A Cochrane review included 4 studies with a total of 527 subjects undergoing knee arthroscopy. The relative risk (RR) of thrombotic events (all sites, diagnosed both clinically and by ultrasound) was 0.16 (95% CI 0.05 to 0.52; 4 studies, n=527) comparing any type of (LMWH) versus no treatment. The number needed to treat to prevent one thrombotic event was 17. All thrombotic events but one (pulmonary embolism in the LMWH group) were distal venous thrombosis. Only one study reported clearly episodes of DVT that were first noticed through clinical diagnosis, and 3 patients out of 239 had clinical symptoms of DVT (all in the control group); the calculated RR was 0.15 (95% CI 0.01 to 2.85). Adverse events were more common in the intervention group than in the control group (RR 2.04, 95% CI 1.21 to 3.44). There were 66 episodes of adverse events. The number needed to harm was 20 for any adverse events.
Comment: The quality of evidence is downgraded by study quality (unclear allocation concealment and blinding), and by imprecise results (wide confidence intervals).