Escalating HIV infections among women in the developing world require alternative women-controlled prevention options, because many are unable to negotiate condom use with their partners as a result of socioeconomic, cultural, and other reasons.1 In recent years, research has focused on topical microbicides, which can be discreetly applied to the vagina before sexual intercourse to prevent HIV and other sexually transmitted infections (STIs). A number of vaginal microbicides are currently under investigation, but several phases of clinical testing are necessary to determine the safety and effectiveness of each product.2 Large-scale efficacy trials are expensive and must be conducted in countries with high HIV incidences, which are also more likely to have limited resources and capacity.
Sub-Saharan Africa is home to nearly 29 million of the 40 million people estimated to be living with HIV worldwide.3 With the highest HIV prevalence and incidence among women and their male partners in this region, the region is an optimal choice for establishing phase 2b/III microbicide or other HIV prevention trial sites.
Overall, the design of phase 2b/III microbicide trials depends on 4 key parameters: HIV incidence, participant accrual and retention rates, duration of the follow-up period, and minimal attrition rates as a result of pregnancy in the targeted population. The HIV Prevention Trials Network (HPTN) conducted a site preparedness study to evaluate these parameters in 4 potential phase 2/IIb microbicide trial sites.
Study Design and Recruitment
This prospective cohort study was conducted at potential phase 2/IIb microbicide trial sites in Durban and the district of Hlabisa, South Africa; Lusaka, Zambia; and Moshi, Tanzania. All sites followed their required community approval and entry processes before recruitment in the trial. Women in Durban and Hlabisa were recruited from surrounding communities through community meetings and door-to-door recruitment. In Lusaka, women were recruited from the lower middle class socioeconomic communities of Chilenje and Kamwala through community meetings and local family planning clinics. In Moshi, 50% of the women were recruited from a family planning clinic representing a more general population and 50% were recruited from a pool of high-risk bar and hotel workers.4 Female fieldworkers visited these bars and hotels, where they provided the women brief information about the study and invited them to the study clinic for more detailed information. Through this study, all sites had the opportunity to enter the respective communities and assess their interest in HIV prevention research and, in particular, future vaginal microbicide trials.
Women were eligible for this study if they met the following criteria: sexually active, HIV-negative at screening, willingness to provide written consent and follow study procedures, absence of clinical STIs (women with treatable STIs were treated before enrollment), not pregnant with intention to maintain their nonpregnant status, and residing in and around the study area for a minimum of 1 year. Women with gynecologic abnormalities were excluded from the study. Women who provided written consent were screened for HIV and STIs and provided with safe sex counseling and condoms. Women were reimbursed for time, travel, and refreshments. The reimbursement amounts per visit were US $10 for Durban and Hlabisa, US $4 to $8 (depending on length of visit) for Lusaka, and US $1.50 plus reimbursement of transport costs for Moshi.
Participants were followed monthly for 12 months. At each monthly visit, participants completed an interval medical/menstrual history and underwent a pregnancy test. HIV/STI risk reduction counseling and condoms were provided. At each quarterly visit, additional HIV/STI tests and assessments were performed. Because this was a planning study for a future large-scale trial, the procedures for HIV and pregnancy testing were conducted according to our future phase 2/IIb design. A face-to-face structured interview was administered every month to ascertain HIV risk behaviors such as condom use, number of partners, and practice of anal sex. Appropriate local language and acceptable behavioral questions were tested in preparation for the large-scale trial.
