Sub-Saharan Africa has the highest prevalence and incidence of HIV-1 infection in the world [1,2], and the region also has high fertility rates, with approximately 14 million pregnancies annually considered unintended . As an example, in Kenya, HIV-1 seroprevalence is 7.1%, 60% of HIV-1-infected adults are women, the total fertility rate is 4.7 births per woman, and it has been estimated that there is a 25% unmet need for family planning among married women [4–6]. Studies [7,8] in Kenya have reported low contraceptive uptake among HIV-1-seropositive women. For women with or at risk for HIV-1 infection, dual-contraceptive use (i.e., use of an effective noncondom contraceptive method in conjunction with condoms) is the optimal strategy to prevent both pregnancy and HIV-1 transmission. Moreover, for HIV-1-seropositive women, prevention of unwanted pregnancies through contraceptive use is an important adjunctive, cost-effective, method to reduce mother-to-child transmission of HIV-1 [6,9–11]. Unfortunately, family planning and HIV-1 care programs are often not integrated, and few examples of successful programs to increase contraceptive uptake within the context of HIV-1 prevention have been described [8–12]. In addition, few reports have described the importance and opportunity to involve male partners in family planning decision-making.
Recent population surveys and mathematical modeling studies have found that stable HIV-1-serodiscordant couples – that is, in which one partner is HIV-1 infected and the other is HIV-1 uninfected – may be responsible for a majority of new HIV-1 transmissions in Africa . Thus, HIV-1-serodiscordant couples should be considered a high-priority population for both HIV-1 prevention efforts and family planning interventions. Here, we describe a successful multipronged intervention to increase contraceptive use among HIV-1-serodiscordant couples enrolled in an HIV-1 prevention trial in Thika, Kenya.
Population and procedures
The Partners in Prevention HSV/HIV Transmission Study  was a phase III, randomized, placebo-controlled, clinical trial of daily acyclovir herpes simplex virus type 2 (HSV-2) suppressive therapy provided to HIV-1-seropositive members within heterosexual HIV-1-serodiscordant couples. The primary aim of the trial was to assess the effect of acyclovir on HIV-1 transmission to the seronegative partners. The trial enrolled participants at 14 African sites in East and southern Africa between December 2004 and May 2007, with individual study sites initiating enrollments in a staged fashion between December 2004 and June 2006. The last study follow-up visits were completed by October 2008. The present study reports on contraceptive use and pregnancy incidence from the Thika, Kenya site, where the supplementary family planning intervention was implemented and compares findings with composite results from the other three Kenyan study sites (located in Eldoret, Kisumu, and Nairobi). All Kenyan sites recruited from urban, peri-urban, and rural areas in a radius up to 150 km from the study site.
For the Partners in Prevention HSV/HIV Transmission Study, couples were eligible if they reported at least three episodes of vaginal intercourse during the 3 months prior to enrollment. HIV-1-seropositive partners were at least 18 years of age, seropositive for HSV-2, had a CD4 cell count of at least 250 cells/μl and no history of AIDS-defining conditions, and were not on antiretroviral therapy (ART). HIV-1-infected women who were pregnant at the time of study screening were excluded from enrollment. HIV-1-seronegative partners were at least 18 years of age.
Study visits occurred monthly for HIV-1-seropositive participants and quarterly for HIV-1-seronegative participants, for a maximum of 24 months (until trial closure). Visits included individual and couple risk reduction counseling, with provision of free male and female condoms. HIV-1-infected women who became pregnant were discontinued from study medication during the pregnancy but remained in the study. HIV-1-infected participants who met national guidelines for initiation of ART during follow-up were referred to local HIV-1 care clinics.
All participants provided written informed consent for study participation. The protocol was approved by institutional review boards at the University of Washington, Kenyatta National Hospital, Kenya Medical Research Institute, and Moi University.
Family planning interventions at the Thika study site
At all study sites, contraceptive methods were offered on-site or by referral on voluntary basis as a part of routine clinical care. During its first year of operation (June 2006–May 2007), the Thika site offered injectable depot medroxyprogesterone acetate and oral contraceptive pills free-of-charge at the research clinic, whereas other methods were offered by referral. In June 2007, site staff instituted a combination of interventions (see the list given below) to increase dual contraceptive uptake (i.e., condoms along with another effective contraceptive method) in part to minimize the time off study drug among HIV-1-seropositive female participants, a protocol requirement, and to prevent unplanned pregnancies among both HIV-1-seropositive and seronegative participants in Thika, Kenya.
