Search

Background Many studies aggregate prescription opioid misuse (POM) and heroin use into a single phenotype, but emerging evidence suggests that their genetic and environmental influences may be partially distinct. Methods In total, 7164 individual twins (84.12% complete pairs; 59.81% female; mean age = 30.58 years) from the Australian Twin Registry reported their lifetime misuse of prescription opioids, stimulants, and sedatives, and lifetime use of heroin, cannabis, cocaine/crack, illicit stimulants, hallucinogens, inhalants, solvents, and dissociatives via telephone interview. Independent pathway models (IPMs) and common pathway models (CPMs) partitioned the variance of drug use phenotypes into general and drug-specific genetic (a), common environmental (c), and unique environmental factors (e). Results An IPM with one general a and one general e factor and a one-factor CPM provided comparable fit to the data. General factors accounted for 55% (a = 14%, e = 41%) and 79% (a = 64%, e = 15%) of the respective variation in POM and heroin use in the IPM, and 25% (a = 12%, c = 8%, e = 5%) and 80% (a = 38%, c = 27%, e = 15%) of the respective variation in POM and heroin use in the CPM. Across both models, POM emerged with substantial drug-specific genetic influence (26–39% of total phenotypic variance; 69–74% of genetic variance); heroin use did not (0% of total phenotypic variance; 0% of genetic variance in both models). Prescription sedative misuse also demonstrated significant drug-specific genetic variance. Conclusions Genetic variation in POM, but not heroin use, is predominantly drug-specific. Misuse of prescription medications that reduce experiences of subjective distress may be partially influenced by sources of genetic variation separate from illicit drug use.

BackgroundDrug classes are grouped based on their chemical and pharmacological properties, but prescription and illicit drugs differ in other important ways. Potential differences in genetic and environmental influences on the (mis)use of prescription and illicit drugs that are subsumed under the same class should be examined. Opioid and stimulant classes contain prescription and illicit forms differentially associated with salient risk factors (common route of administration, legality), making them useful comparators for addressing this etiological issue.MethodsA total of 2410 individual Australian twins [Mage = 31.77 (s.d. = 2.48); 67% women] were interviewed about prescription misuse and illicit use of opioids and stimulants. Univariate and bivariate biometric models partitioned variances and covariances into additive genetic, shared environmental, and unique environmental influences across drug types.ResultsVariation in the propensity to misuse prescription opioids was attributable to genes (41%) and unique environment (59%). Illicit opioid use was attributable to shared (71%) and unique (29%) environment. Prescription stimulant misuse was attributable to genes (79%) and unique environment (21%). Illicit stimulant use was attributable to genes (48%), shared environment (29%), and unique environment (23%). There was evidence for genetic influence common to both stimulant types, but limited evidence for genetic influence common to both opioid types. Bivariate correlations suggested that prescription opioid use may be more genetically similar to prescription stimulant use than to illicit opioid use.ConclusionsPrescription opioid misuse may share little genetic influence with illicit opioid use. Future research may consider avoiding unitary drug classifications, particularly when examining genetic influences.

BackgroundAnorexia nervosa (AN) is a psychiatric disorder with complex etiology, with a significant portion of disease risk imparted by genetics. Traditional genome-wide association studies (GWAS) produce principal evidence for the association of genetic variants with disease. Transcriptomic imputation (TI) allows for the translation of those variants into regulatory mechanisms, which can then be used to assess the functional outcome of genetically regulated gene expression (GReX) in a broader setting through the use of phenome-wide association studies (pheWASs) in large and diverse clinical biobank populations with electronic health record phenotypes.MethodsHere, we applied TI using S-PrediXcan to translate the most recent PGC-ED AN GWAS findings into AN-GReX. For significant genes, we imputed AN-GReX in the Mount Sinai BioMe™ Biobank and performed pheWASs on over 2000 outcomes to test the clinical consequences of aberrant expression of these genes. We performed a secondary analysis to assess the impact of body mass index (BMI) and sex on AN-GReX clinical associations.ResultsOur S-PrediXcan analysis identified 53 genes associated with AN, including what is, to our knowledge, the first-genetic association of AN with the major histocompatibility complex. AN-GReX was associated with autoimmune, metabolic, and gastrointestinal diagnoses in our biobank cohort, as well as measures of cholesterol, medications, substance use, and pain. Additionally, our analyses showed moderation of AN-GReX associations with measures of cholesterol and substance use by BMI, and moderation of AN-GReX associations with celiac disease by sex.ConclusionsOur BMI-stratified results provide potential avenues of functional mechanism for AN-genes to investigate further.

BackgroundDespite abundant research on the potential causal influence of childhood maltreatment (CM) on psychological maladaptation in adulthood, almost none has implemented the discordant twin design as a means of examining the role of such experiences in later disordered gambling (DG) while accounting for genetic and family environmental confounds. The present study implemented such an approach to disentangle the potential causal and familial factors that may account for the association between CM and DG.MethodsParticipants were 3750 twins from the Australian Twin Registry [Mage = 37.60 (s.d. = 2.31); 58% female]. CM and DG were assessed separately via two semi-structured telephone interviews. Random-intercept generalized linear mixed models were fit to the data; zygosity, sex, educational attainment, childhood psychiatric disorder, adult antisocial behavior, and alcohol use disorder (AUD) were included as covariates.ResultsNeither quasi-causal nor familial effects of CM predicted DG after adjusting for covariates. Educational attainment appeared to reduce the risk of DG while AUD appeared to increase risk; evidence also emerged for familial effects of antisocial behavior on DG. Post-hoc analyses revealed a familial effect of CM on antisocial behavior, indicating that the association between CM and DG identified in unadjusted models and in prior studies may be accounted for by genetic and shared family environmental effects of antisociality.ConclusionsThese findings add to the meager literature showing that CM does not exert a causal effect on DG, and present novel evidence that familial effects of antisocial behavior may account for the association between CM and DG identified in extant non-twin research.

