Of 4.2 million new adult HIV-1 infections worldwide in 2004, almost half were in women, with the vast majority infected through heterosexual transmission . As women become increasingly affected by HIV, a safe, effective, acceptable female-initiated method of HIV prevention is urgently needed to decrease HIV sexual transmission between women and men. While the male condom is a very effective method of HIV prevention, there are many situations where women lack the power to negotiate condom use . Condoms are rarely used in stable relationships  and are not appropriate for couples desiring pregnancy . Topical vaginal microbicides designed to prevent sexual HIV transmission are being developed; more than 25 candidate products are currently in the development pipeline and 5 candidates are in large efficacy trials .
The real-world effectiveness of any microbicide will depend on both the clinical efficacy under ‘ideal use’ conditions, and on women's and couples’ willingness to consistently use the product. Evaluation of the acceptability of candidate microbicide products during early stages of clinical trials can provide useful information for product formulation; in later stages it can provide important data to inform marketing and promotion messages.
A recent review of microbicide acceptability data highlighted the lack of studies focusing on male partners and couples . Most studies that included men have focused on hypothetical rather than actual product use [7–9]. The few studies  where men were asked about actual use were phase I studies in which condom use was required by protocol, and men were not directly exposed to the product. Microbicides are used in the context of sexual partnerships and decision making, and men's views about these products will have a major impact on their eventual acceptance and use.
We report on the acceptability of Carraguard, the Population Council's lead non-contraceptive microbicide candidate, among couples enrolled in a 6-month safety and acceptability trial in Chiang Rai, Thailand. The risk of HIV transmission among married couples in Chiang Rai was relatively high in the early 1990s; HIV-1 seroprevalence peaked at 17.3% in male army conscripts and at 10.8% in young primigravidas . Infection rates had declined by the time of this trial.
Acceptability data were collected as part of a 6-month randomized, placebo-controlled, triple-blinded safety study of Carraguard among low risk, heterosexual couples in Chiang Rai, Thailand. Following a detailed study introduction session, individual and couple counselling, eligibility screening, and completion of a comprehension test and informed consent, couples were randomized to use either Carraguard or placebo gel. Participants were asked to apply study gel up to 1 h before sexual intercourse for each act of intercourse during the trial period, including during menses if they had intercourse at that time. They were advised that the gel was not designed for anal or oral use, and were instructed to use condoms if they felt they were at risk for sexually transmitted infections (STI) or HIV. Women participants were asked to record intercourse, gel use, and condom use in a coital diary. Couples attended the study clinic monthly for an interview, genital examination, and laboratory testing for reproductive tract infections.
Carraguard, a 3% carrageenan derived from seaweed, is a large, negatively charged polymer which inhibits uptake of HIV into target cells . The placebo gel formulation contains 2.5% methyl cellulose. The placebo looks, feels, smells, and tastes the same as Carraguard; both are clear gels with similar viscosity. Both gels were packaged in opaque Microlax-type single-use applicators (Norden Pac International AB, Kalmar, Sweden) designed to dispense 4 ml of gel at each application.
Safety data from this trial and from a trial of Carraguard use among low-risk women showed no product-related risk of genital lesions, disrupted vaginal flora, or other adverse effects [12–13].
Couples were recruited from family planning clinics and through community meetings coordinated with community health centre staff and village health volunteers. A total of 55 HIV-negative couples enrolled in the trial. Couples were eligible to participate if both members were monogamous (reported no other sexual partner in the previous year, and expressed no intent to have another partner during the study period), HIV-negative and free of sexually transmitted infections at screening, not regular condom users (i.e., used condoms < 25% of the time), and planned to have vaginal sexual intercourse together at least once per week on average.
