Diagnosis of hypertension based on ambulatory blood pressure monitoring: are 24 hours of sampling enough?

In chronic hypertensive patients, the correlation between blood pressure (BP) and target organ damage, cardiovascular risk, and prognosis is closer for ambulatory BP monitoring (ABPM) than for clinical measurements. So far, most studies have been performed with ABPM for 24 hours. Thus, the criteria for diagnosis of hypertension, mainly based on mean BP values, have been established from data gathered at 15 to 30-min intervals for just one day. Sampling requirements for BP, however, have not been carefully taken into consideration. We have examined the impact of duration and frequency of BP sampling in the reproducibility of mean BP values. We studied 2603 patients with grade 1–2 essential hypertension (1257 men) and a control group of 917 normotensive volunteers (495 men), 51.9±13.8 years of age. ABPM was performed at 20-min intervals from 07:00 to 23:00 hours and at 30-min intervals at night for 48 consecutive hours. Physical activity was simultaneously monitored every minute by wrist actigraphy, and the information used to determine diurnal and nocturnal means of BP for each patient according to individual resting time. BP series were decimated to generate shorter series with data sampled at 1, 2, 3, or 4-hour intervals for 48 hours, as well as at the original sampling rate for just the first 24 hours. Bland-Altman plots indicate that the diurnal, nocturnal and 24-hour means of BP obtained from the original BP series are better reproduced with data sampled up to 3-hour intervals for 48 hours than by data sampled at 20–30 minute intervals for just one day. The range of the difference in the estimation of the 24-hour mean BP value between the original and the decimated series was about 15 mm Hg for data sampled every 3 hours. This range increased to an unacceptable very high 38 mm Hg (−18 to 20 mm Hg) for data sampled every 20–30 min for just 24 hours. The amount of potential error in the estimation of the daily mean was similar for the second 24 hours of sampling (−21 to 18 mm Hg). This study demonstrates that reproducibility of mean BP values is mainly dependent on duration of sampling but not on sampling rate. Results further indicate that, due to a poor estimation of the true mean BP value, ABPM for 24 hours seem to be insufficient for a proper diagnosis of hypertension and for the evaluation of treatment efficacy. Am J Hypertens (2004) 17, 42A–42A; doi: 10.1016/j.amjhyper.2004.03.101