Cervicovaginal and Rectal Fluid as a Surrogate Marker of Antiretroviral Tissue Concentration: Implications for Clinical Trial Design

Quantifying tissue drug concentrations can yield important information during drug development, but complicates pharmacokinetic study design. Mucosal fluids collected by direct aspiration (cervicovaginal fluid; CVF) or swab (rectal fluid; RF) might be used as tissue concentration surrogates, but these relationships are not well characterized.

Mucosal fluids were positively correlated with tissue concentrations and including plasma concentrations improved the regression models in most cases. Dose linearity in CVF, but not RF, suggests a saturation process in lower gastrointestinal tract tissue. These findings suggest that mucosal fluid and plasma concentrations may be used for qualitative inference of tissue concentrations for these antiretrovirals.

Forty-nine healthy women, given a single oral dose of tenofovir, maraviroc, emtricitabine, or raltegravir at 50%-200% of the treatment dose, provided 13 plasma, 12 CVF, 12 RF and one cervical, vaginal and rectal tissue biopsy over 48 hours. Relationships between these paired samples were characterized by linear and multiple linear regression. Adjusted r values were used to select the final predictive models.

CVF exposure increased linearly with dose for all antiretrovirals (r(2) ≥ 0.23, P ≤ 0.02) except raltegravir (r(2) = 0.08, P = 0.19). In RF, only emtricitabine increased linearly with dose (r(2) = 0.27, P = 0.01). For all antiretrovirals, CVF and RF concentrations significantly correlated with mucosal tissue concentrations (female genital tract r(2) ≥ 0.37, rectal tissue (2)r ≥ 0.50, P ≤ 0.001). In the final multivariate models, plasma and fluid concentrations were both associated with FGT concentrations for all antiretrovirals (r(2) ≥ 0.81, P < 0.001). The same was noted for rectal tissue (r(2) ≥ 0.58, P < 0.001) except for tenofovir, for which RF alone was predictive of tissue concentration (r(2) = 0.91, P < 0.001). Conclusions: Mucosal fluids were positively correlated with tissue concentrations and including plasma concentrations improved the regression models in most cases. Dose linearity in CVF, but not RF, suggests a saturation process in lower gastrointestinal tract tissue. These findings suggest that mucosal fluid and plasma concentrations may be used for qualitative inference of tissue concentrations for these antiretrovirals.