International Journal of Urology (2005) 12, 948–952 Blackwell Science, LtdOxford, UKIJUInternational Journal of Urology0919-81722005 Blackwell Publishing Asia Pty LtdSeptember 2005129948952Original Article Risk factors for leaks following RRPVJ Gnanapragasam et al. Correspondence: Vincent J Gnanapragasam PHD FRCS(ED), Prostate Research Group, University of Newcastle Upon Tyne, Northern Institute for Cancer Research, Paul O’Gorman Building, Medical School, Framlington Place, Newcastle Upon Tyne, NE2 4HH, UK. Email: v.j.gnanapragasam@ncl.ac.uk Received 11 December 2004; accepted 5 April 2005. Original Article Identification and validation of risk factors for vesicourethral leaks following radical retropubic prostatectomy VINCENT J GNANAPRAGASAM, PAUL BAKER, GEOFFREY P NAISBY AND DAVID CHADWICK Department of Urology and Radiology, James Cook University Hospital, Middlesbrough, UK Aims: To identify risk factors for anastomotic leaks at cystography following radical retropubic prostatectomy (RRP). Methods: In phase 1 the records of a 107 RRP patients were reviewed. Data collected included comorbidity, pathological factors and intra- and postoperative complications. From these, risk factors were identified that were associated with a leak at cystography. In phase 2 (n = 46) we prospectively tested if the risk factors identified could predict an anastomotic leak. Results: In phase 1 the only identifiable risk factors were that of a difficult anastamosis, an unsatisfactory intraoperative test flush and the presence of a urinary tract infection. One or more of these events were found in 17/25 (68%) of the patients who leaked (P < 0.0001). Of the eight leaks missed, five were classed as minimal and did not require repeat cystography. Within the prospective phase 2 cohort one or more risk factors were present in 7/10 (70%) of the patients who leaked. In contrast, the identified risk factors were present in only 5/36 (13.8%) of the patients who did not leak (P < 0.001). The specificity of the test was 86.1% with a sensitivity of 70%. This gave a positive predictive value of 58.3% and a negative predictive value of 91.1%. Three leaks (two minimal and one moderate) would have been missed but 31 (67.3%) patients would have avoided an unnecessary radiological study. Conclusion: Using a retrospective and prospective cohort of patients we have shown that a cystogram following RRP can be safely avoided in the absence of the identified risk factors. Key words cystogram, radical retropubic prostatectomy, risk factors, vesicourethral leak. Introduction The use of a urethral catheter to facilitate healing of the vesicourethral anastomosis is an essential part of a radical retropubic prostatectomy (RRP). The timing of catheter removal however, varies considerably from center to center. Crucial to the decision of safe catheter withdrawal is the occurrence of a leak at the vesicourethral anastomosis. This can be assessed by a cystogram performed just prior to removal of the catheter. 1 A number of studies have now suggested that early catheter withdrawal can be undertaken with few major complications. 2–6 The necessity for a cys- togram prior to catheter removal has also been the subject of much debate. While early catheter removal is feasible there are clearly a number of leaks which can only be detected by cystog- raphy. Among recently published studies Souto et al. reported a cystographic leak rate of 28.5% in patients who had catheters removed on days 4–5 post-surgery. 7 Lepor et al. found that 25% of patients had a demonstrable radio- graphic leak on day 7 post-surgery. 2 In these studies how- ever, there are a majority of men who will not leak at cystography and could have had the catheter removed safely without imaging. In this context, the identification of risk factors which can predict a leak at cystography would be very useful. In our center it has been routine practice to perform a cystogram just prior to catheter removal. The aim of this study was to rationalize the use of cystograms by identifying potential risk factors that were associated with a leak. We then prospectively tested the value of these risk factors in identifying those patients who could have had their catheters removed without the need for imaging studies. Methods In the first phase of the study we reviewed the notes of 107 patients (Group 1) who underwent RRP for localized pros- tate cancer. All procedures were performed by a single surgeon (DC) using a standard technique. Preoperative data collected included significant past medical history and medication (such as ischemic heart disease, diabetes, renal failure and the use of steroids), previous prostatic surgery and reported lower urinary tract symptoms (LUTS). Intra- operative notes were reviewed for complications and the integrity of the anastomosis which was routinely assessed with an intraoperative test flush of the catheter. An unsat- isfactory test was defined as leakage of water through the anastamosis when the catheter was flushed using 60 mL of saline. The test was deemed satisfactory if there was no visible water in the pelvis and if all 60 mL could be aspi- rated back into the syringe. The anastamosis was per-