Pelvic examinations were performed to detect vaginal infections. Papanicolaou (PAP) smears were collected for all women at enrollment. Colposcopic evaluations were conducted at selected sites in accordance with the CONRAD/World Health Organization (WHO) manual for standardization of colposcopy for the evaluation of vaginal products.5 Women with genital abnormalities detected through colposcopy were not enrolled. In Durban, Hlabisa, and Lusaka, urine samples were collected for diagnosis of Neisseria gonorrhoeae (NG) and Chlamydia trachomatis (CT) using the BDProbe Tec ET assay (Becton Dickinson, Sparks, MD). In Moshi, genital swabs were collected for detection of NG by Gram stain and culture and CT by means of an enzyme-linked immunosorbent assay (ELISA) test. One of the limitations of the study was that in Moshi, we were only able to perform the less sensitive ELISA test for NG and CT because of lack of availability of more sensitive BDProbe Tec assay tests. At all sites, wet preparation assays were performed for detection of bacterial vaginosis (BV), Trichomonas vaginalis (TV), and Candida. Vaginal swabs collected at quarterly visit pelvic examinations were sent to the HPTN Central Laboratory based at Johns Hopkins University (Baltimore, MD) for Gram stain testing. BV diagnosis was based on a Nugent score value of 7 to 10 during microscopic assessment. Women were only treated for symptomatic vaginal infections.
Venous blood samples were collected for HIV and syphilis testing. Before enrollment, HIV testing was performed using 2 rapid ELISA tests and a confirmatory Western blot test if the initial samples were positive. Follow-up HIV testing was performed using a rapid ELISA test and confirming a positive rapid test result with a Western blot from the same specimen and from an additional blood draw. All follow-up HIV-positive results were confirmed by the HPTN Central Laboratory.
The Tanzanian and Zambian sites were granted ethical approval by their local ethics committees to recruit women aged 16 years and older. At the South African sites, approval was granted to recruit women aged 18 years and older. All women with treatable STIs received counseling and treatment as per local guidelines. Male partners were invited to the study site for treatment or were given contact cards to access STI treatment. Women who were identified as HIV-positive at screening were referred to local health care facilities for care and support. Women who seroconverted during the trial remained in the study and were provided with ongoing counseling and referral to local health care facilities for further care at the end of the study.
The protocol and informed consent forms were approved by the respective ethics committees at each site.
A sample size of 240 per site was set to achieve a half-width of approximately 3.2% for the exact 95% confidence interval (CI) for the HIV incidence (based on a 12-month retention rate of 95% and a true rate of 5.1% HIV seroincidence). Descriptive statistics were used to summarize demographic and sexual behavior characteristics. Changes from baseline to follow-up were assessed by the McNemar test for dichotomous variables and by a t test or Wilcoxon signed rank test, whichever was appropriate, for continuous variables. CIs were based on the binomial distribution for prevalence rates and the Poisson distribution for incidence rates. Because women with STIs and other genital infections at all sites received effective treatment, chlamydial infection, gonorrhea, and trichomoniasis infections detected during the follow-up period were considered new infections. For syphilis, because the test result can remain positive for a few weeks after successful treatment, we considered as new syphilis infections during follow-up only the positive test results that were preceded by a negative test result at the previous visit. Infections of BV and Candida were calculated as point prevalence at each quarterly visit. All descriptive and inferential statistical analyses were performed by the Statistical Center for HIV/AIDS Research and Prevention (SCHARP) using SAS 9.1.3 (SAS Institute, Cary, NC) and StatXact 6.2 (Cytel Software Corporation, Cambridge, MA).
Key community leaders (generally men), political and traditional, were approached to solicit their interest in the proposed research in their community. With approval, community meetings were held to engage the broader community on the aims and objectives of the study. The community included men, women, and representatives from nongovernmental organizations and community-based organizations. Health service providers in the area were also invited because referral for health care would be through these service providers. Communities were supportive of the research undertaken. In Hlabisa, for example, there was a shortage of home-based care providers and lay counselors. Through the research program, we trained 150 individuals from the community as lay counselors and caregivers and linked them to health service providers for future training and community support. Community advisory boards (CABs) were set up and trained on HIV prevention trials, microbicide research, and ethics of research. All respective communities expressed interest in the present study, because many were unaware of the accurate HIV estimates in their area. Community feedback meetings were held regularly to provide updates on the study and to address any concerns they may have had.