BackgroundGambling disorder (GD), recognized in Diagnostic and Statistical Manual of Mental Disorders, Version 5 (DSM-5) as a behavioral addiction, is associated with a range of adverse outcomes. However, there has been little research on the genetic and environmental influences on the development of this disorder. This study reports results from the largest twin study of GD conducted to date.MethodsReplication and combined analyses were based on samples of 3292 (mean age 31.8, born 1972–79) and 4764 (mean age 37.7, born 1964–71) male, female, and unlike-sex twin pairs from the Australian Twin Registry. Univariate biometric twin models estimated the proportion of variation in the latent GD liability that could be attributed to genetic, shared environmental, and unique environmental factors, and whether these differed quantitatively or qualitatively for men and women.ResultsIn the replication study, when using a lower GD threshold, there was evidence for significant genetic (60%; 95% confidence interval (CI) 45–76%) and unique environmental (40%; 95% CI 24–56%), but not shared environmental contributions (0%; 95% CI 0–0%) to GD liability; this did not significantly differ from the original study. In the combined analysis, higher GD thresholds (such as one consistent with DSM-5 GD) and a multiple threshold definitions of GD yielded similar results. There was no evidence for quantitative or qualitative sex differences in the liability for GD.ConclusionsTwin studies of GD are few in number but they tell a remarkably similar story: substantial genetic and unique environmental influences, with no evidence for shared environmental contributions or sex differences in GD liability.

BackgroundThe genetic component of Cannabis Use Disorder may overlap with influences acting more generally on early stages of cannabis use. This paper aims to determine the extent to which genetic influences on the development of cannabis abuse/dependence are correlated with those acting on the opportunity to use cannabis and frequency of use.MethodsA cross-sectional study of 3303 Australian twins, measuring age of onset of cannabis use opportunity, lifetime frequency of cannabis use, and lifetime DSM-IV cannabis abuse/dependence. A trivariate Cholesky decomposition estimated additive genetic (A), shared environment (C) and unique environment (E) contributions to the opportunity to use cannabis, the frequency of cannabis use, cannabis abuse/dependence, and the extent of overlap between genetic and environmental factors associated with each phenotype.ResultsVariance components estimates were A = 0.64 [95% confidence interval (CI) 0.58–0.70] and E = 0.36 (95% CI 0.29–0.42) for age of opportunity to use cannabis, A = 0.74 (95% CI 0.66–0.80) and E = 0.26 (95% CI 0.20–0.34) for cannabis use frequency, and A = 0.78 (95% CI 0.65–0.88) and E = 0.22 (95% CI 0.12–0.35) for cannabis abuse/dependence. Opportunity shares 45% of genetic influences with the frequency of use, and only 17% of additive genetic influences are unique to abuse/dependence from those acting on opportunity and frequency.ConclusionsThere are significant genetic contributions to lifetime cannabis abuse/dependence, but a large proportion of this overlaps with influences acting on opportunity and frequency of use. Individuals without drug use opportunity are uninformative, and studies of drug use disorders must incorporate individual exposure to accurately identify aetiology.

BackgroundVulnerability to depression can be measured in different ways. We here examine how genetic risk factors are inter-related for lifetime major depression (MD), self-report current depressive symptoms and the personality trait Neuroticism.MethodWe obtained data from three population-based adult twin samples (Virginia n = 4672, Australia #1 n = 3598 and Australia #2 n = 1878) to which we fitted a common factor model where risk for ‘broadly defined depression’ was indexed by (i) lifetime MD assessed at personal interview, (ii) depressive symptoms, and (iii) neuroticism. We examined the proportion of genetic risk for MD deriving from the common factor v. specific to MD in each sample and then analyzed them jointly. Structural equation modeling was conducted in Mx.ResultsThe best fit models in all samples included additive genetic and unique environmental effects. The proportion of genetic effects unique to lifetime MD and not shared with the broad depression common factor in the three samples were estimated as 77, 61, and 65%, respectively. A cross-sample mega-analysis model fit well and estimated that 65% of the genetic risk for MD was unique.ConclusionA large proportion of genetic risk factors for lifetime MD was not, in the samples studied, captured by a common factor for broadly defined depression utilizing MD and self-report measures of current depressive symptoms and Neuroticism. The genetic substrate for MD may reflect neurobiological processes underlying the episodic nature of its cognitive, motor and neurovegetative manifestations, which are not well indexed by current depressive symptom and neuroticism.

BackgroundPrior research has documented shared heritable contributions to non-suicidal self-injury (NSSI) and suicidal ideation (SI) as well as NSSI and suicide attempt (SA). In addition, trauma exposure has been implicated in risk for NSSI and suicide. Genetically informative studies are needed to determine common sources of liability to all three self-injurious thoughts and behaviors, and to clarify the nature of their associations with traumatic experiences.MethodsMultivariate biometric modeling was conducted using data from 9526 twins [59% female, mean age = 31.7 years (range 24–42)] from two cohorts of the Australian Twin Registry, some of whom also participated in the Childhood Trauma Study and the Nicotine Addiction Genetics Project.ResultsThe prevalences of high-risk trauma exposure (HRT), NSSI, SI, and SA were 24.4, 5.6, 27.1, and 4.6%, respectively. All phenotypes were moderately to highly correlated. Genetic influences on self-injurious thoughts and behaviors and HRT were significant and highly correlated among men [rG= 0.59, 95% confidence interval (CI) (0.37–0.81)] and women [rG= 0.56 (0.49–0.63)]. Unique environmental influences were modestly correlated in women [rE= 0.23 (0.01–0.45)], suggesting that high-risk trauma may confer some direct risk for self-injurious thoughts and behaviors among females.ConclusionsIndividuals engaging in NSSI are at increased risk for suicide, and common heritable factors contribute to these associations. Preventing trauma exposure may help to mitigate risk for self-harm and suicide, either directly or indirectly via reductions in liability to psychopathology more broadly. In addition, targeting pre-existing vulnerability factors could significantly reduce risk for life-threatening behaviors among those who have experienced trauma.

BackgroundChildhood maltreatment (CM) has consistently been linked with adverse outcomes including substance use disorders and adult sexual revictimization. Adult sexual victimization itself has been linked with psychopathology but has predominately been studied in women. The current investigation examines the impact of CM and co-occurring psychopathology on adult sexual victimization in men and women, replicating findings in three distinct samples.MethodWe investigated the association between continuous CM factor scores and adult sexual victimization in the Childhood Trauma Study (CTS) sample (N = 2564). We also examined the unique relationship between childhood sexual abuse (CSA) and adult sexual victimization while adjusting for co-occurring substance dependence and psychopathology. We replicated these analyses in two additional samples: the Comorbidity and Trauma Study (CATS; N = 1981) and the Australian Twin-Family Study of Alcohol Use Disorders (OZ-ALC; N = 1537).ResultsAnalyses revealed a significant association with CM factor scores and adult sexual victimization for both men and women across all three samples. The CSA factor score was strongly associated with adult sexual victimization after adjusting for substance dependence and psychopathology; higher odds ratios were observed in men (than women) consistently across the three samples.ConclusionsA continuous measure of CSA is independently associated with adult sexual trauma risk across samples in models that included commonly associated substance dependence and psychopathology as covariates. The strength of the association between this CSA measure and adult sexual victimization is higher in magnitude for men than women, pointing to the need for further investigation of sexual victimization in male community samples.

