Wallace, Andrea BS; Thorn, Mitchell MS; Maguire, Robin A. BS; Sudol, Kristin M. BA; Phillips, David M. PhD
A NUMBER OF RESEARCH groups are involved in the research and development of a topical formulation that women could use vaginally for protection against infection by HIV and other sexually transmitted pathogens. Such formulations are termed “microbicides.” Currently, it is estimated that there are approximately 65 microbicides in various stages of development.1 There is, however, a growing concern about the feasibility, both logistical and financial, of conducting efficacy (phase III) trials that will yield definitive results. A major concern is that subjects enrolled in the trial are asked to use the study product, microbicide or placebo, outside of the clinical setting, thereby eliminating the ability of trial staff to monitor whether the study product is actually used.
Many factors could influence a subject’s ability or willingness to use the study product. Trial participants might be unable to negotiate use of the product, just as women are sometimes unable to negotiate the use of condoms. Fear of abuse, financial insecurity, and a need to prove fertility are all factors in women’s negotiation abilities. Although exposed to the product, a subject’s partner could dislike perceived differences in sensation during intercourse. The partner, other family members, friends, or even the study participant herself could have or could develop fears or superstitions about the product. Faced with these difficulties, the subject might not be willing to tell the study clinicians that she has not complied with the product use regimen. She also might not want to appear disrespectful or uncooperative. She could fear loss of study benefits such as medical care and could fear that she will be dropped from the study if she reports noncompliance. In addition, a woman might not fully understand the product use regimen until after she has been enrolled, at which time she could feel that there is some cultural taboo associated with the regimen. This myriad of circumstances could have a serious impact on compliance. As a result, assessment of microbicide effectiveness would be difficult.
Of the 4 microbicide phase III effectiveness clinical trials conducted thus far,2 none have successfully proven effectiveness. Often, this is attributed to the possible toxicity of the study product. However, in each case, there was no means by which to measure a participant’s compliance with regard to product use other than by the participant’s self-reporting.
Study products that cannot be administered within the clinical setting such as contraceptives are problematic insofar as the participant’s self-reporting might not be consistent with actual compliance of product use. For this reason, it has been theorized that the level of effectiveness of a product not administered within the clinical setting could be considerably lower than the anticipated level of product effectiveness. Reduction in the level of effectiveness can be attributed to the ability or willingness of the participant to use the product, as well as the extent to which she actually does use it.
Six microbicide phase III effectiveness clinical trials are currently in the planning stages. All trials could be seriously jeopardized if subjects fail to use the study product consistently. Consequently, trials are being designed to increase the interaction between trial staff and participants and to encourage consistent product use.
In a number of the planned microbicide trials, the product is supplied in a single-use applicator. To address concerns over proper disposal of the applicators, study participants are encouraged to return all used and unused applicators to the study site. It is conceivable that “used” applicators returned empty might actually have been emptied ex vivo. In other words, the subject could have cheated by removing the gel from the supposed “used” applicator without actually having inserted the gel. Therefore, a method by which study clinicians could determine whether the applicators had been inserted or merely emptied ex vivo is critical in determining the level of participant compliance.
The assays described here are designed so that a distinction could be made between applicators that have been used in compliance with the study regimen and those that have not. Although the assays cannot distinguish if the applicators are used before, after, or outside of sexual intercourse, they can distinguish whether an applicator has been exposed to the vagina. The rationale is that exposure to the vagina is likely to be a reasonable measure of compliance. The information yielded by the assay could be useful as an exclusion criterion for microbicide trial participants that are chronically not complying with the product use regimen. In addition, the results could be predictive of the overall product use compliance rate within a population. Social intervention could then be used in an effort to increase overall compliance. Furthermore, the assay could aid in the analysis or the interpretation of data collected in a microbicide effectiveness trial. This article describes 3 possible analytical methods. The comparison of the 3 methods identifies the most rapid, direct, and accurate analytical assay, which could be carried out in the clinical setting to evaluate participant compliance to study product use.