A total of 84 staff members were employed in the study as a whole. Staff complement included study leaders and coordinators, HIV counselors, clinical trial nurses, clinicians, laboratory personnel, data managers, data coordinators, field staff for recruitment, and community liaison officers for maintaining regular contact with the community. Quality control and quality assurance officers were trained to monitor data quality internally: case report forms (CRFs) and laboratory tests. Study staff at each of the sites were provided training on human subjects research protection, good clinical practice (GCP), development of HIV/STI education and community training material, data collection (completion of CRFs as per GCP), development of HIV/STI counseling scripts, and opportunities to update their counseling skills through formal courses and training. Nurses were trained on speculum examinations and clinical management. For those sites conducting colposcopy, clinical staff received colposcopy training. Laboratory staff members were trained to conduct HIV rapid tests, pregnancy tests, and good laboratory practice (GLP). Staff members were also trained to address data queries with short turn-around time. Project leaders or managers with a master's or doctoral degree were trained to be clinical trialists and given the opportunity to lead, take ownership of the project, and manage staff at the clinic site under the supervision of the study principal investigator.
A total of 2099 women were approached for screening, and 2044 (97%) of these women were willing, able, and of legal age to sign the informed consent form. Of the 2044 women who consented to screening, 960 (47%) women were eligible for the study (2 of whom did not enroll because of accrual completion). The main reasons for ineligibility included (1) HIV infection from self-report or HIV testing (n = 619 [30%]), (2) screening procedures that exceeded the 30-day screening window as per protocol (n = 161 [8%]), (3) sexually active criterion unmet (1 act of vaginal intercourse in the past 3 months, n = 95 [5%]), (4) diagnosed with a current STI requiring treatment according to WHO guidelines (n = 94 [5%]), and (5) pregnant (n = 75 [4%]). Some women who were not enrolled had multiple reasons for ineligibility. Of the 960 women eligible to be enrolled, 38 (4%) had unsuccessful screening attempts before eligibility, with 23 (61%) of those 38 women initially ineligible because of screening procedures that exceeded the 30-day screening window. A total of 958 women were enrolled.
Overall, 2.2 participants were screened for every participant enrolled in the study, and this ratio was remarkably similar across the sites: 2.2:1 in Hlabisa and Moshi and 2.3:1 in Durban and Lusaka.
Accrual and Retention Rates
Nine hundred fifty-eight women were enrolled in the study between June 2003 and October 2004: 240 each from Durban and Moshi and 239 each from Hlabisa and Lusaka. The South African sites reached their targeted accrual within 7 months (6.3 months in Durban and 6.7 months in Hlabisa). The Lusaka and Moshi sites reached their targets in 7.1 months and 8.3 months, respectively. The highest 12-month participant retention rate was recorded in Durban (97%). Retention rates were also high at Hlabisa (94%), Lusaka (93%), and Moshi (86%). Nineteen percent of the women lost at the Moshi site were from the select group of bar and hotel workers compared with 7% of women from family planning clinics (P < 0.001).
The mean overall age of enrolled participants was 28.6 years (Table 1), ranging from 27.0 to 32.2 years at each site. A significantly higher proportion of women were married in Lusaka and Moshi (62%) than in Durban or Hlabisa (30%; P < 0.01). Correspondingly, higher proportions of women were unmarried and living with their partners in Durban and Hlabisa (69%) than in Lusaka or Moshi (21%; P < 0.01). Polygamous husbands or partners were more common at the 2 South African sites (26%) than in Lusaka and Moshi (9%; P < 0.01).
Given the wealth of data generated, a separate report on behavioral characteristics is currently being written. For this article, a brief overview of behavioral differences at enrollment and 12-month follow-up of 883 participants is provided (Table 2).
At baseline, women reported an average of 1.7 acts of vaginal intercourse in the past week (interquartile range [IQR]: 0.0 to 2.0 acts). The mean number of vaginal sex acts dropped significantly between baseline and month 12 to 1.5 acts in the past week (IQR: 0.0 to 2.0 acts; P < 0.01).
