M. C. Neale

A population-based twin study in women of smoking initiation and nicotine dependence

BACKGROUNDnThe development of drug dependence requires prior initiation. What is the relationship between the risk factors for initiation and dependence?nnnMETHODSnUsing smoking as a model addiction, we assessed smoking initiation (SI) and nicotine dependence (ND) by personal interview in 1898 female twins from the population-based Virginia Twin Registry. We developed a twin structural equation model that estimates the correlation between the liability to SI and the liability to ND, given SI.nnnRESULTSnThe liabilities to SI and ND were substantially correlated but not identical. Heritable factors played an important aetiological role in SI and in ND. While the majority of genetic risk factors for ND were shared with SI, a distinct set of familial factors, which were probably partly genetic, solely influenced the risk for ND. SI was associated with low levels of education and religiosity, high levels of neuroticism and extroversion and a history of a wide range of psychiatric disorders. ND was associated with low levels of education, extroversion, mastery, and self-esteem, high levels of neuroticism and dependency and a history of mood and alcohol use disorders.nnnCONCLUSIONSnThe aetiological factors that influence SI and ND, while overlapping, are not perfectly correlated. One set of genetic factors plays a significant aetiological role in both SI and ND, while another set of familial factors, probably in part genetic, solely influences ND. Some risk factors for SI and ND impact similarly on both stages, some act at only one stage and others impact differently and even in opposite directions at the two stages. The pathway to substance dependence is complex and involves multiple genetic and environmental risk factors.

Testing structural equation models for twin data using LISREL

Simple genetic models can be fitted to twin data using software packages such as LISREL (Jöreskog and Sörbom, 1986a). After discussion of data preparation and routine checks on possible violation of assumptions of the twin method, we illustrate univariate, bivariate, and multivariate genetic models which can be tested in cross-sectional twin data using LISREL. These include models for cohort or cohabitation effects, genotype x sex interaction, and certain types of genotype x environment interaction and genotype-environment correlation.

Genetic risk factors for major depression in men and women: similar or different heritabilities and same or partly distinct genes?

BACKGROUNDnAlthough women are at consistently greater risk for major depression (MD) than men, it is unclear whether sex modifies the aetiological impact of genetic factors on MD. Is the heritability of MD different in men and women? Do the same genetic risk factors predispose to MD in the two sexes?nnnMETHODSnWe obtained a lifetime history of MD by personal interview on two occasions from 6672 individual twins and 2974 complete twin pairs. Three diagnostic criteria of increasing narrowness were employed: DSM-III-R, DSM-III-R plus impairment and Washington University. To increase power by controlling for unreliability of assessment, we evaluated sex differences on genetic risk for MD using a structural equation measurement model.nnnRESULTSnUsing DSM-III-R criteria, but not the two narrower definitions, heritability of MD was significantly greater in women than in men. In the three diagnostic systems, the genetic correlation in liability to MD in men and women was estimated at between +0.50 and +0.65. These estimates differed significantly from unity for the two broader definitions.nnnCONCLUSIONnUsing broad but not narrower definitions of illness, genetic factors play a greater role in the aetiology of MD in women than in men. The genes that influence risk for MD in the two sexes are correlated but are probably not entirely the same. These results raise the possibility that, in linkage and association studies, the impact of some loci on risk for MD will differ in men and women.

Assortative mating for major psychiatric diagnoses in two population-based samples

BACKGROUNDnPrevious studies on assortment for psychiatric disorders have reported discrepant findings. We aimed to test whether there is a significant association for psychiatric diagnoses, including alcoholism, generalized anxiety disorder, major depressive disorder, panic disorder and phobias between husbands and wives in two population-based samples. We further evaluated whether marital resemblance occurs primarily within or across psychiatric disorders and if assortment for psychopathology is primary or secondary to assortment for correlated variables.nnnMETHODSnA model for mate selection addressed whether the correlation between mates for psychiatric disorders arises from direct assortment (primary homogamy) or through correlation with other variables for which assortment occurs (secondary homogamy) or through cross-variable assortment. The model accounted for within-person co-morbidity as well as across-spouse data.nnnRESULTSnFindings suggested that a moderate degree of assortment exists both within and across psychiatric diagnoses. Only a small amount of the observed marital resemblance for mental illness could be explained by assortment for correlated variables such as age, religious attendance and education. Similar results were obtained for the two samples separately and confirmed in their joint analysis, revealing that the co-morbidity and assortment findings, except for the marital correlation for age, religious attendance and education, replicate across samples.nnnCONCLUSIONSnSignificant but moderate primary assortment exists for psychiatric disorders. The bias in twin studies that have ignored the small amount of assortment is negligible.

