Ecsa.ucl.ac.be

Regular Article
An Evaluation of the Absolute and
Relative Stability of Alexithymia in
Patients with Major Depression
Olivier Lumineta R. Michael Bagbyb Graeme J. Taylorc aDepartment of Psychology, University of Louvain and Belgian National Fund for Scientific Research, Louvain,Belgium; bDepartment of Psychiatry, University of Toronto and Centre for Addiction and Mental Health – ClarkeSite, and cDepartment of Psychiatry, University of Toronto and Mount Sinai Hospital, Toronto, Ont., Canada Key Words
with antidepressant medication. Paired t tests and corre- Alexithymia W Depression W Temporal stability lational analyses were performed to evaluate absolutestability and relative stability in alexithymia. Hierarchicalregression analyses were then used to assess the degree Abstract
to which the relative stability in alexithymia scores was Background: Previous studies demonstrating an associa-
related to the severity of depressive symptoms, and the tion between alexithymia and depression have led to the degree to which changes in alexithymia scores could proposal that alexithymia may be a state-dependent be attributed to changes in depression scores. Results:
phenomenon rather than a stable and enduring person- Alexithymia scores changed significantly from baseline ality trait. Several longitudinal studies have provided to follow-up, indicating a general lack of absolute stabili- support for a trait view of alexithymia, but most of these ty. There was, however, strong evidence of relative sta- studies evaluated absolute stability only (i.e., the extent bility, as alexithymia scores at baseline correlated signifi- to which alexithymia scores change over time) and did cantly with alexithymia scores at follow-up and were not examine the relative stability of alexithymia (i.e., the also a significant predictor of follow-up alexithymia extent to which relative differences among individuals scores, after partialling the effects of depression severity.
remain the same over time) in the context of changes in Conclusions: Although alexithymia scores may change
illness symptomatology. The present study evaluated in the presence of large changes in the severity of both absolute stability and relative stability of alexithy- depressive symptoms, the finding of relative stability of mia in depressed patients who experienced a marked alexithymia supports the view that this construct is a sta- reduction in the severity of depressive symptoms. Meth-
ble personality trait rather a state-dependent phenome- ods: Forty-six psychiatric outpatients with major depres-
sion were assessed for alexithymia and depression with the 20-item Toronto Alexithymia Scale and the HamiltonRating Scale for Depression at the start of treatment(baseline) and after 14 weeks of treatment (follow-up) Centre for Addiction and Mental Health – Clarke Site Tel. +1 416 535 8501, ext. 6939, Fax +1 416 979 6821, E-Mail [email protected] Introduction
Contrasting findings were obtained in a similar study by Honkalampi et al. [19], who followed 169 psychiatric Over the past decade numerous studies have yielded outpatients with DSM-III-R diagnoses of major depres- empirical evidence that alexithymia is associated with a sion or other depressive disorders for a period of 6 variety of medical and psychiatric disorders, in particular months. The patients were treated with antidepressant with substance use disorders, eating disorders, panic dis- medication and some also received benzodiazepine drugs order, essential hypertension, functional gastrointestinal and/or psychotherapy. Instead of analyzing TAS-20 disorders, and inflammatory bowel disease [1, 2]. Uncer- scores as a continuous variable, however, these re- tainty remains, however, as to whether alexithymia is a searchers categorized their patients into alexithymic and predisposing or vulnerability factor that influences the nonalexithymic groups based on a TAS-20 cutoff score of onset or course of these disorders or merely a state reac- 661 for alexithymia. Whereas 39% of their depressed tion to the presence of a disorder or any accompanying patients were alexithymic at baseline, only 23% were still alexithymic at follow-up. Regression analyses performed Arguments for alexithymia being a state-dependent on the entire sample revealed that BDI scores explained phenomenon are based, in part, on evidence that the 20% of the variation in TAS-20 scores at baseline and widely used Toronto Alexithymia Scale (TAS) [6] and its 42% at follow-up. Honkalampi et al. [19] concluded from revised 20-item version (TAS-20) [7] correlate positively these analyses that alexithymia is a secondary phenome- with measures of anxiety, depression, somatization, and non rather than a stable personality trait among depressed neuroticism (negative affectivity) in studies using cross- patients. This conclusion may be premature, however, sectional designs [1, 8–12]. Other research, however, has since the association between BDI and TAS-20 change demonstrated that the alexithymia construct is distinct and separate from the constructs of depression [13], Although two other longitudinal studies have yielded somatization [14], and neuroticism [1].
