• depressive personality disorder;
  • dysthymic disorder;
  • harm avoidance;
  • major depressive disorder;
  • Temperament and Character Inventory


  1. Top of page
  2. Abstract

Abstract  The authors evaluated the trait/state issues of harm avoidance in depressive-spectrum disorders and its predictive potential for antidepressant response. Subjects with Diagnostic and Statistical Manual of Mental Disorders (4th edn; DSM-IV) major depressive disorder (n = 39), dysthymic disorder (n = 37), depressive personality disorder (n = 39), and healthy control subjects (n = 40) were evaluated with the Temperament and Character Inventory and the 17-item Hamilton Depression Rating Scale (HDRS-17) at baseline and after a 12 week antidepressant treatment period. Higher harm avoidance scores predicted lesser improvement in subjects with dysthymic disorder and major depressive disorder, as determined by lesser decrease in HDRS-17 scores. Mean harm avoidance scores in depressed subjects were consistently greater than those in healthy controls, controlling for age, gender and diagnosis. Mean harm avoidance scores decreased significantly in all depressive-spectrum disorders after treatment, but still remained higher than harm avoidance scores in control subjects. The present study reports that harm avoidance is a reliable predictor of antidepressant treatment in subjects with major depressive disorder and dysthymic disorder and that harm avoidance is both trait- and state-dependent in depressive-spectrum disorders.


  1. Top of page
  2. Abstract

Harm avoidance (HA), one of the three proposed biogenetic temperament dimensions in Cloninger's theory of temperament and character,1,2 has consistently been reported to be elevated in subjects with major depression.3–11 Harm avoidance scores positively correlated with the severity of depression in most reports.3,6,7,10,12,13 In addition, HA scores have also been found to decline following the treatment of depression.14–17

A group of researchers have studied the possible role of specific temperamental patterns of HA, novelty seeking (NS), and reward dependence (RD) in predicting response to antidepressant treatment. Joyce et al. first classified depressive subjects into eight temperamental types according to combination of HA, NS, and RD and correlated them with treatment response in 84 subjects with major depression.15 However, the predictive capability of these temperamental types was not replicated in a later study in subjects with major depression (n = 199).18

Another line of research has focused on the correlation of each temperament with response to antidepressant medication.12–14,16,18–20 Some studies have reported that lower HA score was correlated with better response to antidepressants. Joffe et al. reported on a sample of 40 outpatients with non-psychotic major depression who were taking desipramine/imipramine treatment for 5 weeks. They found that scores on the HA dimension were significantly lower in antidepressant responders (n = 31) after 3 months.14 Chien and Dunner reported, in a study of 63 outpatients with major depression who were taking paroxetine and fluoxetine, that responders had a greater decrease in HA scores compared to non-responders upon termination of the 12 week trials.16 Tome et al. noted, in a sample of 48 patients with major depression who were taking paroxetine and/or pindolol, that low scores in the temperament dimension of HA had a better outcome at 6 weeks.20

There also have been studies (with differing results) on the correlation of HA and treatment response of depressive disorder. In a study on 52 patients with major depression, Nelsen and Dunner found that HA scores in treatment-resistant patients (n = 26) were lower than those for treatment-responsive patients after a number of different treatments (i.e. electroconvulsive therapy (ECT), lithium augmentation, thyroid augmentation, psychotherapy, and hospitalization).19 Later, Newman et al. found, in a study of 199 outpatients with major depressive disorder who were treated with fluoxetine for 8 weeks, that there was no significant correlation of pretreatment HA scores and treatment response.18

Besides HA, there have been studies that reported the correlation of RD or NS with response to antidepressant medication. Nelson and Cloninger reported that RD scores were significantly different between responders and non-responders.12,13 Tome et al. have also reported that high RD scores and high NS scores predicted a better outcome.20 Nelsen and Dunner found that treatment-resistant patients (n = 26) had lower NS scores as well as HA scores relative to treatment-responsive patients.19

