The idea of a product that women could apply vaginally to protect themselves against HIV and other sexually transmitted infection (STIs) has held great appeal for over two decades. Recent disappointing findings in effectiveness trials have triggered intense debate and discussion as to the future direction of the field, particularly in regards to category of agent to be tested and the way that agents should be prioritized for clinical investigation .
The first agent to be seriously considered as a candidate vaginal microbicide was nonoxynol-9, a detergent-based spermicide that had anti-HIV activity in vitro and in animal models. Initial studies suggested that nonoxynol-9 was well tolerated [2–8], although there was also some indication that it caused genital irritation  and epithelial disruption . Subsequent phase III studies failed to demonstrate any benefit [11–13], and clinical research into nonoxynol-9 was ultimately halted when it was found to cause an increase in the risk of HIV infection when applied frequently , leading to a recommendation in 2001 by the World Health Organization against its use as a microbicide . These findings have been comprehensively summarized in recent reviews [16,17].
The discovery that nonoxynol-9 actually increased the risk of HIV infection, the very outcome it was intended to prevent, presented the field of microbicide development with a major challenge, as ultimate safety was now only demonstrated by the same large-scale trials that were needed to prove its effectiveness. Over 30 other products have now been investigated clinically , including six that have reached effectiveness trials. At least 10 products are under clinical investigation, including three in ongoing effectiveness trials . Evaluation ceased on a number of products, either due to the results of effectiveness trials [19–23] or other considerations.
Clinical trials have used a wide range of toxicity measures [24–26] to assess a microbicide's safety before entering phase III trials. The emphasis has been on local endpoints that indicate some form of disturbance or harm to the genital tract, but systemic toxicity has also been assessed. In order to gain a better understanding of the findings from the safety investigations that have been carried out on products other than nonoxynol-9, we undertook a systematic review of published studies.
Participants and methods
Studies of candidate microbicides other than nonoxynol-9 were included in the review if they were published in a peer-reviewed journal, and reported on at least one safety outcome other than HIV infection in a randomized trial. Studies were excluded if they did not have a control group or the only comparison group was nonoynol-9, enrolled only male participants, or reported on the investigation of a product for the treatment of a genital infection. Also excluded were effectiveness studies, studies that reported only on the distribution of a microbicide in the vagina, and investigations in which the microbicide was used as coating for a condom or in conjunction with a diaphragm.
The literature search and data extraction were undertaken by two of the authors (I.Y.M. and I.M.P.). The methodology has been reported in detail elsewhere . We obtained the names of candidate microbicide products from abstracts of the Microbicides 2006 and 2008 conferences, the web site of the Alliance for Microbicide Development , and review papers [24–26]. Using the names of potential products as keywords, a combined search of the following electronic databases was conducted in September 2008: Cochrane Central Register of Controlled Trials (third quarter 2008); EMBASE (1980–2008, week 36); Ovid MEDLINE (1950 to August week 4, 2008). If the number of references obtained from the initial product name keyword search was greater than 1000, the search was narrowed by using [(the product name) AND (microbicide OR safety)].
The following keywords were used (those searches that were narrowed using microbicide OR safety are marked *): Acidform; Astroglide; benzalkonium chloride*; BufferGel; Carbomer 9748; carrageenan or Carraguard OR PC-213 OR PC-503 OR PC-515; cellulose acetate phthalate; cellulose sulfate OR cellulose sulphate OR Ushercell; chlorhexidine*; cyanovirin; cyclodextrin*; dextran sulphate OR dextran sulfate*; dextrin sulphate OR dextrin sulfate OR Emmelle; ethanol AND emollient; GEDA; Invisible Condom OR sodium lauryl sulfate OR sodium lauryl sulphate*; lactobacilli OR lactobacillus OR CTV-05*; lemon juice; lime juice; menfegol OR TS-88 OR Neo Sampoon; MIV-150 OR MIV150 OR MIV 150; Mucocept; nonoxynol-9 OR COL-1492; PC-815 OR PC815 OR PC 815; polystyrene sulfonate OR polystyrene sulphonate; Praneem; PRO2000 OR PRO-2000 OR PRO 2000 OR polynaphthalene sulfonate; PSC-Rantes OR PSC Rantes; Savvy OR C31G OR Glyminox; sodium dodecyl sulphate OR sodium dodecyl sulfate*; SPL7013 OR Vivagel; Surete; tenofovir OR PMPA*; terameprocol OR M4N OR EM-1421; TMC120 OR TMC-120 OR TMC 120 OR dapivirine; UC-781 or UC781 or UC 781; vinegar.
