Meta-analysis of genome-wide association studies for personality.

Personality can be thought of as a set of characteristics that influence people's thoughts, feelings and behavior across a variety of settings. Variation in personality is predictive of many outcomes in life, including mental health. Here we report on a meta-analysis of genome-wide association (GWA) data for personality in 10 discovery samples (17,375 adults) and five in silico replication samples (3294 adults). All participants were of European ancestry. Personality scores for Neuroticism, Extraversion, Openness to Experience, Agreeableness and Conscientiousness were based on the NEO Five-Factor Inventory. Genotype data of ≈ 2.4M single-nucleotide polymorphisms (SNPs; directly typed and imputed using HapMap data) were available. In the discovery samples, classical association analyses were performed under an additive model followed by meta-analysis using the weighted inverse variance method. Results showed genome-wide significance for Openness to Experience near the RASA1 gene on 5q14.3 (rs1477268 and rs2032794, P=2.8 × 10(-8) and 3.1 × 10(-8)) and for Conscientiousness in the brain-expressed KATNAL2 gene on 18q21.1 (rs2576037, P=4.9 × 10(-8)). We further conducted a gene-based test that confirmed the association of KATNAL2 to Conscientiousness. In silico replication did not, however, show significant associations of the top SNPs with Openness and Conscientiousness, although the direction of effect of the KATNAL2 SNP on Conscientiousness was consistent in all replication samples. Larger scale GWA studies and alternative approaches are required for confirmation of KATNAL2 as a novel gene affecting Conscientiousness.

Genome-wide association scan for five major dimensions of personality.

Personality traits are summarized by five broad dimensions with pervasive influences on major life outcomes, strong links to psychiatric disorders and clear heritable components. To identify genetic variants associated with each of the five dimensions of personality we performed a genome-wide association (GWA) scan of 3972 individuals from a genetically isolated population within Sardinia, Italy. On the basis of the analyses of 362 129 single-nucleotide polymorphisms we found several strong signals within or near genes previously implicated in psychiatric disorders. They include the association of neuroticism with SNAP25 (rs362584, P=5 x 10(-5)), extraversion with BDNF and two cadherin genes (CDH13 and CDH23; Ps<5 x 10(-5)), openness with CNTNAP2 (rs10251794, P=3 x 10(-5)), agreeableness with CLOCK (rs6832769, P=9 x 10(-6)) and conscientiousness with DYRK1A (rs2835731, P=3 x 10(-5)). Effect sizes were small (less than 1% of variance), and most failed to replicate in the follow-up independent samples (N up to 3903), though the association between agreeableness and CLOCK was supported in two of three replication samples (overall P=2 x 10(-5)). We infer that a large number of loci may influence personality traits and disorders, requiring larger sample sizes for the GWA approach to confidently identify associated genetic variants.

High neuroticism and low conscientiousness are associated with interleukin-6.

BACKGROUND: High Neuroticism and low Conscientiousness are frequently implicated in health-risk behaviors, such as smoking and overeating, as well as health outcomes, including mortality. Their associations with physiological markers of morbidity and mortality, such as inflammation, are less well documented. The present research examines the association between the five major dimensions of personality and interleukin-6 (IL-6), a pro-inflammatory cytokine often elevated in patients with chronic morbidity and frailty. METHOD: A population-based sample (n=4923) from four towns in Sardinia, Italy, had their levels of IL-6 measured and completed a comprehensive personality questionnaire, the NEO-PI-R. Analyses controlled for factors known to have an effect on IL-6: age; sex; smoking; weight; aspirin use; disease burden. RESULTS: High Neuroticism and low Conscientiousness were both associated with higher levels of IL-6. The findings remained significant after controlling for the relevant covariates. Similar results were found for C-reactive protein, a related marker of chronic inflammation. Further, smoking and weight partially mediated the association between impulsivity-related traits and higher IL-6 levels. Finally, logistic regressions revealed that participants either in the top 10% of the distribution of Neuroticism or the bottom 10% of conscientiousness had an approximately 40% greater risk of exceeding clinically relevant thresholds of IL-6. CONCLUSIONS: Consistent with the literature on personality and self-reported health, individuals high on Neuroticism or low on Conscientiousness show elevated levels of this inflammatory cytokine. Identifying critical medical biomarkers associated with personality may help to elucidate the physiological mechanisms responsible for the observed connections between personality traits and physical health.