Acceptability data collection
During clinic visits at day 14, month 1, month 3, and month 6, research staff interviewed male and female participants separately using a structured questionnaire. Interviewers asked about the volume of gel, applicator ease of use, timing of gel use, effect on sexual pleasure or frequency, overall rating of the product, and also asked open-ended questions about best or worst aspects of the gel. Men were also asked if they ever inserted the study gel for their partner. At the 6-month closing visit additional questions were asked about whether opinions about the gel changed over time, covert use, hypothetical characteristics of an ideal microbicide, and willingness to use or recommend use in the future.
Additional acceptability and use dynamics outcomes included compliance with study gel use assessed by questionnaire and coital diaries, and tracking of returned used and unused applicators. Applicators returned as ‘used’ were weighed but no assay was conducted to verify the applicators had been inserted in the vagina. Female participants referred to their coital diaries at each visit to complete questions about frequency of sex and gel and condom use. Men relied on recall alone to answer these questions.
Data were entered using EpiInfo 6.04, converted into SAS (SAS 8.2, SAS Institute, Cary, North Carolina, USA) data files, and analysed using SAS and Stata (Stata 8.0, StataCorp, College Station, Texas, USA). As there were no significant differences in sociodemographic status or acceptability responses between the Carraguard and placebo gel groups, the groups were combined for this analysis. For continuous variables, the mean and standard deviation are presented for women, men, and the difference between them, along with the correlation between women's and men's responses. For nominal and ordinal variables, we present frequency responses by sex, percent agreement, and kappa statistics for the concordance within couples. The kappa statistic reflects the proportion of agreement corrected for chance; i.e., corrected for the level of agreement expected by chance alone.
Differences between males and females on factors were tested using the chi-square test for marginal homogeneity; for a binary response factor this is equivalent to McNemar's test. Factors associated with acceptability among men were evaluated using the 95% confidence interval for the estimated relative risk. Unless otherwise stated, responses from the closing study visit are presented.
Description of study population
A total of 84 couples were screened and 55 enrolled in the trial. Median age was 33 years for women and 37 years for men, with mean years of education of 5.9 years among women and 6.3 years among men (Table 1). The most common occupation was farming (40.0% of women, 49.1% of men). The most common current contraceptive methods were oral pills (40.0%) and female sterilization (32.7%). Vaginal product use, condom use, and oral sex within the previous year were infrequently reported, and no couples reported anal sex within the previous year. Retention was high, with 52 couples plus one woman (without her partner) completing a full 6 months of follow up.
Adherence and gel use dynamics
Adherence to gel use was high throughout the study. In both Carraguard and placebo groups, approximately 80% of the women used gel for at least 95% of their acts of vaginal intercourse; the mean number of applicators used per week was between 2.2 and 3.1 at all visits. One woman reported concomitant condom and gel use during the study. Women and men reported a similar number of episodes of vaginal intercourse, with a mean of 9.95 per month reported by women and 10.1 per month reported by men (r = 0.93). The number of episodes of intercourse where gel was used was also similar between women and men: 9.7 per month reported by women and 9.7 per month reported by men (r = 0.94).
Women reported inserting gel a mean of 8.9 min before sex; men reported that their partner inserted gel a mean of 10.1 min before sex. A total of 41% of men ever inserted gel for their partner; 13 women (24.1%) reported ever using gel during menses.
For questions that were asked at multiple visits during the study, a summary of trends in responses over time is shown in Fig. 1. Significant linear trends were observed over the study period, with an increasing proportion of women and men reporting that gel use increased sexual pleasure and frequency, and an increasing proportion of women reporting that the volume of gel was too great. There was no significant change in the overall rating of gel by either women or men, which remained high throughout follow up.
The following results refer to responses from the closing visit of the study. A majority of both women (53.8%) and men (63.5%) felt that the volume of gel was just right, but 46.1% of women and 36.5% of men felt it was too great (Table 2). All found the applicator either appealing (women 52.8%; men 67.9%) or neutral. Only one woman found the gel difficult to apply.
