Rainfall and Human Leptospirosis Incidence: A Systematic Review of Ecological Temporal Studies

Study design and reporting guideline

This study was conducted as a systematic review to synthesize evidence on the relationship between rainfall and human leptospirosis incidence. The review was designed and conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 Statement. The review protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO; Registration No. CRD420261330699).

Eligibility criteria

Inclusion and exclusion criteria were developed to ensure that the selected studies were aligned with the objectives of this review. Because this review aimed to evaluate the temporal relationship between rainfall and human leptospirosis incidence, including the characteristics of rainfall exposure, lag effects, and seasonal patterns, only ecological studies incorporating a temporal component were considered eligible. Studies using time-series, longitudinal ecological, spatial-temporal, or spatiotemporal designs were included, whereas ecological cross-sectional studies without temporal analysis were excluded because they do not allow assessment of temporal associations between rainfall exposure and disease occurrence ( Table 1).

Information sources and search strategy

A systematic literature search was conducted in PubMed, Scopus, ScienceDirect, and Google Scholar to identify studies evaluating the relationship between rainfall and human leptospirosis incidence. The initial search was conducted between September 23 and September 30, 2025. To ensure the inclusion of the most recent evidence, updated searches were subsequently performed until June 2026 using the same search strategy.

The search strategy was developed using a combination of controlled vocabulary and free-text keywords. In PubMed, Medical Subject Headings (MeSH) terms and title/abstract keywords were used to maximize retrieval of relevant studies. For scopus, ScienceDirect, and Google Scholar, equivalent free-text search terms were applied according to the search functionalities of each database. The search strategy was designed to capture studies examining rainfall-related exposures and their association with human leptospirosis incidence.

Search terms were organized around three main concepts: leptospirosis, rainfall-related exposures, and meteorological factors. Keywords related to leptospirosis included “leptospirosis,” “Leptospira,” and “Leptospira infection.” Rainfall-related terms included “rainfall,” “rain,” “precipitation,” “heavy rainfall,” “extreme rainfall,” “rainfall variability,” “rainfall intensity,” “flood,” and “flooding.” Additional meteorological terms included “weather,” “climate variability,” and “meteorological factors.” Boolean operators (AND, OR) were used to combine search concepts and construct database-specific search strategies.

Database-specific search syntaxes were adapted to the indexing systems and search functionalities of each database while maintaining the same core search concepts. Searches were limited to articles published between January 2005 and June 2026. Detailed search strategies for all databases are provided in the Extended Data repository.

Study selection process

The study selection process was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. All records identified through database searching were imported into Rayyan software for reference management and study screening. After duplicate records were removed, titles and abstracts were independently screened by two reviewers (AFA and MKA) according to the predefined eligibility criteria.

Studies considered potentially eligible were retrieved in full text and independently assessed by the same reviewers. Disagreements arising during both the title/abstract screening and full-text assessment stages were resolved through discussion and consensus. If consensus could not be reached, the final decision was made through consultation with AAS and FP.

Reasons for exclusion at the full-text assessment stage were documented and recorded. The complete study selection process, including the number of records identified, screened, excluded, and included, is presented in the PRISMA 2020 flow diagram ( Figure 1).

Figure 1. PRISMA 2020 flow diagram of study selection.

Data extraction

Data were independently extracted by three reviewers (MKA, RKG, and JM) using a standardized data extraction form developed for this review. Extracted information included the first author, publication year, study location, study period, ecological study design, rainfall-related exposure measures, leptospirosis outcome measures, statistical analysis methods, reported lag periods or temporal exposure windows, seasonal patterns, rainfall thresholds or anomalies, and key epidemiological findings. Information on effect estimates was also extracted, including the type of effect measure reported (e.g., relative risks (RR), incidence rate ratios (IRR), adjusted incidence rate ratios (aIRR), regression coefficients (β), correlation coefficients (r or ρ), cross-correlation coefficients, posterior effect estimates, and other reported measures of association), together with their corresponding effect estimates, confidence intervals, credible intervals, or p-values when available. The direction of association between rainfall and leptospirosis occurrence was also recorded. Any discrepancies in data extraction were resolved through discussion among the three reviewers. If consensus could not be reached, consultation with SY and MN was undertaken to reach a final decision. Extracted data were subsequently summarized in evidence tables and used to characterize rainfall metrics, temporal lag structures, seasonal patterns, and epidemiological associations reported across the included studies.

