Genetic and environmental contributions to depressive personality disorder in a population-based sample of Norwegian Twins

Journal of Affective Disorders 99 (2007) 181 – 189 www.elsevier.com/locate/jad

Research report

Genetic and environmental contributions to depressive personality disorder in a population-based sample of Norwegian Twins Ragnhild Elise Ørstavik a,⁎, Kenneth S. Kendler c,d , Nikolai Czajkowski a , Kristian Tambs a , Ted Reichborn-Kjennerud a,b a

d

Department of Mental Health, Norwegian Institute of Public Health, Norway b Institute of Psychiatry, University of Oslo, Norway c Virginia Institute for Psychiatric and Behavioral Genetics, Medical College of Virginia/Virginia Commonwealth University, Richmond, VA, USA Departments of Psychiatry and Human Genetics, Medical College of Virginia/Virginia Commonwealth University, Richmond, VA, USA Received 26 April 2006; received in revised form 5 September 2006; accepted 6 September 2006 Available online 17 October 2006

Abstract Background: Depressive personality disorder (DPD) was introduced in DSM-IV as a new category requiring further study. The aim of this study was to estimate genetic and environmental contributions to DPD in a population-based twin sample, and include data on criteria performance, prevalence and diagnostic overlap. Methods: Axis I and Axis II diagnoses were obtained by structured interviews in a population-based sample of 2794 young adult twins. Statistical analyses included correlation and factor analysis based on polychoric correlation coefficients, and diagnostic overlap applying adjusted odds ratios. Contributions from additive genetic and common and unique environmental influences to the liability to DPD were computed using structural equation modelling, applying a multiple threshold variable. Results: Liability to DPD could best be explained by additive genetic and unique environmental factors, with heritability estimates of 49% (95% CI 0.41–0.57) in females and 25% (95% CI 0.12–0.40) in males. The best-fitting model indicated that some of the genes contributing to DPD differ between men and women. Chronbach's alpha was 0.87. 2.0% of participants fulfilled the criteria for DPD, and overlap was most pronounced for dysthymic disorder and avoidant personality disorder. Limitations: Low prevalence rates and subsequent inclusion of subthreshold criteria could have influenced parameter estimates, especially in males. Conclusions: DPD was almost twice as heritable in females as in males, comparable to previous studies on major depression. The proposed criteria showed good measurement properties, and DPD was not completely subsumed within any other disorder. © 2006 Elsevier B.V. All rights reserved. Keywords: Depressive personality disorder; Personality; Heritability; Twin study

1. Introduction

⁎ Corresponding author. Norwegian Institute of Public Health, Box 4404, Nydalen N-0403 Oslo, Norway. Tel.: +47 22 04 22 00. E-mail address: [email protected] (R.E. Ørstavik). 0165-0327/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2006.09.011

The concept of depressive personality was first introduced in psychiatric literature at the beginning of the previous century. Various descriptions have been proposed, but most emphasize traits of gloominess, worry,

182

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189

pessimism and self-criticism (Phillips et al., 1995). Depressive personality disorder (DPD) was, however, first introduced in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (American Psychiatric Association, 1994) as a new category requiring further study. The proposed diagnosis requires the fulfilment of five of seven criteria. Despite previous suggestions that DPD is a common disorder (Klein, 1999), there is little data to support this claim. Only one previous study has reported the prevalence of DSM-IV DPD in a small, non-clinical sample: Ryder et al. (2001) found that 17 of 368 (4.7%) young adults fulfilled the criteria for DPD. To date, we are unaware of any larger, population-based studies on DPD prevalence, and such data are in demand (Ryder et al., 2005). The personality disorder (PD) work group for DSMIV addressed several issues regarding DPD (Phillips et al., 1995), including the reliability of diagnostic tools, criteria performance, and the overlap between DPD and other PDs and mood disorders. Several studies have addressed the latter question; with some concluding that the overlap is too great to justify a diagnostic entity for DPD (Bagby et al., 2003; Ryder et al., 2002), and others the opposite (Klein and Shih, 1998; Markowitz et al., 2005; McDermut et al., 2003; Phillips et al., 1998). The conflicting results demonstrate that this problem is far from solved and justify further studies. An important diagnostic validation criterion is that of familial aggregation (Robins and Guze, 1970). However, the number of twin studies on PDs and dimensionally defined PD traits is limited: Results from a Norwegian study based on clinical samples showed heritabilities for DSM-III-R PDs in the range of 0.28–0.78, but did not include DPD (Torgersen et al., 2000). In a study on dimensionally defined PD traits, Jang et al. (1996) found heritabilities ranging from 0.38 to 0.48 for facet scales relevant to DPD such as anhedonia, pessimism and guilt proneness. Numerous studies have shown that DSM PDs can be predicted by the five-factor model for personality (NEO-PIR) (Costa and McCrae, 1992), for review see (Saulsman and Page, 2004; Widiger and Costa, 2002). DPD correlates highly (0.50–0.75) with both the higherorder trait of Neuroticism and the lower order trait of Depression (Dyce and O'Connor, 1998), which again has been shown to be heritable (Jang et al., 1998). We are unaware of any prior family, twin or adoption study of DPD. The main aim of this study was to explore the relative contribution of genetic and environmental factors to the liability of DSM-IV DPD in a large, population-based sample of Norwegian twins, including possible sex differences. We also estimated the preva-

