Efficacy
A planned interim analysis on 769 patients was reported. The median overall survival based on 220 deaths revealed a positive trend for the sorafenib group (>18 months), compared to the placebo (14.7 months) with a hazard ratio of 0.72. However, this difference did not reach statistical significance as the threshold for this analysis had been set to a level of 0.0005 (Fleming O'Brian design). In April 2005, 216 of 452 patients (48%) on placebo were crossed over to sorafenib after a positive PFS analysis, which may affect overall survival results.r
Median progression-free survival in the sorafenib group doubled against placebo (24 weeks vs 12 weeks, HR 0.44) at 3 months post-randomisation. A statistically significant improvement of patient well being reported by the FACT-G and FACT-KSI-10 scores was seen for sorafenib.r
Escudierr |
Sorafenib
(n=384) |
Placebo
(n=385) |
Hazard ratio p-value |
Median overall survival based on 220 deaths |
greater than 18 months |
15 months |
0.72 (p=0.018)* |
Median progression-free survival |
24 weeks |
12 weeks |
0.44(p<0.000001) |
Progression free at 3 months post-randomisation |
288/384
(75%) |
166/385
(43%) |
NR |
Confirmed partial response (RECIST) |
7% |
0% |
NR |
* this difference did not reach statistical significance as the thresh hold for this analysis had been set to a level of 0.0005 (Fleming O'Brien design)
Kaplan-Meier Analysis of Overall Survival and Progression-free Survival:r
Panel A: overall survival among 903 patients in May 2005 when patients receiving placebo were allowed to switch to sorafenib
Panel B: shows the probability of overall survival among the same patients in November 2005
Panel C: shows the probability of progression-free survival among 769 patients in January 2005 (P<0.001)
Panel D: shows the probability of progression-free survival among all 903 patients, according to a review by investigators in May 2005 (P<0.001).
© New England Journal of Medicine 2007
In the final analyses, the OS of patients receiving sorafenib was comparable with that of patients receiving placebo (17.8 vs 15.2 months respectively; HR=0.88; p=0.146); however, when post-cross-over placebo survival data were censored, the difference became significant (17.8 vs 14.3 months respectively; HR=0.78; p=0.029).r