BackgroundVarious studies support the inclusion of cannabis withdrawal in the diagnosis of cannabis use disorder (CUD) in the upcoming DSM-5. The aims of the current study were to (1) estimate the prevalence of DSM-5 cannabis withdrawal (criterion B), (2) estimate the role of genetic and environmental influences on individual differences in cannabis withdrawal and (3) determine the extent to which genetic and environmental influences on cannabis withdrawal overlap with those on DSM-IV-defined abuse/dependence.MethodThe sample included 2276 lifetime cannabis-using adult Australian twins. Cannabis withdrawal was defined in accordance with criterion B of the proposed DSM-5 revisions. Cannabis abuse/dependence was defined as endorsing one or more DSM-IV criteria of abuse or three or more dependence criteria. The classical twin model was used to estimate the genetic and environmental influences on variation in cannabis withdrawal, along with its covariation with abuse/dependence.ResultsOf all the cannabis users, 11.9% met criteria for cannabis withdrawal. Around 50% of between-individual variation in withdrawal could be attributed to additive genetic variation, and the rest of the variation was mostly due to non-shared environmental influences. Importantly, the genetic influences on cannabis withdrawal almost completely (99%) overlapped with those on abuse/dependence.ConclusionsWe have shown that cannabis withdrawal symptoms exist among cannabis users, and that cannabis withdrawal is moderately heritable. Genetic influences on cannabis withdrawal are the same as those affecting abuse/dependence. These results add to the wealth of literature that recommends the addition of cannabis withdrawal to the diagnosis of DSM-5 CUD.

BackgroundFew genetically informative studies have examined the effects of different types of trauma on risk for depression over time. The aim of the present study was to examine the relative contributions over time of assaultive trauma, non-assaultive trauma, and familial effects to risk for depression.MethodHistories of depression and trauma were obtained during structured diagnostic interviews with 5266 (mean age 29.9 years, s.d.=2.4) members of a volunteer Australian twin panel from the general population. Age at first onset of a DSM-IV major depressive episode was the dependent variable. Associations of depression with traumatic events were examined while accounting for the temporal sequence of trauma and depression and familial effects.ResultsAssaultive traumatic events that occurred during childhood had the strongest association with immediate and long-term risk for depression, and outweighed familial effects on childhood-onset depression for most twins. Although men and women endorsed equal rates of assaultive trauma, women reported a greater accumulation of assaultive events at earlier ages than men, whereas men reported a greater accumulation of non-assaultive events at all ages.ConclusionsEarly exposure to assaultive trauma can influence risk for depression into adulthood. Concordance for early trauma is a significant contributor to the familiality of early-onset depression.

SummaryBackgroundMost of our knowledge about borderline personality disorder features has been obtained through the study of clinical samples. Although these studies are important in their own right, they are limited in their ability to address certain important epidemiological and aetiological questions such as the degree to which there is a genetic influence on the manifestation of borderline personality disorder features. Though family history studies of borderline personality disorder indicate genetic influences, there have been very few twin studies and the degree of genetic influence on borderline personality disorder remains unclear.MethodData were drawn from twin samples from The Netherlands (n=3918), Belgium (n=904) and Australia (n=674). In total, data were available on 5496 twins between the ages of 18 and 86 years from 3644 families who participated in the study by completion of a mailed self-report questionnaire on borderline personality disorder features.ResultsIn all countries, females scored higher than males and there was a general tendency for younger adults to endorse more borderline personality disorder features than older adults. Model-fitting results showed that additive genetic influences explain 42% of the variation in borderline personality disorder features in both men and women and that this heritability estimate is similar across The Netherlands, Belgium and Australia. Unique environmental influences explain the remaining 58% of the variance.ConclusionGenetic factors play a role in individual differences in borderline personality disorder features in Western society.

Background.Genetic influences contribute significantly to co-morbidity between conduct disorder and substance use disorders. Estimating the extent of overlap can assist in the development of phenotypes for genomic analyses.Method.Multivariate quantitative genetic analyses were conducted using data from 9577 individuals, including 3982 complete twin pairs and 1613 individuals whose co-twin was not interviewed (aged 24–37 years) from two Australian twin samples. Analyses examined the genetic correlation between alcohol dependence, nicotine dependence and cannabis abuse/dependence and the extent to which the correlations were attributable to genetic influences shared with conduct disorder.Results.Additive genetic (a2 = 0.48–0.65) and non-shared environmental factors explained variance in substance use disorders. Familial effects on conduct disorder were due to additive genetic (a2 = 0.39) and shared environmental (c2 = 0.15) factors. All substance use disorders were influenced by shared genetic factors (rg = 0.38–0.56), with all genetic overlap between substances attributable to genetic influences shared with conduct disorder. Genes influencing individual substance use disorders were also significant, explaining 40–73% of the genetic variance per substance.Conclusions.Among substance users in this sample, the well-documented clinical co-morbidity between conduct disorder and substance use disorders is primarily attributable to shared genetic liability. Interventions targeted at generally reducing deviant behaviors may address the risk posed by this shared genetic liability. However, there is also evidence for genetic and environmental influences specific to each substance. The identification of these substance-specific risk factors (as well as potential protective factors) is critical to the future development of targeted treatment protocols.

Background. The use of cannabis and other illicit drugs (OIDs) and their co-morbid misuse are frequently reported in the literature. Correlated vulnerabilities and causal or gateway influences have been implicated in this association. We investigated the source of this co-morbidity between cannabis use (experimentation, early and repeated use, and problems) and OID experimentation and problems using genetic models proposed by Neale and Kendler (American Journal of Human Genetics 1995, 57, 935–953).Method. In a sample of 4152 same-sex male and female adult Australian twin individuals, we fit 13 genetically informative models of co-morbidity to data on experimentation, early use, repeated use of cannabis and co-morbid OID experimentation, and to abuse/dependence (A/D) problems with cannabis and OIDs.Results. Model-fitting results suggest that common genetic, shared and unique environmental factors are responsible for the association between cannabis experimentation, early use, repeated use and A/D problems and OID experimentation or problems. The liability causation model, which is a reduced form of the correlated vulnerabilities model, also fit very well. In women, we found evidence for high-risk cannabis experimenters and repeated users to be at increased risk for OID experimentation, despite being below the risk threshold on the liability distribution for OID experimentation (extreme multiformity).Conclusions. Co-morbid cannabis and OID use and misuse are due partly to a common predisposition to substance use disorders. Putative causal effects could not be ruled out. These models warrant further research, so that features of the correlated vulnerabilities model and the gateway models can be studied jointly in a single series of adaptive nested models.

