Kilmarx, Peter H MD*†; van de Wijgert, Janneke H. H. M. PhD§; Chaikummao, Supaporn BHP*; Jones, Heidi E MPH‡; Limpakarnjanarat, Khanchit MD, MPH*; Friedland, Barbara A MA‡; Karon, John M PhD‖; Manopaiboon, Chomnad MA*; Srivirojana, Nucharee MA¶; Yanpaisarn, Somboonsak MD#; Supawitkul, Somsak MD**; Young, Nancy L MS*; Mock, Philip A MApp Stats*; Blanchard, Kelly MS‡; Mastro, Timothy D MD*†
The HIV/AIDS pandemic continues unabated.1 Currently available HIV prevention tools are limited and not feasible for many women.2 The need to expand the range of HIV prevention tools available, especially those that women can initiate, is urgent. Microbicides are products that are applied topically inside the vagina or rectum to prevent infection with HIV and other sexually transmitted infections (STIs). They can be formulated as gels, creams, suppositories, or vaginal rings and may or may not be contraceptive. They can be used alone or in combination with a physical barrier method to provide increased protection or backup in case of barrier failure.
Carraguard (PC-515) is the Population Council's lead noncontraceptive candidate microbicide. The active pharmaceutic ingredient is a sulfated polysaccharide mixture of λ- and κ-carrageenan derived from red seaweed. Carrageenan is inexpensive and stable and has been used for centuries as a food thickener and emulsifier in topical preparations. Its mucosal absorption is negligible. Carraguard binds to pathogens and host cells, acting as a fusion and entry inhibitor. This differentiates it from surfactants such as nonoxynol-9, which disrupts cell membranes, including host epithelial cells. Preclinical studies indicated that Carraguard gel and its matching placebo methylcellulose gel are nonirritating and nonteratogenic in animals and have no effect on lactobacilli and spermatozoa in vitro and that Carraguard is effective against HIV and several other STI pathogens in vitro and in animal systems.3-7 Neither Carraguard nor placebo has any adverse effects on latex condoms or deleterious chemical or physical interaction with silicon-based lubricants (R. Maguire, October 11, 2005, personal communication). Two phase 1 safety trials of closely related prototypes of Carraguard showed no signs of irritation in women related to gel use.8,9
Between 1999 and 2002, two phase 2 expanded safety and acceptability trials of Carraguard were conducted in South Africa and Thailand. The results of the Thai trial are reported here.
We conducted a randomized, placebo-controlled, triple-blind clinical trial of Carraguard in Chiang Rai, northern Thailand10 as a collaboration of the Thai Ministry of Public Health, the US Centers for Disease Control and Prevention (CDC), the Chiang Rai Provincial Public Health Office, Chiang Rai Hospital, and the Population Council. Thailand was chosen because of the Thai Government's interest in vaginal microbicides, existing collaborations, and availability of required infrastructure. The primary objective was to assess the safety and acceptability of Carraguard when applied vaginally at least 3 times per week for 12 months. The study design followed recommendations by the International Working Group on Microbicides.11,12
Participants were recruited from family planning clinics, including the Chiang Rai Municipal Clinic and 10 district health centers within 20 km from the town of Chiang Rai. Women were eligible to participate if they were in good health, aged 18 years or older, HIV-negative, not pregnant or planning a pregnancy in the next year, had lived and planned to remain in the area for at least 1 year, and able to achieve a score of 80% on a study comprehension test. Women were excluded if they had had a delivery or abortion within the last 6 weeks, a male sex partner known to be HIV-positive, a recent history of genital surgery or nonmenstrual vaginal bleeding during sex, persistent gynecologic morbidity, an abnormal Pap smear at screening, or latex allergy; were participating in another vaginal product study; or were injecting recreational drugs. Women who were diagnosed with a symptomatic reproductive tract infection (RTI) at screening could be enrolled after treatment. Women with asymptomatic bacterial vaginosis (BV) or a yeast infection were eligible for enrollment.