Dimensional representations of DSM-IV cluster B personality disorders in a population-based sample of Norwegian twins: a multivariate study

BACKGROUNDnThe personality disorders (PDs) in the dramatic cluster B [antisocial (ASPD), histrionic (HPD), narcissistic (NPD) and borderline (BPD)] demonstrate co-morbidity. However, the degree to which genetic and/or environmental factors influence their co-occurrence is not known and, with the exception of ASPD, the relative impact of genetic and environmental risk factors on liability to the cluster B PDs has not been conclusively established.nnnMETHODnPD traits were assessed in 1386 Norwegian twin pairs between the age of 19 and 35 years using the Structured Interview for DSM-IV Personality Disorders (SIDP-IV). Using the statistical package Mx, multivariate twin models were fitted to dimensional representations of the PDs.nnnRESULTSnThe best-fitting model, which did not include sex or shared family environment effects, included common genetic and environmental factors influencing all four dramatic PD traits, and factors influencing only ASPD and BPD. Heritability was estimated at 38% for ASPD traits, 31% for HPD traits, 24% for NPD traits and 35% for BPD traits. BPD traits had the lowest and ASPD traits the highest disorder-specific genetic variance.nnnCONCLUSIONnThe frequently observed co-morbidity between cluster B PDs results from both common genetic and environmental influences. Etiologically, cluster B has a substructure in which ASPD and BPD are more closely related to each other than to the other cluster B disorders.

Testing hypotheses about direction of causation using cross-sectional family data.

We review the conditions under which cross-sectional family data (e.g., data on twin pairs or adoptees and their adoptive and biological relatives) are informative about direction of causation. When two correlated traits have rather different modes of inheritance (e.g., family resemblance is determined largely by family background for one trait and by genetic factors for the other trait), cross-sectional family data will allow tests of strong unidirectional causal hypotheses (A and B are correlated “because of the causal influence of A on B” versus “because of the causal influence of B on A”) and, under some conditions, also of the hypothesis of reciprocal causation. Possible sources of errors of inference are considered. Power analyses are reported which suggest that multiple indicator variables will be needed to ensure adequate power of rejecting false models in the presence of realistic levels of measurement error. These methods may prove useful in cases where conventional methods to establish causality, by intervention, by prospective study, or by measurement of instrumental variables, are infeasible economically, ethically or practically.

The analysis of parental ratings of children's behavior using LISREL.

A common procedure for assessing childrens behavior is to obtain parental ratings of the child. Since the ratings obtained are a function of both parent and child, disentangling the childs phenotype from that of the rater becomes an important methodological problem. For the analysis of genetic and environmental contributions to childrens behavior, solutions to this are available when multiple raters, e.g., two parents, rate multiple children, e.g., twins. This paper describes and illustrates simple LISREL models for the analysis of parental ratings of childrens behavior. We show how the assumption that mothers and fathers are rating the same behavior in children can be contrasted with the weaker alternative that parents are rating correlated behaviors. Given the stronger assumption, which appears adequate for ratings of childrens internalizing behavior problems, the contribution of rater bias and unreliability may be separated from the shared and nonshared environmental components of variation in a behavior genetic analysis.

Fitting genetic models with LISREL: hypothesis testing.

A brief introduction to the mathematical theory involved in model fitting is provided. The properties of maximum-likelihood estimates are described, and their advantages in fitting structural models are given. Identification of models is considered. Standard errors of parameter estimates are compared with the use of likelihood-ratio (L-R) statistics. For structural modeling, L-R tests are invariant to parameter transformation and give robust tests of significance. Some guidelines for fitting models to data collected from twins are given, with discussion of the relative merits of parsimony and data description.

A Model System for Analysis of Family Resemblance in Extended Kinships of Twins

The “Virginia 30,000” comprise 29,698 subjects from the extended kinships of 5670 twin pairs. Over 80 unique correlations between relatives can be derived from these kinships, comprised of monozygotic (MZ) and dizygotic (DZ) twins and their spouses, parents, siblings, and children. This paper describes the first application of a fairly general model for family resemblance to data from the Virginia 30,000. The model assesses the contributions of additive and dominant genetic effects in the presence of vertical cultural inheritance, phenotypic assortative mating, shared twin and sibling environments, and within-family environment. The genetic and environmental effects can be dependent on sex. Assortment and cultural inheritance may be based either on the phenotype as measured or on a latent trait of which the measured phenotype is an unreliable index. The model was applied to church attendance data from this study. The results show that the contributions of genes, vertical cultural inheritance, and genotype-environment covariance are all important, but their contributions are significantly heterogeneous over sexes. Phenotypic assortative mating has a major impact on family resemblance in church attendance.

Parental treatment and the equal environment assumption in twin studies of psychiatric illness

The validity of the twin method depends on the equal environment assumption (EEA)--that monozygotic (MZ) and dizygotic (DZ) twins are equally correlated in their exposure to environmental factors of aetiological importance for the trait under study. Parents may treat MZ twins more similarly than DZ twins thereby potentially violating the EEA. We tested this hypothesis for four common psychiatric disorders (major depression, generalized anxiety disorder, phobia, and alcoholism) in a population-based sample of female-female twin pairs where analyses indicate sufficient statistical power meaningfully to test the EEA. Mothers and fathers beliefs about their twins zygosity disagreed with assigned zygosity in approximately 20% of cases, often because of what they were told about their twins zygosity at their birth. By structural equation model-fitting, we found no evidence that mothers or fathers perceived zygosity influenced twin resemblance for any of the disorders. Compared to parents of DZ twins, parents of MZ twins were more likely to report that, in rearing their twins, they emphasized their similarities more than their differences. However, by model-fitting, mothers and fathers approach to raising twins had no significant influence on twin resemblance for the four examined psychiatric disorders. These results suggest that the differential treatment of MZ and DZ twins by their parents is unlikely to represent a significant bias in twin studies of these major psychiatric disorders.