OBJECTIVES: To assess whether neoadjuvant or concurrent platinum-based chemotherapy improves the survival of patients with locally advanced bladder cancer, and to determine whether there is any evidence that chemotherapy is more or less effective within well-defined subgroups of patients. PATIENTS AND METHODS: A formal meta-analysis (overview) was carried out using updated individual data from 479 patients (301 deaths) from four randomized trials comparing local definitive treatment alone with neoadjuvant or concurrent single-agent cisplatin followed by local definitive treatment. Further summary data were available from a similar randomized trial of cisplatin and doxorubicin in 325 patients (127 deaths). RESULTS: Combined analysis of the individual patient data gave an overall hazard ratio of 1.02 in favour of local therapy alone (P = 0.845, 95% confidence interval = 0.81-1.26), representing a 2% increase in the relative risk of death with the use of chemotherapy. When this analysis was supplemented by data from the only trial for which individual patient information was not available, the hazard ratio was 0.91 in favour of chemotherapy (P = 0.328, 95% confidence interval = 0.75-1.10), representing a 9% reduction in the relative risk of death. The only prognostic factor for which the evidence suggested a differential treatment effect (interaction) across groups was age (chi-square test for trend = 3.833, P = 0.05), with younger age groups (< 60 years) exhibiting a possible effect in favour of chemotherapy. CONCLUSIONS: Despite a meta-analysis of all known randomized trials, there is still insufficient information to obtain a definitive answer to the question of whether neoadjuvant cisplatin-based chemotherapy improves the survival of patients with locally advanced bladder cancer. Such chemotherapy cannot therefore be currently recommended for routine use and any planned clinical trial should include a 'no chemotherapy' control arm.
OBJECTIVES: To assess whether neoadjuvant or concurrent platinum-based chemotherapy improves the survival of patients with locally advanced bladder cancer, and to determine whether there is any evidence that chemotherapy is more or less effective within well-defined subgroups of patients. PATIENTS AND METHODS: A formal meta-analysis (overview) was carried out using updated individual data from 479 patients (301 deaths) from four randomized trials comparing local definitive treatment alone with neoadjuvant or concurrent single-agent cisplatin followed by local definitive treatment. Further summary data were available from a similar randomized trial of cisplatin and doxorubicin in 325 patients (127 deaths). RESULTS: Combined analysis of the individual patient data gave an overall hazard ratio of 1.02 in favour of local therapy alone (P = 0.845, 95% confidence interval = 0.81-1.26), representing a 2% increase in the relative risk of death with the use of chemotherapy. When this analysis was supplemented by data from the only trial for which individual patient information was not available, the hazard ratio was 0.91 in favour of chemotherapy (P = 0.328, 95% confidence interval = 0.75-1.10), representing a 9% reduction in the relative risk of death. The only prognostic factor for which the evidence suggested a differential treatment effect (interaction) across groups was age (chi-square test for trend = 3.833, P = 0.05), with younger age groups (< 60 years) exhibiting a possible effect in favour of chemotherapy. CONCLUSIONS: Despite a meta-analysis of all known randomized trials, there is still insufficient information to obtain a definitive answer to the question of whether neoadjuvant cisplatin-based chemotherapy improves the survival of patients with locally advanced bladder cancer. Such chemotherapy cannot therefore be currently recommended for routine use and any planned clinical trial should include a 'no chemotherapy' control arm.