An estimated 150 million women worldwide use hormonal contraception to prevent pregnancy . Over 15 million women are living with HIV, most of whom are of childbearing age and, similar to other women, need access to safe and effective contraception . In many resource-limited countries, access to effective and safe contraception takes on greater importance, owing to the risk of perinatal HIV transmission, high rates of maternal mortality and many other maternal and infant benefits of birth spacing . Although available data are conflicting and scarce, some recent studies [4,5] in humans and animals suggest that hormonal contraception may accelerate HIV disease progression. Data from studies in macaques originally suggested progesterone as the likely cause. Macaques implanted with progesterone and then infected with simian HIV had higher viral loads and faster progression to AIDS than macaques receiving placebo . Additionally, women in a cohort recruited in Mombasa, Kenya, who were using depo-medroxyprogesterone acetate (DMPA) at the time of HIV acquisition had higher viral load set points, a marker for more rapid disease progression .
The Mother-to-Child Transmission (MTCT)-Plus Initiative provided support to 12 clinical programs in Africa and one in Asia (Thailand) to implement HIV/AIDS care and treatment to families identified through perinatal HIV prevention services. Programs were initiated between 2002 and 2004 for family-centered care before such services became more widely available through national programs in Africa and Asia. Enrolled patients in MTCT-Plus Initiative were followed through March 2008. We used data from MTCT-Plus Initiative to investigate whether exposure to hormonal contraception was associated with more rapid HIV disease progression among women not yet on antiretroviral therapy (ART).
The MTCT-Plus Initiative
We obtained data for this analysis from the MTCT-Plus Initiative, a multicountry HIV care and treatment program that utilized a woman-centered, family-focused service model. The design and progress of the Initiative has been described previously (see www.mtctplus.org) [8,9]. At each clinical site, pregnant or postpartum HIV-infected women receiving programs to prevent MTCT (PMTCT) services were enrolled into ongoing HIV care regardless of their HIV disease stage. Women were enrolled either during pregnancy or postpartum and HIV-exposed children, infected partners and other family members were also enrolled into the program.
Women received a package of HIV primary healthcare services, including regular clinical examinations, CD4+ cell counts performed at local laboratories every 6 months, assessment for opportunistic infections with appropriate provision of prevention and treatment for these infections and support to enhance adherence and retention. Barrier and nonbarrier contraceptive methods were provided, either onsite or by referral, with specific methods based on women's preference and local availability. ART initiation was based on national or WHO guidelines. Typically, criteria for starting ART were WHO stage IV disease, CD4+ cell count less than 200 cells/μl, or WHO stage II or III disease with a CD4+ cell count less than 350 cells/μl. In January 2005, the MTCT-Plus guidance changed so that stage II disease patients with a CD4+ cell count of less than 350 cells/μl were no longer eligible for ART. Patients who were not eligible for ART were followed every 6 months. All patients on ART were seen monthly for 6 months and then every 3 months thereafter if clinically stable. Patients who did not return for a scheduled visit were actively followed up in the community, although the methods used varied by program. Repeat CD4+ cell counts, serum chemistry and hemoglobin were performed every 6 months or as deemed necessary by the clinicians.
Upon enrollment into the MTCT-Plus programs, clinicians at each site completed a short, standardized form, which included basic demographic and socioeconomic information, partnership status and obstetric history. In addition, at enrollment and each follow-up visit, information on the occurrence of opportunistic infections, WHO staging and initiation or continuing use of ART were collected. For women, the standardized follow-up form also included items on use of contraception (including use of specific barrier and nonbarrier methods) and pregnancy status (including pregnancy loss since the previous visit), both based on the women's self-reports. The specific contraceptive methods offered by each site varied considerably and each country used its own guidelines for use of contraceptives on ART. Contraceptives were not provided by the MTCT-Plus Initiative, although condoms were provided for all patients.
Statistical and analytical methods
Our primary analytic hypothesis was that women who used hormonal contraceptives would have a more rapid rate of HIV disease progression than those who did not. We also hypothesized that progesterone was the causative agent, and therefore, that women with progesterone-only exposure would fare worse than those exposed to combined estrogen–progesterone methods. We thus used the following categories to describe contraceptive exposure:
1. Implants/injectables: available data did not differentiate between the two; both are progesterone-only exposures.
2. Oral contraceptive pills (OCPs): although most oral contraceptives are combined preparations, some sites prescribed a progesterone-only ‘mini pill’ to women during their initial months of breastfeeding. Available data did not allow us to discriminate between progesterone-only and combined oral contraceptives, but anecdotal reports from sites indicate that the majority of exposure in this category was to combination regimens.