support for the stability of alexithymia, the sample sizes Studies using longitudinal designs have also yielded were small, and the populations studied were patients conflicting findings as to whether alexithymia is a stable with psychoactive substance dependence [20] and under- and enduring personality trait or a phenomenon that graduate university students [21] rather than patients arises secondary to depression or other psychological dis- with depressive disorders. Furthermore, the follow-up tress. Salminen et al. [3], for example, evaluated the sta- intervals in these two studies were rather short (between 4 bility of alexithymia scores over a 1-year period in 54 psy- chiatric outpatients diagnosed primarily with anxiety and There are two important limitations to the above longi- mood disorders. Sixty-five percent of the patients re- tudinal studies. First, these studies evaluated the absolute ceived psychiatric treatment (mostly pharmacotherapy stability of alexithymia only and did not distinguish it and supportive psychotherapy) during the year, while the from the concept of relative stability. Absolute stability remaining 35% received no treatment. An abbreviated refers to the extent to which personality scores change version of the Brief Symptom Inventory (BSI)) [15] indi- over time, whereas relative stability indicates the extent cated a significant decrease in psychological distress at the to which the relative differences among individuals re- end of this 1-year period, whereas alexithymia, measured main the same over time [22]. Second, most studies exam- by the TAS, did not change significantly.
ining absolute change or stability on measures of alexithy- More recently, Saarijärvi et al. [16] followed 149 psy- mia did not examine whether any of the small (i.e., non- chiatric outpatients with DSM-III-R diagnoses of major significant) changes that did occur were related to changes depression for a period of 1 year. In this study, 94% of the in psychological distress. Only Pinard et al. [20] examined patients received treatment with psychotherapy and/or and reported correlations between alexithymia change pharmacotherapy. The patients completed the BSI and scores and psychological distress change scores; these cor- the TAS-20 prior to treatment and after 1 year and were relations were significant, suggesting that alexithymia is assessed for depression with the Beck Depression Inven- associated with state-dependent distress.
tory (BDI) [17] and the Hamilton Rating Scale for De- The concept of absolute stability in personality is pression (HRSD) [18]. Whereas scores on the depression derived primarily from clinical studies with the use of and psychological distress measures had decreased signifi- change in mean scores over the course of a clinical trial. In cantly at the follow-up testing, TAS-20 scores showed no these studies, statistically significant change in personali- significant change at the end of 1 year.
ty test scores over the course of treatment is regarded as indication of no stability. In contrast, personality re- For each group, alexithymia, measured by the TAS-20, searchers generally emphasize the degree of relative sta- and depression and anxiety, measured by the Hospital bility in personality scores assessed over longer periods of Anxiety and Depression Scale [27], were assessed at base- time; this type of stability is estimated through the use of line and at the 6-month follow-up. The relative stability of measures of covariation, such as correlation coefficients.
