Alexithymia and self-directedness as predictors of psychopathology and psychotherapeutic treatment outcome

Available online at www.sciencedirect.com

ScienceDirect Comprehensive Psychiatry 62 (2015) 34 – 41 www.elsevier.com/locate/comppsych

Alexithymia and self-directedness as predictors of psychopathology and psychotherapeutic treatment outcome Jan Terock a,⁎, Deborah Janowitz a , Carsten Spitzer b , Martin Miertsch a , Harald J. Freyberger a , Hans Jörgen Grabe a a

Department of Psychiatry and Psychotherapy, University Medicine Greifswald, HELIOS Klinikum Stralsund, Germany b Asklepios Fachklinikum Tiefenbrunn, Rosdorf, Germany

Abstract Objective: Alexithymia, a common personality style of patients seeking psychotherapeutic help, is associated with illness severity and negative treatment outcome in various mental disorders. Still, it remains unclear how alexithymia influences psychopathology and the therapeutic processes. In previous studies, a strong association of alexithymia with self-directedness (SD), a dimension of Cloninger’s Temperament and Character Inventory (TCI) has been shown. In this study, we investigated the interaction of alexithymia and SD, and their impact on general psychopathology and on treatment outcome. Method: 716 consecutively admitted day-clinic outpatients were examined at admission (t0) and discharge (t1). The Toronto Alexithymia Scale 20 (TAS-20), the SD subscale of the TCI and the Symptom Checklist 90 (SCL-90-R) were administered. Linear regression analyses were performed to calculate associations and the predictive power of TAS-20 and SD on psychopathology at admission and treatment outcome. ANOVA was used to calculate interactions of TAS-20 and SD on treatment outcome. A general linear model was applied to compare the outcome of four subgroups, defined by high/low TAS-20 and SD scores. Results: Regression analyses revealed significant prediction of the baseline General Severity Index (GSIt0) by TAS-20 (df = 4, 711; Beta: 0.385; p b 0.001) and SD (Beta: −0.365; p b 0.001). The whole model accounted for 41% of the explained variance. On subscale level, the ‘Difficulties in identifying feelings’ facet (DIF) of TAS-20 was the strongest predictor of GSIt0 (Beta: 0.478, P b 0.001) and GSIt1 (Beta: 0.072, p = 0.049). Therapeutic outcome measured by GSIt1 was significantly predicted by SD (df = 5, 710; Beta: −0.065; p = 0.041), but not by TAS-20 (Beta: 0.042; p = 0.179). Change scores (Δ) of TAS-20 and SD predicted GSIt1 (df = 5, 710; TAS-20Δ Beta: −0.268; p b 0.001; SDΔ Beta: 0.191; p b 0.001) as well as GSIΔ (df = 5, 710; TASΔ Beta: 0.384; p b 0.001; SDΔ: −0.274; p b 0.001) significantly. ANOVA revealed no significant interactions of TAS-20 and SD at admission on the treatment outcome (p N 0.05). Conclusion: Low SD was shown to be a common problem of alexithymic patients and both, alexithymia and SD were highly associated with general symptom severity. SD was found to have a greater impact on treatment outcome while adjusting for baseline GSI. Alexithymia and SD act as independent factors with no significant interaction in their impact on psychopathology at admission and discharge. As different interventions were shown to improve SD scores in previous studies, SD may represent a relevant psychotherapeutic target, worthy to be addressed especially in alexithymic patients. Future studies should investigate other dimensions of the TCI, especially harm avoidance and reward dependence. © 2015 Elsevier Inc. All rights reserved.

1. Introduction In the understanding of state-dependent psychopathology and its treatment, personality styles are supposed to play a key role. Alexithymia has been conceptualized as a relatively ⁎ Corresponding author at: Department of Psychiatry and Psychotherapy, University Medicine Greifswald, HELIOS Klinikum Stralsund, Rostocker Chaussee 70, 18437 Stralsund. Tel.: +49 3831 45 2162; fax: +49 3831 45 2165. E-mail address: [email protected] (J. Terock). http://dx.doi.org/10.1016/j.comppsych.2015.06.007 0010-440X/© 2015 Elsevier Inc. All rights reserved.

time-stable personality style including the reduced ability to realize, identify and express one’s emotions, but also a concrete cognitive style of thinking and communication [1]. Previous studies found an association of alexithymia with all subscales of SCL-90-R [2,3], and especially with mood and anxiety disorders [4]. For example, Honkalampi et al. and Saarijärvi et al. found alexithymia to be strongly associated with depression [3,5], and, vice versa, that depressed patients show high rates of alexithymia ranging from 11% to 48%. In comparison, there is a relative paucity of studies