1. Training of clinical and counseling staff on contraceptive methods, including practical demonstrations and discussions of common myths and barriers to contraceptive use.
2. Provision of free contraceptive methods [oral contraceptive pills, injectables, implants, and intrauterine devices (IUDs)] to study participants at the research clinic.
3. Use of contraceptive appointment cards with clear dates for renewal of the time-dependent methods (e.g., injectable depot medroxyprogesterone acetate) to avoid lapses in hormonal contraception.
4. Designation of one staff member to ensure that staff received ongoing training in contraceptive counseling and that sufficient contraceptive supplies were always available on-site.
5. Introduction of checklists in chart notes to remind staff to discuss and provide contraceptive methods during study visits.
6. Weekly meetings with clinicians, counselors, and pharmacy staff to share their experiences discussing contraception with study participants.
7. Discussion of challenges to contraceptive uptake with study couples individually and in psychosocial support groups. The insights from these sessions were reported back to the study team to strengthen contraceptive messages for other couples.
8. Involvement of male partners during contraceptive counseling sessions during routine study visits.
9. Review of unintended pregnancies among HIV-1-seropositive women to identify reasons why these pregnancies were not avoided.
The Kisumu and Eldoret sites did not provide noncondom methods at the research clinic and referred women to nearby medical facilities, whereas the Nairobi site offered injectable depot medroxyprogesterone acetate and oral contraceptive pills free-of-charge at the research clinic.
Data were analyzed using SAS version 9.2 (SAS Institute, Cary, North Carolina, USA). Analyses were conducted using data collected at quarterly study visits. The primary study measure was self-reported use of contraception other than condoms, defined as current use of an IUD, surgical method, injectable, implantable, or oral hormonal methods. Contraceptive use and pregnancy incidence were compared between time periods (before and after June 2007, when the multipronged intervention was initiated). Comparisons within the Thika site study population, within the other Kenyan sites, and between Thika and the other Kenyan sites were performed. Use of noncondom contraception over time was analyzed using generalized estimating equations. Pregnancy incidence was analyzed via Andersen–Gill proportional hazards modeling to account for pregnancy as a potential recurrent event. Adjustment for potential confounding by age and number of children did not result in meaningful changes; thus, only unadjusted rates are presented.
Between June 2006 and April 2007, the Thika site enrolled 213 HIV-1-serodiscordant couples, including 159 (74.7%) in whom the female partner was the HIV-1-seropositive partner. The median age of HIV-1-seropositive and seronegative women was 30.2 [interquartile range (IQR) 26.3–34.0] and 30.9 years (IQR 26.0–37.7), respectively, and 89.3 and 90.8%, respectively, had at least one living child (Table 1). Most women were married to their study partner (91.2 and 94.6%, respectively). Unprotected sex during the month prior to enrollment was reported by 25.2% of HIV-1-seropositive women and 25.9% of HIV-1-seronegative women. The median duration of follow-up was 18 months (IQR 15–24) for HIV-1-seropositive women and 18 months (IQR 18–24) for HIV-1-seronegative women, and HIV-1-seropositive and seronegative women each completed a median of six visits (IQR 5–7 and 6–7, respectively). In Thika, 66.9% of visits by HIV-1-seropositive and 68.7% of visits by HIV-1-seronegative women occurred after June 2007.
A total of 1216 couples were enrolled by the other three Kenyan sites (Eldoret, Kisumu, and Nairobi) from December 2004 to May 2007. Of these, 832 (68.4%) were couples in whom the HIV-1 seropositive was woman. Enrollment characteristics, including age, marital status, and frequency of unprotected sex, were similar for female participants from the three Kenyan sites as compared with participants at the Thika site (data not shown), except that HIV-1-seropositive participants were slightly younger at the other Kenyan sites (median 28.2 years, IQR 24.0–33.0) as compared with the Thika site (Mann–Whitney P = 0.002). Participants from the other Kenyan sites had longer median follow-up (24 months for both HIV-1-seropositive and seronegative women) than participants at the Thika site, reflecting that the Thika site was the last site to initiate the study.
