Is intravesical chemotherapy for superficial bladder cancer still justified?


Hartwig Schwaibold, Bristol Urological Institute, Department of Urology, Southmead Hospital, Westbury-on-Trym, Bristol BS10 5NB, UK.


European Organization for Research and Treatment of Cancer


UK Medical Research Council


South-west Oncology Group


transurethral resection


carcinoma in situ.


The use of intravesical chemotherapy in the management of superficial bladder cancer has been controversial in recent years. Lamm et al.[1] reported the apparent failure of intravesical chemotherapeutic agents to influence the long-term course of the disease. A meta-analysis of 22 controlled comparisons of intravesical therapy and surgery combining 3899 patients calculated a 14% decrease in short-term recurrence rates in patients treated with adjuvant intravesical chemotherapy compared with surgery alone [1]. However, that decrease occurred mainly in the first few years after treatment. Long-term results were less encouraging, suggesting that intravesical chemotherapy is unable to decrease the risk of tumour progression, and hypothesizing that intravesical chemotherapy might itself be carcinogenic, counteracting its short-term benefit.

In view of these findings, we review intravesical therapy in superficial bladder cancer, assessing first the widely accepted data before examining the important comparative trials of chemotherapy and administration of BCG.


In 1996 the European Organization for Research and Treatment of Cancer (EORTC) and the UK Medical Research Council (MRC) groups published a meta-analysis of their prospective randomized studies comparing intravesical chemotherapy with surgery alone [2]. In terms of time to recurrence, Kaplan–Meier analysis of the 2535 recruited patients revealed a statistically significant advantage of intravesical chemotherapy instillation compared with transurethral resection (TUR) alone (P < 0.01).

According to published reports, there is no doubt that intravesical chemotherapy has a statistically significant beneficial effect on time to recurrence and recurrence rate of superficial bladder cancer, as noted by the AUA guideline panel on nonmuscle-invasive bladder cancer [3]. However, the most critical endpoint of adjuvant prophylaxis is the increase in tumour-free survival due to reduction in disease recurrence or progression. In these respects, neither the Kaplan–Meier analysis by EORTC/MRC nor the meta-analysis by Lamm et al. showed any statistically significant difference between therapy and control arms [1,2]. To our knowledge, no available data show a statistically significant beneficial influence of intravesical chemotherapy on tumour progression in protocols comparing it with TUR alone [3]. The possible reasons for this will be discussed later.


Following its introduction in 1975, BCG is now used worldwide as a vaccine against tuberculosis and in the management of certain malignancies. Its mechanism of action is still not fully understood, but direct contact of BCG with the urothelium is essential. Fibronectin-binding proteins on the surface of the mycobacteria bind to receptors on the surface of the urothelial cells, initiating intracellular signalling which results in cell-mediated antitumour activity [4]. Lamm et al.[1] analysed six studies comparing surgery alone and combined with BCG immunotherapy. This meta-analysis found a 42% significant difference in recurrence in favour of BCG treatment.


BCG has been shown to be superior to chemotherapy in the treatment of CIS. Rates of complete response to intravesical chemotherapy in the treatment of CIS were 30–70%[5]. In BCG trials [6–13] the mean complete response rate was >70%, with a maximum of >80% (Table 1) [6–13]. However, Ovesen et al.[11] reported a relatively low complete response rate of 52% in 4 years. This appeared to be related to a high proportion of secondary CIS and high-risk patients with increased p53 nuclear accumulation. An associated overall progression rate of 45% was also reported. A prospective randomized trial by the South-west Oncology Group (SWOG) reported an advantage of BCG over doxorubicin [10]. The results of these studies suggest that BCG instillation must be considered the treatment of choice for patients with CIS.

Table 1.  Response to BCG of CIS
ReferenceNo. of patientsRate of complete remission, %Follow-up, months
Brosman [6] 338260
Herr et al.[7] 477218
Kavoussi et al.[8] 327124
Reitsma [9]1537018
Lamm et al.[10] 647039
Ovesen et al.[11] 605248
Takashi et al.[12] 307056
Losa et al.[13] 707174


