Beware the BCG Failures: A Review of One Institution's Results

European Urology European Urology 42 (2002) 542±546 Beware the BCG Failures: A Review of One Institution's Results C. Richard W. Lockyer a,*, James E.C. Sedgwick b, David A. Gillatt a a Bristol Urological Institute, Southmead Hospital, Westbury-on-Trym, Bristol, BS10 5NB, UK b Department of Public Health Sciences, Guy's, King's and St. Thomas' School of Medicine, London SE1 3QD, UK Accepted 26 August 2002 Abstract Objective: To review the outcome of all super cial transitional cell (TCC) bladder cancer treated with intravesical Bacille Calmette-Guerin (BCG) at one institution and, in particular, the prognosis for those patients who gained little bene t from BCG therapy. Patients and Methods: The notes of 122 patients treated with BCG over a nine-year period were reviewed. The following details were recorded: time of diagnosis; time of decision to treat with BCG; results of cystoscopies before and after BCG; duration of follow up; time of progression if occurred, mortality and cause of death. Results: Complete follow up data was available for 112 patients. At a median follow up of 23 months (range 3±107) 57 patients (51%) remained free of tumour, 30 (27%) had progressed and 18 (16%) had died of transitional cell carcinoma. There was a signi cant association between a positive initial check cystoscopy and subsequent progression (p < 0:001) and disease speci c mortality (p < 0:001). Of the 35 patients who had a positive cystoscopy after BCG treatment 21 (60%) progressed and 14 (40%) died of transitional cell carcinoma compared with 9 (12%) and 4 (5%) of the 77 with a negative cystoscopy. Adjusted odds ratios for progression and death from TCC for patients with a positive initial check cystoscopy were 21 and 13, respectively. Conclusion: In our series the patients found to have tumour at the initial check cystoscopy following intravesical BCG had a poor prognosis. This should be remembered when considering treatment options and counselling patients. Follow up of all BCG patients need to be rigorous and protocols would help to unify the treatment patients receive. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Transitional cell carcinoma; Bacille Calmette-Guerin; Recurrence; Disease progression; Prognosis 1. Introduction The majority of bladder tumours are transitional cell carcinomas and are super cial at presentation. Initial transurethral resection usually removes the tumour and allows histological con rmation and determination of the grade and stage. Around 70% of these tumours recur [1], most remain super cial, however some progress. The risk factors for progression are well documented and include high grade; stage T1; multifocality * Corresponding author. Tel. 44-117-959-5690; Fax: 44-117-950-2229. E-mail address: richard@lockyer-uk.freeserve.co.uk (C.R.W. Lockyer). at presentation; the presence of CIS; and size of the initial tumour [1,2]. The success of intravesical Bacille Calmette-Guerin (BCG), as a treatment which aims to reduce recurrences and possibly progression, has been well documented [3]. It has also been acknowledged that intravesical BCG produces side effects in a signi cant number of patients [4] and therefore many clinicians reserve it for super cial bladder cancer with the most potential for progressive disease. However, once this non-invasive therapy is commenced, it is easy to forget the dangers for patients with this subgroup of transitional cell carcinoma. We have reviewed the practice and the outcomes of intravesical BCG at Southmead Hospital since regular 0302-2838/02/$ ± see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S0302-2838(02)00491-8

