Kullo peripheral artery disease series#

026) and a 1.9-fold increased odds of PAD for each 10-unit increase in peripheral AIX (OR, 1.94 95% CI, 1.18-3.20 P =. This corresponds to a 2.1-fold increased odds of PAD for each 10-unit increase in central AIX normalized to 75 beats/min (OR, 2.07 95% CI, 1.09-3.93 P =. 009) were independently associated with higher odds of PAD. The multivariable models revealed that central AIX normalized to 75 beats/min (odds ratio, 1.08 95% confidence interval, 1.01-1.15 P =. The resulting multivariable model adjusted for several known risk factors for PAD and included the following: age, race, hypertension, diabetes, hyperlipidemia, pack-years, and eGFR ( Tables III and IV). 10 were added to the final model, and adjustment for age and race was also made because they are frequent confounders. All covariates with a resulting P value <.

Kullo peripheral artery disease series#

To determine whether central AIX and peripheral AIX were associated with PAD independent of potential confounders, a series of univariate logistic regression analyses were performed between PAD and each potential confounder while adjusting for either central AIX or peripheral AIX. In instances in which related variables met criteria for inclusion (eg, systolic blood pressure and hypertension), the underlying disorder (eg, hypertension) was preferentially selected to avoid overfitting the model. ABI was also excluded for similar reasons. All variables in Table I were screened for inclusion in the multivariable model except for coronary artery disease and prior peripheral revascularization because these variables were used to separate participants into controls and PAD patients and thus were completely collinear with PAD. To determine which variables to include in the multivariable model, a univariate screening analysis was performed with PAD as the outcome while adjusting for either central AIX at 75 beats/min or peripheral AIX as each covariate was added. Multivariable logistic regression models were used to test whether central AIX and peripheral AIX were independently associated with PAD after adjustment for potential confounders. Log values of CRP were used because of its right-skewed distribution. Two-tailed unpaired t-tests and Fisher exact test were used to test for differences between PAD and control groups on variables of interest. Journal of Vascular Surgery Cases, Innovations and Techniquesĭata analysis was performed using Stata version 14 (StataCorp, College Station, Tex).Journal of Vascular Surgery: Venous and Lymphatic.