The ability of the toe-brachial index to predict the outcome of treadmill exercise testing in patients with a normal resting ankle-brachial index

Background: Peripheral arterial disease (PAD) in the presence of a normal ankle-brachial index (ABI) can be diagnosed noninvasively by measuring a postexercise ABI or by measuring the toe-brachial index (TBI). Methods: This was a prospective comparative study. Over a period of 30 months, a total of 415 patients who were referred with the suspicion of vascular claudication and resting values of 0.91 ≤ ABI<1.40 were further evaluated for the resting TBI and postexercise ABI by treadmill testing. Results: A total of 325 (39%) of the 830 investigated limbs had a low TBI (≤0.70), and 505 (61%) had a normal TBI. Of the limbs with a low TBI, 160 (49%) had PAD according to a postexercise ABI versus 165 (33%) of the limbs with normal TBI. The overall agreement in PAD classification between the two methods was 500/830 (60%) with a Cohen's kappa = 0.166 (95% CI: 0.096–0.232). The data showed an inverse correlation between the magnitude of the TBI decrease, as well as the resting ABI, and the probability of an abnormal postexercise ABI. On average, limbs with a low TBI had a lower resting ABI than patients with a normal TBI (1.07 ± 0.09 vs. 1.13 ± 0.10, P < 0.001). The groups with a low TBI had a significantly higher ratio of abnormal test results than patients with a normal TBI, in limbs with ABI (0.96–1.00) and ABI >1.10 (P ≤ 0.022 for both), but there were no statistically significant differences found in other ABI intervals (P > 0.200 for all). Conclusions: The magnitude of the TBI reduction correlates with an increased probability of an abnormal postexercise ABI. However, this is due in part to limbs with a low TBI having a lower resting ABI on average than limbs with a normal TBI, which also correlates with the probability of an abnormal exercise test result. This study shows that the TBI and the postexercise ABI are not interchangeable for establishing a PAD diagnosis.