Prevention and treatment of bleeding complications in patients receiving vitamin K antagonists, part 2: Treatment

Oral anticoagulant therapy with coumarins is widelyused for the prevention and treatment of venous andarterial thrombosis. The most common complication ofvitamin K antagonist therapy is bleeding, with majorbleeding events occurring in 1–3% of patients annually.Although a number of potential predictors for bleedinghave been described, little research is available to pro-vide evidence-based guidelines for the prevention ofbleeding. To address this knowledge gap, we assembleda panel of international experts and posed a series offocused clinical questions. The experts were asked toperform a systematic literature review and summarize theresults of that review within the context of their clinicalquestion. In many cases, data were lacking and theexperts were asked to supplement their answer with clin-ical expertise. To minimize bias the reviews were vettedby three internationally recognized scholarly bodies. Ourgoal in this project is to provide ‘‘best evidence’’ forclinicians faced with the problem of minimizing the riskof bleeding in patients on vitamin K antagonist therapy.The number of patients receiving vitamin K antagonisttherapy increases annually and probably exceeds three mil-lion in the United States. Despite monitoring and carefuldose adjustment, the time spent in the therapeutic Interna-tional Normalized Ratio (INR) range varies from below 47%to a maximum of 81% of the time [1]. Major bleedingevents occur in up to 3% of ‘‘standard risk’’ patients treatedwith an oral vitamin K antagonist in one year. Despite thehigh frequency of use of vitamin K antagonists, very littleresearch has been performed examining techniques toreduce the risk of bleeding. In particular, the role of clinicalbleeding prediction rules is uncertain and further study ofthe role of drug and food interactions, the impact of patienteducation, the risk to benefit ratio of combining differentantithrombotic drugs, the role of pharmacogenomic testing,and on the optimal management of patients presentingwith elevated INR values is required. However, despitethese needs clinicians are still required to initiate andmonitor vitamin K antagonists while mitigating the risks ofbleeding and thrombosis. This article provides recommen-dations designed to assist clinicians in estimating andmanaging bleeding risk in patients receiving oral vitamin Kantagonists.Clinical questions were formulated by three of theauthors (WA, MC, DG) and authors were selected basedon their published experience. In each case, a North Ameri-can and Italian expert were asked to co-author theresponse to the question. The authors were asked to per-form a comprehensive literature review and make specificrecommendations on the basis of this review in concertwith their clinical experience. Given the anticipated lack ofevidence, the recommendations are largely opinion-based.To increase the rigor of the recommendations, each sectionwas carefully reviewed by members of the Italian Societyfor Studies on Haemostasis and Thrombosis, the Federa-tion of Centers for the Diagnosis of Thrombosis and Sur-veillance of Antithrombotic Therapies, and the Anticoagula-tion Forum.To provide readers with information on the strength of therecommendations, we have asked the authors to use theAmerican College of Chest Physicians GRADE system toevaluate the quality of the recommendations [2]. In thissystem, grades of 1A, 1B, 1C, 2A, 2B, and 2C are possible.A 1A recommendation is a strong recommendation basedon evidence from randomized trials or exceptional qualityobservational data which is thought to apply to mostpatients, in most circumstances. A 2C recommendation is avery weak recommendation, based on poor quality evi-dence, and for which there is doubt as to which patientsthe recommendation applies. Our goal with this project is toboth provide guidance to clinicians and to stimulate addi-tional research designed to improve the quality of evidenceon which treatments for vitamin K antagonist associatedbleeding are based.Clinical Scenario 1A 79-year-old woman with atrial fibrillation and well-con-trolled hypertension is assessed for anticoagulant therapy.She has no history of diabetes, congestive heart failure,stroke, or transient ischemic attack (TIA). Two years agoshe had an upper gastrointestinal bleed treated with a pro-