J. Terock et al. / Comprehensive Psychiatry 62 (2015) 34–41

investigating the impact of alexithymia on psychotherapeutic treatment outcome. While McCallum et al. and Ogrodniczuk et al. found the ‘difficulties identifying feeling’ facet of alexithymia to predict treatment outcome in complicated grief, Leweke et al. found alexithymia only to be a mild outcome predictor in a sample with various psychiatric disorders whereas Spek et al. found no correlation of TAS-20 and treatment outcome in subthreshold depression [6–9]. Grabe et al. found alexithymia to be associated with higher GSI scores at admission and discharge [10]. However, a significant treatment response in the alexithymic group of patients was also seen. The mechanism of how alexithymia interacts with more state-dependent psychopathology remains unclear. Previous studies have found relationships between alexithymia and personality features as measured by Cloninger’s Temperament and Character Inventory (TCI). The TCI is based on Cloninger’s biosocial model with 4 temperament dimensions which are understood as more genetically determined and stable over time. Additionally, it includes 3 character dimensions which are considered as being influenced through learning processes throughout the lifespan, e. g. in therapeutic processes. The temperament dimensions include novelty seeking (NS), harm avoidance (HA), reward dependence (RD) and persistence (P). The character dimensions consist of self-directedness (SD), cooperativeness (CO) and selftranscendence (ST). Grabe et al. found low self-directedness (SD), low reward-dependence (RD) and, to a minor degree, harm avoidance (HA), as independent predictors for alexithymia [11]. Picardi et al., too, found low SD, low RD and high HA and, additionally, high cooperativeness (CO) to be associated with TAS-20 total score [12]. Lee et al. investigated the association of personality traits with alexithymia and their mediation through depression and anxiety in pathway analyses. Their results showed low SD, low RD and high CO to be the strongest factors to increase TAS-20 total scores. Lee et al. suggest, that low autonomous self-concept and self-confidence in subjects with low SD may be related to reduced emotion recognition and regulation in alexithymic subjects. Additionally, impoverished fantasy in alexithymia is named to be possibly related to decreased resourcefulness in low SD [13]. There has been extensive research on the relationship of the TCI and psychopathological states and various studies showed associations of especially HA and SD and psychopathology. For example, Jylhä et al. found in a community based sample a correlation of HA and a negative correlation of SD with depression [14]. Izci et al. examined a sample of patients with panic disorder [15]. They found high HA, low SD and low CO to be associated with panic disorder. Regarding the prediction of outcome, Cloninger et al. showed in a prospective study that high HA and low SD can also contribute to the prediction of change in depression [16]. Other studies found an inverse association of SD and treatment outcome of social phobia [17], bulimia nervosa [18] and obsessive–compulsive disorder [19]. Conrad et al.

35

investigated the impact of both, alexithymia and personality traits measured by TCI on state-dependent psychopathology. They found in their study the ‘difficulties identifying feelings’ facet of alexithymia to be a strong predictor of all aspects of psychopathology while low SD was the strongest predictor of obsessionality, depression, interpersonal sensitivity and psychoticism [20]. Thus, low SD seems to be among the most important factors of the TCI with impact on psychopathology, treatment response, and, additionally with a strong relationship with alexithymia. Taken together, these studies found evidence for an association of alexithymia and low SD with psychopathology and, to a minor degree, for a prediction of psychotherapeutic treatment outcome. Still, to our knowledge, no study has examined possible interactions of these two factors in their influence on psychopathology and treatment outcome. In this study we sought to examine the relationship of alexithymia and SD in their impact on psychopathology and outcome of a psychotherapeutic treatment program. Our hypotheses were as follows: (i) Patients with alexithymia show lower scores of SD than non-alexithymic patients. (ii) High scores of TAS-20 and low scores of SD at admission are associated with more severe psychopathology compared to non-alexithymic subjects and patients with higher SD. (iii) High scores of TAS-20 and low scores of SD at admission predict a poorer therapeutic outcome compared to subjects with low TAS-20 and higher SD (iv) Alexithymia and SD are improved after treatment and this improvement is associated with an improvement of GSI. (v) Alexithymia and SD interact in their impact on psychopathology and treatment outcome.

2. Materials and methods 1089 patients were consecutively admitted to 6 psychiatric outpatient’s day-clinics, all part of the University Hospital of Greifswald, Department of Psychiatry and Psychotherapy. 991 patients gave their written informed consent prior to inclusion in the study and completed the baseline testing. 275 patients were excluded from the sample due to discontinuation of the planned treatment program and/ or not completing the testing at discharge, leaving 716 patients with fulfilled treatment program and complete test dataset to form the study sample. Table 1 provides data comparing basic sociodemographic and diagnostic characteristics between the final sample and drop-outs. Patients’ main diagnoses were made according to ICD-10 based diagnostic evaluation. Table 2 gives an overview of main diagnoses in the whole sample and the subgroups. Taking all psychiatric comorbidities into account, percentages of psychiatric diagnoses in the sample of patients completing the treatment program were as follows: Alcohol/ drug dependence and abuse: 62 (8.7%); depressive disorders: 604 (84.4%); anxiety and somatoform disorders: 131 (18.3%), eating disorders: 4 (0.6%) and personality disorders

36

J. Terock et al. / Comprehensive Psychiatry 62 (2015) 34–41

Table 1 Sociodemographic characteristics and diagnostic categories of the patients with complete treatment and testing and drop-outs. PCT (N = 716)

Drop-outs (N = 275)

Statistics (PCT vs. Drop-outs)

Age (mean, SD) Sex

42.3(12.0)

43.9 (12.1)

df = 989; p = 0.058 χ2 = 0.19; df = 1, 989; p = 0.89

Male Female Marital status

204(28.5%) 512(71.5%)

80 (29.1%) 195 (70.9%)

Single Married Separated Married again Divorced Widowed Treatment diagnosis

232 (32.4%) 304 (42.5%) 40(5.6%) 8 (1.1%) 106(14.8%) 26(3.6%)

χ2 = 8.06; df = 5. 985; p = 0.506

Alcohol/Drug 13(1.8%) dependence Depressive 572 (79.9%) Episode(s) Anxiety/somatoform 84 (11.7%) disorders Eating disorders 2 (0.3%) Personality disorder 45 (6.3%)

84 (30.5%) 132 (48%) 7 (2.5%) 3 (1.1%) 34 (12.4%) 15 (5.5%) χ2 = 3.88; df = 5, 985; p = 0.274 2 (0,7%) 209 (76.0%) 35 (12.7%) 0 (0%) 29 (10.6%)

PCT = Patients completing treatment.