Women could be screened for eligibility up to 3 times, with no more than 1 month between the last screening and enrollment visits. Eligible consenting women were randomized at enrollment: half to Carraguard and half to placebo. After randomization, women were asked to insert the first applicator at the clinic and to insert 1 applicator at least 3 times per week (approximately every other day, with or without sex) at home thereafter. They were also asked to insert gel within 1 hour before sex and to use a condom every time sex took place. They were encouraged to continue gel use during menses and to use contraception. They were allowed to use other vaginal products (eg, vaginal douches, medicinal products, tampons) during trial participation, but the use of such products was carefully documented. Both gels came in identical single-use Microlax-type applicators (Norden Pac International AB, Kalmar, Sweden) designed to dispense approximately 4 mL of gel. Women were given coital diaries to track product and condom use and sexual activity and were asked to bring the diaries as well as used and unused applicators to the clinic. Gel use could be temporarily or permanently interrupted by clinic staff or participants. If women became pregnant, they were asked to discontinue gel use but could remain in the study.
Participants were asked to return to the clinic 14 days and 1 month after enrollment and monthly thereafter for a total of 12 months and at any time they needed additional supplies of gel or condoms or were experiencing any problems. Sexual activity, product and condom use, acceptability, and symptoms were assessed by structured interviewing (aided by consulting coital diaries) and by counting and weighing returned applicators at each visit. Mucosal safety was evaluated by visual inspection of the cervix, vaginal walls, and external genitalia during pelvic examinations at each visit. Colposcopy was done routinely at the day 14 and month 1 visits for the first 25 women but at the discretion of the study nurses thereafter.13,14 Each woman could have multiple abnormal genital findings at each visit; all were included and reported. Times of occurrence and resolution were defined as the interval midpoint between the previous and current visits.
Most RTIs were tested for at screening and at months 1 through 12, including BV and yeasts (Gram stain microscopy and Nugent scoring15), pregnancy (urine human chorionic gonadotropin [hCG] test); Trichomonas vaginalis (InPouch culture test; Biomed Diagnostics, San Jose, CA), and Neisseria gonorrhoeae and Chlamydia trachomatis (COBAS Amplicor polymerase chain reaction [PCR]; Amplicor, Branchburg, NJ, with minor modifications to eliminate test inhibition by study gels16). Blood specimens were collected at screening and at months 1, 3, 6, 9, and 12 and were tested for HIV (Genetic Systems HIV 1/2 Enzyme Immunoassay; Redmond, WA, with confirmation by NovaPath HIV-1 Immunoblot; BioRad Diagnostics Group, Hercules, CA) and Treponema pallidum (Macro-Vue rapid plasma reagin card test; Becton-Dickinson, Franklin Lakes, NJ, with confirmation by T. pallidum passive particle agglutination test; Fujirebio Diagnostics, Tokyo, Japan). Pap smears were evaluated at screening and at month 12 using the Class System.17 At the end of the study, stored blood specimens from screening and exit visits were tested for herpes simplex virus (HSV)-2, and, if positive, HSV-1 (HerpeSelect HSV-2 IgG enzyme-linked immunosorbent assay [ELISA] and HSV-1/2 Immunoblot; Focus Technologies, Cypress, CA).
Participant identification numbers were randomized into 6 color-coded groups to ensure sufficient blinding using the pseudorandom number generator of SPSS (version 7.0; SPSS, Inc., Chicago, IL). The first 25 women (providing data for the first interim analysis) were block randomized separately before randomizing the remaining women. After eligibility was confirmed, the next consecutive-numbered opaque envelope was opened by staff to determine group assignment. Participants and study staff were blinded throughout the trial, including data analysis. With 152 woman-years of observation, the study had approximately 80% power to detect a 1.6-fold or higher increase in the rate of genital findings with epithelial disruption, based on an observed rate of 39 episodes per 100 woman-years in the placebo group (2-sided test, α = 0.05). Two interim data analyses were planned and conducted, the first after 25 woman-months of data and the second after 89 women-years of data. The interim results were reviewed by an independent data safety and monitoring board, which decided in both cases that the trial could continue as planned.
