Objective: This study prospectively assessed the diagnostic accuracy of a novel bilateral photoplethysmography toe pulse measurement technique for the detection of significant lower limb peripheral arterial disease.
Method: Bilateral photoplethysmography toe pulse measurements were compared with the ankle-brachial pressure index (ABPI) gold standard reference. Pulse wave analysis techniques extracted timing, amplitude, and shape characteristics for the great toes and their right-to-left side differences. These characteristics were compared with previously obtained normative ranges, and the accuracy was assessed for all significant disease (ABPI <0.9) and higher-grade disease (ABPI <0.5). Measurements were collected in a controlled environment within a tertiary vascular surgical unit for 111 subjects (age range, 42-91 years), of whom 48 had significant lower limb peripheral arterial disease and 63 were healthy. Subjects were matched in age, sex, height, body mass index, and heart rate. Diagnostic performance was assessed using diagnostic sensitivity, specificity, accuracy, negative-predictive and positive-predictive value, and the kappa statistic representing agreement between techniques beyond chance.
Results: The degree that pulse shape fell beyond the normal range of normalized pulse shapes was at the threshold of substantial to almost perfect agreement compared with ABPI for significant disease detection (diagnostic accuracy, 91% [kappa = 0.80]; sensitivity, 93%; specificity, 89%), and with 90% accuracy (kappa = 0.65) for higher-grade disease detection. Pulse transit time differences between right and left toes also had substantial agreement with ABPI, with diagnostic accuracy of 86% for significant disease detection (pulse transit time to pulse foot [kappa = 0.71] and to pulse peak [kappa = 0.70]) and reached at least 90% for these for the higher-grade disease. The performance ranking for the different pulse features mirrored an earlier pilot study. With the shape and pulse transit time measurements, the negative-predictive values of the 5% disease population screening-prevalence level were at least 99% and had positive-predictive values of at least 98% for the 90% disease-prevalence level for vascular laboratory referrals.
Conclusion: This simple-to-use technique could offer significant benefits for the diagnosis of peripheral arterial disease in settings such as primary care where noninvasive, accurate, and diagnostic techniques not requiring specialist training are desirable. Improved diagnosis and screening for peripheral arterial disease has the potential to allow identification and risk factor management for this high-risk group.