The accuracy of automated blood pressure measuring devices in patients with controlled atrial fibrillation

OBJECTIVE: To determine whether automated blood pressure measuring devices can measure blood pressure accurately in patients with atrial fibrillation. DESIGN: Comparison of the accuracy of two electronic sphygmomanometers Takeda UA-751 (Takeda) and Copal UA-251 (Copal) and two ambulatory blood pressure monitors Accutracker 1 (Accutracker) and SpaceLabs 90207 (SpaceLabs) with that of a trained observer using a Hawksley random-zero sphygmomanometer (Hawksley), using the sequential same-arm technique. SETTING: University teaching hospital: medical wards and outpatient department. SUBJECTS: Twenty-eight patients, mean +/- SD age 72 +/- 9 years, blood pressure range 90-158/40-96 mmHg, in atrial fibrillation with a controlled ventricular rate. MAIN OUTCOME MEASURES: The proportion of machine readings > 5 mmHg different from the Hawksley reading was compared with that obtained by three sequential Hawksley measurements. The variability of each measuring method was assessed by determining the SDD for the paired readings from each device. RESULTS: Five per cent of Takeda, 5% of Copal, 14% of Accutracker and 21% of SpaceLabs readings could not be obtained. Sequential testing with the Hawksley resulted in an accuracy at the 5-mmHg level of (systolic/diastolic) 79/79%, compared with 64/54% (P < 0.05 for diastolic) for the Takeda, 68/75% (NS) for the Copal, 50/36% (P < 0.01 for both) for the Accutracker and 50/29% (P < 0.01 for systolic, P < 0.001 for diastolic) for the SpaceLabs. Intrapatient variability, as assessed by SDD, was 8.3/8.6 mmHg for the Hawksley, similar to that for the Copal (7.7/7.3 mmHg) but higher for the Takeda (11.2/19.7 mmHg), the Accutracker (22.4/26.3 mmHg) and the SpaceLabs (7.5/14.8 mmHg). CONCLUSIONS: Accurate measurement of blood pressure with an electronic device is possible in patients who have atrial fibrillation; the Copal UA-251 provides a satisfactory level of accuracy. However, the marked difference between devices and the limited accuracy of the other machines tested here demonstrates the need to ensure that such devices are of proven accuracy in this patient group.