MTN-033: a Phase 1 Study Comparing Applicator versus “as Lubricant” Delivery of Rectal Dapivirine Gel

ABSTRACT

Data to inform behaviorally congruent delivery of rectal microbicides as lubricants are scant. Dapivirine (DPV) is a nonnucleoside reverse transcriptase inhibitor which has been demonstrated to be well-tolerated and efficacious in multiple clinical trials when used in a vaginal ring formulation. DPV gel administered rectally with an applicator was found to be well-tolerated in a phase 1 clinical trial. MTN-033, a single site, open label, sequence randomized, crossover study, enrolled HIV-negative men to receive 0.05% DPV gel intrarectally using an applicator (2.5 g) and self-administered on an artificial phallus as lubricant (up to 10 g). The study evaluated the pharmacokinetics (in plasma, rectal fluid, and mucosal rectal tissue), safety, acceptability, and pharmacodynamics of DPV gel when applied rectally. Statistical comparisons between methods of application were performed using mixed effects models or Wilcoxon’s signed rank tests. Sixteen participants used DPV gel by applicator and 15/16 participants used gel as lubricant (mean, 1.8 g; SD, 0.8). DPV plasma AUC_0–24h after use as lubricant was estimated to be 0.41 times the AUC_0–24h (95% CI 0.24, 0.88) after use with applicator. While DPV was quantifiable in plasma and luminal fluid, it was not quantifiable in tissue for both applicator and as lubricant administration. No related adverse events (AE) were reported, and 15/15 participants felt the gel was easy to use. Evidence of local delivery and systemic absorption of DPV when dosed as an anal lubricant supports the feasibility and potential for development of lubricant-delivered rectal microbicides. There were no safety concerns associated with use of DPV gel and participants reported finding it easy to use. However, lower DPV exposure in plasma and lack of quantifiable DPV in rectal tissue indicate that higher potency, concentration, and longer half-life antiretrovirals with optimized formulations will be needed to achieve protective tissue concentrations.

INTRODUCTION

Oral, vaginal, and injectable formulations of antiretroviral drugs provide safe and efficacious options for HIV preexposure prophylaxis (PrEP) for diverse populations of persons at risk of HIV acquisition (1–6). All these strategies, however, fail to meet the needs of persons whose primary risk of HIV acquisition is unprotected receptive anal intercourse (RAI) and those who desire HIV prevention only at times (“on-demand”) and in anatomic locations (nonsystemic rectal delivery) of risk. On-demand dosing proved efficacious using oral PrEP formulations (3). Rectal microbicides offer local, nonsystemic PrEP, especially for MSM, a population disproportionately impacted by HIV and whose risk is largely attributable to condomless RAI (7). The monthly dapivirine ring demonstrated 30% efficacy overall in phase III studies for prevention of vaginal HIV acquisition, though risk reduction was likely greater among participants who use the ring more consistently. Phase 3 post hoc subgroup analyses suggested a trend toward greater risk reduction among highly adherent women, 75% and higher (1, 8). Open-label extension studies of the product supported that trend, showing increased product use and, accordingly, suggesting higher efficacy of about 50% overall, according to modeling data. Given that MSM who engage in RAI very frequently use sexual lubricants in conjunction with anal sex, rectal microbicide delivery as a sexual lubricant would be ideal – an adherence-enhancing, behaviorally congruent strategy (9). Despite this unmet PrEP need, no rectal microbicide has advanced to efficacy trials.

Phase 1 and 2 rectal gel microbicide studies have delivered gel rectally using an applicator designed for vaginal use, which generally has not been well accepted and does not mimic lubricant use (10, 11). For example, MTN-017 compared 1% tenofovir (TFV) gel (daily and on-demand pericoital use) with daily oral PrEP and demonstrated that product adherence achieved with TFV gel use (specifically pericoital use) was comparable with daily oral PrEP (9). Prevention methods that can be delivered in a behaviorally congruent manner – medicated lubricants or douches already commonly used with anal sex – may have greater acceptability and, by extension, greater effectiveness.