Most women (66.0%) and men (71.7%) reported that gel use resulted in increased sexual pleasure as well as increased frequency of sex (women 54.7%; men 62.3%). One man found that gel use had a negative effect on sexual pleasure and decreased frequency of sex. A large majority (84.9% of women; 92.4% of men) reported that sex with gel was different from sex without gel.
Only 15.1% of women believed that it would be possible to use gel without their partner noticing. Men were more likely to believe that covert use was possible, with 43.4% stating that their partner could use the gel without their noticing. All the women in the study and all but one of the men would recommend the gel to a friend if it were available.
When asked about desired characteristics of a microbicide product, a large majority of women and men viewed extra lubrication as very important (60.4% and 81.1% respectively) (Table 3). Conversely, almost none felt it was important for a microbicide to dry the vagina. There were several characteristics that men were significantly more likely than women to rate as important: a product that makes the vagina tight (rated ‘very important’ by 43.4% of women and 90.6% of men); a product that slows the time to ejaculation (rated ‘very important’ by 22.6% of women and 71.7% of men); a product that smells good (rated ‘very important’ by 32.1% of women and 77.3% of men); and a product that tastes good (rated ‘very important’ or ‘somewhat important’ by 28.3% of women and 56.6% of men). Both women and men preferred a microbicide that was contraceptive (75.5% of women, 83.0% of men) and also believed their partner would be more likely to use a microbicide that was contraceptive (71.1% of women, 84.9% of men). At the same time, 90.6% of women and 96.2% of men felt that if microbicides protected against both disease and pregnancy, then men would be less willing to use condoms.
For both opinions about study gel and desired product characteristics, the concordance within couples was generally very low, with kappa statistics almost all below 0.20 and several below zero. For example, there were negative kappa statistics for whether covert gel use was possible, whether extra lubricant was an advantage, whether good taste was an important attribute of the gel, or whether a contraceptive or non-contraceptive microbicide was preferred by either the respondent or partner.
In responding to open-ended questions on what they liked about the gel, women participants most frequently mentioned increased lubrication during sex (83.0%) and a cool sensation (24.5%). Other aspects mentioned by women included making sex more pleasurable, a clean sensation, and the fact that there was no irritation or ‘allergy’. The only negative aspect commonly mentioned was that the gel was too messy, too wet, or too watery (56.6%). Despite this common complaint, only two women reported using a vaginal hygiene or douching product during the study. Two women (3.8%) felt the gel was too viscous, and two that it smelled ‘medicinal’. In men's responses to these open-ended questions, most (80.8%) also mentioned lubrication as an aspect of the study gel that they liked. Other common favourable gel features reported by men included: no side effects or irritation, making sex more pleasurable for both partners, prolonging the time to ejaculation, and the gel being clear, colourless, odourless, and giving a cool sensation. Regarding dislikes about the gel, 24 (45.2%) men stated it was too wet or too messy, and two (3.8%) found it too viscous.
Factors associated with acceptability
Acceptability among women, as assessed by an overall gel rating of ‘liked somewhat’ or ‘liked very much’, was so high (92.4%) that it was not possible to identify sociodemographic factors or relationship characteristics associated with acceptability. Women who reported less consistent gel use, i.e. the 11 women who reported gel use in less than 95% of sex acts, were less likely to report liking the gel ‘somewhat’ or ‘very much’ (72.7% of inconsistent users versus 97.7% of consistent users, P = 0.01). Among men, when acceptability was defined as liking the gel ‘somewhat’ or ‘very much’, having a higher level of education and using a contraceptive method other than sterilization were both associated with lower acceptability, with relative risks of 0.5 [95%confidenceinterval(CI),0.3–0.7] and 0.6 (95% CI, 0.5–0.8) respectively. Age, age difference between partners, and duration of partnership were not associated with acceptability.
Adherence to use of Carraguard and placebo gel was high and sustained throughout this study of low-risk couples. Overall acceptability was very high among both men and women, with most reporting that they liked the gel and would recommend it to a friend in the future. A majority of both men and women reported that gel use was associated with increased frequency of intercourse and increased sexual pleasure.