lence of DPD and its co-occurrence with Axis I mood disorders and other Axis II disorders to address questions about the need for an independent diagnosis of DPD. 2. Methods 2.1. Sample Subjects included in this study were recruited from The Norwegian Institute of Public Health Twin Panel (NIPHTP). NIPHTP consists of twins identified through information in the national Medical Birth Registry, established January 1, 1967, which receives mandatory notification of all live and stillbirths of at least 16 weeks gestation. The current panel includes information on 15,370 like and unlike sexed twins born from 1967– 1979. During that time period the percentage of pairs for which both twins survived to age 3 ranged from 82% to 89%. Two questionnaire studies have been conducted; in 1992 (twins born 1967–1974) and in 1998 (twins born 1967–1979). Altogether, 12,700 twins received the second questionnaire, and 8045 responded after one reminder (response rate 63%). The sample included 3334 pairs and 1377 single responders. The NIPHTP is described in detail elsewhere (Harris et al., 2002). Data for the present report derive from an interview study of Axis I and Axis II Psychiatric Disorders. Participants were recruited among 3153 complete pairs who in the second questionnaire agreed to participate in the interview study, and 68 pairs who were drawn directly from NIPHTP. Of these 3221 eligible pairs, 0.8% was unwilling or unable to participate, and in 16.2% of pairs only one twin agreed to the interview. 38.2% did not respond after two contacts requesting participation (the maximum number of contacts allowed in the licence obtained from The Regional Committee for Medical Research Ethics). Altogether 2794 twins (44% of those eligible) were interviewed for the assessment of PDs. Our final sample consisted of 1022 males and 1772 females; 221 monozygotic male (MZM) pairs, 116 dizygotic male (DZM) pairs, 448 monozygotic female (MZF) pairs, 261 dizygotic female (DZF) pairs, 340 dizygotic opposite sex (DZO) pairs and 22 single responders. The mean age of attendants was 28.2 years (range 19–36). Approval was received from The Norwegian Data Inspectorate and The Regional Committee for Medical Research Ethics, and written informed consent was obtained from all participants after complete description of the study. Zygosity was initially determined by questionnaire items previously shown to categorize correctly 97.5% of

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189

183

pairs (Magnus et al., 1983). In all but 385 like-sexed pairs, zygosity was also determined by molecular methods based on the genotyping of 24 microsatellite markers. 17 of these pairs with DNA information (2.5%) were found to be misclassified by the questionnaire data and were corrected. From these data, we estimate that in our entire sample, zygosity misclassification rates are less than 1%, a rate which is unlikely to substantially bias results (Neale, 2003a).

reasons were done over the telephone. Each twin in a pair was interviewed by different interviewers. Inter-rater reliability was assessed by two raters scoring 70 audiotaped interviews. The number of subjects with categorically classified PDs was too low to calculate Kappa coefficients. Instead, intraclass correlations for the scaled PDs were used. The correlation for DPD was +0.96 if subthreshold criteria were included and +0.78 if a positive criterion was defined as a score of ≥2.

2.2. Measurements

2.3. Statistics

The Structured Interview for DSM-IV Personality (SIDP-IV) (Pfohl et al., 1995), a comprehensive semistructured diagnostic interview, was used for the assessment of all DSM-IV axis II diagnoses. The instrument includes non-pejorative questions organized into topical sections (e.g. “social relationships”, “work style”, “emotions”) rather than disorders. This allows for a more natural flow of the interview and increases the likelihood that useful information from related questions may be taken into account when rating criteria within that section. The specific DSM IV criteria associated with each set of questions is rated as follows: 0 = “not present or limited to rare isolated examples”, 1 = “subthreshold — some evidence of the trait, but it is not sufficiently pervasive to consider the trait present”, 2 = “present — the trait is clearly present for most of the last five years, 3 = “strongly present — the trait is associated with subjective distress or some impairment in social or occupational functioning, or intimate relationships”. The interview uses the “five year rule”; behaviors, cognitions and feelings that predominated for the last 5 years are considered representative of the individual's long-term personality functioning. A Norwegian version of the computerized Composite International Diagnostic Interview M-CIDI was used to assess Axis I disorders (Wittchen and Pfister, 1997). This comprises all DSM-IV axis I diagnoses and ICD-10 diagnoses (lifetime and 12 month prevalence). Both SIDP and CIDI have been used in numerous previous studies including some from Norway (Kringlen et al., 2001; Torgersen et al., 2001). Interviewers were mostly psychology students in the final part of their training and experienced psychiatric nurses, and were trained by professionals (one psychiatrist and two psychologists) with extensive experience with the instrument. The interviewers were followed up closely individually during data collection, and regular meetings were held with all interviewers to discuss potential problems. Interviews were carried out between June 1999 and May 2004. Most were conducted face-toface, except for 231 interviews (8.3%) that for practical