On the basis of the title and abstract, studies that potentially fulfilled the review criteria were reviewed in full. Citations within articles were used to identify further studies for potential inclusion. Each study was assigned a unique identification code. For each study that was found to satisfy the review criteria, we abstracted information on the design and findings. Initial review of the studies identified a wide range of safety endpoints. We categorized them into groups, based on biological criteria.
The terminology from the WHO manual for standardization of colposcopy  was used to define genital signs with intact epithelium (erythema, oedema, plaque/white finding, petechiae or ecchymosis), and genital signs with disrupted epithelium (peeling, ulceration, abrasion and laceration). Superficial and deep findings were not differentiated in the review. The category of urogenital symptoms was made up of irritation (burning or itching), discharge, bleeding and pain. Systemic safety laboratory assessments included haematological findings, renal and hepatic function, coagulation markers and electrolytes. Genital signs or symptoms not listed above were grouped as ‘other’ or ‘unspecified’ in each category. Other microbiology, including trichomoniasis, gonorrhoea, chlamydia, syphilis, candidiasis, bacterial vaginosis, vaginal microflora alterations and urinary tract infections (UTIs) were recorded separately. Bacterial vaginosis was defined as either a Nugent score above 6, a Gram stain consistent with bacterial vaginosis or clinical criteria consistent with bacterial vaginosis.
As studies reported findings in a variety of ways, it was necessary to take a pragmatic approach to their analysis. No distinction was made as to whether or not events were described as being related to the study product(s). Total events data were used if both related and unrelated events were available. We assessed endpoints based on findings at the end of treatment, not as a comparison with baseline findings in the same individual. Denominators were based on per protocol analyses unless specified otherwise. Where possible, data for each outcome were extracted as the number of treated women affected, and analysed by calculating a relative risk compared to the number in the corresponding control group. If there were no affected women in either the treated or control group, a continuity correction of 0.5 was applied to allow the calculation of a risk estimate. No calculation was made if there were no affected women in both treated and control groups.
In some studies, a particular variable was not measured or results were presented in a manner that prevented their assimilation into the definitions above. Also, in various studies, data were either not reported as number of women for single variable endpoints such as bacterial vaginosis, or symptoms or signs were reported as a composite endpoint such as urogenital symptoms. Rather than omit these studies from the analysis, they were included and analysed by the number of events. In one study (D4; Table 1), urogenital symptoms were reported per woman for intermenstrual bleeding and as number of events for other urogenital symptoms. In this case, data were combined and analysed as events, with one affected woman regarded as constituting one ‘event’. Another study (CS2; Table 1) reported urogenital symptoms and genital signs with intact epithelium per woman each as a combined endpoint rather than individual signs or symptoms, and here data were analysed per woman.
Statistical tests and confidence intervals (CIs) for the relative risks were based on chi-squared analysis when numbers of affected women were reported and on Poisson analysis for numbers of events. We used a meta-analysis approach to calculate risk estimates, combined across dose levels within studies and across studies, using either a fixed-effects or random-effects model (DerSimonian-Laird) . Heterogeneity among studies was assessed using Cochran's Q-statistic and quantified with the I 2-statistic, and considered statistically significant if P was less than 0.10. In the absence of significant heterogeneity, combined risk estimates were reported using a fixed-effects model. All other statistical tests were considered significant if P was less than 0.05, and analyses were performed using STATA version 10.0 (Stata Corporation, College Station, Texas, USA).
We identified 21 trials (reported in 24 publications) of 11 products that met the inclusion criteria [30–53] (Table 1). For seven products, there was a single study reported. All except two of the trials were published since 2000.
Vehicle product was the comparator in nine trials, KYJelly in five, methylcellulose gel in three and hydroxyethyl cellulose gel in two trials. Water and an unspecified placebo gel were each used in one study. Three studies also included no treatment arms (C2, D3 and D4) and three studies (CS1, PS1 and S1) included a nonoxynol-9 arm (not further described in this review). One study compared two different volumes of product (CS1) and eight studies compared multiple concentrations of product (D1, D3, D4, L1, P1, PS1, TM1 and UC1).
