In this thesis we applied various twin models on a population-based sample of young adult Norwegian twins to investigate genetic and environmental contributions to a selection of dimensional representations of DSM-IV personality disorders (PDs), long-term sick leave (LTSL) and disability pension (DP). We also investigated to what extent LTSL and DP, as well as LTSL and a selection of PDs, share genetic and environmental risk factors in common.
Knowledge of the heritability of DSM-IV cluster C PDs corrected for measurement error has been lacking. In Paper 1, we investigated genetic and environmental contributions to dimensional representations of DSM-IV avoidant and dependent PD, using both a semi-structured interview and a self-report questionnaire conducted at a different time-point. The heritability for both PDs was in the upper range of what has previously been found. No evidence of shared environmental effects or sex differences was found for these PDs. The results further indicated that the interview measure had higher specificity for the genetic liability to these PDs than the questionnaire measure.
Few studies have investigated the heritability of LTSL and DP, and none has used a genetically informative design to investigate the structure of common and specific genetic and environmental contributions to these phenotypes. In Paper 2, we found substantial heritability for LTSL and DP. The genetic and environmental risk factors for LTSL and DP were mainly overlapping, but we also found evidence for a genetic factor of moderate size that was not shared in common between them. The specific genetic factor, as well as extreme scores on the shared genetic factor, may explain why some progress from LTSL to DP. We did not find evidence for sex differences, shared environmental effects or sibling interaction. These results indicate that familial transmission of these phenomena is mainly due to genetic factors.
The association between PDs and LTSL has been largely unexplored, and no studies have investigated the association with a genetically informative design. In Paper 3, we found that dimensional representations of DSM-IV schizotypal, paranoid and borderline PD were uniquely and significantly associated with LTSL. Subsequent twin models showed that the association between these PDs and LTSL was almost entirely due to genetic factors shared in common between the phenotypes. Genetic contributions to the selected PDs accounted for 20% of the heritability of LTSL. The results indicated that the association between PDs and LTSL was non-causal and probably due to genetic confounding, although the design we used was not sufficient for firm conclusions to be drawn on this matter.