Analysis of blood pressure rhythms for drug efficacy evaluation

This study analyzes some methods of evaluating the effects of antihypertensive drugs on blood pressure circadian rhythm. We reviewed four different approaches: hourly averages, trough-to-peak ratio, cosinor method, and Fourier series applied to the same data to prove the time course of the effects of isradipine administered once daily. A total of 141 patients of both sexes (mean age 53 years, range 30-76) with mild to moderate essential hypertension were enrolled in this multicenter trial after a 2-week placebo run-in. Treatment with isradipine SRO 5 mg/day administered between 8 and 9 AM was started. Each patient underwent ambulatory BP monitoring at the time of entry and after 6 weeks of treatment. Calculation of hourly averages showed decreases after 4 AM, and from about 8-9 AM, when the drug was administered, and the decreases practically did not vary until about 10 PM. Subsequently, the decreases became smaller and indicated reduced drug activity. However, this hypothesis no longer held after 4 AM. The trough-to-peak ratio was calculated including hourly averages after the dose divided by the lowest hourly average. Both systolic and diastolic blood pressure showed constant reduction from 3 PM (time of peak) to 11 PM. However, after 11 PM, higher trough-to-peak ratios paradoxically occurred due to a major reduction obtained with placebo, and the negative percentages just before the next dose cannot be attributed to treatment. Applying the cosinor method, maximal values were greatly underestimated, nocturnal values were overestimated, and the absolute maximum occurred in proximity to the minimum relating to postprandial dip. The generalized cosinor model, known as Fourier partial series, was always curtailed to the third harmonic. Fourier analysis was able to describe the daily trend of BP both before and after isradipine administration. We used statistical tests to determine if the differences described by the models were significant. The tests indicated significant mean level reductions after therapy and no appreciable amplitude and phase-related variations. The nocturnal and periawakening intervals, in which BP changed linearly, oscillated between 3 and 4 hours. Within these intervals BP rose by 13-16 mm Hg or fell by 14-19 mm Hg. The intervals, ranges, and speed differences were tested and were never significant. The medication tested was effective only at the mean level, the variables used to characterize the time course of its effect remaining unchanged. There was no significant phase shift of the curves, and BP rhythm was preserved.