Internal consistency of the DPD criteria were calculated using Chronbach's α and confirmatory factor analysis (seven criteria loading on one common factor), both based on polychoric correlation coefficients. Categorical diagnoses of DPD (five or more criteria above threshold (≥2)) were used for prevalence estimates and evaluation of co-occurrence with other diagnoses. Odds ratios (OR)s were calculated applying Generalized Estimating Equations with alternating logistic regressions in SAS (version 9.1) (SAS Institute, 2005), to adjust for the non-independency of twin pairs. Twin analyses on categorical data are based on the liability-threshold model, which assumes that the liability for a categorical characteristic is continuously and normally distributed in the population. Individuals who exceed a theoretical threshold express the disorder. In this population-based sample, the prevalence rates for categorically defined diagnoses were too low to permit reliable twin analyses. To increase statistical power, we included sub-threshold criteria (N 1) when defining a positive trait. This assumes that the classification (0–3) represent different degrees of severity on a single dimension, and can be modelled as a bivariate normal density. We tested the single liability threshold model (Vink et al., 2005) for each zygosity group. The model fit well (i.e., p ≥ .05) in all but two of the 35 tests (5 × 7), a pattern consistent with chance expectations (Feild and Armenakis, 1974). Secondly, we used a dimensional approach, constructing an ordinal variable from the number of the seven DPD criteria. Because the number of subjects who endorsed most of the DPD criteria was small, we collapsed the upper categories of the summed score (4–7 criteria fulfilled) resulting in an ordinal variable with five categories. To test the assumption that the number of positive criteria for DPD represents different degrees of severity of the disorder, we applied the same test as described above. The model fit well in all five analyses. In the basic twin model, the variation in an observed trait is assumed to arise from individual differences in

184

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189

liability caused by additive genetic influences (A), common or shared environmental factors (C) and individualspecific or unique environmental factors (E). The relative differences in these latent factors can be derived from the different genetic correlations in MZ and DZ twins: A is perfectly correlated in MZ twins, because they are genetically identical. DZ twins share on average 50% of their segregating genes, giving a correlation of 0.5. Common environmental factors are environmental factors that make twins similar, and have a correlation of 1.0 in both zygosity groups. Unique environmental factors are, by definition, those that make twins differ from each other, and include random error of measurement. The raw data option in the structural equation modelling program Mx (Neale et al., 2003b) was used for estimation of polychoric correlations, model fitting and for testing homogeneity of thresholds within twin pairs and across zygosity groups and gender. Genetic and environmental influences on DPD traits were estimated using maximum likelihood estimation (MLE). This option in Mx does not provide an overall test of goodness-of-fit for the model, but alternative models can be compared by their different chi-squares relative to the difference in their degrees of freedom (df ). According to the principle of parsimony, models with fewer parameters are preferable if they do not provide significantly worse fit. An alternative method for model comparison that combines parsimony with explanatory power is Akaike's Information Criterion (AIC) (Akaike, 1987), calculated as: χ2 − 2df. In twin models which include all five zygosity groups including DZ opposite sex (DZO) twins, sex differences in genetic and environmental effects can be explored. These are addressed by specifying different parameters

for males and females and a correlation between their genetic effects (rg). In models that explore both qualitative and quantitative genetic differences (some or all of the genes that influence the liability to a trait or disorder differ between men and women), rg is estimated freely. In quantitative models (the same genes influence the liability to a trait or disorder, but the effect of these genes differs in magnitude), rg is fixed to 1. These models are then compared to models with no sex effects (parameters for males and females are constrained to be equal and rg = 1). Twin studies rely upon the equal environments assumption (EEA), i.e. that MZ and DZ pairs are equally correlated in their exposure to environmental influences on the trait or disorder under study. If the EEA is violated, greater similarity in monozygotic twins could be due to increased environmental similarity instead of greater genetic similarity (Rose et al., 1990). The dependent variable in our analyses of the EEA was the number of endorsed personality disorder traits (PDT) in twin 2. Two variables that reflected, respectively, similarity of childhood and adult environments were constructed: The similarity of childhood environment was indexed by two items that assessed the years that the twins were in the same class at school and the years the twins lived in the same residence, and adult environment by three items that inquired about the frequency of in person and telephone contact during the last year and the distance between their current residences. We controlled for main effects of zygosity, sex, age and level of environmental similarity, and added shared environmental effects (the main effect of PDT score in twin 1 on twin 2) and genetic effects (interaction of PDT score in twin 1⁎zygosity to predict PDT score in twin 2). We

Table 1 Frequency (%) of depressive personality disorder traits (score 0–3) in 1022 male 1772 female twins Trait

0 M

F

M

F

M

F

M

F

1. Usual mood is dominated by dejection, gloominess, cheerlessness, joylessness, unhappiness 2. Self-concept centers around beliefs of inadequacy, worthlessness, and low self-esteem 3. Is critical, blaming, and derogatory toward self

