Share this article on:

A Randomized, Comparative Safety Study of a Prefilled Plastic and User-Filled Paper Applicator With Candidate Microbicide Tenofovir 1% Gel

Cohen, Jessica A. MHS, CIP*; Brache, Vivian BS; Foster, Jennifer PhD*; Cochon, Leila BS; Callahan, Marianne MEd; Schwartz, Jill MD

Sexually Transmitted Diseases: June 2013 - Volume 40 - Issue 6 - p 476–481
doi: 10.1097/OLQ.0b013e3182927ab1
Original Study

Background: A bridging study was performed to compare the safety, dose delivery, and acceptability of a prefilled plastic and user-filled paper applicator to assess whether a low-cost, user-filled, paper applicator could serve as a delivery option for tenofovir (TFV) 1% vaginal microbicide gel.

Methods: The study used a randomized crossover design with 25 healthy women randomized to begin with the prefilled or user-filled applicator. Within each study arm, participants delivered two 4.0-mL doses of TFV 1% gel vaginally for 7 days, with one dose delivered at the clinic each morning and a second dose delivered at home each evening. To assess the primary objective, applicator safety, colposcopy examinations were performed at 2 time points in each study arm.

Results: There were no colposcopic findings or adverse events attributable to either applicator. One case of vulvovaginal candidiasis was considered possibly related to gel use. On average, the user-filled applicator delivered 96% of the target dose, with 85% of doses falling within ±10% of the average dose volume. Participants found both applicators comparable for ease of use, insertion, and dispensing gel, with 60% of participants preferring the user-filled applicator.

Conclusions: This study suggests that both applicators are safe, and most women delivered TFV with the user-filled applicator as directed. Participants found both applicators acceptable, with a slight majority preferring the user-filled applicator. Incorporating a low-cost, user-filled, paper applicator to deliver TFV could help reduce costs and improve access to TFV 1% gel, especially in resource-limited settings heavily impacted by HIV.

A study comparing plastic prefilled and paper user-filled applicators for tenofovir gel found both applicators safe and acceptable. Most women delivered accurate volumes of tenofovir with the user-filled applicator.

From the *PATH, Seattle, WA; †Profamilia, Santo Domingo, Dominican Republic;and ‡CONRAD, Arlington, VA

The authors thank Breanne Grady, Project Coordinator at PATH, and Chalyce Grace, Clinical Trial Monitor at CONRAD, for their contributions to the study. The authors also thank the staff at Profamilia for helping implement the study as well as the study participants whose participation made this study possible.

Sources of support: This project was made possible by the generous support of the American people through the US Agency for International Development (USAID) under the terms of the Technologies for Health Cooperative Agreement No. GPH-A-00-01-00005. CONRAD provided the tenofovir gel for this study with support from USAID.

Conflict of interest: There are no conflicts of interest among any of the authors.

The contents are the responsibility of PATH and do not necessarily reflect the views of USAID or the US government.

Registration: Clinical Trials. Government Registry No.: NCT01283555.

Correspondence: Jessica Cohen, MHS, CIP, PATH, PO Box 900922, Seattle, WA 98109. E-mail:

Received for publication July 25, 2012, and accepted February 11, 2013.

Previous clinical studies of tenofovir (TFV) 1% gel, a topical antiretroviral microbicide manufactured by CONRAD (Arlington, VA), suggest that TFV is safe, acceptable, and effective in reducing the risk of HIV and herpes simplex virus type 2 infection.1–3 A phase 2b trial demonstrated that a 4.0-mL intravaginal dose of TFV gel administered before and after sex reduced HIV acquisition by an estimated 39% overall and by 54% in women with high adherence.1

These trials, like many others evaluating topical microbicide gels, have used a prefilled, plastic vaginal applicator to minimize variability and ensure accurate and precise dosing. However, the price of the prefilled plastic applicator has been identified as the highest cost component of the TFV product4 and could, therefore, pose a barrier for widespread access and use, especially in low-resource settings greatly impacted by HIV.

PATH identified a user-filled paper applicator as an alternative for microbicide delivery owing to its lower cost and reduced burden on waste disposal systems. The paper applicator is significantly less expensive than the prefilled applicator. Ease of disposal is also an important benefit because the paper applicator can be disposed of by burning, a common method in many resource-limited settings.

