Modeling comparative effectiveness studies: An example using a phase IV intravenous nicardipine versus labetalol in patients with uncontrolled hypertension trial

Abstract

Background: Hypertension is a common acute presentation that requires effective medication for control. Few comparative effectiveness trials exist to guide which agents offer the most efficient response. Our objective was to perform a design simulation to determine if a future study comparing the effect of intravenous (IV) nicardipine or labetalol was warranted.

Methods: We created a predictive model using known clinical responses to currently recommended dosing of nicardipine and labetalol. For nicardipine, we used a three-compartment, weight-normalized pharmacokinetic nonlinear mixed-effects model. For labetalol, BP and HR changes were modeled as a function of time, group (pretreated or untreated within 24 hours), and total dose. Clinically relevant BP and HR changes were defined as >15% and >20% of baseline, respectively. At least 500 patients were simulated and results compared to known clinical data.

Results: Our models demonstrated that the rate of clinically relevant blood pressure drop within 30 minutes of initiation in the nicardipine group would be 61% versus 14-19%, without with a >20% HR decrease, for labetalol.

Conclusions: BP and HR models of antihypertensives can be designed to predict the published data reasonably well. In this fashion the need for comparative effectiveness trials can be assessed.