Association of clinical attributes and treadmill walking performance in patients with claudication due to peripheral artery disease

Background: Patients with claudication secondary to peripheral artery disease have a substantial impairment in walking capacity. This study evaluated factors suspected to be correlated with treadmill walking performance in an effort to gain insights into the pathophysiology of the impairment. Methods: A multivariate model was developed to define the associations between clinical and laboratory biomarkers with treadmill peak walking time (PWT) in patients enrolled in three clinical trials. The model was initially developed in a cohort of 385 patients from one trial using 23 candidate-independent variables and then tested in the combined data from the other two trials (351 patients). Results: The final model was built from 14 variables that met the predefined univariate criteria of P < .15. Main effects remaining in the model were age, resting ankle-brachial index, smoking status, hypertension, statin use, country (United States vs non-United States countries), and high-sensitivity C-reactive protein. The model was highly statistically significant (P < .0001) but explained only a limited portion of the population heterogeneity (r(2) = 0.173). The main effects plus interaction terms had an r(2) = 0.2178. The main effects model was tested in an independent cohort of 351 patients from two other clinical trials in peripheral arterial disease that did not include high-sensitivity C-reactive protein. The model successfully fit the data set, based on prospectively defined root mean squared error and was statistically significant (P = .0005) but had lower overall explanatory power than in the index cohort (r(2) = 0.0687). Conclusions: As expected, age and ankle-brachial index contributed to exercise limitation among patients with PAD. The association of C-reactive protein, hypertension, and smoking with PWT is consistent with a role for inflammation or oxidative stress in determining treadmill walking performance. In contrast to previous reports from smaller and more homogenous populations, clinical attributes and biomarkers explain only a small portion of PWT heterogeneity.