Alcohol abstinence is particularly challenging for sufferers of alcohol dependence, and relapse precipitated by craving is common. The identification of biological and psychological risk factors for craving can provide valuable insights to the improvement of therapeutic techniques for relapse prevention. Alcoholism-related phenotypes which had been shown to have strong genetic basis, such as alcohol consumption, personality, and hormone regulation, serve as candidates for such risk factors. The present study aims at exploring (1) whether cortisol, leptin and ghrelin are potential hormonal biomarkers for alcohol craving, (2) the relationships between personality traits, heavy drinking onset, craving, and relapse risk, and (3) the association of candidate gene polymorphisms with alcoholism in an Australian Caucasian population.
Two hundred and twenty alcohol-dependent subjects were recruited from a 5-day alcohol withdrawal management programme, out of which 195 subjects completed the required procedures, including private interview, self-report questionnaires for personality trait assessment, and collection of saliva and fasting blood after alcohol craving measurement in the morning of a day close to the end of the detoxification treatment. Subjects were followed-up for one month after discharge for their alcohol abstinence status. Salivary cortisol, plasma ghrelin and leptin levels were quantified by ELISA. Genotyping was done by PCR-RFLP and high resolution melting (HRM).
The alcohol-dependent subjects were categorized according to gender and heavy drinking onset age by cluster analysis as the latter displayed potential multimodal distribution. Earlier onset group subjects demonstrated higher daily alcohol consumption, frequency of family history of alcohol problems, and frequency of lifetime other substance use compared to the late-onset group subjects.
Both global (measured by Obsessive Compulsive Drinking Scale; OCDS) and state (measured by Alcohol Urge Questionnaire; AUQ) alcohol craving levels were highest in the early-onset group, followed by the mid-onset group, and lowest in the late-onset group, and this trend persisted despite the global craving level dropping significantly during treatment. In smokers, state smoking and alcohol craving correlated positively (Spearman’s ρ = .325, p < 0.001). State foodcraving correlated positively with appetite-regulating hormones acylated ghrelin (Spearman’s ρ = .228, p = 0.057) and unacylated ghrelin (Spearman’s ρ = .202, p = 0.042). In contrast, pretreatment OCDS Obsessionality correlated negatively with these two hormones (acylated ghrelin: Spearman’s ρ = -.245, p = 0.044; unacylated ghrelin: Spearman’s ρ = -.214, p = 0.028). Additionally, in-treatment OCDS Obsessionality was also significantly correlated with acylated ghrelin (Spearman’s ρ = -.298, p = 0.028). Salivary cortisol and plasma leptin did not significantly correlate with any alcohol craving measurements.
Selected scales from NEO PI-R, Temperament and Character Inventory (TCI), and Behavioural Inhibition System/Behavioural Approach System (BIS/BAS) were classified into five broader personality factors: Determination, Punishment Sensitivity, Impulsivity, Sociability and Reward Responsiveness. Alcohol craving levels (state/global, pre-/in-treatment) were significantly correlated with some of these factor scores. One-month relapse risk after discharge was possibly negatively correlated with Sociability (p = 0.086). Subsequent cluster analysis on the alcoholic subjects by the five personality factor scores yielded three distinct clusters: (1) high Impulsivity, Punishment Sensitivity and Reward Responsiveness, with low Determination, (2) high Punishment Sensitivity, with low Sociability, Determination and Reward Responsiveness, and (3) high Determination, Sociability and Reward Responsiveness, with low Punishment Sensitivity and Impulsivity. Among these clusters, Cluster 1 had the highest percentage of early-onset subjects and the highest percentage of heavy drinking onset attribution to endogenous emotional instability and to excitement/fun seeking. Moreover, the longest abstinence ever maintained was significantly different among clusters (Log-rank test p = 0.027): shortest for Cluster 1, followed by Cluster 2, and longest for Cluster 3. However, as Cluster 1 subjects were youngest at recruitment, possible “maturing-out” effect of personality should be considered.
Using a case-control association study approach with additional cases from post-mortem brain tissues and another Brisbane study (total of 529 alcoholic cases and 381 controls), we observed significant differences in allele and/or genotypic frequencies between alcoholics and controls in candidate gene polymorphisms of COMT, RGS4, GABAA subunit gene cluster, SLC6A4, and LEPR, but they did not remain significant after permutation-based correction for multiple comparisons. Significant differences in haplotype frequency were found in RGS4 and GABAA subunit gene cluster SNPs, but the low frequencies of the potential risk haplotype should be noted. Subsequent logistic regression analysis showed ANKK1 TaqIA and RGS4 SNP7 might contribute to earlier drinking onset in females.
To conclude, this study has demonstrated the relationships between alcohol use history, craving, hormones, and personality traits in a cross-sectional sample of in-patient alcoholics. It has also exhibited the variability of alcoholic patients in terms of personality and craving levels, thus providing insights for personal treatment design that could maximize patients’ acceptance and responsivity. Moreover, our results also suggest that the ghrelin signalling pathway may be a plausible pharmacotherapeutic target as it may regulate the appraisal of rewarding values between alcohol and food during alcohol abstinence. Finally, ANKK1 and RGS4 SNPs may be associated with earlier drinking onset in females, but the phenotypes related require further investigations.