A Systematic Review Evaluating the Effects of Treatment Duration on Bleeding Due to Factor-specific Oral Anticoagulant Therapy

Oral factor-specific anticoagulants have been a highlight of thrombosis research over the past few years. Apart from predictable pharmacokinetic properties, multiple randomized control trials have demonstrated that they are non-inferior to other traditional anticoagulants in terms of therapeutic activity. These new anticoagulants seem to have a lower bleeding risk when compared with conventional therapy offering the potential for safer therapy over longer durations of treatment.

This is a systematic review to determine (1) the homogeneity of bleeding-classifications among the randomized trials, and (2) the relationship between bleeding event and treatment duration.

Ovid MEDLINE databases from 1946 to May 2012 were searched using apixaban, betrixaban, dabigatran, edoxaban, or rivaroxaban as keywords. Reviews, case reports, subgroup analyses, ex-vivo, in-vitro, or animal studies were excluded; only randomized controlled trials were included in the final review. Data regarding the study design, study setting, therapeutic intervention, bleeding classification and bleeding outcomes were extracted.

The studies were categorized into two major groups: (1) non-operative studies for the treatment of patients with atrial fibrillation, acute coronary syndrome, deep vein thrombosis, or pulmonary embolism, (2) perioperative studies evaluating patients after knee or hip replacement. Criteria in the Cochrane guidelines were used to define the risk of bias. The definitions of bleeding classification were assessed for homogeneity. Results from studies which had homogeneous bleeding classifications were pooled and the correlation between bleeding events and the duration of treatment was analyzed using the correlation coefficient (R²value). For each drug, the standard dose(s) that were commonly used in the latest clinical trials were chosen to be included in the final analysis.

522 publications were found in the primary search. Two independent investigators identified 43 eligible trials. In general, bleeding events were classified into major and non-major bleeding, with clinically relevant non-major bleeding as a subset of non-major bleeding in newer trials. The definition of major bleeding was relatively homogenous for both non-operative and perioperative studies, although the classification of non-major bleeding events was more heterogeneous.

Given that major bleeding events were consistently reported in most studies and the criteria of the classification are comparable, the rate of major bleeding was pooled to evaluate against the duration of treatment. The R² value was 0.784 for non-operative studies, suggesting that the rate of major bleeding events was moderately associated with the duration of treatment with an event rate of 2.6 bleeds per year. In contrast, the R²value was −0.398 for perioperative studies, suggesting that the duration of treatment was not a determining factor for the rate of major bleeding events in the perioperative population.

Although the classification of bleeding events is evolving, the definition of major bleeding has been consistent enough among studies allowing direct comparison and pooled analysis. For patients with thromboembolisms requiring long term anticoagulant therapy at full therapeutic dose for more than 12 months, the literature showed that the risk of having a major bleed was 2.6% per year. Therefore, the decision for long term anticoagulant use should be balanced against the ongoing thrombosis risk. After the acute phase when patients have lower prothrombotic risk, dose reduction should be considered to reduce the risk of bleeding complications. On the other hand, for patients requiring anticoagulant prophylaxis after major surgery, the risk of having major bleed in the immediate post-operative period is not completely dependent on the duration of treatment due to other confounding factors such as the surgical intervention itself. If certain patients are at a high risk of thrombosis after surgery, extended duration of anticoagulant prophylaxis can be safely considered. The results of this systemic review should help physicians and future investigators reducing the bleeding complications associated with new oral factor-specific anticoagulants, thereby improving their therapeutic and prophylactic values.

Chan:BMS: Chair of Adjudication Committee, Chair of Adjudication Committee Other; Aventis: Chair of Steering Committee, Chair of Steering Committee Other; Boehringer ingelheim: Member of DMSB for Clinical Trial, Member of DMSB for Clinical Trial Other; Bayer: Consultancy.