3. No exposure to exogenous hormones, which included women who reported using no contraception along with those who used nonhormonal methods (intrauterine device, diaphragm, sterilization and lactational amenorrhea).
Condom use was frequently reported in this cohort and condoms were used both alone and in conjunction with other methods. Women who reported condom use alone were categorized as having a nonhormonal exposure. Women who used condoms in conjunction with another method were categorized according to the other method. In our multivariable analyses, we treated condom use as an independent variable and controlled for it.
Women were eligible for this analysis if they were between 15 and 54 years of age and not on or eligible for ART. Because of the normal physiologic hormonal perturbations and hemodilution associated with pregnancy, we excluded women who were currently pregnant or within 90 days of delivery but allowed them to enter the analysis cohort after this period through left-censoring. Women who became pregnant during follow-up were right-censored at the time the pregnancy was first noted in the clinical record. Women who were lost to follow-up or withdrawn were censored at their last visit date. Baseline CD4+ cell count and WHO stage were defined as the measurement closest to the date of enrollment in MTCT-Plus Initiative or, for women who enrolled antepartum, the measurement closest to the date of becoming eligible for this analysis (e.g., 3 months postpartum).
For the mortality outcome, the follow-up time was censored when ART was initiated. The primary outcome of HIV disease progression for the women was defined as either becoming eligible for ART or death. ART eligibility was defined according to the programmatic CD4+ cell count and WHO disease stage criteria cited above. Women who started ART for any reason were assumed to be eligible irrespective of their CD4+ cell count or WHO disease stage.
Statistical analyses were performed with SAS version 9.1.4 (SAS Institute, Cary, North Carolina, USA) and R software version 2.4.1 (http://www.r-project.org). When assessing baseline characteristics among analysis groups, we compared continuous variables with a nonparametric Wilcoxon rank sum test to evaluate differences in medians. We compared dichotomous and categorical variables with the Pearson chi-square test statistic.
We estimated hazard ratios for our definition of HIV disease progression outcome using Cox proportional hazards regression. We analyzed contraceptive exposure in the Cox models in two ways. In the initial method analysis, a given patient's contraceptive exposure was determined by the first recorded dispensation. Under this assumption, for example, a patient who initially reported using OCPs would be categorized as exposed to OCPs, even if the patient switched to another method at some point during follow-up. In the time-varying exposure analysis, a given patient's contraceptive exposure(s) was allowed to vary over time in the Cox model. Thus, an individual woman who switched among contraceptive methods was allowed to contribute person-time to both contraceptive exposure categories in the proportional hazards regression. In the multivariable analysis, we adjusted for initial condom usage, baseline CD4 cell count, BMI, WHO disease stage, age and site.
The outcomes we report were a part of a preplanned analysis and the definitions used were informed by prior work. The analysis plan demanded a P value of less than 0.01 for statistical significance in this study to allow for multiple comparisons. The conduct of the MTCT-Plus Initiative as a service delivery program with data collection for monitoring and evaluation purposes was approved by the Institutional Review Board of Columbia University.
Between 1 August 2002 and 31 December 2007, a total of 7846 women were enrolled in 13 MTCT-Plus programs. In total, 4109 (52%) women met eligibility criteria for this analysis and contributed 5911 person-years of follow-up. The median follow-up time contributed was 379 days [interquartile range (IQR) 121–833]. Reasons for exclusion from this analysis were that 2107 women were eligible for ART at their initial visit, 864 were enrolled while pregnant or within 90 days of delivery and did not have follow-up data beyond this period, 103 had no information on contraceptive exposure and 306 did not have a baseline CD4+ cell count that allowed assessment of HIV disease progression. An additional 357 patients could not be categorized due to inconclusive contraceptive information (Fig. 1).
The median age of women in the cohort was 27 (IQR 24–31) years, median parity was 2 (IQR 1–3) and 63% of women were married. Generally, the women were healthy with a median BMI of 23 (IQR 21–26) kg/m2 and a median CD4+ cell count of 429 (IQR 280–621) cells/μl, and 65% of the women were WHO disease stage I (Table 1).
At baseline, 3064 (75%) women reported using either no contraception or a nonhormonal method, whereas 823 (20%) reported using implants/injectables and 222 (5%) reported using OCPs. Contraceptive choices differed by site. The most common exposures other than hormonal contraception were condoms, lactational amenorrhea, female sterilization and no contraceptives (Fig. 2).
