Several clinical trials failed to demonstrate efficacy of microbicides in reducing human immunodeficiency virus acquisition. Lack of protocol adherence and product use may have contributed to these disappointing results.1–3 Information is usually obtained in face-to-face interviews, participant-kept diaries, and audio computer-assisted self-interviews—all of which can overestimate adherence and product use.4 Objective methods to measure product use are essential in the determination of product efficacy. Several studies have used trypan blue1,5,6 and FD&C Blue No. 11,7–9 staining techniques to detect vaginal mucus as a proxy for vaginal insertion of applicators and product use.
Previous studies have evaluated vaginal staining of low-density polyethylene (LDPE) Microlax (Norden Pac, Kalmar, Sweden)-like applicators1,5,7,9 and HTI Plastics (Lincoln, NE) polypropylene applicators.6,8 Different plastics exhibit different surface qualities, which may influence the retention of vaginal mucus in a distinct streaking pattern. Dye staining reached high accuracy for the identification of vaginal mucus on gel-filled polypropylene HTI applicators6 as well for gel-filled LDPE Microlax applicators.1,5,7,9 However, Austin et al questioned the sensitivity of FD&C Blue No. 1 for HTI Plastics polypropylene applicators—only 81% to 95% of used and 86% to 93% of unused applicators were correctly identified—when compared against gram-stained smears from applicators correctly identifying 99% of used applicators.8 No information exists for the suitability of dye staining of applicators filled with nongel formulations, and for a third type of plastic—the linear LDPE—which differs structurally from conventional LDPE and has a narrower molecular weight distribution and significantly different rheological properties.10
We analyzed trypan blue staining results of applicators reported as vaginally inserted during 3 clinical trials—using 2 types of plastics (linear LDPE and HTI polypropylene) and 2 product formulations (Carbopol based gel [Noveon, Inc., Cleveland, OH] and a powdered live biotherapeutic product). The phase 1 VivaGel trial (NCT003311032) used HTI polypropylene vaginal applicators. It was a placebo-controlled, double-masked, multisite trial conducted in San Francisco (California, USA) and Kisumu (Kenya), evaluating the safety and tolerability of SPL7013 3% w/w gel (Vivagel® by Starpharma, Melbourne, Australia)11 administered twice daily for 2 weeks in 54 healthy women. The phase 1 LACTIN-V (Osel, Inc., Mountain View, CA) trial (NCT00537576)12 and phase 2a LACTIN-V trial (NCT00635622)13 used linear LDPE applicators in placebo-controlled double-blind trials in San Francisco (California, USA), evaluating the safety and colonization efficiency of Lactobacillus crispatus CTV-05 for prevention of bacterial vaginosis, administered to 12 healthy women and 24 women with bacterial vaginosis, over 2 and 4 weeks, respectively. Details for the 3 studies are published elsewhere.11–13
All 3 trials used identical procedures. At enrollment, the first applicator was always administered in the presence of a clinician, then inserted into the original plastic pouch, labeled, and stored on-site to serve as the positive control. Participants were provided with the remaining applicators, and were instructed to not rinse them after insertion, to repackage each applicator in its original plastic pouch, to place them in a sealable bag, to store at room temperature, and to return them at each follow-up visit. Applicators were tested at their respective sites in Kisumu and San Francisco. Each participant's applicators were tested in 1 batch at her last study visit, along with her positive control and an unused empty applicator kept on-site as a negative control. Consequently, even the first applicator used was stored for a maximum of 2 to 4 weeks, depending on the duration of enrollment in the respective studies. Following standard operating procedures, negative and positive control applicators and all reportedly used applicators were individually sprayed with a 1.0% trypan blue solution in normal saline for 5 seconds, rinsed with tap water for another 5 seconds, placed on a rack, dried overnight, and examined the following day. The same trained clinical staff member read all applicators at the Kisumu site, and a second one read all applicators for the 3 trials at the San Francisco site. Applicators with a streaking along the applicator barrel, which is characteristic for the highly viscous vaginal mucus, were considered used (Fig. 1). Applicators from the LACTIN-V trials with occasional clearly distinguishable blue colored powder remnants just on the tip, but without streaking along the barrel, were considered unused. Because applicators filled with placebo or active product contained the same carrier substance (Carbopol gel in the VivaGel study, preservation matrix in the LACTIN-V trials), no separate analysis comparing staining results between placebo and active arms was performed.
