Phenobarbital treatment has been observed to be negatively associated
with bladder cancer risk in a few studies. It has been suggested that
phenobarbital may induce drug-metabolizing enzymes that detoxify the b
ladder carcinogens found in cigarette smoke. We examined the relations
hip of barbiturate use to bladder cancer risk and the potential modify
ing effect of cigarette smoking in a large cohort of Kaiser Permanente
Medical Care Program members with computerized pharmacy prescriptions
and smoking information. Newly diagnosed bladder cancers were identif
ied among individuals in the study cohort by linkage with data from ca
ncer registries. The overall standardized incidence ratio associated w
ith barbiturate use was 0.71 [95% confidence interval (CT), 0.51-0.99]
, Among current smokers, former smokers, and never smokers, the standa
rdized incidence ratios were 0.56 (95% CI, 0.23-1.16), 0.68 (95% CI, 0
.27-1.40), and 1.04 (95% CI, 0.48-1.98), respectively. Although our es
timates were imprecise, the finding of an inverse association between
barbiturate treatment and bladder cancer risk only among current and f
ormer cigarette smokers is consistent with the hypothesis that treatme
nt with these medications induces drug-metabolizing enzymes that deact
ivate bladder carcinogens found in cigarette smoke.