Only four of the 21 studies involved more than 2 weeks of product exposure. The median study duration was 14 days (range 5–365). A total of 1465 women, including 173 HIV-infected women, participated in the studies. The median sample size was 59 women (range 14–180).
Genital signs with disrupted epithelium
Twenty of the 21 studies reported this outcome. No results were available for T1. Five studies reported no affected women in any of the trial arms (CS2, D1, D3, L1, P2 and TM1) and one (D4) reported no events in either trial arm for two of the four comparisons. In study S1, there were almost four times more genital signs with disrupted epithelium detected among recipients of C31G [relative risk (RR) 3.67, 95% CI 1.02–13.14] than among placebo recipients. This result differs from the corresponding one reported in the published study, as our analysis combined both deep and superficial de-epithelialization. For UC781, the CI for the combined estimate of relative risk across dose levels was very wide, but its lower limit slightly exceeded 1.0 (RR 5.75, 95% CI 1.05–31.39). CIs were generally wide, reflecting the small study sizes and infrequency of this outcome (Fig. 1). One exception was the result of the combined analysis for Carraguard, which was based on a substantial number of events and resulted in a CI of 0.75 around the point estimate of 0.67.
Genital signs with intact epithelium
All studies apart from the terameprocol study (TI) reported this outcome. One study reported no affected women in either trial arms (CS5). One study (D4) reported no new events in either trial arm for two of four comparisons, and one study (L1) reported no new events in either trial arm for one of two comparisons. In study P1, the risk was significantly increased for women receiving 4% PRO2000 (RR 2.08, 95% CI 1.12–3.87), when events from each of the three post-treatment time points were added together. When only the results from the visit 1 week after discontinuing product were evaluated, there was no difference (data not shown). CIs were generally wide (Fig. 2), again with the exception of the estimate for Carraguard.
The combined estimate across dose levels and two studies indicated that there were significantly more genital findings with intact epithelium in PRO2000 recipients than in placebo recipients (RR 1.68, 95% CI 1.08–2.60; Fig. 2).
This outcome could not be assessed for the P2, C1, C2 and M1 studies, as the data on urogenital symptoms were not separately available for each study arm. There were events in all study arms of the other 17 studies apart from the control arm of T1 and the control arm of D1, but no significant differences between product and control groups. Based on the combined estimate across dose levels, recipients of UC781 were found to have reported urogenital symptoms significantly less frequently than those receiving HEC gel (RR 0.34, 95% CI 0.17–0.69) (Fig. 3).
As most studies (17) reported that bacterial vaginosis had been measured as an outcome and data were able to be analysed in 12, bacterial vaginosis was evaluated separately as an endpoint. Of the five that could not be analysed, two studies (C2 and CS2) did not report data separately for each study arm for bacterial vaginosis and three studies did not report results numerically (D1, M1 and P2). There were no events in any arm in two studies (D3 and UC1) and no events in two of three comparisons in TM1 and no events in one of two comparisons in L1. There was significantly less bacterial vaginosis among recipients of DS compared with no gel at the post-treatment visit (RR 0.57, 95% CI 0.35–0.91) in one study (D4; Fig. 4).
On the basis of the combined estimate, a lower incidence of bacterial vaginosis was found in DS recipients than placebo or no-gel groups (RR 0.61, 95% CI 0.42–0.88). However, this was a combined analysis of only two studies (D2 and D4) (Fig. 4).
Fifteen studies reported that one or more of the STIs trichomoniasis, gonorrhoea, chlamydia and syphilis had been measured as an outcome (data not shown). No new STIs occurred in nine studies. Results were not interpretable in five studies (CS4, CS5, D4, M1 and P2), as data were not separately reported for each study arm. In the one Carraguard trial (C1) where new STIs were reported, there was no difference between groups in the incidence of the four STIs combined (RR 0.56, 95% CI 0.16–1.90).
Sixteen trials measured candidiasis. Results were not interpretable in eight studies and there were new cases in eight studies (C1, C2, CS3, CS5, D2, D4, L1 and UC1). In these trials, candidiasis was uncommon (other than in C1) and did not differ in frequency between study arms.
