Cloninger's Temperament Dimensions and Longitudinal Alcohol Use in Early Midlife: A Northern Finland Birth Cohort 1966 Study

Temperament is theorized to be an important factor contributing to the development of alcohol use disorder, but longitudinal studies on how temperament is related to alcohol use among general population in midlife are scarce. Our aims were to investigate potentially reciprocal associations between temperament and changes in alcohol use from age 31 to 46 using prospective birth cohort data.

Within the Northern Finland Birth Cohort 1966, alcohol use and temperament were studied at ages 31 and 46. Participants (N = 5,274) were classified into moderate users, abstainers and heavy users based on their mean alcohol use (g/d). Additionally, participants were categorized as steady users, reducers, or increasers. Multinomial regression analyses were conducted with Cloninger's Temperament and Character Inventory (TCI) scores as factors influencing alcohol use using moderate and stable users as reference groups. Reciprocity of relations was assessed with cross-lagged structural equation modeling.

Temperament and alcohol use are rather stable in midlife. Novelty seeking (NS) predicted heavy use (OR = 1.4; CI: 1.3 to 1.6 for men, OR = 1.3; CI: 1.1 to 1.5 for women) and increasing use (OR = 1.2; CI: 1.1 to 1.4 for men, OR = 1.1; CI: 1.0 to 1.3 for women), whereas low NS predicted abstaining among women (OR = 0.7; CI: 0.6 to 0.8). High harm avoidance (HA) predicted abstaining (OR = 1.3; CI: 1.1 to 1.5) for men. Low persistence (P) among men predicted both abstaining (OR = 0.9; CI: 0.7 to 0.98) and heavy use (OR = 0.9; CI: 0.8 to 0.98). Among women, low reward dependence (RD) predicted heavy use (OR = 0.8; CI: 0.7 to 0.9). Among TCI scores, only NS predicted increasing use in the cross-lagged models.

Temperament has an impact on alcohol use in midlife. Of the TCI dimensions, only NS seems to predispose to increased alcohol use and problem use throughout life. Additionally, RD among women and P among men are significant factors from a life-course perspective. Our results did not support Cloninger's theory on type I alcoholism, as HA showed no relation to problematic alcohol use in midlife.