Of note is the finding that a rising proportion of both men and women over time reported increased pleasure and frequency of sex with gel use. It is possible that studies with shorter periods of follow up may not adequately capture views of products intended for use over the longer term; on the other hand, it is also possible that ‘first impressions’ may be more important in determining future decisions about using a vaginal product in non-research settings.
The reporting of frequency of sex and gel use was relatively well correlated between partners. Since this information was collected differently among women and men (diary cards for women and recall for men) it is not surprising there are some discrepancies within couples. Concordance within couples for evaluation of study gel, as measured by kappa scores, was almost universally low: almost all scores were lower than 0.20; >0.60 is usually considered substantial agreement . This suggests that the two members of a couple were little more likely to agree on gel attributes than would be expected by chance alone; indeed, for those variables with negative kappa scores, couples agreed with each other less frequently than would be expected by chance. Thus, although on many attributes men and women overall had similar views, there was little tendency for couples to agree with their partner any more than with another individual. This suggests that study designs eliciting views of one member of a couple as a proxy for the other may be of limited value.
However, kappa scores are strongly influenced by prevalence: for variables where the prevalence of a given response is very high or very low, then the upper limit of kappa is much less than 1 and low kappa values are to be expected . Thus, for example, when 84.9% of women and 92.4% of men responded that sex was different when using gel, a low kappa score (0.07) is to be expected. Despite the low kappa score, 81.1% of couples overall agreed that sex with gel was different.
There was an important difference of opinion between women and men about whether the gel could be used without the male partner knowing: only 15.1% of women agreed that they could, compared to 43.4% of men, and the kappa score was −0.12 for within-couple agreement. The possibility of covert use by women has been one key rationale for the need for microbicides, particularly in settings where women may lack power to negotiate condom use. While few of the women in this study felt covert use would be possible, many of their partners disagreed. Covert use may be generally more acceptable in casual relationships than in stable partnerships [16,17], but is likely to be needed by women in both situations, including, for example, HIV-discordant relationships where condom use cannot be negotiated. An alternative to difficult discussions with partners about disease risk and trust that has been proposed by participants in some microbicide acceptability studies is ‘indirect’ covert use . In this case, a woman would inform her partner she is using a product but suggest the reason is not disease prevention but rather contraception or enhancement of sexual pleasure. Although marketing this kind of pretext may not be appropriate for a variety of reasons, it may certainly be possible to promote ‘dual use’ for microbicides: both disease prevention and enhancement of pleasure (and ‘triple’ use for those with contraceptive properties). Indeed, some researchers have argued that the first step in product innovation should be the identification of characteristics that enhance sexual pleasure, only then followed by the addition of HIV, STI and/or pregnancy prevention features . Several studies of men's attitudes to potential reproductive health products have noted that men may place particularly high importance on the impact of these products on sexual pleasure [8,19].
Although both men and women participants in this study reported that the increased lubrication associated with gel use had a positive effect on sexual frequency and pleasure, this may not be consistent across cultures. The preference in many sub-Saharan African countries for ‘dry sex’ [7–9,20] may require both a variety of formulations and locally tailored marketing messages in order to ensure widespread acceptability.
Our study had a number of limitations. First, this was a relatively small safety trial, and we enrolled low-risk couples in stable partnerships. Recruitment was challenging and our study population represents a select group where both partners were willing to enter the trial for a 6-month period. This limits the generalizability of our findings, as the experience of a higher-risk target population may be quite different. Second, adherence and acceptability among study participants were so high that we were unable to explore the influence of male partner attitudes or attitude discordance among partners on use behaviour or acceptability. Third, study participants were infrequent condom users, making it difficult to compare the acceptability of microbicides with that of condoms. Finally, acceptability data were collected in face-to-face interviews with clinic research nurses; there may have been a tendency for participants to give what they considered to be ‘desirable’ responses.
