Abstract

Background The association between international-normalised-ratio (INR) correction and mortality in patients with major bleeding on vitamin-K-antagonists (VKA) is important for evaluating the efficacy of reversal agents for oral anticoagulants. Objectives We evaluate if INR correction (defined as ≤1.3) following intervention in major bleeding on VKA is associated with better survival, and if there is a dose-response relationship between Vitamin K (VK) and INR correction. Methods Data on patients' characteristics, haematological management and 30-day outcomes reported by 32 UK hospitals (October 2013–August 2016) were analysed. Associations between INR correction and: (a) 30-day mortality; (b) VK dose were estimated using multivariable logistic regression, using multiple imputation to handle missing INR values. Results Of 1771 patients, 77%, 73% and 33% received prothrombin-complex-concentrate (PCC), VK (92% intravenous) and red cells and fresh frozen plasma transfusion respectively. Proportionally more intracranial haemorrhage (ICH) cases (87%) than non-ICH cases (69%) received PCC. VK administration did not vary by ICH group, with 10 mg (33%) and 5 mg (28%) doses being the most common. Higher doses of VK (10 mg) were more likely to correct INR than lower doses (5 mg). Post-intervention INR > 1.3 in treated patients was associated with 3.2 (95%CI: 2.1–4.9) times higher odds of death within 30 days, compared with INR ≤ 1.3, with no difference between ICH and non-ICH. Conclusions INR correction after intervention to manage major bleeding on VKA is associated with better survival. Higher VK doses (10 mg) improve INR correction more than lower doses (5 mg) in major bleeding, but further studies are warranted to compare the relative benefits/risks of 5 mg versus 10 mg doses.