Studies e.g.,8 have used unassessed controls based on the premise that, for low prevalence disorders, this approach only modestly reduces power.9 However, opioids are among the most highly addictive drugs10,11 with high rates of progression from misuse to dependence12 and thus the main constraint on the prevalence of opioid dependence may be the prevalence of opioid misuse. The extent to which genetic factors contribute to liability at various stages of opioid addiction (e.g., initiation, regular use, and dependence), and are shared between stages, is not well characterized.5 Thus, the use of unassessed, predominately unexposed controls might be problematic for identifying genetic effects expressed after the initiation of opioid misuse. Importantly, significant effects of common SNPs manifesting at intermediate and later stages of addiction would be missed in comparisons to unexposed controls. Similarly, comparisons to assessed, unexposed controls are more useful to examine shared liability for initiation and dependence.