Three trials assessed vaginal flora changes. In one CS trial (CS4), there was a decrease in polymorphonuclear cells and H2O2 positive lactobacilli in both arms. The use of CS was associated with a significant increase in the risk of Escherichia coli and Staphylococcus aureus detection and a decrease in the risk of ureaplasma. DS had no significant influence on vaginal flora and there was essentially no difference in levels of H2O2 producing lactobacilli before and after exposure to any study product (D1). At one site of the PRO2000 study (P1), there was a decrease in lactobacilli in product and placebo arms 2 h after use, which returned to baseline levels 7 days after dosing.
Seven studies specifically documented UTI. These were uncommon in all seven trials.
Other safety outcomes
Systemic laboratory parameters were assessed in 13 studies (data not shown). For four studies, the results could not be analysed, as data were not separately available for each study arm for systemic laboratory abnormalities. There were no laboratory toxicities reported in four studies (CS1, D1, PS1 and P1). In the remaining five studies (CS3, CS4, D2, D3 and UC1), there were few laboratory toxicities reported, with no significant differences between product and placebo users. All toxicities were mild except one severe activated partial thromboplastin time rise in a 1% DS user (D3).
Five studies assessed systemic product absorption (T1, TM1, P1, D1 and UC1). In three studies, there was no evidence of plasma absorption of product (T1, P1 and D1). There was measurable plasma TMC120 in 75% of women on day 7 among all product arms. Two women receiving 1% UC781 had detectable but extremely low plasma levels of UC781 (<2.5 ng/ml).
This systematic review of published randomized controlled safety trials of candidate microbicides other than nonoxynol-9 has found that some 1465 women have participated in these trials over the past 10 years. Most studies were of short duration, with limited numbers of participants. Only four trials were longer than 2 weeks and only five trials enrolled more than 100 women. There were very few findings of significant difference between women in active trial arms and women in control arms. The CIs around effect measures were generally very wide, including both increases and decreases in the rates of potentially clinically significant events.
A total of 1465 women in trials were included in this review and a total of 90 genital epithelial disruptions in the product arms and 62 in the placebo or no treatment arms were detected. It is instructive to compare this cumulative experience with that of nonoxynol-9. Nine trials involving 5096 women, reported in 10 papers, were included in an earlier systematic review of trials of nonoxynol-9 for prevention of HIV and other STIs . Bacterial vaginosis, STIs, HIV and genital lesions were assessed in a number of trials and meta-analyses showed that the overall increase in risk of HIV, although not statistically significant, was higher for those women using nonoxynol-9 and that the risk of genital lesions was significantly higher among women using nonoxynol-9 (RR 1.18, 95% CI 1.02–1.36). There were over 400 genital ulcer events in both the nonoxynol-9 and the placebo arms.
The increased risk of HIV acquisition that was observed in some trials of nonoxynol-9 as a candidate microbicide is now generally attributed to genital disruption associated with nonoxynol-9 at higher doses. The background rate of any genital disruption in women has been estimated to be around 3% . To detect an increase in this rate to 10%, with 80% power, a randomized trial would require 222 participants in each of two trial arms. Even the detection of an increase to 20%, with 80% power, would require 66 participants in each arm.
The results of this systematic review need to be interpreted in the context of the methodologies used by trials covered by the review. All studies were randomized, with well defined study populations. In an earlier review , we highlighted the variation across studies in the evaluation and reporting of safety endpoints in these and other studies. Urogenital symptoms and signs were evaluated routinely in nearly all studies and colposcopy was conducted in the majority of safety studies. In contrast, microbiological assessments were highly variable across studies. Few studies used culture to evaluate effects on vaginal microflora or reported on UTIs or candidiasis. A variety of microbiological tests was used in the published studies, which also may have impacted on the comparability of such results.
The review is subject to a number of limitations. Our search strategy may have led to the omission of candidate microbicides. Studies of safety in men were not included. We also excluded safety data that have been published only in conference abstracts. Published results from some studies could not be analysed in a standardized manner for this review, because they were not reported in a way that allowed them to be categorized into one or more of the three composite endpoints that we used (urogenital symptoms and genital signs with either intact or disrupted epithelium). We did not evaluate systemic clinical findings other than laboratory toxicities.