However, limitations in study design make the interpretations of these findings difficult. First, all previous studies of HA were done with depressed subjects who either had or were presumed to have had substantial comorbid psychiatric diagnoses of axis I8,13,15 or II.7,8,15,17 In addition, exclusion criteria were not comprehensive enough to cover psychiatric diagnoses and there has been insufficient description of comorbidity.3,5,6,10,12–14,16,18,19,21–23 Considering the reports of high HA with obsessive–compulsive disorder,21,24 panic,5,8,21,25–29 post-traumatic stress disorder,30,31 generalized anxiety disorder,21,25,26,28,29,32 agoraphobia,25,33 social phobia,21,34 eating disorder,35–37 and personality disorders,38,39 comorbidity issues should be taken into consideration when interpreting the results of the Temperament and Character Inventory. In addition, Mulder et al. demonstrated that the temperamental patterns differed by the comorbid psychiatric disorders in subjects with major depression.8 Also, only a few of the studies have used structured or semistructured diagnostic interviews to identify both axis I and II comorbid psychiatric disorders.8,17

Second, HA findings have been reported primarily in subjects with major depression or depressive episode. Only one study uses a sample of dysthymia17 and there is no study in depressive personality disorder even though the latter two (i.e. dysthymia and depressive personality disorder), are even more prevalent than major depressive disorder. Hellerstein et al. studied a sample of 355 early onset dysthymia subjects.17 They found that HA scores of dysthymia subjects were higher than community norms and decreased significantly after treatment with either sertraline or imipramine.

The purpose of the current study was to systematically examine the trait/state patterns of HA, before and after antidepressant treatment (12 week duration), in all Diagnostic and Statistical Manual of Mental Disorders (4th edn; DSM-IV) axis I and II depressive-spectrum disorders (major depressive disorder, MDD; dysthymic disorder, DD; depressive personality disorder, DPD), all without comorbid axis I or II psychiatric diagnosis. We also intended to explore the possible capability of HA levels before treatment to predict treatment response. In addition, HA was assessed in healthy control subjects before and at the end of the treatment period for the depressed subjects.

Based on previous reports of higher HA scores in depressed subjects,3–13 we hypothesized that HA scores would be greater in depressed subjects compared to those in healthy control subjects and that it would decrease significantly after successful antidepressant treatment. In addition, based on previous reports of correlations between lower HA at baseline,14,20 greater decrease in HA after treatment and better treatment response, we hypothesized that this would hold in subjects with MDD and possibly in subjects with DD and DPD.


  1. Top of page
  2. Abstract

Subjects and diagnostic procedures

Female subjects with potential depression aged 20–41 years were recruited from 243 patients who were referred for current depression from March 1996 through May 2000 at the Psychiatric Outpatient Clinic, Seoul National University Hospital, and through referrals from the school clinics and consultation services at four women's and five coed universities in Seoul metropolitan area, South Korea. Initially, both female and male subjects with potential depression were recruited. However, the male composition ranged from 5% to 15% depending on the recruiting organizations. Consequently, only female subjects were included in the statistical analysis to enhance the homogeneity of study subjects, because gender differences in HA have been reported both in depressed40 and psychiatrically healthy subjects.1

Diagnoses of DSM-IV MDD (n = 39) and DD (n = 37) were determined by the Korean version of the Structured Clinical Interview for DSM-IV (Patient Version (SCID-I/P, version 2.0).41,42 The presence of DSM-IV DPD (n = 39) was determined by the Korean version of the Diagnostic Interview for Depressive Personality (DID).43–45

Exclusion criteria were (i) current or lifetime comorbid DSM-IV axis I disorders, as determined by of the SCID-I/P; (ii) comorbid DSM-IV axis II personality disorders, as determined by Diagnostic Interview for Personality Disorders (DIPD);46 and (iii) concurrent neurological or other significant medical illnesses and past history of brain trauma, encephalitis, seizure, or attention-deficit hyperactive disorder/learning disabilities, as evaluated by history interviews, school reports, physical examination, and laboratory testing (complete blood count, liver function test, serology, and urine analysis).

All volunteer control subjects were recruited through the use of advertisements and carefully evaluated by history interviews, physical examination, and laboratory testing to rule out concurrent neurological or other significant medical illnesses. Healthy control subjects (n = 40; mean age: 28.5 ± 3.4 years) were without current or lifetime DSM-IV axis I or II disorders and DPD, as determined by the SCID-I/P, the DIPD, and the DID.

