IN THE FACE OF CONTINUED spread of the human immunodeficiency virus (HIV)-1 epidemic, there is a need for development and evaluation of female-controlled methods for HIV-1 prevention. Vaginal virucides, if proven to be safe and efficacious, may be used without the male partner's participation and may be affordable to women in developing countries who are most in need of such agents.
Nonoxynol-9 (N-9), the most widely used vaginal spermicide worldwide, has been shown in vitro to be cidal for HIV-1,1,2 although results from two clinical studies on the effect of N-9 on HIV-1 acquisition in women yielded conflicting results.3–5 The placebo-controlled trial of an N-9 sponge by Kreiss et al. led to an appreciation that N-9 may be toxic to the epithelium of the genital tract. Subsequent studies of a vaginal suppository containing 150 mg of N-9 found epithelial disruption in Thai women exposed to frequent N-9 doses,6 and in women in the Dominican Republic with daily use.7
One product developed with the potential mucosal toxicity of N-9 in mind, Advantage-24 (Columbia Research Laboratories, Inc., Rockville Centre, NY), contains 3.5% N-9 in a polymer-based gel. The gel base, polycarbophil, adheres to epithelial cells, allowing once-daily application independent of the number of times a woman has intercourse. This product is packaged in a single-use applicator to deliver 1.5 g of gel containing 52.5 mg of N-9. Once-daily application and a low dose of N-9 are ideal characteristics for a product considered for use in HIV-1 infection prevention.
This trial was designed to evaluate the safety and mucosal toxicity of once-daily dosing of this N-9 gel in a cohort of female sex workers in Mombasa, Kenya. This safety study was conducted in preparation for a phase III efficacy trial of HIV-1 infection prevention in this population.
The study was conducted as a randomized, double-blinded, placebo-controlled, two-period crossover trial. HIV-1-seronegative women who were participants in an ongoing cohort study of HIV-1 seroincidence in commercial sex workers in Mombasa, Kenya were eligible to participate.8 To minimize any risk of HIV-1 infection to trial participants, only those women who reported 100% condom use at their last cohort study visit, were free from sexually transmitted diseases (STDs) at their last cohort study visit, and who agreed to use condoms throughout the trial were enrolled.
After giving informed consent, subjects underwent a detailed structured interview to collect demographic and behavioral information. A general physical examination was performed, followed by pelvic examination and colposcopy according to the methods outlined in the World Health Organization (WHO) Manual for the Standardization of Colposcopy for the Evaluation of Vaginally Administered Products.9 All physical examinations were performed by a single observer (CES) with experience in colposcopy.
Those women who had no clinical signs of STDs, who had no visual or colposcopic evidence of genital epithelial disruption, and who were not clinically pregnant were randomized to a sequence of either N-9 followed by placebo, or vice versa. Subjects were instructed to insert one applicator of study gel once daily for 14 days. Subjects were evaluated by interview and physical examination, including pelvic examination and colposcopy, at the end of 7 and 14 days of product administration. The first product period was followed by a washout period of at least 14 days. During the second product period, women were crossed over to use of the alternative product. Subjects were seen after 7 and 14 days of product use for evaluation, as outlined for the first period of the trial. Any subject experiencing epithelial disruption was withdrawn from that period of the trial, given appropriate treatment if indicated, and followed weekly until all lesions resolved.
Compliance with product use and sexual activity were monitored through the use of a coital/product use log, interview, and enumeration of unused gel applicators returned. Any discrepancies between the coital log, interview answers, and number of unused applicators were clarified with the subject.
Written informed consent was obtained from all study participants. This study received ethical approval by the institutional review boards of the University of Washington and the University of Nairobi.
Data Management and Analysis
Data were entered into a 486 desktop computer using the SPSS Data Entry program (SPSS Inc., Chicago, IL) with range checks and skip rules in place. All data were verified by comparing line listings of data entered with the subject files.
The primary outcome measure was genital epithelial disruption. As defined by the WHO manual, epithelial disruption included ulceration and abrasion.9 Other lesions not involving disruption of the genital epithelium were secondary outcome measures and included erythema, edema, petechiae, ecchymosis, and subepithelial hemorrhage with swelling. All epithelial abnormalities were defined as described in the WHO manual.9 The lesions were assessed both visually and colposcopically.
