Seven RCTs were included (2,304 patients, including 1,156 receiving short-term therapy and 1,148 receiving long-term therapy).
Thrombotic recurrence at 12 months (7 RCTs; 2,304 patients including 1,148 and 1,156 on short- and long-term therapy, respectively).
Recurrence rates were greater in those receiving short-term compared to long-term therapy (11.1 versus 6.4%).
Long-term treatment significantly reduced the incidence of recurrence compared to short duration, regardless of the statistical method used; RR 0.60 (95% CI: 0.45, 0.79, p<0.001). No significant heterogeneity was found (p=0.16).
Limiting analysis to the three most recent and adequately designed trials also showed a significant reduction in patients receiving long-term compared to short-term therapy; RR 0.45 (95% CI: 0.32, 0.63, p<0.001). No significant heterogeneity was found (p=0.73). Major bleeding (3 RCTs; 1,823 patients).
Major bleeding was more common in those receiving long-term compared to short-term therapy (weighted mean of major bleeding was 1.1 versus 0.7%).
Duration of anticoagulant therapy (short- versus long-term) had no statistically-significant effect on the incidence of major bleeding on any statistical method used; RR 1.43 (95% CI: 0.51, 4.01, p=0.5). No significant heterogeneity was found (P=0.34).
Influence of risk factors.
Permanent risk factors and idiopathic VTE (4 RCT;, 671 patients and 652 patients on short- and long-term therapy, respectively): recurrence rates were greater in the short compared to the long duration of therapy groups (13.8 versus 6.7%).
Long-term treatment significantly reduced the incidence of recurrence compared to short-term treatment; RR 0.48 (95% CI: 0.34, 0.68, p<0.001). No significant heterogeneity was found (p=0.39).
Temporary risk factors and idiopathic VTE (4 RCTs; 276 and 284 patients on short- and long-term therapy, respectively).
Recurrence rates were less than in patients with permanent risk factors, and were greater with short-term compared to long-term therapy (5.3 versus 1.4%).
There was no significant difference between long- and short-term treatment at the p>0.01 level; RR 0.34 (95% CI: 0.13, 0.93, P=0.035). No significant heterogeneity was found (p=0.91).
Data did not allow subgroup analysis using haemorrhage as the outcome, or analysis by DVT site (proximal or calf).