After complete description of the study to the subjects, written informed consent was obtained.

Evaluation of harm avoidance and other variables

All subjects completed the Korean version of the Temperament and Character Inventory2 before and after the 12 week treatment period. Reliability and validity of the Korean version of the Temperament and Character Inventory were successfully tested in a recent study using a sample of 851 non-clinical subjects.47 Other assessed variables included level of depression using the 17-item Hamilton Depression Rating Scale (HDRS-17),48 level of education, and socioeconomic status.

All patients were treated according to the treatment protocol of a large antidepressant study. Sertraline was started at 50 mg/day, and the dose was adjusted at the investigator's discretion (50–150 mg range) with a 12 week mean dose of 73.5 mg. Adequate therapeutic response was defined as a 50% or more reduction in HDRS-17 score and a final score of <10 at any time during the first 12 weeks of treatment. Six subjects (MDD n = 1, DD n = 2, and DPD n = 3) changed medication to paroxetine due to side-effects.

Sample size calculations

Sample size calculation was based on expected HA differences before and after treatment with an alpha level of 0.05. A medium-effect size difference was assumed based on previous reports. To provide an expected statistical power of 0.85 or greater for detecting differences in the HA scores before and after treatment, the sample size for each group needed to have a minimum of 30 for each group at post-treatment.49

Statistical analysis

Group differences in demographic and clinical variables involving continuous data (age, height, weight, years of education, social class, and HDRS scores) were computed using one-way anova with post-hoc Scheffe tests. Group comparisons involving categorical data (marital status, and the presence of mood disorders in first-degree relatives) were assessed using Fisher's exact test for n × k table.

Between-diagnostic group differences in HA scores were computed using multiple linear regression analyses at baseline and after treatment, controlling for age.

Within-diagnostic group differences between baseline and after treatment in HA scores were compared using cross-sectional time-series regression analysis (generalized estimation equation model: GEE) considering that HA scores at baseline and after treatment within diagnostic group are intercorrelated. The method permits adjustment for covariates and allows robust estimation of standard errors. Age was controlled in all GEE models.

To test the hypothesis that baseline HA scores may play a role in predicting treatment outcome, multiple linear regression analyses were used (independent variable: baseline HA scores; dependent variable: decrease in HDRS scores following treatment) in each diagnostic group. Age and the baseline HDRS scores were controlled as covariates. In addition, similar analyses were repeated with modified dependent variables of percentage reduction in HDRS scores after treatment (multiple linear regression analysis).

Because there have been reports, although inconsistent, on the predictive capability of RD for treatment response in subjects with depression,12,13,15,20 RD was originally included in construction of a regression model. However, RD was dropped due to non-significance.

Statistical significances were defined at the 0.05 level for the primary analysis and at 0.01 for auxiliary analyses, two-tailed. Stata 6.0 for Windows (Stata, College Station, TX, USA) was used for all computations.


  1. Top of page
  2. Abstract

Demographic and clinical characteristics in all diagnostic groups at baseline

There were significant group differences between diagnostic groups in age (one-way anova: F3,151 = 21.8, P < 0.001; Scheffe test: MDD > DD, control > DPD), height (one-way anova: F3,151 = 5.7, P = 0.001; Scheffe test: DPD > DD, MDD) and weight (one-way anova: F3,151 = 3.4, P = 0.019; Scheffe test: MDD > DPD; Table 1). There were no significant differences in gender, educational level, marital status, socioeconomic status, and the presence of mood disorders in first-degree relatives between groups. All multivariate analyses hereafter were conducted controlling for age.

Table 1.  Subject profile (mean ± SD)
VariablesHealthy control n = 40DPD n = 39DD n = 37MDD n = 39
  • DPD, depressive personality disorder; DD, dysthymic disorder; MDD, major depressive disorder; HRSC, Hollingshead and Redlich social class; HDRS, Hamilton Depression Rating scale.

  • One-way ANOVA: F(3,151) = 21.8, P < 0.001; Scheffe test: MDD>DD, healthy control>DPD.

  • One way ANOVA: F(3,151) = 5.7, P = 0.001; Scheffe test: DPD>DD, MDD.