alexithymia was demonstrated by very high correlations In these studies, statistically significant test-retest correla- between TAS-20 scores measured at baseline and follow- tions indicate that the personality trait is stable. As an up for the total sample (r = 0.95, p ! 0.001) as well as example, some clinical researchers argue that the person- within each of the groups (unchanged, r = 0.95, improved, ality trait of neuroticism is not stable as scores on mea- r = 0.96, worsened, r = 0.95, all p ! 0.001). Moreover, sures of neuroticism and depressive symptoms typically while significant changes were observed within the groups decrease significantly following recovery from a depres- for both anxiety and depression (i.e., small changes for the sive episode [23]. In contrast, on the basis of large test- unchanged group, decreases for the improved group, and retest correlations, most personality researchers conclude increases for the worsened group), the degree of alexithy- that neuroticism demonstrates considerable stability over mia was not influenced by the state effects of the level of time [24]. To evaluate whether neuroticism or any other disease activity. The demonstration of relative stability of personality characteristic represents a potential vulnera- alexithymia in the context of change in disease activity bility factor for a medical or psychiatric illness, however, supports the view that alexithymia could serve as a possi- it is necessary to demonstrate the relative stability of the characteristic in the context of acute change. This was Further support for the relative stability of alexithymia done in a recent study by Santor et al. [22], who assessed comes from studies demonstrating test-retest reliability of test-retest correlations between self-report ratings for neu- self-report measures of the construct. Using data collected roticism and extraversion completed by depressed pa- from various nonclinical populations at two separate tients when they were acutely depressed and 5 weeks later, times, investigators have reported Pearson correlation when recovered. Although the patients’ neuroticism coefficients ranging from 0.77 after 3 weeks to 0.83 after 3 scores decreased (as did their depression scores) and months for the TAS-20 [7, 28], and from 0.75 after 5 extraversion scores increased significantly over 5 weeks of weeks to 0.68 after 7 months for the TAS [6, 29].
treatment, indicating a lack of absolute stability, the test- Despite the accumulation of data that supports view- retest correlations for neuroticism and extraversion scores ing alexithymia as a stable personality trait, methodologi- indicated a high degree of relative stability. Regression cal and statistical limitations compromise the interpreta- analyses also indicated that changes in neuroticism and bility of the findings from some of these studies. For extraversion scores over the course of treatment were example, in some studies the magnitude of change in dis- modestly or not at all accounted for by changes in depres- tress or illness was of relatively small size or not assessed [16, 21, 26], or treatment effects were not specified [3]. In It is interesting to note that unlike neuroticism and addition, almost all previous studies evaluated either the extraversion, which are considered enduring personality absolute stability or the relative stability of alexithymia traits despite their lack of absolute stability (but because rather than both types of stability. To our knowledge, only of evidence of relative stability), alexithymia is regarded one study has examined both absolute and relative stabili- by some researchers [25] as merely a state reaction even ty of alexithymia. This was a study in which 41 patients though data from several studies indicate that changes in with bulimia nervosa were administered the TAS and the mean alexithymia scores are nonsignificant in the context HRSD, and assessed for bulimic symptomatology, before of acute changes in levels of psychological distress. More- and after 10 weeks of treatment with an antidepressant over, there are also preliminary data that support the rela- medication [30]. Although there was a significant im- tive stability of alexithymia. Porcelli et al. [26] evaluated provement in bulimic symptomatology at the end of 10 the relative stability of alexithymia in 104 patients with weeks, there were no significant changes in mean alexithy- inflammatory bowel disease (IBD) over a 6-month period.
mia and depression scores. The correlation between TAS The patients were treated for IBD but not for the psycho- scores at baseline and at follow-up was significant (r = logical distress usually associated with the disease. Six 0.58, p ! 0.001). Although these results support the abso- months after the initial assessment, three groups of pa- lute and relative stability of alexithymia, as Schmidt et al.
tients were distinguished based on changes in the level of [30] point out, it is possible that the continuing high alexi- disease activity (unchanged, improved, and worsened).
