Efficacy of three BCG strains (Connaught, TICE and RIVM) with or without secondary resection (re-TUR) for intermediate/high-risk non-muscle-invasive bladder cancers: results from a retrospective single-institution cohort analysis

Purpose (I) To evaluate the clinical efficacy of three different BCG strains in patients with intermediate-/high-risk non-muscle-invasive bladder cancer (NMIBC). (II) To determine the importance of performing routine secondary resection (re-TUR) in the setting of BCG maintenance protocol for the three strains. Methods NMIBCs who received an adjuvant induction followed by a maintenance schedule of intravesical immunotherapy with BCG Connaught, TICE and RIVM. Only BCG-naïve and those treated with the same strain over the course of follow-up were included. Cox proportional hazards model was developed according to prognostic factors by the Spanish Urological Oncology Group (CUETO) as well as by adjusting for the implementation of re-TUR. Results n = 422 Ta-T1 patients (Connaught, n = 146; TICE, n = 112 and RIVM, n = 164) with a median (IQR) follow-up of 72 (60–85) were reviewed. Re-TUR was associated with improved recurrence and progression outcomes (HRRFS: 0.63; 95% CI 0.46–0.86; HRPFS: 0.55; 95% CI 0.31–0.86). Adjusting for CUETO risk factors and re-TUR, BGC TICE and RIVM provided longer RFS compared to Connaught (HRTICE: 0.58, 95% CI 0.39–0.86; HRRIVM: 0.61, 95% CI 0.42–0.87) while no differences were identified between strains for PFS and CSS. Sub-analysis of only re-TUR cases (n = 190, 45%) showed TICE the sole to achieve longer RFS compared to both Connaught and RIVM. Conclusion Re-TUR was confirmed to ensure longer RFS and PFS in intermediate-/high-risk NMIBCs but did not influence the relative single BCG strain efficacy. When routinely performing re-TUR followed by a maintenance BCG schedule, TICE was superior to the other strains for RFS outcomes. Supplementary Information The online version contains supplementary material available at 10.1007/s00432-021-03571-0.


Introduction
Non-muscle invasive bladder cancers (NMIBCs) represent a heterogeneous category of tumors associated with high recurrence (30-80%) and progression (25-50%) rates, depending on the risk profile, and leading to cancer death after bladder-sparing treatment within 5 years in about 16-23% of cases (Ferlay et al. (2012); CompeÃÅrat et al. 2015). The highest risk subtype of NMIBC represented by high-grade T1 (HG T1) can reach an almost 40% rate of recurrence and 20% of progression at 5 years, despite adjuvant intravesical therapy with Bacillus Calmette-Guérin (BCG) (Bosch and Alfred 2011). From the initial introduction more than 35 years ago, BCG remains the gold standard organ-sparing option for patients classified as intermediate/high risk according to European Association of Urology (EAU) Guidelines (Babjuk et al. 2020;Witjes et al. 2020). Several series have demonstrated BCG's superiority when compared with transurethral resection of bladder tumor (TURBT) alone or in combination with intravesical chemotherapy (Shelley et al. 2004;Han and Pan 2006;Malmström et al. 2009).
For optimal efficacy, BCG should be given in a maintenance schedule to prevent recurrence in patients with intermediate-/high-risk tumors, and evidence suggests that the three-year maintenance protocol originally described by Donald Lamm (2000) is more effective than 1-year Oddens et al. 2013). Despite this, many patients appropriate for intravesical immunotherapy will not complete the treatment protocol as planned due to side effects, efficacy, compliance and unfortunately shortage of BCG availability. The current worldwide BCG shortage has made the comparison of different strains, and potentially strain substitution, particularly of interest. To date, few studies have been conducted that compare strain efficacy, optimal dose and toxicity in the clinical setting with many limitations including variable maintenance schedules and BCG doses. Regarding induction protocol, a Dutch randomized clinical trial (RCT) (Vegt et al. 1995) observed RIVM to be superior in terms of recurrence-free survival (RFS) compared to TICE, and a second RCT in the same setting demonstrated Connaught to better prevent recurrence when compared to TICE (Rentsch et al. 2014). The largest European cohort of high-risk NMIBCs, published by Witjes et al. (2016), compared the 2 most widely used BCG strains, Connaught and TICE, and clearly revealed the independent importance of maintenance and how with this schedule TICE resulted in better recurrence-free outcomes.
An important frequently omitted variable among the available trials is information regarding re-staging procedures (re-TUR) and their potential influence on survival outcomes. In particular, for NMIBCs, residual tumor rates may vary between 33 and 76% for all cases, including 27-72% and 33-78% for Ta and T1 tumors respectively (Cumberbatch et al. 2018). Also, 7-30% of these patients are understaged after initial TURBT, increasing up to 45% when resection does not include detrusor muscle (Shindo et al. 2014). As a result, both intravesical chemotherapy and BCG do not appear to reliably compensate for inadequate resection and are not recommended to replace secondary resection, especially for the highest risk category of HG T1 (Mack et al. 2001;Herr 2005).
To explore this unresolved topic related to potential efficacy differences among BCG strains, we reviewed our historical cohort of NMIBC patients who were assigned to the most used BCG strains (Connaught, TICE and RIVM). The main objective of our analysis was to verify the existence of survival differences among these three stains, with a secondary goal of analyzing the weighted importance of secondary resection procedures in these patients.

