Regular Article Psychother Psychosom  An Evaluation of the Absolute and Relative Stability of Alexithymia in Patients with Major Depression Olivier Luminet R

Regular Article Psychother Psychosom An Evaluation of the Absolute and Relative Stability of Alexithymia in Patients with Major Depression Olivier Luminet R - Description

Michael Bagby Graeme J Taylor Department of Psychology University of Louvain and Belgian National Fund for Scientific Research Louvain Belgium Department of Psychiatry University of Toronto and Centre for Addiction and Mental Health Clarke Site and ID: 30299 Download Pdf

154K - views

Regular Article Psychother Psychosom An Evaluation of the Absolute and Relative Stability of Alexithymia in Patients with Major Depression Olivier Luminet R

Michael Bagby Graeme J Taylor Department of Psychology University of Louvain and Belgian National Fund for Scientific Research Louvain Belgium Department of Psychiatry University of Toronto and Centre for Addiction and Mental Health Clarke Site and

Similar presentations


Download Pdf

Regular Article Psychother Psychosom An Evaluation of the Absolute and Relative Stability of Alexithymia in Patients with Major Depression Olivier Luminet R




Download Pdf - The PPT/PDF document "Regular Article Psychother Psychosom An..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.



Presentation on theme: "Regular Article Psychother Psychosom An Evaluation of the Absolute and Relative Stability of Alexithymia in Patients with Major Depression Olivier Luminet R"— Presentation transcript:


Page 1
Regular Article Psychother Psychosom 2001;70:254260 An Evaluation of the Absolute and Relative Stability of Alexithymia in Patients with Major Depression Olivier Luminet R. Michael Bagby Graeme J. Taylor Department of Psychology, University of Louvain and Belgian National Fund for Scientific Research, Louvain, Belgium; Department of Psychiatry, University of Toronto and Centre for Addiction and Mental Health Clarke Site, and Department of Psychiatry, University of Toronto and Mount Sinai Hospital, Toronto, Ont., Canada R. Michael Bagby, PhD Centre for Addiction and Mental

Health Clarke Site 25 College Street Toronto, Ont. M5T 1R$ %Canada) Tel. +1 (1) 535 $5 1, e+t. ),3,, -a+ +1 (1) ,., )$21, /0Mail michael_bagby3camh.net ABC -a+ + (1 )1 3 ) 12 3( /0Mail karger3karger.ch 444.karger.com 5 2 1 S. 6arger A7, Basel 3331, 9 19 . 5 25(:1..5 9 Accessible online at; 444.karger.com9journals9pps Key Words Alexithymia W Depression W Temporal stability Abstract Background: Previous studies demonstrating an associa- tion between alexithymia and depression have led to the proposal that alexithymia may be a state-dependent phenomenon rather than a stable and enduring

person- ality trait. Several longitudinal studies have provided support for a trait view of alexithymia, but most of these studies evaluated absolute stability only (i.e., the extent to which alexithymia scores change over time) and did not examine the relative stability of alexithymia (i.e., the extent to which relative differences among individuals remain the same over time) in the context of changes in illness symptomatology. The present study evaluated both absolute stability and relative stability of alexithy- mia in depressed patients who experienced a marked reduction in the severity of

depressive symptoms. Meth- ods: Forty-six psychiatric outpatients with major depres- sion were assessed for alexithymia and depression with the 20-item Toronto Alexithymia Scale and the Hamilton Rating Scale for Depression at the start of treatment (baseline) and after 14 weeks of treatment (follow-up) with antidepressant medication. Paired t tests and corre- lational analyses were performed to evaluate absolute stability and relative stability in alexithymia. Hierarchical regression analyses were then used to assess the degree to which the relative stability in alexithymia scores was related

to the severity of depressive symptoms, and the degree to which changes in alexithymia scores could be attributed to changes in depression scores. Results: Alexithymia scores changed significantly from baseline to follow-up, indicating a general lack of absolute stabili- ty. There was, however, strong evidence of relative sta- bility, as alexithymia scores at baseline correlated signifi- cantly with alexithymia scores at follow-up and were also a significant predictor of follow-up alexithymia scores, after partialling the effects of depression severity. Conclusions: Although alexithymia scores

