Reduced-dose bacillus Calmette-Guérin (BCG) in an era of BCG shortage: real-world experience from a tertiary cancer centre

Abstract

Introduction

The global bacillus Calmette-Guerin (BCG) shortage has prompted the use of alternative treatment strategies for patients with intermediate- and high-risk non-muscle invasive bladder cancer (NMIBC). One such strategy is split-vial dosing which allows several patients to be treated from a single vial of BCG. We evaluated the impact of one-third dose BCG on oncological outcomes in a large cohort of patients with NMIBC treated with adequate BCG in a real-world setting.

Patients and Methods

We performed an institutional review board-approved review of patients with NMIBC treated with adequate BCG at our institution between 2000 and 2019. BCG dose was recorded for each patient and patients were stratified according to whether they had received one-third dose (1/3D) BCG or full dose (FD) BCG. Time to recurrence, time to progression and cancer-specific survival were estimated using Kaplan-Meier methods.

Results

Of 563 patients with NMIBC treated with adequate BCG, 150 (26.6%) received 1/3D and 413 (73.4%) received FD. The use of 1/3D BCG did not adversely affect time to recurrence (p=0.449) or time to progression (p=0.716). When stratified by 2021 EAU prognostic factor risk group, 1/3D BCG was not associated with worse time to progression. Cancer-specific survival was similar between patients receiving 1/3D and FD BCG (p=0.320).

Conclusion

The use of 1/3D BCG was not associated with adverse oncological outcomes in a large cohort of patients receiving adequate BCG for intermediate- and high-risk non-muscle invasive bladder cancer. Based on this real world experience, risk-stratified split-vial dosing may represent a valuable approach for other institutions facing BCG shortages whilst also providing reassurance to patients who may be concerned about suboptimal outcomes.

Introduction

Bacillus Calmette-Guérin (BCG) immunotherapy has been the gold-standard treatment for intermediate- and high-risk non-muscle invasive bladder cancer (NMIBC) for almost half a century. Several meta-analyses have demonstrated its superiority over intravesical chemotherapy for recurrence (Bohle 2003, Shelley 2004, Han 2006, Malmstrom 2009) and, BCG also reduces, or at least delays, progression to muscle-invasive disease (Sylvester 2002; Bohle 2004). Current guidelines support a risk-stratified schedule for maintenance therapy, with 1- and up to 3-years of BCG recommended for patients with intermediate- and high-risk disease, respectively (Babjuk 2021, Chang 2016).

In 2011, the United States Food and Drug Administration (US FDA) suspended the production of BCG Connaught following the discovery of mould at the Sanofi Pasteur manufacturing plant (Messing 2017). This event was the catalyst for a series of problems culminating in several periods of global BCG shortage. Shortly following the suspension of BCG Connaught production, Merck, the sole manufacturer and supplier of BCG Tice to the US market, faced its own production difficulties (Messing 2017). Further compounding matters, Sanofi Pasteur subsequently announced that it would indefinitely cease BCG production in mid-2017 (Sanofi Pasteur 2016). As the Connaught strain supplied a significant proportion of the global market, the end of its production has led to a BCG shortage both in the US and worldwide.

These shortages have prompted the use of several alternative treatment strategies including rationing BCG, using intravesical chemotherapy and upfront cystectomy (Mostafid 2015). In accordance with American Urological Association guidelines, our institution elected to treat patients with one-third dose (1/3D) BCG, allowing a single vial of BCG to be divided amongst three patients. This approach is based on a randomized trial from the European Organization for Research and Treatment of Cancer (EORTC) that showed no difference in progression rates between full dose and one-third dose BCG (Oddens 2013). Herein, we evaluate the impact of 1/3D BCG on oncological outcomes in NMIBC patients treated with adequate BCG in a real-world setting.

Patients and Methods

This study was conducted with approval from the University of Texas M.D. Anderson Cancer Center institutional review board. We identified 631 patients between January 2000 and December 2019 with NMIBC (cTa, cTis, cT1) who received ‘adequate’ BCG as defined by the United States Food and Drug Administration (US FDA 2018), International Bladder Cancer Group (Kamat 2016) and European Association of Urology (EAU 2021): at least 5 of 6 induction instillations plus 2 additional instillations (either as part of 3 planned maintenance instillations or 6 planned repeat induction instillations), all taking place within a 6-month period. A total of 68 patients were excluded for the following reasons: primary treatment with BCG plus interferon alpha (n=35), unknown dosing (n=21) and inclusion in trials (n=12).

