The first 1000 cases of laparoscopic radical prostatectomy in the UK: evidence of multiple ‘learning curves’

Abstract

To report the initial experience of one surgeon, with contemporary experience of both open radical prostatectomy (ORP) and reconstructive laparoscopy, in laparoscopic radical prostatectomy (LRP) in 1000 patients, and to investigate the rate of change of various outcome variables for this procedure with time. Between March 2000 and December 2007, 1000 consecutive patients with clinical stage T < or = 3aN0M0 prostate cancer underwent LRP, either supervised (17%) or performed (83%), by one surgeon. The median prostate-specific antigen (PSA) level was 7.0 (1-50) ng/mL and median Gleason sum 6 (4-10); the clinical stage was T1 in 46.9%, T2 in 49.8% and T3 in 3.3%. The median (range) operative duration was 177 (78-600) min. There was one conversion (patient 8) to open surgery. The median blood loss was 200 (10-1300) mL and four patients were transfused (0.4%). The median postoperative hospital stay was 3.0 (3-28) nights. The median catheterization time was 10.0 (0.8-120) days. There were 48 complications (4.8%) requiring surgical intervention in 33 (3.3%) patients, 58% of these as a day-case admission. The positive margin rates according to d'Amico risk groups were: low, 9.1%; intermediate, 20.3%; and high, 36.8%. The overall positive margin rate was 13.3%. The PSA level was < or =0.1 mg/L at 3 months in 99.1% of patients. At a mean follow-up of 27.7 (3-72) months, 96.1% of patients were free of biochemical recurrence. In patients with a follow-up of > or =24 months potency rates peaked in the series at 86% for all men and 94% for men aged < or =65 years, and continence rates at 98% before declining thereafter in men with a shorter follow-up. The learning curve for operating time and blood loss was overcome within the first 100-150 cases, but complication and continence rates took 150-200 cases to reach a plateau. The longest learning curve was for potency, which did not stabilize until 700 cases. These learning curves are likely to be considerably shorter when surgeons are taught in departments with a high throughput of cases but both surgeons and patients should be aware of them. In view of these findings, the authors recommend that LRP should not be self-taught and should be learned within an immersion teaching programme. Even then, a large surgical volume is likely to be needed to maintain clinical outcomes at the highest level.