Correction of meta-analysis after retraction of article on the effects of anti-androgens in polycystic ovary syndrome (eClinicalMedicine (2023) 63, (S2589537023003395), (10.1016/j.eclinm.2023.102162))

In our published article,1 we evaluated the use of anti-androgens (e.g., finasteride, flutamide, spironolactone) compared with various interventions in the management of polycystic ovary syndrome (PCOS). Our meta-analysis concluded that anti-androgens should not be preferred over combined oral contraceptive pills (COCPs) for treating hyperandrogenism, but may be considered for treating hirsutism when COCPs are contraindicated, poorly tolerated or ineffective after six months.1 It has come to our attention that one of the articles included in our meta-analysis, titled “Comparison of finasteride versus flutamide in the treatment of hirsutism” published in 1999 by Falsetti et al.2 was retracted in January 2024 by the European Journal of Endocrinology due to concerns about data integrity.3 In response to this retraction, we have re-examined our meta-analyses to assess whether exclusion of this study would influence our overall conclusions. The study by Falsetti et al.2 was one of two studies included in our meta-analysis comparing different dosing regimens of anti-androgens (finasteride, flutamide) in adult women with PCOS.1 By removing one of these two studies, the findings from this particular analysis, which suggested that daily doses were more effective for reducing hirsutism (−3.48 [−4.58, −2.39], p < 0.0001) and androstenedione levels (−0.30 ng/ml [−0.50, −0.10], p = 0.004), are no longer applicable and should be excluded from interpretation. It is important to note that this particular comparison has little clinical relevance because flutamide (examined in the study by Falsetti et al.2) is not generally recommended for managing hyperandrogenism in PCOS due to safety concerns, particularly the risk of hepatotoxicity.1,4 Additionally, concerns regarding the study by Falsetti et al.2 had already been identified prior to its retraction, as it was rated as having a high risk of bias, with GRADE assessment indicating a “very low certainty” of evidence for this comparison. Nevertheless, the following revisions are made to the paper: • In the abstract, the excerpt: “Daily use was more effective for hirsutism than use every three days (3.48 [4.58, 2.39], p < 0.0001, I2 = 1%) and resulted in lower androstenedione levels (0.30 ng/ml [0.50, 0.10], p = 0.004; I2 = 0%)” is revised to: “There were no differences between daily and intermittent anti-androgen use based on a single study”.• In the results (page 9), the paragraph under the subtitle ‘Anti-androgen comparisons', is replaced with the following paragraph: “One high RoB study5 compared finasteride regimens (2.5 mg daily versus 2.5 mg every 3 days for 5–10 months) and found that both regimens improved hirsutism scores, with no differences between groups. Intermittent low-dose finasteride is therefore as effective as continuous (daily) administration for improving hirsutism in PCOS.”• In the discussion (page 12), the excerpt: “Flutamide has also been associated with hepatotoxicity which was reflected in this systematic review where three women dropped out from the study by Falsetti et al.2 due to flutamide-induced hepatotoxicity” is revised to: “Flutamide has also been associated with hepatotoxicity.”In summary, the retraction of the study by Falsetti et al.2 requires the exclusion of one specific meta-analysis comparing anti-androgens, but this does not meaningfully alter the conclusions of our paper, nor does it affect the key findings. The broader recommendations therefore remain valid–that anti-androgens should not be used for hyperandrogenism in PCOS, but may be considered for hirsutism when COCPs are contraindicated, poorly tolerated or ineffective.ContributorsSA and AM reviewed the impact of the retracted study and drafted the letter. All authors reviewed and edited the letter, and approved the final version for publication.