Patients and methods
After Institutional Review Board (IRB) approval, we performed a retrospective cohort study on intermediate-/highrisk BCG-naïve NMIBC patients who received induction followed by maintenance with three different BCG strains. Data were also collected on patients who underwent re-TUR within 2-8 weeks following primary resection.
Each participant enrolled in the study had signed an informed consent before undergoing intravesical BCG therapy according to the European Association of Urology (EAU) and Good Clinical Practice (GCP) Guidelines, and the ethical principles of the latest version of the Declaration of Helsinki.
Patients with muscle-invasive disease (≥ T2), upper tract urothelial cancer (UTUC), non-urothelial carcinoma, previous BCG, incomplete/missing data, or who did not initially receive BCG were excluded. The study design is summarized in Supplementary Fig. 1.
Recurrence was defined as tumor relapse in the bladder or prostatic urethra, regardless of tumor stage. Progression was defined as ≥ T2 tumor relapse in the bladder or prostatic urethra. The cause of death was determined from death certificates and chart review.
Full BCG consisted of induction and 7 maintenance courses. Six-week induction started 2-3 weeks after staging TURBT or re-TUR. Maintenance therapy was 3 weekly instillations every 3 months for the first two schedules and then every 6 months (Lamm 2000). Patients were only included if they were treated with the same BCG strain throughout follow-up. The three BCG strains were: BCG-Connaught (ImmuCyst®, Sanofi Pasteur, France) BCG-TICE (OncoTICE®, MSD, USA) BCG-RIVM (Medac®, D-20354, Germany). BCG choice was made on the basis of availability, price and supply.

Statistical analysis
Pearson chi-square test or Fisher's exact test measured the association between variables. Kruskal-Wallis test measured association among quantitative variables. Kaplan-Meier method tested univariate effect of BCG strain on survival outcomes. Log-rank test assessed subgroup differences adjusted for multiple comparison (Dunn-Sidak) when appropriate. For the three strains, study end points were recurrence-and progression-free survival (RFS, PFS), defined as months for any stage/grade to relapse (RFS) and months for rise to T2 or higher stage (PFS). An additional end point was cancer-specific survival (CSS).
Times to events were calculated by taking the date of starting BCG as time zero. Patients without an event were censored at the last follow-up. Cox proportional hazards multivariable regression analysis adjusted for the number of CUETO prognostic factors in BCG maintenance patients (Fernandez-Gomez et al. 2009). This model adjusted for age (< 60, 60-70 and ≥ 70 years), gender, prior recurrence, tumor number (< 3 vs. ≥ 3 lesions), T category (Ta vs. Tis/ T1), concomitant CIS, and tumor grade (G1/2 vs. G3). Adjustments for re-TUR procedures, and separate analyses in patients who received it or not, compared the benefit in outcomes among the strains. Statistical analysis was performed Stata version 16.1 (Stata Corporation, College Station, TX, USA) with statistical significance set as p < 0.05.  (Fig. 1a).
When adjusting for CUETO predictors, as well as for CUETO plus re-TUR, we did not find a statistically significant difference among the three strains in time to progression, but a trend towards significance was observed for TICE and RIVM compared to Connaught (HR 0.62, 95% CI 0.36-1.1 and 0.65, 95% CI 0.40-1.04 respectively; Supplementary Tables 2b-3b).
In the sub-analysis of re-TUR groups, when looking at only those who received it, no differences were demonstrated among all three stains for time to progression. When Re-TUR was not performed, RIVM was superior to Connaught (HR 0.45, 95% CI 0.26-0.83) (Fig. 2b).

