The observations of correlations strictly on the phenotypic level require some explanation. The correlation between AD-CrC and MAXDRINKS is expected: the larger the value for MAXDRINKS, the higher risk to have more alcohol dependence criteria. The positive correlation between AD-CrC and flushing is perhaps more surprising as it is in effect opposite to the genetic relationship between these phenotypes. The most likely explanation for this positive correlation is, we believe, simply that drinking more alcohol causes flushing directly in many subjects.
