For mild/moderate acne, adapalene + BPO, clindamycin + BPO, and adapalene alone are most effective, in that order

Reference

Stuart B, Maund E, Wilcox C, et al. Topical preparations for the treatment of mild-to-moderate acne vulgaris: systematic review and network meta-analysis. Br J Dermatol 2021;185(3):512-525.

Synopsis

This interdisciplinary team of UK researchers began by asking patients with acne what was most important to them in a study of acne treatments. The patients stated that self-reported improvement was more important than investigator-reported outcomes. Thus, the primary outcomes of this study were the proportion of patients who reported moderate or better improvement, and the proportion who withdrew from the study or stopped using the study medication due to adverse events. This network meta-analysis searched carefully and identified 40 randomized trials evaluating 12 topical agents or topical combinations in 18,089 patients; 79% of the studies were from North America or Europe. Only studies that reported acne severity and in which fewer than half of patients had severe acne were included. The studies were a mix of active drug versus vehicle cream, or active drug versus a different active drug. A network meta-analysis reports both a meta-analysis of direct comparisons, as well as a meta-analysis of both direct and indirect comparisons. The authors found that adapalene plus benzoyl peroxide (BPO) was most effective compared with vehicle cream based on meta-analyses of both direct and indirect comparisons (odds ratio [OR] 3.65; 95% CI 2.58 - 5.15), with clindamycin plus BPO coming in second (OR 2.98; 2.22 - 4.01) and adapalene third (OR 2.44; 1.66 - 3.60). Limiting the analysis to only studies that made a direct comparison with vehicle cream had similar findings (the technical term for this is "coherence"), which gives us greater confidence in the overall results. With regard to withdrawals due to adverse events, adapalene plus BPO had the highest rate (OR 2.93; 1.69 - 5.08), followed by BPO alone (OR 1.59; 0.98 - 2.57) and clindamycin plus BPO (OR 1.44; 0.75 - 2.72). Although absolute probabilities and numbers needed to treat or harm aren't reported, the researchers state that withdrawals due to adverse events were uncommon.