This study has several important findings. First, asymptomatic individuals living in a region hyperendemic for leptospirosis had a high rate of seropositivity (at any level) for leptospiral infection (38% of 314 participants). Almost 60% of the seropositive individuals had evidence of recent sub-clinical infection, as indicated by MAT titer ≥1/800. Second, and of unique interest, a novel 16S rDNA hybridization assay used to screen urine samples for the presence of leptospiral DNA found that almost 5% of healthy people living in a rural Amazonian community were urinary shedders of Leptospira but did not have serological or clinical evidence of recent infection. Third, we found that both pathogenic and intermediately pathogenic Leptospira persistent infected the renal tubules of humans. Such observations have not been reported previously and are particularly notable because they demonstrate that inapparent leptospiral infection is common and frequently leads to shedding of organisms in urine. The long-term clinical significance of this finding remains to be determined.
The occurrence of leptospirosis, and indeed many infectious diseases, depends on several interacting variables. These include favorable environmental conditions, the density of local reservoir host populations, the type and frequency of exposure, exposure to infectious doses of the etiologic agent, the virulence of the infecting strain, and the lifestyle preferences and susceptibility of individuals within the exposed human population 
. In the context of this zoonotic infection, the density of local animal reservoir populations is likely an important determinant of the extent to which the environment may become contaminated by leptospires through urine from chronically infected carriers. When environmental conditions are ideal and background contamination is prevalent, social practices that predispose to infection, and the virulence of local strains are significant factors that affect the incidence of the disease 
. To date, there is no evidence that humans contribute to environmental contamination with Leptospira
, but the data presented here do not rule out this possibility.
Exposure to Leptospira
in this rural Amazonian study population was common (~39% were serologically positive at any MAT titer) with many subjects having evidence of recent sub-clinical infection. However, the serological data presented here need to be interpreted with caution: in an endemic setting, a high individual MAT titer (≥1
800) and/or IgM positivity are not reliable indicators of recent or current infection as antibodies may persist for prolonged periods 
. The high background exposure rates and relative absence of severe disease in this hyper-endemic region do suggest that long-term urinary shedding may occur more frequently here than elsewhere, where natural immunity may not be as common.
It is generally accepted that humans can excrete leptospires from weeks to months after infection 
. However the data presented here indicate that humans may excrete Leptospira
for periods exceeding a year; extending previous understandings of the carrier state. Ten asymptomatic individuals without clinical (no febrile episodes in more than a year) or serological evidence of recent exposure were found to be shedding either pathogenic or intermediately pathogenic Leptospira
in their urine. Although these persons may have been recently sub-clinically infected and either failed to produce anti-leptospiral antibodies or all produced ‘false-negative’ serology, these explanations seem unlikely. It is more likely that they represent long-term renal asymptomatic shedders of Leptospira
, regardless of whether patients were subclinically infected or had acute illness. However, a prospective study would be needed to assess this possibility. Nonetheless, prospective observational studies of such patients are required to confirm this hypothesis
Our data also suggest that women (especially mature women) are more likely to develop long-term renal carriage of Leptospira than are men; with a significant increase in incidence with age in women, possibly reflecting increased exposure with age or alternatively increased susceptibility. It is possible that the conclusion that women are more likely to be long-term asymptomatic urinary shedders than men may reflect a bias in the study, considering its relative underrepresentation of men. However, this observation may also reflect increased susceptibility of women to persistent leptospiral kidney infections; the reasons for this are unclear. However in our study population, the MAT titer was significantly lower in women than in men perhaps indicating that men are able to mount a more effective immune response than are women. Alternatively, men may have persistently higher antibody titers as a result of more frequent exposure due to work or recreational practices. Such possibilities require prospective study to address.
The long-term consequences of human renal infection by Leptospira need to be explored, in particularly the effect of persistent infection on renal function and electrolyte balance. Moreover, the nature of the infecting strains needs to be more carefully explored as some strains may be more likely than others to result in persistent renal infections in humans. Although we have identified the species of the infecting strains in the present study, other methods that are able to identify serovars, particularly isolation, will be more informative.