may change in the presence of large changes in the severity of depressive symptoms, the finding of relative stability of alexithymia supports the view that this construct is a sta- ble personality trait rather a state-dependent phenome- non. Copyright  2001 S. Karger AG, Basel
Page 2
Stability of Ale+ithymia Psychother Psychosom 2 1?. ;25(2) 255 Introduction Over the past decade numerous studies have yielded empirical evidence that ale+ithymia is associated 4ith a variety of medical and psychiatric disorders, in particular 4ith substance use disorders, eating disorders,

panic dis0 order, essential hypertension, functional gastrointestinal disorders, and inflammatory bo4el disease A1, 2B. Cncer0 tainty remains, ho4ever, as to 4hether ale+ithymia is a predisposing or vulnerability factor that influences the onset or course of these disorders or merely a state reac0 tion to the presence of a disorder or any accompanying distress A35B. Arguments for ale+ithymia being a state0dependent phenomenon are based, in part, on evidence that the 4idely used Toronto Ale+ithymia Scale %TAS) A)B and its revised 2 0item version %TAS02 ) A.B correlate positively 4ith measures

of an+iety, depression, somatiDation, and neuroticism %negative affectivity) in studies using cross0 sectional designs A1, $12B. Other research, ho4ever, has demonstrated that the ale+ithymia construct is distinct and separate from the constructs of depression A13B, somatiDation A1(B, and neuroticism A1B. Studies using longitudinal designs have also yielded conflicting findings as to 4hether ale+ithymia is a stable and enduring personality trait or a phenomenon that arises secondary to depression or other psychological dis0 tress. Salminen et al. A3B, for e+ample, evaluated the sta0 bility of

ale+ithymia scores over a 10year period in 5( psy0 chiatric outpatients diagnosed primarily 4ith an+iety and mood disorders. Si+ty0five percent of the patients re0 ceived psychiatric treatment %mostly pharmacotherapy and supportive psychotherapy) during the year, 4hile the remaining 35E received no treatment. An abbreviated version of the Brief Symptom Fnventory %BSF)) A15B indi0 cated a significant decrease in psychological distress at the end of this 10year period, 4hereas ale+ithymia, measured by the TAS, did not change significantly. More recently, SaarijGrvi et al. A1)B follo4ed 1(, psy0

chiatric outpatients 4ith DSM0FFF0R diagnoses of major depression for a period of 1 year. Fn this study, ,(E of the patients received treatment 4ith psychotherapy and9or pharmacotherapy. The patients completed the BSF and the TAS02 prior to treatment and after 1 year and 4ere assessed for depression 4ith the Beck Depression Fnven0 tory %BDF) A1.B and the Hamilton Rating Scale for De0 pression %HRSD) A1$B. Hhereas scores on the depression and psychological distress measures had decreased signifi0 cantly at the follo40up testing, TAS02 scores sho4ed no significant change at the end of 1 year.

Contrasting findings 4ere obtained in a similar study by Honkalampi et al. A1,B, 4ho follo4ed 1), psychiatric outpatients 4ith DSM0FFF0R diagnoses of major depres0 sion or other depressive disorders for a period of ) months. The patients 4ere treated 4ith antidepressant medication and some also received benDodiaDepine drugs and9or psychotherapy. Fnstead of analyDing TAS02 scores as a continuous variable, ho4ever, these re0 searchers categoriDed their patients into ale+ithymic and nonale+ithymic groups based on a TAS02 cutoff score of )1 for ale+ithymia. Hhereas 3,E of their depressed patients

