Stepwise multiple regression analyses were applied to 50 raltegravir pharmacokinetic profiles from 50 HIV patients with the goal to identify limited sampling strategies for the prediction of drug area under the time-concentration curve (AUC(0-12)). Raltegravir single sampling point-based equations failed to reliably predict daily drug exposure. Conversely, different algorithms based on few samples and associated with good correlation, acceptable bias, and imprecision with the measured raltegravir AUC(0-12) were identified. These models could used to predict raltegravir exposure for clinic or research purposes.