Between June 2006 and May 2007, HIV-1-seropositive women in Thika used noncondom contraception at 31.5% of visits (117/372), whereas after June 2007 (i.e., between June 2007 and September 2008), use rose to 64.7% of all visits (486/751) [odds ratio (OR) 4.0, 95% confidence interval (CI) 3.0–5.3] (Fig. 1). Among HIV-1-seronegative women in Thika, there was also a significant increase in use of noncondom methods of contraception, from 28.6% of visits (34/119) prior to June 2007 to 46.7% of visits (122/261) after June 2007 (OR 2.2, 95% CI 1.4–3.5). The most common contraceptive method used was injectable methods (reported at 31.7% of visits by HIV-1-seropositive women and 23.2% of visits by HIV-1-seronegative women) followed by oral methods (7.2 and 9.2%, respectively), tubal ligation or other surgical procedure (6.8 and 5.0%, respectively), implants (5.8 and 1.3%, respectively), and IUDs (2.2 and 2.4%, respectively).
In the other Kenyan sites, noncondom contraceptive use did not change substantially in the time periods before and after June 2007. In these sites, among HIV-1-seropositive women, contraceptive use increased from 15.6 (523/3344) to 22.3% (604/2712) of visits (OR 1.5, 95% CI 1.3–1.9), whereas among HIV-1-seronegative women, use decreased slightly from 13.6 (187/1373) to 12.7% (176/1389) of visits (OR 0.9, 95% CI 0.7–1.3). Thus, after June 2007, noncondom contraceptive use was substantially higher among HIV-1-seropositive women from Thika (64.7%) as compared with the other Kenyan sites (22.3%) (OR 6.4, 95% CI 4.6–8.9). Similarly, after June 2007, noncondom contraceptive use among HIV-1-seronegative women was higher at Thika (46.7%) than at the other Kenyan sites (12.7%) (OR 6.0, 95% CI 3.4–10.7).
Reported condom use was high throughout the entire study follow-up period for both HIV-1-seropositive and seronegative women from all sites (87.8 and 91.4% of sex acts, respectively).
Among HIV-1-seropositive women in Thika, pregnancy incidence declined after June 2007, from 13.5 to 8.7 per 100 woman-years (seven events in 52 woman-years to 16 events in 183 woman-years, respectively). Among HIV-1-seronegative women in Thika, pregnancy incidence also declined from 21.1 to 11.0 per 100 woman-years (three events in 14 woman-years to seven events in 64 woman-years, respectively). However, at the other Kenyan sites, pregnancy incidence increased during this same time period: from 16.8 to 21.9 per 100 woman-years among HIV-1-seropositive women (101 events in 602 woman-years to 140 events in 640 woman-years, respectively) and from 14.6 to 19.7 per 100 woman-years (34 events in 233 woman-years to 64 events in 326 woman-years, respectively) among HIV-1-seronegative women. Within the Thika study population (i.e., including both HIV-1-seropositive and seronegative women in the analysis), pregnancy incidence was significantly lower after June 2007 as compared with before June 2007 (hazard ratio 0.2, 95% CI 0.1–0.6). In addition, after June 2007, pregnancy incidence was significantly lower at Thika as compared with the other Kenyan sites (hazard ratio 0.5, 95% CI 0.3–0.8), whereas there was no statistically significant difference in pregnancy incidence for Thika versus the other sites for the time period before June 2007 (hazard ratio 1.0, 95% CI 0.5–1.9).
Within the setting of a biomedical HIV-1 prevention trial among Kenyan HIV-1-serodiscordant couples, we implemented a multipronged strategy focused on increasing contraceptive uptake. Our efforts substantially increased dual-contraceptive use – that is, use of condoms along with another effective method – among both HIV-1-seropositive and seronegative women in HIV-1-serodiscordant partnerships. Pregnancy incidence declined during this same time period. Given the low utilization of both barrier and nonbarrier contraceptive services in many African populations, particularly within stable sexual partnerships, successful initiatives promoting family planning and dual-contraceptive use may inform programmatic efforts to prevent unplanned pregnancies and decrease HIV-1 risk.