Reproductive hormones in aging men. II. Basal pituitary gonadotropins and gonadotropin responses to luteinizing hormone-releasing hormone.

Although basal serum LH and FSH levels have been shown to increase with age in men, there is evidence that aging may impair pituitary gonadotropin secretion to some extent. We have measured serum LH and FSH before and after the iv injection of 100 micrograms LRH in 69 healthy men, divided into 3 age groups: A (25-49 yr; n = 24), B (50-69 yr; n = 23), and C (70-89 yr; n = 22). We have previously demonstrated these men to have unaltered serum androgens and estrogens over the entire age range. Only log-transformed measures were analyzed in order to equalize variance among groups. Because we found increasing age to be associated with significant rises in basal levels of LH and FSH, and because the basal level affects the magnitude of response, we used repeated measures analysis of variance to compare LRH peak responses with basal gonadotropin levels. The presence of a significant interaction term would indicate an effect of age on response beyond that expected from the effect of basal level alone. The mean LRH responses (log peak--log basal) decreased, with significant interaction terms for both LH (P = 0.001) and FSH (P less than 0.0001) responses. This means that the function relating basal to peak gonadotropin concentrations changed significantly with age. Peak LH responses in the youngest group (A) occurred equally as often 15 and 30 min after LRH injection, but in the oldest group (C) there were no 15 min peaks, and 40% of the peak values occurred at 60 min (P less than 0.01). FSH peaks were equally distributed at 30, 60, and 120 min in all 3 groups. Total testicular volumes did not decrease significantly with age. We conclude that there is a small but statistically significant tendency for LH and FSH responses to LRH to be both diminished and delayed in healthy aging men. The mechanisms responsible for this apparent pituitary gonadotropic hypofunction remain to be clarified.

Evaluation of pituitary gonadotropic function in men: value of luteinizing hormone-releasing hormone response versus basal luteinizing hormone level for discrimination of diagnosis.

We examined the serum LH response to LRH in 40 normal men and 38 men with various forms of gonadal dysfunction in an attempt to determine whether the LH response to LRH was more useful than the basal LH level alone for categorizing pathophysiological subgroups of gonadal dysfunction. The subgroups studied included hypogonadotropic hypogonadism, delayed puberty, idiopathic oligospermia, and primary hypogonadism. Log transformation of all values was done in order to normalize the data. Spearman's correlation analysis showed that increased basal LH was associated with increased incremental LH responses. However, our preliminary analysis of LRH response data suggested that the changes in LH were smaller in the secondary hypogonadal and delayed puberty groups then could be accounted for by the differences in basal LH level. Accordingly, we used various statistical techniques to test whether, and under what conditions, LRH testing provides information beyond that evident from examination of basal LH values. We found that the relationship between basal LH and the LH response to LRH differs in men with and without hypothalamic-pituitary dysfunction and that this difference can be exploited to improve the discrimination of men with pituitary hypogonadism from normals.

Do the dimensions of the temperament and character inventory map a simple genetic architecture? Evidence from molecular genetics and factor analysis.

OBJECTIVE: It has been reported that the human temperament dimensions of novelty seeking and harm avoidance are associated with polymorphisms in the D(4) dopamine receptor gene (D4DR) and the serotonin-transporter-linked promoter region (5-HTTLPR), respectively. Although these findings are consistent with Cloninger's hypothesized psychobiological model of temperament and character, many studies failed to replicate these findings. In the present study the authors tested whether the psychobiological model taps the genetic architecture of personality by exploring associations between these candidate genes and the dimensions of the Temperament and Character Inventory and by examining its phenotypic structure. METHOD: Of the 946 male and female participants in the Baltimore Longitudinal Study of Aging to whom the Temperament and Character Inventory was administered, 587 were genotyped for a polymorphism with a 48-base-pair repeat in the D4DR gene and 425 were genotyped for a 44-base-pair insertion or deletion in the 5-HTTLPR polymorphism. RESULTS: There was no significant association between D4DR polymorphisms and novelty seeking. The authors also failed to find an association between 5-HTTLPR polymorphisms and harm avoidance. The factor structure of the Temperament and Character Inventory did not reveal the hypothesized phenotypic structure. CONCLUSIONS: This investigation produced no support for the temperament-character model at either the biological or psychological level.

Solid ground in the wetlands of personality: a reply to Block.