Materials and Methods
Twist-off top Microlax-type applicators (Norden Pac, Kalmar, Sweden) that had been filled with 7 mL of a 2.5% methylcellulose gel were used. Manufacturing of the methylcellulose gel and applicator filling was carried out at Clean Chemical Sweden AB (Borlange, Sweden).
Twenty women participated in the study. Each study participant was given 2 applicators, each in a 50-mL plastic centrifuge tube (VWR). One tube was labeled “FOR vaginal use” and the other labeled “NOT for vaginal use.” Women were instructed to use the applicator in the tube labeled “FOR vaginal use” by twisting off the top, inserting the elongated (5 cm) tip of the applicator into their vagina, expelling all of the contents by squeezing the tube of the applicator (also 5 cm in length) into their vagina, and returning the “used” applicator to its respective centrifuge tube. For the applicator-containing tube labeled “NOT for vaginal use,” participants were instructed to twist off the top, not to insert the applicator vaginally, to expel the contents ex vivo, and to return the “control” applicator to its respective centrifuge tube. All applicator-containing tubes were returned to the laboratory for analytical evaluation.
Visual Observation of Applicators
One tube was labeled “FOR vaginal use” and the other tube was labeled “NOT for vaginal use.” The applicator-containing tubes were collected over a 1-month period, such that some applicators were analyzed within a few days after their return, whereas others were analyzed up to 1 month after return to the laboratory.
Staining and Examining Applicators.
Several basic and acid dyes were tested for staining applicators. All of the dyes stained vaginal mucous on applicators. However, some dyes stained the microbicide Carraguard a slightly different color than our inactive methylcellulose formulation. None of these dyes were ultimately chosen, because they would confound the blinding of our studies if clinic workers could distinguish between the active and inactive formulations. The dye trypan blue was selected because it stained Carraguard and methylcellulose the same color. For staining, the large portion of each applicator was held while the tip was sprayed with a solution of 1% trypan blue for 5 seconds. After 10 seconds, the applicator was washed for a few seconds under running cold tap water and dried overnight. Four judges examined the applicators, looking for smearing and consistency of color. Figure 2B demonstrates visually the differences between an unused applicator (left) and a used applicator (right) after staining.
The 40 applicators were randomly coded and labeled with a black indelible marker by 1 individual. Only this individual knew which applicators came from the tubes labeled “FOR vaginal use” and which applicators came from tubes labeled “NOT for vaginal use.” This individual did not serve as a judge.
Four examiners served as judges. Each judge examined the applicators in private both before and after staining. Each judge evaluated the applicators independently without consulting the other examiners. Based on visual evaluation, examiners recorded whether they thought each applicator had been used or had served as a control and gave the results to the individual who numbered the applicators. Judges were instructed not to discuss their findings with anyone.
The microbiologic analysis is based on detection of growth of Lactobacillus. Growth medium was made by dissolving 55 g of powdered Difco Lactobacillus MRS Broth (Becton Dickinson, Sparks, MD) in 1 L of distilled water. Broth was autoclaved and cooled to room temperature. Ten milliliters of medium was added to 40 sterile 50-mL centrifuge tubes (VWR, Westport, CT). The tip of each applicator was placed in each tube. Tubes were swirled for a few seconds and incubated at 37°C. After 24 hours, 200-μL samples from each tube were placed into wells of a 96-well flat-bottom plate (NUNC-Immuno MaxiSorp; NUNC, Rochester, NY). Optical density was measured at 530 nm on a Dynex Multiplate Reader (Dynex, Chantilly, VA).
To examine whether growth of lactobacilli could be used to determine whether an applicator tip had been in the vagina, we exposed Lactobacillus culture medium to the tips of 20 used applicators and 20 unused applicators. This was done after the tips had been stained, because staining was found to have no effect on Lactobacillus. The optical density in 24 hours increased for medium in which tips from used applicators were placed (Fig. 1A), but it did not increase in density for media in which unused applicator tips had been placed (Fig. 1B). All of the used applicators had optical density readings above 0.4, (range, 0.46–1.05). All of the unused applicators had readings below 0.3 (range, 0.24–0.30).
