CA184-043

CA184-043 randomised 799 men with at least one bone metastasis from castration-resistant prostate cancer that had progressed after docetaxel to a single 8 Gy fraction of bone-directed radiotherapy followed by ipilimumab 10 mg/kg or placebo every 3 weeks for up to four doses, with maintenance every 3 months for non-progressors.

Median overall survival was 11.2 months (95 percent confidence interval 9.5 to 12.7) with ipilimumab and 10.0 months (8.3 to 11.0) with placebo, hazard ratio 0.85 (0.72 to 1.00, p=0.053). It missed significance, and the proportional hazards assumption was violated (p=0.0031), which is the interesting part. A piecewise hazard model showed the hazard ratio changed with time: 1.46 (1.10 to 1.95) in the first 5 months, 0.65 (0.50 to 0.85) from 5 to 12 months and 0.60 (0.43 to 0.86) beyond 12 months. Ipilimumab appeared to kill men early and help those who survived the early period.

The toxicity explains the first phase. Grade 3 to 4 immune-related adverse events occurred in 101 of 393 men (26 percent) against 11 of 396 (3 percent), including diarrhoea in 64 (16 percent) against seven (2 percent) and colitis in 18 (5 percent) against none. Four deaths were attributed to study drug toxicity, all with ipilimumab.

The companion trial CA184-095, in chemotherapy-naive men without visceral metastases, also failed to improve overall survival. No prospectively defined subgroup was ever found in which ipilimumab helps men with prostate cancer.