940 (92.0) 898 (87.9) 838 (82.0) 899 (88.0) 721 (70.5) 864 (84.5) 794 (77.7)

1575 (88.9) 1377 (77.7) 1253 (70.7) 1331 (75.7) 1384 (78.1) 1351 (76.2) 1227 (69.2)

62 (6.1) 94 (9.2) 118 (11.5) 81 (7.9) 220 (21.5) 119 (11.6) 142 (13.9)

117 (6.6) 255 (14.4) 303 (17.1) 227 (12.8) 301 (17.0) 272 (15.3) 307 (17.3)

17 (1.7) 24 (2.3) 47 (4.6) 30 (2.9) 67 (6.6) 25 (2.4) 62 (6.1)

153 (2.9) 95 (5.4) 140 (7.9) 114 (6.4) 81 (4.6) 101 (5.7) 151 (8.5)

3 (0.3) 6 (0.6) 19 (1.9) 12 (1.2) 14 (1.4) 14 (1.4) 24 (2.3)

28 (1.6) 45 (2.5) 76 (4.3) 190 (5.1) 6 (0.3) 148 (2.7) 87 (4.9)

4. Is brooding and given to worry 5. Is negativistic, critical, and judgmental towards others 6. Is pessimistic 7. Is prone to feeling guilty or remorseful

1

2

3

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189

185

Table 2 Comorbidity (n (%)) between depressive personality disorder (DPD) and other personality and mood disorders in a population-based sample of 2794 twins

3. Results

Diagnosis

Chronbach's α for the seven diagnostic criteria was 0.87. In confirmatory factor analysis, six of the seven traits loaded on the common factor with values ranging from 0.32 to 0.60. However, the trait “negativistic, critical and judgmental towards others” (criterion five, Table 1) loaded only + 0.09, indicating a weak relationship to the overall DPD construct.

Without DPD (2739)

Personality disorders Paranoid 11 (0.4) Schizoid 1 (0.0) Schizotypal 1 (0.0) Any cluster Aa 13 (0.5) Antisocial 9 (0.3) Borderline 5 (0.2) Histrionic 2 (0.1) Narcissistic 0 (0.0) Any cluster Ba 15 (0.5) Avoidant 36 (1.3) Dependent 4 (0.1) Obsessive– 63 (2.3) compulsive Any cluster Ca 99 (3.6) Any PD other than 120 (4.4) DPDa Mood disordersb Dysthymia Major depressive disorder a b

31 (1.1) 365 (13.3)

With DPD (55)

OR (95% CI)

3 (5.5) 2 (3.6) 0 (0.0) 4 (7.3) 0 (0.0) 6 (10.9) 0 (0.0) 1 (1.8) 6 (10.9) 23 (41.8) 3 (5.5) 6 (10.9)

15.9 (4.4–57.8) 69.0 (9.8–484.6) 15.5 (5.0–47.9) 61.6 (17.2–221.0)

21.6 (7.5–61.9) 54.4 (30.2–98.0) 40.8 (9.0–185.7) 4.7 (1.7–12.9)

27 (49.1) 29 (52.7)

25.1 (14.4–43.6) 23.7 (13.4–41.8)

17 (30.9) 28 (50.9)

36.6 (18.8–71.3) 6.7 (3.9–11.3)

Numbers do not add up, as some patients have more than one PD. Lifetime diagnosis.

controlled for the correlational structure of our data in these analyses using independent estimating equations as operationalized in the SAS procedure GENMOD (SAS Institute, 2005). We tested whether the DPD trait score in Twin 1 interacted with the measure of contact to predict DPD trait score in Twin 2. A violation of the EEA would mean that Twin 1 DPD trait score would be a better predictor of Twin 2 DPD trait score in twins with high versus low contact. We found no significant effects of child or adult contact on DPD (interaction terms exceeded 0.10). Table 3 Maximum likelihood estimates of twin correlations of depressive personality disorder subthreshold traits collapsed into five categories Zygosity

Number Polychoric correlation (95% CI) of pairs

Monozygotic males Dizygotic males Monozygotic females Dizygotic females Dizygotic opposite sex

221

0.28 (0.12–0.42)

116 448

0.09 (−0.13–0.30) 0.49 (0.40–0.57)

261 340

0.30 (0.16–0.42) 0.05 (−0.08–0.18)

3.1. Internal consistency of diagnostic criteria

3.2. Prevalence of DPD and comorbidity with Axis I and Axis II disorders Fifty-five subjects (2.0%, 2.8% of females and 0.5% of males, p b 0.001) fulfilled the full criteria for DPD. Of these, one was MZM, 29 MZF, 13 DZF and 12 DZO (six males). There were two concordant pairs among the MZFs and one among the DZFs. Prevalences of the seven criteria (scores 0–3) in male and female attendances are given in Table 1. All criteria except number 5 were more frequent among females than males. Twenty-nine subjects (52.7%) also met the criteria for at least one other PD (OR 24.3, 95% CI 13.9–42.6 compared to those without DPD). The greatest overlap was with avoidant PD (Table 2). Patients fulfilling the diagnostic criteria for DPD had substantial increased risk of having a mood disorder (Table 2), and a lifetime diagnosis of major depressive disorder (MDD). However, most patients with DPD did not meet the criteria for dysthymic disorder (DD). Table 4 Model-fitting results for depressive personality disorder in 2794 twins − 2ll