Discussion
The genomic changes over 40-50 years which characterized several sub-strains may have attenuated potential efficacy (Gan et al. 2013). This may have led to imbalances in survival outcomes worldwide, especially in the BCG shortage era. Other trials compare strains (Huang et al. 2017), but study design and implementation of induction therapy alone led to mixed results in determining superiority. Of note, even though RIVM is the third most commonly used BCG strain worldwide (Gsponer et al. 2012), it remains the least studied.
We compared a large sample of patients treated with the three most representative BCG strains. To our knowledge, this is the first series with direct comparison of survival outcomes in these cohorts. Our patients were consistently treated with standardized maintenance. We also determined re-TUR's importance on survival outcomes and the relative effect combined with specific strains.
The literature shows that re-TUR on high-risk and selected intermediate-risk NMIBCs improves outcomes (Krajewski et al. 2020). Eroglu et al. (2020) demonstrated the benefit of re-TUR on RFS and PFS and showed that re-TUR was an independent determinant of overall survival. We also found that re-TUR was independently associated with improvement in both recurrence and progression outcomes.
We found both TICE and RIVM superior to Connaught for prolonging DFS. Witjes et al. showed that maintenance TICE performs better than Connaught in HG T1 patients (Witjes et al. 2016). Connaught has a higher earlier immune response during induction but loses immune response efficacy during maintenance. Rentsch et al. (2014) showed Connaught conferred significantly greater 5-year RFS compared to TICE (p = 0.0108) only with a sole induction course. Mice studies suggest that Connaught induced greater initial immune response than TICE. TICE seems to reach its optimum response over time, and therefore with maintenance shows longer DFS.
We found a similar advantage of RIVM over Connaught and a similar percentage of RFS for RIVM to what has previously been seen (Krajewski et al. 2018). Few other studies have reported RIVM's comparative efficacy in NMIBC treatment. RIVM has been studied in comparison with intravesical chemotherapy or alone, not allowing for comparison (Vegt et al. 1995;Krajewski et al. 2018;Sengiku et al. 2013;Kaisary 1987). RIVM has demonstrated excellent RFS and PFS outcomes and 70% completion rate, suggesting good tolerability (Farah et al. 2014).
PFS was similar to that in the literature (Witjes et al. 2016;Nicolazzo et al. 2017 Oct;Nicolazzo et al. 2019;D'Andrea et al. 2020). No strain in our analysis was independently associated with longer PFS. We did observe a consistent trend towards and advantage of TICE and RIVM for PFS and CSS when compared to Connaught (Supplementary Tables 2b-2c). BCG efficacies were almost overlapping after CUETO risk factors ± re-TUR adjustments, thus precluding definitive conclusions on the effect of re-TUR on strain performance in a real-life setting. But we observed two interesting outcomes regarding TICE and RIVM when our analyses were stratified according to the subgroups who had received re-TUR. TICE was the sole strain to improve RFS in re-TUR patients while RIVM provided longer PFS and CSS in the subgroup of patients who had not received re-TUR (Fig. 2a, b).
These findings corroborate that NMIBCs submitted to re-TUR followed by maintenance might achieve RFS with any of these three strains. But when no re-TUR is performed, our results suggest RIVM provides a potential intrinsic protection against MIBC progression and death from BCa which merits further evaluation. To our knowledge, this is the first series reporting a direct comparison among these strains in a maintenance setting. We cannot point to specific survival advantages, but our data suggest future avenues incorporating re-TUR.
A potential explanation for the consistently better performance of TICE and RIVM compared to Connaught may be better tolerability ). This was seen in our higher drop-off rate in the Connaught and was clearly associated with a median of fewer total instillations.
Our study has limitations, including the limitations of retrospective studies. This was not a randomly allocated head-to-head investigation. There may be selection biases between the groups that we could not account for. There was an imbalance between the groups who underwent re-TUR. Nevertheless, our findings are consistent with previous studies and use long-term follow-up to compare three different strains in the same study. Moreover, the strict eligibility criteria, strict BCG protocol homogeneity, and consistent maintenance allowed us to reliably explore the impact of re-TUR.

Conclusions
Our study showed the RFS benefit of both TICE and RIVM compared to Connaught when administered with a maintenance protocol. We also corroborated the importance of performing routine re-TUR in intermediate-/high-risk NMIBCs. We did not find significant differences between TICE and RIVM for the analyzed survival outcomes. Stratifying our data for re-TUR revealed some benefits of TICE for RFS and RIVM for PFS end points. Future trials are needed on this topic in the BCG shortage era.