Although our family planning intervention was not randomized, we were able to evaluate the uptake of dual contraceptive methods and pregnancy incidence by a time series comparison from the Thika site, using data from the year prior to the intervention. In addition, we compared Thika data with that from the other three Kenyan sites that participated in this same clinical trial and which provided ‘standard’ contraceptive services and referrals to community family planning programs. In both comparisons, we found that dual-contraception uptake was substantially higher among both the HIV-1-seropositive and seronegative women in the Thika cohort during the intervention period. Major components of the Thika family planning intervention were provision and enhanced promotion of free contraceptive services at the research clinic and staff training in family planning methods and counseling. Our experience is that both of these components were instrumental in achieving high rates of contraceptive use by study participants. Prior studies have shown that accessibility is an important factor to increase contraceptive use , and our results illustrated a potential opportunity to provide contraception in a research clinic, which women are attending for other purposes. In addition, many healthcare workers lack knowledge of effective family planning options, particularly for women with HIV-1 [10,11], and we found that staff training in contraceptive methods and counseling for HIV-1-serodiscordant couples was an essential part of our intervention.
A potentially valuable role for male involvement in family planning provision has been suggested by previous studies [7,14]. We followed both members of HIV-1-serodiscordant couples as part of our clinical trial procedures, and thus were able to integrate male partner participation in contraceptive counseling into routine quarterly study visits. Condom use was high among these HIV-1-serodiscordant couples, for whom counseling of the HIV-1-positive partners was done monthly and counseling of both partners was done quarterly; approximately 90% of sex acts were reported to be protected by condoms.
In addition to a significant increase in contraceptive use for both HIV-1-seropositive and seronegative women during the intervention period at the Thika site, pregnancy incidence was also lower in Thika during the intervention period, both compared with the preintervention period and compared with the other Kenyan sites during the intervention period. We did not assess fertility intentions in this cohort and future studies among HIV-1-serodiscordant couples should assess fertility planning and contraceptive decision-making. Interventions to increase contraceptive uptake may benefit from follow-up visits to focus on factors that lead to early discontinuation of dual contraception, including myths, side effects, and changing fertility desires.
In our population, HIV-1-seropositive women had a higher contraceptive uptake as compared with HIV-1-seronegative women. There are several possible explanations for this finding. First, HIV-1-seropositive women were seen monthly, whereas HIV-1-seronegative women were seen quarterly, and thus HIV-1-seropositive women had greater opportunity to discuss family planning intentions, receive counseling about contraceptive options, and identify acceptable contraceptive methods. Second, HIV-1-seropositive women may have had a greater desire to avoid unwanted pregnancies in order to prevent possible HIV-1 transmission to a child. Finally, our clinical trial protocol required discontinuation of study drug for HIV-1-seropositive women who became pregnant, which may have been an incentive for study staff to focus family planning messages more strongly towards HIV-1-seropositive women.
Many HIV-1 prevention clinical trials are currently being conducted among women at risk for HIV-1 and women in HIV-1-serodiscordant couples in sub-Saharan Africa . Biomedical prevention trials, such as those testing oral antiretroviral medications and candidate vaginal microbicides, frequently require that women who become pregnant during follow-up should be removed from the investigational product or discontinued from the study . High number of pregnancies can result in a substantial proportion of study follow-up being unexposed to study product, thereby reducing the statistical power for the trial to demonstrate efficacy. Our results are particularly relevant to clinical trial populations, in which periodic study visits can permit opportunities for ongoing counseling regarding family planning options.
Strengths of this study include the relatively large sample size, longitudinal follow-up, and inclusion of both HIV-1-seropositive and HIV-1-seronegative women. We were able to compare Thika with the other Kenyan sites, where the same clinical trial procedures were performed during the same time period. A limitation was that our intervention package was comprised of a combination of services and thus it is not possible to evaluate the impact of each separately. Our intervention was conducted within a clinical trial setting, which limits the generalizability of our findings to other family planning and HIV-1 prevention and care programs with fewer resources and less frequent follow-up.
In summary, we found a substantial increase in dual-contraceptive uptake, and a concomitant decrease in incident pregnancy, among women in HIV-1-serodiscordant partnerships after an intervention to integrate family planning promotion into a clinical trial setting. Our results suggest that a multipronged intervention, including training of counselors and clinical staff, counseling sessions that include the male partner, and provision of free contraceptive services on-site, can lead to a significant increase in uptake of contraceptive methods without decreasing condom use. This approach may be able to be adapted from the research clinic setting to HIV-1 service delivery settings.
