Re: A Randomized, Placebo- Controlled Trial of Zoledronic Acid in Patients With

The results of the well-conducted trial by Saad et al. (1) have prompted many to consider routine use of zoledronic acid to reduce the incidence of skeletal-related events in minimally symptomatic or asymptomatic men with hormone-refractory prostate cancer. This is a somewhat troublesome propo-sition, because all men in this situation are not at risk of a skeletal-related event. In the placebo arm of the trial (1), the rate of skeletal-related events was 44.2%, and thus more than half the men participating in the trial never experi-enced a skeletal-related event. Higher rates of fatigue, anemia, myalgia, fever, lower limb edema, dizziness, and weight loss were reported by men who received zoledronic acid than by men who re-ceived placebo. Thus, there is a good rationale for avoiding this treatment in men at little risk for a skeletal-related event. Uni- and multivariable analyses of baseline patient characteristics and treat-ment received are usually included in reports of large, randomized trials. Typ-ically these analyses identify indepen-dent predictors of the primary outcome and adjust the treatment effect for chance imbalances in these predictors, increasing the precision of the treatment effect (2). Compared with men who re-ceived placebo, men who received zoledronic acid at 4 mg had fewer prior skeletal-related events (30.8 % versus 37.5%) and had longer median survival (546 days versus 464 days), suggesting subtle differences between these groups at baseline that may have influenced the observed rates of skeletal-related events (3,4). Even small differences in impor-tant predictors can have a major impact if the variable is important enough (2). Little is known about predictors of skeletal-related events in men with hormone-refractory prostate cancer, but clearly such information could help clinicians identify those patients at highest risk for a skeletal-related event and those most likely to benefit from zoledronic acid therapy (5). We request that the authors provide these data to help clinicians de-termine the most logical use of this drug for their patients