C.R.W. Lockyer et al. / European Urology 42 (2002) 542±546 543 use commenced in the early 1990s. Of particular interest were those patients for whom BCG was of little therapeutic bene t. 2. Methods Patients treated with BCG initially received 81 mg Connaught strain BCG weekly for 6 consecutive weeks. The BCG was dissolved in 50 ml normal saline and then instilled into the bladder via a urethral catheter and held for 1 hour. Some patients received maintenance therapy in the form of one installation of 81 mg Connaught BCG as above at 3, 6, 12, 18 and 24 months as decided by their specialist. All treatments were carried out in the outpatient department and recorded in a patient log. The notes of the 122 patients treated with BCG since 1991 were retrieved and reviewed by a single author (CRWL). The following details were recorded: time of diagnosis; time of decision to treat with BCG; results of cystoscopies before and after BCG; duration of follow up; time of progression if occurred, mortality and cause of death. The use of urine cytology was not universal in these patients, but the results were noted if present. Patients were grouped according to their histology at the cystoscopy immediately prior to BCG. A patient was de ned as having responded to BCG if the rst check cystoscopy showed no evidence of tumour. If no tissue was taken for histological con rmation, then recurrence was de ned by the cystoscopic ndings. Progression was de ned as disease stage T2 or greater or the need for radical therapy such as cystectomy or radiotherapy. The need for radical therapy for persistent or recurrent disease, in the absence of documented T2 disease, was subjective and depended upon the treating urologist's judgement. The proportion of patients with different outcomes (progression, disease speci c mortality) was compared in different prognostic groups. Logistic regression was performed to examine the relationship between initial response to BCG therapy and outcome, controlling for CIS, stage, grade, multifocality, maintenance therapy, age, sex and duration of follow up. 3. Results Complete follow up data was available for 112 patients. Of the ten without full data, nine sets of notes were unavailable and one patient has not yet completed BCG therapy and undergone initial check cystoscopy. At a median follow up of 23 months (range 1±107), 57 patients (51%) remained free of tumour, 30 (27%) had progressed and 18 (16%) had died of transitional cell carcinoma. Outcomes for each grade and stage are summarised in Table 1. There was no grade available for two and no stage for four of the papillary tumours. Patients were separated into those who responded and those who did not, according to the result of the rst check cystoscopy and the results are shown in Table 2. There was a signi cant association between a positive initial check cystoscopy and subsequent progression (p < 0:001) and disease speci c mortality (p < 0:001). Of the 35 patients who had a positive cystoscopy after BCG treatment 21 (60%) progressed and 14 (40%) died of transitional cell carcinoma compared with nine (12%) and four (5%) of the 77 with a negative cystoscopy. Two years of single shot maintenance intravesical BCG was given to 38 (34%) of the patients at 3, 6, 12, 18 and 24 months, whereas no maintenance was given to 61 (54%). For a variety of reasons including disease progression, patient choice and side effects, 13 (12%) received some maintenance, but not the full 2-year course. In those given maintenance 16% progressed and 8% died compared with 26% and 12%, respectively in those with no maintenance. Only eight patients in our series had a second course of intravesical BCG. When compared with patients with a negative initial check cystoscopy, those with a positive cystoscopy had a relative odds of death from transitional cell carcinoma of 13 (95% con dence interval 3±57), and for progression of 21 (5±84) after controlling for grade, stage, multifocality, age, sex and follow up. Results were similar when time since therapy was used instead of length of follow up, to take account of patients having a shorter follow up because they died. Results of the logistic regression are summarised in Table 3. Table 1 Recurrence, progression and mortality for patients treated with BCG at Southmead Grade/stage (n) Median follow up in months (range) Patients recurrence free(%) Patients with recurrence (%) Patients with progressio Mortality from TCC (%) All (112) 23 (3±107) 57 (51) 55 (49) 30 (27) 18 (16) CIS (57) 23 (3±104) 34 (60) 23 (40) 16 (28) 10 (18) CIS primary (29) 28 (3±104) 21 (72) 8 (28) 5 (17) 4 (14) G3 (21) 16 (5±75) 13 (62) 8 (38) 8 (38) 3 (14) G2 (60) 26 (3±107) 23 (38) 37 (62) 15 (25) 10 (17) T1 (41) 23 (5±107) 21 (51) 20 (49) 12 (29) 7 (17) Ta (38) 20 (3±96) 13 (34) 25 (66) 11 (29) 6 (16) No grade available for two patients and no stage for four patients.