Δx2

Δdf

p-value

AIC

Qualitative and quantitative sex limitation models I ACE, rg free 8048.78 2781 – – II AE, rg free 8049.44 2783 0.66 2 III CE, rg free 8064.89 2783 16.12 2

– 0.72 0.000

– −3.34 12.12

Quantitative sex limitation models IV ACE, rg = 1 8050.97 2782 V AE, rg = 1 8053.57 2784

2.20 4.60

1 3

0.14 0.20

0.20 − 1.40

No sex limitation models VI ACE, rg = 1 8064.33 8064.31 VII AE, rg = 1

15.56 15.54

4 5

0.004 0.008

7.56 5.56

df

2785 2786

All models are compared to the full ACE model (general sex limitation, rg free). A; additive genetic effects, C; common environmental effects, E; unique environmental effects. rg; correlation between sexes for additive genetic factors, AIC; Akaike's information criterion. Best fitting model in bold.

186

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189

3.3. Twin correlations

effects on DPD (Model II). The parameter estimates for the full and best-fitting model are given in Table 5.

Cross-twin polychoric correlation coefficients with 95% CI, are given in Table 3. Despite relatively wide CIs, the difference between MZ and DZ correlations suggest genetic effects for both genders. The correlation between twins of opposite sexes was lower than average for same sex DZ twins, indicating possible sex differences in genetic liability to DPD traits. 3.4. Model fitting We found no significant differences in thresholds within same-sex pairs or between MZ and DZ twins of same gender in the different zygosity groups. However, a model constraining thresholds to be equal across sexes resulted in a significant deterioration in fit (Δχ2 = 70.10, p b 0.0001). Consequently, models were run with separate thresholds for males and females. Results from the model fitting are summarized in Table 4. The basic (full) model, with which the nested models were compared, was an ACE model specifying different genetic and environmental parameters for the two genders with the correlation between genetic effects in males and females (rg) estimated freely. Constraining rg to be 1 resulted in deterioration in fit (Models IV and V). When model V (AE model with rg = 1) was directly compared to the corresponding AE model with rg estimated freely (Model II), the difference was statistically significant (Δx2 = 3.94, Δdf = 1, p b 0.05). Models that constrained both genetic and environmental parameters to be equal in males and females (models VI and VII) resulted in a significant deterioration in fit, and were rejected. The best fitting model according to AIC included additive genetic and unique environmental paths only, and allowed for both quantitative and qualitative sex-differences in genetic Table 5 Parameter estimates and 95% confidence intervals for full and best fit models for depressive personality disorder in 2794 twins

Males

Females

Full model

Best fit

rg = 0.27 (0.00–1.00)

rg = 0.23 (0.00– 0.99)

A

C

E

A

E

0.26 (0.00– 0.40) 0.37 (0.06– 0.56)

0.00 (0.00– 0.30) 0.12 (0.00– 0.38)

0.74 (0.60– 0.88) 0.52 (0.44– 0.66)

0.26 (0.12– 0.40) 0.49 (0.41– 0.57)

0.74 (0.60– 0.88) 0.51 (0.43– 0.59)

Abbreviations: See Table 4.

4. Discussion To our knowledge, this is the first population-based genetic study on DPD as defined in DSM-IV. The criteria showed good inter-rater reliability and internal consistency. Overlap with other PDs and mood disorders were substantial, but DPD was not completely subsumed in any other diagnosis. Genetic factors accounted for approximately 50% and 25% of the variance of DPD in females and in males, with both quantitative and qualitative sexdifferences. In a previous study on criteria performance for all DSM-IV PDs, Farmer and Chapman (2002) found a Kappa for inter-rater reliability of 0.94 for DPD, and an internal consistency (α) of 0.81. We used intraclass and polychoric correlations which gives slightly higher estimates than Kappa scores and Pearson's correlation, but the results were still fairly similar. Our finding that the fifth criterion (critical towards others) relates poorly to the overall DPD construct is consistent with the findings of Farmer et al (Farmer and Chapman, 2002). The prevalence of DPD in this sample was 2.0%. To our knowledge, prevalence of DPD by DSM IV criteria has only been addressed in one previous study: Among 368 students participating in an interview study to validate the criteria for DPD and it's overlap with dysthymic disorder, 17 (4.6%) was diagnosed with DPD (Ryder et al., 2001). Despite differences in sample sizes and selections, our results indicate that DPD is less common than previously assumed (Klein, 1999; Ryder et al., 2001). The overlap between DPD and other personality disorders was substantial, and most pronounced for avoidant PD. However, most patients with DPD did not meet the criteria for avoidant PD. These results are similar to those found in a previous study based on a clinical sample (Markowitz et al., 2005). A considerable proportion of the patients fulfilling the criteria for DPD also met criteria for DD. However, the overlap was in the lower range of those found previously (Klein and Shih, 1998; Markowitz et al., 2005; McDermut et al., 2003; Phillips et al., 1998; Ryder et al., 2001, 2006). We also found a clear correlation between DPD and a lifetime diagnosis of major depressive disorder. The direction and cause of these relationships need to be explored in further studies. The results from our heritability estimates indicate a substantial genetic contribution to DPD in females and a more modest genetic influence in males. As this has