Data were double-entered and verified in Epi Info 6.04d (Centers for Disease Control and Prevention [CDC], Atlanta, GA, and World Health Organization [WHO], Geneva, Switzerland, 2001). Adverse events (AEs) were coded by an independent Population Council physician using the WHO Adverse Reaction Dictionary adapted to describe genital findings in greater detail.18 All primary analyses were by intent-to-treat. Data were analyzed using SPSS (version 8.2; SAS Institute, Inc., Cary, NC) and Stata (version 7.0; StataCorp, College Station, TX). Two-sample statistical tests included the Fisher exact test for categoric outcomes and the Wilcoxon-Mann-Whitney test for continuous outcomes. Logistic regression was used to identify predictors for study dropout. Generalized estimating equation (GEE) models were used to analyze compliance with the study schedule, product use, and prevalence of end points at follow-up visits. The structure of the working correlation matrix was chosen based on the extension of the Akiake information criterion to GEE models.19 Anderson-Gill counting process proportional hazards models for repeated events were used to analyze the incidence of end points, and Kaplan-Meier curves were used to verify that the proportional hazards assumption was appropriate.20
Women received reimbursement to cover transportation and time spent for each visit. This study was approved by Ethical Committees of the CDC, the Population Council, and the Thai Ministry of Public Health. The Chiang Rai Microbicide Research Community Advisory Group helped to ensure that the research was responsive to the needs of the community.
Between February and November 2000, 261 women were screened and 165 were randomized, with 83 in the Carraguard group and 82 in the placebo group (Fig. 1). Data collection was completed in November 2001. Retention was high, with 148 (89.7%) women completing the full year in the study, resulting in 152 woman-years of data (75 and 77 woman-years in the Carraguard and placebo groups, respectively). Among all women randomized, at least 75% of women in each group attended all their visits. The proportion of women who completed the study was higher in the placebo group than in the Carraguard group (95% vs. 84%; P = 0.04). Ten women withdrew early-8 in the Carraguard group and 2 in the placebo group (P = 0.10): 5 because of spousal disapproval (all in Carraguard group), 3 who relocated, 1 who attributed a urinary tract infection to gel use, and 1 who died of breast cancer. Seven women were lost to follow-up (5 in the Carraguard group and 2 in the placebo group; P = 0.44). Study dropout was not associated with experiencing an RTI, genital findings, perceived side effects of gel, or poor self-reported acceptability.
Women in the Carraguard and placebo groups did not differ significantly for any of the sociodemographic characteristics recorded, except for educational level, or in reports of baseline sexual and vaginal product use behaviors (Table 1). Enrolled women were 32 years of age on average (range: 19-50 years of age) and averaged between 6 and 7 years of education. Almost all women were married, lived with their husbands, had at least 1 child, had 1 sexual partner, and were using at least 1 method of family planning at baseline. In terms of clinical and laboratory findings, women in the placebo group were more likely to report having taken any medication in the month before screening (77% vs. 57%; P = 0.01) and were more likely to test HSV-2-seropositive at screening (58% vs. 43%; P = 0.04).
Study Gel and Condom Use During the Trial
Reported compliance with gel use was high throughout the study, with between 87% and 96% of the women returning at least 3 used applicators per week at each study visit in each group, with no differences between the 2 groups. The mean number of applicators used per week for all visits combined was 4.0 in the Carraguard group (range: 1.1-7.8 applicators) and 4.2 (range: 1.0-7.1 applicators) in the placebo group. Reported gel and condom use during vaginal sex was also high. The median number of sex acts for women who completed all 12 months was 71. Study gel was used in 84% of vaginal sex acts in the Carraguard group and in 91% in the placebo group. Condom use was reported for 71% of vaginal sex acts in the Carraguard group and for 62% in the placebo group. There were no statistically significant differences between groups in frequency of sex or the proportions of vaginal sex acts during which gel and condoms, gel with or without condoms, and condoms with or without gel were used (data not shown). The proportion of women using gel during menses (mostly in the absence of sex) increased over time from 30% at month 1 to 49% at month 6 and 54% at month 12. A total of 64 women reported at least 1 interruption of gel use during the study, 32 in each group (P = 1.00). They reported a total of 80 gel use interruptions, 40 in each group (P = 1.00), with no significant differences in mean duration and reasons for interruptions. The most common reasons for gel interruptions were not having the gel while away from home (15 interruptions), developing a genital complaint (9 interruptions), not visiting the study clinic within 2 weeks from scheduled visit (8 interruptions), and experiencing a serious adverse event (SAE; 8 interruptions). Almost all women reported throughout the study to have squeezed all the gel out of each applicator every time. Applicator weight data showed that, on average, women dispensed 5.0 mL of gel into the vagina.
