Sub-Saharan African randomised clinical trials into male circumcision and HIV transmission: methodological, ethical and legal concerns

J Law Med. 2011 Dec;19(2):316-34.

Abstract

In 2007, WHO/UNAIDS recommended male circumcision as an HIV-preventive measure based on three sub-Saharan African randomised clinical trials (RCTs) into female-to-male sexual transmission. A related RCT investigated male-to-female transmission. However, the trials were compromised by inadequate equipoise; selection bias; inadequate blinding; problematic randomisation; trials stopped early with exaggerated treatment effects; and not investigating non-sexual transmission. Several questions remain unanswered. Why were the trials carried out in countries where more intact men were HIV-positive than in those where more circumcised men were HIV-positive? Why were men sampled from specific ethnic subgroups? Why were so many participants lost to follow-up? Why did men in the male circumcision groups receive additional counselling on safe sex practices? While the absolute reduction in HIV transmission associated with male circumcision across the three female-to-male trials was only about 1.3%, relative reduction was reported as 60%, but, after correction for lead-time bias, averaged 49%. In the Kenyan trial, male circumcision appears to have been associated with four new incident infections. In the Ugandan male-to-female trial, there appears to have been a 61% relative increase in HIV infection among female partners of HIV-positive circumcised men. Since male circumcision diverts resources from known preventive measures and increases risk-taking behaviours, any long-term benefit in reducing HIV transmission remains uncertain.

MeSH terms

  • Africa South of the Sahara
  • Circumcision, Male* / ethics
  • Developing Countries
  • Female
  • HIV Infections / prevention & control*
  • HIV Infections / transmission*
  • Humans
  • Informed Consent
  • Male
  • Randomized Controlled Trials as Topic / ethics
  • Randomized Controlled Trials as Topic / methods*