Twelve RCTs (18 comparisons) were eligible for inclusion. The total sample size was over 506 (sample size was not reported for one study). Trial quality was low. Only one RCT reported adequate allocation concealment and only half clearly reported blinded outcome assessment. Median loss to follow up was four per cent (range 0 to 42 per cent).
Antiandrogens versus placebo
Hirsutism scores were significantly lower in the intervention group (WMD -3.9, 95% CI: -5.4, -2.3, NNT=3, five comparisons); there was no heterogeneity. Subgroup analysis showed no statistically significant differences in effect between spironolactone, flutamide and finasteride.
Antiandrogens versus OCPs
One RCT reported no statistically significant difference between the comparison groups.
Antiandrogens versus metformin
Scores were significantly lower in the intervention group (WMD -3.7, 95% CI: -6.8, -0.6, three RCTs), but with marked heterogeneity (I2=80%), partly explained by a larger effect associated with flutamide (two RCTs) than with spironolactone (one RCT).
Antiandrogens as add-ons
No statistically significant difference was found between OCP plus antiandrogens (spironolactone, finasteride and CPA) and OCP alone (five comparisons). Flutamide plus metformin was associated with significantly lower scores than metformin alone (WMD -4.6, 95% CI: -7.9, -1.3, NNT=2), but with moderate heterogeneity (I2=40%) that was unexplained by subgroup analyses.
Patient self-assessments (three RCTs):
Patient assessments were congruent with clinician assessments in two RCTs that reported both measures. A third RCT reported a trend towards improved outcomes in the intervention group (p=0.09).
Other subgroup analyses were reported in the review.