Follow-up data during the N-9 and placebo arms were compared using statistical methods designed for matched-pairs analysis, allowing each woman to serve as her own control. McNemar's test was used for binary variables and the Wilcoxon signed ranks test was used for continuous variables. These methods require complete follow-up on all subjects, so only those subjects with complete follow-up in both arms of the trial were included in these calculations. However, data on epithelial disruption, the main outcome variable, were included for all subjects. Generalized estimating equations with a binary link and exchangeable correlation structure were used to estimate the risk of epithelial disruption associated with N-9 use. This method estimates risk while adjusting for correlation between multiple observations in a given subject, and does not require complete follow-up.10
Sixty subjects were randomized in this trial, 30 to receive N-9 followed by placebo, and 30 to receive placebo followed by N-9. The median age was 30 years. Subjects reported a median of two sexual partners and two acts of intercourse in the week before enrollment. There were no demographic or behavioral differences at enrollment between subjects in the two randomization sequences (data not shown).
Fifty-two subjects (87%) were evaluated at all four follow-up visits (i.e., after 7 and 14 days of each study period) and were evaluable in the matched-pairs analyses. Three subjects completed only the first period of the trial and did not return for the second crossover period. Two subjects missed the day 7 examination of the first period, one the day 14 examination of the first period, and one both day 7 and 14 examinations of the second period. One subject was withdrawn from her first study period because of a genital abrasion, and this same subject did not begin the second study period because of pregnancy. Thus, follow-up data were incomplete on eight subjects. Compliance with gel use as measured by reported applicator use and enumeration of unused applicators was 98% for N-9 and 97% for placebo.
No significant differences in reported sexual exposure characteristics occurred during follow-up (Table 1). Median sexual frequency per week was 1.8 during the N-9 period and 2.5 during the placebo period. The overall rates of genital symptoms reported by subjects were similar during the N-9 and placebo periods. Genital complaints were reported by 48% of women during the use of N-9 and 38% during use of placebo. Although the rate of genital complaints was high during both periods, no individual symptom was significantly different between the N-9 and placebo periods.
Physical examination findings during the N-9 and placebo periods were similar, with the exception of median vaginal pH, which was lower during the placebo period than the N-9 period (Table 2). There were no differences in inguinal adenopathy, vaginal discharge, cervical mucopus, cervical friability, pelvic tenderness, or cervical ectopy.
No significant differences in epithelial abnormalities were detected during follow-up, as shown in Table 3. Colposcopy was performed for only 64% of follow-up examinations because of frequent interruptions in municipal electricity supply. For the 144 visits that included both colposcopic and visual examination, correlation of results using the two detection methods was 99% (Spearman rank correlation coefficient, 0.97). In only one instance was a lesion (petechiae) detected colposcopically but not visually. The overall number of abnormal epithelial findings was low, and there were no differences between N-9 and placebo periods. Petechiae, erythema, and edema were the only nondisruptive epithelial lesions identified, and they occurred with similar frequency during use of N-9 and placebo.
Epithelial abnormalities that involved disruption of epithelial integrity are described in detail in Table 4. Of four such events, three occurred while subjects were receiving placebo, and one occurred during N-9 treatment. The first two disruptions described in Table 4 were clinically consistent with recurrent herpes simplex virus. The third disruption was clinically consistent with the diagnosis of candidal vulvitis and was confirmed by the presence of yeast elements on microscopy. The fourth epithelial break was clinically consistent with chancroid.
Only three epithelial breaks are reflected in Table 3 because one subject who had a vulvar abrasion did not go on to receive N-9 in the second period of the trial owing to pregnancy. This subject was thus not included in the matched-pairs analysis. To avoid bias introduced by excluding a subject without complete follow-up, generalized estimating equations, which allowed the inclusion of all visits regardless of completeness of follow-up, were used. The risk of epithelial disruption associated with N-9 use was 0.33 (95% confidence interval, 0.03–3.26).
In this trial of once-daily application of a vaginal gel containing 52.5 mg of N-9, we observed no evidence of genital epithelial toxicity after 7 and 14 days of use. Evaluation was designed to detect subtle evidence of epithelial damage, as outlined by the WHO.9 Using these methods, we detected no genital epithelial damage, either ulcerative or nonulcerative, with this product.
In previous studies, N-9 has been associated with significant genital mucosal toxicity. Although N-9 had been licensed and widely used for more than 30 years, its potential toxicities in high-risk populations were first publicized by Kreiss et al., who detected an increased risk of genital ulceration in N-9 sponge users.3 The N-9 sponge contained 1 g of N-9 in a vehicle that was designed to be left in place for 6 hours after intercourse. The mean weekly sponge use in the study population was 14. It was unknown whether the excess ulceration observed was due to high N-9 concentration, prolonged exposure, high frequency of use, the sponge vehicle per se, protective effects of the placebo, or the sexual behavioral characteristics of the population.
