As noted above, case-control designs focus on identifying factors for increased disease risk and increase power to identify risk factors by oversampling cases and consequently enriching the sample for ‘risk’ factors. Early genetic studies of addiction frequently contrasted drug-dependent cases and age/sex matched unaffected population controls. However, pathogenesis of substance use disorder (SUD) is a multi-stage process, progressing first through drug use, then abuse, and ultimately psychological and physiological dependence and compulsive pursuit of drugs (addiction), with varying influence of genes and environment on individual variation by stage. Including controls who have not ever used drugs hinders focusing on a specific stage of dependence. Study designs that use controls who report drug exposure but do not satisfy criteria for SUD (addiction) likely deliver greater power to detect allelic associations pertaining to risk for dependence, since it is unclear if controls who have never used would develop dependence if exposure occurred. Consistent with this observation, the recent successes in addiction genetics have been in studies focusing on variation within drug-exposed populations or included controls with significant exposure (Cornelis et al., 2011;