Observation of Unstained Applicators.
Each judge independently made assessments as to whether the applicator had or had not been in the vagina. Of 40 applicators, 21 applicators were judged correctly by all 4 of the observers. Ten were judged correctly by 3 of the 4 observers. Seven were judged correctly by 2 of the 4 observers. The remaining 2 were judged incorrectly by all of the 4 observers.
Observation of Stained Applicators.
As described in the “Methods” section, the 20 applicators that had been exposed to the vagina (used) and the 20 that had been squeezed out but not exposed to the vagina (unused) were dried, stained, and washed. Applicators were stained as described in the “Methods” section. Four masked observers then independently viewed the 40 applicators. Thirty-nine of the applicators were judged correctly as having been exposed or not exposed to the vagina. The remaining applicator was judged correctly by 3 of the 4 observers.
Three methods were tested to determine if an applicator had been exposed to the vagina. Visual examination of unstained applicators proved somewhat successful, as evidenced by the fact that all 4 examiners made correct judgments in 50% of the visual examination cases. Visual examination of applicators, therefore, could be of some value as a preliminary tool.
Growth of lactobacilli was 100% predictive. However, it has been shown that lactobacilli can be cultured from only approximately 75% of women.3 Women with vaginitis tend not to have culturable lactobacilli.4 Thus, in an area of high prevalence of sexually transmitted infections and vaginitis such as South Africa, the percent of women with culturable lactobacilli could be low.
For these reasons, staining applicators is the method of choice. It is inexpensive, simple, rapid, and accurate. The method requires no equipment and could be carried out by an unskilled person in a minimal clinical setting in the developing world. In this study, applicators were allowed to dry so that they could be judged independently by 4 different judges at different times. In the clinical setting, however, drying of applicators might not be necessary. Applicators could simply be sprayed with the dye and viewed immediately by a single judge. This method allows for accurate distinguishing of used and unused applicators (Fig. 2).
Although staining of applicators can differentiate used applicators from unused applicators, the method cannot determine whether the applicators have been used properly. For example, the method cannot determine whether the applicator was used by the subject or by another woman. Likewise, staining of applicators does not allow one to determine whether the subject used the applicator before or after intercourse or, indeed, whether she had intercourse at all. Thus, although staining of applicators is valuable in determining if the applicator was used, it is not synonymous with compliance.
There are several ways that compliance tests could be implemented in a clinical trial. Vaginal applicator use could be compared among different populations. An accurate correlation between self-reporting and frequency of microbicide use could be drawn. Data on applicator use could also be used as exclusion criteria to eliminate noncompliant subjects from a trial. Determination of compliance might also be used in data analysis where compliance could be correlated to efficacy. In addition, compliance testing could be used in studies to determine the impact of a social intervention strategy on compliance.
In conclusion, visual examination of stained applicators is accurate and simple. The only method currently used to determine compliance is self-reporting. Indirect evidence suggests that self-reporting is unreliable. It is hoped that the method of staining applicators will be a valuable tool to accurately measure compliance. Without such assays, it is possible that efficacy trials of microbicides could be seriously flawed.
1. Harrison P, Rosenberg Z, Bowcut J. Topical microbicides for disease prevention: Status and challenges. Clin Infect Dis 2003; 36:1290–1294.
2. Van Damme L, Ramjee G, Alary M, et al. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: A randomised controlled trial. Lancet 2002; 360:971–977.
3. Eschenbach DA, Davick PR, Williams BL, et al. Prevalence of hydrogen peroxide-producing Lactobacillus species in normal women and women with bacterial vaginosis. J Clin Microbiol 1989; 27:251–256.
4. Hill GB, Eschenbach DA, Holmes KK. Bacteriology of the vagina. Scand J Urol Nephrol Suppl 1984; 86:23–39.