Risk of bias assessment

Risk of bias was assessed using the Navigation Guide Systematic Review Methodology for environmental health studies. The assessment evaluated eight domains: selection bias, exposure assessment, outcome assessment, confounding, incomplete outcome data, selective reporting, conflict of interest, and other sources of bias. Risk of bias assessments were independently conducted by two reviewers (MKA and SSA) using predefined assessment criteria. Any disagreements were resolved through discussion between the reviewers. If consensus could not be reached, a third reviewer (LNR) was consulted to make the final decision.

Each domain was rated as low risk, probably low risk, probably high risk, or high risk of bias according to the Navigation Guide recommendations. An overall risk of bias judgment was subsequently assigned to each included study based on the combined assessment across all domains. Detailed study-level risk of bias assessments and justifications are available in the Extended Data repository.

Data synthesis

Given the substantial heterogeneity in study designs, analytical approaches, rainfall exposure definitions, temporal lag structures, outcome measures, and reported effect estimates, a quantitative meta-analysis was not considered appropriate. Studies differed considerably in their rainfall metrics (e.g., cumulative rainfall, rainfall anomalies, rainfall intensity, extreme rainfall events, and flooding-related indicators), temporal units of analysis (daily, weekly, monthly, and annual), statistical methods, and effect measures. Therefore, findings were synthesized narratively. Data synthesis focused on effect estimates and the direction of association between rainfall and leptospirosis occurrence, temporal lag structures reported between rainfall exposure and leptospirosis outcomes, and seasonal patterns associated with disease occurrence.

Reported effect estimates were extracted and summarized descriptively. Because studies reported diverse effect measures, including relative risks (RR), incidence rate ratios (IRR), adjusted incidence rate ratios (aIRR), regression coefficients (β), correlation coefficients (r or ρ), posterior effect estimates, and other measures of association, quantitative pooling was not undertaken. When multiple effect estimates were reported within a study, the estimate derived from the final model, best-fitting model, or primary rainfall-related analysis identified by the study authors was extracted. Corresponding confidence intervals, credible intervals, or p-values were also recorded when available.

Temporal lag structures were synthesized by extracting lag periods reported as statistically significant, epidemiologically relevant, or identified by the original study authors as the principal lag associated with leptospirosis occurrence. When studies reported multiple statistically significant lag periods, all relevant lag structures were recorded and included in the synthesis. Lag periods were subsequently summarized according to their reported temporal windows (e.g., weeks or months) to identify recurring patterns across studies.

Seasonal patterns were synthesized descriptively by extracting information on recurring rainfall-associated peaks, wet-season trends, monsoon-related patterns, flooding periods, and other temporal rainfall anomalies associated with leptospirosis occurrence. Seasonal findings were summarized according to the temporal patterns reported by the original study authors rather than being reclassified using a standardized seasonal framework. The direction of association was classified as positive, negative, mixed, or no significant association based on the findings reported in each study. Risk-of-bias assessments were considered during the interpretation of findings, with greater emphasis placed on evidence derived from studies judged to have lower risk of bias.