Despite condom counseling, reported condom use actually declined during the study. Compared with enrollment, by the month 12 visit, the proportion of women reporting no condom use had increased to 72%, with a corresponding decrease in condom use with every sex act to 20% of women (P = 0.02; see Table 2).
Of the 958 women enrolled, 800 (84%) reported using at least 1 type of contraception at baseline, including injectables (n = 313 [33%]), male condoms (n = 310 [32%]), pills (n = 148 [15%]), natural methods (n = 130 [14%]), and surgical sterilization (n = 61 [6%]). There were 17 (2%) women using female condoms and 1% or less of women reporting use of the following methods: Norplant (Wyeth Pharmaceuticals, Madison, NJ), intrauterine devices, traditional methods, and the diaphragm. By month 12, there was a 10% decrease (P < 0.01) in the proportion of women reporting natural methods. The proportion of women who reported using male condoms as a contraceptive method declined by 6% at month 12 (P < 0.01), and there was a slight but significant 2% decrease in reported use of oral contraceptive pills (P = 0.04). Some women reported multiple forms of contraception, for example, male condoms and natural methods.
Pregnancy Prevalence and Incidence
Pregnancy prevalence during screening was 4%, and the pregnancy incidence rate in the study as a whole was 20.2 per 100 women-years (wy), which varied from a low of 17.4 per 100 wy in Lusaka to a high of 22.7 per 100 wy in Moshi. The pregnancy incidence per 100 wy was highest among women aged 18 to 25 years in Durban (28.4), Lusaka (22.9), and Moshi (32.3). In Hlabisa, however, more women aged 26 to 35 years (35.8) became pregnant. Pregnancy incidence per 100 wy was also highest among women reporting baseline contraceptive use of male condoms (28.6), natural methods (26.7), pills (23.0), female condoms (21.5), injectables (11.4), and surgical sterilization (3.3). None of the women who became pregnant were infected with HIV during the study. Of the 188 pregnancies detected, 105 (56%) resulted in a pregnancy outcome during study follow-up. The mean length of completed pregnancies was 114 days (range: 16 to 291 days), suggesting that a substantial proportion of pregnancies were not carried to term. Table 3 shows the frequency and duration of pregnancies during follow-up.
Baseline HIV/STI Prevalence
The overall HIV prevalence rate was 32.5% (95% CI: 30.4 to 34.8; Table 4). The highest prevalence was in Durban, and the lowest was in Hlabisa. The highest HIV prevalence was in the 26- to 35-year-old age group (see Table 4).
The overall prevalences of NG, CT, syphilis, and TV were 4.7% (95% CI: 3.7 to 5.9), 5.4% (95% CI: 4.3 to 6.6), 6.1% (95% CI: 5.0 to 7.4), and 6.4% (95% CI: 5.0 to 8.1), respectively. Table 5 shows the prevalence of STIs and other reproductive tract infections (RTIs) at the 4 sites. The limited detection of NG and CT in Moshi could be attributed to low sensitivity of the test used.
Follow-Up STI/HIV Incidence
Overall STI incidence rates per 100 wy were 11.1 (95% CI: 9.0 to 13.3), 12.5 (95% CI: 9.9 to 15.1), 7.5 (95% CI: 5.2 to 9.8), and 31.9 (95% CI: 27.5 to 36.3) for NG, CT, syphilis, and TV, respectively (Table 6). The incidence of NG in Durban was nearly double that of Hlabisa and approximately 3 and 6 times more than in Lusaka and Moshi, respectively. A similar pattern was seen in infections with CT.
