Relationship between abdominal pain subgroups in the community and psychiatric diagnosis and personality - A birth cohort study

Introduction: It is unclear if there is a causal link between psychiatric disorders and unexplained chronic gastrointestinal (GI) symptomatology. The role of personality is also in dispute. We aimed to assess the association of these factors with functional GI symptoms in a birth cohort study. Methods: The Dunedin birth cohort is well characterised and has been followed-up prospectively to age 26 (n=980). Measured were upper and lower GI symptoms over the prior year at age 26 using a validated questionnaire, psychiatric diagnoses at ages 18 and 21 by standardised interview applying DSM-III-R criteria, and personality at age 18 using the Multidimensional Personality Questionnaire (MPQ). Natural symptom groupings were identified using factor analysis and k-means clustering. The association of these clusters and psychiatric diagnoses or personality was assessed by logistic regression. Results: The k-means analysis produced a six-cluster solution, which was made up of a health group, and five disease clusters defined by higher than average scores on a single symptom. A diagnosis of depression at age 18 or 21 years was associated with increases in the odds of 1.69 (95% CI: 1.27-2.25) for all GI, of 2.16 (95% Cl: 1.12-4.16) for dysmotility and of 2.07 (95% Cl: 1.13-3.80) for constipation, but not with the other clusters. Similar results were observed with respect to anxiety disorders for the odds of GI overall (OR=1.42, 95% Cl: 1.01-1.99) and constipation (OR=2.11, 95% Cl: 1.17-3.79). The personality subscales were not strongly linked; membership of any diseased cluster was associated with a reduced odds of being in the fourth quartile for the well-being scale (OR=0.64, 95% CI: 0.46-0.88) but increased odds of being in the fourth quartile for the social potency scale (OR=1.64, 95% Cl: 1.18-2.28). Conclusions: In a young adult community sample, unexplained GI symptoms appear to be linked to psychiatric disorders but personality differences were minimal. (C) 2003 Elsevier Inc. All rights reserved.