Lack of agreement between office and ambulatory blood pressure responses to hydrochlorothiazide

BACKGROUND: Differences between the antihypertensive responses to drug therapy measured by office blood pressure (OBP) and ambulatory blood pressure monitoring (ABPM) techniques have been noted but rarely analyzed. We studied whether the OBP and 24-h ABPM responses to hydrochlorothiazide differ and, if so, the relevance of these differences. METHODS: The OBP and ABPM responses to hydrochlorothiazide (25 mg/d, for 4 weeks) were measured in 228 subjects with essential hypertension, and mean responses were compared between methods using the Student paired t test. To assess variation in the agreement between OBP and ABPM responses among subjects, the limits of agreement were calculated as the mean difference between OBP and ABPM responses +/-2 standard deviations. RESULTS: The mean systolic OBP response was 4.8 mm Hg greater than the response measured by ABPM (-14.3 v -9.5 mm Hg, P < .001), and the mean diastolic OBP response was 2.1 mm Hg greater than the response measured by ABPM (-7.5 v -5.5, P < .001). The limits of agreement between the OBP and ABPM responses ranged from -18.7 to +28.2 mm Hg for systolic response and from -12.9 to +17.1 mm Hg for diastolic response. The systolic and diastolic OBP and ABPM responses were in opposite directions in 22.8% and 23.7% of the subjects, respectively. CONCLUSIONS: Compared to ABPM, OBP overestimates the mean systolic and mean diastolic blood pressure responses to hydrochlorothiazide. Variation among subjects in the magnitude and direction of responses renders OBP an unreliable predictor of ABPM responses.