The burden of NMIBC includes high financial costs to the healthcare system, significant risk of recurrence that necessitates life-long invasive surveillance, and uncertainty of possible progression that would prompt future radical operative intervention, especially in the highest-risk patients. Strategies to reduce the risk of recurrence and progression, including intravesical chemotherapy and immunotherapy, have been shown to be effective. [4, 12] However, none of these are without risk of potential significant side effects. In our current study we sought to utilize postoperative CSBI in a fashion similar to MMC, as an immediate, one-time postoperative treatment following surgery. This strategy avoids the toxicity of intravesical chemotherapy, as well as the inconvenience of an overnight hospital stay for prolonged CSBI.
In our cohort, however, post-operative CSBI for two hours was not equivalent to a single dose of perioperative MMC. Given the small numbers of patients in the low risk subgroup, we combined patients from low risk and intermediate risk groups for analysis. In the low and intermediate risk patients, there was a significant improvement in DFS with MMC compared with CSBI. In fact, CSBI performed no better than no perioperative treatment. In the high risk subgroup, a similar trend was observed. In our study the absolute risk reduction of postoperative MMC compared with no treatment at one year was 12.3%, which is similar to what is reported in the literature (11.7%). [4, 13] This benefit of MMC holds true even in our Cox multivariable model.
With respect to the efficacy of CSBI, our data stands in contrast to results published by others, albeit with some important differences in study design. Onishi et al. performed a non-randomized study comparing 18–22 h of post-operative CSBI to a full year of induction and maintenance MMC in patients with European Organization for Research and Treatment of Cancer (EORTC) intermediate risk NMIBC and showed no difference in several outcomes, including recurrence-free rates, time to first recurrence, and frequency of recurrences. [10] In this manuscript, the authors alluded to a planned prospective study that was recently published. [14] In their follow-up study, 227 patients with primary EORTC low- to intermediate-risk (all LG) NMIBC were randomized 1:1 to receive CSBI for 18 h or a single dose of 30 mg of MMC in 30 mL of saline. After a median follow-up of 37 months, 29% of patients experienced a recurrence. Recurrence-free rates at 1, 3, and 5 years were similar between the CSBI and MMC groups on Kaplan-Meier analysis. Subgroup analysis showed no difference when stratified between the low- and intermediate-risk tumors. Adverse events were also compared and the MMC group was found to have significantly higher rates of gross hematuria, irritative bladder symptoms, and dysuria (including retention). While the equivalence of CSBI and MMC demonstrated by Onishi et al. could be explained in part by patient selection (all LG patients), we did not replicate this result even in the low and intermediate risk subgroups of our cohort. One important difference in our protocols is the dose of MMC, which was the standard 40 mg in our study and 30 mg in the study by Onishi et al. The most striking difference between our studies, however, is in the duration of CSBI. We intentionally restricted CSBI to two hours to limit the need for overnight hospital stays. While similarly efficacious to one instillation of MMC, CSBI used by Onishi et al. was titrated over 18 h, and it was not reported how many of these patients required an overnight stay. While the authors debate the cost advantages of saline compared with MMC, we question whether this may be offset by even a small fraction of patients requiring overnight admissions for CSBI. Nevertheless, this data demonstrates that in addition to a standard dose of 40 mg of MMC, duration may be an important component of the efficacy of CSBI in preventing tumor cell re-implantation.
Our results also appear to conflict with the results of a recent meta-analysis utilizing individual patient data from randomized trials comparing immediate intravesical instillation of various chemotherapy agents to TURBT alone or instillation of control solution (saline or water). [5] Upon closer examination, however, we are unable to compare the protocols included as published in the meta-analysis or in the original manuscripts to our brief post-operative irrigation protocol. Of the 13 included studies, the use of post-operative irrigation was only documented as consistently used in four of these studies. Irrigation protocols were not detailed in the meta-analysis and review of the original data could not identify specific protocols. Furthermore, at least one study utilized distilled water for irrigation, which has the theoretical advantage of an osmotic cytotoxic effect but the disadvantages of being hypotonic. Therefore, despite a 21% relative reduction in recurrences found in this meta-analysis with use of post-operative irrigation alone, we can only cautiously compare this result with our data without more detailed information about the irrigation protocols used.