thymia scores were due to persistent symptoms as the majority of patients had not fully recovered by the end of 39.07 years (SD = 10.92) and the mean level of education was 16.11 years (SD = 2.27). The mean number of previous major depressiveepisodes was 2.29 (SD = 1.19), and the mean age at onset of the first In the present study we evaluated both the absolute sta- major depressive episode was 25.43 years (SD = 12.39). The inter- bility and the relative stability of alexithymia in a group of viewers who administered the HRSD at follow-up were ‘masked’ psychiatric outpatients with major depression who were with respect to patients’ baseline and follow-up TAS-20 scores, but being treated with antidepressant medication over a 14- week period. Large changes in depressive symptoms were anticipated as previous investigators treating depressed Alexithymia was assessed with the TAS-20, which is a revised and patients with the same treatment protocol have reported improved version of the original TAS. It has demonstrated reliability major reductions in HRSD scores over even a shorter and validity, and is currently the most widely used measure of the period (5 weeks) [31]. We also examined the more specific alexithymia construct [1, 7, 33]. The severity of depressive symptoms issue concerning the degree to which the relative stability was assessed with the 17-item version of the HRSD. The HRSD is aclinician-related, semistructured interview and is the most widely of alexithymia may be accounted for by relative stability used and accepted measure of depression in clinical studies with in depression severity. In addition, we examined how depressed patient samples [18]. It is employed in most clinical trials changes in alexithymia are related to changes in depres- of depression to detect change in depression severity [34].
Statistical AnalysesThe following specific issues regarding the stability of alexithy- mia during an acute treatment phase for depression were addressed: (1) whether relative stability in alexithymia can be observed in a con-text in which absolute change in alexithymia scores over a short peri- od of time (14 weeks) is likely; (2) the degree to which the relative The subjects were drawn from a clinical database of outpatients stability in alexithymia scores is related to severity of depressive who were referred for treatment at a Depression Clinic located in a symptoms; (3) the degree to which changes in alexithymia scores can university-based tertiary care facility. All patients involved in the be attributed directly to changes in depression scores.
clinical database are treated with a standardized treatment protocol The first issue was addressed using paired t tests to assess absolute that involves pretreatment administration of the Structured Clinical change in TAS-20 and HRSD scores after 14 weeks of treatment with Interview for DSM-IV Axis I Disorders, Patient Edition (SCID-P) antidepressant medication. This analysis was supplemented by cal- [32], the HRSD, and a variety of self-report measures. To be eligible culating effect sizes (Cohen’s d) to determine the magnitude of for entry into the clinical database, all patients must meet DSM-IV change. Pearson correlations and intraclass correlations were used to diagnostic criteria for a primary, nonpsychotic, major depression as assess the relative stability of TAS-20 and HRSD scores across time determined by the SCID-P, score 16 or greater on the HRSD, not (i.e., test-retest). The association between TAS-20 and HRSD scores have a concurrent active medical illness (as determined by the refer- at baseline and follow-up was also assessed using Pearson correla- ring physician), and be free of antidepressant medications for at least 2 weeks (4 weeks for fluoxetine). Patients receive one of several anti- The second issue was examined using hierarchical regression depressant medications, using standardized treatment protocols in analyses, as was done in the study by Santor et al. [22]. Given that accordance with prespecified dose algorithms; they are routinely measures of alexithymia have been shown to have moderate to strong reassessed with the HRSD after 14 weeks of treatment. The various correlations with measures of depression in cross-sectional designs and possible medications and dose ranges are: (a) bupropion 100– (i.e., within but not across time) [8–12], the relative stability among 300 mg/day, (b) citalopram 20–60 mg/day, (c) clomipramine 75–300 individuals’ differences in alexithymia from baseline to follow-up mg/day, (d) desipramine 50–300 mg/day, (e) fluoxetine 20–60 mg/ may be attributable to individual differences in depression severity day, (f) imipramine 75–300 mg/day, (g) moclobemide 300–600 mg/ at either baseline or follow-up. To this end, we built a regression day; (h) nefazodone 150–600 mg/day, (i) paroxetine 10–60 mg/day, model in which follow-up alexithymia scores served as the criterion (j) phenelzine 35–90 mg/day, (k) sertraline 50–200 mg/day, (l) tranyl- variable. The predictor variables, in order of entry were: (1) a ‘block’ cypromine 20–60 mg/day and (m) venlafaxine 75–300 mg/day. Pa- containing baseline and follow-up depression scores, and (2) baseline tients receive one or two trials of antidepressant medication at ade- alexithymia scores. Showing that variance in follow-up alexithymia quate doses selected at the discretion of the treating psychiatrist.