Overview of included studies

A total of 22 studies met the inclusion criteria and were included in the final synthesis.^15–36 The included studies were conducted across diverse geographical regions, including Asia, South America, Oceania, Europe, and the Indian Ocean region. Most studies were undertaken in tropical and subtropical areas where leptospirosis is endemic, whereas only a limited number of studies originated from temperate regions. Of the 22 included studies, 17 employed an ecological time-series design, while five used ecological spatiotemporal time-series approaches. The included studies demonstrated considerable heterogeneity in the definition of rainfall exposure and the measurement of leptospirosis outcomes. Rainfall was evaluated using daily, weekly, monthly, and annual measurements, as well as more specific indicators such as cumulative rainfall, rainfall anomalies, number of rainy days, extreme rainfall indices, and the Standardized Precipitation–Evapotranspiration Index (SPEI). Outcomes included laboratory-confirmed leptospirosis cases, incidence rates at different administrative levels, and leptospirosis-related hospitalizations. A variety of analytical approaches were used to evaluate the association between rainfall and leptospirosis occurrence, including Poisson and negative binomial regression models, generalized additive models (GAMs), distributed lag non-linear models (DLNMs), ARIMA/ARIMAX models, wavelet analyses, machine-learning approaches, and Bayesian spatiotemporal modelling. Differences in exposure definitions, temporal scales, outcome measures, and statistical methods indicate substantial methodological heterogeneity across studies. Detailed characteristics of the included studies are presented in Table 2.^15–36

Association between rainfall and leptospirosis occurrence

The synthesis of findings indicated that increased rainfall was generally associated with increased leptospirosis occurrence, although the magnitude of the reported effects varied according to the rainfall indicators, analytical scales, and statistical approaches used. Associations were reported using a range of effect measures, including incidence rate ratios (IRRs), relative risks (RRs), regression coefficients, correlation coefficients, and Bayesian effect estimates. Among studies reporting relative risk measures, effect sizes ranged from relatively small increases in risk for each incremental increase in rainfall to substantially higher risks under conditions of greater rainfall exposure. Positive effect estimates ranged from an adjusted incidence rate ratio (aIRR) of 1.001 (95% CI: 1.001–1.002) to a relative risk (RR) of 2.45 (95% CI: 1.80–3.33). In addition, long-term spatial analyses consistently identified positive associations between rainfall and leptospirosis occurrence across different analytical approaches and measurement scales. Similar positive associations were also observed in studies using correlation-based and time-series analyses. Reported association strengths ranged from moderate to very strong, with correlation coefficients varying between 0.34 and 0.92. These findings indicate a consistent temporal relationship between increased rainfall and increased leptospirosis occurrence across diverse geographical and epidemiological contexts. Although positive associations were commonly reported, heterogeneity in findings was observed across studies. Some studies found no statistically significant association between precipitation and leptospirosis occurrence (IRR 1.00; 95% CI: 0.99–1.01), whereas others reported differing directions of association across locations, with positive associations in one setting and negative associations in another. These findings suggest that the relationship between rainfall and leptospirosis occurrence may vary across geographical contexts. A summary of the reported effect estimates and directions of association for each included study is presented in Table 3.

Temporal patterns of rainfall-associated leptospirosis

The synthesis of temporal patterns indicated that the effects of rainfall on leptospirosis occurrence may emerge across a broad range of time intervals, from immediate effects to several months after exposure. Reported lag patterns could be broadly categorized into short lags (0–3 weeks), intermediate lags (1–3 months), and long lags (>3 months). Lag periods within the 1–3 month range were reported across diverse geographical settings and analytical approaches, suggesting that increases in leptospirosis occurrence frequently follow a period of delay after increased rainfall. In addition to lag effects, relatively consistent seasonal patterns were observed across multiple study settings. Increased leptospirosis occurrence generally coincided with rainy seasons, monsoon periods, or annual rainfall cycles. Temporal synchronization between peak rainfall and peak leptospirosis occurrence was also reported in various geographical contexts, with increases in cases occurring after periods of elevated rainfall. However, seasonal patterns were not observed uniformly across all study settings, reflecting temporal variation among geographical and climatic contexts. Overall, the reported temporal patterns indicate that increases in leptospirosis occurrence may occur at different time intervals following increased rainfall and often correspond to local seasonal rainfall cycles. A summary of the lag structures and seasonal patterns reported by the included studies is presented in Table 4.