Women in the three initial exposure groups were largely similar with respect to clinical and demographic characteristics. Compared with women without initial hormonal exposure, women initially exposed to either implants/injectables or OCPs had higher BMIs, slightly higher hemoglobin levels, were less likely to be widowed and were less likely to report condom use; although statistically significant, none of these differences seemed clinically significant (Table 1). Women initiating implants/injectables were also more educated than women without initial hormonal exposure and less likely to have disclosed their HIV status (Table 1).
Switching among contraceptive methods
Women who commenced OCPs were much more likely to switch methods at least once compared with women starting no contraception/nonhormonal contraception or implants/injectables. Of the 222 women starting OCPs, 124 (56%) switched methods at least once with a median time to switch of 132 days (IQR 63–333); 291 of the 823 (36%) patients starting implants/injectables switched at least once with a median time to switch of 230 days (IQR 92–588). Of the 3064 women starting no contraception/nonhormonal contraception, 613 (20%) switched at least once for a median time to switch of 175 days (IQR 84–357). Of the 1785 women who reported no contraception use at baseline, 1170 (66%) adopted some contraceptive method during follow-up.
Composite outcome: HIV disease progression
We defined disease progression as becoming eligible for/starting ART or death. In the cohort of 3229 women who entered the analysis ineligible for ART, 944 (29%) met these criteria at some point during follow-up for a rate of 18.3 per 100 woman-years of follow-up [95% confidence interval (CI) 17.1–19.5]. Predictors for our primary outcome were WHO disease stage and CD4+ cell count (Table 2). In either crude or adjusted analysis, neither exposure to implants/injectables [adjusted hazard ratio (AHR) 1.0, 95% CI 0.8–1.1] nor exposure to OCPs (AHR 0.8, 95% CI 0.6–1.1) was associated with this outcome. Kaplan–Meier analyses indicated no difference among the three initial contraceptive exposures for the combined outcome (data not shown). In adjusted time-varying analyses, neither implants/injectables exposure (AHR 1.0, 95% CI 0.8–1.1) nor exposure to OCPs (AHR 0.8, 95% CI 0.6–1.1) was associated with accelerated HIV disease progression compared with women in the no/nonhormonal group (Table 2).
Individual components of composite outcome
Hormonal contraception (whether implants/injectables or OCPs) was not associated with becoming eligible for/starting ART. Among the 3229 women who were not eligible for ART upon entering the analysis cohort, 924 (29%) became eligible for/started ART during the study period (crude rate: 17.9 per 100 woman-years, 95% CI 16.8–19.1). In both crude and adjusted analyses, risk factors for becoming eligible or starting ART were CD4+ cell count between 200 and 350 cells/μl (AHR 5.3, 95% CI 4.6–6.1) and WHO disease stage II and III (AHR 1.6, 95% CI 1.3–1.9 and AHR 2.8, 95% CI 2.1–3.7, respectively) at baseline (Table 2).
During the study period, 63 (2%) women died over 5911 woman-years of follow-up (crude mortality rate: 1.1 per 100 woman-years, 95% CI 0.8–1.4). Exposure to hormonal contraception was not associated with death (Table 2). In both crude and adjusted analyses, death was strongly associated with low CD4+ cell count and advanced WHO disease stage at baseline. Compared with women with a CD4+ cell count of more than 350 cells/μl, women with baseline CD4+ cell count between 200 and 350 cells/μl and those with a CD4+ cell count of less than 200 cells/μl were more likely to die (AHR 3.3, 95% CI 1.7–6.3 and AHR 17, 95% CI 7.6–36, respectively). Compared with women who were WHO disease stage I at entry, women who were disease stage III and IV were more likely to die (AHR 3.3, 95% CI 1.5–7.2 and AHR 17, 95% CI 6.6–42.2, respectively). These results did not change in adjusted time-varying analysis (Table 2).
To investigate variations between MTCT-Plus Initiative programs, a separate time-varying analysis was conducted for each individual program, although site was a factor included in the analyses described above. These adjusted analyses used our primary outcome of HIV disease progression, in which an event was defined as either becoming eligible for ART or death. Forest plots of the hazard ratios comparing each hormonal contraceptive categorization to the patients using no/nonhormonal contraceptives are depicted in Fig. 3(a) and (b).
Notably, none of the sites showed a significantly increased risk of HIV disease progression comparing the implants/injectables group to the no hormones/nonhormonal group or comparing the OCP group to the no/nonhormonal group.
