Diagnostic accuracy of automated oscillometric determination of the ankle-brachial index in peripheral artery disease

Objective: Peripheral artery disease (PAD) is underdiagnosed in primary care settings, partly because of limited accuracy of the Doppler ankle-brachial index (ABI). This study aimed to assess the diagnostic accuracy of an automated oscillometric ABI device compared with reference standard vascular laboratory Doppler ABI equipment and to examine the influence of oscillometric errors on performance. Methods: Consecutive patients who were referred to a large general hospital for an ABI measurement were invited to participate. In each patient, the oscillometric analysis was followed by the Doppler analysis. Legs with incompressible ankle arteries were excluded from analysis. ABI values were compared using the Bland-Altman method. Oscillometric errors were defined as the incapacity of the oscillometer to report a value of ABI. A receiver operating characteristic curve was constructed, and the area under the curve was computed. Results: A total of 201 patients participated. The Bland-Altman plot showed a mean difference of 0.05 +/- 0.12 (limits of agreement, -0.20 to 0.29), representing a small ABI overestimation after oscillometry. Oscillometric errors occurred more commonly in limbs with PAD than in limbs without PAD (28% and 7%, respectively; P < .001). Considering a 0.9 threshold and after oscillometric error exclusion, the oscillometric ABI showed a 74% sensitivity and a 97% specificity for a diagnosis of PAD. When oscillometric errors were considered as abnormal ABIs, sensitivity increased to 86% and specificity was maintained at 95%. The receiver operating characteristic curve showed an area under the curve of 0.96. The best oscillometric ABI cutoff point was 1.00. Using this threshold and when considering oscillometric errors as abnormal ABIs, sensitivity improved to 94% while maintaining specificity at 92%. Conclusions: Oscillometric ABI showed good diagnostic accuracy compared with the reference standard. However, the high incidence of oscillometric errors and the challenges to correctly interpret readings may limit the use of the oscillometric method in PAD diagnosis