Published data on direct randomized comparisons of BCG immunotherapy with intravesical chemotherapy (Table 2) are somewhat confusing [10,14–21]. Some studies suggested no difference in recurrence between BCG and chemotherapy [15,17,19]. A prospective randomized EORTC trial showed that mitomycin C reduced tumour progression significantly relative to BCG [20]. Other studies, mainly from the USA and Scandinavia, reported a consistent advantage of BCG over doxorubicin and mitomycin C (Table 2). Reasons for the discrepancies include the different characteristics of the study populations and differences in treatment protocols. Some studies included consecutive patients, many of whom had low-grade primary tumours with a low risk of recurrence [15,17,19,20]; others recruited high-risk patients, including those who already had recurrent disease before entering the studies [10,14,16,18]. Some studies used different intravesical therapy protocols for duration and intensity of instillation in their BCG and chemotherapy arms. For instance, one of the trials compared nine BCG instillations over 4 months with 36 mitomycin C instillations over 2 years (Table 2). These discrepancies heavily influenced the comparison of adjuvant-therapy studies and could seriously affect the results of published meta-analyses.

Table 2.  A summary of the results of prospective randomized studies, and the inclusion criteria and treatment schedules, of BCG vs intravesical chemotherapy in Ta, T1 bladder transitional cell carcinoma, 1990–2001
ReferenceSubstanceN patientsRecurrence rate, %PFollow-up, monthsInclusion criteriaTreatment schedules
  • NS, not significant

  • *

    Reduction in recurrences/100 patient-months by prophylaxis;

  • †Recurrence rate per year;

  • ‡statistically significant advantage of mitomycin referring to progression rate in patients with no CIS;

  • ¶Statistical advantage of BCG arms in terms of time to first recurrence.

Lamm et al.[10]BCG262630.01565≥2 recurrences11×/2years
Doxorubicin 83   16×/1year
Rintala et al.[14]BCG 91−23.3*0.001228≥2 recurrences27×/2years
Mitomycin C −9.01*   27×/2years
Melekos et al.[15]BCG12932.2NS33Ta, T115×/2years
Epirubicin 40.3   15×/2years
Lamm et al[16]BCG37740.30.01730≥2 recurrences17×/1year
Mitomycin C 54.3   17×/1year
Vegt et al.[17]BCG-RIVM43746NS36Ta, T16×/6weeks
BCG-Tice 64NS  6×/6weeks
Mitomycin C 43P < 0.01 vs BCG-Tice  9×/6months
Lundholm et al.[18]BCG25051P < 0.0339≥3 recurrences20×/2years
Mitomycin C 66   20×/2years
Krege et al.[19]BCG21424.5NS20Ta, T1, excluding9×/4months
Mitomycin C 26.7  primary Ta, G136×/2years
Witjes et al.[20]BCG344 0.164NS (for recurrence rate)86Ta, T19×/5months
Mitomycin C  0.1570.006  6×/6weeks
van der Meijden  et al.[21]BCG83735<0.00142intermediate27×/3years
BCG + INH 360.009 and high risk27×/3years
Epirubicin 51   27×/3years

Huncharek and Kupelnick [22], in a recent meta-analytic re-evaluation of prospective randomized trials on intravesical chemotherapy and BCG immunotherapy, contended that previous chemotherapy treatment biased the results of studies of recurrence rate in favour of the BCG arms, as recurrences after previous chemotherapy appear less responsive to drug therapy than those in chemotherapy-naïve patients. Therefore the authors concluded that clinical trials directly comparing intravesical BCG to intravesical chemotherapy should stratify by the presence or absence of previous chemotherapy.

Also important are the empirical doses and regimens of both intravesical chemotherapy and BCG, which will influence the effectiveness of the agents, and the outcome of the studies. For instance, maintenance BCG (which was not used in some studies) has been reported to be more effective than induction therapy only [23]. Similarly, Au et al.[24] concluded that variable and incomplete response to mitomycin C is partly due to inadequate drug delivery. They identified a pharmacologically optimized intravesical mitomycin C treatment with statistically significantly better efficacy, using 40 mg in 20 mL of sterile water, fluid restriction for 8 h before intravesical treatment, eliminating residual urine before instilling mitomycin C, and alkalinization of urine with sodium bicarbonate to stabilize the mitomycin C.