As findings were merged to create these endpoints, it was necessary to combine results that were variously reported in published studies as number of events or number of women affected. Our statistical methods were devised with the goal of including effect measures from as many studies and endpoints as possible. In creating combined estimates across dose levels within single studies, we did not take account of the fact that all dose levels generally were compared to the same control group. As a consequence, the CIs for the combined estimate involving such studies will have been artificially reduced to some extent. Our statistical methods should be seen as pragmatic rather than statistically ideal, but they provide valid information on overall directions of effect, study power and pattern of results across trials. Furthermore, we did not differentiate safety results according to whether or not participants were sexually active or their HIV status, and we assumed equal follow-up time per participant.
Although HIV incidence is now viewed as the ultimate safety endpoint for potential microbicides, as well as the only endpoint that can be used to assess effectiveness, a number of outcome variables routinely measured in microbicide safety trials are recognized from observational research to be associated with increased risk of HIV transmission. Particular examples include epithelial disruption, bacterial vaginosis, vaginal candidiasis and vaginal pH [14,55]. It is, therefore, essential that safety trials are able to provide a clear indication as to whether or not the study product has had an effect on one or more of these endpoints. Because the relationship between these endpoints and the risk of HIV transmission is not known with any precision, it is difficult to specify an effect size that should preclude progression to larger scale effectiveness trials. It is, nevertheless, clear from our review that the current trials are not able to detect quite substantial effects, as they produce wide CIs for most endpoints. Furthermore, the relatively short duration of the safety trials does not provide confidence that study products will not cause important toxicities after longer term usage. In the future, consensus may emerge on specific safety markers that will guide decisions about moving agents to effectiveness trials, but for the moment, it is not known which markers are most relevant.
In conclusion, we believe that researchers, regulators and microbicide developers must agree on standard ways of designing safety trials and reporting on their results, to maximize the interpretability of their findings. There has in fact been progress in this direction, through the creation of expert panels to develop, evaluate and validate biomarkers and standardize the assessment of microbicide safety . Safety trials for microbicides remain a key step in the development pathway, but will benefit from better specification of their objectives, with corresponding strengthening of their design to ensure that these objectives are met.
The National Centre in HIV Epidemiology and Clinical Research is funded by the Australian Government Department of Health and Ageing, and is affiliated with the Faculty of Medicine, The University of New South Wales. Its work is overseen by the Ministerial Advisory Committee on AIDS, Sexual Health and Hepatitis. IMP is funded by a National Health and Medical Research Council PhD scholarship.
All authors contributed to the design of the study and the development and critical revision of the manuscript. I.M.P. collected the data, performed the statistical analysis and interpreted the data. I.Y.M. collected and interpreted the data. M.O.F. performed the meta-analysis and generated the forest plots. M.G.L. provided statistical expertise, D.N.A. provided microbiological and epidemiological expertise and L.V.D. provided clinical and epidemiological expertise. J.M.K. devised the study concept and provided epidemiological expertise. I.M.P. drafted the manuscript, to which all authors contributed. All authors have seen and approved the final version of the manuscript.
I.M.P., I.Y.M. and J.M.K. have received funding to undertake microbicide safety trials in collaboration with Starpharma Ltd, as part of a research program funded by the United States National Institutes of Health. M.G.L. has received research grants, consultancy and/or travel grants from Abbott, Boehringer Ingelheim, Bristol-Myers Squibb, Gilead, GlaxoSmithKline, Janssen-Cilag, Johnson & Johnson, Merck Sharp & Dohme, Pfizer and Roche. L.V.D., D.N.A. and M.O.F. have no conflict of interest.
1. Grant RM, Hamer D, Hope T, Johnston R, Lange J, Lederman MM, et al
. Whither or wither microbicides? Science 2008; 321:532–534.
2. Rustomjee R, Karim QA, Abdool Karim SS, Laga M, Stein Z. Phase 1 trial of nonoxynol-9 film among sex workers in South Africa. AIDS 1999; 13:1511–1515.
3. Van Damme L, Niruthisard S, Atisook R, Boer K, Dally L, Laga M, et al
. Safety evaluation of nonoxynol-9 gel in women at low risk of HIV infection. AIDS 1998; 12:433–437.