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189

not been addressed in previous studies, comparisons are restricted to studies of PD traits and dimensions on normal personality dimensions: In a Canadian study based on questionnaires, the authors examined genetic and environmental contributions to 18 basic dimensions of PDs (Livesley et al., 1993), and subsequently included 69 facet traits (Jang et al., 1996). The heritability estimates for the traits closest to those in the proposed DPD diagnosis (anhedonia, pessimism, guilt proneness and trait anxiety) was 0.40–0.50 (Jang et al., 1996). These numbers correspond closely to the heritability estimate for females in the present study (0.49) (Table 5). DPD can be predicted by the five-factor model for normal personality (Widiger and Costa, 2002): In a study of 614 undergraduates, DPD was shown to correlate 0.72 with the higher-order trait Neuroticism and 0.75 with the lower-order facet score Depression (Dyce and O'Connor, 1998). Correlations between DPD and three of the other lower-order Neuroticism facets were above 0.50. In a cross-cultural study, examining heritability of the residual specific variance in facet-level traits based on the five factor model using the NEO-PIR) (Costa and McCrae, 1992), heritability for the Depression facet score was found to be 0.44 (Jang et al., 1998) — again close to those for DPD in the present study. Consistent with other studies of normal and abnormal personality, we found little evidence for shared environmental influences on the liability to DPD. However, the power of the classical twin study when using ordinal data compared to continuous data is limited (Neale et al., 1994). Small effects (b10–15%) are difficult to detect with realistic sample sizes, even when including subthreshold values. Our results indicate both quantitative and qualitative gender differences in the heritability of DPD traits. This should be interpreted with some caution as the prevalence in males was very low, and the confidence intervals wide. Nonetheless, our findings are interesting in light of corresponding findings with regards to mood disorders. Studies on gender differences in the heritability of major depression and depressive symptoms have shown conflicting results (Agrawal et al., 2004; Kendler et al., 2001; McGuffin et al., 1996). However, the most powerful study to date (Kendler et al., 2001) found evidence for both qualitative and quantitative gender differences, and these results have recently been replicated in a Swedish twin group of more than 15,000 complete pairs (Kendler et al., 2006). The results from the present study suggest that trait-like aspects of depression show the same pattern of heritability as depressive states.

187

The results from our study must be interpreted in light of several limitations: First, because of the low prevalence of DPD, we did not examine the fully syndromal versions of the diagnosis. Instead we examined a number of criteria using a low threshold of endorsement. Since twin analysis are based on the single liability threshold model, it should in principle make no difference if the variable studied is categorical or dimensional as long as the multiple threshold test indicates that this variable reflects the same liability that underlay the categorical diagnosis. Regarding the use of subthreshold criteria, we tested the results of the heritability analyses applying threshold criteria only. In females, where the prevalence rate was sufficiently high to perform the analysis, the results were comparable to those obtained including subthreshold criteria (results available on request). Furthermore, many have argued that personality disorders are best conceptualized as dimensional rather than dichotomous constructs (Skodol et al., 2005). Secondly, these results were obtained from a particular sample of young Norwegians, and may or may not be extrapolated to other age cohorts or ethnic groups. The participants in our sample were twins, and the results cannot necessarily be generalized to the general population. However, a previous study on personality in twins failed to show any systematic differences compared to non-twin samples (Johnson et al., 2002). Third, although we included a large number of twins, substantial attrition was observed in this sample from the birth registry through three waves of contact consisting of two questionnaires and a personal interview. We report detailed analyses of the predictors of non-response across waves elsewhere (Harris et al, in preparation). Briefly, cooperation was strongly and consistently predicted by female sex, monozygosity, older age, and higher educational status, but not symptoms of mental disorder. In particular, we assessed personality disorder traits at the second questionnaire with 91 self-report items. We used these items to predict the PD score in the interview. The polychoric correlation between the scores based on the questionnaire and those based on the interview for DPD was 0.54. Controlling for demographic variables, the weighted score from the second wave questionnaire did not predict participation in the personal interview ( p = 0.33). Finally, twins were interviewed only once. Although we demonstrated high inter-rater reliability, the test–retest reliability might be considerably lower (Markowitz et al., 2005; Zanarini et al., 2000). In twin analyses, measurement errors are reflected in E, so an improved diagnostic reliability could result in higher heritability estimates. In spite of these limitations, this study contributes important findings with regards to previously raised