544 C.R.W. Lockyer et al. / European Urology 42 (2002) 542±546 Table 2 Recurrence, progression and mortality in patients with a positive or negative initial cystoscopy following BCG treatment Grade/stage (n) Median follow up in months Initial check cystoscopy Progression (%) Mortality from TCC (%) All (112) 23 Negative 77 (69) 9 (12) 4 (5) Positive 35 (31) 21 (60) 14 (40) CIS (57) 23 Negative 43 (75) 8 (19) 4 (9) Positive 14 (25) 8 (57) 6 (43) CIS primary (29) 28 Negative 24 (83) 2 (8) 1 (4) Positive 5 (17) 3 (60) 3 (60) G3 (21) 16 Negative 16 (76) 4 (25) 1 (6) Positive 5 (24) 4 (80) 2 (40) G2 (60) 26 Negative 37 (62) 3 (8) 2 (5) Positive 23 (38) 12 (52) 8 (35) T1 (41) 23 Negative 29 (71) 4 (14) 2 (7) Positive 12 (29) 8 (67) 5 (42) Ta (38) 20 Negative 22 (58) 2 (9) 1 (5) Positive 16 (42) 8 (50) 5 (31) No grade available for two patients and no stage for four patients. Table 3 Progression and mortality for patients by result of rst check cystoscopy: logistic regression Outcome Negative cystoscopy Positive cystoscopy Odds ratio (95% CI) Adjusted odds ratio (95% CI) a Disease progression 9 (12) 21 (60) 11 (4±30) 21 (5±84) Death from TCC 4 (5) 14 (40) 12 (4±41) 13 (3±57) a Adjusted for grade, stage, multifocality, age, sex and follow up. Of the 30 patients progressing, 13 underwent cystectomy at a median time of 1 year post BCG; 10 were referred for radiotherapy and 7 patients had only palliative or conservative procedures. The histology at cystectomy revealed T2 or greater disease in six patients. 4. Discussion In general BCG is only used at our institution for high grade, recurrent transitional cell carcinomas, whereas other groups have reported results relating to lower grade disease. When compared with these groups, our recurrence and progression rates appear less than satisfactory. However, in this series 51% of patients had evidence of CIS, 45% were either primary CIS or grade three tumours and 49% of the papillary lesions were T1. Malmstrom [5] also treated many high-grade tumours and noted an overall recurrence of 53% and progression in 19% compared to our 49% and 27%, respectively. The aggressive nature of transitional cell carcinoma in patients who require BCG should not be forgotten. Without intravesical therapy 30% of T1 and 45% of grade 3 super cial bladder cancer will progress [6]. Catalona [7] previously reported the risks and bene ts of repeated courses of intravesical BCG. On entering therapy patients had a 7% risk of invasive cancer, 5% risk of metastatic disease and 77% chance of being rendered free of tumour. After failing one course these changed to 11%, 14% and 58%, respectively, and after failing two courses the risk of invasion was 30%, risk of metastatic disease 50% and chances of being clear of tumour were only 20%. We have examined the outcome of patients based on the ndings at the rst cystoscopy after the initial course of BCG. Those with a negative cystoscopy had a relatively good prognosis with 12% progressing and 5% dying of their disease. However, those with a positive cystoscopy were in a perilous position, 60% of them progressed and 40% died of transitional cell carcinoma. These patients were approximately 21 times as likely to progress and 13 times as likely to die of their disease after accounting for confounding factors such as grade, stage and age. The pattern of

C.R.W. Lockyer et al. / European Urology 42 (2002) 542±546 545 progression and mortality was similar for grades 2 and 3 and stages Ta and T1. Merz [8] has previously analysed the early failures of 115 patients, with CIS, treated with intravesical BCG. He de ned early failure as persistent or recurrent disease within 9 months of treatment rather than 3 months as here. Fifteen of 23 (65%) patients with early failure progressed which compares similarly to 21 of 35 (60%) of our own patients. Solsona [9] reported 191 patients with high risk super cial bladder cancer treated with intravesical therapy. Seventy- ve received BCG and 116 received either mitomycin C or doxirubicin. Regardless of treatment, progression occurred in 39% of those with no tumour at the 3 month check cystoscopy, compared with 80% of those with tumour at the 3 month check. He found the median time to progression for those with tumour at 3 months to be less than 1 year. Our results are similar for those with tumour at the rst check, but our progression is much lower for those without tumour at the rst check. This might, in part, be explained by our shorter median follow up of 23 months compared to Solsona's 73 months. Catalona [7] showed a 20% increase in response for patients treated with a second course of BCG. Only eight patients in our series received a second course. A wider use of the second course may have reduced progression and mortality in patients with a positive