Background. This study examined the relationships between self-reported childhood sexual abuse (CSA) and drug-related outcomes in an Australian twin panel.Method. A semi-structured psychiatric interview was conducted in 1996–2000 by telephone with young adult Australian twins (mean age 29·9 years). Data reported here are from 6050 twins who responded to both CSA and drug-related items.Results. A history of CSA was associated with significant risk for subsequently occurring regular smoking and use of each illicit drug class. Further CSA-associated risk was found among regular users, for nicotine and alcohol dependence, and among illicit drug users, for abuse/dependence of most drug classes. In same-sex discordant pairs, significant risk for regular smoking and illicit drug use was found in twins with a history of CSA compared to their non-abused co-twins. Similar analyses for abuse/dependence found significant risk for opioids, any illicit drug, and any non-cannabis illicit drug. CSA was associated with significantly earlier drug use. Despite the association of CSA with risk for early-onset cannabis use and regular smoking, risks for illicit drug outcomes associated with CSA and with either form of early-onset use combine in near-additive fashion.Conclusions. CSA is associated with risk for subsequently occurring regular smoking and illicit drug use and abuse/dependence. Risks for drug use are mildly attenuated with control for familial contributions; similar risks for abuse/dependence remain significant for opioids and for illicit drugs combined across classes. Although we found evidence of earlier onset drug use with CSA, risks associated with CSA and with early-onset use combine in a largely additive manner.

Background. Children of alcoholics are significantly more likely to experience high-risk environmental exposures, including prenatal substance exposure, and are more likely to exhibit externalizing problems [e.g. attention deficit hyperactivity disorder (ADHD)]. While there is evidence that genetic influences and prenatal nicotine and/or alcohol exposure play separate roles in determining risk of ADHD, little has been done on determining the joint roles that genetic risk associated with maternal alcohol use disorder (AUD) and prenatal risk factors play in determining risk of ADHD.Method. Using a children-of-twins design, diagnostic telephone interview data from high-risk families (female monozygotic and dizygotic twins concordant or discordant for AUD as parents) and control families targeted from a large Australian twin cohort were analyzed using logistic regression models.Results. Offspring of twins with a history of AUD, as well as offspring of non-AUD monozygotic twins whose co-twin had AUD, were significantly more likely to exhibit ADHD than offspring of controls. This pattern is consistent with a genetic explanation for the association between maternal AUD and increased offspring risk of ADHD. Adjustment for prenatal smoking, which remained significantly predictive, did not remove the significant genetic association between maternal AUD and offspring ADHD.Conclusions. While maternal smoking during pregnancy probably contributes to the association between maternal AUD and offspring ADHD risk, the evidence for a significant genetic correlation suggests: (i) pleiotropic genetic effects, with some genes that influence risk of AUD also influencing vulnerability to ADHD; or (ii) ADHD is a direct risk-factor for AUD.

Background. Genetic influences have been shown to play a major role in determining the risk of alcohol dependence (AD) in both women and men; however, little attention has been directed to identifying the major sources of genetic variation in AD risk.Method. Diagnostic telephone interview data from young adult Australian twin pairs born between 1964 and 1971 were analyzed. Cox regression models were fitted to interview data from a total of 2708 complete twin pairs (690 MZ female, 485 MZ male, 500 DZ female, 384 DZ male, and 649 DZ female/male pairs). Structural equation models were fitted to determine the extent of residual genetic and environmental influences on AD risk while controlling for effects of sociodemographic and psychiatric predictors on risk.Results. Risk of AD was increased in males, in Roman Catholics, in those reporting a history of major depression, social anxiety problems, and conduct disorder, or (in females only) a history of suicide attempt and childhood sexual abuse; but was decreased in those reporting Baptist, Methodist, or Orthodox religion, in those who reported weekly church attendance, and in university-educated males. After allowing for the effects of sociodemographic and psychiatric predictors, 47% (95% CI 28–55) of the residual variance in alcoholism risk was attributable to additive genetic effects, 0% (95% CI 0–14) to shared environmental factors, and 53% (95% CI 45–63) to non-shared environmental influences.Conclusions. Controlling for other risk factors, substantial residual heritability of AD was observed, suggesting that psychiatric and other risk factors play a minor role in the inheritance of AD.

Background. Whether current criteria used to define nicotine dependence are informative for genetic research is an important empirical question. The authors used items of the DSM-IV and of the Heaviness of Smoking Index to characterize the nicotine dependence phenotype and to identify salient symptoms in a genetically informative community sample of Australian young adult female and male twins.Method. Phenotypic and genetic factor analyses were performed on nine dependence symptoms (the seven DSM-IV substance dependence criteria and the two Heaviness of Smoking Index (HSI) items derived from the Fagerström Tolerance Questionnaire, time to first cigarette in the morning and number of cigarettes smoked per day). Phenotypic and genetic analyses were restricted to ever smokers.Results. Phenotypic nicotine dependence symptom covariation was best captured by two factors with a similar pattern of factor loadings for women and men. In genetic factor analysis item covariation was best captured by two genetic but one shared environmental factor for both women and men; however, item factor loadings differed by gender. All nicotine dependence symptoms were substantially heritable, except for the DSM-IV criterion of ‘giving up or reducing important activities in order to smoke’, which was weakly familial.Conclusions. The salient behavioral indices of nicotine dependence are similar for women and men. DSM-IV criteria of tolerance, withdrawal, and experiencing difficulty quitting and HSI items time to first cigarette in the morning and number of cigarettes smoked per day may represent the most highly heritable symptoms of nicotine dependence for both women and men.