scores can be predicted by baseline alexithymia scores beyond the Antidepressant medications are changed if the patient experiences effects due to baseline and follow-up depression scores demonstrates side effects or fails to respond to the first trial. In the present study, that the relative stability of alexithymia scores cannot be accounted no patients were treated with concurrent psychotherapy. All patients for in terms of depression severity.
in this study were treated with either moclobemide, paroxetine, ser- The third issue was also examined using hierarchical regression analysis. While absolute change in both depression and alexithymia For the purpose of this study, patients entering the clinical data- scores might be adequately detected using paired t tests, it is unclear base were also administered the TAS-20. After 46 patients (19 men, from this type of analysis to what extent change in alexithymia scores 27 women) had completed the TAS-20 at baseline and at the 14-week can be accounted for by changes in depression severity. To assess for follow-up, the TAS-20 was removed from the battery of measures this potential effect, we built a regression model in which change routinely collected. The mean age of this group of 46 patients was scores in alexithymia (i.e., baseline TAS-20 scores minus follow-up TAS-20 scores) served as the criterion variable, and change scores in Estimating the Relative Stability of Alexithymia depression (i.e., baseline HRSD scores minus follow-up HRSD scores) served as the predictor variable. All analyses were performed As indicated above, we used hierarchical regression using the Windows version of the Statistical Package for SocialSciences (SPSS) 10 [35].
analyses to examine the extent to which stability in alexi-thymia scores may be accounted for by individual differ-ences in depression severity. The criterion variable was follow-up TAS-20 scores, and the predictor variables wereHRSD scores at baseline and follow-up and TAS-20 Gender differences were examined for alexithymia, A statistically significant amount of variance in follow- severity of depression and duration of current depressive up TAS-20 scores was predicted from the block including episode, total number of previous depressive episodes, HRSD scores at baseline and follow-up [R2 = 0.15, F(2, age at baseline, marital status, years of education, and 42) = 3.58, p ! 0.05, ß = 0.39]. TAS-20 scores at follow-up, socioeconomic status. None of these comparisons were however, were predicted by TAS-20 scores at baseline found to be significant (p ! 0.05); therefore, all the results beyond the variance explained by depression severity for the men and women were considered together.
[R2chg = 0.33, Fchg(1, 41) = 26.41, p ! 0.0001, ß = 0.59].
These results provide strong evidence for relative stability Absolute Change Scores and Relative Stability Next, we assessed absolute change scores for depres- sion severity and alexithymia. The mean HRSD score at Modelling Change in Alexithymia and Depression baseline was 20.87 (SD = 3.76) and at follow-up was 9.56 For this analysis change scores for the TAS-20 from (SD = 8.30), reflecting a substantial decrease in depres- baseline to follow-up served as the criterion variable and sion severity from baseline to follow-up; mean change change scores in HRSD from baseline to follow-up served score = –11.31, t(45) = 10.54, p ! 0.0001. There was also a as the predictor variable. Only 3% of the variance in significant decrease in TAS-20 scores. The mean TAS-20 change scores for the TAS-20 was accounted for by the score at baseline was 51.91 (SD = 16.50) and at follow-up variance in change scores for the HRSD, which was statis- was 46.80 (SD = 16.50); mean change score = –5.11, t(45) tically nonsignificant [F(1, 43) = 2.50, p 1 0.12]. These = 2.61, p ! 0.01. Effect sizes (Cohen’s d) were calculated results suggest that changes in TAS-20 scores cannot be for the TAS-20 and HRSD mean change scores. For the attributed to changes in HRSD scores.