The five-factor model (FFM) of personality offers a structural organization of personality traits in terms of 5 broad factors. J. Block's (1995) critique of the FFM failed to recognize the utility of a trait taxonomy and the intent of research designed to test the 5-factor hypothesis. In a number of instances he omitted reference to empirical evidence that addresses concerns he raised; this evidence shows strong support for the FFM beyond the lexical and questionnaire traditions he reviews. Many of his suggestions for improving the quality of personality research are valuable, but are likely to be more fruitful when used in conjunction with established knowledge about the structure of personality traits: the FFM.

Influence of the normal personality dimension of neuroticism on chest pain symptoms and coronary artery disease.

For at least the last 200 years it has been suspected that somatic manifestations of psychological distress play a role in the medical recognition and treatment of coronary artery disease (CAD). The cardiovascular system is intricately linked to the experience of emotion, and these links may explain how and when neuroticism can cloud the diagnosis of cardiovascular disease. A possible source of anginal symptoms in the absence of angiographically documented CAD is high standing on the personality dimension of neuroticism, which is a broad dimension of individual differences in the tendency to experience negative, distressing emotions and to possess associated behavioral and cognitive traits. A brief review of the clinical cardiologic literature on chest pain is presented, with special attention to distinguishing true angina pectoris from pseudoangina and related syndromes. After a brief description of the major dimensions of normal personality, especially the domain of neuroticism, empirical evidence is reviewed on 1,191 adult men and women who 10 years earlier had made chest pain or discomfort reports part of the National Health and Nutrition Examination Survey. The findings illustrate neuroticism's links to illness and disease. When the mean initial neuroticism levels of the chest pain groups were compared, significantly higher initial levels of neuroticism were found for those who reported any pain or discomfort. As hypothesized, logistic regression results on myocardial infarction death indicated no increased risk due to neuroticism. Neuroticism was related to increased somatic complaints, including chest pain or angina-like complaints, but was not causally or etiologically related to objective signs or pathophysiologic evidence of disease, especially CAD.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of age, hypertension history, and neuroticism on health perceptions.

The effects of measured blood pressure, history of hypertension diagnosis, age, and neuroticism on number of somatic complaints and self-rated health were examined in a sample of 970 non-health-care-seeking adult men and women. Significant differences in number of somatic complaints and self-rated health were found due to age, neuroticism, and history of hypertension diagnosis. Measured blood pressure, however, was unrelated to both measures of health perception. With the exception of the effect of neuroticism on somatic complaints, the effects of the independent variables on health perceptions were rather small in magnitude and explained only small proportions of the variance. Age differences had a particularly weak effect on health perceptions, accounting for less variance than either neuroticism or history of hypertension diagnosis. A significant interaction of neuroticism with awareness of hypertension was found, but only for number of somatic complaints. These results suggest that health perception is a complex, multidimensional construct. The relatively weak influence of hypertension diagnosis on health perception may account for the difficulties in maintaining patient compliance with antihypertensive treatment.

Using college alumni populations in epidemiologic research: the UNC Alumni Heart Study.

The UNC Alumni Heart Study (UNCAHS) is a prospective study of the role of psychosocial factors, in particular hostility, in the development of coronary heart disease. The target population is composed of persons who completed the Minnesota Multiphasic Personality Inventory while attending the University of North Carolina in the mid-1960s. Logistic regression analyses were used to determine whether hostility, demographic and other variables were significant determinants of the subjects' locatability and participation. It was found that MMPI hostility scores at initial testing were unrelated to either potential or actual locatability or participation. Thus there is no evidence that hostility is the source of selection bias in the UNCAHS. Selection into the study was predicted by age, sex, degree status and variables concerned with the conditions under which the MMPI was administered. It is concluded that follow-up studies of college cohorts may have study-specific sources of selection bias.

Covariance of personality, neurocognition, and schizophrenia spectrum traits in the community.

