April 2020 HSC Section 4 - Plastic and Reconstructive Problems

J A M A CAD D ERMATOL J ULY 2017

Adil and Godwin

Fig 1. Article identification, assessment, and selection.

measure of variance that could be used in a meta- analysis. Overall, all treatments were superior to placebo ( P \ .00001) in the 5 meta-analyses ( Fig 2 , A-E ). Most studies were conducted with male subjects. We found sufficient data to assess 2% minoxidil solution twice daily for women. Meta-analysis of these studies showed a mean difference of 1 12.41 hairs/cm 2 in the 2% minoxidil group compared with placebo treatment. The treatments that showed a mean difference in hair count listed from highest to lowest for men are finasteride 1 mg daily (18.37 hairs/cm 2 ), LLLLT (17.66 hairs/cm 2 ), 5% minoxidil twice daily (14.94 hairs/cm 2 ), and 2% minoxidil twice daily (8.11 hairs/cm 2 ). Heterogeneity was negligible (I 2 = 0%) in all groups except for finasteride versus placebo (I 2 = 91%; P \ .001). Although statistical hetero- geneity was, for the most part, not an issue with these studies and the interventions were deliv- ered in a very consistent way across studies, the

populations did vary somewhat from study to study, and there was variability in how the outcomes were assessed. We opted to use the random effects model, which is more conserva- tive than the fixed effect model, because of this clinical heterogeneity. Because there were only a few studies for each intervention, we used all of the studies to create a funnel plot looking for publication bias. Fig 3 suggests the possibility of a publication bias, which might mean some negative studies have not been published. No serious side effects were reported in any of the studies. A small number of study participants re- ported decreased libido with finasteride and dutas- teride. Most studies were industry supported, which should be considered when interpreting the results. DISCUSSION This meta-analysis strongly suggests that minox- idil, finasteride, and LLLLT are effective for

90