Abstract 3115: Ankle Brachial Pressure Index (ABI) Is A Better Predictor Of Cardiovascular Mortality Than Toe-Brachial Index (TBI) Or Abnormal Doppler Flow Curves In Both Diabetic And Non-diabetic Subjects: The Hoorn Study.
Introduction: Individuals with peripheral arterial disease (PAD) have a high risk of future cardiovascular events. Many studies have demonstrated that low (<0.9) Ankle Brachial Index (ABI) predicts cardiovascular mortality in both diabetic and non-diabetic subjects. However, in diabetic subjects the measurement of the ABI is complicated by the presence of medial arterial calcification which results in falsely elevated ABI-values despite the presence of significant PAD. Therefore, we investigated whether the Toe-Brachial Index (TBI) or abnormal Doppler flow curves would be better predictors of PAD-associated cardiovascular mortality (CVM).
Methods: Associations between measures of PAD and CVM were studied in an age-, sex,- and glucose tolerance stratified sample (n = 631) of a population based cohort aged 50 –75 years followed for 15 years. Measures included Doppler assisted ABI, flow velocity tracings of the femoral, popliteal and crural arteries and measurement of the TBI. Predictors of CVM were determined by Cox proportional hazards multiple regression analysis. Results are described as hazards ratios (HR) and 95% confidence intervals (CI).
Results: At baseline ABI < 0.9 was present in 10.5% (6.9% in non-diabetic vs 17.2% in diabetic individuals), TBI < 0.7 was present in 21.0% (18.9% vs 24.3%), and abnormal Doppler flow curves in 20.3% (16.0% vs 29.3%). After 15 years of follow-up 141 patients had died of cardiovascular causes. After adjusting for age, sex, hypertension, total and HDL-cholesterol, triglycerides, BMI and smoking, HR (95% CI) were 3.19 (1.91–5.34) P<0.001 for ABI<0.9, 2.80 (0.97– 8.05) P=0.056 for presence of abnormal flow-curves, and 1.58 (0.97 – 2.57) P=0.068 for TBI<0.7. When entering the three measures simultaneously in the regression model only ABI<0.9 was an independent predictor of CVM, HR=2.81 (1.47 – 5.39) P=0.002. In diabetic individuals this effect was as least as strong, HR=3.30 (1.39 – 7.99) P=0.007.
Conclusion: Although an ABI< 0.9 may underestimate the presence of PAD in individuals with type 2 diabetes, this measure is a powerful independent predictor of CVM in both diabetic and non-diabetic subjects. Assessment of the ABI should be used more often to identify high-risk patients, especially in the diabetic population.