Safety: Genital Findings
Of the 37 women with at least 1 disrupted genital finding during visual inspection of the cervix, vaginal walls, and external genitalia, 26 had such a finding at only 1 visit, 8 at 2 visits, 2 at 3 visits, and 1 at 4 visits. Only 1 woman had epithelial disruption at consecutive monthly visits (superficial disruption on the external genitalia at 4 consecutive visits). Table 2 shows incidence rates for genital findings with and without epithelial disruption by treatment group and location on the genitalia. Genital findings with intact epithelium were more common than those with disrupted epithelium. Numbers of incident events and incidence rates in each category were similar in the 2 groups. Group assignment and number of applicators used per week were not associated with the prevalence or incidence of disrupted or intact genital findings at any location (Table 3).
Safety: Vaginal Flora Changes
Table 4 shows the prevalence and incidence of BV and yeast infections, with no significant differences between the 2 groups. A total of 22% of women had BV and 38% had a yeast infection at least once during the study. At most visits, the percentage of women with normal vaginal flora (Nugent score: 0-3) was similar in both groups (68%-82% of women at each study visit), but Carraguard users were more likely to have intermediate flora (Nugent score 4-6) and placebo users were more likely to have BV (Nugent score: 7-10). These differences reached statistical significance at months 7 (P = 0.04) and 12 (P = 0.05). Group assignment and the number of applicators used per week were not significantly associated with the prevalence or incidence of symptomatic or asymptomatic BV.
Safety: Other Clinical Findings
Only women with normal (class 1) Pap smears at screening were enrolled. Women were tested again at month 12: 1 woman (in the Carraguard group) had a class 3 Pap smear (dysplasia), and 29 women (17 cases in the Carraguard group and 12 cases in the placebo group; P = 0.20), or 20%, had a class 2 Pap smear (atypical, inflammatory, or uterine cells seen). For comparison, in the population of all women screened (ie, in the absence of gel use), the prevalence of class 2 Pap smears was 31% and that of class 3/4 smears was 3%. There were no significant differences between the 2 groups in any other clinical examination findings or self-reported symptoms (data not shown).
Safety: Adverse Events
Fifteen SAEs were reported during the trial (8 in the Carraguard group and 7 in the placebo group), but only 1 was judged as possibly related to study gel (a kidney infection in the Carraguard group). The other 14 SAEs included hospitalization for normal labor and delivery (4 SAEs), motorcycle accident (2 SAEs), appendicitis (2 SAEs), cesarean section, premature contractions during pregnancy, epileptic seizures, kidney stones, broken wrist, and death attributable to breast cancer. There were a total of 2091 AEs during the trial; almost all were judged by study nurses to be unrelated or probably unrelated to study gel. Table 4 shows that the total number of AEs and AEs coded to the female reproductive system were similarly distributed between the 2 groups. Seventeen AEs were judged as possibly or probably related to study gel and were also similarly distributed (Table 4). A total of 437 AEs were judged as moderate or severe (236 in the Carraguard group and 201 in the placebo group), and the remainder were judged as mild. Only 1 moderate case of vaginal itching, burning, and pain in the placebo group and no severe AEs were judged possibly or probably related to gel use.
Responses to acceptability questions at months 1, 6, and 12 were evaluated. There were no significant differences in responses to any of the acceptability questions between the 2 groups. Most women at each of the visits reported that they liked the study gel or felt neutral about it; only 2 women ever said that they disliked it. More than 60% of the women in each group at each of these visits said that the volume of gel was just right, but 29% to 39% said that the gel volume was too much. Most women said that gel application was easy (64%-77%) and that the applicator was appealing or neutral (96%-98%). Most women said that the gel had no effect on sexual pleasure (70%-78%) or frequency of sex (85%-88%).