TAS-20, the change in mean scores from baseline to fol-low-up produced an effect size of 0.78; for the HRSD, theeffect size was more than 4 times greater (Cohen’s d = Discussion
To assess evidence of relative stability, we computed The present study evaluated both the absolute stability correlations between TAS-20 scores at baseline and fol- and the relative stability of alexithymia in the context of low-up and HRSD scores at baseline and at follow-up.
significant changes in the severity of depressive symp- Positive and significant correlations were found between toms. The results indicate that, although significant baseline and follow-up TAS-20 scores (r = 0.64, p ! changes in absolute levels of depression severity and alexi- 0.0001; intraclass correlation r = 0.63, p ! 0.0001) and thymia were observed, alexithymia still evidenced a high between baseline and follow-up HRSD scores (r = 0.50, degree of relative stability. This was shown first by a posi- p ! 0.001; intraclass correlation r = 0.38, p ! 0.005). How- tive and highly significant correlation between alexithy- ever, while the TAS-20 was not correlated significantly mia scores at treatment initiation (baseline) and after 14 with the HRSD at baseline (r = 0.09, p = NS, explaining weeks of aggressive treatment with antidepressant medi- approximately 1% of the variance), a positive correlation cation. Stronger evidence for relative stability was pro- was observed at follow-up (r = 0.38, p ! 0.01, explaining vided by the hierarchical regression analyses, indicating approximately 14% of the variance), suggesting that any that variance in follow-up alexithymia scores can be pre- relative stability for TAS-20 scores could be attributed to dicted from alexithymia scores at baseline over the effects shared variance with the relative stability in HRSD contributed by baseline and follow-up depression severity While several previous studies showed evidence of score moves from one discrete category to another in rela- absolute stability in alexithymia [3, 16, 20, 21], the present results indicated absolute changes in alexithymia, One unexpected finding in the present study was the which could suggest that alexithymia is a state reaction.
absence of a significant correlation between the measures The magnitude of absolute change in alexithymia scores of alexithymia and depression at baseline, although the was small, however, in comparison with the magnitude of two measures did correlate significantly at follow-up. This absolute change in depression severity. In contrast, in the finding can probably be explained by the truncated distri- earlier study of patients with major depression who bution of scores for the HRSD at baseline, which ranged showed absolute stability in alexithymia at the end of a from 16 to 29. This distribution was the result of the 1-year period [16], the magnitude of change in depression inclusion criterion that the patients with major depression severity was small, perhaps because the majority of pa- needed to score greater than 16 to be diagnosed as clinical- tients were treated less aggressively with only low-dosage ly depressed. At follow-up, no constraint was imposed on antidepressant medication and/or supportive discussions HRSD scores, and the resultant distribution was much with a mental health professional once or twice a month broader than it was at baseline. The narrower range of [J.K. Salminen, personal commun., August 3, 2000]. In scores at baseline compared to the range of scores at fol- the present study, absolute change in alexithymia was not low-up most likely constrained the upper limit of the cor- unexpected as a large decrease in depression severity was anticipated in response to intensive pharmacotherapy.
Although the present study shows strong evidence for Although our findings need to be replicated with larger relative stability of the alexithymia construct, this does samples of depressed patients, it seems that absolute sta- not necessarily mean that alexithymic characteristics can- bility of alexithymia is unlikely in the presence of large not be modified by specific therapeutic interventions.