Clinical researchers have observed in relatives of schizophrenic individuals abnormal personality traits resembling the psychopathology of schizophrenia. Further similarities have been observed in correlations between measures of brain function, including attention and executive abilities, and these personality psychopathologies. However, two methodologic factors might account for the covariation of these 'schizophrenia spectrum' personality traits and measures of brain function. Clinical selection bias (Berkson's bias) might result in subjects with overlapping conditions being more likely to be studied, and normal personality attributes could affect performance on neurobehavioral tasks. This study investigated relationships between neurobehavioral correlates of schizophrenia, clinical schizophrenia spectrum personality traits, and normal personality dimensions in the five-factor model of personality. To avoid Berkson's bias, subjects expected to have a high probability of spectrum traits were recruited from the Baltimore Epidemiologic Catchment Area Survey community sample. About 40% of the sample were found to have DSM-IIIR Schizotypal, Schizoid, or Paranoid Personality Traits or Disorders. Schizophrenia Spectrum traits showed significant associations with personality dimensions of the five factor model, particularly Openness to Experience and Neuroticism. In ordinary linear regression models, after adjustment for a number of normal personality characteristics, Schizotypal Personality Traits were still strongly associated with perseverative responses on the Wisconsin Card Sorting Test (WCST). In logistic regression models, subjects with Schizotypal, Schizoid, or Paranoid Traits differed in terms of normal personality profiles and WCST performance.

Relationship of trait anger to resting blood pressure: a meta-analysis.

A series of meta-analyses were conducted to assess whether anger is related to essential hypertension. The present review also considered the relevance of the distinction between anger experience and anger expression, the effect of participant selection bias, and the white-coat hypertension effect for the anger-blood pressure (BP) association. Anger experience was correlated with elevated BP, but the relationship was small and highly variable. When positive effects emerged, both participant selection and the reliability of BP measurement posed interpretational problems. Persons high in anger are not merely exhibiting elevated BP in response to testing, so a white-coat effect is not evident. Being labeled as hypertensive may contribute to higher anger scores, however. The review suggests lines of future research concerning associations between trait anger and blood pressure.

Potential for hostility and dimensions of anger.

Recent reviews have linked Potential for Hostility derived from the Structured Interview (SI) to coronary artery disease, independent of the global Type A pattern. The present study examined the construct validity of Potential for Hostility ratings by correlating Potential for Hostility with 21 scales from four widely used anger/hostility measures: 7 scales from the Anger Self-Report, 8 scales from the Buss-Durkee Hostility Inventory, the total score from the Novaco Anger Inventory, and 5 scales from the Multidimensional Anger Inventory. The pattern of correlations revealed that Potential for Hostility was significantly related to scales reflecting awareness and arousal of anger, particularly the verbal expression of anger. To identify underlying anger dimensions, the 21 scales were factor-analyzed. Examination of two and three rotated principal components confirmed previous solutions. The first component, representing anger-arousing and -eliciting situations and anger awareness, was labeled Experience of Anger. The second component, consisting of scales dealing with either physical assault or verbal expression of anger, was labeled Expression of Anger. When a third factor was retained, it contained scales of suspicion, mistrust-suspicion, and guilt: It was therefore labeled Suspicion-Guilt. Potential for Hostility was correlated only with the Expression of Anger factor in the two-factor solution; in the three-factor solution, Potential for Hostility was correlated equally with the Experience of Anger and Expression of Anger factors but was not correlated with the Suspicion-Guilt factor. The implications of these results for the assessment of hostility are discussed.

Does the Illness Behavior Questionnaire measure abnormal illness behavior?

Abnormal illness behavior (AIB) has been proposed as a construct measuring the inappropriate or maladaptive modes of responding to one's state of health, and the Illness Behavior Questionnaire (BQ; Pilowsky, 1975) was designed to measure this construct. Previous studies using small samples have failed to agree on the factor structure of this questionnaire. The present paper examines the factor structure of the Illness Behavior Questionnaire and critically evaluates the interpretation of its dimensions as well as the construct of AIB. A factor analysis of responses from 1,061 health care and nonhealth care seeking subjects yielded six interpretable factors which substantially replicated Pilowsky's previous results. Six scales were calculated and correlated with several personality measures. The results indicated that the Illness Behavior Questionnaire is saturated with neuroticism, a dimension known to be related to excessive medical complaints. But excessive medical complaints cannot be equated with hypochondriasis or AIB in the absence of objective medical information. In the absence of evidence for the discriminant validity of the IBQ, its use as a diagnostic device is unwarranted. Treating elevated IBQ scores as indicators of abnormal illness behavior without corroborating medical information may be more misleading than accepting patients' symptom reports at face value.

Social learning influences on menstrual symptoms and illness behavior.

