Parkhomenko, Alexander

Abstract Background Many patients with atrial fibrillation (AF) have comorbid vascular disease. The effects of oral anticoagulation (OAC) in AF patients with vascular disease, however, have not been widely studied. Purpose To investigate the impact of OAC in AF patients with (Vasc) or without (nVasc) concomitant vascular disease. Methods GARFIELD-AF is the largest multinational, prospective AF registry. The study comprised 51,574 GARFIELD-AF patients with newly diagnosed AF, 13,365 Vasc and 38,209 nVasc patients. All patients who reported coronary artery disease, aortic or peripheral artery disease, acute coronary syndromes, myocardial infarction, stenting, or coronary artery bypass graft were classified as having vascular disease. Adjusted hazard ratios were obtained via Cox proportional-hazards models to quantify the association of vascular disease with selected endpoints. Comparative effectiveness analyses were restricted to patients enrolled from April 2013-September 2016 (when NOACs became widely available) and who were eligible for anticoagulation (CHA2DS2-VASc ≥2 excl. gender). To evaluate the safety and efficacy of different anticoagulation strategies in Vasc and nVasc patients, propensity score using an overlap weighting scheme was applied. Weights were applied to Cox proportional-hazards models to estimate the effects of OAC vs No OAC and NOAC vs VKA. Results Vasc patients were older (median (Q1; Q3): 72.0 (65.0; 79.0) vs 70.0 (62.0; 78.0) and more often male (62.0 vs 53.6%). Vasc patients had a higher rate of comorbidities including heart failure, hypertension, and diabetes. Vasc patients received less OAC (62.8 vs 68.3%). NOACs were less common compared with nVasc patients (23.8% vs 28.7%) but a similar proportion of VKAs was observed in both (39.0% vs 39.6%). Antiplatelet monotherapy was more common in Vasc (31%) than nVasc (18%) patients. At 2-years, Vasc was associated with a higher risk of all-cause (HR [95% CI]: 1.30 [1.16–1.47]) and cardiovascular mortality (1.59 [1.28–1.97]). OACs significantly lowered the risk of all-cause mortality and stroke in nVasc patients (0.72 [0.63–0.82] and 0.64 [0.49–0.84], respectively), but not in nVasc patients. OACs led to a numerical increase in major bleeding in Vasc patients (1.32 [0.90–1.93]) and a significant increase in major bleeding in nVasc patients (1.40 [1.03–1.90]) (Figure 1). Compared with VKAs, NOACs did not significantly improve the risk of outcomes in nVasc patients. In Vasc patients however, NOACs significantly lowered the risk of all-cause mortality (0.74 [0.61–0.90]) and major bleeding (0.45 [0.29–0.70]) compared with VKAs (Figure 2). Conclusion AF patients with vascular disease have worse long-term outcomes than those without. They receive less often OAC, specifically NOAC, and more antiplatelet agents. The beneficial effects of NOAC over VKA are much more pronounced in patients with than in those without vascular disease. Funding Acknowledgement Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): This work was supported by the Thrombosis Research Institute (London, UK).

Abstract Aims Most clinical risk stratification models are based on measurement at a single time-point rather than serial measurements. Artificial intelligence (AI) is able to predict one-dimensional outcomes from multi-dimensional datasets. Using data from Global Anticoagulant Registry in the Field (GARFIELD)-AF registry, a new AI model was developed for predicting clinical outcomes in atrial fibrillation (AF) patients up to 1 year based on sequential measures of prothrombin time international normalized ratio (PT-INR) within 30 days of enrolment. Methods and results Patients with newly diagnosed AF who were treated with vitamin K antagonists (VKAs) and had at least three measurements of PT-INR taken over the first 30 days after prescription were analysed. The AI model was constructed with multilayer neural network including long short-term memory and one-dimensional convolution layers. The neural network was trained using PT-INR measurements within days 0–30 after starting treatment and clinical outcomes over days 31–365 in a derivation cohort (cohorts 1–3; n = 3185). Accuracy of the AI model at predicting major bleed, stroke/systemic embolism (SE), and death was assessed in a validation cohort (cohorts 4–5; n = 1523). The model’s c-statistic for predicting major bleed, stroke/SE, and all-cause death was 0.75, 0.70, and 0.61, respectively. Conclusions Using serial PT-INR values collected within 1 month after starting VKA, the new AI model performed better than time in therapeutic range at predicting clinical outcomes occurring up to 12 months thereafter. Serial PT-INR values contain important information that can be analysed by computer to help predict adverse clinical outcomes.