Background. We aimed to explore the prevalence and factor structure of premenstrual symptoms in a sample of Australian twins; to investigate phenotypic associations between reported premenstrual symptoms, personality and reproductive dimensions; and to identify the relative contributions of genes and environment to premenstrual symptoms and the extent of genetic and environmental covariation with the personality trait Neuroticism and lifetime major depression.Method. Seven hundred and twenty female twin pairs (454 monozygotic and 266 dizygotic) from the Australian National Health and Medical Research Council Twin Register reported on experience of 17 premenstrual symptoms during the previous 12 months. In the same questionnaire twins also responded to questions on symptom states, and personality dimensions including neuroticism. Interview data enabling diagnosis of lifetime history of DSM-IV major depression were also available. We fitted univariate and multivariate genetic models to the data.Results. Most frequently reported symptoms were breast tenderness/pain and bloating/weight gain, followed by affective symptoms. Twelve-month prevalence was 2·4% for the combination of symptoms and functional interference meeting a very rough approximation of DSM-III-R criteria for late luteal dysphoric disorder. Principal factor analysis identified a single premenstrual (PMS) factor. Additive genetic influences (44% of total variance) were identified for PMS. Although we found genetic correlations of 0·62 between reported PMS and neuroticism, and 0·70 with lifetime major depression, 39% of the genetic variance of PMS was not explained by these factors.Conclusions. Our findings support the existence of genetic influences on premenstrual symptoms, but we were unable to distinguish between liability to symptom experience and symptom reporting. Retrospective reporting may have contributed to our finding that PMS genes were shared in part with neuroticism and liability to lifetime major depression.

Background. We examined the relationship between childhood sexual abuse (CSA), and interviewees' recollections of pathogenic parenting, testing for possible retrospective biases in the recollections of those who have experienced CSA.Methods. Information about CSA, parental divorce and interviewees' recollections of parental rejection, parental overprotection and perceived autonomy (as assessed through a shortened version of the Parental Bonding Instrument) was obtained through telephone interviews with 3626 Australian twins who had also returned self-report questionnaires several years earlier. Recollections of parental behaviours were compared for individuals from pairs in which neither twin, at least one twin, or both twins reported CSA.Results. Significant associations were noted between CSA and paternal alcoholism and between CSA and recollections of parental rejection. For women, individuals from CSA-discordant pairs reported levels of parental rejection that were significantly higher than those obtained from CSA-negative pairs. The levels of parental rejection observed for twins from CSA-discordant pairs did not differ significantly from those obtained from CSA-concordant pairs, regardless of respondent's abuse status. For men from CSA-discordant pairs, respondents reporting CSA displayed a tendency to report higher levels of parental rejection than did respondents not reporting CSA. Other measures of parenting behaviour (perceived autonomy and parental overprotection) failed to show a clear relationship with CSA.Conclusions. The relationship between CSA and respondents' recollections of parental rejection is not due solely to retrospective bias on the part of abused individuals and, consistent with other studies, may reflect a pathological family environment with serious consequences for all siblings.

Background. This study was designed to determine lifetime prevalence of psychiatric disorders among twins who reported childhood sexual abuse (CSA), and to compare these rates with those among non-abused co-twins. The contribution of familial and individual-specific factors to reported sexual abuse was also examined.Method. Information about lifetime psychopathology and substance use was obtained by structured telephone interviews with 5995 Australian twins. Twins who reported a history of childhood sexual abuse (CSA) were contrasted on lifetime psychopathology with subjects without such a history; in addition, comparisons were made between same-sex twin pairs discordant for CSA.Results. A history of CSA was reported by 5·9% of the women and 2·5% of the men. In the sample as a whole, those reporting CSA were more likely to receive lifetime diagnoses of major depression, conduct disorder, panic disorder and alcoholism, and were more likely to report suicidal ideation and a history of suicide attempt. Abused women, but not men, were also more likely to report social phobia. When comparisons were restricted to non-abused co-twins, no differences in psychopathology were seen. However, rates of major depression, conduct disorder and suicidal ideation were higher if both co-twins were abused than if the respondent alone reported CSA. Model-fitting indicated that shared environmental factors influenced risk for reported CSA in women, but not in men.Conclusion. The association between CSA and psychopathology arises at least in part through the influence of shared familial factors on both risk of victimization and risk of psychopathology.

Background. Substantial evidence exists for an important genetic contribution to alcohol dependence risk in women and men. It has been suggested that genetically determined differences in alcohol sensitivity may represent one pathway by which an increase in alcohol dependence risk occurs.Methods. Telephone interview follow-up data were obtained on twins from male, female and unlike-sex twin pairs who had participated in an alcohol challenge study in 1979–81, as well as other pairs from the same Australian twin panel surveyed by mail in 1980–82.Results. At follow-up, alcohol challenge men did not differ from other male twins from the same age cohort on measures of lifetime psychopathology or drinking habits; but alcohol challenge women were on average heavier drinkers than other women. Acomposite alcohol sensitivity measure, combining subjective intoxication and increase in body-sway after alcohol challenge in 1979–81, exhibited high heritability (60%). Parental alcoholism history was weakly associated with decreased alcohol sensitivity in women, but not after adjustment for baseline drinking history, or in men. High alcohol sensitivity in men was associated with substantially reduced alcohol dependence risk (OR=0·05, 95% CI 0·01–0·39). Furthermore, significantly decreased (i.e. low) alcohol sensitivity was observed in non-alcoholic males whose MZ co-twin had a history of alcohol dependence, compared to other non-alcoholics. These associations remained significant in conservative analyses that controlled for respondents' alcohol consumption levels and alcohol problems in 1979–81.Conclusions. Men (but not women) at increased genetic risk of alcohol dependence (assessed by MZ co-twin's history of alcohol dependence) exhibited reduced alcohol sensitivity. Associations with parental alcoholism were inconsistent.

Background. Conflicting evidence exists on causes of vulnerability to post-natal depression. We investigated genetic and environmental influences on variation in post-natal depressive symptoms (PNDS) following first live birth, and sources of covariation with the personality trait Neuroticism and lifetime major depression occurring post-natally (DEP-PN) and at other times (DEP-XPN) to test for shared genetic influences.Method. Retrospective interview and questionnaire data from 838 parous female twin pairs (539 monozygotic, 299 dizygotic) from the Australian National Health and Medical Research Council volunteer adult twin register were used for multivariate genetic model-fitting. Data on PNDS were evaluated for consistency with diagnostic interview assessment.Results. Genetic factors explained 38% of variance in PNDS (95% confidence interval 26–49%) and 25% of the variance in interview-assessed DEP-PN. The genetic correlation between PNDS and lifetime major depression (DEP-PN and DEP-XPN) was low (rg = 0·17, 95% confidence interval = 0·09–0·28), suggesting that the questionnaire was measuring a construct other than post-natally occurring major depression, possibly post-natal dysphoria. Associations between PNDS and obstetric factors were very modest.Conclusions. Findings suggest modest genetic influences on major depression occurring post-natally. Independent and stronger genetic influences identified for post-natal symptomatology or dysphoria (PNDS) justify further investigation.