There are limited data on the use of rectal microbicides as sexual lubricants. Drug distribution may be impacted based on delivery modality. Intrarectal application of a radiolabeled gel using manual application as lubricant for gel delivery demonstrated significantly lower colonic retention compared to applicator-based administrations (3.4% versus 94.9% of a 10-mL dose retained, respectively) (12). These data highlight the importance of product formulation and delivery mechanism to provide required antiretroviral drug retention and coverage in the colon.

Dapirivine (DPV), a nonnucleoside reverse transcriptase inhibitor, proved effective when formulated as a vaginal microbicide ring to reduce the risk of HIV infection in two phase 3 trials of African women (1) and is recommended by the World Health Organization for women who are at substantial risk of HIV. MTN-026 evaluated a 0.05% DPV gel administered rectally using the vaginal applicator and showed delivery of DPV sufficient to at least transiently reduce the infectivity of rectal biopsy specimens challenged with HIV ex vivo; however, colorectal tissue DPV concentrations were below those seen in cervicovaginal tissue after vaginal dosing of the same gel (13). In MTN-033, the present study, we compare applicator to “as lubricant” delivery of the same 0.05% DPV gel to rectal tissue to explore the feasibility of dosing DPV gel as a behaviorally congruent sexual lubricant.

RESULTS

Sixteen healthy, participants living without HIV enrolled in the study, all of whom identified as male. Mean age was 29.7 years (Table 1). Fifteen participants (94%) completed both doses per protocol. One participant discontinued early due to an anorectal sexually transmitted infection after their applicator dose (but prior to as lubricant) leaving a total of 16 applicator doses and 15 as lubricant doses for analysis (Fig. 1). Participants completed the coital simulation device (CSD) gel application within 5 min. The estimated amount of gel inserted when using the CSD was between 0 and 3.0 mL, with mean (SD) of 1.8 mL (0.8 mL) or 0.9 mg DPV (0.4 mg). Gel volume administered with the applicator was assumed to be 2.5 mL or 1.25 mg DPV.

TABLE 1.

Demographic characteristics of participants enrolled in MTN-033

Demographic	Sequence Aa (n = 8)	Sequence Bb (n = 8)	Overall (n = 16)
Age (yrs)
Mean (SD)	29.0 (9.1)	30.4 (10.1)	29.7 (9.3)
Range	19−45	19−47	19−47
Participant-identified gender, n (%)
Male	8 (100%)	8 (100%)	16 (100%)
Race, n (%)
Asian	2 (25%)	0 (0%)	2 (13%)
Black or African American	0 (0%)	2 (25%)	2 (13%)
White	5 (63%)	5 (63%)	10 (63%)
Latino or Hispanic	1 (13%)	1 (13%)	2 (13%)
Other	1 (13%)	1 (13%)	2 (13%)

^aDPV gel by applicator followed by DPV gel as lubricant.

^bDPV gel as lubricant followed by DPV gel by applicator.

FIG 1.

CONSORT diagram for the disposition of participants in the MTN-033 phase 1 trial.

Pharmacokinetics.

Plasma DPV concentration rose to a peak by 2 and 2.5 h after administration for the applicator and as lubricant use, respectively, and fell below the lower limits of quantification (LLOQ) by 24 h after administration for both methods (Fig. 2). Peak plasma DPV concentration was higher for use with an applicator (median, 319 pg/mL; interquartile range [IQR], 247 to 603), than use as lubricant, (median 85 pg/mL; IQR, 54 to 218; P < 0.01). The area under the concentration-time curve to 24 h after dosing (AUC_0–24h) for plasma after use as lubricant was estimated to be 0.41 times the AUC_0–24h after use with an applicator (95% CI 0.24–0.88, P-value < 0.001; Table 2) without adjustment for dose administered. Adjusting for the dose administered for each dosing method, the bioavailability of rectal DPV as lubricant relative to applicator was a median of 0.51 (IQR, 0.29 to 1.14; Table 2).

FIG 2.

