Benefits and risks for the individual, anticoagulation for patients with atrial fibrillation

Abstract

In this thesis we explored the individual benefit risk (BR) balance for oral anticoagulants (OACs) used in patients with atrial fibrillation (AF) to prevent ischaemic stroke and other thromboembolisms.We found that compared to vitamin K antagonists (VKAs), non-vitamin K antagonists (NOACs) are equally effective in the prevention of ischaemic stroke. However, the adjusted hazard ratio (HR) for major bleeding was 2.07 with a 95% confidence interval (95% CI) of 1.27-3.38 for NOACs compared with VKAs, which was mainly attributed by the increased risk of gastrointestinal bleeding (HR 2.63, 95% CI 1.50-4.62). No major differences in characteristics were found over time between the different OAC user groups. The uptake of NOACs was much greater in the US than in the UK (79% versus 21% in 2013 of total OAC users). A reduced risk of ischaemic stroke was found in the US (adjusted HR 0.72 (95% CI, 0.61-0.90)) and an increased risk was found in the UK (adjusted HR 1.31 (95% CI, 1.04-1.65)). When conducting a formal BR assessment, it was found that warfarin is more favourable than placebo for the prevention of ischaemic stroke in patients with AF. Despite the fact that the uncertainty around input data was high, the model was found to be robust with a probability of a favourable BR balance for warfarin against placebo of 0.99. The risk scores all performed modestly (c statistic ~ 0.6) with the ATRIA risk score performing best. The ATRIA risk score correctly reclassified patients to a lower risk category. Low absolute stroke rates were found in patients with a CHA2DS2-VASc risk score of >0, which is the cut-off for treatment with an OAC according to the current guidelines. We found that whilst the overall harm-benefit balance is positive, the harm-benefit balance varies strongly amongst patients. Several characteristics are associated with an unfavourable harm-benefit balance (harm-benefit balance <0). Patients with a history of anaemia (adjusted odds ratio [OR] 7.19, 95% CI 6.19-8.36), alcohol misuse (OR 3.35, 95% CI 2.62-3.63), major bleed (OR 7.34, 95% CI 6.51-8.29), liver failure (OR 6.66, 95% CI 4.40-10.06), cancer (OR 2.44, 95% CI 2.28-2.61) and thrombocytopenia (OR 22.06, 95% CI 12.44-39.12) all have an increased risk of having an unfavourable harm-benefit balance. A method was proposed for clustering of INR patterns. A total of 6 different clusters were found and the most unstable pattern showed a ten-fold increased risk of mortality (OR 10.7, 95% CI 8.27-13.85) and a substantially increased risk of stroke (OR 3.42 95% CI 2.08-5.63), major bleed (OR 1.60, 95% CI 1.13-2.26) and minor bleed (OR 2.13, 95% CI 1.39-3.27). ORs were generally lower in clusters with more stable INR patterns over time. We confirmed that amiodarone increases the risk of bleeding when used simultaneously with warfarin (incidence rate ratio (IRR) 4.94, 95% CI 2.27-10.73). Of the statins, only simvastatin showed an increased risk of major bleed (IRR 2.60 95% CI, 1.57-4.31). We advocate that treatment should be better tailored to the patient, taking individual absolute risk and preference in to account.