PATH and Profamilia, in collaboration with CONRAD, conducted this bridging study to determine whether the user-filled applicator had similar safety, dose delivery, and acceptability profiles to the prefilled applicator. The primary objective of this study was to compare the effect of the 2 applicators delivering TFV gel on symptoms and signs of irritation of the external genitalia, cervix, and vagina as seen on colposcopy after 7 days of twice-daily use. The secondary objectives included evaluating the dosing accuracy and precision as well as the acceptability of the 2 applicators. The results from this study could support the approval of TFV gel with this lower-cost applicator, thereby reducing costs and improving access to the first anti-HIV microbicide in settings where women urgently need new HIV prevention options.

Back to Top | Article Outline


Study Methods and Population

The study was conducted between January 2011 and June 2011 at the Profamilia Clinic in Santo Domingo, Dominican Republic. Profamilia is a nonprofit, nongovernmental organization that provides reproductive and primary health care. In a randomized, crossover study design, 25 participants were consented, enrolled, and randomized to an initial applicator assignment (either the prefilled or the user-filled applicator). Within each study arm, participants were asked to deliver 2 doses of TFV gel each day for 7 days—with one dose delivered under observation at the clinic each morning and a second dose delivered by the participant at home each evening. A 21-day minimum washout period (including menses) separated the study arms. Figure 1 describes the sequence of the study visits.

Participants were recruited from among the clientele of the Profamilia clinic. All participants were women aged 18 to 50 years. The main inclusion criteria were premenopausal, in good health as determined by self-report and baseline clinical testing, having regular menstrual cycles, not breast-feeding, at low risk for sexually transmitted infections, not pregnant or at risk of pregnancy, and willing to abstain from sexual intercourse and use of vaginal products during the two 7-day periods in which the applicators were used.

The sample size of 25 participants was selected based on previous safety studies and allowed for identification and comparison of colposcopic findings as the primary outcome.

The protocol was reviewed and approved by the institutional review board of Profamilia, the National Bioethics Health Council of the Dominican Republic, and the Research Ethics Committee of PATH.

Back to Top | Article Outline

Study Products

The products under investigation were a user-filled, paper applicator manufactured by Tekpak, Inc (Marion, AL) and a prefilled, plastic applicator manufactured by HTI Plastics (Lincoln, NE). Tenofovir gel was delivered by both applicators. It is a clear, transparent, viscous gel. The strength (1%) and dosage (4.0 mL) were equivalent to those used in previous clinical trials.

The user-filled applicator consists of a paper barrel and plunger. A mechanical stop and a fill line are included to help prevent users from filling the applicator beyond the intended dose of 4.0 mL (Fig. 2). Participants were also provided a multidose tube containing 100 g of TFV gel for filling the applicator. These tubes are aluminum, lined with epoxy, and come with a white polyethylene screw cap equipped with a puncture tip. Aluminum tubes were selected to minimize the “suck back” of air being pulled into the tubes when squeezing pressure is released.

The prefilled, polypropylene applicators were prefilled to deliver 4.0 mL of gel (Fig. 3). The design consists of a barrel, plunger, piston, and snap cap.

Back to Top | Article Outline

Product Administration

After the baseline colposcopy for study arm 1, participants were randomly assigned to the order of applicator use using envelopes that matched the assigned participant number. The randomization allocation sequence was generated by PATH; investigators at Profamilia enrolled participants and assigned participant numbers. Based on this assignment, Profamilia investigators opened the matching randomization envelope and then distributed the appropriate applicator kit to the participant along with instructions for use. The first dose of gel was then delivered by the participant in the presence of a female study investigator. The participant returned to the clinic the following 6 mornings to deliver their morning dose. The daily clinic visits provided study staff with the opportunity to directly observe and record applicator use. The second daily applicator use was conducted at home and recorded in a log. The log was checked the following morning by study staff.