4ere ale+ithymic at baseline, only 23E 4ere still ale+ithymic at follo40up. Regression analyses performed on the entire sample revealed that BDF scores e+plained 2 E of the variation in TAS02 scores at baseline and (2E at follo40up. Honkalampi et al. A1,B concluded from these analyses that ale+ithymia is a secondary phenome0 non rather than a stable personality trait among depressed patients. This conclusion may be premature, ho4ever, since the association bet4een BDF and TAS02 change scores 4as not assessed. Although t4o other longitudinal studies have yielded support for the stability of

ale+ithymia, the sample siDes 4ere small, and the populations studied 4ere patients 4ith psychoactive substance dependence A2 B and under0 graduate university students A21B rather than patients 4ith depressive disorders. -urthermore, the follo40up intervals in these t4o studies 4ere rather short %bet4een ( and , 4eeks). There are t4o important limitations to the above longi0 tudinal studies. -irst, these studies evaluated the absolute stability of ale+ithymia only and did not distinguish it from the concept of relative stability. Absolute stability refers to the e+tent to 4hich personality

scores change over time, 4hereas relative stability indicates the e+tent to 4hich the relative differences among individuals re0 main the same over time A22B. Second, most studies e+am0 ining absolute change or stability on measures of ale+ithy0 mia did not e+amine 4hether any of the small %i.e., non0 significant) changes that did occur 4ere related to changes in psychological distress. Only Pinard et al. A2 B e+amined and reported correlations bet4een ale+ithymia change scores and psychological distress change scores? these cor0 relations 4ere significant, suggesting that ale+ithymia is

associated 4ith state0dependent distress. The concept of absolute stability in personality is derived primarily from clinical studies 4ith the use of change in mean scores over the course of a clinical trial. Fn these studies, statistically significant change in personali0 ty test scores over the course of treatment is regarded as
Page 3
25) Psychother Psychosom 2 1?. ;25(2) Iuminet9Bagby9Taylor indication of no stability. Fn contrast, personality re0 searchers generally emphasiDe the degree of relative sta0 bility in personality scores assessed over longer periods of time? this

type of stability is estimated through the use of measures of covariation, such as correlation coefficients. Fn these studies, statistically significant test0retest correla0 tions indicate that the personality trait is stable. As an e+ample, some clinical researchers argue that the person0 ality trait of neuroticism is not stable as scores on mea0 sures of neuroticism and depressive symptoms typically decrease significantly follo4ing recovery from a depres0 sive episode A23B. Fn contrast, on the basis of large test0 retest correlations, most personality researchers conclude that neuroticism

demonstrates considerable stability over time A2(B. To evaluate 4hether neuroticism or any other personality characteristic represents a potential vulnera0 bility factor for a medical or psychiatric illness, ho4ever, it is necessary to demonstrate the relative stability of the characteristic in the conte+t of acute change. This 4as done in a recent study by Santor et al. A22B, 4ho assessed test0retest correlations bet4een self0report ratings for neu0 roticism and e+traversion completed by depressed pa0 tients 4hen they 4ere acutely depressed and 5 4eeks later, 4hen recovered. Although the

patientsJ neuroticism scores decreased %as did their depression scores) and e+traversion scores increased significantly over 5 4eeks of treatment, indicating a lack of absolute stability, the test0 retest correlations for neuroticism and e+traversion scores indicated a high degree of relative stability. Regression analyses also indicated that changes in neuroticism and e+traversion scores over the course of treatment 4ere modestly or not at all accounted for by changes in depres0 sion scores. Ft is interesting to note that unlike neuroticism and e+traversion, 4hich are considered enduring

personality traits despite their lack of absolute stability %but because of evidence of relative stability), ale+ithymia is regarded by some researchers A25B as merely a state reaction even though data from several studies indicate that changes in mean ale+ithymia scores are nonsignificant in the conte+t of acute changes in levels of psychological distress. More0 over, there are also preliminary data that support the rela0 tive stability of ale+ithymia. Porcelli et al. A2)B evaluated the relative stability of ale+ithymia in 1 ( patients 4ith inflammatory bo4el disease %FBD) over a )0month

period. The patients 4ere treated for FBD but not for the psycho0 logical distress usually associated 4ith the disease. Si+ months after the initial assessment, three groups of pa0 tients 4ere distinguished based on changes in the level of disease activity %unchanged, improved, and 4orsened). -or each group, ale+ithymia, measured by the TAS02 , and depression and an+iety, measured by the Hospital An+iety and Depression Scale A2.B, 4ere assessed at base0 line and at the )0month follo40up. The relative stability of ale+ithymia 4as demonstrated by very high correlations bet4een TAS02 scores