147 patients (20.5%). Among the personality disorders (PDs), the diagnosis of borderline personality disorder (BPD) was by far the most important: 87 patients were diagnoses as suffering from BPD, which corresponds to 59.2% of all PDs. Other PDs diagnosed were: Histrionic PD (11 patients), Obsessive–compulsive PD (11), Avoidant PD (15), Dependent PD (6) and Narcissistic PD (9). Additional 8 patients were given the diagnosis of combined PD according to criteria of ICD-10. 2.1. Instruments As part of the clinical routine patients were asked to complete the German versions of the revised Symptom Checklist 90 (SCL-90-R), the Toronto Alexithymia Scale (TAS-20) and the subscale for self-directedness, part of the Temperament and Character inventory (TCI), at admission and at discharge. After explanation of the study the patients gave written informed consent to use their psychometric data for research purposes. The SCL-90-R is a 90-item self-report scale widely used in clinical practice and in psychotherapy research. It includes nine subscales reflecting different facets of psychopathology. Using these subscales, a summary score reflecting the global psychopathological distress (Global Severity Index, GSI) can be generated. Validity and reliability of the original version as well as the German version have been shown [21–23]. Alexithymia was assessed using the German version of the TAS-20 [24], a self-rating scale composed of 20 items

representing three different factors: (1) Difficulty in identifying feeling, (2) difficulty in describing feeling and (3) externally oriented thinking [25]. Self-directedness as proposed by Cloninger et al. represents a character dimension included in the TCI, a 240-item, forced-choice self-report scale. We used the corresponding subscale of the German version (Richter et al., 2001). 2.2. Treatment setting and program Patients were treated as outpatients in day-clinics located in different towns in the north-east of Germany. In general we applied psychodynamically oriented psychotherapy. The duration of the treatment program ranged regularly between 6 and 8 weeks, depending on individual needs. We included art therapy, group therapies and individual psychotherapeutic sessions in the treatment program. Thus, biographical work, subsequent derivation of individual emotional needs and identification of interpersonal problems formed the basis of the therapy. According to individual needs of the patient, but also as a matter of the specialization of the therapist (medical doctor or psychological psychotherapist) elements of cognitive behavioral therapy like cognitive restructuring and exposition trainings were included and may have been the main focus of the treatment for some patients. The daily treatment program comprised 30 min of sports/ gymnastics in the morning. Two units (one unit equals 50 min) of body and movement therapy, 4 units of art therapy, and 1 unit of relaxation therapy (according to the principles of Jacobson) were offered per week. Group therapies formed a central element of the treatment, with 1.5 h three times per week. We applied a psychodynamic approach with focus on current interpersonal conflicts. The development of insight and clarification of different emotional states and responses in stressful situations were important aims. Another focus was the communication of psychoeducational aspects. Special dedication was given to the identification of the subjects’ individual goals and values and their adjustment according to the needs of the current personal situation. Another important aspect was the clarification and verbalization of individual emotions. Psychotropic medication was offered if clinically indicated. 2.3. Statistical analysis Descriptive statistics were performed with t tests or χ 2 tests (two-tailed). ANOVA (with age and sex as covariates) was used to compare GSI, SD and TAS-20 scores at admission and follow-up. Additionally, effect sizes (Cohen’s d) were calculated. Drop-outs and patients with complete treatment and testing were compared in basic sociodemographic data and treatment diagnoses. Afterwards, we conducted a split of the sample by TAS-20 scores: alexithymic patients (AP) with TAS-20 t0 N 60 and non-alexithymic patients (NAP) with

J. Terock et al. / Comprehensive Psychiatry 62 (2015) 34–41

37

Table 2 Sociodemographic characteristics and diagnostic categories of the total sample and comparison between alexithymic and non-alexithymic patients.

