Contrary findings ignored
In Garenne (Reference Garenne2023b), six countries with circumcision ‘prevalence’ of 12.7–71.2% were studied. However, these prevalence figures incorrectly conflate voluntary medical male circumcision (VMMC) with traditional male circumcision (TMC), which can increase risk of human immunodeficiency virus (HIV) infection through use of unsterilised contaminated instruments on multiple youths, some already infected (Brewer et al. Reference Brewer, Potterat, Roberts and Brody2007, Reference Brewer, Potterat, Roberts and Brody2009; Ndiwane Reference Ndiwane2008). TMC of 16–20-year-olds is common in Lesotho where VMMC, but not TMC, is associated with HIV risk reduction (Coburn et al. Reference Coburn, Okano and Blower2013; Carrasco et al. Reference Carrasco, Rosen, Maile, Manda, Amzel and Kiggundu2020; Makatjane et al. Reference Makatjane, Hlabana and Letete2016). These observations render the study’s finding of no association questionable.
Confounding
These studies ignore confounding from antiretroviral therapy (ART). ART was rolled out alongside VMMC in sub-Saharan Africa concurrently with VMMC. While ART can reduce HIV infection, it increases HIV prevalence because people who previously would have died from acquired immunodeficiency syndrome-related illnesses now live with HIV (Shafer et al. Reference Shafer, Nsubuga, Chapman, O’Brien, Mayanja and White2013; Zaidi et al. Reference Zaidi, Grapsa, Tanser, Newell and Bärnighausen2013). The articles should have documented incidence. A modelling study found VMMC was the third most effective intervention after ART and condoms (Johnson et al. Reference Johnson, Meyer-Rath, Dorrington, Puren, Seathlodi, Zuma and Feizzadeh2022). Furthermore, uptake of VMMC has been high in young adolescent boys who are not yet sexually active (UNAIDS and WHO 2021), leading to a delay between VMMC and reduction in new HIV cases.
Data limitations
The primary data sources use self-reported circumcision status. Self-reporting is unreliable. This could affect any study using such data, but it seems to be a particular issue with Lesotho, which is a country Garenne focused on in the first paper we criticise. A study in Lesotho found that only half of men claiming to be circumcised actually were, and a further 26.6% were only partially circumcised (Thomas et al. Reference Thomas, Tran, Cranston, Brown, Kumar and Tlelai2011). Partial TMC is also common in Malawi (Renne et al. Reference Renne, Perry, Corneli, Chilungo and Umar2016).
Selective and misleading literature citations
The studies ‘cherry-pick’ often dated opinion pieces that fail to consider some of the issues described above. The articles ignore reviews of >30 studies linking VMMC to HIV risk reduction (Siegfried et al. Reference Siegfried, Muller, Volmink, Deeks, Egger, Low, Weiss, Walker and Williamson2003; Addanki et al. Reference Addanki, Pace and Bagasra2008). The studies also fail to point out that most subjects in Connolly et al. (Reference Connolly, Simbayi, Shanmugam and Nqeketo2008) had TMC, not VMMC. And findings in Rosenberg et al. (Reference Rosenberg, Goméz-Olivé, Rohr, Kahn and Bärnighausen2018) are attributed to self-selection, not ineffectiveness of VMMC.
Van Howe’s meta-regression analyses in 2015 are cited (Van Howe Reference Van Howe2015), but not the detailed critique undermining his statistics (Morris et al. Reference Morris, Barboza, Wamai and Krieger2018). Van Howe’s reply in 2018 (Van Howe Reference Van Howe2018) was rebutted (Morris et al. Reference Morris, Barboza, Wamai and Krieger2017). A meta-analysis (Van Howe Reference Van Howe1999) was discredited (Moses et al. Reference Moses, Nagelkerke and Blanchard1999; O’Farrell & Egger Reference O’Farrell and Egger2000) and became a textbook example of how not to do a meta-analysis (Borenstein et al. Reference Borenstein, Hedges, Higgins and Rothstein2009). Yet the articles cite it (Garenne & Matthews Reference Garenne and Matthews2019; Garenne Reference Garenne2023a,Reference Garenneb). All other meta-analyses confirm VMMC is effective against female-to-male HIV transmission (O’Farrell & Egger Reference O’Farrell and Egger2000; Weiss et al. Reference Weiss, Quigley and Hayes2000; Byakika-Tusiime, Reference Byakika-Tusiime2008; Lei et al. Reference Lei, Liu, Wei, Yan, Yang, Song, Yuan, Lv and Han2015; Sharma et al. Reference Sharma, Raison, Khan, Shabbir, Dasgupta and Ahmed2018; Farley et al. Reference Farley, Samuelson, Grabowski, Ameyan, Gray and Baggaley2020) but these are ignored.
Conclusion
These recent articles on VMMC and HIV are problematic. The issues described above are well-known to researchers in the field but are ignored. Unfortunately, the articles are now being cited (Garenne Reference Garenne2023c) to support an opposition to VMMC. Most authorities find that VMMC is biologically- and cost-effective against HIV infection in Africa (Farley et al. Reference Farley, Samuelson, Grabowski, Ameyan, Gray and Baggaley2020; Bershteyn et al. Reference Bershteyn, Mudimu, Platais, Mwalili, Zulu, Mwanza and Kripke2022; Bansi-Matharu et al. Reference Bansi-Matharu, Mudimu, Martin-Hughes, Hamilton, Johnson, ten Brink, Stover, Meyer-Rath, Kelly, Jamieson, Cambiano, Jahn, Cowan, Mangenah, Mavhu, Chidarikire, Toledo, Revill, Sundaram, Hatzold, Yansaneh, Apollo, Kalua, Mugurungi, Kiggundu, Zhang, Nyirenda, Phillips, Kripke and Bershteyn2023).
Funding statement
None.
Competing interests
Brian J. Morris is a member of the Circumcision Academy of Australia, a not-for-profit, government registered, medical society that provides accurate, evidence-based information on male circumcision to parents, practitioners, and others, as well as contact details of doctors who perform the procedure in Australia and New Zealand. This author’s interest in male circumcision for disease prevention began after he co-invented and competitively patented the first use of PCR for viral detection, applying this technology to human papillomavirus detection in screening for cervical cancer, the prevalence of which is lower in women with circumcised male partners.
Ethical standard
Not applicable.