The overall incidence of HIV was 3.8 per 100 wy (95% CI: 2.6 to 5.2; see Table 6). The HIV incidence rate in the 3 southern African sites was significantly higher than that in the eastern African site (P = 0.03). HIV incidence was 5.0 per 100 wy (95% CI: 2.6 to 8.8) in Durban and 6.0 per 100 wy (95% CI: 3.3 to 10.0) in Hlabisa, both of which were about twice the rate recorded in Lusaka, 2.6 per 100 wy (95% CI: 1.0 to 5.7). The incidence of HIV in Moshi was half the rate in Lusaka, 1.3 per 100 wy (95% CI: 0.3 to 3.8). Overall, HIV incidence was highest among women aged 18 to 25 years, and this was true for all sites except Moshi (Table 7). Women with incident HIV infection were at a significantly increased risk of other STIs such as CT (odds ratio [OR] = 5.6, 95% CI: 2.2 to 14.7; P < 0.01) and NG (OR = 4.1, 95% CI: 1.2 to 13.9; P = 0.02). Also, women with BV had a significantly increased hazard rate (HR) for HIV infection (HR = 4.8, 95% CI: 1.9 to 12.1; P < 0.01). Participants in the South African sites were 3.5 times more likely to be at risk of HIV than participants in Moshi and Lusaka (P = 0.03). Detailed analysis of STIs as a risk factor for HIV is to be published elsewhere. Point prevalence of RTIs at quarterly visits based on Nugent criteria are provided in Table 8. The prevalence of BV and Candida did not decline significantly during the course of the study.
This longitudinal prospective cohort study underscores the value of site preparedness studies for planning of large-scale HIV prevention clinical trials. Similar preparatory studies have been reported from Uganda (Rakai),6,7 Tanzania (Mbeya),8 and Kenya (Kericho).9 The present study allowed trial sponsors and researchers to gain insight into community support for HIV prevention research and, in particular, future microbicide trials, demographics of the study population and their sexual behavioral characteristics, and HIV/STI prevalence and incidence in the targeted population. We were able to identify appropriate community entry and approval processes and identify recruitment areas for future trials. The high retention rate suggests that retention strategies used in this preparedness activity were good and likely to be useful for future trials.10
In addition, the study provided an opportunity to test and finalize standard operating procedures, case report forms, and culturally acceptable behavior questions to be used in future trials. More importantly, it built clinical trial research capacity at sites, including development of young scientists, counselors, field workers, and community liaison officers, to conduct large-scale efficacy trials.
The speed of accrual and retention at all sites was commendable, mainly because of the great emphasis placed on community partnerships and involvement by the HPTN. The comparatively lower retention rate at Moshi could be attributed to the mobility of high-risk women4 recruited in the study, because 93% of participants from family planning clinics were retained after 12 months of follow-up compared with 81% of bar and hotel workers (P < 0.01).
Despite monthly HIV risk reduction counseling and promotion of condoms, we were less successful in inducing safer sexual practices over the course of the study. It was apparent that condom use and its negotiation of use were still a challenge for women.11 Univariate analysis showed that there was no increased risk of HIV for participants who failed to use a condom with every vaginal sex act in past 3 months (HR = 1.00, 95% CI: 0.4 to 2.7). One continuing concern among some HIV prevention experts is that high condom counseling and condom provision in prevention trials may have an impact on our ability to determine the efficacy of the intervention. Our results suggest that this effect may be minimal in populations in which condom use is unlikely to be high. This observational study demonstrates that even without other interventions such as a microbicide, sustaining condom use was still a challenge; hence, the issue of condom migration in presence of less effective HIV prevention options may be of little concern. It is possible that condom use may increase in trials with long follow-up periods, although this was not the case in the present study.12 We believe that couple counseling may be the key to enhancing condom use and that counseling women alone may not be ideal for ensuring safe sex behavior.