measured at baseline and follo40 up for the total sample %r K .,5, p ! . 1) as 4ell as 4ithin each of the groups %unchanged, r K .,5, improved, r K .,), 4orsened, r K .,5, all p ! . 1). Moreover, 4hile significant changes 4ere observed 4ithin the groups for both an+iety and depression %i.e., small changes for the unchanged group, decreases for the improved group, and increases for the 4orsened group), the degree of ale+ithy0 mia 4as not influenced by the state effects of the level of disease activity. The demonstration of relative stability of ale+ithymia in the conte+t of change in disease

activity supports the vie4 that ale+ithymia could serve as a possi0 ble vulnerability factor for FBD. -urther support for the relative stability of ale+ithymia comes from studies demonstrating test0retest reliability of self0report measures of the construct. Csing data collected from various nonclinical populations at t4o separate times, investigators have reported Pearson correlation coefficients ranging from ... after 3 4eeks to .$3 after 3 months for the TAS02 A., 2$B, and from ..5 after 5 4eeks to .)$ after . months for the TAS A), 2,B. Despite the accumulation of data that supports vie40

ing ale+ithymia as a stable personality trait, methodologi0 cal and statistical limitations compromise the interpreta0 bility of the findings from some of these studies. -or e+ample, in some studies the magnitude of change in dis0 tress or illness 4as of relatively small siDe or not assessed A1), 21, 2)B, or treatment effects 4ere not specified A3B. Fn addition, almost all previous studies evaluated either the absolute stability or the relative stability of ale+ithymia rather than both types of stability. To our kno4ledge, only one study has e+amined both absolute and relative stabili0 ty of

ale+ithymia. This 4as a study in 4hich (1 patients 4ith bulimia nervosa 4ere administered the TAS and the HRSD, and assessed for bulimic symptomatology, before and after 1 4eeks of treatment 4ith an antidepressant medication A3 B. Although there 4as a significant im0 provement in bulimic symptomatology at the end of 1 4eeks, there 4ere no significant changes in mean ale+ithy0 mia and depression scores. The correlation bet4een TAS scores at baseline and at follo40up 4as significant %r K .5$, p ! . 1). Although these results support the abso0 lute and relative stability of ale+ithymia, as

Schmidt et al. A3 B point out, it is possible that the continuing high ale+i0 thymia scores 4ere due to persistent symptoms as the
Page 4
Stability of Ale+ithymia Psychother Psychosom 2 1?. ;25(2) 25. majority of patients had not fully recovered by the end of the 1 04eek treatment program. Fn the present study 4e evaluated both the absolute sta0 bility and the relative stability of ale+ithymia in a group of psychiatric outpatients 4ith major depression 4ho 4ere being treated 4ith antidepressant medication over a 1(0 4eek period. Iarge changes in depressive symptoms 4ere anticipated

as previous investigators treating depressed patients 4ith the same treatment protocol have reported major reductions in HRSD scores over even a shorter period %5 4eeks) A31B. He also e+amined the more specific issue concerning the degree to 4hich the relative stability of ale+ithymia may be accounted for by relative stability in depression severity. Fn addition, 4e e+amined ho4 changes in ale+ithymia are related to changes in depres0 sion severity. Method Subjects The subjects 4ere dra4n from a clinical database of outpatients 4ho 4ere referred for treatment at a Depression Clinic located in

a university0based tertiary care facility. All patients involved in the clinical database are treated 4ith a standardiDed treatment protocol that involves pretreatment administration of the Structured Clinical Fntervie4 for DSM0FL A+is F Disorders, Patient /dition %SCFD0P) A32B, the HRSD, and a variety of self0report measures. To be eligible for entry into the clinical database, all patients must meet DSM0FL diagnostic criteria for a primary, nonpsychotic, major depression as determined by the SCFD0P, score 1) or greater on the HRSD, not have a concurrent active medical illness %as determined