At the Kisumu VivaGel site, all 885 distributed HTI polypropylene applicators (100%) were reported as used, and all were returned emptied at subsequent study visits. At the San Francisco VivaGel site, 423 of the 440 distributed applicators (96.1%) were returned emptied and reportedly used. During the phase 1 LACTIN-V study using linear LDPE applicators, all 60 distributed applicators were returned emptied and reportedly used. For the phase 2a LACTIN-V study, 157 of 168 applicators (93.5%) were returned emptied and reportedly used.
For both the HTI polypropylene and the linear LDPE applicators, all positive controls stained positive and all negative controls stained negative. Of the 1308 reportedly used HTI polypropylene applicators at the 2 VivaGel sites combined, 1271 applicators stained positive (97%). The 37 applicators stained negative were obtained from 12 participants, most of whom failed to use 1 or 2 applicators properly (of 28 applicator doses each), but 2 participants from the Kisumu site misrepresented 9 and 13 noninserted but emptied applicators, respectively, as vaginally used. Fifty-seven of the 60 emptied and reportedly used linear LDPE applicators (95%) in the phase 1 study of LACTIN-V trial stained positive, and 3 applicators from 3 participants stained negative. Of the 157 reportedly used linear LDPE applicators in the phase 2a study of LACTIN-V, 135 stained positive (86%). The 22 applicators stained negative were obtained from 9 different participants, 3 of whom each reported 5 or 6 noninserted emptied applicators as used (of 7 total doses).
Our results support the use of trypan blue staining as an objective, cost-effective, and sufficiently accurate method to detect vaginal exposure of gel-filled HTI polypropylene applicators as well as the previously untested linear LDPE applicators filled with a powdered product. The 1308 reportedly used HTI polypropylene applicators from the VivaGel study tested constitute a sample size significantly larger than that tested in previous studies,8 whereas the 217 returned applicators from the LACTIN-V studies provided data on previously untested linear LDPE applicators and on nongel formulations. Following evidence that previously established trypan blue as an accurate indicator of vaginal exposure1,5,6,9 for 2 plastics, this methodology was primarily used to verify participant-reported product use in clinical trial settings, and not to gain data on the accuracy of applicator staining. Other study designs, which compare staining results with a gold standard, are preferable for this purpose.5,7,8
The lower proportion of positively stained applicators (86%) for the phase 2a LACTIN-V study is likely due to lower actual than reported product use, and not a failure of the staining method. This assumption is based on higher rates of observed nonadherence for other protocol-specified requirements, such as failing to return used applicators, missed follow-up visits, and self-reported sex during periods of requested abstinence. Participants reporting less product use also showed less vaginal colonization with L. crispatus CTV-05.13 Additionally, the majority of misrepresented vaginal applicator use (16 of 22 applicators stained negative) were from only 3 phase 2a LACTIN-V participants who, based on the staining results, presumably routinely emptied applicators extravaginally and used only 1 or 2 doses of 7. Similarly, only 2 of the 36 VivaGel participants accounted for 22 of the 37 applicators stained negative. Although 15% to 25% of participants per study misrepresented vaginal use, most of these 24 participants did so for 1 to 2 applicators, and only 5 of the total 90 study participants routinely averted product use.
Although Austin suggested that more frequently applied amounts of the viscous gel may interfere with the formation of the characteristic mucus streaking on the smooth applicator surface,8 the VivaGel data do not support this hypothesis. After application of 3.5 g of gel twice daily, 97% of reportedly used applicators stained positive when using trypan blue on similar HTI polypropylene applicators.
