Assessing the efficacy and safety of sequential intravesical gemcitabine and docetaxel – does time from transurethral resection of bladder tumour to induction matter?

Abstract

Clinical practice guidelines recommend adjuvant intravesical BCG for high-risk and, in many cases, intermediate-risk non-muscle-invasive bladder cancer (NMIBC) [1, 2]. However, BCG induction is generally delayed 2–4 weeks following transurethral resection of bladder tumour (TURBT) to allow bladder re-epithelisation and mitigate the risk of disseminated BCG infection [2-4]. In contrast, early use of intravesical chemotherapy following TURBT carries minimal theoretical and observed risk [1, 5]. Clinical practice guidelines do not explicitly recommend any prescribed waiting period post-TURBT, and many patients will receive an immediate postoperative dose of chemotherapy at the time of TURBT barring any suspected bladder perforation [1, 2].

Sequential intravesical gemcitabine and docetaxel (Gem/Doce), has emerged as a promising alternative to BCG for both high- and intermediate-risk NMIBC [6, 7]. However, the effect of timing between TURBT and Gem/Doce induction on recurrence and treatment-related toxicity has yet to be explored. Thus, we sought to evaluate the timing of induction Gem/Doce after TURBT with regards to efficacy and toxicity.

Following Institutional Review Board approval, we retrospectively reviewed all patients with treatment-naïve, high- or intermediate-risk NMIBC treated with Gem/Doce at the University of Iowa from 2013 to 2024. Inclusion criteria were an intention to complete six weekly Gem/Doce instillations and follow-up surveillance. Prior to induction, patients with visible tumour received complete resection or ablation and perioperative chemotherapy was administered agnostic of tumour grade. Time from last TURBT to first instillation of induction Gem/Doce was stratified by quartile. Risk status was as defined by AUA criteria [1]. A complete induction course was defined as ≥five/six instillations.

The Gem/Doce treatment followed a previously described protocol [7]. Maintenance therapy was initiated for up to 2 years or until unacceptable toxicity or progression. Surveillance was performed as previously described [7]. Patients with significant treatment delays (typically >8 weeks post-TURBT) routinely underwent repeat cystoscopy to identify potential interval recurrences. In this scenario, any additionally needed TURBT effectively reset the interval between resection and induction, thus minimising the potential for bias related to undetected disease at the time of treatment initiation.

A total of 315 patients received Gem/Doce induction: 231 high-risk and 84 intermediate-risk. Induction Gem/Doce occurred at a median (range) of 35 (6–178) days and patients had a median (range) follow-up of 21 (2.1–106) months. Time from TURBT to first dose of induction Gem/Doce was stratified by quartile ([Q1]: 6–25 days, [Q2]: 26–34 days, [Q3]: 35–48 days, [Q4]: 49–178 days). The median (interquartile range [IQR]) number of Gem/Doce doses administered during induction was 6 (6–6) across all quartiles. Regarding maintenance therapy, the number of doses administered was similar across quartiles, with a median (IQR) of 9 (2–21) doses in Q1, 9 (3–18) in Q2, 9 (2–18) in Q3, and 11 (1–21) in Q4. Both receipt of index TURBT at our institution and receipt of perioperative chemotherapy were associated with earlier treatment initiation (both P < 0.01). Otherwise, clinicopathological characteristics were similar across quartiles.

Using the Kaplan–Meier method, there were no significant differences in recurrence-free survival (RFS) when time to Gem/Doce induction was stratified by quartile (Fig. 1). The 2-year RFS (95% CI) for Q1 was 73% (60–82%), for Q2 was 80% (68–88%), for Q3 was 79% (67–88%), and for Q4 was 72% (58–82%). Upon univariate regression analysis, time from TURBT to Gem/Doce induction did not significantly affect RFS outcomes when stratified as quartiles (P = 0.80), with hazard ratios (95% CIs) of 0.75 (0.39–1.44), 1.02 (0.54–1.94), and 0.96 (0.51–1.79) for Q2, Q3, and Q4, respectively (with Q1 as the reference). When analysed as a continuous variable, the hazard ratio was 1.00 (95% CI 1.00–1.01), with P = 0.52. These results were consistent for high-grade RFS among patients with high-risk disease prior to Gem/Doce.

Adverse events (AEs) occurred similarly across quartiles, ranging from 34% in Q2 to 37% in Q1 (P = 0.58). Univariate regression analysis indicated high-risk patient status (odds ratio [OR] 1.78, 95% CI 1.06–3.00; P = 0.03) and earlier treatment year (OR 1.11, 95% CI 1.01–1.22; P = 0.04) were associated with higher odds of developing AEs. Time to induction (P = 0.81), pre-treatment T1-containing (P = 0.13), and perioperative chemotherapy (P = 0.25) were not significant factors. Notably, bladder spasms appeared to be the predominant AE contributing to the higher AE rates in high-risk patients. Since our institution began offering Gem/Doce to high-risk patients earlier in the study period, it is likely that evolving management strategies for bladder spasms at our institution influenced this observation. Upon multivariable analysis, treatment year became non-significant (P = 0.05) while high-risk status remained independently associated with increased AEs (P = 0.04). This could reflect more extensive resections taking place in the high-risk group and subsequently increased bladder irritation during intravesical therapy.

