Exploring the relationship between alcohol use disorders and internalising disorders in longitudinal cohorts of adolescents

Sammanfattning: The literature supports a relationship between alcohol use disorders and internalising disorders, but how they relate to each other is unclear. Both disorders, as well as their comorbidity increase in adolescence, making adolescence of particular interest. The main objective of the present thesis was to investigate the relationship between alcohol use disorders and internalising disorders in longitudinal cohorts of adolescents. I was interested in the relationship in terms of causality, developmental growth relationships, latent growth trajectories, gender differences, whether the relationship differs with age and whether externalising disorders had an impact on the relationship. To assess these aims, two longitudinal, self-reported, prospective cohorts of adolescents were available for the present thesis. Several models addressing different issues were applied to assess the relationship of alcohol use disorders and internalising disorders. The direction of effect between the phenotypes was assessed using cross-lagged models. Latent growth models were applied in order to investigate the development of each phenotype across time. Parallel growth models of the latent growth curves assessed whether the development of one phenotype was predictive of the other phenotype. Growth mixture models were used in order to establish whether latent trajectory classes of each phenotype were present and whether trajectories of each phenotype were associated with the other phenotype. Finally, externalising symptoms were added as a covariate to the cross-lagged models in order to assess whether externalising symptoms impacted the relationship between alcohol use/problems and internalising symptoms. Additionally, all analysis assessed for gender differences as well as differences between alcohol use and alcohol problems in their relationship to internalising symptoms. Moreover, all analysis was performed on both samples in order to test whether the results were consistent across different age groups, different settings and different geographical locations. Results did indicate a significant relationship between alcohol use disorders and internalising disorder, but the relationship differed by age, gender and sample. However, when growth of the traits was accounted for, a bidirectional relationship was observed. In addition, a weak or a negative relationship was indicated between alcohol use and internalising disorders. Our final most comprehensive model of latent trajectories indicated similar relationships of alcohol use disorders and internalising disorders in both the samples. Evidence for latent trajectories varying with severity was indicated for alcohol use, alcohol use disorders and internalising symptoms. Evidence indicated a significant relationship between alcohol use disorders and internalising disorders. Several different pathways explained the association between alcohol problems and internalising symptoms; there was not one pathway which accounts for the relationship. The two main pathways observed were a model of self-medication (consistent in both samples) and one of a bidirectional relationship, driven by alcohol problems in some individuals and internalising symptoms in other individuals. In a sample of US college students, internalising symptoms and alcohol problems indicated a bidirectional relationship. In a school cohort of mid to late adolescents in Northern Ireland, depression predicted alcohol problems in males while alcohol problems acted as the main driver of a comorbid relationship of the phenotypes in females. Pathway presented tended to depend on factors such as age, gender, whether anxiety was assessed or not, and characteristics of the sample. Finally, externalising symptoms did not significantly account for the relationship between alcohol problems and internalising symptoms but did account for the positive relationships between alcohol use and internalising symptoms and enhanced the negative. Our results provided a possible explanation to the inconsistencies in the literature, either by suggesting that several pathways are present between the two disorders or that the association is so weak that it is hard to pick up on.