Age (mean, SD) Sex Male Female Marital status Single Married Separated Married again Divorced Widowed Treatment diagnosis Alcohol/Drug dependence Depressive Episode(s) Anxiety/somatoform disorders Eating disorders Personality disorder

Total sample (N = 716)

Alexithymic (N = 231)

Non-alexithymic (N = 485)

Statistics (AP vs. NAP)

42.3(12.0)

40.6(11.1)

43.1(12.4)

df = 714; p = 0.008 χ 2 = 6.9; df = 1, 714; p = 0.009

204(28.5%) 512(71.5%)

51 (22.1%) 180 (77.9%)

153 (31.5%) 332 (68.5%)

232 (32.4%) 304 (42.5%) 40(5.6%) 8 (1.1%) 106(14.8%) 26(3.6%)

79 (34.2%) 102 (44.2%) 11 (4.8%) 4 (1.7%) 29 (12.5%) 6 (2.6%)

153 (31.6%) 202 (41.6%) 29 (6.0%) 4 (0.8%) 77 (15.9%) 20 (4.1%)

χ 2 = 4.3; df = 5. 710; p = 0.506

χ 2 = 4.7; df = 5, 710; p = 0.317 13(1.8%) 572 (79.9%) 84 (11.7%) 2 (0.3%) 45 (6.3%)

3 (1.3%) 180 (77.9%) 28 (12.1%) 0 (0%) 20 (8.7%)

10 (2.1%) 392 (80.8%) 56 (11.5%) 2 (0.4%) 25 (5.2%)

AP = Alexithymic patients; NAP = Non-alexithymic patients.

TAS-20 t0 b 61) were extracted and tested for sociodemographic differences (Table 2). We stratified the patients into four groups according to high/low TAS-20 and SD admission-scores (median split). A general linear model with repeated measures and GSI before and after treatment as dependent variable was applied (adjusted for sex and age). In order to assess the association of alexithymia and self-directedness with psychopathology at admission, age, sex, TAS-20 t0 and SDt0 were entered in a regression analysis with GSIt0 as dependent variable. To assess the predictive power of TAS-20 t0 and SDt0 on GSIt1, adjusted for GSIt0, we applied another linear regression with age, sex, TAS-20 t0, SDt0, and GSIt0 in the model. To specifically assess the role of the three subscales of TAS-20, two additional regression analyses were performed, now using the three factors ‘difficulties in identifying feeling’ (DIF), ‘difficulties in describing feelings’ (DDF) and ‘externally oriented thinking’ instead of the TAS-20 total score in the model. Dependent variables were GSIt0 and GSIt1, respectively. ANOVA was used to calculate the interaction of TAS-20 t0 and SDt0: First, with GSIt0, second with GSIΔ as dependent variable (sex and age as covariates). We performed two more regression analyses to assess the predictive power of ΔTAS-20 and ΔSD on the outcome: First with GSIt0 as outcome variable, second with ΔGSI as outcome variable, adjusted for age, sex and GSIt0.

3. Results At admission, 32.3% of the patients were alexithymic. In concordance with hypothesis (i), alexithymic patients showed to have significantly lower SD at t0 as well as at t1

(p b 0.001). Table 3 provides detailed psychopathological characteristics of the whole sample and the subgroups. To test the reliability of results derived from the SD subscale of the TCI, we calculated Cronbach’s Alpha. The score of 0.761 indicates sufficient internal consistence of the subscale we used. In order to test hypothesis (ii), we performed a regression analysis with TAS-20 t0 and SDt0 as independent variable and GSIt0 as dependent variable. TAS-20 t0 (df = 4, 711; Beta: 0.385 p b 0.001) as well as SDt0 significantly (df = 4, 711; Beta: −0.365, p b 0.001) predicted GSIt0. To investigate hypothesis (iii), we calculated the predictive power of TAS-20 t0 and SDt0 for the treatment outcome in terms of GSIt1 using linear regression analysis. Only SDt0 proved to significantly predict the treatment outcome after correction for GSIt0 (df = 5, 710; Beta: −0.065; p = 0.041), while the predictive power of TAS-20 t0 remained not significant (Beta: 0.042; p = 0.179) (Table 3). Thus, our third hypothesis was only partially confirmed. Table 4 shows detailed results of both regression analyses. Additional analyses of the subscales of TAS-20 showed that all factors significantly predicted GSIt0, while DIF had the strongest predictive power (Beta: 0.478, p b 0.001). EOT was inversely associated with GSIt0 (Beta: −0.062; p = 0.033). In the prediction of GSIt1 by the subscales of TAS-20, only DIF remained significant after adjusting for GSIt0 (Beta: 0.072, p = 0.049), whereas DDF (Beta: 0.047, p = 0.142) and EOT (Beta: −0.045, p = 0.097) did not show significant impact (Table 4). Hypothesis (iv) was confirmed using two regression analyses: First, change scores instead of baseline measures were included as independent variables, and GSIt1 served as outcome measure. TAS-20Δ (df = 5, 710; Beta = −0.268; p b 0.001), SDΔ (Beta = 0.191; p b 0.001) and GSIt0 (Beta = 0.770; p b 0.001) showed significant predictive power for GSIt1 (Table 5). Second, GSIΔ was applied as

38

J. Terock et al. / Comprehensive Psychiatry 62 (2015) 34–41

Table 3 Summary of psychometric measures. Total sample N = 716 Mean

AP N = 231 SD

NAP N = 485

Mean

SD

Mean

AP vs. NAP SD

GSI t0

1.41

0.66

1.85

0.62

1.2

0.56

GSI t1

1.13

0.71

1.53

0.73

0.94

0.61

GSIΔ

0.28

0.49

0.33

0.53

0.26

0.47

SD t0

25.42

8.08

21.12

6.72

27.52

7.84

SD t1

26.7

8.18

22.81

7.7

28.55

7.74

TAS-20 t0

55.45

11.20

67.68

5.81

49.61

7.95

TAS-20 t1

54.14

11.92

62.83

9.5

49.98

10.64

p d p d p d p d p d p d p d

b = b = = = b = b = b = b =

0.001 −1.12 0.001 −0.91 0.082 −0.14 0.001 0.85 0.001 0.74 0.001 −2.66 0.001 1.25

AP = Alexithymic patients; NAP = Non-alexithymic patients.

