Method of mean value calculation as an additional source of variability in ambulatory blood pressure measurement

There is no consensus on how the summary mean blood pressure (BP) values should be calculated in ambulatory BP (ABP) monitoring. We report the absence of agreement between two common methods of calculation, either using the arithmetic mean of all valid individual measurements, or the average of hourly means.

ABP recordings were made with SpaceLabs 90207 monitors. The means of hourly means, as reported by SpaceLabs 92506 software, were compared to arithmetic means calculated independently after raw data extraction. A total of 422 ABP recordings (n = 134 for normotensive subjects and n = 288 for hypertensive patients) were eligible for comparison. Agreement between both methods was assessed according to the Bland-Altman method.

Mean 24-h systolic BP (SBP) was significantly lower when calculated by the hourly mean method in both normotensive subjects (-0.9 mm Hg (95% confidence interval (95% CI): -1.0 to -0.8); limits of agreement: -2.2 to +0.4 mm Hg) and hypertensive patients (-1.0 mm Hg (95% CI: -1.2 to -0.9); limits of agreement: -3.7 to +1.6 mm Hg). In hypertensive patients, daytime SBP/diastolic BP was slightly higher with the hourly mean method (0.4 mm Hg (95% CI: +0.3 to +0.5)/+0.4 mm Hg (95% CI: +0.3 to +0.4)) than with the arithmetic mean method. Although small, these differences in daytime BP resulted in misdiagnosis of uncontrolled hypertension in eight (2.8%) recordings.

Significant differences exist between the mean of hourly means and arithmetic means methods, especially for 24-h BP. For daytime BP, the difference is small, but nevertheless results in the lack of agreement between the two methods used to diagnose uncontrolled hypertension.