Atrial Fibrillation Registry for Ankle-brachial Index Prevalence Assessment: Collaborative Italian Study. (ARAPACIS)

• ABI PREVALENCE [ Time Frame: 1 year ] [ Designated as safety issue: No ]
  To estimate peripheral artery disease prevalence (defined by an Ankle-brachial index <=0.9) in AF patients.
  
  

• Vascular Events [ Time Frame: 3 years ] [ Designated as safety issue: No ]
  To Estimate ischemic cardiovascular and cerebro-vascular events (fatal or non-fatal ) incidence in AF patients with or without peripheral artery disease
  
  

Study design: Prospective longitudinal study

Methods and Materials: The investigators planned to assess at baseline and at scheduled follow up visits :

1. Ankle-Brachial Index measurement
2. Anamnestic clinical information and Anthropometric measurements
3. Echocardiogram (volume size), electrocardiogram (AF type)
4. Outcome events such as nonfatal or fatal acute myocardial infarction, target lesion or vessel revascularization nonfatal or fatal ischemic stroke, transient ischemic attack, death from any cardiac or vascular cause, death from any cause Study duration: 3 years follow-up Statistical methods: The prevalence will be calculated by exact confidence intervals (Wilson method). The cumulative incidence will be calculated by the product-limit estimator of Kaplan-Meyer and presented with confidence intervals at 95%. The incidences and prevalences will be then adjusted through appropriate multivariate analysis (using the Cox proportional hazards model and logistic model) that will take into account the effect of potential confounders. Similarly, the effect-center presence will be checked and possibly removed. Secondary endpoints will be assessed by using Log-rank test method, and by the Cox model (with time-dependent effects) multivariate analysis.

Subgroups analysis will be also conducted for patients with first onset of AF or recurrent AF Sample size: The investigators plan to include in the study n = 3,000 AF patients, with competitive recruitment between centers involved in the study. The sample size was calculated assuming an expected prevalence of 19% at time zero, and in order to obtain a confidence interval 95% to prevail at time zero whose distance from the edge is less than or equal to 1.4%. This sample size yields a power greater than 99.9% for the secondary endpoint, assuming an event rate of 19% for patients with ABI <=0.9, and 10% for patients with ABI >0.9.An interim analysis showed an ABI prevalence, calculated by exact confidence intervals, of 21% in patients with AF, it is considered to interrupt the enrollment, as the observed prevalence is greater than two percentage points higher than that assumed. The sample size is amended as follows: a sample of 2027 patients leads to the expected prevalence of 21% with a confidence interval width of 3.5. This sample size has no impact on the power of the secondary objective.