Sixteen RCTs (n=1,406, range 40 to 240) were included. Overall trial quality was low: computer-generated randomisation was used in two studies; allocation concealment was reported in eight studies; blinding of patients was stated in two studies and blinding of outcome assessors in five studies; and withdrawals were rarely reported. Follow-up in half of the studies was 12 months (range 0.7 to 48 months). There was no indication of publication bias.
For B-TURP compared with M-TURP at 12 months there were no clinically relevant differences in efficacy (Q(max) WMD 0.72mL/s, 95% CI 0.08 to 1.35; five studies). Treatment of 50 patients (95% CI 33 to 111) and 20 patients (95% CI 10 to 100) with B-TURP resulted in one fewer case of transurethral resection syndrome (RD 2.0%, 95% CI 0.9% to 3.0%; 15 studies) and one fewer case of clot retention (RD 5.0%, 95% CI 1.0% to 10%; seven studies). Moderate heterogeneity (I2=49%) was present for the clot retention analysis and absent for other comparisons. There was no significant difference between: operation times; transfusion rates; retention rates after catheter removal; and urethral complications.
Compared with B-TURP, M-TURP resulted in significantly longer irrigation (WMD 8.75h, 95% CI 6.8 to 10.7; three studies).
Sensitivity analyses and subgroup analyses were reported.