OBJECTIVE: To compare the effects on office blood pressure and home blood pressure of placebo and active drug administration. DESIGN: After a 2-week wash-out period, patients with mild-to-moderate hypertension entered a 2-week single-blind placebo period and then a 4-week double-blind period. Patients were randomly assigned to be administered either 2 mg trandolapril once daily or its placebo in a 2:1 proportion. Office blood pressure was measured by a physician at the end of each period, using a mercury sphygmomanometer (mean of three consecutive measurements). Home blood pressure was measured during the last week of each period according to standard procedure carefully taught to each patient by the physician. Compliance was checked by using electronic pill boxes. RESULTS: Data for 34 of the 44 patients who entered the study were eligible for analysis. Baseline systolic blood pressure/diastolic blood pressure were significantly (P = 0.0001/P = 0.0001) higher for office blood pressure (161/101 mmHg) than they were for home blood pressure (145/93 mmHg). There was no statistically significant difference between the placebo and active-treatment groups at baseline. During the single-blind period, blood pressures measured at the office and at home did not change significantly. Office blood pressure decreased by 2.7 +/- 10 mmHg for systolic blood pressure and by 0.5 +/- 4 mmHg for diastolic blood pressure whereas home blood pressure increased by 0.8 +/- 6 mmHg for systolic blood pressure and by 0.7 +/- 4 mmHg for diastolic blood pressure. During the double-blind period, office blood pressure fell significantly with trandolapril treatment (systolic by 10.2 +/- 12 mmHg, diastolic by 8.3 +/- 6 mmHg; P = 0.0005/0.0001, versus single-blind placebo period) but this decrease was not significantly different (P = 0.45/0.92) from the fall in members of the placebo group (systolic by 6.9 +/- 9 mmHg, diastolic by 8.0 +/-6 mmHg; P = 0.04/0.002, versus single-blind placebo period). Thus, no antihypertensive effect of trandolapril was demonstrated. The fall lin home blood pressure with trandolapril treatment was significant (systolic by 10.7 +/- 8 mmHg, diastolic by 5.8 +/- 5 mmHg; both P = 0.0001, versus single-blind placebo period) and was significantly greater (P = 0.0004/0.004) than the minimal change observed with placebo (systolic fell by 0.2 +/- 5mmHg, diastolic fell by 0.6 +/- 4 mmHg; P = 0.90/0.62, respectively, versus single-blind placebo period). The evening decrease in home blood pressure was similar to the morning decrease in home blood pressure in members of the trandolapril-treated group. The resulting morning:evening decrease in blood pressure ratio was 0.83 for diastolic blood pressure and 0.95 for systolic blood pressure. For the subgroup of responders, mean of individual ratios was 0.77 +/- 0.43 for diastolic blood pressure and 0.70 +/- 0.39 for systolic blood pressure. CONCLUSION: The placebo effect observed with office blood pressure measurements does not occur with home blood pressure measurements. Expected treatment effect can alter a physician's blood pressure readings. The precision of measurements is greater with home blood pressure (there is a lower SD). Use of home blood pressure measurements increases the power of comparative trials, allowing one either to study fewer subjects or to detect a smaller difference in blood pressure.