Background. Medically unexplained chronic fatigue in childhood may cause considerable disability and (by definition) its cause remains unclear. A study of fatigue in healthy twins has been undertaken to examine whether or not genetic factors play a part.Method. A questionnaire survey of the main carers of an epidemiological population-based sample of 670 twin pairs who were asked about periods of unexplained and disabling fatigue in their twins. Out of 1340 individuals a period of disabling fatigue was reported for 92 (6·9%). Thirty-three (2·5%) reported disabling fatigue for more than 1 month. Zygosity could be confidently assigned in 98% of the sample providing 278 monozygotic (MZ) and 378 dizygotic (DZ) pairs. These data were analysed using a structural equation modelling approach.Results. The results showed that disabling fatigue in childhood is highly familial with an MZ tetrachoric correlation (rMZ) of 0·81 and a DZ tetrachoric correlation (rDZ) of 0·59, for fatigue lasting at least a week. The most acceptable model using Akaike's information criteria, was one containing additive genetic effects (A) and shared environment (C) plus residual (or non-shared) environment (E). For fatigue lasting at least a month rMZ was 0·75 and rDZ 0·47. The most acceptable model included just A and E. However, the role of shared environment could not be conclusively rejected.Conclusions. Unexplained disabling fatigue in childhood is substantially familial. Both genetic and shared environmental factors are worth further exploration in a search for the causes.

Background. Prolonged fatigue syndromes have been proposed as prevalent and disabling forms of distress that occur independently of conventional notions of anxiety and depression.Methods. To investigate the genetic and environmental antecedents of common forms of psychological and somatic distress, we measured fatigue, anxiety, depression and psychological distress in 1004 normal adult twin pairs (533 monozygotic (MZ), 471 dizygotic (DZ)) over 50 years of age.Results. Familial aggregation of psychological distress, anxiety and fatigue appeared to be due largely to additive genetic factors (MZ[ratio ]DZ ratios of 2·12–2·69). The phenotypic correlations between the psychological measures (distress, anxiety and depression) were moderate (0·67–0·79) and higher than that between fatigue and psychological distress (0·38). Multivariate genetic modelling revealed a common genetic factor contributing to the development of all the observed phenotypes (though most strongly for the psychological forms), a second independent genetic factor also influenced anxiety and depression and a third independent genetic factor made a major contribution to fatigue alone. In total, 44% (95% CI 25–60%) of the genetic variance for fatigue was not shared by the other forms of distress. Similarly, the environmental factor determining psychological distress made negligible contributions to fatigue, which was underpinned largely by its own independent environmental factor.Conclusion. This study supports the aetiological independence of prolonged fatigue and, therefore, argues strongly for its inclusion in classification systems in psychiatry.

Background. Genetic influences on alcoholism risk are well-documented in men, but uncertain in women. We tested for gender differences in genetic influences on, and risk-factors for, DSM-III-R alcohol dependence (AD).Method. Diagnostic follow-up interviews were conducted in 1992–3 by telephone with twins from an Australian twin panel first surveyed in 1980–82 (N=5889 respondents). Data were analysed using logistic regression models.Results. Significantly higher twin pair concordances were observed in MZ compared to DZ same-sex twin pairs in women and men, even when data were weighted to adjust for over-representation of well-educated respondents, and for selective attrition. AD risk was increased in younger birth cohorts, in Catholic males or women reporting no religious affiliation, in those reporting a history of conduct disorder or major depression and in those with high Neuroticism, Social Non-conformity, Toughmindedness, Novelty-Seeking or (in women only) Extraversion scores; and decreased in ‘Other Protestants’, weekly church attenders, and university-educated males. Controlling for these variables, however, did not remove the significant association with having an alcoholic MZ co-twin, implying that much of the genetic influence on AD risk remained unexplained. No significant gender difference in the genetic variance in AD was found (64% heritability, 95% confidence interval 32–73%).Conclusions. Genetic risk-factors play as important a role in determining AD risk in women as in men. With the exception of certain sociocultural variables such as religious affiliation, the same personality, sociodemographic and axis I correlates of alcoholism risk are observed in women and men.

BackgroundGenetic studies in adults indicate that genes influencing the personality trait of neuroticism account for substantial genetic variance in anxiety and depression and in somatic health. Here, we examine for the first time the factors underlying the relationship between neuroticism and anxiety/depressive and somatic symptoms during adolescence.MethodThe Somatic and Psychological Health Report (SPHERE) assessed symptoms of anxiety/depression (PSYCH-14) and somatic distress (SOMA-10) in 2459 adolescent and young adult twins [1168 complete pairs (35.4% monozygotic, 53% female)] aged 12–25 years (mean=15.5±2.9). Differences between boys and girls across adolescence were explored for neuroticism, SPHERE-34, and the subscales PSYCH-14 and SOMA-10. Trivariate analyses partitioned sources of covariance in neuroticism, PSYCH-14 and SOMA-10.ResultsGirls scored higher than boys on both neuroticism and SPHERE, with SPHERE scores for girls increasing slightly over time, whereas scores for boys decreased or were unchanged. Neuroticism and SPHERE scores were strongly influenced by genetic factors [heritability (h2)=40–52%]. A common genetic source influenced neuroticism, PSYCH-14 and SOMA-10 (impacting PSYCH-14 more than SOMA-10). A further genetic source, independent of neuroticism, accounted for covariation specific to PSYCH-14 and SOMA-10. Environmental influences were largely specific to each measure.ConclusionsIn adolescence, genetic risk factors indexed by neuroticism contribute substantially to anxiety/depression and, to a lesser extent, perceived somatic health. Additional genetic covariation between anxiety/depressive and somatic symptoms, independent of neuroticism, had greatest influence on somatic distress, where it was equal in influence to the factor shared with neuroticism.