changes in depression severity. In previous studies with Experienced clinicians recommend individual and/or depressed patients, university students, and eating disor- group psychotherapies that focus on educating patients der patients, changes in psychological distress were rela- about their emotions and on developing imaginative ca- tively small, and the degree of alexithymia did not change pacities as well as an ability to identify and describe sub- significantly [16, 21, 30]. The present study, by showing jective feelings [1, 36, 37]. Indeed, in a recent preliminary clear and substantial changes in symptomatology, pro- study of postmyocardial infarction patients, who received vides a more rigorous test of the stability of the alexithy- once weekly group psychotherapy for 4 months, there was mia construct than do these previous studies. Even in the a significant decrease in the mean TAS score, which there- context of the large change in depressive symptoms, the after remained stable over a 2-year follow-up period (r = relative stability of alexithymia was demonstrated, indi- 0.75) [38]. Further studies are needed, however, to verify cating its status as a stable personality trait and not a the extent to which psychotherapeutic interventions can enable individuals to reduce or adapt to deficits in the Whereas Honkalampi et al. [19] concluded that alexi- thymia does not demonstrate stability in depressed pa-tients, they failed to assess relative stability and did notexamine statistically whether changes in TAS-20 scores Acknowledgements
were correlated significantly with changes in depression Dr. Olivier Luminet is postdoctoral researcher at the Belgian scores. One reason for their failure to assess relative stabil- National Fund for Scientific Research and North Atlantic Treaty ity was their decision to create discrete alexithymic and Organization (NATO) research fellowship. His contribution to this nonalexithymic groups based on a TAS-20 cutoff score.
study was supported by grant FRFC 2.4546.97 and grant FNRS As with other personality constructs, alexithymia is best 1.5.124.00 from the Belgian National Fund for Scientific Research, conceptualized as a dimensional rather than categorical and made while he was an in-resident Post-Doctoral Fellow at theCentre for Addiction and Mental Health, Toronto, Canada.
variable, especially when examining its association withsymptom severity. Moreover, evaluating whether alexi-thymia is stable is best done by assessing if an individual’sscore on a measure of this construct relative to others’scores remains the same over a prescribed interval, evenin the context of acute symptom change. In our opinion,this is preferable to assessing whether an individual’s References
1 Taylor GJ, Bagby RM, Parker JDA: Disorders 13 Parker JDA, Bagby RM, Taylor GJ: Alexithy- 26 Porcelli P, Leoci C, Guerra V, Taylor GJ, Bag- of Affect Regulation: Alexithymia in Medical mia and depression: Distinct or overlapping by RM: A longitudinal study of alexithymia and Psychiatric Illness. Cambridge, Cambridge constructs? Compr Psychiatry 1991;32:387– and psychological distress in inflammatory bowel disease. J Psychosom Res 1996;41:569– 2 Taylor GJ: Recent developments in alexithy- 14 Bach M, Bach D, de Zwaan M: Independency mia theory and research. Can J Psychiatry of alexithymia and somatization. Psychoso- 27 Zigmond AS, Snaith RP: The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 3 Salminen JK, Saarijärvi S, Äärelä E, Tammin- 15 Derogatis LR, Melisaratos N: The brief symp- en T: Alexithymia – state or trait? One-year fol- tom inventory: An introductory report. Psychol 28 Pandey R, Mandal MK, Taylor GJ, Parker low-up study of general hospital psychiatric JDA: Cross-cultural alexithymia: Development consultation out-patients. J Psychosom Res 16 Saarijärvi S, Salminen JK, Tikka J, Toikka T: and validation of a Hindi translation of the 20- Alexithymia and depression: A one-year fol- item Toronto Alexithymia Scale. J Clin Psy- 4 Wise TN, Mann LS, Mitchell JD, Hryvniak M, low-up study in outpatients with major depres- Hill B: Secondary alexithymia: An empirical sion (abstract). J Psychosom Res 2000;48:238.
29 Linden W, Wen F, Paulhus DL: Measuring validation. Compr Psychiatry 1990;31:284– 17 Beck AT, Steer RA: Beck Depression Inventory alexithymia: Reliability, validity, and preva- Manual. San Antonio, Psychological Corpora- lence. Adv Pers Assess 1995;10:51–95.