4. Martin HL Jr, Stevens CE, Richardson BA, Rugamba D, Nyange PM, Mandaliya K, et al
. Safety of a nonoxynol-9 vaginal gel in Kenyan prostitutes: a randomised clinical trial. Sex Transm Dis 1997; 24:279–283.
5. Van Damme L, Chandeying V, Ramjee G, Rees H, Sirivongrangson P, Laga M, et al
. Safety of multiple daily applications of COL-1492 a nonoxynol-9 vaginal gel, among female sex workers. AIDS 2000; 14:85–88.
6. Lapido OA, De Castro MP, Filho LCCT, Coutinho E, Waller DP, Cone F, et al
. A new vaginal antimicrobial contraceptive formulation: phase 1 clinical pilot studies. Contraception 2000; 62:91–97.
7. Coggins C, Elias C, N-9 Formulation Preferences Study Group Committee. Safety of three formulations of nonoxynol-9 containing vaginal spermicides. Int J Gynaecol Obstet 2000; 68:267–268.
8. Roddy RE, Cordero M, Cordero C, Fortney JA. A dosing study of nonoxynol-9 and genital irritation. Int J STD AIDS 1993; 4:165–170.
9. Stafford MK, Ward H, Flanagan A, Rosenstein IJ, Taylor-Robinson D, Smith JR, et al
. Safety study of nonoxynol-9 as a vaginal microbicide: evidence of adverse effects. JAIDS 1998; 17:327–331.
10. Hoffman IF, Taha TE, Padian NS, Kelly CW, Welch JD, Martinson FE, et al
. Nonoxynol-9 100 mg gel: multisite safety study from sub-Saharan Africa. AIDS 2004; 18:2191–2195.
11. Roddy RE, Zekeng L, Ryan KA, Tamoufe U, Weir SS, Wong EL. A controlled trial of nonoxynol-9 film to reduce male-to-female transmission of sexually transmitted diseases. N Engl J Med 1998; 339:504–510.
12. Roddy RE, Zekeng L, Ryan KA, Tamoufe U, Tweedy KG. Effect of nonoxynol-9 gel on urogenital gonorrhea and chlamydial infection: a randomized controlled trial. JAMA 2002; 287:1117–1122.
13. Richardson BA, Lavreys L, Martin Jr HL, Stevens CE, Ngugi E, Mandaliya, et al
. Evaluation of a low-dose nonoxynol-9 gel for the prevention of sexually transmitted diseases: a randomized clinical trial
. Sex Transm Dis
14. Van Damme L, Ramjee G, Alary M, Vuylsteke B, Chandeying V, Rees H, et al
. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomized controlled trial. Lancet 2002; 360:971–977.
15. WHO/CONRAD Technical Consultation on Nonoxynol-9. Geneva: World Health Organization; 9–10 October 2001; Summary Report.
16. Wilkinson D, Ramjee G, Tholandi M, Rutherford G. Nonoxynol-9 for preventing vaginal acquisition of sexually transmitted infections by women from men
. Cochrane Database Syst Rev
17. Hillier SL, Moench T, Shattock R, Black R, Reichelderfer P, Veronese F. In vitro and in vivo: the story of nonoxynol-9. JAIDS 2005; 39:1–8.
18. Alliance for Microbicide Development. [cited 10 August 2007 for search strategy, cited 10 May 2008 for current trials and cited 9 December 2008 for Microbicide Development Strategy and Microbicide Research Working Group]. www.microbicide.org
19. Feldblum PJ, Adeiga A, Bakare R, Wevill S, Lendvay A, Obadaki F, et al
. SAVVY vaginal gel (C31G) for prevention of HIV infection: a randomized controlled trial in Nigeria. PLoS ONE 2008; 3:e1474.
20. Peterson L, Nanda K, Opoku BK, Ampofo WK, Owusu-Amoako M, Boakye AY, et al
. SAVVY(R) (C31G) gel for prevention of HIV infection in women: a phase 3, double-blind, randomized, placebo-controlled trial in Ghana. PLoS ONE 2007; 2:e1312.
21. Johansson E. Results from the phase III Carraguard study. Microbicides Conference; New Delhi, India; 2008.
22. Van Damme L, Govinden R, Mirembe FM, Guédou F, Solomon S, Becker ML, et al
. Lack of effectiveness of cellulose sulfate gel for the prevention of vaginal HIV transmission. N Engl J Med 2008; 359:463–472.
