Injectable Use Associated With Increased Risk Of HIV Transmission

A. Kott

First published online:

Use of hormonal contraceptives, especially the injectable, is associated with an increase in a woman's risk of acquiring HIV or of passing it on to her male partner, new research suggests.1 In a prospective study of HIV-serodiscordant couples in Africa, women who were using hormonal contraceptives had twice the risk of nonusers of contracting HIV from their partner or transmitting the virus to him; the association was driven by the increased risk among injectable users. Moreover, compared with women who were not using hormonal contraceptives, HIV-positive injectable users were more likely to have HIV RNA in their genital tract—a possible mechanism for transmission—and had higher concentrations of the virus.

To examine associations between hormonal contraceptive use and HIV transmission, researchers analyzed data on 3,790 heterosexual, serodiscordant couples who had participated in either of two longitudinal HIV studies conducted in seven African countries from 2004 to 2010. The couples were recruited through community outreach and referrals from HIV testing and treatment centers, antenatal clinics and nongovernmental organizations. Participants had to be 18 or older and sexually active; those with HIV were ineligible if they had an AIDS-related illness, were using antiretroviral therapy or were pregnant.

Participants completed baseline and quarterly evaluations about sexual behaviors for up to 24 months (median, 18 months). Only women completed evaluations about hormonal contraceptive use. At each evaluation, women reported their current contraceptive method; they were assumed to have used the same method throughout the past three months. HIV-negative participants underwent quarterly serological testing, and HIV-positive participants underwent CD4 testing every six months. All participants received HIV counseling, free condoms and contraceptive supplies or referrals.

The researchers used generalized estimating equations to compare participant characteristics during periods of hormonal contraceptive use and nonuse, and Cox proportional hazards regression and marginal structural modeling to identify associations between hormonal contraceptive use and HIV transmission. For each three-month period, a woman was classified as having used a hormonal contraceptive if she had used the injectable or the pill; she was classified as a nonuser if she had used no method or condoms only, or had had a hysterectomy or tubal ligation. The investigators conducted separate analyses for women's acquisition of HIV and their transmission of the virus. Cases of female-to-male transmission were included in analyses only if genetic testing confirmed that both partners had the same strain of HIV, to avoid including transmission to men from outside partners whose use of hormonal contraceptives was unknown. Finally, researchers used logistic and linear regression to compare the prevalence and concentration of HIV RNA in cervical samples from hormonal contraceptive users with those from nonusers.

Among the 3,790 HIV-serodiscordant couples, the female partner was HIV-positive in 2,476 and HIV-negative in 1,314. Most couples were married and had children; the median age was in the mid-30s. HIV-positive participants had median CD4 counts of 455 cells per microliter.

At baseline, 15% of the HIV-negative women and 17% of the HIV-positive women were using a hormonal method; about three-fourths of users were receiving the injectable, and one-fourth were using the pill (4%). Overall, 1,321 women used hormonal contraceptives at some point in the study, although only about half used them during the entire study. During follow-up, men exposed to hormonal contraceptive use by their female partner were more likely than those not exposed to report having had unprotected sex with their partner during the past month (13% vs. 10%), but less likely to report having had sex outside the relationship in the past month (10% vs. 12%). Sexual behavior did not vary by hormonal contraceptive use among HIV-negative women, nor did plasma virus levels and CD4 counts in the infected partner differ by method use in either HIV group.

In multivariate analyses that adjusted for age, pregnancy, unprotected sex and the infected partner's plasma HIV levels, women who were using hormonal methods had twice as great a risk of becoming infected with HIV as women who were not using a hormonal method (adjusted hazard ratio, 2.0). The risk also was elevated among users of the injectable (2.1), but not among pill users, in part because of small sample sizes, according to the researchers. Similarly, the risk of contracting HIV was doubled among men whose partner was using any hormonal method or the injectable (2.0 for each).

Cervical samples taken from 1,691 HIV-positive women indicated that injectable users were more likely than those not using hormonal methods to have detectable levels of viral RNA in their genital tract (adjusted odds ratio, 1.7). In addition, the RNA was present at higher concentrations in injectable users than in nonusers.

The authors note several study limitations. First, they relied on self-reports to assess contraceptive use but did not collect information on contraceptive compliance or brands, whose properties may vary. Additionally, most participants were part of a randomized trial of a therapy to suppress herpes simplex virus type 2, which affects vulnerability to HIV infection (nearly 80% of HIV-infected people in Africa also are infected with herpes simplex virus type 2), and with which nearly all HIV-positive partners were infected.

Although the authors acknowledge that their findings are cause for concern, they note that "the benefits of effective hormonal contraceptive methods are unequivocal and must be balanced with the risk for HIV-1 infection." They recommend combining HIV and contraceptive counseling to educate women about the potential HIV risk that hormonal methods may pose, and about the importance of dual protection with condoms. In an accompanying editorial, Morrison and Nanda reiterated the study's limitations, but noted that other studies have also found possible evidence of a link.2 "The question of hormonal contraceptive use and risk of HIV acquisition remains unanswered after more than two decades," they wrote. "The time to provide a more definitive answer to this crucial public health question is now."—A. Kott


1. Heffron R et al., Use of hormonal contraceptives and risk of HIV-1 transmission: a prospective cohort study, Lancet Infectious Diseases, 2012 (forthcoming).

2. Morrison CS and Nanda K, Hormonal contraception and HIV: an unanswered question, Lancet Infectious Diseases, 2012 (forthcoming).