Preliminary Sexually Transmitted Infection Prevention and Contraceptive Effectiveness
There were no incident HIV or HSV-2 infections during this study. The number of incident bacterial STI infections was also low (Table 5), with no significant differences between groups. Eight women became pregnant during the study, 4 in each group; they discontinued gel use at the time of pregnancy. Five pregnancies ended in normal vaginal deliveries, 2 in induced abortions, and 1 in a cesarean section. All 6 babies were normal at birth.
In this year-long microbicide trial with 165 women, Carraguard use was not associated with abnormal genital findings, abnormal vaginal flora, Pap smear abnormalities, or other abnormal clinical signs or symptoms as compared with use of the placebo.
Genital findings with epithelial disruption are considered important safety end points in microbicide trials because breaches of the genital epithelium are likely to accelerate HIV transmission.11,12 In this study, disrupted findings were rare, tended not to persist, and were similarly distributed between the study arms. Abnormal genital findings with intact epithelium, conversely, were common in the study population but were similarly distributed between the 2 groups. Their clinical significance is not known. A literature review showed baseline rates of colposcopic findings (in the absence of or before any type of study intervention) in between 10% and 50% of sexually active women.21-26 In one of these studies21 and in the study presented in this article (data not shown), genital findings were associated with sexual intercourse, tampon use, the presence of RTIs, and self-reported genital symptoms. We therefore conclude that the genital findings in this study could have been caused by a variety of background factors and fell within a normal range.
Changes in the vaginal flora are tracked in most microbicide trials because vaginal flora disturbances have been associated with increased HIV transmission.11,12,27 In this study, most women in both groups had normal vaginal flora at screening and during the trial. There were no differences between groups in vaginal flora, Pap smear abnormalities, any other clinical examination findings, self-reported symptoms, or AEs. We therefore conclude that Carraguard use is not associated with abnormal vaginal flora or other clinical abnormalities.
Study gel use was similarly high in both groups throughout the trial. Participants in both groups reported high acceptability of gel use. Although women in the Carraguard group were significantly more likely to drop out early than women in the placebo group (partially because of partner disapproval), overall retention rates were high in both groups (84% and 95%) and no evidence was found that early dropout was associated with experiencing an RTI, genital findings, gynecologic complaints, perceived side effects of gel use, or poor self-reported acceptability. We therefore conclude that Carraguard use was acceptable to most (but not all) Thai women.
This study was not designed or powered to determine the efficacy of Carraguard in preventing STIs or pregnancy. As expected, there were no HIV or HSV-2 seroconversions and few incident bacterial STIs and pregnancies.
A few limitations of this study, and of most other microbicide trials, should be noted. The study participants were carefully selected, and are therefore not representative of the general population. Several key variables (eg, gel use adherence, sexual activity, acceptability) were assessed by self-report, and were therefore prone to social desirability bias. Finally, we cannot rule out the presence of ascertainment bias in the clinic examination data.
In summary, we conclude that Carraguard can safely be used an average of 4 times per week with or without vaginal sex and is acceptable to many Thai women. Based on these and other findings, a phase 3 trial is ongoing at 3 sites in South Africa to determine the effect of Carraguard use on HIV and STI prevention.
The authors thank Chris Elias; Jordan Tappero; Philip Guest; Beverly Winikoff; Charlotte Ellertson; Robin Maguire; the staff of the Thailand Ministry of Public Health-US CDC Collaboration Laboratory, Data Management, and Administration Sections; the Chiang Rai Health Club; Chiang Rai Public Health Office; Chiang Rai Hospital; and Population Council offices in Bangkok and New York.
1. United Nations Joint Program on AIDS. Report on the Global HIV/AIDS Epidemic. Joint United Nations Programme on HIV/AIDS
. Geneva, Switzerland: UNAIDS; 2004.
