Eleven RCTs (n=3,343) were included. Six RCTs evaluated LMWH (n=1,801), four evaluated warfarin (n=1,265) and one evaluated UFH (n=277).
Three studies scored 2 or less for quality on the Jadad scale. Two studies scored 5 points and two used an intention-to-treat analysis. Two studies were not evaluable and 6 studies scored 3 or more and were considered to be good quality. Quality criteria were reported for the individual studies.
No significant heterogeneity was found for 1-year mortality and bleeding complications. The results from fixed-effect models are reported.
There was a statistically significant reduction in 1-year overall mortality among patients allocated to anticoagulants compared with control treatments (RR 0.905, 95% CI: 0.847, 0.967, p=0.003). The reduction in 1-year mortality was significant for studies of LMWH (RR 0.877, 95% CI: 0.789, 0.975, p=0.015), but there was no significant difference between warfarin and control treatments (p=0.239). There was no significant difference between UFH and the control treatment in the one study evaluating UFH. There was no significant difference between LMWH and warfarin in 1-year mortality (p=0.418); the ARDs were 8.0% and 3%, respectively.
There was a statistically significant increase in all bleeding complications among patients allocated to anticoagulants compared with control treatments (RR 2.309, 95% CI: 1.928, 2.764, p<0.0001). The increase in bleeding complications was significant for LMWH (RR 2.029, 95% CI: 1.205, 3.417, p=0.008) and warfarin (RR 2.366, 95% CI: 1.954, 2.866, p<0.0001). There was no significant difference between UFH and the control treatment in bleeding complications in the one study evaluating UFH. There was a significantly greater increase in bleeding complications in patients allocated to warfarin compared with LMWH; the ARD was 22.5% versus 2.4% (p<0.0001).
There was a statistically significant increase in major bleeding complications among patients allocated to anticoagulants compared with control treatments (p<0.0001). The increase in major bleeding complications was significant for warfarin (p<0.0001) but not LMWH (p=0.128).
There was no significant difference between anticoagulants and control treatments in fatal bleeding complications (0.49% versus 0.17%, p=0.542). None of the studies reported any cases of fatal pulmonary embolism.
There was no evidence of publication bias.
Other results were also reported.