European Urology

The European Association of Urology (EAU) guidelines on non–muscle-invasive urothelial carcinoma of the bladder provide recommendations for adjuvant treatment after the complete transurethral resection (TUR) of all papillary bladder tumours [1]. After an immediate instillation of chemotherapy in all patients, different treatment strategies are proposed, based on a patient's risk of recurrence and progression to muscle-invasive disease [1] and [2]:

• • In patients at low risk of tumour recurrence and progression, no further instillations are recommended prior to a subsequent recurrence.
• • In patients at high risk of tumour progression, intravesical bacillus Calmette-Guérin (BCG) for 1–3 yr is recommended.
• • In the remaining intermediate-risk patients (ie, patients at an intermediate or high risk of recurrence and an intermediate risk of progression), there is no consensus concerning whether further instillations of chemotherapy or 1–3 yr of BCG should be given; both options are foreseen.

Previous studies and meta-analyses have shown that both intravesical chemotherapy and intravesical BCG reduce the recurrence rate compared with TUR alone [3] and [4] and that BCG with maintenance is more effective than chemotherapy in reducing recurrences [5].

Intravesical chemotherapy has not been shown to reduce the rate of progression to muscle-invasive disease [3]. Although a meta-analysis concluded that BCG reduces the risk of progression to muscle-invasive disease when maintenance is given [6], controversy over the use of maintenance BCG exists. The largest study comparing maintenance to no maintenance employed the somewhat subjective end point of disease worsening instead of actual tumour progression [6] and [7]. Additionally, the meta-analysis did not use individual patient data and the need for maintenance BCG was demonstrated by indirect comparisons (ie, the relative benefit of BCG was greater in the presence than in the absence of maintenance) [6].

The value of BCG relative to chemotherapy is also controversial with regard to preventing or delaying progression in the intermediate-risk patients who have a probability of recurrence of ≥50% but a probability of progression of only about a 10% [2]. A meta-analysis comparing BCG to mitomycin C (MMC) found that BCG with maintenance reduced the rate of progression to muscle-invasive disease; however, this meta-analysis was not based on individual patient data [8]. It included a study that used disease worsening rather than progression as an end point, and it included several nonrandomised studies. In the meta-analysis of progression in all randomised trials, no conclusions could be drawn in the subgroup of trials comparing BCG with MMC [6].

In this issue of European Urology, Järvinen and colleagues present the long-term results of a randomised trial in 89 patients comparing 2 yr of treatment with intravesical BCG or intravesical MMC in frequently recurring stage TaT1 bladder cancer patients without carcinoma in situ [9]. Most patients included in this trial are intermediate risk. This study, which recruited patients from 1984–1987, has a number of methodological limitations, including a small sample size (89 patients), an incomplete central pathology review, and a low concentration of MMC (0.2 mg/ml) by today's standards. It does have an uncommonly long median follow-up of >19 yr in patients who are still alive, and therein lies the study's interest. Sixty-two patients (70%) recurred, with a statistically significant difference in favour of BCG for time to first recurrence (p = 0.005). But even with such a long follow-up, only 14 patients (16%) progressed to muscle-invasive disease, 4 of whom (9%) were on BCG and 10 of whom (22%) were on MMC (p = 0.10). A total of 72 patients (81%) died, 36 on each treatment, with 13 deaths (15%) due to bladder cancer (4 [9%] on BCG and 9 [20%] on MMC; p = 0.20). Despite such a long follow-up, the study remains underpowered to detect realistic differences in progression and disease-specific survival.

Table 1 compares the percentages of patients progressing and dying in the present study to the results reported in several meta-analyses. Patients entered in the studies included in these meta-analyses have somewhat heterogeneous patient characteristics; however, many of these patients were intermediate risk, and the percentage of patients progressing in the meta-analyses is similar to what would be expected in intermediate-risk patients. The studies themselves have different but relatively short durations of follow-up. Compared with the meta-analyses, the longer duration of follow-up in the current study has increased by only about 4–7% of the percentage of patients progressing and by 5–8% of the percentage of patients dying due to bladder cancer. The longer follow-up has, in fact, decreased the relative percentage of patients dying due to bladder cancer, since more patients die due to other causes during the extended follow-up.

The relatively low percentage of intermediate-risk patients who progress and die due to bladder cancer makes it very difficult to detect a difference in treatment efficacy for these long-term end points in studies comparing BCG with MMC, should such a difference exist. The size of any potential differences may be diminished by heterogeneous treatment regimens and schedules, including whether maintenance BCG is given or not, short durations of follow-up, the possibility of crossover from MMC to BCG prior to progression, and the decision to carry out a cystectomy prior to progression.

Simply increasing the number of patients studied will not remove these methodological limitations. In a recent individual patient data meta-analysis of nine trials comparing BCG to MMC (74% of the patients were intermediate risk), the conclusions were similar to those in the current study. Although maintenance BCG delayed recurrence as compared with MMC, only nonsignificant trends in favour of BCG were observed for progression and for both overall and bladder cancer–specific survival in the 1880 patients for whom these data were available [10]. Although many studies have been criticised for their short follow-up, the current study has shown that long-term follow-up >10 yr is unlikely to yield many additional progressions or deaths due to disease in intermediate-risk patients [9].

Because BCG reduces the recurrence rate as compared with chemotherapy, patients may be spared the psychological stress associated with a recurrence and may have better quality of life; however, the increased toxicity of BCG as compared with chemotherapy must also be taken into account.

Based on currently available data, one cannot definitively conclude that BCG improves the long-term progression and disease-specific survival rates as compared with chemotherapy in intermediate-risk patients. The absolute differences that have been reported are small. This finding is not surprising; although the 4863 patient meta-analysis reported a relative reduction of 27% in the overall odds of progression when BCG was used, there was an absolute reduction of just 4% because the overall progression rate was only 12% [6].

Given the very large number of patients that would be required to show what appears to be, at most, a small difference in progression and disease-specific survival, it seems unlikely that any new large randomised trials comparing BCG with chemotherapy will be undertaken in intermediate-risk patients or that an update of existing studies will provide significant new information. Recurrence and safety are more realistic end points for studies in intermediate-risk patients. Additionally, recurrence rates, and perhaps even progression and disease-related death rates, may be even lower in current practice as compared with published figures because of recent changes in day-to-day clinical practice: the use of fluorescent cystoscopy and TUR of bladder tumour (TURBT), re-TURBT, an immediate instillation of chemotherapy after TURBT, and optimisation of the delivery of MMC.

The advantages and disadvantages of the two treatment options, chemotherapy and BCG, need to be considered for each individual patient on a case-by-case basis, taking into account the patient's overall condition; previous treatment; history of recurrences; and risk of progression, which in intermediate-risk patients is about 10%. In primary patients, one could start with MMC and later switch to BCG if the patient continues to recur on chemotherapy. In patients with a history of recurrences on chemotherapy, BCG could be offered, especially if there is a worsening of stage or grade. Unfortunately, we cannot accurately predict which of the intermediate-risk patients will progress. Molecular markers represent a hope for the future, but we are not there yet.

Because the results of previous trials and meta-analyses comparing intravesical chemotherapy to intravesical BCG have not yielded clear conclusions for long-term end points, the EAU guidelines continue to recommend either BCG with 1–3 yr of maintenance or a maximum of 1 yr of chemotherapy in the intermediate-risk patients, at least until new, more effective treatment modalities become available [1].

Conflicts of interest

The author has nothing to disclose.