Objectives To assess the clinical effectiveness of thigh length versus knee length antiembolism stockings for the prevention of deep vein thrombosis (DVT) in surgical patients.

Design Systematic review and meta-analysis using direct methods and network meta-analysis.

Methods Previous systematic reviews and electronic databases were searched to February 2014 for randomised controlled trials (RCTs) of thigh length or knee length antiembolism stockings in surgical patients. Study quality was assessed using the Cochrane Risk of Bias Tool. The primary outcome was incidence of DVT. Analysis of the DVT data was performed using ORs along with 95% CIs. The I² statistic was used to quantify statistical heterogeneity.

Results 23 RCTs were included; there was substantial variation between the trials and many were poorly reported with an unclear risk of bias. Five RCTs directly comparing thigh length versus knee length stockings were pooled and the summary estimate of effect favouring thigh length stockings was not statistically significant (OR 1.48, 95% CI 0.80 to 2.73). 13 RCTs were included in the network meta-analysis; thigh length stockings with pharmacological prophylaxis were more effective than knee length stockings with pharmacological prophylaxis, but again results were not statistically significant (OR 1.76, 95% credible intervals 0.82 to 3.53).

Conclusions Thigh length stockings may be more effective than knee length stockings, but results did not reach statistical significance and the evidence base is weak. Further research to confirm this finding is unlikely to be worthwhile. While thigh length stockings appear to have superior efficacy, practical issues such as patient acceptability may prevent their wide use in clinical practice.

Systematic review registration number CRD42014007202.