Prevalence of peripheral artery disease by abnormal ankle-brachial index in atrial fibrillation: Implications for risk and therapy

To the Editor: Nonvalvular atrial fibrillation (NVAF) is the most common sustained arrhythmia encountered in clinical practice and is associated with a 5-fold increased risk for stroke (1). Moreover, patients with NVAF often suffer from atherosclerotic complications such as acute myocardial infarction (AMI) (2). Peripheral artery disease (PAD) is an established marker of systemic atherosclerosis but its prevalence in NVAF is still unclear. We reasoned that inclusion of ankle-brachial index (ABI), which is an established tool for diagnosis of PAD (3), in the CHA2DS2-VASc (4) score would better define the prevalence of vascular disease. Toaddress this issue, the ItalianSociety of InternalMedicine (SIMI) established an Italian registry documenting ABI inNVAF patients. The Atrial Fibrillation Registry for the ARAPACIS (Ankle- brachial Index Prevalence Assessment: Collaborative Italian Study) study is an independent research project involving all Regional Councils of SIMI. The first objective of the study was to estimate the prevalence of ABI 0.90 in NVAF patients. Consecutive patients with NVAF referred to internal medicine wards were eligible for the enrollment. Enrollment started in October 2010 and continued until October 30, 2012. Patients were enrolled if they were 18 years or older and had a diagnosis of NVAF, recording during the qualifying admission/consultation or in the preceding 12 months, and if it was possible to obtain the ABI measurement. Exclusion criteria included the following: acquired or congenital valvular AF, active cancer, disease with life expectancy <3 years, hyperthyroidism and pregnancy. We initially planned to include 3,000 patients. The Data and Safety Monitoring Board (Online Appendix) decided to perform an interim analysis to assess the prevalence of ABI in the enrolled populationsdas a higher than expected prevalence of low ABI was detecteddand decided to interrupt the patients’ enrollment. The sample size was amended as follows: a sample of 2,027 patients leads to the expected prevalence of 21% with a 95% confidence interval width of 3.5% (StataCorp LP, College Station, Texas). Among the 2,027 NVAF patients included in the study, hyper- tension was detected in 83%, diabetes mellitus in 23%, dyslipidemia in 39%, metabolic syndrome in 29%, and smoking in 15%. At least 1 atherosclerotic risk factor was detected in 90% of patients. The NVAF population was at high risk for stroke, with only 18% having a CHA2DS2-VASc score of 0 to 1, while 82% had a risk 2. Despite this, 16% were untreated with any antith- rombotic drug, 19% were treated with antiplatelet drugs (APs), and 61% with oral anticoagulants (OAC); 4% of patients were treated with both APs and OAC. Among the AF population, 428 patients (21%) had ABI 0.90 (69%); 204 patients (10%) had ABI 1.40 (Fig. 1). ABI recorded only in 1 leg was excluded from the analysis (n ¼ 14). ABI 0.90 progressively increased from paroxysmal to permanent NVAF (18%, tensive (88% vs. 82%; p ¼ 0.032), diabetic (34% vs. 20%; p < 0.0001), or smokers (20% vs. 14%; p ¼ 0.0008), or to have experi- enced transient ischemic attack or stroke (17% vs. 10%; p < 0.001). 21%, 24%; p ¼ 0.0315). NVAF patients with ABI 0.90 were more likely to be hyper- NVAF patients with ABI 0.90 had a higher percentage of CHA2DS2-VASc score 2 compared with those with ABI >0.90 (93% vs. 82%; p < 0.0001). significantly associated with a smoking habit (odds ratio [OR]: 1.99; 95% confidence interval [CI]: 1.48 to 2.66; p < 0.0001), diabetes (OR: 1.93; 95% CI: 1.51 to 2.46; p < 0.0001), age class 65 to 74 years (OR: 2.05; 95% CI: 1.40 to 3.07; p < 0.0001), age Logistic regression analysis demonstrated that ABI 0.90 was class 75 years (OR: 3.12; 95% CI: 2.16 to 4.61; p < 0.0001), and history of previous transient ischemic attack/stroke (OR: 1.64; 95% CI: 1.20 to 2.24; p ¼ 0.002). Vascular disease, as assessed by the history elements of CHA2DS2VASc score, was recorded in 17.3% of patients; inclu- sion of ABI 0.90 in the definition of vascular disease yielded a total prevalence of 33%. A higher prevalence of vascular disease was detected if ABI 0.90 was included in the CHA2DS2VASc score (Fig. 1). CHA2DS2VASc including ABI 0.90 was more associated with previous stroke (43%; OR: 1.85; 95% CI: 1.41 to 2.44; p < 0.0001) compared to CHA2DS2VASc with ABI 0.91 to 1.39 (23%; OR: 1.52; 95% CI: 1.10 to 2.11; p ¼ 0.0117). To the best of our knowledge, there is no large-scale study that specifically examined the prevalence of ABI 0.90 in NVAF. In our population, 21% had ABI 0.90 indicating that NVAF is often associated with systemic atherosclerosis. The CHADS2 has been recently refined with the CHA2DS2- VASc score, which includes vascular disease as documented by a history of AMI, symptomatic PAD, or detection of atheroscle- rotic plaque in the aortic arch (4). Comparison of vascular prevalence as assessed by CHA2DS2- NVAF patients. Inclusion of ABI 0.90 in the definition of vascular disease greatly increased the prevalence of vascular disease, which increased from 17.3% (based on history alone) to 33% (based compared with 1,381 patients, who had an ABI of 0.91 to 1.39 to better define the risk profile ofNVAFpatients with an up-grading of the risk score in each CHA2DS2-VASc score category. This may have important therapeutic implications if the new score could be tested prospectively, as a higher number of NVAF patients would on ABI) in the entire population. If ABI 0.90 was encompassed in the definition of vascular disease of CHA2DS2-VASc score the prevalence of vascular disease increased in every risk class. Inclusion of ABI0.90 in theCHA2DS2-VASc score allowed us VASc score and/or ABI 0.90 is of interest to define the poten- tially positive impact of measuring ABI in the management of potentially be candidates for an anticoagulant treatment by measuring ABI. A prospective study is, therefore, necessary to validate the risk score of this new definition of vascular disease. In conclusion, this study provides the first evidence that one-fifth of NVAF patients had an ABI 0.90, indicating that it may represent a simple and cheap method to better define the prevalence of vascular disease in NVAF.