The high HIV and STI prevalence in the southern African sites was as expected.3,13-18 The prevalence of HIV in the 2 sites in the KwaZulu-Natal province of South Africa corresponds with the current HIV infection rates reported among pregnant women.3,13 In Zambia, the HIV prevalence rate among pregnant women aged 15 to 44 years has been estimated to be 18% to 20%, whereas our findings from Lusaka among nonpregnant women suggest a much higher prevalence rate than previously reported.16,20 In Moshi, the data are consistent with the current estimates of HIV among antenatal women in Tanzania.20-22
The overall HIV incidence of 3.8 per 100 wy was lower than estimated but was nonetheless high enough to warrant development of clinical trial sites for HIV prevention in southern Africa. We had based our sample size on HIV incidence estimates from each site on the basis of what was available. The lower than expected HIV incidence suggests that future trial designs need to be conservative in the calculation of sample size to avoid altering designs in the midst of trial implementation. This could be through increasing duration of trial follow-up, increasing participant numbers at existing sites, or developing additional sites. These activities are likely to be expensive and time-consuming; thus, determining accurate HIV incidence is crucial in developing microbicide effectiveness trial sites. The lower than expected HIV incidence at the Moshi site facilitated the decision not to use Moshi as a site in the HPTN 035 Phase 2/IIb trial.23
The recent closure of trial sites in Ghana and Nigeria because of lower than predicted HIV incidence (≤2%) highlights the value of preparatory studies or other accurate predictor tests for determining HIV incidence before implementation of large-scale efficacy trials.24 Recent data from these studies suggest that participants in HIV prevention trials are likely to have lower HIV incidence compared with those not participating. The authors attribute this to frequent condom counseling, treatment of STIs, and risk reduction counseling provided in HIV prevention trials (Family Health International, personal communication, September 2007). Planning for future trials needs to determine the potential impact of the standard-of-care package provided through trial sites on HIV incidence rates observed in the trial.
For efficacy trials, it is vital that we target appropriate populations. Our data suggest that the incidence of HIV is higher among women in the general population in southern Africa. In Moshi, HIV seroconversion, albeit a small number, was only seen among the hotel and bar workers. It may be possible that an inappropriate population was targeted in Tanzania.
Pregnancy is another issue of great concern in microbicide trials. In most of these trials, pregnancy is an exclusion criterion and the study product is discontinued once pregnancy is confirmed because of unknown teratogenic effects of the product. A high incidence of pregnancy, and therefore interruption of study intervention, results in loss of statistical power to determine product effectiveness. HIV prevention trials must consider the following issues with regard to pregnancy: (1) restrict enrollment to participants who are willing to use reliable contraception for the study duration; (2) monitor contraception provision and adherence at study sites; (3) determine the frequency of pregnancy testing; (4) potentially allow for use of the product during pregnancy provided that sufficient data are available on the teratogenic effects of the product; and (5) for long-term follow-up studies, allow women to resume postpartum product use. Recent data from Durban suggest that contraceptive uptake at prevention trial sites is high25 and that it may be feasible to provide and monitor contraception use during the trial.
The high incidence of pregnancy in this study reflects to a certain extent the accuracy of reported contraceptive use by the participants. We speculate that the high pregnancy incidence may be reflective of incorrect reporting of contraceptive use and nonadherence to contraceptive methods. Furthermore, many women who became pregnant reported condom use as a means of contraception. Other contraceptive use also declined in the study, which supports the suggestion that reported contraceptive use at enrollment may not be accurate.
It is possible that the high frequency of pregnancy testing (monthly) may have detected subclinical pregnancies, which do not reach full term. Our data clearly demonstrate that not all positive pregnancies reached full-term during the course of the study. One of the limitations with the data is that we do not know if loss of pregnancies observed was attributable to chemical pregnancies or to spontaneous or elective abortion. It is thus advisable for the field to assess whether frequent (monthly) pregnancy testing is essential for prevention trials, because we may be inadvertently discontinuing women from product use, thus affecting the power of the study to show efficacy. Furthermore, obtaining pregnancy data in preparatory studies not only advises on sample size for large-scale trials but provides an estimate of participant attrition attributable to pregnancy.