  • §

    One-way ANOVA: F(3,151) = 3.4, P = 0.019; Scheffe test: MDD>DPD.

  • One-way ANOVA: F(3,151) = 220.9, P < 0.001; Scheffe test: MDD>DD>DPD>healthy control.

  • Family history of mood disorders (depression, DD, bipolar disorder and cyclothymia) is produced for 36 healthy control subjects, 35 DPD subjects, 33 DD subjects and 35 MDD subjects at baseline. No significant difference in prevalence between depressive spectrum disorders, eitiher individually or combined, as compared to the healthy control group.

Age (years) 28.5 ± 3.4 25.2 ± 3.4 28.5 ± 3.9 32.2 ± 4.6
Gender (F/M)40/039/037/039/0
Height (cm)160.1 ± 4.5161.8 ± 3.2158.9 ± 3.9158.2 ± 4.5
Weight§ (kg) 50.3 ± 6.9 49.3 ± 5.6 51.1 ± 6.8 53.7 ± 6.2
Education (years) 13.8 ± 1.9 14.3 ± 1.7 13.6 ± 1.6 13.3 ± 1.6
Marital status (n, %)
 Never married24 (60.0)29 (74.4)24 (64.9)17 (43.6)
 Married16 (40.0) 8 (20.5)12 (32.4)19 (48.7)
 Divorced0 (0.0)2 (5.1)1 (2.7)3 (7.7)
HRSC 2.0 ± 0.7 2.1 ± 0.8 2.2 ± 1.1 1.9 ± 1.0
HDRS 4.3 ± 2.1 12.1 ± 2.8 15.9 ± 2.7 18.2 ± 2.7
Family history of mood disorder (n, %):
 Subjects with
 0 relatives with mood disorder34 (94.4)27 (77.1)27 (81.8)28 (80.0)
 1 relative with mood disorder2 (5.6) 7 (20.0) 6 (18.2) 5 (14.3)
 ≥2 relatives with mood disorder0 (0.0)1 (2.9)0 (0.0)2 (5.7)

There were also significant group differences in baseline HDRS scores between diagnostic groups (one-way anova: F3,151 = 220.9, P < 0.001; Scheffe test: MDD > DD > DPD > control).

HA scores in DPD, DD, MDD and control groups, at baseline and after treatment

At baseline, HA mean scores significantly differed between depressive spectrum disorder groups (no significant differences among DPD, DD, MDD) and the healthy control group, controlling for age (DPD, MDD, DD > control: multiple linear regression coefficent > 8.34, t > 7.67, d.f. = 150, P < 0.001; Fig. 1).


Figure 1. Harm avoidance scores before and after treatment in healthy control subjects and subjects with depressive personality disorder, dysthymic disorder, and major depressive disorder. X axis: diagnostic group; Y axis: Harm Avoidance scores. Box, mean standard error; bars, 95% confidence intervals.

Download figure to PowerPoint

Following treatment, HA mean scores significantly differed between depressive spectrum disorder groups and the healthy control group, controlling for age (multiple linear regression coefficient > 3.91, t > 4.08, d.f. = 150, P < 0.001). There were also significant differences among depressive spectrum groups (DPD > MD, DD: multiple linear regression coefficient > 4.05, t > 3.89, d.f. = 150, P < 0.001).

Within-diagnostic group differences between baseline and following treatment were as follows. There were significant ‘time’ differences in DPD, DD, MDD, controlling for age (generalized estimation equation: coefficient = −3.33, z = −4.00, P < 0.001; coefficient =−4.38, z = −5.45, P < 0.001; coefficient = −5.49, z =−5.66, P < 0.001, respectively). Age was a significant predictor of time-related difference in HA scores only in the MDD group (coefficient = −0.26, z = − 2.47, P = 0.013). There was no significant ‘time’ difference in healthy control group.

Spearman's correlation coefficients between HA and HDRS scores in depressive spectrum disorders (n = 115) were 0.38 (t = 5.06, d.f. = 113, P < 0.01) at baseline and 0.24 (t = 3.01, d.f. = 113, P < 0.01) following treatment.

