Semaglutide is widely used for metabolic disease, but cutaneous adverse events are inconsistently reported. This systematic review and meta-analysis evaluated randomized comparisons of semaglutide with placebo, usual care, or non-semaglutide active treatment. The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD420261464289). PubMed, Europe PubMed Central, trial registers, sponsor reports, and regulatory sources were examined through September 9, 2026; source verification was completed on September 21, 2026. Meta-analysis required directly available, source-verified participant event counts and denominators in both groups. Reported alopecia or hair loss and source-defined participant-level injection-site reactions were analyzed separately using common-effect Mantel-Haenszel odds ratios (ORs) without continuity correction. Risk of bias and evidence certainty were assessed for the primary syntheses. Thirty-one trials contributed verified outcome data. Twenty trials formed the primary placebo/common-control evidence set, of which 19 contributed relative-effect estimates and one reported zero injection-site reactions in both groups. Eleven additional trials provided active-comparator or design-specific evidence. Across nine comparisons, alopecia or hair loss was reported in 191/11,564 semaglutide recipients and 36/9,799 comparators (OR: 2.92, 95% confidence interval (CI): 1.99-4.27; p < 0.001; low-certainty evidence). The association persisted after exclusion of the largest cardiovascular outcomes trial with restricted ascertainment; random-effects analyses supported the primary interpretations. Eleven informative comparisons reported injection-site reactions in 156/5,066 semaglutide recipients and 114/3,483 comparators (OR: 0.86, 95% CI: 0.67-1.11; p = 0.256; very-low-certainty evidence). Active comparators and design-specific results were described separately. Semaglutide may increase reports of alopecia or hair loss; its comparative effect on injection-site reactions remains uncertain. Available-count synthesis does not resolve selective non-reporting or incomplete event ascertainment. The findings do not establish a hair-loss mechanism or injection-site safety equivalence.