Median and interquartile range of (A) dapivirine (DPV) concentration in blood plasma for participants at all time points and (B) DPV concentration in rectal fluid swabs for 8 participants at 1 h, 8 participants at 4 h, and all participants at 24 h postdosing, after rectal dosing of the 2.5% DPV gel with an applicator (2.5 g of the gel) and using a coital simulation device (CSD, up to 10 g).

TABLE 2.

Median and interquartile range (IQR) peak concentration (C_max), time to peak concentration (T_max) and area under the concentration-time curve from 0 to 24 h (AUC_0–24) of dapivirine in plasma

Parameter	Applicator (n = 16)	CSD (n = 15)e
Dapivirine dose, mga	~1.25	0.90 (0.55, 1.25)
Cmax, pg/mL	319 (247, 603)	85 (54, 218)
Tmax, hours	2.0 (1.5, 2.8)	2.5 (2.0, 3.0)f
AUC0–24, pg·h/mL	3,545 (3,098, 6,158)	1,337 (720, 2,401)
AUC Ratiob		0.41 (0.24, 0.88)
Dose-normalized AUC0–24c, pg·h/mLb	3,545 (3,098, 6,158)	1,991 (1,648, 3,614)g
Relative bioavailabilityd		0.51 (0.29, 1.14)g

^aBased on the standard amount of 0.05% DPV gel provided with the applicator (2.5 mg) or the weighted amount administered with the CSD.

^bAUC_0-24 (CSD)/AUC_0–24 (applicator) for each participant.

^cAUC_0-24/(2.5 mg DPV dose/dose delivered) for each participant.

^d(AUC_0–24 [CSD]/weight [CSD])/(AUC_0–24 [applicator]/weight [applicator]) for each participant, where weight (CSD) is the estimated amount of gel administered with the CSD and weight (applicator) is the standard targeted amount of gel administered with the applicator (2.5 g).

^eOnly 15 out of the 16 participants received DPV gel using the CSD, as one participant discontinued the study early and completed only the first product use period.

^fExcludes one additional participant who had dapivirine below the LLOQ at all time points after using the CSD.

^gExcludes one additional participant for whom an invalid amount of DPV gel used with the CSD was reported.

One hour after gel administration, rectal fluid DPV concentrations were numerically higher, but not statistically different (two-sided P value = 0.688; Fig. 2). When comparing dosing as lubricant (median, 16.1; IQR, 8.8 to 53.6 ng/mg) to dosing with the applicator (median, 9.2; IQR, 3.6 to 18.2 ng/mg). Four-hour rectal fluid DPV concentrations were significantly lower when dosed as lubricant (median, 3.13; IQR 0.8, 6.1 ng/mg) than when dosed with the applicator (median, 13.7; IQR, 1.8, 20.9; P value = 0.008). Rectal fluid DPV was quantifiable in 9/16 (56%) and 8/15 (53%) samples taken 24 h after gel dosing with applicator and as lubricant, respectively; no statistical test was performed.

DPV was quantifiable in only 1/62 postdose rectal biopsy specimens (DPV concentration, 0.9 ng/mg); therefore, no statistical comparisons were performed.

Pharmacodynamics.

Statistically significant reduction in HIV infectivity of colorectal tissue explants was only observed 4 h after dosing and only with applicator dosing, with a mean change in log₁₀ cumulative p24 (IQR) of −0.97 pg/mL/mg (−1.57 to −0.37; Fig. 3). This decrease was primarily driven by a few observations (six biopsy specimens collected from two participants 4 h after applicator dosing) that showed HIV replication suppression to below the limit of quantification (BLQ) p24 values. These biopsy specimens, along with a few others that also showed BLQ p24 values (four biopsy specimens collected 1 h after applicator dosing and one biopsy specimen collected 4 h after application as lubricant), came from some of the participants with the highest plasma DPV concentrations at those time points (Fig. 4), though DPV and p24 concentration were weakly and negatively correlated (Spearman correlation coefficient, −0.24; P = 0.021).

FIG 3.