In conclusion, our data indicate that Carraguard is highly acceptable to low-risk heterosexual men and women in Thailand. This study is the first to report on both men's and women's views of a candidate microbicide in the context of actual use, without concurrent condom use. This allowed valuable insights into microbicide use dynamics and the reactions to gel use from both members of participating couples. The positive response to study gel from both men and women bodes well for the acceptability of this or similar products once they are marketed. Further research of this nature among more diverse populations and types of partnerships are critical, and a full assessment of ‘real-world’ use-effectiveness is likely to require licensing and widespread use of a microbicide product.
2. Elias CJ, Coggins C. Female-controlled methods to prevent sexual transmission of HIV. AIDS 1996; 10:S43–S51.
3. Ali MM, Cleland J, Shah IH. Condom use within marriage: a neglected HIV intervention. Bull World Health Org 2004; 82:180–186.
4. McCormack S, Hayes R, Lacey CJN, Johnson AM. Microbicides in HIV prevention. Br Med J 2001; 332:410–413.
6. Mantell JE, Myer L, Carballo-Dieguez A, Stein Z, Ramjee G, Morar NS. Microbicide acceptability research: current approaches and future directions. Soc Sci Med 2005; 60:319–330.
7. Ramjee G, Gouws E, Andrews A, Myer L, Weber AE. The acceptability of a vaginal microbicide among South African men. Int Fam Plan Perspect 2001; 27:164–170.
8. Coggins C, Blanchard K, Friedland B. Men's attitudes towards a potential vaginal microbicide in Zimbabwe, Mexico and the USA. Reproductive Health Matters 2000; 8:132–141.
9. van de Wijgert JHHM, Khumalo-Sakatukwa GN, Coggins C, Dube SE, Nyamapfeni P, Mwale M, et al
. Men's attitudes toward vaginal microbicides and microbicide trials in Zimbabwe. Int Fam Plan Perspect 1999; 25:15–20.
10. Bentley ME, Fullem AM, Tolle EE, Kelly CW, Jogelcar N, Srirak N, et al
. Acceptability of a microbicide among women and their partners in a 4-country phase I trial. Am J Public Health 2004; 94:1159–1164.
11. Kilmarx PH, Supawitkul S, Wankrairoj M, Uthaivoravit W, Limakarnjanarat K, Saisorn S, et al
. Explosive spread and effective control of human immunodeficiency virus in northernmost Thailand: the epidemic in Chiang Rai province, 1988–99. AIDS 2000; 14:2731–2740.
12. 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: Findings from a phase II clinical trial. JAIDS epub ahead of print, Oct 2, 2006.
13. Kilmarx PH, Wankrairoj M, Achalapong J, Blanchard K, Chaikummao S, Connolly C, et al. Safety of Carraguard® use by heterosexual men in a six-month clinical trial in Thailand. XV International Conference on AIDS
. Bangkok, July 2004 [abstract ThOrB1374].
14. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159–174.
15. Kraemer HC, Bloch DA. Kappa coefficients in epidemiology: An appraisal of a reappraisal. J Clin Epidemiol 1988; 41:959–968.
16. Koo HP, Woodsong C, Dalberth BT, Viswanathan M, Simons-Rudolph A. Context of acceptability of topical microbicides: sexual relationships. J Soc Issues 2005; 61:67–93.
17. Woodsong C. Covert use of topical microbicides: Implications for acceptability and use. Int Fam Plan Perspect 2004; 30:94–98.
18. Severy L, Newcomer S. Critical issues in contraceptive and STI acceptability research. J Soc Issues 2005; 61:45–65.
19. Grady WR, Klepinger DH, Neson-Wally A. Contraceptive characteristics: the perceptions and priorities of men and women. Fam Plan Perspect 1999; 31:168–175.
20. Braunstein S, van de Wijgert J. Cultural Norms and Behavior Regarding Vaginal Lubrication During Sex: Implications for the Acceptability of Vaginal Microbicides for the Prevention of HIV/STIs. New York: Population Council; 2003.