Case-control studies of OUD often have a significant caveat: a person cannot become opioid dependent if they are never exposed to opioids. In many case-control studies, controls are defined based solely on the DSM definition. Some controls may therefore have significant genetic risk for OUD but never develop the disorder due to lack of drug exposure. To mitigate this issue, Gelernter et al. analyzed only opioid-exposed individuals in the case-control GWAS. No significant findings were found. However, a subsequent study by Cheng et al. analyzed only opioid-exposed European-Americans in a larger cohort that included samples from the previous GWAS [40*]. When comparing 1290 subjects with opioid dependence to 1768 opioid-exposed controls, a SNP ~110kb downstream of RGMA was found to be significantly associated with dependence (rs12442183, p = 1.3 × 10−8). Analysis of microarray data from the frontal cortex suggested that rs12442183 was an expression quantitative trait locus (eQTL) for the RGMA gene [40*]. The previously identified hits in KCNG2, KCNC1 and APBB2 associated with symptom count were not identified in this study, despite the overlapping sample sets, supporting the idea that limiting controls to only opioid-exposed individuals significantly changes GWAS results [40*].
