Fifteen studies were included in meta-analyses (1,040 total knee replacement procedures in 991patients). All the studies clearly defined the appropriate outcome measures, but baseline comparability was poorly demonstrated. Half of studies did not clearly define the eligibility criteria. Nones of the studies used intention-to-treat analysis. In most trials randomisation was poorly described and blinding was inadequate.
Total knee replacement with a tourniquet was associated with a significant reduction in intra-operative blood loss (WMD -269.22mL, 95% CI -88.06 to -450.39; six studies) compared with non tourniquet-assisted surgery. There were no significant differences in total blood loss, postoperative blood loss measured from drainage systems, blood transfusion rates, operative time and length of hospital stay between the two groups. Significant heterogeneity was observed in all these outcomes.
There appeared to be a higher incidence of wound haematoma, peroneal nerve palsy, superficial wound healing disorders, blisters, deep vein thrombosis, and pulmonary embolism in the tourniquet group compared with the non tourniquet group.