Rainfall characteristics associated with increased leptospirosis risk

The synthesis of findings indicated that increased leptospirosis risk was associated not only with the amount of rainfall but also with specific rainfall characteristics. These characteristics included extreme rainfall events, accumulated rainfall over defined periods, positive rainfall anomalies, wetter-than-normal conditions, and rainfall-related flooding events. Extreme rainfall was consistently associated with increased leptospirosis occurrence. Indicators used across studies included rainfall exceeding the 90th percentile (34.5 cm/month), daily rainfall of 175–250 mm, extreme rainfall exceeding 100 mm/week, and threshold-based extreme rainfall indices of more than 50 mm/day. Increased risk was also reported during periods of heavy rainfall (≥16 cm/week), intense rainfall (≥32 cm/week), and very heavy rainfall (≥63 cm/week). These findings suggest that elevated leptospirosis risk was observed across a range of extreme rainfall indicators. Leptospirosis risk was also associated with accumulated rainfall over longer periods, including six-week cumulative rainfall, monthly rainfall, and annual precipitation. Higher risk was further observed under conditions of positive rainfall anomalies of +20 mm/week above expected seasonal levels and wetter-than-normal conditions measured using the Standardized Precipitation–Evapotranspiration Index (SPEI-3), with values up to 2.79. Rainfall-related flooding was also reported as an important characteristic accompanying increased leptospirosis risk. Overall, these findings indicate that increased leptospirosis risk is influenced not only by the quantity of rainfall but also by its intensity, accumulation, deviation from normal conditions, and associated hydrological impacts.

Risk of bias assessment

Risk of bias was assessed for all 22 included studies using the Navigation Guide Systematic Review Methodology. Overall, seven studies were judged to have a low risk of bias, five were classified as probably low risk of bias, nine were rated as probably high risk of bias, and one study was judged to have a high risk of bias. Detailed study-level risk of bias assessments are available in the Extended Data repository. Across the included studies, confounding was the domain most frequently associated with elevated risk of bias. Although many studies adjusted for selected meteorological variables, several were unable to account for other determinants of leptospirosis transmission, including flooding characteristics, sanitation conditions, land-use patterns, rodent reservoirs, occupational exposures, and socioeconomic factors. Additional concerns included limited geographical coverage and potential exposure misclassification. In contrast, exposure assessment and outcome assessment were generally rated as low or probably low risk of bias. Rainfall exposure data were predominantly obtained from official meteorological agencies, monitoring systems, or validated environmental datasets, whereas leptospirosis outcomes were primarily derived from surveillance databases, hospital records, or laboratory-confirmed cases. Selective reporting and conflict of interest were also generally assessed as low risk of bias. Overall, the risk of bias assessment suggests that the included studies provided reasonably reliable evidence regarding the association between rainfall and leptospirosis occurrence. Nevertheless, residual confounding remained an important methodological concern and should be considered when interpreting the findings of this review.

Limitations

This review has several limitations that should be considered when interpreting the findings. First, the literature search was restricted to articles published in English and Indonesian and indexed in four databases, namely PubMed, Scopus, ScienceDirect, and Google Scholar. In addition, grey literature was not included in the search strategy; therefore, it is possible that some relevant studies were not identified. Second, the reviewed studies exhibited substantial heterogeneity in the definition and measurement of rainfall exposure, including differences in temporal scales, lag structures, and the use of various rainfall indicators and hydrometeorological indices. Variations in the analytical approaches employed across studies also limited direct comparisons of the reported effect sizes. This heterogeneity precluded the conduct of a quantitative synthesis through meta-analysis, and therefore the findings of this review are presented as a narrative synthesis. Furthermore, most of the included studies used ecological or time-series observational designs, which allow the identification of patterns of association at the population level but cannot be used to draw direct causal inferences at the individual level. Consequently, the findings should be interpreted with caution, as associations observed at the population level may not necessarily reflect relationships at the individual level, a limitation commonly referred to as the ecological fallacy. The observed associations may also have been influenced by other environmental factors that frequently interact with rainfall, such as temperature, humidity, flooding, land use, and animal reservoir dynamics, which were not consistently controlled for across studies. Nevertheless, this review successfully integrates evidence from diverse geographical and epidemiological contexts and provides a more comprehensive understanding of the relationship between rainfall and leptospirosis occurrence, including the temporal patterns and rainfall characteristics associated with increased disease risk.