188

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189

questions on the placement of DPD in DSM-V. Further studies are needed to explore to which extent the overlap between DPD and related Axis II and Axis I disorders are due to common genetic or environmental factors. Acknowledgements The work was supported in part by grants MH-068643 from the National Institutes of Health, and by grants from The Norwegian Research Council, The Norwegian Foundation for Health and Rehabilitation, The Norwegian Council for Mental Health, The European Commission under the program 'Quality of Life and Management of the Living Resources' of 5th Framework Program (no. QLG2-CT-2002-01254). Genotyping on the twins was performed at the Starr Genotyping Resource Centre at the Rockefeller University. We are very thankful to the twins for their cooperation. References Agrawal, A., Jacobson, K.C., Gardner, C.O., Prescott, C.A., Kendler, K.S., 2004. A population based twin study of sex differences in depressive symptoms. Twin Res. 7, 176–181. Akaike, H., 1987. Factor-analysis and AIC. Psychometrica 52, 317–332. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. American Psychiatric Association, Washington DC. Bagby, R.M., Ryder, A.G., Schuller, D.R., 2003. Depressive personality disorder: a critical overview. Curr. Psychiatry Rep. 5, 16–22. Costa, P.T., McCrae, R.R., 1992. Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual. Psychological Assessment Resourches, Odessa, FI. Dyce, J.A., O'Connor, B.P., 1998. Personality disorders and the five-factor model: a test of facet-level predictions. J. Pers. Disord. 12, 31–45. Farmer, R.F., Chapman, A.L., 2002. Evaluation of DSM-IV personality disorder criteria as assessed by the structured clinical interview for DSM-IV personality disorders. Compr. Psychiatry 43, 285–300. Feild, H.S., Armenakis, A.A., 1974. On use of multiple tests of significance in psychological research. Psychol. Rep. 35, 427–431. Harris, J.R., Magnus, P., Tambs, K., 2002. The Norwegian Institute of Public Health Twin Panel: a description of the sample and program of research. Twin Res. 5, 415–423. Jang, K.L., Livesley, W.J., Vernon, P.A., Jackson, D.N., 1996. Heritability of personality disorder traits: a twin study. Acta Psychiatr. Scand. 94, 438–444. Jang, K.L., McCrae, R.R., Angleitner, A., Riemann, R., Livesley, W.J., 1998. Heritability of facet-level traits in a cross-cultural twin sample: support for a hierarchical model of personality. J. Pers. Soc. Psychol. 74, 1556–1565. Johnson, W., Krueger, R.F., Bouchard Jr., T.J., McGue, M., 2002. The personalities of twins: just ordinary folks. Twin Res. 5, 125–131. Kendler, K.S., Gardner, C.O., Neale, M.C., Prescott, C.A., 2001. Genetic risk factors for major depression in men and women:

similar or different heritabilities and same or partly distinct genes? Psychol. Med. 31, 605–616. Kendler, K.S., Gatz, M., Gardner, C.O., Pedersen, N.L., 2006. A Swedish national twin study of lifetime major depression. Am. J. Psychiatry 163, 109–114. Klein, D.N., 1999. Depressive personality in the relatives of outpatients with dysthymic disorder and episodic major depressive disorder and normal controls. J. Affect. Disord. 55, 19–27. Klein, D.N., Shih, J.H., 1998. Depressive personality: associations with DSM-III-R mood and personality disorders and negative and positive affectivity, 30-month stability, and prediction of course of Axis I depressive disorders. J. Abnorm. Psychology 107, 319–327. Kringlen, E., Torgersen, S., Cramer, V., 2001. A Norwegian psychiatric epidemiological study. Am. J. Psychiatry 158, 1091–1098. Livesley, W.J., Jang, K.L., Jackson, D.N., Vernon, P.A., 1993. Genetic and environmental contributions to dimensions of personality disorder. Am. J. Psychiatry 150, 1826–1831. Magnus, P., Berg, K., Nance, W.E., 1983. Predicting zygosity in Norwegian twin pairs born 1915–1960. Clin. Genet. 24, 103–112. Markowitz, J.C., Skodol, A.E., Petkova, E., Xie, H., Cheng, J., Hellerstein, D.J., Gunderson, J.G., Sanislow, C.A., Grilo, C.M., McGlashan, T.H., 2005. Longitudinal comparison of depressive personality disorder and dysthymic disorder. Compr. Psychiatry 46, 239–245. McDermut, W., Zimmerman, M., Chelminski, I., 2003. The construct validity of depressive personality disorder. J. Abnorm. Psychol. 112, 49–60. McGuffin, P., Katz, R., Watkins, S., Rutherford, J., 1996. A hospitalbased twin register of the heritability of DSM-IV unipolar depression. Arch. Gen. Psychiatry 53, 129–136. Neale, M.C., 2003a. A finite mixture distribution model for data collected from twins. Twin Res. 6, 235–239. Neale, M.C., Boker, S.M., Xie, G., Maes, H.H., 2003b. Mx: Statistical Modeling. Dept. of Psychiatry, Virginia Commonwealth University Medical School: Box 980126, Richmond, VA 23298. Neale, M.C., Eaves, L.J., Kendler, K.S., 1994. The power of the classical twin study to resolve variation in threshold traits. Behav. Genet. 24, 239–258. Pfohl, B., Blum, N., Zimmerman, M., 1995. Structured Interview for DSM-IV Personality (SIDP-IV). University of Iowa, Department of Psychiatry, Iowa City. Phillips, K.A., Hirschfeld, R.M., Shea, M.T., Gunderson, J.G., 1995. Depressive personality disorder. In: Livesley, W.J. (Ed.), The DSM-IV Personality Disorders. The Guilford Press, New York, pp. 287–302. Phillips, K.A., Gunderson, J.G., Triebwasser, J., Kimble, C.R., Faedda, G., Lyoo, I.K., Renn, J., 1998. Reliability and validity of depressive personality disorder. Am. J. Psychiatry 155, 1044–1048. Robins, E., Guze, S.B., 1970. Establishment of diagnostic validity in psychiatric illness: its application to schizophrenia. Am. J. Psychiatry 126, 983–987. Rose, R.J., Kaprio, J., Williams, C.J., Viken, R., Obremski, K., 1990. Social contact and sibling similarity: facts, issues, and red herrings. Behav. Genet. 20, 763–778. Ryder, A.G., Bagby, R.M., Dion, K.L., 2001. Chronic, low-grade depression in a nonclinical sample: depressive personality or dysthymia? J. Pers. Disord. 15, 84–93. Ryder, A.G., Bagby, R.M., Schuller, D.R., 2002. The overlap of depressive personality disorder and dysthymia: a categorical problem with a dimensional solution. Harv. Rev. Psychiatry 10, 337–352.