Patients were only included if pathological information was available from their index TURBT specimen and if they had completed at least one follow-up cystoscopy post-induction BCG at our institution within 6 months from initiation of therapy. A genitourinary pathologist within our institution centrally reviewed all pathology specimens used in clinical treatment decisions. Surveillance schedules were standardized among all providers and based on available NMIBC guidelines.

Medication records were reviewed to determine BCG dosage. The TICE^® BCG strain (Merck, USA) containing 1 to 8 × 10⁸ colony forming units was used. Patients were defined as either FD (50mg) or 1/3D (16.5mg) on the basis of the dosage received to ensure adequate BCG treatment (i.e. at least 5 out of 6 induction instillations plus 2 additional instillations). The preparation of 1/3D was done by dissolving one vial with 150ml of saline and taking 50ml of BCG-Tice. The suspension was instilled into the bladder via a catheter in the absence of contraindications. Following removal of the catheter, the patient was instructed to retain the BCG for 2 hours.

Study variables in our dataset included age, gender, primary versus recurrent tumour, tumour size and number, T stage, the presence of concomitant carcinoma in situ (CIS) and grade. The primary endpoints were time to recurrence (calculated as the number of months from first BCG instillation to the date of biopsy-proven disease recurrence) and time to progression to muscle-invasive bladder cancer or metastasis (calculated as the number of months from first BCG instillation to the date of the first increase to stage T2 or higher). Unrelated deaths and patients without recurrence or progression were censored at the date of death or last follow-up.

The study population was summarized using descriptive statistics. Pearson’s chi-squared tests or Fisher’s exact tests, and Wilcoxon rank-sum tests or t-tests were used to analyze categorical and continuous data, respectively. The Kaplan-Meier method was used to estimate the time to recurrence and progression. Patients were stratified according to BCG dose. Event-free probabilities were stratified separately in patients receiving 1/3D and FD BCG. The log rank test was used to assess the statistical significance of these results. For all analyses a p value <0.05 was considered to indicate statistical significance. All statistical analyses were performed with the Stata/SE, version 16.1 (Stata Corp. LP, College Station, TX).

Results

Five hundred and sixty-three patients were treated with adequate BCG at our institution. Of these, 150 (26.6%) received 1/3D BCG and 413 (73.3%) received FD BCG. Table 1 contains a summary of patient and tumour characteristics. Clinicopathological characteristics were similar between 1/3D and FD groups with three exceptions: 1/3D patients were more likely to have tumours that were solitary (p=0.045), low grade (p=0.010) and not associated with concomitant CIS (p=0.017). However, when patients were stratified according to 2021 EAU prognostic factor risk groups (Babjuk 2021), there was no significant difference in distribution of risk (p=0.063). The median number of BCG doses received by patients in the 1/3D and FD groups was 18 and 21 doses, respectively (p=0.133). Time from index transurethral resection of bladder tumour to first BCG instillation was the same in groups (median 53 days; p=0.263).

Table 1.

Patient and tumour characteristics

	Number (%)
	1/3D (n=150)	FD (n=413)	p-value
Age at diagnosis			0.146
Median (IQR)	69.00 (63.00 – 76.00)	68.00 (61.00 – 75.00)
Gender			0.131
Female	24 (16.0)	90 (21.8)
Male	126 (84.0)	323 (78.2)
Grade			0.010
HG	127 (84.7)	380 (92.0)
LG	23 (15.3)	33 (8.0)
Clinical T stage			0.153
Ta	78 (52.0)	193 (46.7)
Tis	14 (9.3)	26 (6.3)
T1	58 (38.7)	194 (47.0)
Maximum tumour diameter			0.975
<3cm	64 (42.7)	189 (45.8)
≥3cm	71 (47.3)	211 (51.0)
Unknown	15 (10.0)	13 (3.2)
Lymphovascular invastion			0.999
Absent	149 (99.3)	408 (98.8)
Present	1 (0.7)	5 (1.2)
Concomitant CIS			0.017
Absent	124 (83.7)	301 (72.9)
Present	26 (17.3)	112 (27.1)
Focality			0.045
Solitary	87 (58.0)	200 (48.4)
Multifocal	63 (42.0)	213 (51.6)
Prostatic urethra involvement			0.931
Absent	143 (95.3)	393 (95.2)
Present	7 (4.7)	20 (4.8)
Primary vs. recurrent tumour			0.831
Primary	101 (67.3)	282 (68.3)
Recurrent	49 (32.7)	131 (31.3)
Variant histology			0.587
Present	5 (3.3)	18 (4.4)
Absent	145 (96.7)	395 (95.6)
Re-TUR performed			0.448
No	58 (38.7)	145 (35.2)
Yes	92 (61.3)	267 (64.8)
2021 EAU prognostic risk group			0.063
Intermediate	44 (29.7)	97 (23.8)
High	85 (57.4)	225 (55.1)
Very high	19 (12.8)	86 (21.1)
Total BCG doses Median (IQR)	18 (14–24)	21 (15–24)	0.133