Participants were advised to repackage used applicators in the individual wrappers and then place in a storage bag containing other wrapped used applicators. In theory, loosely rewrapped used applicators could contaminate noninserted applicators during storage. However, such contamination, as well as other caused by rubbing against skin or exposure to saliva (of lower viscosity than vaginal mucus), cannot mimic the characteristic streaking pattern of mucus along the barrel of the applicator (Fig. 1). Moreover, Wallace et al found that even if unwrapped inserted and noninserted applicators were bagged together, the accuracy for predicting use and nonuse was 97% and 96%, respectively.7
A strength of this study is the inclusion of known negative and positive control applicators for each participant. Limitations are the lack of a second independent reviewer, and the fact that staining results were anticipated to be largely positive in a clinical trial setting. Other specifically designed studies compared the accuracy of staining methods with that of other standards like Gram staining8 or optical density measurement of lactobacilli growth,5 which detect vaginal cells and reach slightly higher accuracy levels of approximately 100%, but require more training and resources than dye staining.5–8 Although the reviewer knew the negative and positive controls, observation bias is not likely to have been significant, as the reviewer at each site was well trained, followed a standard procedure, and the difference between applicators stained positive and those stained negative was readily apparent (Fig. 1). Other studies validating this technique on different applicator types have found low rates of interobserver variability,5–7,9 further allaying concerns about the potential for significant observation bias.
Although dye-based assays may not offer a level of sophistication that currently developed “smart” applicators with embedded microchips or biomarkers for product tracing could provide,14,15 staining with trypan blue of HTI polypropylene and linear LDPE applicators, as well as the previously tested conventional LDPE Microlax applicators, provide an adequate, low-cost, reproducible, and easy-to-train method to verify self-reported vaginal insertion in clinical trial settings.
1. Skoler-Karpoff S, Ramjee G, Ahmed K, et al.. Efficacy of Carraguard for prevention of HIV infection in women in South Africa: A randomised, double-blind, placebo-controlled trial. Lancet 2008; 372:1977–1987.
2. Van Damme L, Govinden R, Mirembe FM, et al.. Lack of effectiveness of cellulose sulfate gel for the prevention of vaginal HIV transmission. N Engl J Med 2008; 359:463–472.
3. Padian NS, van der Straten A, Ramjee G, et al.. Diaphragm and lubricant gel for prevention of HIV acquisition in southern African women: A randomised controlled trial. Lancet 2007; 370:251–261.
4. Turner AN, De Kock AE, Meehan-Ritter A, et al.. Many vaginal microbicide trial participants acknowledged they had misreported sensitive sexual behavior in face-to-face interviews. J Clin Epidemiol 2009; 62:759–765.
5. Wallace AR, Thorn M, Maguire RA, et al.. Assay for establishing whether microbicide applicators have been exposed to the vagina. Sex Transm Dis 2004; 31:465–468.
6. Hogarty K, Kasowitz A, Herold BC, et al.. Assessment of adherence to product dosing in a pilot microbicide study. Sex Transm Dis 2007; 34:1000–1003.
7. Wallace AR, Teitelbaum A, Wan L, et al.. Determining the feasibility of utilizing the microbicide applicator compliance assay for use in clinical trials. Contraception 2007; 76:53–56.
8. Austin MN, Rabe LK, Hillier SL. Limitations of the dye-based method for determining vaginal applicator use in microbicide trials. Sex Transm Dis 2009; 36:1000–1003.
9. Katzen LL, Fernández-Romero JA, Sarna A, et al.. Validation of a dye stain assay for vaginally inserted hydroxyethylcellulose-filled microbicide applicators. Sex Transm Dis 2011; 38:1050–1055.
10. Furuyima A. Relationship between molecular characteristics and physical properties of linear low density polyethylenes. Pure Appl Chem 1985; 57:823–832.
11. Cohen CR, Brown J, Moscicki AB, et al.. A phase i randomized placebo controlled trial of the safety of 3% SPL7013 gel (VivaGel) in healthy young women administered twice daily for 14 days. PLoS One 2011; 6:e16258.
12. Hemmerling A, Harrison W, Schroeder A, et al.. Phase 1 dose-ranging safety trial of Lactobacillus crispatus
CTV-05 for the prevention of bacterial vaginosis. Sex Transm Dis 2009; 36:564–569.
13. Hemmerling A, Harrison W, Schroeder A, et al.. Phase 2a study assessing colonization efficiency, safety, and acceptability of Lactobacillus crispatus
CTV-05 in women with bacterial vaginosis. Sex Transm Dis 2010; 37:745–750.
14. Mauck CK. Biomarkers for semen exposure. Sex Transm Dis 2009; 36(suppl 3):S81–S83.
15. Mauck CK, van der Straten A. Using objective markers to assess participant behavior in HIV prevention trials of vaginal microbicides. J Acquir Immune Defic Syndr 2008; 49:64–69.