Global review maps serologic exposure among occupationally exposed workers

A systematic review and meta-analysis found that 18.3% of occupationally exposed workers included in studies from around the world had detectable antibodies against Leptospira bacteria. The estimate was based on 42 studies involving 11,155 participants, of whom 2,419 were seropositive. The researchers reported that the global figure had very low certainty because estimates varied sharply among studies, confidence intervals were often wide, and funnel-plot asymmetry raised the possibility of small-study effects. The result therefore indicates substantial evidence of occupational exposure but should not be treated as a uniform prevalence for every occupation or region.

The analysis synthesized studies published from January 2011 through February 9, 2026, when searches were conducted in LILACS, MEDLINE, Scopus, Web of Science, PubMed and ScienceDirect without language restrictions. Eligible reports were original peer-reviewed studies of occupationally exposed people, with accessible full texts, serological assessments and both sample sizes and numbers of seropositive participants. Two reviewers independently screened records and assessed methodological quality with the Joanna Briggs Institute Critical Appraisal Checklist for Studies Reporting Prevalence Data. The protocol followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and was registered in PROSPERO as CRD420261356167.

Researchers used a random-effects generalized linear mixed model with a random intercept and logit transformation, estimating between-study variance by maximum likelihood. The model yielded a pooled seroprevalence of 18.3% with a 95% confidence interval of 11.7% to 27.4%. A simpler calculation that divided all seropositive workers by all participants produced 21.7%, but the review distinguished this crude proportion from the model-based estimate because it gave greater weight to larger studies and did not account for within-study sampling variance or between-study heterogeneity. Individual study estimates ranged from 0% to 100%, and heterogeneity was considerable: I² was 96.7%, with a 95% confidence interval of 96.1% to 97.2%.

Geographic region accounted for a measurable part of the variation. Pooled seroprevalence was 33% in Asia, with a 95% confidence interval of 25% to 42%, and 27% in Africa, with a very wide interval of 9% to 58%. Lower pooled estimates were obtained for Oceania at 11% (5% to 21%), the Americas at 9% (4% to 20%) and Europe at 2% (0% to 9%). The formal comparison across continents was statistically significant, and continent explained 25.68% of heterogeneity in univariable meta-regression. These estimates nevertheless remained highly variable within regions, with I² values of 93.0% in Asia, 97.7% in Africa, 94.6% in the Americas and 95.7% in Oceania. The European estimate came from only four studies, while the African interval was particularly broad, so the figures cannot support precise geographic rankings.

Neither diagnostic method nor broad occupational category explained a measurable proportion of heterogeneity. Twenty-eight studies used the microscopic agglutination test, nine used enzyme-linked immunosorbent assay and five used both. Pooled estimates were 18% for microscopic agglutination testing, 14% for ELISA and 29% for the combination, but the formal subgroup test did not reach the 0.05 significance threshold (p = 0.0575). Diagnostic method also was not a significant moderator in meta-regression. Five occupational groupings—animal-related work, sanitation and waste, agriculture and fishing, environmental, mining and forestry work, and mixed or other work—produced pooled estimates between 11.4% and 21.0%, with no statistically significant differences among them. The broad groupings may have obscured differences in exposure intensity, duration, tasks, personal protective equipment and workplace conditions that were reported inconsistently across studies.

Diagnostic thresholds also did not fully explain the differences. Among 24 microscopic agglutination studies with identifiable positivity cut-offs, pooled seroprevalence was 22.6% for cut-offs below 1:100 and 11.0% for cut-offs of 1:100 or higher. Although the lower threshold was associated numerically with more seropositive workers, the subgroup difference was not statistically significant, and the threshold category explained only 1.05% of heterogeneity. Four of the 28 microscopic agglutination studies did not report an interpretable cut-off. The review also could not quantitatively compare serovar-panel composition because panels differed and were reported inconsistently, leaving potentially important differences in antigen coverage and locally circulating serovars unresolved.

Publication year showed a negative association with seroprevalence in univariable meta-regression, with a coefficient of −0.146 and a 95% confidence interval of −0.285 to −0.007. This variable accounted for 5.81% of the observed heterogeneity, but the relationship was not independent of geography. When publication year and continent were included together, the temporal coefficient was attenuated to −0.108, with a confidence interval of −0.237 to 0.022 and a p-value of 0.1023. The authors therefore cautioned against interpreting the unadjusted pattern as evidence of a worldwide decline in occupational exposure; it may partly reflect changes in the geographic distribution of the studies included over time.

Sensitivity analyses indicated that the overall pooled estimate was not dominated by a single unusual study. Removing a potentially influential Nigerian study reduced the pooled estimate from 18.3% to 16.7%, while considerable heterogeneity remained. Excluding the three studies with the most extreme observed proportions—10 of 10 seropositive, 252 of 263, and none of 347—changed the pooled estimate to 17.1%. Excluding the five smallest studies, leaving 37 studies with 10,799 participants, produced an estimate of 17.5%. These analyses support the numerical stability of the summary, although they did not remove the underlying variation among populations and settings.

Funnel-plot asymmetry was significant in Egger’s regression test (p = 0.0053), but the direction of the association differed from the usual pattern of small studies with high estimates being preferentially published. Smaller studies instead tended to report lower prevalence estimates, suggesting possible small-study effects rather than evidence that publication bias inflated the pooled result. Such tests are difficult to interpret in a highly heterogeneous evidence base, and asymmetry remained after extreme studies were removed. Applying an adapted Grading of Recommendations Assessment, Development and Evaluation approach, the authors rated certainty in the overall and principal subgroup estimates as very low because of inconsistency, imprecision and possible small-study effects. The review concludes that the results provide evidence of substantial but geographically variable occupational exposure, while standardized surveillance, more detailed exposure reporting and integration of human, animal and environmental surveillance are needed to produce estimates that are more precise and generalizable.

Subject of Research: Global occupational exposure to Leptospira based on a systematic review and meta-analysis of 42 studies

Article Title: Global seroprevalence of Leptospira spp. infection in workers with occupational risk: a systematic review with meta-analysis

Article References: Batista, A. I. V., Barbosa, B. M., Araújo, G. C. M., dos Santos, T. G., da Costa Filho, A. A., dos Santos, J. R. S., Parentoni, R. N., Campana, E. H., da Silva Neto, F. B., de Azevedo, S. S., Carvalho, L. R. R. A., & Brasil, A. W. D. L. (2026). Global seroprevalence of Leptospira spp. infection in workers with occupational risk: a systematic review with meta-analysis. Frontiers in Public Health, 14, Article 1931298. https://doi.org/10.3389/fpubh.2026.1931298

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.1931298

Keywords: Leptospira, leptospirosis, occupational exposure, seroprevalence, zoonoses, systematic review, meta-analysis, epidemiology, occupational health, One Health, Global, infection