Actionable evidence: Injectable contraceptives increase the risk of HIV infection

I was thinking of a new category of posts which I would be calling actionable evidence. Actionable evidence is that evidence which you can immediately put into practice; these are of direct clinical relevance. So to start this off I have selected a very important study that came out in the Lancet Infectious Diseases titled “Use of hormonal contraceptives and risk of HIV-1 transmission: a prospective cohort study”. Here is the main outcome of the study –

Risk of HIV acquisition in women

	Hazard ratio (95% CI) Adjusted Cox proportional hazards regression analysis	Hazard ratio (95% CI) Adjusted marginal structural models analysis
Any Hormonal Contraception	1.98 (1.06-3.68)*	1.84 (0.98-3.47)
Injectable	2.05 (1.04-4.04)*	2.19 (1.01-4.74)*
Oral	1.80 (0.55-5.82)	1.63 (0.47-5.66)
	*significant

Risk of transmission from women to men

	Hazard ratio (95% CI) Adjusted Cox proportional hazards regression analysis	Hazard ratio (95% CI) Adjusted marginal structural models analysis
Any Hormonal Contraception	1.97 (1.12-3.45)*	2.05 (1.12-3.74)*
Injectable	1.95 (1.06-3.58)*	3.01 (1.47-6.16)*
Oral	2.09 (0.75-5.84)	2.35 (0.79-6.95)
	*significant

Recommendation of the authors - Women should be counseled about potentially increased risk of HIV-1 acquisition and transmission with hormonal contraception, especially injectable methods, and about the importance of dual protection with condoms to decrease HIV-1 risk.

Comment –

Injectable contraceptives are used as suitable contraceptive method by many women. It has the advantage of ease of use. For example, DMPA (depot medroxyprogestrone acetate), one of the injectable contraceptives, can be taken at 3 monthly intervals. You don’t have to worry about contraceptives during these 3 months which is a problem with the oral pill, which has to be taken daily and one might forget to take it.

The study was of prospective cohort design with African serodiscordant (one partner with HIV infection and other partner free from infection) couples, though the majority of the data is taken secondarily from a randomized control trial. Basically the authors recruited HIV serodiscordant couples and subsequently they tested the HIV uninfected partner to see if they are getting infected. After doing this, they compared if the rate of infection  was more in seronegative subjects who were using hormonal contraceptives or in case of male subjects whose female partners were using hormonal contraceptives with those who were not using hormonal contraceptives. They followed the subjects for 24 months with mean follow up in seronegative women being 18 months and mean follow up of 18.7 months for seronegative men.

Overall this is a well conducted study, though there are few issues related to the methods. Firstly, the contraceptive use was self reported. Secondly, the use of injectable hormonal contraception was quite low among the participants but the study was adequately powered. Thirdly, the incidence of HIV infection was quite low. Fourthly, it is not clear to me whether the assessors were blinded to the nature of contraception use by the participants.

Before you ask the obvious, the authors controlled for the number of unprotected sexual intercourse. Even after controlling for unprotected sexual intercourse, injectable hormonal contraceptives significantly increased the risk of HIV infection. This suggests that something other than just unsafe sex is at play here.

Bottom line – Injectable hormonal contraceptives might increase the risk of HIV infection, evidence from observational study. All women using injectable hormonal contraceptives should be counseled to use condoms during intercourse to prevent HIV infection. Randomized controlled trial needed to see if the association holds.

Heffron R, Donnell D, Rees H, Celum C, Mugo N, Were E, de Bruyn G, Nakku-Joloba E, Ngure K, Kiarie J, Coombs RW, Baeten JM, & for the Partners in Prevention HSV/HIV Transmission Study Team (2011). Use of hormonal contraceptives and risk of HIV-1 transmission: a prospective cohort study. The Lancet infectious diseases PMID: 21975269