P19.019.C An European pharmacogenomic study of response to opioids in advanced cancer patients identified variants associated with efficacy and toxicity

Background/Objectives: Cancer patients usually receive opioids to control pain but unfortunately, 10-20% of patients do not benefit from treatment and experience side effects. Genetics might explain this interindividual variability in the response to opioids. The aim of this genome-wide association study (GWAS) is to identify new genetic markers of opioid efficacy and toxicity. Methods: European cancer patients receiving morphine, oxycodone, buprenorphine, fentanyl were recruited (n = 2060). Data about efficacy (pain intensity, PI) and toxicity (nausea-vomiting score, NVS) were collected; DNA samples were genotyped using Axiom PMRA arrays. Linear regression between genotypes and NVS or PI were performed, using PLINK software. We also used the REGENIE pipeline, based on a machine learning algorithm, as an alternative method. Sex, age, study, country and opioid were included in the models as covariates. Results: GWAS identified 4 and 7 variants associated with PI and NVS, respectively (P-value < 1.0 × 10-6). REGENIE found 5 variants associated with PI and 33 variants associated with NVS (P-value < 1.0 × 10-7), including rs111539671, intronic variant of S100Z gene, above the genome-wide significance threshold (P-value < 5.0 × 10-8). Conclusions: This is the first GWAS for response to opioids performed in more than 2000 patients, individually genotyped. We did not detect any associations reaching the genome-wide significance threshold for the PI phenotype while, for the toxicity phenotype, we obtained a significant genome-wide association. These preliminary results, requiring further validation, highlight the need of analyzing larger cohorts, with homogeneous efficacy and toxicity data. Grant References: AIRC MFAG 2019-ID.22950 project. Conflict of Interest: None declared