outcome measure, resulting in an even stronger impact of TAS-20Δ and SDΔ on the treatment outcome, relative to GSIt0 (df = 5, 710; TAS-20Δ: Beta = 0.384, p b 0.001; SD Δ: Beta = −0.274, p b 0.001; GSIt 0 : Beta = 0.230; p b 0.001) (Table 5). ANOVA analyses of interactions between TAS-20 t0 and SDt0 were performed to investigate hypothesis (v). We did not find significant interactions between TAS-20 t0 and SDt0 in their impact on GSIt0 (df = 289, 426; p = 0.405), GSIt1 (df = 121, 594; p = 0.498) and GSIΔ (df = 121, 594; p = 0.688), whereby the fifth hypothesis was rejected. In addition to the previous tests, we used another approach to present the associations and interactions of TAS-20 t0, SDt0 and GSI-scores: We stratified the patients into four groups according to alexithymic/non-alexithymic and high/low SD t0-scores. For each group, a regression analysis with repeated measures (and adjustment for sex and age) was performed, using GSI as dependent variable, measured at t0 and t1. We found a decrease of GSI scores in all groups after treatment (Pillai-Spur 0.054; F = 40.308; p b 0.001). No significant interaction between TAS-20, SD and change of GSI was found (df = 3, 712; Pillai-Spur = 0.01; F = 2.46; p = 0.061) (Fig. 1).

4. Discussion In our study, we found that low SD was significantly associated with alexithymia. As both, alexithymia and SD were strongly associated with the initial general psychopathology, alexithymic patients low in SD formed a patient group particularly at risk of suffering from high general symptom severity. Only SDt0 but not TAS-20 t0 contributed to the prediction of treatment outcome while adjusting for GSIt0 scores. On the subscale level, the ‘difficulties in

identifying feelings’ factor had the strongest association with psychopathology and was the only factor of alexithymia to significantly predict treatment outcome. Looking at the change scores, TAS-20 as well as SD showed a strong association with the treatment outcome. No interaction in the impact on psychopathology and the prediction of treatment outcome between alexithymia and SD was found. The concept of alexithymia was first introduced as a pattern of cognitive and affective characteristics predominantly found among patients with somatoform disorders and associated with unsatisfactory response to psychodynamic psychotherapy [1]. It has been further developed, now emphasizing deficits in cognitive processing and regulation of emotions [26]. Different mechanisms of how alexithymia may lead to negative emotional states have been suggested: The ability to perceive, cognitively process and communicate one’s feelings was shown to be crucial for emotion Table 4 Multiple linear regression modeling.

ANOVA

Predictor Age Sex TAS-20 SD GSIt0 DIF DDF EOT

GSIt0

GSIt1

F = 125.04 P b 0.001 Adjusted R 2 0.41 Standard β 0.085⁎⁎ −0.098⁎⁎ 0.385⁎⁎⁎ −0.365⁎⁎⁎ N/A 0.478⁎⁎⁎ 0.085⁎ −0.062⁎

F = 176.41 P b 0.001 Adjusted R 2 0.55 Standard β 0.079⁎⁎ 0.011 0.042 −0.065⁎ 0.688⁎⁎⁎ 0.072⁎ 0.047 −0.045

Prediction of psychopathology and outcome by baseline measures: GSIt0 and GSIt1, predicted by age, sex, TAS-20t0 and SDt0 (and GSIt0). ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

J. Terock et al. / Comprehensive Psychiatry 62 (2015) 34–41 Table 5 Multiple linear regression modeling.

ANOVA

Age Sex ΔTAS-20 Δ SD GSIt0

GSIt1

ΔGSI

F = 322.4 P b 0.001 Adjusted R 2 0.692 0.025 0.005 −0.269⁎⁎⁎ 0.192⁎⁎⁎ 0.77⁎⁎⁎

F = 85.97 P b 0.001 Adjusted R 2 0.373 −0.036 −0.007 0.384⁎⁎⁎ −0.274⁎⁎⁎ 0.226⁎⁎⁎

Prediction of outcome by change scores: GSI follow-up and GSIΔ by age, sex, TAS-20Δ, SDΔ and GSI baseline. *p b 0.05. ⁎⁎p b 0.01. ⁎⁎⁎ p b 0.001.

regulation [27]. Alexithymic persons experience emotional distress as somatic symptoms which cannot be interpreted in a meaningful way, thus impairing the individual to adapt to stressful situations. Especially the ‘difficulties in identifying feelings’ factor showed to be a strong predictor of psychopathology, underlining the importance of insight and the identification of one’s feelings for the regulation of emotions [20]. Additionally, alexithymic patients are more regularly involved in social conflicts, as the differentiation of one’s own emotions and those of other persons is necessary to manage social situations [28]. Alexithymia has been reported to be associated with primitive defense styles like projection, denial and passive–aggressive behavior [29,30]. Therefore, alexithymia is considered to go along with the reduced ability to manage strong affections and interpersonal relations. Ogrodniczuk et al. suggested that these impairments cause rather global and undifferentiated emotions which have lost their value to guide one’s adaptive behavior [8]. Continuing these thoughts, it is considerable that the identification of differentiated feelings represents a pre3.000

