Corticosteroids for treatment of leptospirosis

Leptospirosis is a bacterial disease caused by Leptospira spp, a zoonotic pathogen spread via contaminated soil and water. Corticosteroids have been used for the treatment or prevention of severe manifestations of disease, but the indications for their use and treatment efficacy remain uncertain. This review evaluates the existing evidence for the use of corticosteroids in leptospirosis from randomised trials.

To assess the benefits and harms of corticosteroids versus no intervention, no intervention beyond standard of care, or placebo for the treatment of people with leptospirosis.

Electronic searches in the Cochrane Hepato-Biliary Group Controlled Trials Register, Cochrane Central Register of Controlled Trials in the Cochrane Library, MEDLINE, Embase, LILACS, Science Citation Index Expanded, Conference Proceedings Citation Index - Science, and other resources were conducted. We searched online clinical trial registries to identify unpublished or ongoing trials, and reviewed reference lists from the identified publications for potential trials. We contacted authors of identified trials, relevant individuals, and organisations for additional information. The last search date was 10 April 2025.

We considered the inclusion of randomised clinical trials of any trial design which assessed corticosteroids for the treatment of leptospirosis. We imposed no restrictions on age, sex, occupation, comorbidity of trial participants, or outcomes reported. We looked for trials assessing corticosteroids irrespective of type, route of administration, dosage, and schedule versus no intervention, placebo, or no intervention beyond standard care. We included trials meeting any of these criteria, irrespective of the manuscript's primary language.

We adhered to Cochrane methodology. Data entry and analysis were facilitated by the use of the Review Manager. The primary outcomes were all-cause mortality and the proportion of individuals experiencing serious adverse events. The secondary outcomes were quality of life, the proportion of individuals experiencing non-serious adverse events, days of hospitalisation, and the proportion of individuals experiencing Jarisch-Herxheimer reactions. We employed the risk of bias 2 tool (RoB 2) to assess the bias risk of included trials. We used the GRADEPro software to evaluate the certainty of evidence. We presented dichotomous outcomes as risk ratios (RR) and continuous outcomes as mean differences (MD), both accompanied by their corresponding 95% confidence intervals (CI). We applied a random-effects meta-analysis for the primary analysis and a fixed-effect model for the sensitivity analyses. Our primary outcome analyses included trial data at the longest follow-up. We analysed the outcome data regardless of the risk of bias.

Four randomised trials were included in this review, with a pooled total of 253 participants. Each of the trials compared a corticosteroid (prednisolone, hydrocortisone, a combined treatment regimen of dexamethasone and prednisolone, or methylprednisolone) versus no intervention, no intervention beyond standard of care, or placebo. Participants in three trials received similarly administered co-interventions as standard of care, and had no further intervention in the fourth trial. All participants were recruited from populations presenting to general hospitals in leptospirosis endemic settings. The ages of the participants ranged from six years to 65 years. Depending on the trial, the treatment duration ranged from over four hours to seven days. All included trials were judged to have either some concerns or to be at high risk of bias. The certainty of evidence for all evaluated outcomes was judged to be very low. Quality of evidence was downgraded for risk of bias arising from the randomisation process, measurement of outcome, and selection of reporting of results; indirectness of evidence due to choice of intervention; inconsistency due to different point estimates and unexplained heterogeneity; and imprecision attributable to confidence intervals (CI) crossing clinically important thresholds, failure to meet optimal information size, or an upper/lower CI boundary more than three risk ratios. Corticosteroids compared with no intervention beyond standard of care or placebo may have little to no effect on all-cause mortality (RR 1.04, 95% CI 0.38 to 2.80, I² = 0%, 3 trials, 123 participants, very low-certainty evidence) and on the proportion of individuals experiencing serious adverse events (RR 1.15, 95% CI 0.32 to 4.11, I² = 62%, 3 trials, 123 participants, very low-certainty evidence), but the evidence is very uncertain. Corticosteroids compared to no intervention beyond standard of care or placebo may increase the proportion of individuals experiencing non-serious adverse events (RR 2.00, 95% CI 0.21 to 18.98, 1 trial, 22 participants, very low-certainty evidence), but the evidence is very uncertain. Corticosteroids compared to no intervention beyond standard of care or placebo may decrease the number of days of hospitalisation (MD 0.46, 95% CI -1.81 to 2.73, I² = 83%, 3 trials, 123 participants, very low-certainty of evidence), but the evidence is very uncertain. Corticosteroids compared to no intervention may reduce the risk of Jarisch-Herxheimer reaction events (RR 0.13, 95% CI 0.04 to 0.41, 1 trial, 130 participants, very low-certainty evidence), but the evidence is very uncertain. None of the four trials assessed health-related quality of life. We have listed one trial registered as 'randomised' in studies awaiting classification because we could not identify further information. We have listed one trial in the ongoing section because trial recruitment has not yet started.

Based on the very low certainty of evidence attributable to our analyses, we do not know whether corticosteroids compared with no intervention, no intervention beyond standard of care, or placebo, reduce all-cause mortality, increase the risk of serious or non-serious adverse events, decrease days of hospitalisation, or decrease the proportion of people experiencing Jarisch-Herxheimer reaction events. None of the four trials assessed health-related quality of life. There is a lack of harmonised treatment strategies, clinically relevant outcome definitions, and definitive and rigorously designed randomised trials to support the use of corticosteroids for leptospirosis. Future research should focus on these evidence gaps.