This study investigated whether the 'surface ratio', a novel index to characterize long-acting antihypertensive agents, would provide a more reproducible estimate of the duration of the antihypertensive effect than the more commonly used trough-to-peak ratio. In 66 hypertensive patients (diastolic pressure on conventional measurement > 95 mmHg), the ambulatory blood pressure was measured on a placebo at baseline and 2 months later, while the patients took 10 mg lisinopril once a day between 7 p.m. and 11 p.m. Diurnal treatment effect curves were obtained by subtracting the blood pressure at baseline from the corresponding value at 2 months for all time intervals considered in the analysis. In order to calculate the surface ratio, the area under the treatment effect curve was divided by the product of the maximal blood pressure lowering effect and the dosing interval (24 h). Reproducibility of the trough-to-peak and surface ratios was investigated by the Bland and Altman techniques. At 2 months, lisinopril reduced (+/- standard deviation) the 24 h pressure by 13 +/- 16 mmHg systolic and by 8 +/- 8 mmHg diastolic (p < 0.001). According to the usual approach, disregarding inter-individual variability, the trough-to-peak ratio was 0.7 for systolic and diastolic pressure. When in individual patients diurnal treatment effects curves with a 1 h resolution were investigated, the median trough-to-peak ratio was 0.30 for systolic pressure (5th-95th percentile interval [PI]: -0.51, 0.82) and 0.28 for diastolic pressure (PI: -0.37, 0.78); the corresponding values for the surface ratio were 0.33 (PI: 0.03, 0.58) and 0.30 (PI: -0.01, 0.55). In the same manner, the trough-to-peak ratios and surface ratios became larger when the individual blood pressure profiles were progressively smoothed by substituting 1 h averages by 2 h moving averages, 2 h averages, 3 h moving averages or by 3 h averages. The distributions of the trough-to-peak ratios and surface ratios were non-normal in 37 of 40 instances (p < 0.001, Shapiro-Wilk's test). Consistency was higher (p < 0.001) for the surface than for the trough-to-peak ratios. The within-subject reproducibility of the surface ratios tended to be superior to that of the corresponding trough-to-peak ratios. In conclusion, the surface ratio provides an index of the duration of action of antihypertensive agents. Moreover, in the present patients, the surface ratio tended to be characterized by a higher within-subject reproducibility than the trough-to-peak ratio.