*** 2.500

***

***

***

2.000

***

1.500 1.000 0.500 0.000 Group 1

Group 2

Group 3

GSI t0 Mean (SD)

Group 4

Total Sample

GSI t1 Mean (SD)

Fig. 1. Data of general linear model (GLM) after extracting of four subgroups: Group 1: AP, low SD, (N = 125); Group 2: AP, high SD (N = 106); Group 3: NAP, low SD (N = 255); Group 4: NAP, high SD (N = 230). Significant differences (***p b 0.001) appear within each group between GSIt0 and GSIt1. No significant interaction between groups and change of GSI was found. (AP = Alexithymic patients: TAS-20 total score N60; NAP = Non-alexithymic patients: TAS-20 total score b61; “high/low” SD: Median split).

39

condition for the development of individual goals and the according behavior, which are features of SD. Despite the significant relevance for the treatment outcome, only few studies have examined psychological interventions aiming at the improvement of alexithymic symptoms. Speaking generally, it is believed that insight-oriented or interpretive approaches demand skills that are rather underdeveloped in alexithymic patients and that more directive or supportive approaches fit better with the alexithymic style of thinking. Ogrodniczuk et al. reported of greater reduction of alexithymia after supportive psychotherapy compared to therapy with an interpretive approach [31]. In a comparison of supportive therapy and traditional psychodynamic therapy, both approaches were associated with poor outcome in alexithymic patients [8]. Still, results are few and show some inconsistencies while alexithymic patients remain challenging in clinical treatment. SD is described as the ability to control and adapt behavior according to chosen goals. Persons with low SD typically feel that most things in their lives are influenced by other persons or factors beyond their control. They are characterized as immature, destructive and socially badly integrated [32]. They develop an irresponsible and unreliable behavior in social situations. Cloninger described SD as closely related to the concept of self-efficacy, suggesting that people with low SD show a reduced ability to solve difficult tasks and situations [33]. Thus, considering the characteristics of social behavior, the style of ego-defense and the regulation of one’s emotions, alexithymia and low SD seem to share various features. There is converging evidence that SD can be successfully addressed in psychotherapeutic treatment programs: Anderson et al. could show that a program of cognitive–behavioral therapy for bulimia nervosa patients could improve SD scores, which predicted positive treatment outcome [34]. Dalle Grave et al. found SD to be improved after CBT for eating disorders [35]. Mörtberg et al. compared in a sample of patients with social phobia cognitive–behavioral interventions with treatment as usual finding an increase of SD only in the CBT group [36]. However, despite the close relationship of SD and alexithymia, no studies have examined if an improvement of SD goes along with an improvement of alexithymia. Our results show that low SD is a common problem among alexithymic patients and that both personality traits independently predict psychopathology. Especially the difficulties in identifying feelings factor of alexithymia proved to be a major predictor of psychopathology. These findings confirmed our hypotheses (i) and (ii) and replicated results of various studies that found a strong inverse correlation of TAS-20 and self-directedness [11,37]. Grabe et al. found particularly low resourcefulness and low responsibility of SD to be linked with lack of fantasy and difficulties in identifying feelings, which conceivably results in a reduced ability to cope with emotionally stressful situations [11]. Picardi et al. suggested a relationship of the deficit in emotion regulation, a feature of alexithymia and the

40

J. Terock et al. / Comprehensive Psychiatry 62 (2015) 34–41

reduced ability to control and adapt behavior according to chosen goals, which characterizes low SD [12]. They further speculate that impoverished fantasy life is related to a reduced resourcefulness, found in individuals low in SD. In summary, these findings bring further support for the assumption of a close relationship between alexithymia and SD and highlight the role of deficits in emotion regulation and maladaptive coping skills in the etiology of psychopathology, especially depression. Regarding the prediction of treatment outcome, SD but not TAS-20 total score emerged as significant predictor after adjusting for GSIt0 scores, whereby our third hypothesis was only partially confirmed. On a subscale level, DIF proved to significantly predict of treatment outcome. The result that alexithymic patients, although showing higher GSI scores at both measurement times, could benefit from treatment to a similar degree as non-alexithymic patients, is contrary to previous studies reporting poor treatment outcome in psychodynamical therapies [6]. In our study, although therapies were psychodynamically oriented, CBT elements were included in the treatment approach, which may have influenced the outcome in a beneficial way. Furthermore, these findings lend support to previous findings by Grabe et al., who found a significant symptom reduction in alexithymic patients in a modified psychodynamic treatment setting, where therapists take in a more active part and give specific attention to the identification and verbalization of feelings [10]. Our results emphasize the importance to strengthen SD in the psychotherapeutic process and are consistent with earlier findings showing pretreatment SD scores to predict response to cognitive–behavioral therapy [38] and to pharmacological antidepressant therapy [39]. These findings bring up the question if TAS-20 and SD scores could be improved during treatment and if a possible improvement of TAS-20 and SD is followed by a more beneficial treatment outcome in terms of GSIt1. Our results revealed that the change scores for TAS-20 and SD have significant predictive power for GSIt1 also after adjusting for GSIt0, thus confirming hypothesis (iv). Although the observed effects were moderate, they emphasize the crucial role of SD for the treatment outcome. However, taking the association of alexithymia and SD into account, it may be useful to focus on alexithymic symptoms, especially the identification of feelings in order to enable the development of personal goals and thus improve SD. CBT interventions which have been shown in previous studies to increase SD scores could be applied in addition. Still, to our knowledge, no treatment programs specifically aiming at the improvement of SD are available. Considering the relevance of SD for the treatment outcome, therapeutic approaches specifically aiming at the improvement of SD are needed. Contrary to our fifth hypothesis, our results did not show an interaction between alexithymia and SD in their impact on the psychopathology and treatment outcome (GSIt0: df = 289, 426, p = 0.405; GSIt1: df = 121, 594 p = 0.498 and GSIΔ: df = 121, 594; p = 0.688). This indicates that the concepts of