In contrast with intravesical chemotherapy, some controlled BCG studies (Table 3) [10,25,26] described evidence of decreased disease progression. The Memorial Sloan-Kettering Cancer Center study noted a statistically significant prolongation of the interval to progression in 86 high-risk patients, where progression in most cases was determined by the presence or absence of multifocal recurrence. Of the patients in the TUR-alone arm, 42% finally underwent cystectomy, compared with 26% of those treated with BCG. Tumour-specific mortality increased from 19% in patients given BCG to 32% in controls (P = 0.032) [25]. The SWOG study [10] compared BCG with doxorubicin and found a significant influence on progression (defined as increased stage or extent of the disease) in favour of BCG. Similarly, Pagano et al.[26] reported that muscle invasion decreased from 17% to 4% (P < 0.01). All three studies were cited in reviews as giving evidence for the beneficial influence of BCG on tumour progression. However, in 1997, Herr's group published its long-term results and noted an identical progression rate of 53% after TUR and BCG [27]. Pagano et al.[26] studied 133 patients and found a significant reduction in progression rate with BCG (4% vs 17%) compared with the control group of TUR alone. However, the follow-up was only 21 months, which as shown above, is much too short for viable data to be obtained.

Table 3.  The beneficial effect of intravesical BCG instillation on progression of Ta,T1 bladder tumours in prospective randomized studies
ReferenceBCG vsN patientsBenefit for BCG arm, PFollow-up, months
Herr et al.[25]TUR 86Progression-free survival, < 0.0172
Lamm et al.[10]Doxorubicin 131Progression-free survival, 0.0165
Pagano et al.[26]TUR133Progression rate, < 0.00121

The SWOG study [10] was initiated in 1985 and included 262 evaluable patients, i.e. 131 with papillary Ta or T1 disease, and the remainder with CIS. The median progression-free survival was 10.4 months in patients treated with BCG and 22.5 months in the control group. However, a thorough analysis of the original paper reveals that progression-free survival was defined as the time to progression or recurrence, and that the exact number of patients who had tumour progression was not shown. No survival benefit was found for BCG.


More recently, Sylvester et al.[23] conducted a meta-analysis of published results of randomized clinical trials of all intravesical chemotherapeutic agents vs BCG. They concluded that intravesical BCG significantly reduces the risk of progression after TUR in patients with superficial bladder cancer, but only in those who receive maintenance BCG treatment. Published abstracts were included in this meta-analysis, with the overall mean follow-up being only 2.5 years.

In two published meta-analyses comparing intravesical BCG and mitomycin C for superficial bladder cancer (pTa or pT1 of any grade), Bohle et al.[28] suggested that BCG was better than mitomycin C in preventing tumour recurrence and that BCG with maintenance was better for preventing tumour progression [29].

As to the latter, Griffiths and Parmar [30], in an editorial comment, pointed out issues about the methods of the above meta-analyses: The papers by Bohle and Bock [29] shared some studies with that of Sylvester et al.[23], and so should not be regarded as independent evidence. Sylvester et al. did not present a subgroup analysis of BCG maintenance vs mitomycin C; Bohle and Bock included two non-randomized comparisons of BCG and mitomycin C, the exclusion of which would give a statistically insignificant result; and neither meta-analysis was based on individual patient data, which may have enhanced the effect of BCG.

In contrast, the Cochrane review and meta-analysis of six randomized trials of BCG vs mitomycin C including 1527 patients (693 mitomycin C and 834 BCG) by Shelley et al.[31] indicated significantly less tumour recurrence in the BCG arm, but only for high-risk patients, but no difference in terms of disease progression or survival.


The apparent failure of intravesical therapy (irrespective of the substance used) to influence disease progression is surprising in the light of the evident reduction in tumour recurrence brought about by different intravesical agents. Normally it would be considered that tumour progression could not take place without recurrence. However, the beneficial influence of intravesical therapy on recurrence seems to have no effect on tumour progression. As to this apparent contradiction, three major points must be considered:

  • • The risk of progression in many of the trials was relatively low because of the inclusion of unselected combinations of patients with Ta, T1 tumours.
  • • In most comparative studies, almost all patients who had been initially randomized into the control arms left the study after developing their first recurrence and were managed as deemed appropriate by the attending urologist, i.e. with some form of intravesical instillation. Therefore, for progression, long-term analysis of prospective randomized trials compare early vs delayed instillation and not intravesical prophylaxis vs TUR alone.
  • • We are probably confronted with two different entities of ‘superficial recurrent bladder tumours’, i.e. one which is relatively benign, and tends to recur often with no progression, whereas the other group will progress towards muscle-invasive or high-grade disease quite often with the first recurrence.

In conclusion, intravesical prophylaxis decreases the recurrence rate and time to recurrence in patients with superficial bladder cancer. BCG can be considered the ‘gold standard’ for managing CIS, but for papillary Ta, T1 bladder tumours differences between the therapy options seem minimal, and in our view, any long-term influence on disease progression remains to be proven.


None declared.