Table 1.
Description of all newly included studies (published since 15 January 2014), ordered by publication year, for systematic review update on use of various hormonal contraceptive methods among women at risk of HIV acquisition.
| First author, publication year, location | Design, purpose, period of data collection | Number enrolled, description of population | Results (point estimate [adjusted, where available] and 95% CIs) | Met criteria for being considered ‘informative but with important limitations’? |
| Direct evidence from HC versus no-HC studies | ||||
| Kapiga 2013 [24], Tanzania | Cohort; to assess feasibility, retention, and appropriateness of population for future HIV prevention trials. Recruitment 2008–2010 | 2229 women working in hotels, restaurants, bars, guesthouses or shops selling traditionally-brewed beer, or food-sellers at makeshift facilities in northern Tanzania | OCs and other non-DMPA HC (+/− condoms) adjIRR: 0.68 (0.23–2.04); DMPA (+/− condoms) adjIRR: 1.63 (0.75–3.52) | No, did not control for condom use (and selection process for inclusion of covariates in the statistical model was not based on assessing HC-HIV acquisition) |
| Dube 2014 [22], Mozambique (Beira) | Cohort; to measure HIV incidence in women at higher risk of HIV and assess the feasibility of recruiting and retaining them as research participants. Data collection 2009–2012 | 411 women enrolled (387 contributed follow-up data). HIV-seronegative women aged 18–35 with at least two sexual partners in the past month, recruited from schools and places where women typically meet potential sexual partners | OCs/injectables adjHR: 1.2 (0.4–4.0) | No, unclear measurement of exposure (did not distinguish between HC methods) |
| Crook 2014 [21], South Africa, Uganda, Tanzania, Zambia | Cohort; RCT to assess safety and efficacy of Pro2000 candidate microbicide for HIV prevention. Enrollment 2005–2008 | 9385 HIV-negative women aged 16+ (18+ in South Africa and Zambia) from a range of settings. In Uganda (9% of analytic population), women were recruited as serodiscordant couples | Time-updated covariate model including time-varying exposure: DMPA adjHR 1.45 (1.09–1.93); NET-EN adjHR: 1.20 (0.84–1.69); OCs (likely COCs) adjHR: 0.90 (0.63–1.26); IPW model; DMPA adjHR: 1.49 (1.06–2.08); NET-EN adjHR: 1.31 (0.86–1.99); OC (likely COCs) adjHR: 1.00 (0.62–1.61); Additional models in Table 4 of [21] | Yes |
| Feldblum 2014 [23], Mozambique (Chókwè) | Cohort; to measure HIV incidence prospectively, and to assess the site's ability to enroll and retain the cohort. Data collection 2010-2012 | Enrolled 479 HIV-seronegative women aged 18–35, who were sexually active in the last month, willing to adhere to study visit requirements, and planning to reside in Chókwè for duration of study, recruited from community venues with young women who engage in risky sexual behavior | OCs/injectables Crude HR: 0.4 (0.1–1.3) | No, unclear measurement of exposure (did not distinguish between HC methods; no time-varying HC exposure) and no adjustment for condom use |
| Wall 2015 [27], Zambia (Lusaka) | Cohort; prospective study of serodiscordant couples. Data collection 1994–2012 | In this analysis, 1393 M+ F− serodiscordant couples recruited from couples voluntary counseling and testing services | HIV infections genetically linked to cohabitating male partner; Implants adjHR: 0.96 (0.29–3.14); DMPA adjHR: 1.34 (0.85–2.12); COCs adjHR: 1.39 (0.90–2.15); Linked and unlinked infections; Implants adjHR: 1.08 (0.53–2.20); DMPA adjHR: 1.19 (0.81–1.73); COCs adjHR: 1.29 (0.92–1.80) | Yes |
| McKinnon 2015 [25], Kenya (Nairobi) | Cohort; to estimate HIV incidence & risk factors in a program catering to FSW. Enrollment 2008–2011 | Enrolled 3951 HIV-uninfected FSWs from broths, bars, clubs, and the street, as well as providing cards to clinic attendees to distribute to their peers | DMPA adjHR: 5.12 (1.98–13.22) | No, unclear measurement of exposure (no time-varying HC exposure; reference group contains unclear number of women using other HC) |