Median and interquartile rage (boxplots) of the log₁₀ cumulative, weight-adjusted p24 levels from up to three rectal tissue explant supernatants per participant time point, collected at baseline (all participants), 1 h (8 participants), and 4 h (8 participants) after rectal application of the DPV gel. The gray horizontal bars represent the interquartile range (dark gray) and range (light gray) of the weight-adjusted LLOQ of cumulative p24 concentrations.

FIG 4.

Log₁₀ cumulative, weight-adjusted p24 levels from three rectal tissue explant supernatants per participant time point against the log₁₀ DPV plasma concentration from serum collected within 30 min of the rectal biopsy specimen. Markers are unique to each participant. The gray horizontal bars represent the interquartile range (dark gray) and range (light gray) of the weight-adjusted LLOQ of cumulative p24 concentrations, while the vertical dashed line represents the LLOQ for DPV concentration in blood plasma (20 pg/mL). Postdose concentrations below the LLOQ were assigned a value equal to LLOQ/2 and predose baseline concentrations are shown separately (left), for reference.

Secondary endpoint: safety.

There were a total of seven adverse events (AEs) reported by three participants, five after applicator use (one participant) and two after as lubricant use (two participants). Grade 2 or higher AEs were reported by two participants, one occurring with each dosing method. None of the reported AEs were considered related to study product.

Histology evaluation.

Most participants had scores of 0 or 1 (0 = no abnormality, 1 = mononuclear cell infiltrate) with only one score of 2 (neutrophilic infiltrate-lamina propria) with the applicator. With the CSD, there was one score of 3 (neutrophilic infiltrate-epithelium) and one score of 4 (crypt destruction). The change in rectal tissue histology scores relative to baseline showed no evidence of a statistically significant difference between the methods of application (P = 0.145).

Acceptability.

From the 15 participants who completed the study, most (n = 14; 93.3%) indicated that they would be “likely” or “very likely” to use a gel that provided some protection against HIV within 12 h before and up to 12 h after RAI. After dosing, all reported usage of the gel to be “easy” or “very easy.” Most participants (n = 12; 80%) also expressed a willingness to use twice as much volume during sex if it was required to achieve protection against HIV. Participants noted that both methods of administration would be acceptable if an HIV prevention gel was found to be efficacious; detailed acceptability data have been published elsewhere (14).

DISCUSSION

Prior phase 1 and 2 studies of rectal microbicides (TFV gel and DPV gel) used the standard vaginal HTI applicator to administer gel and therefore were unable to provide data informing the use of a rectal microbicide as lubricant. MTN-033 is the first study to assess comparative pharmacokinetics (PK) and pharmacodynamics (PD) of a candidate rectal microbicide when administered as a lubricant versus using a standard applicator. It provides crucial data to inform the safety and potential efficacy of rectal microbicides when used as sexual lubricants. An ideal rectal microbicide for HIV prevention could be used in a behaviorally congruent manner such as in the form of a sexual lubricant. In this study, gel was overall well tolerated and generally well liked. Further acceptability data have been published (14).

Plasma DPV concentrations achieved with use of rectal DPV gel as lubricant were approximately 41% of those achieved with applicator use. Similar relationships were seen with 24 h AUC between use of DPV gel as lubricant and with applicator. Adjusting for the volume of gel retained resulted in a more favorable bioavailability of 51% for as lubricant use compared to applicator. Plasma DPV PK were consistent with findings from the MTN-026 (phase 1 rectal DPV PK study) which demonstrated detectable plasma DPV concentrations, but overall low DPV levels in the rectal tissue.