The concept of utilizing irrigation for eradication of residual tumor cells following surgery for cancer is not a new concept, nor is it limited to urology or even endoscopic surgery. Surgeons have traditionally irrigated surgical sites to mechanically wash away debris, dilution of bacterial loads, and as a method of tumor cell lysis, depending on the tonicity of the fluid. A survey in England found that 74% of general surgeons perform intraoperative peritoneal lavage during cancer operations (36% with water, 21% with saline, and 17% with betadine). [15] However, efficacy data on irrigation type is conflicting. Sweitzer et al. designed an experiment in mice to evaluate whether distilled water or sterile saline irrigation could reduce the burden of orthotopically implanted melanoma tumor cells. [16] Unfortunately, they found that neither the mechanical process of irrigation nor the hypotonicity of water reduced the tumor burden. In contrast, Fumito et al. demonstrated the superiority of water irrigation to saline irrigation following laparotomy in a mouse model of colorectal cancer tumor spillage. [17] In head and neck cancer models, both the type of irrigation and type of cancer cell line contributed to efficacy. [18, 19] These and other conflicting data suggest that multiple factors play a role with respect to the eradication of residual tumor burden, potentially related to the microenvironment and tumor cell-specific factors, such as cell adhesion properties and degree of de-differentiation.
The literature does strongly support irrigation following intra-luminal surgery in other surgical fields. For example, Zhou et al. performed a meta-analysis of studies evaluating intra-luminal washout following anterior resection for rectal cancer and concluded that washout leads to reduced rates of local recurrence. [20] In the urologic literature, Moskovitz et al. first postulated in 1987 that intravesical irrigation with distilled water during and after TURBT would lead to fewer recurrences. [21] While several small studies have demonstrated conflicting results regarding the use of water irrigation compared with no perioperative treatment, no studies have compared CSBI to MMC until the aforementioned studies by Onishi et al. [10, 14, 22, 23] Our study is the first to compare a shorter, perioperative duration of CSBI to both MMC and no perioperative treatment, and to evaluate this strategy in a heterogeneous patient population with low, intermediate, and high risk disease.
Our results, however, should be considered within the context of several limitations. Although this was a hypothesis-based study driven by pre-clinical and clinical data, it was not a randomized controlled study, and was limited to the data available in medical records. Furthermore, the study is underpowered and longer term follow-up is required to fully realize the potential differences between treatment groups. It is possible that a larger cohort with longer term follow up could confirm the null hypothesis, suggesting that no difference exists between treatment groups. However, at our institution we are mainly utilizing intravesical gemcitabine based on recently published data that suggests efficacy at a fraction of the cost and with reduced side effects compared with MMC. [24] Consequently, in combination with the current data that suggests inefficacy of 2 h of CSBI, we are unlikely to treat more patients with adjuvant CSBI. Primarily one surgeon (KC) performed CSBI during the study period while most other surgeons in the department utilized either MMC or no additional perioperative therapy. Therefore, referral patterns may have contributed to patient heterogeneity between groups. Despite some baseline differences between treatment groups described in our results, the data remains consistent when controlling for factors such as tumor grade, stage, and recurrence disease, among others, in a multivariable model. A consistent surveillance cystoscopy protocol was not used for all patients and could have helped standardize follow-up and limit detection bias. Finally, we utilized a clinical definition of recurrence that included any suspicious lesion during office cystoscopy that warranted an intervention (usually fulguration), which may have artificially increased our recurrence rates.
Nevertheless, our study comparing perioperative CSBI, perioperative MMC, and no perioperative treatment answers important questions regarding CSBI as prophylaxis following endoscopic resection for NMIBC. While CSBI for two hours postoperatively should not replace current guideline-recommended perioperative MMC, it does appear that longer duration of CSBI may increase its efficacy. [10, 14] Research is needed to determine whether the duration can be reduced to limit the number of additional hospital stays and whether other, novel perioperative instillations may reduce recurrences and limit side effects.