Predicting treatment response (reduction in HDRS scores by baseline HA scores)

Regression modeling was started with independent variables of baseline HA, baseline RD, age and gender. However, contrary to the literature, baseline RD was not a significant predictor of treatment response when entered with HA. Also, because there were no goodness-of-fit differences between models with and without RD, two final regression models were formed as follows: (i) independent (baseline HA, baseline HDRS, and age); dependent (decrease in HDRS scores following treatment); and (ii) independent (baseline HA and age); dependent (% reduction in HDRS scores).

There were significant inverse relationships between HA at baseline and HDRS decrease following treatment in groups with DD (multiple regression coefficient = −0.21, t = −2.62, d.f. = 33, P = 0.013) and MDD (multiple regression coefficient = −0.18, t =−2.33, d.f. = 35, P = 0.026) controlling for age and the baseline HDRS scores (Fig. 2). There was a subthreshold inverse relationship in DPD group (multiple regression coefficient = −0.13, t = −1.84, d.f. = 35, P = 0.074) and no significant relationship in the healthy control group. Results from the second model had similar results to the first model, and therefore was not presented in the paper. Regression models with all temperaments of HA, NS and RD were also constructed. However, NS and RD were not significant predictors for the treatment response, either alone or via interactions. As a result, HA was the only independent factor in the regression model.


Figure 2. Inverse relationships between harm avoidance scores at baseline and the decrease of Hamilton Depression Rating Scale (HDRS) scores following treatment in groups with dysthymic disorder and major depressive disorder. There were significant inverse relationships between harm avoidance scores at baseline and the decrease of HDRS scores following treatment in groups with dysthymic disorder and major depressive disorder, controlling for age and HDRS scores at baseline. Box, mean standard error; bars, 95% confidence intervals.

Download figure to PowerPoint


  1. Top of page
  2. Abstract

There were a number of important differences between the current study and previous investigations of HA in depressed subjects. First, the longitudinal assessment, before and after a 12 week antidepressant treatment period, in the present study complemented cross-sectional measurements of HA in depressed and healthy control groups. Thus, the implication of antidepressant treatment could be more efficiently evaluated. Second, state and trait components of HA could be explained in a more integrated way. Third, depressed subjects and healthy control subjects without any comorbid axis I and II disorders were included in the current study.

Isolating the effects of HA within the depressive spectrum disorders from other mental disorders was essential because HA had been reported to be high also in many psychiatric disorders.5,8,21,24–39 Hence, the current study reported a more stringent test of HA levels in depressive disorders than prior investigations.

The present study showed that HA alone could play an important role in predicting antidepressant treatment. High HA score at baseline was followed by poor improvement of depression symptoms measured by HDRS. Previous studies have tried to identify the predictor of drug response using each subscale itself or combinations of NS, HA, RD.12–16,18–20 The present study was in line with the prior studies that higher HA predicted poor improvement.

Our finding that depressed subjects, both before and after treatment, had significantly higher HA scores than healthy controls was in line with results of previous studies.3–11 The current finding suggested that HA was trait-dependent in depression, and could be considered as a personality risk factor for depression.14 Harm avoidance scores following treatment were significantly higher in DPD groups than in MDD or DD groups. Thus, HA at baseline may be a stronger trait in cognitive and behaviorally defined depression (i.e. DPD) than in more affectively defined depressive disorders (i.e. MDD and DD).

Current findings have also suggested that HA may be, in part, state-dependent as well as trait-dependent because mean HA scores for the MD, DD, and DPD groups after treatment were significantly lower than those before treatment. This perspective is in accord with prior reports.50,51

Past history of depression can induce an alteration of premorbid personality, which, in turn, may affect the measurement of the HA level. However, HA scores in depressed subjects have been reported to be associated with relatively stable biological markers such as platelet serotonergic receptors.52 In addition, a recent sibling study has showed that HA was related to the genetic association of depression.40 Siblings of subjects with major depression have been reported to have higher HA than healthy control subjects, even if they had no current or past depression. Consequently, high HA is likely to play a role as a risk factor for development of depression.

One drawback to the clinical utility of the current study findings, however, was that the clinical state of depression can influence some aspects on personality assessment, such as emotional strength, interpersonal dependency, and extraversion.53–55 Therefore, clarification of the state and trait dependency issues of HA in depressive-spectrum disorders will be needed in the future studies.

