The present study also showed that pre-treatment blood glucose level significantly correlated with the first 8 weeks of during-treatment PHDD. This finding suggests that pre-treatment blood glucose level could predict PHDD during treatment, and therefore that glucose could be able to influence the alcohol use behavior and the related alcohol consumption. These results were obtained over and above the effects of baseline PHDD, gender, and BMI, all of which represent important independent predictors of both drinking and blood glucose levels. Overall, these findings are consistent with preclinical data demonstrating the role of glucose in alcohol preference, alcohol intake, and alcohol-seeking behaviour in animal models. For example, C57BL mice which are hyperglycaemic show a preference for ethanol (Connelly et al., 1983). A comparison of C57 Bl/6j mice (alcohol preferring) and DBA/2j mice (alcohol avoiding) also showed that after the administration of an identical amount of glucose, the C57 Bl/6j mice have significantly higher levels of plasma glucose than the DBA/2j strain (Goas et al. 1979). Furthermore, a functional relationship between this diabetogenic disturbance and alcohol preference was shown in C57 Bl/6j
