[Skip to Content]
[Skip to Content Landing]
November 3, 2004

Androgen Suppression Plus Radiation Therapy for Prostate Cancer

JAMA. 2004;292(17):2084-2085. doi:10.1001/jama.292.17.2084-a

To the Editor: Dr D’Amico et al1 report a statistically significant improvement in survival with the addition of 6 months of androgen suppression therapy to 70-Gy radiation therapy for patients with clinically localized prostate cancer. However, they note a P value of.04 for the log-rank test of overall survival, while the 95% confidence intervals for 5-year survival in the treatment groups overlap widely (80%-95% with androgen suppression and 68%-88% without it). The reason for this discrepancy is not clear. I wonder if the authors are actually reporting a 1-sided P value (2-sided P values are specified in the “Methods” section only for their power calculations), or if the difference reflects the log-rank test’s exaggeration of changes in the tail of the survival curve, where statistical variability is greater due to fewer patients at risk.2 The latter possibility would be suggested if the generalized Wilcoxon test is not statistically significant, since this test adjusts for the number of patients at risk at different times.2 Additionally, the authors do not report the details of their multivariable analysis, including checks for appropriateness of that analysis, so the reader cannot determine whether potentially confounding variables were adequately controlled for.3 This information would be helpful for the proper interpretation of this study.