The current findings are results from female subjects. As a result, further study using both male and female subjects is recommended because gender differences in HA have been reported both in depressed40 and psychiatrically healthy subjects.1

In conclusion, the current study systemically evaluated the role of baseline HA as a potential predictor for treatment response in subjects with a wide range of depression.

Also, the state/trait dependency issues of HA in all axis I and II mood disorders of MDD, DD, and DPD have been further clarified by cross-sectional and longitudinal studies.


  1. Top of page
  2. Abstract

This work was supported by grants from the Korean Research Foundation (KRF-99–003-F00216-F1518). The authors thank Eunjoo Yang, MA for comments on the manuscript and proofreading.


  1. Top of page
  2. Abstract
  • 1
    Cloninger CR. A systematic method for clinical description and classification of personality variants. A proposal. Arch. Gen. Psychiatry 1987; 44: 573588.
  • 2
    Cloninger CR, Svrakic DM, Przybeck TR. A psychobiological model of temperament and character. Arch. Gen. Psychiatry 1993; 50: 975990.
  • 3
    Strakowski SM, Dunayevich E, Keck PE Jr, , McElroy SL. Affective state dependence of the Tridimensional Personality Questionnaire. Psychiatry Res. 1995; 57: 209214.
  • 4
    Cloninger CR, Praybeck T, Svrakic DM, Wetzel R. The Temperament and Character Inventory: A Guide to its Development and Use. Center for Psychobiology of Personality, Washington University, St Louis, 1994.
  • 5
    Ampollini P, Marchesi C, Signifredi R et al. Temperament and personality features in patients with major depression, panic disorder and mixed conditions. J. Affect. Disord. 1999; 52: 203207.
  • 6
    Hansenne M, Pitchot W, Moreno AG, Reggers J, Machurot PY, Ansseau M. Harm avoidance dimension of the Tridimensional Personality Questionnaire and serotonin-1A activity in depressed patients. Biol. Psychiatry 1997; 42: 959961.
  • 7
    Hansenne M, Reggers J, Pinto E, Kjiri K, Ajamier A, Ansseau M. Temperament and character inventory (TCI) and depression. J. Psychiatr. Res. 1999; 33: 3136.
  • 8
    Mulder RT, Joyce PR, Cloninger CR. Temperament and early environment influence comorbidity and personality disorders in major depression. Compr. Psychiatry 1994; 35: 225233.
  • 9
    Tanaka E, Kijima N, Kitamura T. Correlations between the Temperament and Character Inventory and the Self-rating Depression Scale among Japanese Students. Psychol. Rep. 1997; 80: 251254.
  • 10
    Naito M, Kijima N, Kitamura T. Temperament and Character Inventory (TCI) as predictors of depression among Japanese college students. J. Clin. Psychol. 2000; 56: 15791585.
  • 11
    Sato T, Narita T, Hirano S et al. Factor validity of the temperament and character inventory in patients with major depression. Compr. Psychiatry 2001; 42: 337341.
  • 12
    Nelson E, Cloninger CR. Exploring the TPQ as a possible predictor of antidepressant response to nefazodone in a large multi-site study. J. Affect. Disord. 1997; 44: 197200.
  • 13
    Nelson EC, Cloninger CR. The tridimensional personality questionnaire as a predictor of response to nefazodone treatment of depression. J. Affect. Disord. 1995; 35: 5157.
  • 14
    Joffe RT, Bagby RM, Levitt AJ, Regan JJ, Parker JD. The Tridimensional Personality Questionnaire in major depression. Am. J. Psychiatry 1993; 150: 959960.
  • 15
    Joyce PR, Mulder RT, Cloninger CR. Temperament predicts clomipramine and desipramine response in major depression. J. Affect. Disord. 1994; 30: 3546.
  • 16
    Chien AJ, Dunner DL. The Tridimensional Personality Questionnaire in depression: state versus trait issues. J. Psychiatr. Res. 1996; 30: 2127.