BackgroundAssociations between parental depression and offspring affective and disruptive disorders are well documented. Few genetically informed studies have explored the processes underlying intergenerational associations.MethodA semi-structured interview assessing DSM-III-R psychiatric disorders was administered to twins (n=1296) from the Australian Twin Register (ATR), their spouses (n=1046) and offspring (n=2555). We used the Children of Twins (CoT) design to delineate the extent to which intergenerational associations were consistent with a causal influence or due to genetic confounds.ResultsIn between-family analyses, parental depression was associated significantly with offspring depression [hazard ratio (HR) 1.52, 95% confidence interval (CI) 1.20–1.93] and conduct disorder (CD; HR 2.27, CI 1.31–3.93). Survival analysis indicated that the intergenerational transmission of depression is consistent with a causal (environmental) inference, with a significant intergenerational association in offspring of discordant monozygotic (MZ) twin pairs (HR 1.39, CI 1.00–1.94). Logistic regression analysis suggested that the parental depression–offspring CD association was due to shared genetic liability in the parents and offspring. No intergenerational association was found when comparing the offspring of discordant MZ twins [odds ratio (OR) 1.41, CI 0.63–3.14], but offspring of discordant dizygotic (DZ) twins differed in their rates of CD (OR 2.53, CI 0.95–6.76). All findings remained after controlling for several measured covariates, including history of depression and CD in the twins' spouses.ConclusionsThe mechanisms underlying associations between parental depression and offspring psychopathology seem to differ depending on the outcome. The results are consistent with a causal environmental role of parental depression in offspring depression whereas common genetic factors account for the association of parental depression and offspring CD.

BackgroundThree cognitive constructs are risk factors for eating disorders: undue influence of weight and shape, concern about weight and shape, and body dissatisfaction (BD). Undue influence, a diagnostic criterion for eating disorders, is postulated to be closely associated with self-esteem whereas BD is postulated to be closely associated with body mass index (BMI). We understand less about the relationships with concern about weight and shape. The aim of the current investigation was examine the degree of overlap across these five phenotypes in terms of latent genetic and environmental risk factors in order to draw some conclusions about the similarities and differences across the three cognitive variables.MethodA sample of female Australian twins (n=1056, including 348 complete pairs), mean age 35 years (s.d.=2.11, range 28–40), completed a semi-structured interview about eating pathology and self-report questionnaires. An independent pathways model was used to investigate the overlap of genetic and environmental risk factors for the five phenotypes.ResultsIn terms of variance that was not shared with other phenotypes, self-esteem emerged as being separate, with 100% of its variance unshared with the other phenotypes, followed by undue influence (51%) and then concern (34%), BD (28%) and BMI (32%).ConclusionsIn terms of shared genetic risk, undue influence and concern were more closely related than BD, whereas BMI and BD were found to share common sources of risk. With respect to environmental risk factors, concern, BMI and BD were more closely related to each other than to undue influence.

BackgroundObjective binge eating (OBE) and self-induced vomiting (SIV) occur and co-occur across a range of eating disorders but the extent to which the risk factors for these two behaviours overlap is unclear. Examination of this overlap was the focus of the current report.MethodA population of female Australian twins (n=1002), mean age 35 years (s.d.=2.11, range 28–40), participated in three waves of data collection and were assessed for lifetime disordered eating with a semi-structured interview at wave 3 and a self-report questionnaire at wave 1; risk factors were assessed via self-report at waves 1 and 3.ResultsNon-shared environmental influences were the largest contributor to the variance of both OBE and SIV, with a more modest contribution of genetic influences. Between 5% and 14% of the environmental risk factors for OBE and SIV were shared and 27–100% of genetic risk factors were shared. SIV initiation was predicted by higher neuroticism and novelty seeking and lower maternal and paternal care, whilst lower levels of perceived paternal care, higher lifetime BMI, and a wider BMI range predicted OBE initiation. Retrospective correlates associated with both SIV and OBE onset were parental comments about weight, whilst higher levels of parental conflict, expectations and criticism was associated with OBE onset only.ConclusionsThe substantial extent of non-overlap between risk factors for SIV and OBE suggests that each of these behavioural disturbances warrants future investigation in its own right, not only when they occur in conjunction with each other.

BackgroundWe investigated the genetic and environmental contributions to covariation between smoking age-at-onset, cigarette consumption and smoking persistence.MethodMultivariate biometrical modelling methods were applied to questionnaire data from Australian twins and their siblings (14 472 individuals from 6247 families). The contributions of genetic and environmental factors to covariation between the three traits were estimated, allowing for sex differences in both trait prevalence and the magnitude of genetic and environmental effects.ResultsAll traits were moderately heritable in males and females (estimates between 0·40 and 0·62), but there were sex differences in the extent to which additive genetic influences were shared across traits. Twin-specific environmental factors accounted for a substantial proportion of the variance in smoking age-at-onset in females (0·19) and males (0·12), but had little influence (ConclusionsThese results provide some insight into observed sex differences in smoking behaviour, and suggest that searching for pleiotropic genes may prove fruitful. However, further work on phenotypic definitions of smoking behaviour, particularly persistence, is warranted.

Background. Serotonin is a good candidate for major depression. We attempted to replicate the study by Caspi and colleagues [Science (2003) 301, 386–389] which reported a significant interaction between serotonin transporter (5-HTTLPR) genotype and stressful life events when predicting major depression. Method. We typed the serotonin promoter 5-HTTLPR gene in 1206 male and female twins aged 19–78 years (mean=39, S.D.=11). A DSM-IV diagnosis of major depression was available for 1199 twins. Most of these twins had participated in a 1988–1990 study which included a stressful life events inventory and self-report measure of depression based on the SCL-90 and DSSI/sAD. Complete 5-HTT genotype and life events data, self-report symptoms and major depression diagnoses were available for 1091 subjects. We regressed categorical and ordinal measures of depression onto stressful life events and genotype. Results. There were significant main effects for stressful life events but there was no evidence for any effect of 5-HTT genotype, nor a genotyperstressful life event interaction. Conclusions. Regardless of whether our results were based on binary logistic or ordinal regression analyses we found no evidence to support a main effect of 5-HTTLPR, or an interaction between the 5-HTTLPR genotype and stressful life events on major depression. Only 20% of our subjects were aged below 30 years. It is possible that the effect reported by Caspi and colleagues is specific to young people, in which case our study has much less power in this age group.

