Pharmacokinetics and pharmacodynamics of once-daily versus twice-daily raltegravir in treatment-naive HIV-infected patients

Abstract

QDMRK was a phase III clinical trial of raltegravir given once daily (QD) (800-mg dose) versus twice daily (BID) (400 mg per dose), each in combination with once-daily coformulated tenofovir-emtricitabine, in treatment-naive HIV-infected patients. Pharmacokinetic (PK) and pharmacokinetic/pharmacodynamic (PK/PD) analyses were conducted using a 2-step approach: individual non-model-based PK parameters from observed sparse concentration data were determined, followed by statistical analysis of potential relationships between PK and efficacy response parameters after 48 weeks of treatment. Sparse PK sampling was performed for all patients (QD, n = 380; BID, n = 384); selected sites performed an intensive PK evaluation at week 4 (QD, n = 22; BID, n = 20). In the intensive PK subgroup, daily exposures (area under the concentration-time curve from 0 to 24 h [AUC(0-24)]) were similar between the two regimens, but patients on 800 mg QD experienced ~4-fold-higher maximum drug concentration in plasma (C(max)) values and ~6-fold-lower trough drug concentration (C(trough)) values than those on 400 mg BID. Geometric mean (GM) C(trough) values were similarly lower in the sparse PK analysis. With BID dosing, there was no indication of any significant PK/PD association over the range of tested PK parameters. With QD dosing, C(trough) values correlated with the likelihood of virologic response. Failure to achieve an HIV RNA level of <50 copies/ml appeared predominantly at high baseline HIV RNA levels in both treatment arms and was associated with lower values of GM C(trough) in the 800-mg-QD arm, though other possible drivers of efficacy, such as time above a threshold concentration, could not be evaluated due to the sparse sampling scheme. Together, these findings emphasize the importance of the shape of the plasma concentration-versus-time curve for long-term efficacy.

Fig 1

Arithmetic mean (SE) concentration-time profiles for the subset of patients with intensive PK sampling at week 4. For the intensive PK evaluation, samples were collected at the following time points: predose and 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, and 24 h postdose.

Fig 2

Probability of achieving HIV RNA levels of <50 copies/ml as a function of the GM C_trough and stratified by log baseline HIV RNA for the 800-mg-QD treatment arm, showing the PK/PD relationship for the mean baseline viral load (solid) and the 25% and 75% quartiles (dashed). Probability curves are superimposed above observed data (divided by quartiles) for the GM C_trough and the percentage of patients observed with HIV RNA levels of <50 copies/ml. The median GM C_trough value, range of GM C_trough values in the quartile, number of subjects achieving HIV RNA levels of <50 copies/ml, and total number of subjects in each quartile are displayed below each quartile.

Fig 3

GM C_trough and log₁₀ baseline HIV RNA as a predictor for achieving HIV RNA levels of <50 copies/ml for pooled data from 800-mg-QD and 400-mg-BID arms. Patients of BID arm failing to achieve HIV RNA levels of <50 copies/ml (open black circles), BID arm achieving HIV RNA levels of <50 copies/ml (open red triangles), QD arm failing to achieve HIV RNA levels of <50 copies/ml (blue plus signs), and QD arm achieving HIV RNA levels of <50 copies/ml (green asterisks) are indicated. Values along the right side of the plot represent the GM C_trough values in each quartile of baseline viral load for patients who achieved (or failed to achieve) HIV RNA levels of <50 copies/ml. Values in parentheses represent the numbers of patients in each quartile of baseline viral load who achieved (or failed to achieve) HIV RNA levels of <50 copies/ml.

Fig 4

ROC curve for the 800-mg-QD arm. The ROC analysis aims to identify a threshold value of the various parameters which give the best balance between true positive and true negative. The inset table displays the parameter value, sensitivity, and specificity for the value that yields the largest separation between true positive and true negative. Log₁₀ baseline HIV RNA provides the best separation between true-positive (sensitivity) and false-positive (1 − specificity) results, with no PK parameter providing a very sensitive or specific threshold.