23. Halpern V, Wang L, Obunge O, Ogunsola F, Mehta N, Onyejepu N, et al. Effectiveness of cellulose sulfate gel for prevention of HIV: results of the phase III trial in Nigeria
. Special session: 4th IAS Conference on HIV Pathogenesis, Treatment and Prevention
; 2007; abstract no. WESS302.
24. Balzarini J, Van Damme L. Microbicide drug candidates to prevent HIV infection. Lancet 2007; 369:787–797.
25. McGowan I. Microbicides: a new frontier in HIV prevention. Biologicals 2006; 34:241–255.
26. D'Cruz OJ, Uckun FM. Clinical development of microbicides for the prevention of HIV infection. Current Pharm Design 2004; 10:315–336.
27. Millwood I, Poynten IM, Andresen D, Van Damme L, Kaldor J. A systematic review of microbicide safety studies: most published trials are characterized by inadequate statistical power, short duration of exposure, and nonstandardized endpoints [abstract 576]
2008: New Delhi, India; 24–27 February 2008.
28. WHO/CONRAD. Manual for the Standardization of colposcopy for the evaluation of vaginally administered products (update 2004): revised procedure for colposcopy in the development of new vaginal products. Geneva, Switzerland: World Health Organization; 2004.
29. Deeks JJ, Douglas AG, Bradburn MJ. Statistical methods for examining heterogeneity and combining results from several studies in meta-analysis. In: Egger M, Davey-Smith D, Altman DG, editors. Systematic reviews in healthcare. 2nd edn. London: BMJ Books; 2001. pp. 285–312.
30. Kilmarx PH, van de Wijgert JHHM, Chaikummao S, Jones HE, Limpakarnjanarat K, Friedland BA, et al
. Safety and acceptability of the candidate microbicide Carraguard in Thai women. JAIDS 2006; 43:327–334.
31. van de Wijgert JHHM, Braunstein SL, Morar NS, Jones HE, Madurai L, Evans Strickfaden TT, et al
. Carraguard vaginal gel safety in HIV-positive women and men in South Africa. JAIDS 2007; 46:538–546.
32. Kilmarx PH, Blanchard K, Chaikummao S, Friedland BA, Srivirojana N, Conolly C, et al
. A randomized, placebo-controlled trial to assess the safety and acceptability of use of Carraguard vaginal gel by heterosexual couples in Thailand. Sex Transm Dis 2008; 35:226–232.
33. Bollen LJM, Blanchard K, Kilmarx PH, Chaikummao S, Connolly C, Wasinrapee P, et al
. No increase in cervicovaginal proinflammatory cytokines after Carraguard use in a placebo-controlled randomized clinical trial. JAIDS 2008; 47:253–257.
34. Mauck C, Weiner DH, Ballagh S, Creinin M, Archer DF, Schwartz JL. Single and multiple exposure tolerance study of cellulose sulfate gel: a phase I safety an colposcopy study. Contraception 2001; 64:383–391.
35. Malonza IM, Mirembe F, Nakabiito C, Odusoga LO, Osinupebi OA, Hazari K, et al
. Expanded phase I safety and acceptability study of 6% cellulose sulphate vaginal gel. AIDS 2005; 19:2157–2163.
36. El-Sadr WM, Mayer KH, Maslankowski L, Hoesley C, Justman J, Gai F, et al
. Safety and acceptability of cellulose sulphate as a vaginal microbicide in HIV-infected women. AIDS 2006; 20:1109–1116.
37. Schwartz JL, Mauck C, Lai J-J, Creinin MD, Brache V, Ballagh SA, et al
. Fourteen-day safety and acceptability study of 6% cellulose sulfate gel: a randomised double-blind phase I safety study. Contraception 2006; 74:133–140.
38. Doh AS, Ngoh N, Roddy R, Lai JJ, Linton K, Mauck C. Safety and acceptability of 6% cellulose sulfate vaginal gel applied four times per day for 14 days. Contraception 2007; 76:245–249.
39. Stafford MK, Cain D, Rosenstein I, Fontaine EA, McClure M, Flanagan AM, et al
. A placebo-controlled, double-blind prospective study in healthy female volunteers of dextrin sulphate gel: a novel potential intravaginal virucide. JAIDS 1997; 14:213–218.