Background. Somatoform disorders such as neurasthenia and chronic fatigue syndrome are characterized by a combination of prolonged mental and physical fatigue. This study aimed to investigate the heritability of somatic distress and determine whether this dimension is aetiologically distinct from measures of depression and anxiety.Method. Measures of anxiety, depression, phobic anxiety, somatic distress and sleep difficulty were administered in a self-report questionnaire to a community-based sample of 3469 Australian twin individuals aged 18 to 28 years. Factor analysis using a Promax rotation, produced four factors: depression, phobic anxiety, somatic distress and sleep disturbance. Multivariate and univariate genetic analyses of the raw categorical data scores for depression, phobic anxiety and depression were then analysed in Mx1.47.Results. Univariate genetic analysis revealed that an additive genetic and non-shared environmental (AE) model best explained individual differences in depression and phobic anxiety scores, for male and female twins alike, but could not resolve whether additive genes or shared environment were responsible for significant familial aggregation in somatic distress. However, multivariate genetic analysis showed that an additive genetic and non-shared environment (AE) model best explained the covariation between the three factors. Furthermore, 33% of the genetic variance in somatic distress was due to specific gene action unrelated to depression or phobic anxiety. In addition, 74% of the individual environmental influence on somatic distress was also unrelated to depression or phobic anxiety.Conclusion. These results support previous findings that somatic symptoms are relatively aetiologically distinct both genetically and environmentally from symptoms of anxiety and depression.

Background. Many reports have appeared on associations between platelet monoamine oxidase (MAO) activity and susceptibility to psychiatric conditions; principally alcohol dependence but also conduct disorder, other drug use and depression. Recently, it has become apparent that MAO activity is inhibited by some component of cigarette smoke, and smokers have low platelet MAO activity. Since the prevalence of smoking is higher in many of the conditions in which MAO has been implicated, the MAO susceptibility associations may be partly, or entirely, false.Methods. We have measured platelet MAO in 1551 subjects, recruited from the Australian NHMRC Twin Registry, who have provided information on alcohol use and dependence, smoking, conduct disorder, depression, attempted suicide, panic disorder and social phobia.Results. Current smoking reduced platelet MAO activity in a significant and dose-related manner, with no evidence of lower MAO in ex-smokers or in non-smoking subjects with co-twins who smoked. Alcohol use and lifetime DSM-III-R alcohol dependence history were not associated with MAO activity when smoking was taken into account. Depression, panic disorder and social phobia showed no significant associations with platelet MAO activity. Subjects with a history of serious attempts at suicide had low platelet MAO activity; but although the difference from controls was as great as the reduction associated with smoking it was not significant after correction for smoking effects.Conclusions. Although synaptic MAO activity undoubtedly plays a role in psychopathology, the concept that platelet MAO activity is a direct genetic marker of vulnerability to alcohol dependence cannot be sustained.

Background. The study explored the genetic and environmental risk factors for both the behaviours and attitudes characteristic of disordered eating.Methods. In three waves of data collection, information was collected from female twins regarding their eating and attitudes towards eating, weight and shape. The first assessment consisted of a self-report questionnaire (1988–9) with 1682 women. The second assessment consisted of a semi-structured psychiatric interview schedule (1992–3), completed by 1852 women, many of whom had completed Wave 1 assessment. The third assessment, with 325 women chosen from Waves 1 and 2 (1995–6), consisted of a semi-structured interview (the Eating Disorder Examination).Results. As only one twin pair was concordant for lifetime bulimia nervosa at Wave 3 assessment, ordinal measures of all assessments were used in a multivariate genetic analysis. Results indicated that additive genetic and non-shared environmental influences best explained variance in liability to disordered eating, with about 60% (95% CI 50–68) of the variance explained by genetic factors. Comparison with a model allowing for the effects of shared environment indicated genetic factors accounted for a similar degree of variance (59%, 95% CI 36–68).Conclusion. Liability to the development of the behaviours and attitudes characteristic of eating disorders is best explained by both environmental and genetic factors, with covariation between the three measures best explained by a single latent phenotype of disordered eating which has a heritability of 60%.

Background. Although there is considerable support for adverse relationships between states of psychological and somatic distress and immune response, there is little evidence in humans of the relative contribution of genetic and environmental factors.Methods. This study utilized a twin methodology to examine the interplay between psychological distress, fatigue and immune function. We recorded a number of measures of distress, including conventional depression and anxiety as well as the somatic symptom of prolonged fatigue, and immune responsiveness (by delayed-type hypersensitivity skin response) in 124 normal adult twin pairs (79 monozygotic, 45 dizygotic).Results. While there were strong genetic influences on the psychological distress and fatigue factors (only some of which are common to both), familial aggregation of immune responsiveness arose mainly from environmental factors shared by both members of a twin pair. Phenotypic correlations between psychological and immune measures were negligible, but multivariate genetic modelling revealed that these masked larger genetic and environmental correlations of opposite sign. Negative environmental effects of psychological distress and fatigue on immune responsiveness were countered by a positive genetic relationship between psychological distress and immune function.Conclusions. Our study suggests that current psychoneuroimmunological hypotheses in humans need to be modified to place increasing importance on the individual's genotype. In this cohort immune responsiveness varied in response to a complex interplay of genetic and environmental factors. Additionally, although psychological distress and fatigue had some shared genetic determinants, independent genetic and environmental risk factors for fatigue were also identified.

Background. Psychiatric history, familial history of suicide attempts, and certain traumatic life events are important predictors of suicidal thoughts and behaviour. We examined the epidemiology and genetics of suicidality (i.e. reporting persistent suicidal thoughts or a plan or suicide attempt) in a large community-based sample of MZ and DZ twin pairs.Method. Diagnostic telephone interviews were conducted in 1992–3 with twins from an Australian twin panel first surveyed in 1980–82 (N=5995 respondents). Data were analysed using logistic regression models, taking into account twin pair zygosity and the history of suicidality in the respondent's co-twin.Results. Lifetime prevalence of suicidal thoughts and attempts was remarkably constant across birth cohorts 1930–1964, and across gender. Major psychiatric correlates were history of major depression, panic disorder, social phobia in women, alcohol dependence and childhood conduct problems. Traumatic events involving assault (childhood sexual abuse, rape or physical assault) or status-loss (job loss, loss of property or home, divorce), and the personality trait neuroticism, were also significantly associated with suicide measures. Prevalence of serious suicide attempts varied as a function of religious affiliation. After controlling for these variables, however, history of suicide attempts or persistent thoughts in the respondent's co-twin remained a powerful predictor in MZ pairs (odds ratio=3·9), but was not consistently predictive in DZ pairs. Overall, genetic factors accounted for approximately 45% of the variance in suicidal thoughts and behaviour (95% confidence interval 33–51%).Conclusions. Risk of persistent suicidal thoughts and suicide attempts is determined by a complex interplay of psychiatric history, neuroticism, traumatic life experiences, genetic vulnerability specific for suicidal behaviour and sociocultural risk or protective factors.