Abstract

Study objective: To determine the association between a clinician assessment of temperament in early adulthood and cause specific mortality.

Design: Prospective observational study.

Setting: Glasgow University.

Participants: 9239 male former students aged 16–30 (mean 20.5) years who participated in an ongoing health survey from 1948–68. A physician recorded free text assessment of temperament, which seemed to capture aspects of personality (trait) and mental health (state), was coded into: stable, anxious, schizoid, hypomanic, odd, depressed, immature, hypochondriacal, unstable, and obsessive. Associations between temperament and mortality were investigated using Cox proportional hazards models.

Main results: There were 878 deaths. Most students—8342 (90.3%)—were assessed as stable, the remaining 897 (9.7%) having at least one, and 103 (1.1%) having more than one, temperament type. The second most common temperament was anxiety, recorded in 520 (5.6%) students. In multivariable analyses, having at least one temperament type was associated with increased all cause and stroke mortality, hazard ratios (95% confidence intervals): 1.23 (1.01 to 1.50) and 1.95 (1.06 to 3.59) respectively, compared with stable students. Students with more than one temperament type had higher risk of death from: all causes, 2.05 (1.36 to 3.09); stroke, 3.26 (1.01 to 10.56); and cancer, 2.90 (1.62 to 5.20). Anxiety was positively associated with all cause and cancer mortality, respective hazard ratios: 1.36 (1.07 to 1.72) and 1.51 (1.04 to 2.20). Men labelled hypomanic had increased cardiovascular mortality risk, 1.90 (1.05 to 3.44).

Conclusions: Markers of early adult psychological distress are associated with increased mortality. Mechanisms underlying these associations require investigation.