Back to Top | Article Outline

Outcome Measures


To assess applicator safety, colposcopic examinations at the baseline (preapplicator use) and follow-up (after 7 days of twice-daily use) visits for each arm were conducted, for a total of 50 colposcopy examinations for each applicator type. Immediately before the second colposcopy examination in each study arm, participants inserted and removed an empty applicator to identify any immediate irritation that could result from applicator use. Findings were classified by number, type, location, and product relatedness (unrelated, possibly, probably, or definitely related to applicator and/or gel use). Baseline colposcopic findings and iatrogenic findings were classified as non–applicator related. Clinical examination findings, self-reported symptoms, and adverse events (AEs) were also included in the safety analysis. In addition to descriptive analyses of safety outcomes, Fisher exact test was used to compare the difference of outcomes between the 2 applicator types. All colposcopy examinations were performed by the same investigator and were performed according to World Health Organization/CONRAD guidelines.5

Back to Top | Article Outline


Dose delivery data were collected 3 times in each applicator arm, for a total of 75 data points collected for each applicator. At each of these visits, the participant was asked to either fill the user-filled applicator with 4.0 mL of TVF gel or was given a prefilled applicator containing 4.0 mL of TFV gel. For the user-filled applicator, the participant attached the applicator barrel to a tube of TFV gel and then squeezed the gel into the barrel until the plunger automatically stopped and the fill line on the plunger appeared. The investigator then weighed the filled applicator and returned it to the participant to dispense the gel vaginally. After gel expression, the investigator recorded the weight of the emptied applicator. For the prefilled applicator, the participant loaded the plunger into the barrel; the investigator then weighed the applicator and then weighed it again after the gel was expressed vaginally. The calculations in Table 1 were used to measure dose delivery. To calculate the average weight of each empty applicator type, 20 empty user-filled and prefilled (without caps) applicators were weighed. The respective average weight (±SD) was 2.37 (±0.02) and 7.92 (±0.04) g.

To assess applicator acceptability, semistructured interviews were conducted with the participants at the final study visit, for a total of 25 data points. Data about ease of use, comfort, applicator preference, and the overall acceptability of each applicator were collected. Categorical data were summarized by frequencies and percentages, and qualitative data were analyzed to identify categories and themes.

Back to Top | Article Outline


A total of 37 women provided written informed consent and were screened for eligibility. Of these, 11 participants were screened out and 1 participant qualified but was not enrolled because the target enrollment had been reached. A total of 25 participants were enrolled, randomized to the order of product use, completed both study arms, and were included in all analyses (Fig. 4).

Demographics for the 25 participants are shown in Table 2. Thirty-six percent (36%) were between 21 and 30 years of age, and the remaining 64% were between 31 and 45 years of age. Eighty-four percent (84%) had the equivalent of a high school education or less. Seventy-six percent (76%) of participants noted prior experience with applicators (e.g., for vaginal treatments or medications), whereas 80% reported never having used a tampon.

Back to Top | Article Outline

Safety Results

Table 3 summarizes the colposcopic findings and compares the frequency between baseline and follow-up visits. At the baseline colposcopic examinations, a total of 8 findings were reported: 3 petechiae, 2 minor external genitalia lacerations, 1 peeling of the epithelium of the cervical face, and 2 iatrogenic findings, caused by insertion of the speculum before the examination. All findings, except one, were resolved by the following colposcopic examination after 1 week of product use. The only finding that had not resolved was a minor laceration of the genitalia, which was first observed at baseline of arm 1 and did not resolve until the end of arm 2.

After 1 week of product use, 4 colposcopic findings were reported, all at the end of study arm 2. Three findings were peeling of the cervix or posterior vaginal fornix; the remaining finding was an iatrogenic petechiae caused by the speculum during the final visit. None of the findings were considered related to the applicators or drug. The differences of frequency of findings between baseline visit and after 1 week of use were not significant (Table 3).

After 1 week of product use, 10 AEs among 7 women were recorded. All were reported as mild events, except 1 episode of nausea, which was reported as moderate. Of these, only 1 was considered possibly related to gel use—vulvovaginal candidiasis, first reported on visit 9 (after 7 days of gel delivery with the user-filled applicator). No other genitourinary symptoms or clinical examination findings were reported. No AEs were considered possibly related to applicator use; there were no reports of serious AEs.