Median follow-up for the entire cohort was 54 months (IQR: 29.0–87.5 months). Overall, 198 (35%) of the 563 patients experienced tumour recurrence, of whom 51 (34.0%) received 1/3D and 147 (35.6%) received FD. There was no significant difference in time to recurrence between the groups (p=0.449; Figure 1a). Disease progression to muscle-invasive disease or metastasis was seen in 43 (7.6%) patients, of whom 8 (5.3%) received 1/3D and 35 (8.5%) received FD. Progression to muscle-invasive disease alone occurred in 5 (3.3%) and 22 (5.3%) patients receiving 1/3D and FD, respectively. Distant metastases alone occurred in 3(2.0%) and 13 (3.1%) patients receiving 1/3D and FD, respectively. There was no significant difference in time to progression between the groups (p=0.716; Figure 1b). Subgroup analysis of time to progression stratified by 2021 EAU prognostic factor risk groups showed no significant difference between 1/3D and FD groups (Figure 2): intermediate- (p=0.847), high- (p=0.536) and very-high risk (p=0.603). Overall, 60 radical cystectomies were undertaken. Fifteen (10.0%) patients in the 1/3D group underwent cystectomy; of these, 2 (13.3%) had upstaging at the time of cystectomy. Forty-five (10.9%) patients in the FD group underwent radical cystectomy, of whom 18 (40.0%) had upstaging at cystectomy.

Figure 1. Kaplan-Meier curves for a) time to recurrence, b) time to progression and c) cancer-specific survival stratified by 1/3D vs. FD BCG.

*The last patient in the 1/3D group experienced a progression event after the last censoring

Figure 2. Kaplan-Meier curves time to progression stratified by 2021 EAU prognostic factor risk group: a) intermediate-risk, b) high-risk and c) very-high risk.

*The last patient in the 1/3D group experienced a progression event after the last censoring

Overall, 100/563 (17.8%) patients died. In the 1/3D group, 9 (6.0%) patients died; of these 2 (1.3%) deaths were from bladder cancer. Ninety-one (22%) patients in the FD group died, of whom 18 (4.4%) died from bladder cancer. The five-year cancer-specific Kaplan-Meier survival rate was 98.5% (95% CI: 90.0%-99.8%) in the 1/3D group and 95.7% (95% CI: 92.8%-97.5%) in the FD group. There was no significant difference in cancer-specific survival between groups (p=0.320; Figure 1c).

Discussion

Faced with a worldwide BCG shortage, our institution employed split-vial dosing to treat patients with intermediate- and high-risk NMIBC. In the present study, we show that the use of 1/3D BCG was not associated with inferior oncological outcomes compared to FD in patients receiving adequate BCG treatment. To our knowledge, this paper presents the first real-world experience of using split-vial dosing to mitigate the effects of the BCG shortage.

Our results align with the findings of several randomized controlled trials examining the efficacy of reduced dose BCG compared to full dose (Martinez-Pinero 2002, Martinez-Pinero 2005, Oddens 2013). The CUETO group compared full dose (81mg) BCG with a 3-fold (27mg) dose reduction (Martinez-Pinero 2002). There were no differences in recurrence or progression between groups although, on subgroup analysis, patients with multifocal tumours fared better with the full dose. In a subsequent study, the same CUETO group found that in patients with high-risk cancers (G3/T1/Tis), a 3-fold decreased dose of intravesical BCG was as effective as the standard dose (Martinez-Pinero 2005). The EORTC 30962 study examined the use of full and one-third dosing of BCG with either a 1- or 3-year maintenance course in patients with intermediate- and high-risk NMIBC (Oddens 2013). Full dose BCG was not superior to one-third dose BCG (5-year disease-free rate: 61.7% VS. 58.5%, p=0.092). However, subgroup analysis revealed that in high-risk patients, 3 years of maintenance with full-dose BCG significantly reduced recurrence compared to one-third dose (p=0.009).