An important finding from this study is that the very group that is at risk of infection (18 to 25 years old) is also the group in which pregnancy incidence was highest. For effectiveness trials, we need to recruit women at risk of HIV infection. It would be detrimental to the effect size of the intervention to lose such high-risk women through pregnancy.
Our data on HIV and pregnancy are disconcerting, because many young women in this setting are unlikely to opt for HIV prevention as a preference over motherhood. Community and recruitment efforts need to integrate HIV and pregnancy prevention education and counseling. Furthermore, given that fertility is highly regarded in many countries in which HIV infection risk is high, involvement and counseling of the male partners of women in the trial may be an option to reduce pregnancy. In addition, couple counseling may allow for more sustainable education on HIV prevention, use of condoms, adherence to product use, and pregnancy prevention as part of the overall recruitment effort.
Ethical concerns arising from provision of contraception at trial sites include posttrial sustainability. To address this, site preparedness activities should include an assessment of available health care in the community and partnerships with health care providers should be created before trial implementation so that trial sites only provide care that is sustainable.
An obvious ethical imperative that was highlighted by this study was the issue of care for women who were diagnosed with HIV at screening and, more importantly, those who seroconverted after enrollment. Partnerships with health care providers need to be developed before initiation of research to ensure adequate referral and sustainability of HIV prevention and treatment care.
In conclusion, this clinical trial site preparedness study has provided much needed insight into several issues currently debated among clinical trial investigators. Based on our findings, we recommend that before development of clinical trial sites, preparedness studies are undertaken to ascertain community interest and support for the trial to ensure that accrual and retention targets are met. Innovative strategies to promote behavior change are urgently needed. Our findings suggest that high pregnancy rates are likely to have an impact on the power of phase 2b/III trials because of large numbers of women discontinuing product use. We recommend contraception provision as part of care offered at clinical trial sites, provided that posttrial sustainability is ensured. Our data show that HIV incidence is highest among women aged between 18 and 25 years and declines after the age of 35 years. We recommend that cohorts establish an upper age limit capped at 35 years for microbicide trials in our setting. This may not apply to all populations. Furthermore, investigators and sponsors should ensure that strategies for facilitation of HIV care and treatment are arranged before trial implementation for HIV seroconverters. Finally, southern Africa seems to be eminently suited for large-scale HIV prevention effectiveness trials because of high HIV incidence rates in the general population.
The authors express their thanks to the following members of the HTPN 055 Study Team: Medical Research Council, HIV Prevention Research Unit, Durban, South Africa: Roshini Govinden, Nozizwe Dladla-Qwabe, Kodwa Mpepho, Yages Singh, and Neetha Morar; Zambia: Sten Vermund and Muzala Kapina Kanyanga; Moshi, Tanzania: Sarah Chiduo, Longin Barongo, Noel Sam, George Seage, Wafaie Fawzi, Lori Miller, Trong Ao, Tara Daley, Ireen Kiwelu, Basidi Bamba, Estomih Mduma, Msafiri Swai, Charles Mnbando, Margareth Maya, and Lwice Macha; SCHARP: Stacie Kentop, Pete McDonnell, Geoff Minerbo, Karisse Román, and Kayla Stratton; Family Health International: Lydia Soto-Torres and Roberta Black; Central: Anne Coletti; and Division of AIDS Laboratory, US National Institute of Allergy and Infectious Diseases: Edward Livant, Estelle Piwowar-Manning, and Sharon Hillier.
1. Lederman MM, Offord RE, Hartley O. Microbicides and other topical strategies to prevent vaginal transmission of HIV. Nat Rev Immunol
2. Mauck C, Rosenberg Z, Van Damme L. Recommendations for the clinical development of topical microbicides: an update. AIDS
3. UNAIDS Joint United Program on HIV/AIDS. AIDS Epidemic Update
. Geneva, Switzerland: UNAIDS Joint United Program on HIV/AIDS; December 2005.