While data supporting an ideal timing of induction therapy following TURBT are limited, clinical guidelines advocate delayed induction to prevent systemic absorption of intravesical agents. For BCG, the European Association of Urology (EAU) guidelines recommend a 2 week delay following TURBT to avoid systemic mycobacterial infection, while the National Comprehensive Cancer Network (NCCN) guidelines suggest a 3–4 week delay [2, 3]. However, a study evaluating the time interval from TURBT to induction BCG found no difference in oncological outcomes or AEs, suggesting that early intervention of BCG may not be as hazardous as previously assumed [3]. Despite these findings, the rare but serious complication of BCG sepsis remains a feared complication of early induction.

Alternatively, modern, high-molecular weight intravesical chemotherapeutics post-TURBT have demonstrated minimal systemic absorption with any absorbed drug existing transiently and clearing rapidly [5]. This pharmacokinetic profile suggests that concerns over the timing of chemotherapy induction post-TURBT may be less warranted. Here, our findings similarly indicate that early induction Gem/Doce did not increase toxicity, supporting early intervention of Gem/Doce as a streamlined approach to NMIBC management.

Conversely, concerns have been raised over delays in intravesical treatment, often stemming from logistical hurdles encountered during the referral process. These delays have long been associated with worse outcomes, as studies have shown increased oncological risks when initiation of therapy is postponed. For instance, prior research demonstrated that delaying BCG or mitomycin C treatment beyond 7 weeks significantly worsened RFS in patients with intermediate-risk NMIBC, although this effect was not observed in high-risk disease [8]. Alternatively, other investigators found that postponing BCG heightened recurrence and progression risks for T1 tumours [4].

Yet, in our study, we found that delaying administration of Gem/Doce after TURBT did not lead to higher recurrence rates, regardless of disease stage or grade. This challenges the prevailing notion that all intravesical therapy delays carry oncological consequences. We theorise that the mechanism of Gem/Doce may differ from other agents—specifically, that the initial gemcitabine instillation exerts an exfoliative effect on the urothelium, enhancing docetaxel penetration. This dual-agent approach could allow for effective treatment even when initiation is postponed. While the quality of TURBT remains an important factor, our findings suggest that the mechanistic action of Gem/Doce may mitigate concerns surrounding delayed induction. Importantly, while our institutional protocol favours early intervention, our results indicate that an adaptable treatment strategy does not compromise oncological efficacy when appropriate re-cystoscopic evaluation is performed as necessary.

There are limitations to our study. The retrospective analysis is inherently prone to selection bias and unaccounted confounding variables that could influence outcomes. TURBTs at our institution are generally performed with blue light, which may allow longer time period before induction due to completeness of resection. AEs were abstracted from patient charts and may not be fully represent patient experiences. A more thorough evaluation of quality of life would require validated patient-reported outcome questionnaires. Additionally, modifications of institutional intravesical instillation procedures likely resulted in fewer patient-reported AEs over time. Lastly, the data reported here come from a single, large tertiary medical centre with rigorous NMIBC protocols, which may not be fully generalisable elsewhere.

In conclusion, the toxicity and efficacy of Gem/Doce were not significantly impacted by timing of treatment induction following TURBT in a treatment-naïve population. With appropriate re-cystoscopic evaluation when necessary, induction therapy initiation may be tailored to patient and logistical needs.

Conceptualisation, Grant M. Henning, Ian M. McElree, Michael A. O’Donnell; Methodology, Michael A. O’Donnell, Vignesh T. Packiam, Grant M. Henning; Software, Ian M. McElree, Sarah L. Mott; Validation, Michael A. O’Donnell, Vignesh T. Packiam; Formal Analysis, Sarah L. Mott; Investigation, Michael A. O’Donnell, Vignesh T. Packiam, Ryan L. Steinberg, Ian M. McElree; Resources, Michael A. O’Donnell, Sarah L. Mott; Data Curation, Ian M. McElree; Writing – Original Draft Preparation, Ian M. McElree, Sarah L. Mott, Grant M. Henning; Writing – Review and Editing, Michael A. O’Donnell, Helen Y. Hougen, Vignesh T. Packiam, Ryan L. Steinberg; Visualisation, Ian M. McElree; Supervision, Michael A. O’Donnell, Grant M. Henning, Vignesh T. Packiam; Project Administration, Michael A. O’Donnell, Vignesh T. Packiam; Funding Acquisition, Michael A. O’Donnell, Sarah L. Mott.

Informed consent was waived due to the retrospective nature of this project. Data included in this study are not publicly available.

Ian M. McElree, Ryan L. Steinberg, Sarah L. Mott, Grant M. Henning, and Helen Y. Hougen have no disclosures. Vignesh T. Packiam has the following consulting disclosures: Valar Labs, Urogen, Photocure, Veracyte, Ferring, Engene. Michael A. O’Donnell has the following consulting disclosures: Abbott, Photocure, Fidia, Merck, Theralase, Urogen.