In this multicountry cohort of women enrolled in HIV care but not yet on ART, we observed no association between hormonal contraceptive use and HIV disease progression. Other factors known to be associated with disease progression, such as low CD4+ cell count, more advanced clinical stage and low BMI [10,11], were associated with HIV disease progression in this cohort.
Although there are a variety of data available on the impact of hormonal contraception on HIV acquisition, data on the impact of hormonal contraception on HIV disease progression are limited and not consistent. Two studies [7,12] suggest that hormonal contraception may accelerate HIV disease progression, whereas two others [13,14] did not demonstrate such an association. Encouragingly, this multicountry cohort analysis suggests that hormonal contraception does not accelerate HIV disease progression. Although there may be subtle interactions between hormonal contraception and HIV disease progression, we did not observe this in our large cohort.
A particular strength of our study is its multicountry nature. Women from the west, east and southern Africa are included, as well as a cohort of Thai women. Collectively, this represents a large diversity of patients, viral subtypes and contraceptive methods. In addition, the study includes by far the largest number of women in whom this association has been examined. Other strengths are that all the women were prospectively followed on a predetermined schedule and information regarding contraceptive use, as well as laboratory and clinical data, were collected using standardized forms.
A limitation of our data is the inability to disaggregate the individual contributions of injectable contraception and implants. With varying hormonal compositions, these two types of progesterone-only contraceptives may have different biologic effects that could have been masked when they are considered together. Progesterone contained in implants is released more slowly over time with more constant drug levels, whereas injectables are associated with much more fluctuation in drug levels and overall higher systemic hormone concentration . Additionally, women in our analysis were exposed to two different types of injectable contraception: Depo Provera (Pfizer, New York, New York, USA) (DMPA) and Noristerat (Bayer, Newbury, Berkshire, UK) (NET-EN). DMPA, given every 3 months, contains medroxyprogesterone acetate, whereas NET-EN contains norethisterone enanthate and is administered every 2 months. Medroxyprogesterone acetate has a strong affinity for glucocorticoid receptors, which are expressed by a variety of cells of the immune system, whereas norethisterone enanthate does not seem to have such an affinity . Medroxyprogesterone acetate has also been shown to suppress interleukin (IL)-1, IL-2, and IL-6, which are important cytokines in CD4+ cell proliferation [17,18]. An additional limitation of our analysis is the relatively short amount of follow-up time contributed by most women (median: 379 days). Even though the top quartile of follow-up exceeded 2 years, it is possible that longer term hormonal exposure could produce subtle effects on HIV disease progression that simply did not have time to manifest in this cohort.
The lack of association between hormonal contraceptive use and HIV disease progression in this large, prospective, multicountry cohort is reassuring. It is our view that clinicians should not hesitate in prescribing hormonal contraception to HIV-infected women who desire family planning. Of course, medical contraindication must always be considered including potential interactions with some antiretroviral drugs. The benefits of avoiding unplanned pregnancies (and the risks that attend them in resource-constrained settings) outweigh any theoretical risks of hormonal contraception-mediated progression of HIV disease. Although a great deal of recent attention has been paid to tackling HIV in low-income countries, efforts to improve access to wider varieties of contraception methods must continue for all women. In addition, further research is needed to explore the safety and efficacy of contraceptive methods in individuals with HIV infection.
We would like to acknowledge all of the participating MTCT-Plus Initiative staff and patients who worked tirelessly to make this study possible.
The MTCT-Plus Initiative is funded through grants from the following philanthropic foundations: Bill and Melinda Gates Foundation, William and Flora Hewlett Foundation, David and Lucile Packard Foundation, Robert Wood Johnson Foundation, Henry J. Kaiser Family Foundation, John D. and Catherine T. MacArthur Foundation, Rockefeller Foundation, and Starr Foundation. Additional support is provided by the United States Agency for International Development.
Participating MTCT-Plus programs are Formation Sanitaire Urbaine de Yopougon-Attié, Abidjan, Cote d'Ivoire; Nyanza Provincial General Hospital, Kisumu, Kenya; Moi Hospital/Mosoriot Rural Health Center, Eldoret, Kenya; Day Hospitals in Beira and Chimoio, Mozambique; Treatment and Research AIDS Center/Kigali Health Centres, Kigali, Rwanda; Perinatal HIV Research Unit, Chris Hani Baragwanath Hospital, Soweto, South Africa; Langa Clinic, City of Cape Town Health Department, Cape Town, South Africa; Ekuphileni Clinic/Cato Manor, University of KwaZulu-Natal, Durban, South Africa; Mulago Hospital, Kampala, Uganda; St. Francis Nsambya Hospital, Kampala, Uganda; and Chelstone and Mtendere District Health Clinics, Lusaka, Zambia.