  • 17
    Hellerstein DJ, Kocsis JH, Chapman D, Stewart JW, Harrison W. Double-blind comparison of sertraline, imipramine, and placebo in the treatment of dysthymia: effects on personality. Am. J. Psychiatry 2000; 157: 14361444.
  • 18
    Newman JR, Ewing SE, McColl RD et al. Tridimensional personality questionnaire and treatment response in major depressive disorder: a negative study. J. Affect. Disord. 2000; 57: 241247.
  • 19
    Nelsen MR, Dunner DL. Clinical and differential diagnostic aspects of treatment-resistant depression. J. Psychiatr. Res. 1995; 29: 4350.
  • 20
    Tome MB, Cloninger CR, Watson JP, Isaac MT. Serotonergic autoreceptor blockade in the reduction of antidepressant latency: personality variables and response to paroxetine and pindolol. J. Affect. Disord. 1997; 44: 101109.
  • 21
    Kennedy BL, Schwab JJ, Hyde JA. Defense styles and Personality dimensions of research subjects with anxiety and depressive disorders. Psychiatr. Q. 2001; 72: 251262.
  • 22
    Richter J, Eisemann M, Richter G. Temparament and character during the course of unipolar depression among inpatients. Eur. Arch. Psychiatry Clin. Neurosci. 2000; 250: 4047.
  • 23
    Sato T, Hirano S, Narita T et al. Temperament and character inventory dimensions as a predictor of response to antidepressant treatment in major depression. J. Affect. Disord. 1999; 56: 153161.
  • 24
    Pfohl B, Black D, Noyes R Jr, , Kelley M, Blum N. A test of the tridimensional personality theory: association with diagnosis and platelet imipramine binding in obsessive-compulsive disorder. Biol. Psychiatry 1990; 28: 4146.
  • 25
    Fossey M, Roy-Byrne P, Cowley D et al. Personality assessment using the Tridimensional Personality Questionnaire (TPQ) in patients with panic disorder and generalized anxiety disorder. Biol. Psychiatry 1989; 25: 10A13A.
  • 26
    Cowley DS, Roy-Byrne PP, Greenblatt DJ, Hommer DW. Personality and benzodiazepine sensitivity in anxious patients and control subjects. Psychiatry Res. 1993; 47: 151162.
  • 27
    Perna GBL, Caldirola D, Garberi A et al. Personality dimension in panic disorder: state versus trait issues. New Trends Exp. Clin. Psychiatry 1992; 8: 4954.
  • 28
    Starcevic V, Uhlenhuth EH, Fallon S, Pathak D. Personality dimensions in panic disorder and generalized anxiety disorder. J. Affect. Disord. 1996; 37: 7579.
  • 29
    Wingerson D, Sullivan M, Dager S, Flick S, Dunner D, Roy-Byrne P. Personality traits and early discontinuation from clinical trials in anxious patients. J. Clin. Psychopharmacol. 1993; 13: 194197.
  • 30
    Richman H, Frueh BC. Personality and PTSDII: personality assessment of PTSD-diagnosed Vietnam veterans using the cloninger tridimensional personality questionnaire (TPQ). Depress. Anxiety 1997; 6: 7077.
  • 31
    Wang S, Mason J, Charney D, Yehuda R, Riney S, Southwick S. Relationships between hormonal profile and novelty seeking in combat-related posttraumatic stress disorder. Biol. Psychiatry 1997; 41: 145151.
  • 32
    Allgulander C, Cloninger CR, Przybeck TR, Brandt L. Changes on the Temperament and Character Inventory after paroxetine treatment in volunteers with generalized anxiety disorder. Psychopharmacol. Bull. 1998; 34: 165166.
  • 33
    Saviotti FM, Grandi S, Savron G et al. Characterological traits of recovered patients with panic disorder and agoraphobia. J. Affect. Disord. 1991; 23: 113117.
  • 34
    Chatterjee S, Sunitha TA, Velayudhan A, Khanna S. An investigation into the psychobiology of social phobia: personality domains and serotonergic function. Acta Psychiatr. Scand. 1997; 95: 544550.
  • 35
    Brewerton TD, Hand LD, Bishop ER Jr. The Tridimensional Personality Questionnaire in eating disorder patients. Int. J. Eat. Disord. 1993; 14: 213218.