2. van de Wijgert J, Coggins C. Microbicides to prevent heterosexual transmission of HIV: ten years down the road. AIDScience
3. Pearce-Pratt R, Phillips DM. Sulphated polysaccharides inhibit lymphocyte-to-epithelial transmission of human immunodeficiency virus-1. Biol Reprod
4. Maguire RA, Bergman N, Phillips DM. Comparison of microbicides for efficacy in protecting mice against vaginal challenge with herpes simplex virus type 2, cytotoxicity, antibacterial properties, and sperm immobilization. Sex Transm Dis
5. Nahmias AAJ, Kibrick S, Bernfeld P. Effect of synthetic and biological polyanions on herpes simplex virus. Proc Soc Esp Biol Med
6. Zacharapoulos V, Phillips D. Vaginal formulations of carrageenan protect mice from herpes simplex virus infection. Clin Diagn Lab Immunol
7. Maguire RA, Zacharopoulos VR, Phillips DM. Carrageenan-based nonoxynol-9 spermicides for prevention of sexually transmitted infections. Sex Transm Dis
8. Elias CJ, Coggins C, Alvarez F, et al. Colposcopic evaluation of a vaginal gel formulation of iota-carrageenan. Contraception
9. Coggins C, Blanchard K, Alvarez F, et al. Preliminary safety and acceptability of a carrageenan gel for possible use as a vaginal microbicide. Sex Transm Infect
10. Kilmarx PH, Supawitkul S, Wankrairoj M, et al. Explosive spread and effective control of human immunodeficiency virus in northernmost Thailand: the epidemic in Chiang Rai province, 1988-99. AIDS
11. International Working Group on Vaginal Microbicides. Recommendations for the development of vaginal microbicides. AIDS
. 1996;10 (Suppl):S1-S6.
12. Mauck C, Rosenberg Z, van Damme L, for the International Working Group on Microbicides. Recommendations for the clinical development of topical microbicides-an update. AIDS
13. World Health Organization. Manual for the Standardization of Colposcopy for the Evaluation of Vaginally Administered Products: An Update
. Geneva, Switzerland: World Health Organization; 2000.
14. Bollen L, Kilmarx PH, Wiwatwongwana P. Photo Atlas for Microbicide Evaluation
. Bangkok, Thailand: Thailand Ministry of Public Health-US Centers for Disease Control and Prevention Collaboration; 2002.
15. Young NL, Kilmarx PH, Borchardt K, et al. Inhibitory effects of vaginal microbicides on COBAS Amplicor and Gen-Probe Chlamydia trachomatis and Neisseria gonorrhoeae, and InPouch Trichomonas vaginalis tests [abstract]. Int J STD AIDS
. 2001;12(Suppl 2):193.
16. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram stain interpretation. J Clin Microbiol
18. World Health Organization Collaborating Center for International Drug Monitoring. WHO Adverse Reaction Dictionary
. Uppsala, Sweden: World Health Organization; 1991 and 1993 editions.
19. Hardin JW, Hilbe JM. Generalized Estimating Equations
. Boca Raton, FL: Chapman & Hall/CRC; 2003.
20. Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model
. New York: Springer; 2000.
21. Fraser IS, Lahteenmaki P, Elomaa K, et al. Variations in vaginal epithelial surface appearance determined by colposcopic inspection in healthy, sexually active women. Hum Reprod
22. Roddy RE, Cordero M, Cordero C, et al. A dosing study of nonoxynol-9 and genital irritation. Int J STD AIDS
23. Stafford MK, Ward H, Flanagan A, et al. Safety study of nonoxynol-9 as a vaginal microbicide: evidence of adverse effects. J Acquir Immune Defic Syndr Hum Retrovirol
24. Amaral E, Faundes A, Zaneveld L, et al. Study of the vaginal tolerance to Acidform, and acid-buffering, bioadhesive gel. Contraception
25. Mauck C, Weiner DH, Ballagh S, et al. Single and multiple exposure tolerance study of cellulose sulfate gel: a Phase 1 safety and colposcopy study. Contraception
26. Low-Beer N, Gabe R, McCormack S, et al. Dextrin sulfate as a vaginal microbicide: randomized, double-blind, placebo-controlled trial including healthy female volunteers and their male partners. J Acquir Immune Defic Syndr
27. Sewankambo N, Gray RH, Wawer M, et al. HIV-1 infection associated with abnormal vaginal flora morphology and bacterial vaginosis. Lancet
© 2006 Lippincott Williams & Wilkins, Inc.