While humans are considered to be exclusively incidental hosts, animals can be maintenance and/or incidental hosts; maintenance hosts are defined as species in which infection is endemic, of low or no pathogenicity and (as a key factor) transmitted directly to the same species 
. Although human-to-human transmission has been rarely documented, it is unlikely that asymptomatic infected individuals have an important role in disease maintenance and transmission 
. An increased risk of having leptospiral antibodies in households of leptospirosis index cases compared to controls in an epidemic setting has been shown recently 
, but this is most likely related to common environmental exposure risks or genetic susceptibilities rather than direct transmission. In light of these data, further studies should address the possibility that long-term urinary shedders may represent a source of Leptospira
for their families and explore human-human transmission more carefully. Infection in carrier animals is usually acquired at an early age, and the prevalence of chronic excretion in the urine increases with age; we observed a similar trend in this population. Of note, none of the long-term urinary shedders had circulating anti-leptospiral antibodies; this is in accordance with early observations in Leptospira
-carrier mammals, where chronic urinary carriage was associated with low seropositivity to urinary culture rates in asymptomatic well-established serovar-specific carriers 
. Taken together, these observations make us speculate that in regions with high disease transmission, humans can develop some clinical and serological characteristics of asymptomatic urinary carriers, an attribute classically restricted to animals. Further longitudinal studies should address this possibility since the impact on disease transmission and in renal function of the affected individuals are unknown.
The study design had several limitations. First, we relied only on the recall of participants to define absence of fever in 1 year. Men were underrepresented; fewer men were recruited because of a lack of availability at the time of recruitment (most were away working). Thus, no leptospiruric males were detected. This observation suggests that we may have underestimated the overall number of asymptomatic shedders, as men have been typically associated with a higher risk of exposure due to work-related contact and behavioral practices. Another limitation was the initial non-efficient PCR screening strategy. The presence of other bacterial DNA hindered the identification of Leptospira
-positive clones; we detected both Leptospira
DNA in multiple samples, making the selection of colonies harboring the leptospiral 16S gene less efficient, in some instances, several hundred colonies had to be screened; we were unable to sequence the infecting strain in three enrollees due these technical limitations. Though unlikely, it is also possible that in these three instances the dot-blot gave false positive results. A third limitation of this study is that culture isolation of leptospires from urine was not attempted. Future work will be needed to further validate the molecular results presented here, and will use the PCR method to screen patients who then would have urine cultured for Leptospira
. Nonetheless, the deployment of a valid molecular tool to detect leptospiruria represents a new approach to assessing chronic asymptomatic infections in humans without the need for obtaining isolates. Finally, because L. licerasiae
serovar Varillal 
had not been fully characterized nor its epidemiological implications known, this strain was not used as antigen in the MAT panel or ELISA used to study patient sera, nor are these sera available for retrospective analysis.
Few of the published Leptospira
-specific PCR have been applied in clinical or field settings
. Furthermore, detection of bacterial DNA in urine is cumbersome because of the presence of PCR inhibitors and samples are often contaminated by multiple bacterial species whose DNA can interfere with the PCR assay 
Current understanding of host immune responses to Leptospira
or the pathogenesis of leptospirosis remains limited. Naturally acquired immunity that protects against re-infection by Leptospira
does occur and has been shown in animal models. It has been assumed that naturally acquired immunity is humorally-mediated particularly by antibodies against oligosaccharides of leptospiral LPS. Evidence also suggests that antibodies specific to Leptospira
membrane-associated proteins may play a role in host defense 
. We have documented that in this hyperendemic area, in spite of the high levels of environmental exposure to Leptospira
and high prevalence of seropositivity, the prevalence of severe disease is low 
. These observations suggest the possibility that protective immunity against severe disease from repeated infection may develop in areas with high leptospirosis transmission, especially if high frequency of infection leads to cross-serovar protection. Based on the finding that asymptomatic infection and urinary carriage are prevalent in this area where transmission is high and the prevalence of severe disease is low, we suggest that repeated exposure to Leptospira
and asymptomatic infection could induce protective acquired immunity. Longitudinal studies are needed to test this hypothesis.
In conclusion, we have identified a long-term renal shedder group among persons asymptomatically infected with pathogenic and intermediately pathogenic Leptospira. The health implications of long-term renal colonization and whether antibiotic treatment of such patients is required remain to be determined.