  • 36
    Kleifield EI, Sunday S, Hurt S, Halmi KA. Psychometric validation of the Tridimensional Personality Questionnaire: application to subgroups of eating disorders. Compr. Psychiatry 1993; 34: 249253.
  • 37
    Waller DA, Gullion CM, Petty F, Hardy BW, Murdock MV, Rush AJ. Tridimensional Personality Questionnaire and serotonin in bulimia nervosa. Psychiatry Res. 1993; 48: 915.
  • 38
    Svrakic DM, Whitehead C, Przybeck TR, Cloninger CR. Differential diagnosis of personality disorders by the seven-factor model of temperament and character. Arch. Gen. Psychiatry 1993; 50: 991999.
  • 39
    Mulder RT, Joyce PR, Sullivan PF, Bulik CM, Carter FA. The relationship among three models of personality psychopathology: DSM-III-R personality disorder, TCI scores and DSQ defences. Psychol. Med. 1999; 29: 943951.
  • 40
    Farmer A, Mahmood A, Redman K, Harris T, Sadler S, McGuffin P. A sib-pair study of the Temperament and Character Inventory scales in major depression. Arch. Gen. Psychiatry 2003; 60: 490496.
  • 41
    First MB, Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for DSM-IV Axis I Disorders – Clinician Version (SCID-CV). American Psychiatric Press, Washington, DC, 1997.
  • 42
    Hahn OS, Ahn JH, Song SH et al. Development of Korean version of Structured Clinical Interview Schedule for DSM-IV Axis I disorder: interrater reliability. J. Korean Neuropsychiatr. Assoc. 2000; 39: 362372.
  • 43
    Gunderson JG, Phillips KA, Triebwasser J, Hirschfeld RM. The Diagnostic Interview for Depressive Personality. Am. J. Psychiatry 1994; 151: 13001304.
  • 44
    Kwon JS, Kim YM, Chang CG et al. Three-year follow-up of women with the sole diagnosis of depressive personality disorder: subsequent development of dysthymia and major depression. Am. J. Psychiatry 2000; 157: 19661972.
  • 45
    Kim YM. Clinical validity and utility of depressive personality disorder diagnosis (doctoral dissertation). Seoul, Department of Psychology, Yonsei University Graduate School, 1999.
  • 46
    Zanarini MC, Frankenburg FR, Chauncey DL, Gunderson JG. The Diagnostic Interview for Personality Disorders: interrater and test-retest reliability. Compr. Psychiatry 1987; 28: 467480.
  • 47
    Sung SM, Kim JH, Yang E, Abrams KY, Lyoo IK. Reliability and validity of the Korean version of the Temperament and Character Inventory. Compr. Psychiatry 2002; 43: 235243.
  • 48
    Hamilton M. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry 1960; 23: 5662.
  • 49
    Cohen J. Statistical Power Analysis for the Behavioral Sciences. Erlbaum, Hillsdale, NJ, 1988.
  • 50
    Krebs H, Weyers P, Janke W. Validation of the German version of Cloninger's TPQ. Replication and correlations with stress coping, mood measure and drug use. Pers. Individ. Diff. 1997; 24: 805814.
  • 51
    Hirano S, Sato T, Narita T et al. Evaluating the state dependency of the Temperament and Character Inventory dimensions in patients with major depression: a methodological contribution. J. Affect. Disord. 2002; 69: 3138.
  • 52
    Nelson EC, Cloninger CR, Pryzbeck TR, Csernansky JG. Platelet serotonergic markers and Tridimensional Personality Questionnaire measures in a clinical sample. Biol. Psychiatry 1996; 40: 271278.
  • 53
    Hirschfeld RM, Klerman GL, Clayton PJ, Keller MB, McDonald-Scott P, Larkin BH. Assessing personality: effects of the depressive state on trait measurement. Am. J. Psychiatry 1983; 140: 695699.
  • 54
    Liebowitz MR. Newer uses for older psychotropic medications. Hosp. Community Psychiatry 1982; 33: 282286.
  • 55
    Black KJ, Sheline YI. Personality disorder scores improve with effective pharmacotherapy of depression. J. Affect. Disord. 1997; 43: 1118.