Back to Top | Article Outline

Dose Delivery Results

Filling Accuracy and Precision

With regard to filling accuracy, on average, the user-filled applicator was filled with 5.5% more than the target dose of 4.0 mL; the prefilled was loaded with approximately 3% more than the target dose (Table 4). Regarding filling precision, 92% (69/75) of the user-filled applicators were filled within ±5% of the average filled volume, and 100% of the prefilled applicators were filled within ±5% of the average filled volume.

Back to Top | Article Outline

Dosing Accuracy and Precision

Figure 5 illustrates the range of dose delivery volumes observed with the user-filled applicator. On average, participants delivered a 3.83-mL dose of TFV gel, representing 96% of the target dose (4.0 mL) (Table 5). Regarding dosing precision (range around the average expressed volume), 64% (48/75) of applicators delivered volumes within ±5% of the average dose, whereas 85% (64/75) of applicators delivered within ±10% of the average dose.

Of the 7 doses falling below the 10% range for dosing precision, in 1 case, an air bubble was observed in the applicator. In the remaining 6 instances, underdosing was related to not pushing the plunger in completely or gel adhering to the applicator upon removal. Four doses were greater than 10% of the average dose.

Owing to an error in an illustration for the prefilled applicator instructions, participants were instructed to push the plunger too deep into the applicator barrel before vaginal application. This resulted in gel being pushed behind the applicator piston and remaining in the barrel after gel delivery. As a result of this error, the dose delivery data for the prefilled applicator are not valid, and a direct comparison between the dose delivery data from the prefilled applicator and the user-filled applicator cannot be made. Of note, the illustrations for the prefilled applicator were developed for this study. The problem identified here does not apply to other trials using the prefilled applicator.

Back to Top | Article Outline

Applicator Acceptability Results

The results from this study indicate that all women found the user-filled applicator easy to fill and were confident or very confident that they had inserted the correct amount of gel into the applicator (Table 6).

The user-filled and prefilled applicators rated similarly in relation to ease of use, insertion, dispensing, and comfort (Table 7). However, for disposal, 76% of participants preferred the user-filled paper applicator, whereas for storage, 64% of participants preferred the prefilled applicator. When asked to select a preferred applicator overall, 60% preferred the user-filled applicator.

Reasons cited for preferring the user-filled, paper applicator included the thinner diameter and smooth material, leading to easier and more comfortable insertion. Ease of disposal was also noted, with several participants citing burning and disintegration as beneficial for the environment. Among participants who preferred the prefilled, plastic applicator, reasons cited included the comfort of insertion and the fact that it came prefilled.

When asked about future use of the user-filled applicator, all but 1 participant said that they would use it in the future if it came with a gel to prevent HIV, and 100% said that they would recommend this applicator to other women if it came with a gel that helped prevent HIV.

Back to Top | Article Outline


The purpose of this study was to assess and compare the safety, dosing, and acceptability of a user-filled, paper applicator in relation to a prefilled, plastic applicator upon delivery of TFV gel.

These data suggest that the user-filled applicator has a comparable safety profile to the prefilled applicator, with no colposcopic findings or AEs attributable to either applicator. Peeling of the epithelium was the only reported finding after 1 week of product use. However, peeling was also observed at baseline, and the skin below the peeling was completely normal, without any redness that could indicate an inflammatory process.

Peeling has been reported in previous colposcopy studies,6,7 both at baseline and after product use. One of these studies hypothesized that peeling may represent normal cervicovaginal epithelial turnover and that because of moisture in the vagina, dead cells slough as hydrated sheets instead of small, dry flakes desquamated from areas of exposed, drier skin. Although it is biologically plausible that constant use of a gel may further facilitate this desquamation process, our limited data do not support this hypothesis but rather that the peeling observed is a normal variant.

These applicator safety data are similar to previous studies that have evaluated the safety of microbicide applicators. In one study assessing genital findings in a phase 2 microbicide trial of Carraguard, manufactured by FMC Biopolymer (Philadelphia, PA), the authors found no evidence of applicator trauma (using a Microlax-type applicator manufactured by Norden Pac International [Kalmar, Sweden]).8 A second study, evaluating the safety of 3 vaginal applicators (including the prefilled applicator evaluated in this study), identified 4 minor colposcopic findings (petechiae) that were considered possibly related to applicator use, with no difference among applicators.9

Another key consideration for use of the user-filled applicator is whether women can fill and dispense gel from this applicator as directed. Results from this study suggest that women can fill the applicator with similar accuracy, although slightly less precision than the prefilled applicator (which is filled by machine). The dosing accuracy of 96% illustrates that, on average, women delivered a volume close to the target dose (3.83 mL vs. 4.0 mL). Although there were 7 doses under the 10% precision range that lowered the overall dose delivery mean, most doses (85%) fell within a 10% range around the average expressed dose volume.

