Modeling circadian rhythm of diastolic blood pressure in hypertensive patients using 24‐hour ambulatory blood pressure monitoring

Background To develop a model describing the circadian rhythm of the 24‐hour ambulatory blood pressure monitoring (ABPM) measured diastolic blood pressure (DBP) in patients with mild to moderate essential hypertension when no drug is present. Methods 3,771 ABPM measurements from 107 hypertensive patients in a randomized placebo‐controlled study of an antihypertensive agent, assigned to the placebo group (n=24, 3 occasions) or in the placebo run‐in period (n=83), were analyzed by nonlinear mixed effects regression using NONMEM. FOCE with interaction was the estimation method. Results The final model was a two‐cosine term model (24‐ and 12‐hour cycles) having phase shift on clock time, with interindividual variability on the 24‐hour rhythm adjusted mean, the amplitudes of the cosine terms, phase shift, and clock time. Interoccasion variability was also included on the 24‐hour rhythm adjusted mean, the amplitude of the cosine term by 24‐hour cycle and the clock time. The population standard of the 24‐hour rhythm adjusted mean was 95.6 mmHg[95% CI: 93.5–97.7] and its interindividual and interoccasion variability were 9.2 CV%[6.2–12.1] and 3.2 CV%[0.6–5.7], respectively. The residual error was 7.48 mmHg. Conclusions A two‐cosine model was able to characterize the circadian rhythm of ABPM measured DBP in this patient population. This model can be used as the baseline model to characterize the exposure‐response relationship of antihypertensive agents. Clinical Pharmacology & Therapeutics (2005) 77 , P57–P57; doi: 10.1016/j.clpt.2004.12.109