Given that DPV was quantifiable in only a single rectal tissue biopsy specimen (obtained at 1 h from one participant in the applicator arm) no further comparisons were made. It is unknown why tissue concentrations were low despite evidence of drug penetration, as evidenced by quantifiable drug concentrations in plasma. These results were also consistent with DPV rectal tissue concentrations seen in MTN-026 which showed most rectal tissue biopsy specimens after a single dose were below the LLOQ. In MTN-026, DPV levels in rectal biopsy specimens were also lower than levels in cervicovaginal tissue concentrations achieved with DPV vaginal ring used in phase 3 trials that demonstrated protective efficacy against HIV (15). Reasons could include rapid absorption into systemic circulation (where DPV has a longer half-life than in tissue), limited DPV retention in colorectal tissue (which is less lipophilic than vaginal tissue), anatomic differences in transporters or efflux pumps, potentially different impacts of simulated sex on rectal versus vaginal mucosa, and differential assay sensitivity across these matrices. Further investigation into these factors may help explain discrepancies between detectable DPV in plasma and lack of DPV in rectal tissue.

While tissue concentrations observed in rectal biopsy samples were predominantly nonquantifiable, ex vivo efficacy studies of tissue biopsy samples suggested a trend toward increasing systemic DPV concentrations and decreasing levels of p24. p24 was also suppressed in a few biopsy samples with nonquantifiable DPV levels, particularly in samples with the highest DPV plasma concentrations, suggesting that DPV quantitation in tissue may be limited by assay LLOQ, thus, limiting estimation/testing of a concentration-response association. The magnitude of p24 antigen reduction, nearly 1 log₁₀ pg/mL/mg, is similar to previous studies with orally and rectally administered drugs, though of shorter duration (16, 17). This indicates the need for different formulations of the DPV gel to achieve more sustained DPV tissue concentrations as seen with TFV and FTC dosing and higher loading of the gel to overcome the inefficient delivery of DPV dosed as lubricant.

An efficacious rectal microbicide could have potential as a sustainable prevention strategy for populations that may not be able or willing to use oral PrEP or are unable to access oral PrEP. While long-acting injectable and oral PrEP are highly efficacious HIV biomedical prevention strategies, operational challenges, and barriers to access may make these approaches less attractive and less practical for populations at the highest risk for HIV, such as youth and minority populations that struggle to access health care. A recent report of HIV prevention product preferences listing injectable, on-demand oral, and on-demand rectal products as options, all three product types were favored by most of the PrEP naive research participants (18).

This study had several strengths: rigorous PK plasma sampling; directly observed dosing (with applicator) and guided on-site dosing (with the CSD) optimized adherence; and the crossover design allowing each participant to serve as his own control. This is the first study to assess differential PK of a rectally dosed antiretroviral drug between applicator-based dosing and as lubricant dosing using simulated coitus. Participants were also able to control the amount of gel used with as lubricant dosing mimicking close to real life conditions in this research setting. All materials which contacted DPV gel were weighed before and after use which allowed for calculation of retained gel following use of gel as lubricant. While efforts were made to calculate retained gel as accurately as possible, there was one participant for whom the retained gel with CSD use was determined to be 0 g. It is conceivable that residual stool or mucus on the CSD could have confounded these estimates. MTN-033 allowed enrollment of both cisgender males and transgender women, but all 16 of the individuals who ultimately enrolled were cisgender males. Further research is warranted on safety in transgender populations and the impact of gender-affirming hormone therapy on the PK/PD of microbicide candidates. While there are no known histologic sex specific differences in the anorectal mucosa, prior studies indicate that there may be differences in vascularity and perfusion of the anorectum between individuals assigned male at birth versus individuals assigned female at birth (19) that could potentially impact local drug concentrations and PK/PD. A recent study indicated variable explant infectivity in cisgender women compared to MSM (17). It is also conceivable that the manner/position in which the CSD was used could have impacted DPV PK, but these data were not collected as part of the study and may warrant further investigation in future rectal microbicide studies. The difference in the PK observed between the applicator and CSD may have been influenced to some degree by differences in dose of gel delivered (2.5 g with applicator versus a mean of 1.8 g with CSD). These differences could be mediated by factors, including anatomic anorectal differences, baseline anal sphincter tension, and position while using the CSD. Future research to optimize gel dose would also be warranted. Prior studies conducted by Shieh et al. (20) used SPECT/PET imaging to estimate retention of 10 mL radiolabeled DTPA in HEC gel delivered using manual dosing with a CSD versus an applicator (both CSD and applicator were identical to what was used in MTN-033). While their findings suggested that using a CSD resulted in an adequate mucosal luminal distribution of product in the colorectum, there was limited colorectal retention of drug product (3.4%). Differences in study gel viscosity, drug lipophilicity, composition, and other product characteristics could also potentially explain why a greater proportion of gel was retained in the MTN-033 study.

