European Urology

Bladder Cancer

Transurethral resection (TUR) is the mainstay of diagnosis and treatment of bladder cancer. For this reason, repeat TUR has been advocated in the case of an incomplete first resection, and restaging TUR should be always performed in cases in which the first TUR failed to provide all of the needed pathologic data (eg, no or small fragment of muscle in the biopsy) [1], [2], and [3]. With these indications, there is no reason to question the recommendation of a procedure that should always be recommended.

Despite TUR being a standard procedure in the urologic armamentarium, significant variability in TUR quality has been shown [4] and may significantly affect the probability of tumor recurrence and progression. Consequently, performance of a further TUR has been suggested in some cases in which an apparently complete initial resection was performed and muscle was present in the TUR specimen, namely, in case of T1 and/or high-grade cancers.

Much literature on the issue has been published, and several studies have clearly demonstrated that a second TUR can demonstrate the presence of residual cancer in 20–78% of the cases [5] and that pathologic findings from a second TUR can modify patient management in a considerable percentage of patients. Individually, those findings can justify the use of a second TUR, which is currently recommended by all of the most important urologic and oncologic guidelines in case of T1 or high-grade cancers [1], [2], and [3]. Notably, the National Comprehensive Cancer Network guidelines [3] also recommend a second TUR in the presence of lymphovascular invasion in the TUR specimen.

Despite these recommendations, the long-term effect of a second TUR on patient outcome has not been extensively evaluated. Divrik et al previously published and recently updated data from a randomized controlled trial (RCT) evaluating this issue [6], and [7]. In the most recent publication [7], the authors reported on a group of 210 newly diagnosed T1 cancers for which complete TUR was supposed to have been done. Those patients were randomized to a second TUR within 2–6 wk from the initial TUR (105 patients) or to the standard clinical practice as dictated by the pathologic findings at initial resection (105 patients). At a mean follow-up duration of about 66 mo, the authors demonstrated significantly higher probabilities of recurrence-free survival (59% vs 32%, p = 0.0001) and progression-free survival (93% vs 79%, p = 0.0001) in those patients undergoing a second TUR. Moreover, the number of cancer-related deaths was lower in the treatment group, and second TUR was an independent predictor of both recurrence and progression in multivariable analyses, once adjusted for the effects of patient age and sex and tumor number, size, and grade.

From the clinical point of view, the study [7] has some interesting strengths, including accurate patient selection, use of perioperative intravesical chemotherapy, and long follow-up duration. The authors should be commended for having designed such a study in 2000 and performed it through the years. Methodologically speaking, however, Divrik et al's RCT is undoubtedly of low quality due to the presence of a long list of major drawbacks.

First, primary, coprimary and secondary end points were not defined in the text [7], preventing us from understanding the main purpose of the study in the RCT design. Second, it is unclear how sample size was determined. Estimation of the number of patients needed to be enrolled was not planned; if estimation was done, figures were not reported, making it impossible to evaluate the statistical power of the RCT. Third, the method of randomization was not described, preventing estimation of the adequacy of allocation concealment.

Moreover, no blinding procedure was adopted. A sham control might have been used to effectively blind participants in such a study, although the ethical acceptability would be questionable. Considering the objective nature of both disease recurrence and progression, blinding of data collection and data analysis would have been sufficient and easily feasible. Consequently, due to unclear randomization and lack of blinding procedures, the risk of ascertainment bias is not marginal in the study [7].

Finally, the authors [7] failed to report an intention-to-treat (ITT) analysis. Specifically, the authors did not summarize the outcomes of all patients randomized but rather deleted those patients receiving bacillus Calmette-Guérin (BCG) or radical cystectomy following the second TUR (12 cases). In doing that, the authors inadvertently reduced the risk of recurrence and progression of the patients in the treatment arm, and this reduction, by itself, can explain most of the study results. Similarly, patients lost to follow-up (four cases) or who did not complete intravesical chemotherapy (three cases) in the control arm were excluded from analysis, and this exclusion could have underestimated the risk of recurrence and progression in the control arm.

It is highly unlikely that the overall effect of patient exclusions in both arms did not affect the study results, considering that the patients excluded from the treatment group seem to have had high-risk disease. The authors [7] should have analyzed all patients as they were randomized (ITT), keeping for analysis all patients whose follow-up was known and making estimations according to the “best scenario” (no patients with disease recurrence or progression) and the “worst scenario” (all patients with recurrence and progression) or carrying forward the last observation for those lost to follow-up. Additionally, the study does not comply with the CONSORT statement, the shared standard to report RCTs [8], as often happens with RCTs published in the urologic literature [9]. Moreover, no registration for the RCT was found at www.clinicaltrials.gov, ISRCTN (http://isrctn.org/), or the Australian New Zealand Clinical Trials Registry (www.actr.org.au), although this may be understandable, considering that patients were enrolled into the study starting in 2001.

Divrik et al's RCT [7] is affected by several methodological biases, and its findings likely overestimate the effect on a second TUR on long-term tumor recurrence and progression, as often happens with poor-quality RCTs evaluating interventions [10]. The study has to be regarded as a missed chance to fill a gap in our knowledge of bladder cancer prognosis.

Despite the still-unknown extent of the advantages for recurrence and progression, a second TUR has to be recommended in patients with T1 and/or high-grade newly diagnosed cancers. The consistent risk of understaging bladder cancer as well as the inefficacy of adjuvant BCG immunotherapy in the presence of residual disease and the major impact on decision making of any tumor recurrence following BCG strongly support the need for a staging procedure that is as accurate as possible. Tailoring the most appropriate treatment—selecting either intravesical chemo- or immunotherapy or early cystectomy for those who have the highest chance to benefit from the least invasive treatment—is the best thing we can do for our patients. In such a context, repeating a TUR is, unfortunately, the only available tool to aid the decision-making process.