4. Kapiga SH, Sam NE, Shao JF, et al. HIV-1 epidemic among female bar and hotel workers in northern Tanzania: risk factors and opportunities for prevention. J Acquir Immune Defic Syndr
5. CONRAD/WHO. CONRAD/WHO manual for the standardization of colposcopy for the evaluation of vaginal products, update 2004. Available at: http://www.conrad.org/colposcopy.htm
. Accessed December 2006.
6. Wawer MJ, Serwadda D, Gray RH, et al. Trends in HIV-1 prevalence may not reflect trends in incidence in mature epidemics: data from the Rakai population-based cohort, Uganda. AIDS
7. Ahmed S, Lulato T, Wawer M, et al. HIV incidence and sexually transmitted disease prevalence associated with condom use: a population study in Rakai, Uganda. AIDS
8. Riedner G, Hoffmann O, Rusizoka M, et al. Decline in sexually transmitted infection prevalence and HIV incidence in female bar workers attending prevention and care services in Mbeya Region, Tanzania. AIDS
9. Sateren WB, Foglia G, Renzullo P, et al. Epidemiology of HIV-1 infection in agricultural plantation residents in Kericho, Kenya: preparation for vaccine feasibility studies. J Acquir Immune Defic Syndr
10. Govinden R, Ramjee G. Successful strategies for participant accrual and retention [abstract OB13]. Presented at: Microbicides 2006 Conference; 2006; Cape Town.
11. Global Campaign for Microbicides. October 2006. Fact sheet 4. Gender equality in AIDS prevention: why we need prevention options for women. Available at: http://www.global-campiagn.org
. Accessed December 2006.
12. Padian N, Van der Straten A, Ramjee G, et al. Diaphragm and lubricant gel for prevention of HIV acquisition in Southern African women: a randomized controlled trial. Available at: wwwthelancet.com
. Accessed July 13, 2007.
13. National Department of Health. National HIV and Syphilis Antenatal Sero-prevalence Survey in South Africa
. Pretoria, Gauteng. 2005.
14. Buve A, Carael M, Hayes RJ, et al. Multicentre study on factors determining differences in rate of spread of HIV in sub-Saharan Africa: methods and prevalence of HIV infection. AIDS
. 2001;15(Suppl 4):S5-S14.
15. Odei EA, Calleja JMG, Boerma JT. HIV prevalence and trends in sub-Saharan Africa: no decline and large subregional differences. Lancet
16. Dzekedzeke K, Fylkesnes K. Reducing uncertainties in global HIV prevalence estimates: the case of Zambia. BMC Public Health
17. Udjo EO. Estimation of mortality from vital registrations in South Africa. Curr HIV Research
18. Smith A. HIV/AIDS in KwaZulu-Natal and South Africa. AIDS Anal Afr
19. Kaiser Family Foundation. 2006. Global health facts: Zambia. Available at: http://www.globalhealthfacts.org
. Accessed December 2006.
20. Kaiser Family Foundation. 2006. Global health facts: Tanzania
. Available at: http://www.globalhealthfacts.org
. Accessed December 2006.
21. Somi GR, Matee MI, Swai RO, et al. Estimating and projecting HIV prevalence and AIDS deaths in Tanzania using antenatal surveillance data. BMC Public Health
22. Malima KIY, Olsen BE, Matee MI, et al. The silent HIV epidemic among pregnant women within rural Northern Tanzania. BMC Public Health
23. HIV Prevention Trials Network. HPTN035 protocol. Available at: http://www.hptn.org/research_studies/HPTN035StudyDocuments.htm#Protocol
. Accessed December 2006.
24. Feldblum P, Perterson L. FHI phase II/III study of C31G in two Nigerian cities and in Ghana [panel discussion talk]. Presented at: Microbicides 2006 Conference; 2006; Cape Town.
25. Moodley J, Morar NS, Ramjee G. Contraception uptake among microbicide clinical trial participants in Kwazulu-Natal, South Africa [abstract OB14]. Presented at: Microbicides 2006 Conference; 2006; Cape Town.