30Schmidt U, Jiwany A, Treasure J: A controlled MacMurray JP: Alexithymia: Subscales and re- 18 Hamilton M: Development of a rating scale for study of alexithymia in eating disorders.
lationship to depression. Psychother Psycho- primary depressive illness. Br J Soc Clin Psy- 31 Bagby RM, Joffe RT, Parker JDA, Kalemba V, 6 Taylor GJ, Ryan D, Bagby RM: Toward the 19 Honkalampi K, Hintikka J, Saarinen P, Leh- Harkness KL: Major depression and the five- development of a new self-report alexithymia tonen J, Viinamäki H: Is alexithymia a perma- factor model of personality. J Personal Disord scale. Psychother Psychosom 1985;44:191– nent feature in depressed patients? Results from a 6-month follow-up study. Psychother 32 First MB, Spitzer RL, Gibbon M, Williams 7 Bagby RM, Parker JDA, Taylor GJ: The Twen- JBW: Structured Clinical Interview for DSM- ty-Item Toronto Alexithymia Scale. I. Item se- 20Pinard L, Negrete JC, Annable L, Audet N: IV Axis I Disorders, Patient Edition (SCID-P), lection and cross-validation of the factor struc- Alexithymia in substance abusers: Persistence ture. J Psychosom Res 1994;38:23–32.
and correlates of variance. Am J Addic 1996;5: chiatric Institute, Biometrics Research, 1995.
8 Bagby RM, Taylor GJ, Ryan D: Toronto Alexi- 33 Bagby RM, Taylor GJ, Parker JDA: The Twen- thymia Scale: Relationship with personality 21 Martı´nez-Sa´nchez F, Ato-Garcia M, Co´rcoles ty-Item Toronto Alexithymia Scale. II. Conver- Adam E, Huedo Medina TB, Selva España JJ: gent, discriminant, and concurrent validity. J Stability in alexithymia levels: A longitudinal 9 Deary IJ, Scott S, Wilson JA: Neuroticism, analysis on various emotional answers. Pers 34 Snaith RP: Present use of the Hamilton De- alexithymia and medically unexplained symp- pression Rating Scale: Observations on method toms. Pers Individ Diff 1997;22:551–564.
22 Santor DA, Bagby RM, Joffe RT: Evaluating of assessment in research of depressive disor- 10Honkalampi K, Hintikka J, Tanskanen A, Leh- stability and change in personality and depres- ders. Br J Psychiatry 1996;168:594–597.
tonen J, Viinamäki H: Depression is strongly sion. J Pers Soc Psychol 1997;73:1354–1362.
35 Norusis: Statistical Package for the Social associated with alexithymia in the general pop- 23 Hirschfeld MA, Klerman GL, Clayton PJ, Kell- Sciences (SPSS). Basic and Professional Statis- ulation. J Psychosom Res 2000;48:99–104.
er MB, McDonald-Scott P, Larkin BH: Assess- 11 Honkalamp K, Saarinen P, Hintikka J, Virtan- ing personality: Effects of the depressive state 36 Krystal H: Alexithymia and psychotherapy.
en V, Viinamäki H: Factors associated with on trait measurement. Am J Psychiatry 1983; alexithymia in patients suffering from depres- 37 Swiller HI: Alexithymia: Treatment utilizing sion. Psychother Psychosom 1999;68:270– 24 Costa PT Jr, McCrae RR, Zonderman AB, Bar- bano HE, Lebowitz B, Larson DM, Lebowitz B: therapy. Int J Group Psychother 1988;38:47– Cross-sectional studies of personality in a na- mensions of alexithymia and their relation- tional sample. 2. Stability in neuroticism, ex- 38 Beresnevaite M: Exploring the benefits of ships to anxiety and depression. J Pers Assess traversion and openness. Psychol Aging 1986; group psychotherapy in reducing alexithymia in coronary heart disease patients: A prelimina- 25 Ahrens S, Deffner G: Empirical study of alexi- ry study. Psychother Psychosom 2000;69:117– thymia: Methodology and results. Am J Psy-

Source: http://www.ecsa.ucl.ac.be/personnel/luminet/pdf/Stability_alexithymia.pdf