In this study, there was only one event noted where the user-filled applicator was filled beyond the mechanical stop, indicating that although the mechanical stop is not fail proof, it functioned as intended in 99% of uses. The mechanical stop can be adjusted to accommodate different dose volumes. Therefore, if a higher average dose volume is needed to better approximate the target dose volume for TFV, the stop could be moved to allow higher fill volumes.

The dosing data observed with the user-filled applicator should be considered in association with the aluminum tubes that delivered the TFV into the applicator. Aluminum tubes minimize air being sucked back into the tube after pressure on the tube is released, reducing the chance of air bubbles being transferred into applicators. With this combination of tube and applicator, only one air bubble was noted. Tubes made of other materials could result in different dose delivery profiles with the same applicator.

Although women in this study reported both applicators easy to use and all women considered the user-filled applicator easy to fill, the generalizability of these results is limited. Specifically, this study had a small sample size, and the high rate of prior applicator use could also have affected the secondary outcomes of this study. Further research on dose delivery and acceptability of the user-filled applicator with TFV should be conducted with larger study populations in sub-Saharan Africa and elsewhere where TFV will be introduced.

Back to Top | Article Outline


This study suggests that both the user-filled applicator and the prefilled applicator are safe, with neither applicator causing any irritation to the vagina, cervix, or external genitalia. The dose delivery results for the user-filled applicator indicate that women are able to use this applicator as directed, delivering 96% of the target dose on average; most doses (85%) fell within a 10% range around the average dose volume. Women found both applicators acceptable, with a slight majority (60%) preferring the user-filled applicator. Given the importance of ensuring widespread access to microbicides in resource-limited settings, the user-filled, paper applicator could provide an important alternative for delivery of TFV 1% gel, or other microbicide gel formulations by reducing delivery costs and increasing access to important new HIV prevention methods.

Back to Top | Article Outline


1. Karim QA, Karim SSA, Frohlich JA, et al. Effectiveness and safety of tenofovir gel, an antiretroviral microbicide, for the prevention of HIV infection in women. Science 2010; 329: 1168–1174.
2. Hillier SL. Safety and acceptability of daily and coitally dependent use of 1% tenofovir over six months of use. Presented at: Microbicides 2008 Conference; 2008; New Delhi.
3. Mayer KH, Maslankowski LA, Gai F, et al. Safety and tolerability of tenofovir vaginal gel in abstinent and sexually active HIV-infected and uninfected women. AIDS 2006; 20: 543–551.
4. World Health Organization (WHO)/UNAIDS. Next steps with 1% tenofovir gel: Meeting report [WHO Web site]. Available at: Accessed July 9, 2012.
5. WHO, CONRAD. Manual for the standardization of colposcopy for the evaluation of vaginal products, update 2004. Geneva: WHO; 2004. Available at: Accessed July 9, 2012.
6. Schwartz JL, Kovalevsky G, Lai J, et al. A randomized six-day safety study of an antiretroviral microbicide candidate UC781, a non-nucleoside reverse transcriptase inhibitor. Sex Transm Dis 2008; 35: 414–419.
7. Ballagh SA, Mauck CK, Henry D, et al. A comparison of techniques to assess cervicovaginal irritation and evaluation of the variability between two observers. Contraception 2004; 70: 241–249.
8. van de Wijgert JH, Kilmarx PH, Jones HE, et al. Differentiating normal from abnormal rates of genital epithelial findings in vaginal microbicide trials. Contraception 2008; 77: 122–129.
9. Brache V, Cohen JA, Cochon L, et al. Evaluating the clinical safety of three vaginal applicators: A pilot study conducted in the Dominican Republic. Contraception 2006; 73: 72–77.
© Copyright 2013 American Sexually Transmitted Diseases Association