Conflicts of interest: None.
1. Population Reference Bureau. Family planning worldwide 2008 data sheet. Washington, DC; 2008.
2. Joint United Nations Programme on HIV/AIDS (UNAIDS). Report on the global AIDS epidemic. Geneva: WHO; 2008.
3. Norton M. New evidence on birth spacing: promising findings for improving newborn, infant, child, and maternal health. Int J Gynaecol Obstet 2005; 89:S1–S6.
4. Baeten JM, Lavreys L, Overbaugh J. The influence of hormonal contraception use on HIV-1 transmission and disease progression. Clin Infect Dis 2007; 45:360–369.
5. Stringer EM, Antonsen E. Hormonal contraception and HIV disease progression. Clin Infect Dis 2008; 47:945–951.
6. Marx PA, Spira AI, Gettie A, Dailey PJ, Veazey RS, Lackner AA, et al
. Progesterone implants enhance SIV vaginal transmission and early virus load. Nat Med 1996; 2:1084–1089.
7. Baeten JM, Lavreys L, Sagar M, Kreiss JK, Richardson BA, Chohan B, et al
. Effect of contraceptive methods on natural history of HIV: studies from the Mombasa cohort. J Acquir Immune Defic Syndr 2005; 38(Suppl 1):S18–S21.
8. Rabkin M, El-Sadr W, Katzenstein DA, Mukherjee J, Masur H, Mugyenyi P, et al
. Antiretroviral treatment in resource-poor settings: clinical research priorities. Lancet 2002; 360:1503–1505.
9. Rabkin M, El Sadr W. Saving mothers, saving families: the MTCT-Plus initiative – case study
. Perspectives and practice in antiretroviral treatment
. Geneva: WHO; 2003.
10. Langford SE, Ananworanich J, Cooper DA. Predictors of disease progression in HIV infection: a review. AIDS Res Ther 2007; 4:11.
11. Lapadula G, Torti C, Maggiolo F, Casari S, Suter F, Minoli L, et al
. Predictors of clinical progression among HIV-1-positive patients starting HAART with CD4+
T-cell counts > or =200 cells/mm3
. Antivir Ther 2007; 12:941–947.
12. Stringer EM, Kaseba C, Levy J, Sinkala M, Goldenberg RL, Chi BH, et al
. A randomized trial of the intrauterine contraceptive device vs hormonal contraception in women who are infected with the human immunodeficiency virus. Am J Obstet Gynecol 2007; 197:144.e1–144.e8.
13. Cejtin HE, Jacobson L, Springer G, Watts D, Levine A, Greenblatt R, et al
. Effect of hormonal contraceptive use on plasma HIV-1 RNA levels among HIV-infected women. AIDS 2003; 17:1702–1704.
14. Richardson BA, Otieno PA, Mbori-Ngacha D, Overbaugh J, Farquhar C, John-Stewart GC. Hormonal contraception and HIV-1 disease progression among postpartum Kenyan women. AIDS 2007; 21:749–753.
15. Garza-Flores J, Rodriguez V, Perez-Palacios G, Virutamasen P, Tang-Keow P, Konsayreepong R, et al
. A multicentered pharmacokinetic, pharmacodynamic study of once-a-month injectable contraceptives. Different doses of HRP112 and of DepoProvera. World Health Organization Task Force on Long-acting Systemic Agents for Fertility Regulation. Contraception 1987; 36:441–457.
16. Koubovec D, Ronacher K, Stubsrud E, Louw A, Hapgood JP. Synthetic progestins used in HRT have different glucocorticoid agonist properties. Mol Cell Endocrinol 2005; 242:23–32.
17. Bamberger CM, Else T, Bamberger AM, Beil FU, Schulte HM. Dissociative glucocorticoid activity of medroxyprogesterone acetate in normal human lymphocytes. J Clin Endocrinol Metab 1999; 84:4055–4061.
18. Enomoto LM, Kloberdanz KJ, Mack DG, Elizabeth D, Weinberg A. Ex vivo effect of estrogen and progesterone compared with dexamethasone on cell-mediated immunity of HIV-infected and uninfected subjects. J Acquir Immune Defic Syndr 2007; 45:137–143.