initial cystoscopy. However, repeated courses of BCG therapy may delay de nitive radical treatment. The instigation of maintenance therapy in these patients was patchy and some of those who started did not have the full course. This has made it dif cult to draw any rm conclusions about the effect of maintenance therapy. However progression and mortality for those with or without maintenance were similar. This concurs with the SWOG group [10] who found that maintenance given as a single dose at 3, 6, 12, 18 and 24 months, as in our unit, shows no bene t over no maintenance; whereas maintenance given as 3 weekly installations at 3, 6, 12, 18, 24, 30 and 36 months confers signi cant bene ts in terms of recurrence and progression. This study was a retrospective audit of one department's results with patients being treated by several different consultants, as is common in most UK units. No protocol existed for enrolling patients on therapy, for giving maintenance or for those who failed to respond. It is obvious that some patients drifted back to regular review cystoscopies (often exible under local anaesthesia) which meant that their high potential for progression could easily have been overlooked. This might have been compounded by an overstretched service, which meant check cystoscopies were often cancelled due to a lack of beds. Obviously, the postponement of check cystoscopies for those patients with a high potential for progression should be avoided. 5. Conclusions The result of the initial check cystoscopy is a very good predictor of the likely prognosis of patients treated with BCG for super cial bladder cancer. Those patients with tumour at the rst check cystoscopy need careful evaluation, and decisions regarding their future management need to be taken early, at a senior level with perhaps a single consultant responsible for all BCG patients. Patients should be followed up rigorously and the use of protocols would help to standardise treatment. If maintenance therapy is felt necessary, then 3 weekly installations at 3, 6, 12, 18, 24, 30, 36 months should be considered. Despite all efforts some patients with this sub group of transitional cell carcinoma will progress and this should be remembered when considering treatment options and counselling patients; particularly those with a positive cystoscopy after the initial BCG course. Acknowledgements The authors are very grateful to Mrs. Pat Sage for her help in retrieving notes and to Dr. Ros Sexton for proof reading the text. References [1] Parmar MK, Freedman LS, Hargreave TB, Tolley DA. Prognostic factors for recurrence and followup policies in the treatment of super cial bladder cancer: report from the British Medical Research Council Subgroup on Super cial Bladder Cancer (Urological Cancer Working Party). J Urol 1989;142(2 Pt 1):284±8. [2] Pagano F, Garbeglio A, Milani C, Bassi P, Pegoraro V. Prognosis of bladder cancer. I. Risk factors in super cial transitional cell carcinoma. Eur Urol 1987;13(3):145±9. [3] Lamm DL. Long-term results of intravesical therapy for super cial bladder cancer. Urol Clin North Am 1992;19(3):573±80. [4] Lamm DL, Stogdill VD, Stogdill BJ, Crispen RG. Complications of bacillus Calmette-Guerin immunotherapy in 1,278 patients with bladder cancer. J Urol 1986;135(2):272±4. [5] Malmstrom PU, Wijkstrom H, Lundholm C, Wester K, Busch C, Norlen BJ. 5-year followup of a randomized prospective study comparing mitomycin C and bacillus Calmette-Guerin in patients

546 C.R.W. Lockyer et al. / European Urology 42 (2002) 542±546 with super cial bladder carcinoma. Swedish-Norwegian Bladder Cancer Study Group. J Urol 1999;161(4):1124±7. [6] Heney NM, Ahmed S, Flanagan MJ, Frable W, Corder MP, Hafermann MD, Hawkins IR. Super cial bladder cancer: progression and recurrence. J Urol 1983;130(6):1083±6. [7] Catalona WJ, Hudson MA, Gillen DP, Andriole GL, Ratliff TL. Risks and bene ts of repeated courses of intravesical bacillus Calmette-Guerin therapy for super cial bladder cancer. J Urol 1987;137(2):220±4. [8] Merz VW, Marth D, Kraft R, Ackermann DK, Zingg EJ, Studer UE. Analysis of early failures after intravesical instillation therapy with bacille Calmette-Guerin for carcinoma in situ of the bladder. Br J Urol 1995;75(2):180±4. [9] Solsona E, Iborra I, Dumont R, Rubio-Briones J, Casanova J, Almenar S. The 3-month clinical response to intravesical therapy as a predictive factor for progression in patients with high risk super cial bladder cancer. J Urol 2000;164(3 Pt 1):685±9. [10] Lamm DL, Blumenstein BA, Crissman JD, Montie JE, Gottesman JE, Lowe BA, et al. Maintenance bacillus Calmette-Guerin immunotherapy for recurrent TA, T1 and carcinoma in situ transitional cell carcinoma of the bladder: a randomized Southwest Oncology Group Study. J Urol 2000;163(4):1124±9.