed Kernel density distribution of subjects with Leptospira antibodies (N = 489) and all (N = 3,171) subjects, respectively, according to place of residence. The yellow-to-red gradient represents increasing density in smoothing analyses which used 40 meters as the bandwidth. Black circles show the location of subject households. Panel C shows the distribution of the population-adjusted Kernel density estimator for subjects with Leptospira antibodies which was calculated as the ratio of the estimators for subjects with Leptospira antibodies and all subjects. Panel D shows a topographic map generated by the digital terrain model. The yellow line is the level that is 20 meters above the lowest point in the four valleys within the community site. Panels E and F show the distribution of, respectively, open rainwater and sewage drainage systems and accumulated refuse according to size (m²).

Figure 4. Generalized additive models (GAM) of the association between the risk of acquiring Leptospira antibodies and continuous variables of (A) per capita daily household income, (B) level of household in meters above the lowest point in valley, and (C) distance in meters to the nearest open sewer, and (D) reported number of rats sighted in the household environment.

The coefficient, f(infection), in the GAM model is a measure for the risk of acquiring Leptospira antibodies. In Panels A, B, C and D, the x axis intercept values, where f(infection) equals zero, were US$1.70/day, 22 meters, 22 meters and 2 rats, respectively.