Determining risk for severe leptospirosis by molecular analysis of environmental surface waters for pathogenic Leptospira

Abstract

Background: Although previous data indicate that the overall incidence of human leptospirosis in the Peruvian Amazon is similar in urban and rural sites, severe leptospirosis has been observed only in the urban context. As a potential explanation for this epidemiological observation, we tested the hypothesis that concentrations of more virulent Leptospira would be higher in urban than in rural environmental surface waters.

Methods and findings: A quantitative real-time PCR assay was used to compare levels of Leptospira in urban and rural environmental surface waters in sites in the Peruvian Amazon region of Iquitos. Molecular taxonomic analysis of a 1,200-bp segment of the leptospiral 16S ribosomal RNA gene was used to identify Leptospira to the species level. Pathogenic Leptospira species were found only in urban slum water sources (Fisher's exact test; p = 0.013). The concentration of pathogen-related Leptospira was higher in urban than rural water sources (approximately 10(3) leptospires/ml versus 0.5 x 10(2) leptospires/ml; F = 8.406, p < 0.05). Identical 16S rRNA gene sequences from Leptospira interrogans serovar Icterohaemorrhagiae were found in urban slum market area gutter water and in human isolates, suggesting a specific mode of transmission from rats to humans. In a prospective, population-based study of patients presenting with acute febrile illness, isolation of L. interrogans-related leptospires from humans was significantly associated with urban acquisition (75% of urban isolates); human isolates of other leptospiral species were associated with rural acquisition (78% of rural isolates) (chi-square analysis; p < 0.01). This distribution of human leptospiral isolates mirrored the distribution of leptospiral 16S ribosomal gene sequences in urban and rural water sources.

Conclusions: Our findings data support the hypothesis that urban severe leptospirosis in the Peruvian Amazon is associated with higher concentrations of more pathogenic leptospires at sites of exposure and transmission. This combined quantitative and molecular taxonomical risk assessment of environmental surface waters is globally applicable for assessing risk for leptospiral infection and severe disease in leptospirosis-endemic regions.

Figure 1. The Urban Slum Environment of the Belen District of Iquitos

(A and B) Typical houses in the Belen district, built on stilts to avoid flood waters when the river rises. (C) Typical “sanitation” in the Belen Market area that attracts scavenger animals such as dogs and rats. (D) A typical view of the Belen Market area where commerce is conducted. Rats are often seen beneath tables and in gutters. Rats are so abundant in the market areas that they are commonly seen during the day although they are primarily nocturnal. (With kind permission of Springer Science and Business Media)

Figure 2. The Rural Village of Padrecocha, Outside of Iquitos

(A) A quebrada (stream), strongly associated with leptospirosis transmission. (B) A typical well in close association with activities of daily living. (C) The pond of Padrecocha (the cocha) adjacent to the river. (D) Flooding in the village during the high river season after heavy rains. (With kind permission of Springer Science and Business Media)

Figure 3. Quantification of Leptospira in Belen Versus Padrecocha

Bars indicate 95% confidence intervals. Analysis of variance, F = 8.406, p = 0.0003. Newman-Keuls multiple comparison test: Padrecocha versus Belen market, p < 0.001; Padrecocha versus Belen living area, p < 0.01; Belen market area versus Belen living area, p > 0.05.

Figure 4. Comparison of Leptospiral Counts in the Belen Living Area, According to Water Source

Means and 95% CIs are shown. Open gutters and river samples showed significantly higher bacterial counts than the samples from puddles and underground sources (unpaired t-test with Welch's correction: t = 2.284, p < 0.05).

Figure 5. Padrecocha Sequential Sampling

A total of 17 wells and six stream sampling points were tested over time. Filled boxes indicate PCR-positivity on that day (A). Bars indicate percentage of positivity of wells and stream sampled points for each tested day (B). The streams had a statistically significant higher frequency of positivity over time than the wells (χ² = 133.74, p < 0.001). NS, not sampled.

Figure 6. Phylogenetic Analysis of Leptospiral 16s rRNA Gene Sequences

Sequences represent human isolates and clones from environmental water samples obtained during a prospective study of leptospirosis in Peru using a Bayesian approach. The top cluster (A) comprises strains of the most pathogenic leptospiral species; also included are intermediates (B), the new clade (Clade C) (C), and saprophytes (D). Numbers at each node represent the posterior probability of the given node. Nodes without support values had less than 50% posterior probability, but are consistent with the overall branching of the tree. BEL indicates isolates obtained from patients seen in Belen, a busy riverside market community frequented by Rattus spp.; HAI indicates isolates obtained from patients seen at the Hospital de Apoyo; both are within the urban/periurban center of Iquitos (highlighted in red). Human isolates from rural areas (green) have been designated MOR (Moralillo), VAR (Varillal), and ZUN (Zungarococha) consistent with the location of the villages at which the patients were seen and the cultures obtained. Clones from water samples from Belen have been designated BEL (LA, living area; MA, market area), and samples from Padrecocha as PAD. Nodes highlighted with an asterisk (*) have support values less than 50%, but are consistent with the overall branching of the tree. Padrecocha is a rural area with a significantly lower rat population density but with many pigs roaming free.