Validation of an automated measurement method for determination of the ankle-brachial index

Objective. Lower extremity atherosclerotic disease (LEAD) diagnosis is largely based on ankle-brachial index (ABI) recordings. Equipment that could automatically determine ABI may facilitate LEAD identification within a broad range of health services. We aimed to test the measurement properties of an automated oscillometric ABI measurement device (MESI ABPI MD^®) as compared to manual reference ABI measurements in patients with and without LEAD. Design. A total of 153 patients with and without LEAD visiting a vascular surgery clinic underwent manual and automated ABI measurements. In total, 306 limbs were investigated. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated to assess the automated ABI device overall validity, with the manual method as reference. Correlation analysis (Spearman) was used to assess patterns of correlation between measurement methods while Bland-Altman plots were used to quantify measurement agreement. Results. Sensitivity and specificity for the automated ABI device were 75 and 67% whereas PPV and NPV were 72 and 71%, respectively. The correlation coefficient (automated versus manual measurements) was r = 0.552, p < .01. Bland-Altman plots revealed proportional bias and a tendency by the automated device to overestimate lower ABI values and underestimate higher ABI values. The best agreement between automated and manual ABI recordings was observed within the normal ABI range. Conclusions. The ABPI MD^® device performance was unfavorable. The automated device tended to overestimate lower ABI values while underestimating higher values, which may lead to underdiagnosis of LEAD. Our data do not support the use of this automated ABI measurement device in clinical practice.