Pain is a complex human trait. It is likely that the interaction of multiple genes, each with a small individual effect, along with the effect of environmental factors, influences the clinical efficacy of opioids rather than a single gene alone. Polymorphisms in genes coding for the mu-opioid receptor (A118G) and catechol-O-methyl transferase (Val158Met) may be important modulators of opioid efficacy. We assessed joint effects of the OPRM1 and COMT genes in predicting morphine dose for cancer pain relief. We used genotype and clinical data from a pharmacokinetic study of morphine in 207 inpatients treated with stable morphine dose for at least 3 days by Palliative Medicine Specialists. Results showed significant variation in morphine dose requirement by genotype groups: carriers of COMT Val/Val and Val/Met genotype required 63% and 23%, respectively, higher morphine dose compared to carriers of Met/Met genotype (p=0.02). Carriers of OPRM1 GG genotype required 93% higher morphine dose compared to carriers of AA genotypes (p=0.012). When we explored for joint effects, we found that carriers of the OPRM1 AA and COMT Met/Met genotype required the lowest morphine dose to achieve pain relief (87 mg/24 h; 95%CI=57,116) and those with neither Met/Met nor AA genotype needed the highest morphine dose (147 mg/24 h; 95%CI=100,180). The significant joint effects for the Met/Met and AA genotypes (p<0.012) persisted, even after controlling for demographic and clinical variables in the multivariable analyses. Future studies are needed to further characterize the joint effects of multiple genes, along with demographic and clinical variables, in predicting opioid dose.