| Balkus 2016 [29], Malawi, South Africa, USA, Zambia, Zimbabwe | Cohort; RCT to assess safety and efficacy of BufferGel (ReProtect Inc, Baltimore, Maryland, USA) and Pro2000 (Indevus Pharmaceuticals, Lexington, Massachusetts, USA) versus placebo or no gel. Enrollment 2005–2008 | Enrolled 3099 HIV-uninfected, nonpregnant women aged 18 and older who were sexually active | Injectables adjHR: 1.17 (0.70–1.96); OCs adjHR: 0.76 (0.37–1.55) | Yes |
| Morrison 2015 [26,38], IPD meta-analysis (and subanalysis of seven databases), Kenya, Tanzania, Uganda, South Africa | IPD meta-analysis of prospective studies | Full IPD meta-analysis included data on 37 124 sexually active women across 18 datasets; We focus on information from a subanalysis of seven studies previously unpublished studies, to avoid double-counting of component studies that are already included in our review | Full IPD meta-analysis [26]: COC adjHR: 1.07 (0.91–1.25); DMPA adjHR: 1.52 (1.27–1.82); NET-EN adjHR: 1.27 (0.99–1.61); Subanalysis of seven previously unpublished studies (two-stage random effects model) [38]; COC adjHR: 0.79 (0.38–1.64); DMPA adjHR: 1.69 (1.02–2.78); NET-EN adjHR: 1.58 (0.66–3.79) | Yes |
| Byrne 2016 [28] South Africa | Cohort; FRESH study – to understand mucosal immune factors associated with HIV acquisition risk. 2012–2015 | Included data on 432 HIV-uninfected women aged 18–23 recruited by referral from community organizations or from community outreach | Injectables adjHR: 2.93 (1.09–7.86) | No, unclear measurement of exposure (failure to include time-varying exposure information; some women included in the injectable group did not consistently use injectables, some in comparison group used DMPA during follow-up but were considered nonusers). No adjustment for condom use, either at baseline or over time, although authors conducted a Fisher's exact test (P = 0.1539) to assess for differences at baseline in condom use between comparison groups |
| Indirect evidence from head-to-head studies | ||||
| Noguchi 2015 [30], South Africa | Cohort; to investigate the safety and efficacy of three formulations of tenofovir for HIV prevention (VOICE trial). Enrollment and follow-up 2009–2012 | 5029 non-HIV-infected, sexually active, nonpregnant, nonbreastfeeding women without curable genitourinary infections or abnormal renal, hematological, or hepatic functions willing to use effective contraception enrolled in RCT (952 excluded from non-South Africa sites, 936 excluded for not meeting inclusion criteria) | DMPA versus NET-EN adjHR = 1.41 (1.06–1.89) | Yes |
| Morrison 2015 [26] IPD meta-analysis | IPD meta-analysis of prospective studies | Authors performed a subanalysis assessing direct comparisons between HC methods among studies with pertinent data; number of included women not provided | DMPA versus NET-EN adjHR: 1.32 (1.08–1.61) (based on IPD meta-analysis of data from the following studies: [17,21,34,35,37,47,51,55,56]); DMPA versus COC adjHR: 1.43 (1.23–1.67) (based on IPD meta-analysis of data from the following studies: [17,21,24,32–37,39,42,48,51,53,55,56,60]); NET-EN versus COC adjHR: 1.30 (0.99–1.71) (based on IPD meta-analysis of data from the following studies: [17,21,34,35,37,51,55,56]) | Yes |
Note: Please refer to 2014 systematic review for detail on previously included studies [1]. adjHR, adjusted hazard ratio; adjIRR, adjusted incidence rate ratio; CI, confidence intervals; COCs, combined oral contraceptive pills; DMPA, depot medroxyprogesteone acetate; FSW, female sex worker; HC, hormonal contraception; HIV, human immunodeficiency virus; HR, hazard ratio; IPD, individual participant data; IPW, inverse probability weighted; NET-EN, norethisterone enanthate; OCs, oral contraceptive pills; POPs, progestin-only pills; RCT Randomized controlled trial.