Individual smoothing of blood pressure profiles to define responders or non-responders to drugs

OBJECTIVE: To compare different methods of individual therapyh efficacy assessment in order to define responding subjects. METHODS: Hypertensive patients were included in three double-blind clinical trials (placebo versus bisoprolol, lisinopril and amlodipine) and ambulatory blood pressure measurements (four per hour) were performed at the end of each month. We analysed the effect of therapy (placebo minus treatment) according to the following criteria: type of model (hourly mean, moving average, fast Fourier analysis), determination of the time to the peak effect (the lowest value of the modelled blood pressure) and the sampling time around this peak (1, 2,., 24 h). RESULTS: Regardless of the type of model, the level of individual therapy efficacy is significantly higher than that of the overall subjects (group efficiency), when the sampling time around this peak decreases. The proportion of responders decreases as the sampling time used to calculate the drop in blood pressure increases, whatever the kind of model and the threshold used to define responders (5 or 10 mmHg in systolic blood pressure). CONCLUSION: By this method, it is possible to appreciate the percentage of subjects considered individuallyas responders according to the time around the peak. This evaluation complements information given by the trough: peak ratio.