The study findings suggest that DPV gel, when dosed as an anal lubricant, traverses the rectal mucosa resulting in plasma detection. Larger amounts of gel associated with applicator dosing result in transient HIV p24 suppression in rectal biopsy explant studies. While this study supports the feasibility of rectal microbicides delivered as lubricant, the observed 3-fold lower DPV exposure in plasma, lack of detectable DPV in tissue biopsy specimens, and failure to suppress HIV p24 in explant assays indicate formulation changes are necessary to achieve protective tissue concentrations, as informed by vaginal microbicide experience. Our findings support continued development of a lubricant rectal microbicide strategy.

MATERIALS AND METHODS

MTN-033 was a single-site, open-label, dosing method sequence randomized crossover study of 16 participants living without HIV. Each participant was to receive 0.05% dapivirine (DPV) gel by intrarectal dosing using both the vaginal HTI applicator (2.5 g) containing 1.25 mg DPV (administered by study nurse) and by self-administration on an artificial phallus, referred to as the coital simulation device (CSD), to mimic dosing as a lubricant. For as lubricant dosing, up to 10 g (4 × 2.5-g applicators) could be used, as desired. The DPV gel from four applicators (2.5 g each) was expelled into a weighing cup for an initial weight measurement. Participants were instructed to use the desired amount of gel on the CSD, which was used to simulate receptive anal sex, inserting the CSD to its fullest extent in and out of the rectum each second for 5 min. After the CSD was used, the unused gel remaining in the weighing cup was measured to calculate the quantity of gel that the participant chose to use. Participants were randomized in a 1:1 ratio to receive DPV gel by applicator first followed by a washout period and then DPV gel as lubricant (sequence A) or to receive DPV by lubricant first followed by a washout period and then DPV gel by applicator (sequence B). Plasma pharmacokinetics (PK) sampling of all research participants occurred at 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, and 24 h postgel dosing. We randomized participants 1:1 to rectal fluid and biopsy sampling at either 1 and 24 h or 4 and 24 h.

The study was approved by the University of Pittsburgh institutional review board (PRO16050207).

Study objectives.

The primary endpoint was to characterize the systemic and local PK of DPV 0.05% gel applied rectally by the two different methods. Secondary endpoints were to assess the safety of DPV 0.05% gel applied rectally, defined as the proportion of participants with at least one grade 2 or higher AE, and to assess the DPV gel acceptability.

Study population.

Eligible participants were healthy men or transgender women without HIV aged 18 or older at screening with a reported history of RAI in the past year. Individuals with abnormalities of the colorectal mucosa, significant gastrointestinal symptoms (such as a history of rectal bleeding), evidence of anorectal Chlamydia trachomatis (CT) or Neisseria gonorrhoeae (GC) infection, chronic hepatitis B or hepatitis C antibody positivity, or a requirement to use drugs that were likely to increase the risk of bleeding following mucosal biopsy were excluded from the study. Blood was collected for safety laboratory evaluations at study enrollment (HIV serology, complete blood count, creatinine, alanine aminotransferase, and aspartate aminotransferase) and syphilis serology, and also upon study termination (HIV serology, creatinine, alanine aminotransferase, and aspartate aminotransferase).

Study product.

The DPV gel study product contained dapivirine (0.05%), purified water (90.99%), hydroxyethyl cellulose (3.50%), polycarbophil (0.20%), propylene glycol (5.00%), methylparaben (0.20%), propylparaben (0.05%), and sodium hydroxide (0.01%). The DPV gel was developed by the International Partnership for Microbicides (IPM) and manufactured by DPT laboratories.