Two subsequent studies evaluated the epithelial toxicity of a vaginal suppository containing 150 mg of N-9. In Thailand, hourly application of a vaginal suppository for 4 hours on 14 consecutive days resulted in a 43% incidence of epithelial disruption.6 A dose-ranging study conducted in the Dominican Republic found that women using the suppository every other day did not have an increased risk of epithelial disruption compared with placebo users, but more frequent use was associated with a two- to fivefold increased risk of genital epithelial breaks.7 Thus, every-other-day use appeared to be a threshold limit for safety of the 150-mg suppository.
One recently published study compared the epithelial toxicity of Advantage-24, the N-9 product evaluated in the current trial, with the sponge containing 1,000 mg N-9 and a cellulose-based gel containing 100 mg N-9 after once-daily use for 7 days.11 No genital ulceration was seen in the Advantage-24 or sponge groups, and two ulcers were seen in women using the 100-mg N-9 gel. No placebo group was included in this trial.
The current study is the first placebo-controlled trial published to evaluate the epithelial toxicity of Advantage-24. This product contains approximately 5% of the amount of N-9 contained in the sponge, one third of the N-9 used in the suppository trials, and one half of the N-9 contained in the gel studied recently. The delivery vehicle, a polycarbophil polymer gel, may also play a role in limiting the epithelial toxicity of this product. It appears that this product may be less toxic than other formulations that contain higher doses of N-9.
Several limitations of the current trial warrant consideration. First, the study period lasted for only 2 weeks, and it is possible that toxicity may become apparent with longer-term use. Second, the statistical power to detect a small increase in epithelial toxicity was limited in our safety study. Although our estimate of the risk of epithelial disruption associated with N-9 use was 0.33, the 95% confidence interval (0.03–3.26) included one. Third, colposcopy was not available at every examination because of problems with electrical supply. Although colposcopy should ideally be performed in a trial evaluating the epithelial toxicity of a vaginally administered product, in the 144 visits that included both colposcopic and visual inspection, we detected only one lesion colposcopically that was not detected visually. Thus, it is unlikely that significant epithelial toxicity was missed during the noncolposcopic examinations. Finally, this trial addressed the safety of the gel applied once daily. If, in actual practice, the gel is used more than once a day, a different safety profile might be observed.
It is now more than 10 years since the in vitro cidal activity of N-9 against HIV-1 was documented, yet to date only one randomized, clinical trial to evaluate the efficacy of an N-9 preparation in preventing HIV acquisition by women has been published. The current study suggests that daily short-term use of a low-dose N-9 gel is safe with respect to genital epithelial toxicity. Phase III efficacy trials of this product in high-risk populations are now indicated to determine if protection against HIV and other STDs is provided.
1. Hicks DR, Martin LS, Getchell JP, et al. Inactivation of HTLV-III/LAV-infected cultures of normal human lymphocytes by nonoxynol-9 in vitro [letter]. Lancet 1985; 2:1422–1423.
2. Polsky B, Baron PA, Gold JWM, Smith JL, Jensen RH, Armstrong D. In vitro inactivation of HIV-1 by contraceptive sponge containing nonoxynol-9 [letter]. Lancet 1988; 1:1456.
3. Kreiss J, Ngugi E, Holmes K, et al. Efficacy of nonoxynol-9 contraceptive sponge use in preventing heterosexual acquisition of HIV in Nairobi prostitutes. JAMA 1992; 268:477–482.
4. Zekeng L, Feldblum PJ, Oliver RM, Kaptue L. Barrier contraceptive use and HIV infection among high-risk women in Cameroon. AIDS 1993; 7:725–731.
5. Feldblum PJ, Weir SS. The protective effect of nonoxynol-9 against HIV infection [letter]. Am J Public Health 1994; 84:890–891.
6. Niruthisard S, Roddy RE, Chutivongse S. The effects of frequent nonoxynol-9 use on the vaginal and cervical mucosa. Sex Transm Dis 1991; 18:176–179.
7. 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.
8. Martin H, Jackson D, Mandaliya K, et al. Preparation for AIDS vaccine evaluation in Mombasa, Kenya: Establishment of seronegative cohorts of commercial sex workers and trucking company employees. AIDS Res Human Retroviruses 1994; 10: S235–S237.
9. World Health Organization. Manual for the Standardization of Colposcopy for the Evaluation of Vaginally Administered Products. Geneva: World Health Organization, November, 1993.
10. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986; 42:121–130.
11. Poindexter AN, Levine H, Sangi-Haghpeykar H, Frank ML, Grear A, Reeves KO. Comparison of spermicides on vulvar, vaginal, and cervical mucosa. Contraception 1996; 53:147–153.