The Partners in Prevention HSV/HIV Transmission Study was funded by the Bill and Melinda Gates Foundation (grant ID #26469 to the University of Washington).
We thank the principal investigators and study coordinators at the Kenyan sites of the Partners in Prevention HSV/HIV Transmission Study for use of site data for this analysis (Drs Elizabeth Bukusi, Craig Cohen, Carey Farquhar, Kenneth Fife, Grace John-Stewart, James Kiarie, Harrison Tamooh, Edwin Were, and Cosmas Apaka and Josephine Odoyo) for use of their site data. We are grateful to Paul Musingila for assistance with data management.
Conflicts of interest: None.
1. Dunkle KL, Stephenson R, Karita E, Chomba E, Kayitenkore K, Vwalika C, et al
. New heterosexually transmitted HIV infections in married or cohabiting couples in urban Zambia and Rwanda: an analysis of survey and clinical data. Lancet 2008; 371:2183–2191.
2. Richey C, Setty V. Family planning choices for women with HIV. Popul Rep Ser L 2007; 15:5–6.
3. Hubacher D, Mavranezouli I, McGinn E. Unintended pregnancy in sub-Saharan Africa: magnitude of the problem and potential role of contraceptive implants to alleviate it. Contraception 2008; 78:73–78.
4. National AIDS and STI Control Program (NASCOP), Ministry of Health (MOH), Kenya. Kenya AIDS Indicator Survey 2007: preliminary report
. Nairobi, Kenya; 2008.
5. Central Bureau of Statistics (CBS) [Kenya], Ministry of Health (MOH) [Kenya], and ORC Macro. Kenya Demographic and Health Survey 2003
. Calverton, Maryland, USA: CBS, MOH and ORC Macro; 2004.
6. Cohen SA. Hiding in plain sight: the role of contraception in preventing HIV. Guttmacher Policy Rev 2008; 11:2–5.
7. Bii SC, Otieno-Nyunya B, Siika A, Rotich JK. Family planning and safer sex practices among HIV infected women receiving prevention of mother-to-child transmission services at Kitale District Hospital. East Afr Med J 2008; 85:46–50.
8. Mutiso SM, Kinuthia J, Qureshi Z. Contraceptive use among HIV infected women attending comprehensive care center. East Afr Med J 2008; 85:171–177.
9. Mark KE, Meinzen-Derr J, Stephenson R, Haworth A, Ahmed Y, Duncan D, et al
. Contraception among HIV concordant and discordant couples in Zambia: a randomized controlled trial. J Women's Health (Larchmt) 2007; 16:1200–1210.
10. Wilcher R, Petruney T, Reynolds HW, Cates W. From effectiveness to impact: contraception as an HIV prevention intervention. Sex Transm Inf 2008; 84(Suppl 2):ii54–ii60.
11. Reynolds HW, Janowitz B, Wilcher R, Cates W Jr. Contraception to prevent HIV-positive births: current contribution and potential cost-savings in PEPFAR countries. Sex Transm Inf 2008; 84(Suppl 2):ii49–ii55.
12. Shears KH. Family planning and HIV service integration. Potential synergies are recognized. Network 2004; 23:4–8.
13. Lingappa JR, Kahle E, Mugo N, Mujugira A, Magaret A, Baeten J, et al
. Characteristics of HIV-1 discordant couples enrolled in a trial of HSV-2 suppression to reduce HIV transmission: the Partners Study. PLoS ONE 2009; 4:e5272.
14. Homsy J, Bunnell R, Moore D, King R, Malaba S, Nakityo R, et al
. Reproductive intentions and outcomes among women on antiretroviral therapy in rural Uganda: a prospective study. PLoS ONE 2009; 4:e4149.
15. Lingappa JR, Lambdin B, Bukusi EA, Ngure K, Kavuma L, Inambao M, et al
. Regional differences in prevalence of HIV-1 discordance in Africa and enrollment of HIV-1 discordant couples into an HIV-1 prevention trial. PLoS ONE 2008; 3:e1411.
16. Institute of Medicine (IOM). Methodological challenges in biomedical HIV prevention trials
. Washington, District of Columbia, USA: The National Academies Press; 2008.