All phenotypes were based on self-report without confirmation from collateral reports. The remission phenotypes in the current study did not account for past relapses or the possibility of future relapses. Remission was based on at least 12 months without symptoms, and thus might be a premature classification given that risk of relapse is reduced after 5 years of remission and continues to decline thereafter (54). The prevalence of non-abstinent remission in probands was low and the absence of significant findings about non-abstinent remission in this study does not preclude the possibility that it may have a familial association in a sample with more non-abstinent remitters. The sample was selected for high familial risk for alcoholism and therefore results are not necessarily generalizable to AUD samples drawn from population-based data, although our sample may be informative for clinical populations.
