Background: Individual risk of recurrent venous thromboembolism affects patient management and might differ between men and women. We did a meta-analysis to assess from available evidence whether men and women have the same risk of recurrent venous thromboembolism after stopping anticoagulant treatment.
Methods: Eligible articles were identified by searches of MEDLINE (source PubMed, 1966 to February 2005), EMBase (1980 to February 2005), and the Cochrane database 2005, issue 1. Prospective cohort studies and randomised trials were eligible if they included patients with objectively diagnosed venous thromboembolism treated for a minimum of 1 month and followed up for recurrence after anticoagulant treatment was stopped. Data were extracted for study design, study quality, and the number, sex, and age of enrolled patients, risk factors for venous thromboembolism, treatment given, duration of follow-up, and the number of episodes of recurrent venous thrombosis.
Findings: 15 studies (nine randomised controlled trials and six prospective observational studies) enrolling a total of 5416 individuals (2729 men), of whom 816 (523 men) had recurrent venous thromboembolism after stopping treatment, were eligible for inclusion. The pooled estimate of the relative risk (RR) of recurrent venous thromboembolism for men compared with for women was 1.6 (95% CI 1.2-2.0). Significant heterogeneity was shown among individual study findings; however, the higher risk of recurrent venous thromboembolism in men than in women was consistent across predefined subgroups. The relative risk for recurrence in men from randomised trials (RR 1.3; 95% CI 1.0-1.8) was lower than that from observational studies (2.1; 1.5-2.9). The lower risk of recurrent venous thromboembolism in women did not seem to be accounted for by a reduced rate of recurrence after venous thromboembolism associated with oestrogen treatment or pregnancy.
Interpretation: Men seem to have a 50% higher risk than women of recurrent venous thromboembolism after stopping anticoagulant treatment. If confirmed by further prospective studies, this difference in risk of recurrence should be considered when duration of anticoagulant treatment is determined in individual patients.