The positive association that we studied the most was between barbiturates and lung cancer.  It was biologically plausible based on liver tumor promotion demonstrated in experimental animals.  As shown above we could not confirm causality because we could not rule out confounding by cigarette smoking.  Another interesting association was negative and with bladder cancer.  We found some support for a previously proposed hypothesis that barbiturates induce enzymes that detoxify bladder carcinogens in cigarette smoke.