This is the first study to evaluate the PK and PD of a candidate microbicide following coitus and clearly demonstrates that both differ significantly from results obtained in the absence of sexual intercourse. In addition to the significantly diminished antiviral activity (PD) and lower concentrations of PRO 2000 (PK), postcoital CVL had higher pH (mean 5.99 and 5.94 for visits 2 and 3, respectively) compared to CVL obtained in the absence of sex (pH 4.75 and 4.65 for visits 1 and 4, respectively). The postcoital CVL samples also had significantly greater protein concentrations compared to samples obtained in the absence of sex (). However, it is unlikely that these differences contributed to the differences in antiviral activity. No differences in the antiviral activity of PRO 2000 were observed across a broad pH range in vitro 
. Moreover, a 1
5 dilution of pooled seminal plasma (protein concentration ~6 mg/ml) had little effect on viral infection 
The extent of antiviral activity and the optimal distribution of drug required for protection are unknown. The CVL may underestimate the total amount of drug found in the genital tract, some of which may have been lost in the cell pellet during centrifugation, adherent to epithelial cells within the genital tract, or inaccessible to lavage. In addition, drug distribution to non-target cells, could also result in drug dilution and decreased efficacy. Even at Visit 4, only a small percentage of administered PRO 2000 was recovered. However, the unbound drug may be most important. The primary mechanism by which PRO 2000 inhibits HIV infection is binding to viral gp120, although some activity may be attributed to binding to CD4 receptors on target immune cells 
. If during sex, drug is redistributed, bound to cells in the ejaculate (which could include HIV-infected cells) or to target cells in the genital tract, it could provide some anti-HIV activity, although no protection was observed in the MDP301 trial. In contrast, redistribution of PRO 2000 might further diminish any impact against HSV-2, as infection likely occurs closer to the introitus (or outside the vagina) and the drug acts primarily by binding to the HSV-2 envelope glycoprotein B to competitively inhibit viral attachment and entry 
Not only did sexual intercourse result in a reduction in PRO 2000 recovered in CVL, but the in vitro
data ( and 
) suggest that SP interfere with antiviral activity. Even when the PRO 2000 concentrations in postcoital CVL (Visit 3) were similar to the concentration measured in non-postcoital samples (Visit 4) as observed with Couple 4, for example, the antiviral activity was lower.
Important limitations of this study are the single dose design, absence of a placebo gel group, reliance on CVL specimens alone, and small sample size. A single dose may underestimate the potential effectiveness of the drug, which may accumulate in the genital tract following repeated applications. In addition, differences in the duration of intercourse, ejaculate volume or composition may have contributed to variability in antiviral activity detected in postcoital samples. The addition of biopsy samples to CVL collection would provide information regarding drug distribution. Challenging biopsy samples ex vivo with HIV would be especially critical for evaluating microbicides that act intracellularly, such as tenofovir or dapivirine. However, the number, size, and site of sampling will require optimization as there is variability in the infectivity of biopsy tissue.
Although this study was not designed to evaluate endogenous antiviral activity in genital tract secretions, baseline CVL (Visit 1) inhibited both HIV and HSV-2, which is consistent with prior studies 
. Notably, the anti-HIV activity was diminished when experiments were conducted with postcoital CVL (Visit 2), suggesting interference with the endogenous anti-HIV activity. Possibly, the alkaline pH or specific enzymes and proteases in semen inactivate, degrade, or interfere with protective immune mediators. In contrast, the anti-HSV-2 activity persisted in postcoital CVL (Visit 2). Differences in the viscosity between the control buffer and CVL samples may have contributed to the endogenous antiviral activity, although most of the mucous pellets with the cells during centrifugation. In addition, prior work indicates that the anti-HSV activity correlates with the concentration of specific antimicrobial proteins including defensins (human neutrophil peptides 1–3) and lysozyme 
The endogenous anti-HSV activity in this study was less than previously observed 
and may reflect the fact that all women who participated in this study were on hormonal contraception. In a recently completed study, we found that the anti-HSV-2 activity in CVL was reduced among women on hormonal contraception compared to women not using hormonal contraception 
. Larger studies are needed to fully evaluate the genital tract environment both in the absence and following coitus.
In summary, this work demonstrates the feasibility and importance of conducting postcoital studies. The current paradigm of microbicide development should be modified to include postcoital sampling following single and repeated dosing with both active and placebo products and should be expanded to include both CVL and biopsies to more fully define the PK and PD of lead candidates prior to embarking on large-scale efficacy trials.