40. Rosenstein IJ, Stafford MK, Kitchen VS, Ward H, Weber JN, Tylor-Robinson D. Effect on normal vaginal flora of three intravaginal microbicidal agents potentially active against human immunodeficiency virus type 1. J Infect Dis 1998; 177:1386–1390.
41. Low-Beer N, Gabe R, McCormack S, Kitchen VS, Lacey CJ, Nunn AJ. Dextrin sulphate as a vaginal microbicide: randomised double-blind, placebo-controlled trial including healthy female volunteers and their male partners. JAIDS 2002; 31:391–398.
42. McCormack S, Jespers V, Low-Beer N, Gabe R, Kaganson N, Chapman A, et al
. A dose-ranging phase I study of dextrin sulphate, a vaginal microbicide, in HIV-negative and HIV-positive female volunteers. Sex Transm Dis 2005; 32:765–770.
43. Bakobaki JM, Lacey CJ, Bukenya MI, Nunn AJ, McCormack S, Byaruhanga RN, et al
. A randomised controlled safety and acceptability trial of dextrin sulphate vaginal microbicide gel in sexually active women in Uganda. AIDS 2005; 19:2149–2156.
44. Hemmerling A, Potts M, Walsh J, Young-Holt B, Whaley K, Stefanski DA. Lime juice as a candidate microbicide? An open-label safety trial of 10% and 20% lime juice used vaginally. J Women's Health 2007; 16:1041–1051.
45. Goeman J, Ndoye I, Sakho L, Mboup S, Piot P, Karam M, et al
. Frequent use of Menfegol spermicidal vaginal foaming tablets associated with a high incidence of vaginal lesions. J Infect Dis 1995; 171:1611–1614.
46. Mauck CK, Weiner DH, Ballagh SA, Creinin MD, Archer DF, Schwartz JL, et al
. Single and multiple exposure tolerance study of polystyrene sulfonate gel: a phase I safety and colposcopy study. Contraception 2004; 70:77–83.
47. Van Damme L, Wright A, Depraetere K, Rosenstein I, Vandersmissen V, Poulter L, et al
. A phase I study of a novel potential intravaginal microbicide, PRO2000, in healthy sexually inactive women. Sex Transm Inf 2000; 76:126–130.
48. Lacey CJN, Wright A, Weber JN, Profy AT. Direct measurement of in-vivo vaginal microbicide levels of PRO 2000 achieved in a human safety study. AIDS 2006; 20:1027–1030.
49. Keller MJ, Guzman E, Hazrati E, Kasowitz A, Cheshenko N, Wallenstein S, et al
. PRO 2000 elicits a decline in genital tract immune mediators without compromising intrinsic antimicrobial activity. AIDS 2007; 21:467–476.
50. Ballagh SA, Baker JM, Henry DM, Archer DF. Safety of single daily use for one week of C31G HEC gel in women. Contraception 2002; 66:369–375.
51. Khanna N, Dalby R, Connor A, Church A, Stern J, Frazer N. Phase I clinical trial of repeat dose Terameprocol vaginal ointment in healthy female volunteers. Sex Transm Dis 2008; 35:577–582.
52. Jespers VA, Van Roey JM, Beets GI, Buve' AM. Dose-ranging phase 1 study of TMC120, a promising vaginal microbicide, in HIV-negative and HIV-positive female volunteers. JAIDS 2007; 44:154–158.
53. Schwartz J, Kovalevsky G, Lai J, Ballagh SA, McCormick T, Douville K, et al
. A randomized six-day safety study of an antiretroviral microbicide candidate UC781, a nonnucleoside reverse transcriptase inhibitor. Sex Transm Dis 2008; 35:414–419.
54. van de Wijgert JH, Kilmarx PH, Jones HE, Karon JM, Chaikummao S. Differentiating normal from abnormal rates of genital epithelial findings in vaginal microbicide trials. Contraception 2008; 77:122–129.
55. van de Wijgert JHHM, Morrison CS, Cornelisse PGA, Munjoma M, Moncada J, Awio P, et al
. Bacterial vaginosis and vaginal yeast, but not vaginal cleansing, increase HIV-1 acquisition in African women. JAIDS 2008; 48:203–210.