Design and analysis of randomized clinical trials for onchocerciasis, loiasis and mansonellosis: A systematic review.

BACKGROUND

The design and analysis of randomized clinical trials (RCTs) in filarial diseases such as onchocerciasis, loiasis, and mansonellosis pose unique statistical challenges, including skewed endpoints and limited sample sizes. This systematic review summarizes design and analysis approaches of RCTs conducted in these diseases with a focus on the statistical methodology.

METHODS AND FINDINGS: A systematic search was conducted in PubMed and four trial registries to identify RCTs investigating treatments for onchocerciasis, loiasis, and mansonellosis published or registered between 2000 and 2024. We excluded studies focusing on new methods or pharmacokinetics, short reports, and Phase I trials. Forty-four studies met the inclusion/exclusion criteria (23 for onchocerciasis, 16 for loiasis, and 5 for mansonellosis), information was retrieved from the registries, the manuscripts and/or the study protocol. As primary efficacy endpoints, for onchocerciasis studies qualitative endpoints dominated, while quantitative endpoints were more frequently observed for loiasis and mansonellosis. The most frequently reported hypothesis tests for the primary endpoint were the Mann-Whitney U and the chi-squared tests. We found considerable heterogeneity between trials - not only in study-specific parameters such as the number of arms, type of blinding or control group - but also in design parameters or attributes that could be standardized within each disease across studies with similar objectives, such as the primary endpoint, length of follow-up, the analysis method and the primary analysis population.

CONCLUSIONS

Several trials were well-planned with detailed information provided in either the manuscript or the registry. However, for some trials, information was sparse or incomplete, indicating a need for more structured and transparent reporting. Adopting established frameworks such as CONSORT and ICH E9 (R1) estimand approach would enhance transparency and better align trial objectives, analyses, and reported conclusions.