alexithymia and SD represent different processes in emotion regulation and adaption to stressful situations. However, as each trait showed to be independently associated with psychopathology and they frequently co-occur, additive effects of alexithymia and low SD on psychopathology and treatment outcome are likely to emerge. Our results were acquired from a naturalistic treatment setting and included a large sample with a variety of diagnoses. Thus, some limitations need to be considered. First, we decided to investigate the outcome of a mixed sample without focusing on particular diagnoses, comorbidities and interactions between these diagnoses. However, the majority of patients suffered from depressive disorders, forming a relatively homogenous diagnostic sample. Second, concerning the personality factors described by the TCI, we only assessed SD as a major personality dimension in psychotherapeutic processes. Therefore, no information on the other TCI dimensions, especially harm avoidance and reward dependence was available. Future studies should take a closer look at other personality factors of the TCI, especially HA and RD. Third, all applied instruments were self-report questionnaires. Although extensively used in research and clinical practice and well-validated, it may be challenging especially for alexithymic patients, characterized by a reduced affective insight, to report adequately about their emotional state. Finally, the treatment program varied between the patients in its approach. While psychodynamically oriented psychotherapy represented the basic alignment for all dayclinics, for some patients CBT interventions were included and may have even have formed the central therapeutic element for some patients. Thus, conclusions regarding specific treatment approaches may not be drawn. In conclusion, our findings emphasize the predictive value of alexithymia, especially the DIF factor, and SD for psychopathology. Thus, they underline the role of identifying and cognitively process feelings in the etiology of psychopathology, especially depressive states. Low SD is shown to be a common feature of alexithymic patients and a major factor influencing the treatment outcome, worth to be addressed in psychotherapeutic treatments. We suggest that assessing temperament and character traits prior to the beginning of treatment may be particularly useful for alexithymic patients. However, there is still a considerable need for psychotherapeutic approaches aiming at improving SD.

References [1] Sifneos PE. The prevalence of “alexithymic” characteristics in psychosomatic patients. Psychother Psychosom 1973;22:255-62. [2] Hintikka J, Honkalampi K, Lehtonen J, Viinamäki H. Are alexithymia and depression distinct or overlapping constructs? A study in a general population. Compr Psychiatry 2001;42:234-9. [3] Honkalampi K, Hintikka J, Laukkanen E, Viinamäki JLH. Alexithymia and depression: a prospective study of patients with major depressive disorder. Psychosomatics 2001;42:229-34. [4] Hendryx MS, Haviland MG, Shaw DG. Dimensions of alexithymia and their relationships to anxiety and depression. J Pers Assess 1991;56:227-37.