The BCG shortage has forced urologists to choose alternative and often suboptimal treatment strategies that deviate from established recommendations. For example, in France, the national medicines agency made the decision to restrict BCG therapy to patients with high-grade disease and to only provide an induction course (Ourfali 2021). Ourfali et al performed a cost-consequence analysis of the medical and financial impact of these restrictions between 2013–2016 (Ourfali 2021). The authors found a 3-fold increase in the rate of tumour recurrence and the rate of cystectomy was five times higher. These results are not unexpected given the body of evidence demonstrating that maintenance treatment is required to reduce the risk of recurrence and progression (Bohle 2003, Sylvester 2002, Malmstrom 2009). The NIMBUS trial evaluated whether a reduced frequency of standard dose BCG was non-inferior to the standard frequency schedule in patients with high-grade NMIBC (Grimm 2020). The reduced frequency schedule was found to be inferior to the standard schedule with regards to time to first recurrence (HR 0.40; 95% CI 0.24–0.68). At our centre, split-vial dosing allowed us to maintain the SWOG 8507 regimen (Lamm 2000) during periods of BCG shortage; patients received an average of 18 and 21 BCG instillations in the RD and FD groups, respectively. Our results support the belief that it is the duration of BCG treatment rather than the dose which is more important in the event of BCG shortage.

We observed a similar proportion of cystectomies in both groups (RD: 10.0% vs. FD: 13.3%), although upstaging at cystectomy was more frequently seen in patients receiving FD than those receiving RD (40.0% vs. 13.3%). However, patients receiving RD BCG had shorter follow-up than those receiving FD and it may be the case that the number of cystectomies in this group increases with time. Interestingly, a recent SEER database analysis examining patterns of radical cystectomy use during the BCG shortage found that cystectomy rates were higher pre-shortage, contrary to the hypothesis that the opposite would occur (Tully 2021).

The present study has limitations inherent in its retrospective nature including unmeasured confounding. Selection bias may explain the equivalence in outcomes observed between RD and FD groups as patients receiving RD BCG were more likely to have tumours that were low-grade, solitary and not associated with concomitant CIS. However, it should be noted that distribution of EAU prognostic factor risk groupings was similar between FD and RD patients (p=0.063). Furthermore, subgroup analysis of each individual risk group did not reveal any differences in time to progression between patients receiving FD and those treated with 1/3D. Nevertheless, this emphasizes the importance of adopting a risk-stratified approach when managing NMIBC patients during periods of shortage. At our institution, patients with higher risk disease were prioritized to receive FD BCG where possible. This practice is by no means unique to our centre. A recent SEER analysis assessing trends in BCG utilization before and during national BCG shortages showed that BCG use was rationed according to clinical risk in the years following interruptions to BCG supply, with the steepest declines occurring in lower risk patients (Khanna 2019). Another limitation pertains to the classification of BCG dosing; patients were defined as either FD or RD on the basis of the dosage received to ensure adequate BCG treatment (i.e. at least 5 induction doses plus 2 additional instillations). This means that instillations received subsequent to this may have been either FD or RD, providing a potential explanation as to why, in contrast to the EORTC trial (Oddens 2013), RD patients did not have inferior time to recurrence compared to FD patients. Finally, this study does not provide any toxicity data. However, this is was never an objective as the association between dose reduction and lower toxicity is well established (Martinez-Pinero 2002, Oddens 2013).

Although Merck have announced plans to construct a new TICE^® BCG manufacturing facility in the United States, BCG supplies are likely to remain in short supply as it will take several years for the facility to become fully operational (AUA 2020). This study provides real-world evidence that 1/3D BCG is not detrimental to patients. As such, risk-stratified split-vial dosing may represent a valuable approach for other institutions facing shortages whilst also providing reassurance to patients who may be concerned about the impact of the shortage on their care. It is hoped that future BCG shortages might be mitigated through the use of different BCG strains, alternative intravesical therapies or novel immunotherapies. A more preferable scenario would see the development of a viable alternative to BCG with equivalent or improved anti-tumour activity, thus completely circumventing the significant difficulties associated with BCG production.

Conclusion

This study provides real world evidence that one-third dosing is not associated with adverse oncological outcomes in patients with intermediate- and high-risk non-muscle-invasive bladder cancer treated with adequate BCG. Risk-stratified split-vial dosing may represent a valuable approach for other institutions facing shortages whilst also providing reassurance to patients who may be concerned about suboptimal outcomes.