Carol A Prescott1, Jonathan W Kuhn, Nancy L Pedersen. 1. Department of Psychology, University of Southern California, 3620 S McClintock Ave, SGM 501, Los Angeles, CA 90089-1061, USA. cprescot@usc.edu
Abstract
BACKGROUND: Allgulander et al. (Allgulander C, Nowak J, Rice JP (1991) Acta Psychiatr Scand 83, 12) published twin pair analyses of psychiatric hospitalization for like-sex pairs from the Swedish Twin Registry born 1926-1958. As noted in a subsequent letter (Allgulander C, Nowak J, Rice JP (1992) Acta Psychiatr Scand 86, 421), several features of the original study resulted in under-ascertainment of cases and underestimated heritability, particularly for alcoholism. The present report updates the prior results by using 17 additional years of follow-up, including members of opposite-sex twin pairs, and addressing biases arising from cohort effects and from excluding pairs with unknown zygosity. METHODS: Registry records for 29,602 twin pairs born 1926-1958 were matched against national databases of psychiatric and medical hospitalizations from 1972-2000 to obtain ICD diagnostic codes. Zygosity was known for 10,903 opposite-sex pairs and 15,401 like-sex pairs who participated previously in research. Twin-pair resemblance and genetic and environmental variance proportions were estimated for hospitalization for alcoholism, affective disorders, psychosis, and (in females) anxiety disorders. RESULTS: Hospitalization rates during the ascertainment window were: alcoholism: males = 3.67%, females = 0.94%; affective disorders: males = 1.99%, females = 2.75%; anxiety disorders: males = 0.46%, females = 0.74%; and psychotic disorders: males = 1.70%, females = 1.96%. Twins from like-sex pairs with unknown zygosity had significantly higher prevalences than those with known zygosity. Tetrachoric correlations and heritability estimates were affected by the method used to model unknown zygosity and cohort effects. CONCLUSIONS: Inclusion of additional follow-up information, opposite-sex twin pairs, age-adjustment, and use of current ICD definitions yielded higher heritability estimates for alcoholism, anxiety disorders, and psychosis than previously published for this nationally-representative sample of twins from Sweden. The results show that relatively small selection biases can alter twin study results and underscore the importance of addressing under-ascertainment of cases in genetic research based on volunteers.
BACKGROUND: Allgulander et al. (Allgulander C, Nowak J, Rice JP (1991) Acta Psychiatr Scand 83, 12) published twin pair analyses of psychiatric hospitalization for like-sex pairs from the Swedish Twin Registry born 1926-1958. As noted in a subsequent letter (Allgulander C, Nowak J, Rice JP (1992) Acta Psychiatr Scand 86, 421), several features of the original study resulted in under-ascertainment of cases and underestimated heritability, particularly for alcoholism. The present report updates the prior results by using 17 additional years of follow-up, including members of opposite-sex twin pairs, and addressing biases arising from cohort effects and from excluding pairs with unknown zygosity. METHODS: Registry records for 29,602 twin pairs born 1926-1958 were matched against national databases of psychiatric and medical hospitalizations from 1972-2000 to obtain ICD diagnostic codes. Zygosity was known for 10,903 opposite-sex pairs and 15,401 like-sex pairs who participated previously in research. Twin-pair resemblance and genetic and environmental variance proportions were estimated for hospitalization for alcoholism, affective disorders, psychosis, and (in females) anxiety disorders. RESULTS: Hospitalization rates during the ascertainment window were: alcoholism: males = 3.67%, females = 0.94%; affective disorders: males = 1.99%, females = 2.75%; anxiety disorders: males = 0.46%, females = 0.74%; and psychotic disorders: males = 1.70%, females = 1.96%. Twins from like-sex pairs with unknown zygosity had significantly higher prevalences than those with known zygosity. Tetrachoric correlations and heritability estimates were affected by the method used to model unknown zygosity and cohort effects. CONCLUSIONS: Inclusion of additional follow-up information, opposite-sex twin pairs, age-adjustment, and use of current ICD definitions yielded higher heritability estimates for alcoholism, anxiety disorders, and psychosis than previously published for this nationally-representative sample of twins from Sweden. The results show that relatively small selection biases can alter twin study results and underscore the importance of addressing under-ascertainment of cases in genetic research based on volunteers.
Authors: V V McCutcheon; J D Grant; A C Heath; K K Bucholz; C E Sartor; E C Nelson; P A F Madden; N G Martin Journal: Psychol Med Date: 2012-03-16 Impact factor: 7.723
Authors: Gregory John Lydall; Nicholas J Bass; Andrew McQuillin; Jacob Lawrence; Adebayo Anjorin; Radhika Kandaswamy; Ana Pereira; Irene Guerrini; David Curtis; Anna E Vine; Pamela Sklar; Shaun M Purcell; Hugh Malcolm Douglas Gurling Journal: Psychiatr Genet Date: 2011-12 Impact factor: 2.458