J. Terock et al. / Comprehensive Psychiatry 62 (2015) 34–41 [5] Saarijärvi S, Salminen JK, Toikka TB. Alexithymia and depression: a 1-year follow-up study in outpatients with major depression. J Psychosom Res 2001;51:729-33. [6] Leweke F, Bausch S, Leichsenring F, Walter B, Stingl M. Alexithymia as a predictor of outcome of psychodynamically oriented inpatient treatment. Psychother Res 2009;19:323-31. [7] McCallum M, Piper WE, Ogrodniczuk JS, Joyce AS. Relationships among psychological mindedness, alexithymia and outcome in four forms of short-term psychotherapy. Psychol Psychother 2003;76:133-44. [8] Ogrodniczuk JS, Piper WE, Joyce AS. Effect of alexithymia on the process and outcome of psychotherapy: a programmatic review. Psychiatry Res 2011;190:43-8. [9] Spek V, Nyklíček I, Cuijpers P, Pop V. Alexithymia and cognitive behaviour therapy outcome for subthreshold depression. Acta Psychiatr Scand 2008;118:164-7. [10] Grabe HJ, Frommer J, Ankerhold A, Ulrich C, Groger R, Franke GH, et al. Alexithymia and outcome in psychotherapy. Psychother Psychosom 2008;77:189-94. [11] Grabe HJ, Spitzer C, Freyberger HJ. Alexithymia and the temperament and character model of personality. Psychother Psychosom 2001;70:261-7. [12] Picardi A, Toni A, Caroppo E. Stability of alexithymia and its relationships with the “big five” factors, temperament, character, and attachment style. Psychother Psychosom 2005;74:371-8. [13] Lee YJ, Yu S-H, Cho S-J, Cho IH, Koh S-H, Kim SJ. Direct and indirect effects of the temperament and character on alexithymia: a pathway analysis with mood and anxiety. Compr Psychiatry 2010;51:201-6. [14] Jylhä P, Isometsä E. Temperament, character and symptoms of anxiety and depression in the general population. Eur Psychiatry 2006;21:389-95. [15] Izci F, Gültekin BK, Saglam S, Koc MI, Zincir SB, Atmaca M. Temperament, character traits, and alexithymia in patients with panic disorder. Neuropsychiatr Dis Treat 2014;10:879-85. [16] Cloninger CR, Svrakic DM, Przybeck TR. Can personality assessment predict future depression? A twelve-month follow-up of 631 subjects. J Affect Disord 2006;92:35-44. [17] Mörtberg E, Andersson G. Predictors of response to individual and group cognitive behaviour therapy of social phobia. Psychol Psychother 2014;87:32-43. [18] Rowe S, Jordan J, McIntosh V, Carter F, Frampton C, Bulik C, et al. Dimensional measures of personality as a predictor of outcome at 5year follow-up in women with bulimia nervosa. Psychiatry Res 2011;185:414-20. [19] Corchs F, Corregiari F, Ferrão YA, Takakura T, Mathis ME, Lopes AC, et al. Personality traits and treatment outcome in obsessive– compulsive disorder. Rev Bras Psiquiatr 2008;30:246-50. [20] Conrad R, Wegener I, Imbierowicz K, Liedtke R, Geiser F. Alexithymia, temperament and character as predictors of psychopathology in patients with major depression. Psychiatry Res 2009;165:137-44. [21] Derogatis LR. SCL-90-R manual. Baltim Clin Psychom Res 1977.

41

[22] Derogatis LR. SCL-90-R: administration, scoring and procedures manual for the revised version and other instruments of the psychopathology rating scale series. Clin Psychom Res 1992. [23] Franke GH, Derogatis LR. Die Symptom-Checkliste von Derogatis: Manual. SCL-90-R; deutsche Version. Beltz Test; 1995. [24] Bach M, Bach D, de Zwaan M, Serim M, Böhmer F. Validation of the German version of the 20-item Toronto Alexithymia Scale in normal persons and psychiatric patients. Psychother Psychosom Med Psychol 1996;46:23-8. [25] Bagby RM, Taylor GJ, Parker JD. The Twenty-item Toronto Alexithymia Scale–II. Convergent, discriminant, and concurrent validity. J Psychosom Res 1994;38:33-40. [26] Taylor GJ. Recent developments in alexithymia theory and research. Can J Psychiatry 2000;45:134-42. [27] Lane RD. Neural substrates of implicit and explicit emotional processes: a unifying framework for psychosomatic medicine. Psychosom Med 2008;70:214-31. [28] Guttman H, Laporte L. Alexithymia, empathy, and psychological symptoms in a family context. Compr Psychiatry 2002;43:448-55. [29] Parker JD, Taylor GJ, Bagby RM. Alexithymia: relationship with ego defense and coping styles. Compr Psychiatry 1998;39:91-8. [30] Taylor GJ, Bagby RM, Parker JDA. Disorders of affect regulation: alexithymia in medical and psychiatric illness. Cambridge University Press; 1999. [31] Ogrodniczuk JS, Joyce AS, Piper WE. Change in alexithymia in two dynamically informed individual psychotherapies. Psychother Psychosom 2013;82:61-3. [32] Spitzer C, Siebel-Jurges U, Barnow S, Grabe HJ, Freyberger HJ. Alexithymia and interpersonal problems. Psychother Psychosom 2005;74:240-6. [33] Cloninger CR, Svrakic DM, Przybeck TR. A psychobiological model of temperament and character. Arch Gen Psychiatry 1993;50:975-90. [34] Anderson CB, Joyce PR, Carter FA, McIntosh VV, Bulik CM. The effect of cognitive–behavioral therapy for bulimia nervosa on temperament and character as measured by the temperament and character inventory. Compr Psychiatry 2002;43:182-8. [35] Dalle Grave R, Calugi S, Brambilla F, Abbate-Daga G, Fassino S, Marchesini G. The effect of inpatient cognitive–behavioral therapy for eating disorders on temperament and character. Behav Res Ther 2007;45:1335-44. [36] Mörtberg E, Bejerot S, Aberg Wistedt A. Temperament and character dimensions in patients with social phobia: patterns of change following treatments? Psychiatry Res 2007;152:81-90. [37] Celikel FC, Kose S, Erkorkmaz U, Sayar K, Cumurcu BE, Cloninger CR. Alexithymia and temperament and character model of personality in patients with major depressive disorder. Compr Psychiatry 2010;51:64-70. [38] Bulik CM, Sullivan PF, Fear JL, Pickering A. Outcome of anorexia nervosa: eating attitudes, personality, and parental bonding. Int J Eat Disord 2000;28:139-47. [39] Sato T, Hirano S, Narita T, Kusunoki K, Kato J, Goto M, et al. Temperament and character inventory dimensions as a predictor of response to antidepressant treatment in major depression. J Affect Disord 1999;56:153-61.