Safety.

AEs were graded using the Division of AIDS Table for Grading the Severity of Adult and Pediatric Adverse Events, corrected version 2.1, July 2017, and/or Addenda 2 and 3 (Male Genital [Dated November 2007] and Rectal [clarification dated May 2012] Grading Tables for Use in Microbicide Studies; http://rsc.tech-res.com/clinical-research-sites/safety-reporting/daidsgrading-tables). Colorectal tissue biopsy specimens were stained with hematoxylin and eosin and evaluated using a qualitative injury and inflammation score, as previously described (11).

Pharmacokinetics and pharmacodynamics.

DPV concentrations in plasma and rectal tissue biopsy specimens were measured using previously described liquid chromatographic-tandem mass spectrometric (LC-MS/MS) assays (15, 21). Measurements in rectal fluid collected on Dacron swabs were conducted using a similar specimen extraction approach as described for other luminal fluids (22). Assay lower limits of quantification (LLOQ) were 20 pg/mL (plasma), 0.05 ng/sample (tissue) and 0.25 ng/swab (rectal fluid). Tissue biopsy and rectal fluid results were weight-corrected using biopsy specimen and fluid weights, respectively, and reported as ng/mg. Study-specific, weight-adjusted LLOQs were a median of 0.004 ng/mg (IQR 0.004 to 0.005 ng/mg) for rectal tissue, and 0.013 ng/mg (IQR 0.009 to 0.020 ng/mg) for rectal fluid.

Four colorectal tissue “explants” were placed in 0.5 mL tissue culture media and challenged ex vivo with 10⁴ TCID₅₀ HIV-1 (BaL Lysate) for 2 h (19). HIV-1 p24 antigen assays were performed on tissue culture supernatant harvested on days 3, 7, 10, and 14. Cumulative p24 antigen (CUM p24) was calculated as the sum of the four supernatant p24 antigen concentrations with each CUM p24 result divided by the mass of each biopsy specimen (pg/mL/mg) to provide the weight-adjusted CUM p24. Values below the 10 pg/mL LLOQ, defined as below the limit of assay quantification (BLQ), were imputed as LLOQ/2.

Behavioral assessments.

Participants completed behavioral acceptability surveys at enrollment and after each dosing regimen (14). The follow-up assessments explored reactions to the product, applicator, and administration method. In-depth interviews were conducted after the application of each single dose (visit 3 and visit 5).

Statistical analysis.

Descriptive statistics were used to summarize study population demographics. All participants exposed to the study product with at least one of the application methods are included in the main analyses. The area under the concentration-time curve to 24 h after dosing (AUC_0–24h) was calculated using noncompartmental methods. BLQ observations were imputed as half the LLOQ. Geometric mean ratios of DPV concentrations and the AUC_0–24h in plasma for the CSD relative to the Applicator were estimated from a mixed effects model on the log-transformed endpoint. The model includes adjustment for period (and collection time point for the concentration levels) and a random intercept at the participant level, to account for the crossover design of the trial. A similar model was used to estimate the mean difference in log₁₀ cumulative p24 antigen in rectal tissue biopsy specimens after ex-vivo challenge, with a random effect at the participant level to account for variability between biopsy specimens within participants. Carry-over effects were tested with an interaction term for the application method and period. DPV concentrations in samples of rectal fluid, as well as change in score from the histological evaluation of rectal mucosal tissue biopsy samples (relative to baseline) were compared using a Wilcoxon’s signed rank test.

The proportion of participants experiencing at least one AE of grade 2 or higher after use of the study product were compared between methods with a logistic mixed effects model (excluding any AEs occurring before exposure to study product). The proportion of participants giving a positive answer (highest two categories in a 5-point Likert scale) to questions related to liking the study product and ease of use are reported. Unless otherwise noted, one-sided P values testing the null hypothesis of endpoints after use with CSD being as high or higher than with the applicator; 95% confidence intervals are also provided. PK calculations and statistical analyses were performed using SAS and R software (23, 24).