R.E. Ørstavik et al. / Journal of Affective Disorders 99 (2007) 181–189 Ryder, A.G., Bagby, R.M., Marshall, M.B., Costa, P.T., 2005. The depressive personality. In: Rosenbluth, M., Kennedy, S.H., Bagby, R.M. (Eds.), Depression and Personality: Conceptual and Clinical Challenges. American Psychiatric Publishing, Arlington, pp. 65–94. Ryder, A.G., Schuller, D.R., Bagby, R.M., 2006. Depressive personality and dysthymia: evaluating symptom and syndrome overlap. J Affect. Disord. 91, 217–227. SAS Institute, I. (2005). SAS Institute Inc., SAS OnlineDoc Version 9.1.3, Cary, NC: SAS Institute Inc., 2002–2005. In (ed. S. C.D.o.S. NC State University), Cary, NC. Saulsman, L.M., Page, A.C., 2004. The five-factor model and personality disorder empirical literature: a meta-analytic review. Clin. Psychol. Rev. 23, 1055–1085. Skodol, A.E., Oldham, J.M., Bender, D.S., Dyck, I.R., Stout, R.L., Morey, L.C., Shea, M.T., Zanarini, M.C., Sanislow, C.A., Grilo, C.M., McGlashan, T.H., Gunderson, J.G., 2005. Dimensional representations of DSM-IV personality disorders: relationships to functional impairment. Am. J. Psychiatry 162, 1919–1925. Torgersen, S., Lygren, S., Oien, P.A., Skre, I., Onstad, S., Edvardsen, J., et al., 2000. A twin study of personality disorders. Compr. Psychiatry 41, 416–425.

189

Torgersen, S., Kringlen, E., Cramer, V., 2001. The prevalence of personality disorders in a community sample. Arch. Gen. Psychiatry 58, 590–596. Vink, J.M., Willemsen, G., Boomsma, D.I., 2005. Heritability of smoking initiation and nicotine dependence. Behav. Genet. 35, 397–406. Widiger, T.A., Costa, P.T., 2002. Five-factor model personality disorder research, In: Costa, P.T., Widiger, T.A. (Eds.), Personality Disorders and the Five-Factor Model of Personality, 2 ed. American Psychological Association, Washington DC, pp. 59–87. Wittchen, H.U., Pfister, H. (1997) DIA-X-Interviews (M-CIDI): Manual für Screening-Verfahren und Interview; Interviewheft Längsschnittuntersuchung (DIA-X-Lifetime); Ergänzungsheft (DIA-X lifetime); Interviewheft Querschnittuntersuchung (DIA-X 12 Monate); Ergänzungsheft (DIA-X 12 Monate); PC-Programm zur Durchführung des Interviews (Längs-und Querschnittuntersuchung); Auswertungsprogramm. Frankfurt, Germany: Swets and Zeitlinger. Zanarini, M.C., Skodol, A.E., Bender, D., Dolan, R., Sanislow, C., Schaefer, E., et al., 2000. The collaborative longitudinal personality disorders study: reliability of axis I and II diagnoses. J. Pers. Disord. 14, 291–299.