Method of Mean Value Calculation as an Additional Source of Variability in Ambulatory Blood Pressure Measurement

• Published in: American journal of hypertension Vol. 23; no. 7; pp. 725 - 731
• Main Authors: Frank, Michael, Peyrard, Séverine, Bobrie, Guillaume, Azizi, Michel
• Format: Journal Article
• Language: English
• Published: Basingstoke Oxford University Press 01.07.2010; Nature Publishing Group; Oxford University Press (OUP)
• Subjects: 24-h ambulatory blood pressure monitoring; Arterial hypertension. Arterial hypotension; Biological and medical sciences; Blood and lymphatic vessels; Blood Pressure; Blood Pressure - physiology; Blood Pressure Monitoring, Ambulatory; Blood Pressure Monitoring, Ambulatory - instrumentation; Blood Pressure Monitoring, Ambulatory - standards; Blood Pressure Monitoring, Ambulatory - statistics & numerical data; Cardiology and cardiovascular system; Cardiology. Vascular system; Clinical manifestations. Epidemiology. Investigative techniques. Etiology; Human health and pathology; Humans; hypertension; Life Sciences; Mathematical Computing; mean value calculation; Medical sciences; Randomized Controlled Trials as Topic; Reference Values; Software; 24-h ambulatory blood pressure monitoring; blood pressure; mean value calculation; hypertension; Hypertension; Variability; Cardiovascular disease; Arterial pressure; Blood pressure; Ambulatory; Pressure measurement; Monitoring
• ISSN: 0895-7061; 1941-7225; 1941-7225
• DOI: 10.1038/ajh.2010.47

Background 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. Methods 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. Results Mean 24-h systolic BP (SBP) was significantly lower when calculated by the hourly mean method in both normotensive subjects (−0.9mmHg (95% confidence interval (95% CI): −1.0 to −0.8); limits of agreement: −2.2 to +0.4mmHg) and hypertensive patients (−1.0mmHg (95% CI: −1.2 to −0.9); limits of agreement: −3.7 to +1.6mmHg). In hypertensive patients, daytime SBP/diastolic BP was slightly higher with the hourly mean method (0.4mmHg (95% CI: +0.3 to +0.5)/+0.4mmHg (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. Conclusion 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.