The contribution of learning to the adult experience of illness was investigated by asking 351 nursing students how their mothers reacted to menstrual symptoms and cold symptoms during their adolescence and how their mothers behaved when they themselves had menstrual symptoms. Mothers of respondents were independently asked the same questions. Nursing students who had been encouraged to adopt a sick role for menses or whose mothers modeled menstrual distress reported significantly more menstrual symptoms, clinic visits, and disability days for these symptoms as adults. Similarly, those encouraged to adopt a sick role for colds or who lived with a chronically ill person reported more clinic visits and disability days for nongynecological symptoms. Specific types of symptom reports and health care appeared to be learned: Encouragement and modeling of the menstrual sick role were more highly correlated with symptom reports, clinic visits, and absenteeism for menstrual than for nongynecological symptoms, and encouragement of the cold sick role was more highly correlated with nongynecological than with menstrual clinic visits and absenteeism.

Personality and personality disorders.

The mental disorders that most clearly relate to personality are the personality disorders. The purpose of this article is to review the support for the hypothesis that the personality disorders of the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; American Psychiatric Association, 1987) represent variants of normal personality traits. We focus in particular on the efforts to identify the dimensions of personality that may underlie the personality disorders. We then illustrate the relationship of personality to personality disorders using the five-factor model, discuss conceptual issues in relating normal and abnormal personality traits, and consider methodological issues that should be addressed in future research.

Depressive symptoms as a nonspecific, graded risk for psychiatric diagnoses.

Using data from a 16-year follow-up of a nationally representative sample of 6,913 adults, measures of depressive symptoms were used to predict psychiatric diagnoses taken from hospitalization records. In proportional hazards analyses, two measures of depression were significantly associated with subsequent diagnoses of depression and other psychiatric disorders after statistical control for demographic variables and previous history of psychological problems. Depressive symptoms predicted late as well as early occurrence of psychiatric diagnoses and showed a pattern of increasing risk with increasing scores, even below clinical cutoffs. This pattern of results is consistent with the view that depressive symptoms predict future psychiatric disorders largely because they serve as proxy measures of some chronic vulnerability, such as the normal personality dimension of neuroticism.

Validation of the five-factor model of personality across instruments and observers.

Two data sources--self-reports and peer ratings--and two instruments--adjective factors and questionnaire scales--were used to assess the five-factor model of personality. As in a previous study of self-reports (McCrae & Costa, 1985b), adjective factors of neuroticism, extraversion, openness to experience, agreeableness-antagonism, and conscientiousness-undirectedness were identified in an analysis of 738 peer ratings of 275 adult subjects. Intraclass correlations among raters, ranging from .30 to .65, and correlations between mean peer ratings and self-reports, from .25 to .62, showed substantial cross-observer agreement on all five adjective factors. Similar results were seen in analyses of scales from the NEO Personality Inventory. Items from the adjective factors were used as guides in a discussion of the nature of the five factors. These data reinforce recent appeals for the adoption of the five-factor model in personality research and assessment.

Personality in adulthood: a six-year longitudinal study of self-reports and spouse ratings on the NEO Personality Inventory.

Previous longitudinal studies of personality in adulthood have been limited in the range of traits examined, have chiefly made use of self-reports, and have frequently included only men. In this study, self-reports (N = 983) and spouse ratings (N = 167) were gathered on the NEO Personality Inventory (Costa & McCrae, 1985b), which measures all five of the major dimensions of normal personality. Cross-sectional and longitudinal analyses on data from men and women aged 21 to 96 years showed evidence of small declines in Activity, Positive Emotions, and openness to Actions that might be attributed to maturation, but none of these effects was replicated in sequential analyses. The 20 other scales examined showed no consistent pattern of maturational effects. In contrast, retest stability was quite high for all five dimensions in self-reports and for the three dimensions measured at both times in spouse ratings. Comparable levels of stability were seen for men and women and for younger and older subjects. The data support the position that personality is stable after age 30.

Primary traits of Eysenck's P-E-N system: three- and five-factor solutions.

The Eysenck Personality Profiler (EPP), a set of 21 scales measuring primary traits hypothesized to be definers of Neuroticism, Extraversion, and Psychoticism factors, was administered to 229 adults together with the Eysenck Personality Questionnaire--Revised (H. J. Eysenck & S. B. G. Eysenck, 1991) and the Revised NEO Personality Inventory (NEO-PI-R; P. T. Costa & R. R. McCrae, 1992b). Correlations of EPP scales with NEO-PI-R facet scales provided preliminary evidence supporting the convergent and discriminant validity of the EPP scales. However, varimax and targeted validimax factor analyses suggested that some EPP scales were misclassified and that EPP scales could better be understood in terms of the 5-factor model than the intended 3-factor model.