Temperament and character traits and profiles: impact on bipolar disorder risk and onset age

This study investigated how personality (Temperament, Character) influences the occurrence and onset age of Bipolar Disorder (BD). Unlike previous studies, Temperament and Character Inventory (TCI) dimensions as well as profiles were examined regarding BD onset age.

We recruited within 5 years 179 adults with prevalent BD (37.4% males, mean age 48.0 ± 12.0 years) attending a general hospital outpatient clinic, in euthymia during the previous 3 months, and 96 controls (36.5% males, mean age 40.4 ± 12.6 years), a convenience sample of community adults recruited via snowball sampling. All participants completed TCI-140. Associations of personality with diagnosis were investigated using logistic regressions while case-only linear and Cox regressions examined associations of personality with onset age. Analyses included TCI dimensions, continuous and binary (dichotomized utilizing Greek general population medians), and profiles (all possible combinations of Temperament or Character high/low binary dimensions), adjusting for potential confounders.

In logistic regressions, BD risk was associated with higher scores in Novelty Seeking (NS), Harm Avoidance (HA) and Self-Transcendence (ST) and lower in Self-Directedness (SD) and Cooperativeness (CO). In linear and Cox regressions, high NS indicated earlier onset compared to low (B=-4.70 [-8.10, -1.30]; HR = 1.60 [1.15, 2.22]), while high SD indicated delayed onset (B = 5.24 [1.87, 8.62]; HR = 0.57 [0.40, 0.79]). In exploratory profile analyses, Narcissistic (high NS, HA and Reward Dependence [RD]) and Histrionic (high NS, low HA, high RD) Temperament profiles and Cyclothymic (low SD, high CO, high ST) and Melancholic (low SD, CO and ST) Character profiles were associated with earlier BD onset, while the Reliable (low NS, low HA, high RD) Temperament profile and Bossy (high SD, low CO, low ST), Creative (high SD, CO and ST) and Organized (high SD, high CO, low ST) Character profiles were associated with delayed onset. These associations, however, lost statistical significance after correcting for multiple comparisons and should be interpreted cautiously.

This study reaffirmed previous associations of TCI dimensions with BD risk, with two (NS, SD) also significantly relating for the first time to onset age. TCI profiles tentatively provided a more nuanced understanding of BD's onset than dimensions but warrant further investigation in larger samples.