by the refer0 ring physician), and be free of antidepressant medications for at least 2 4eeks %( 4eeks for fluo+etine). Patients receive one of several anti0 depressant medications, using standardiDed treatment protocols in accordance 4ith prespecified dose algorithms? they are routinely reassessed 4ith the HRSD after 1( 4eeks of treatment. The various and possible medications and dose ranges are; %a) bupropion 1 3 mg9day, %b) citalopram 2 ) mg9day, %c) clomipramine .53 mg9day, %d) desipramine 5 3 mg9day, %e) fluo+etine 2 ) mg9 day, %f) imipramine .53 mg9day, %g) moclobemide 3 ) mg9 day?

%h) nefaDodone 15 ) mg9day, %i) paro+etine 1 ) mg9day, %j) phenelDine 35, mg9day, %k) sertraline 5 2 mg9day, %l) tranyl0 cypromine 2 ) mg9day and %m) venlafa+ine .53 mg9day. Pa0 tients receive one or t4o trials of antidepressant medication at ade0 quate doses selected at the discretion of the treating psychiatrist. Antidepressant medications are changed if the patient e+periences side effects or fails to respond to the first trial. Fn the present study, no patients 4ere treated 4ith concurrent psychotherapy. All patients in this study 4ere treated 4ith either moclobemide, paro+etine,

ser0 traline or venlafa+ine. -or the purpose of this study, patients entering the clinical data0 base 4ere also administered the TAS02 . After () patients %1, men, 2. 4omen) had completed the TAS02 at baseline and at the 1(04eek follo40up, the TAS02 4as removed from the battery of measures routinely collected. The mean age of this group of () patients 4as 3,. . years %SD K 1 .,2) and the mean level of education 4as 1).11 years %SD K 2.2.). The mean number of previous major depressive episodes 4as 2.2, %SD K 1.1,), and the mean age at onset of the first major depressive episode 4as 25.(3 years

%SD K 12.3,). The inter0 vie4ers 4ho administered the HRSD at follo40up 4ere NmaskedJ 4ith respect to patientsJ baseline and follo40up TAS02 scores, but 4ere a4are of baseline HRSD scores. Measures Ale+ithymia 4as assessed 4ith the TAS02 , 4hich is a revised and improved version of the original TAS. Ft has demonstrated reliability and validity, and is currently the most 4idely used measure of the ale+ithymia construct A1, ., 33B. The severity of depressive symptoms 4as assessed 4ith the 1.0item version of the HRSD. The HRSD is a clinician0related, semistructured intervie4 and is the most

4idely used and accepted measure of depression in clinical studies 4ith depressed patient samples A1$B. Ft is employed in most clinical trials of depression to detect change in depression severity A3(B. Statistical Analyses The follo4ing specific issues regarding the stability of ale+ithy0 mia during an acute treatment phase for depression 4ere addressed; %1) 4hether relative stability in ale+ithymia can be observed in a con0 te+t in 4hich absolute change in ale+ithymia scores over a short peri0 od of time %1( 4eeks) is likely? %2) the degree to 4hich the relative stability in ale+ithymia

scores is related to severity of depressive symptoms? %3) the degree to 4hich changes in ale+ithymia scores can be attributed directly to changes in depression scores. The first issue 4as addressed using paired t tests to assess absolute change in TAS02 and HRSD scores after 1( 4eeks of treatment 4ith antidepressant medication. This analysis 4as supplemented by cal0 culating effect siDes %CohenJs d) to determine the magnitude of change. Pearson correlations and intraclass correlations 4ere used to assess the relative stability of TAS02 and HRSD scores across time %i.e., test0retest). The

association bet4een TAS02 and HRSD scores at baseline and follo40up 4as also assessed using Pearson correla0 tions. The second issue 4as e+amined using hierarchical regression analyses, as 4as done in the study by Santor et al. A22B. 7iven that measures of ale+ithymia have been sho4n to have moderate to strong correlations 4ith measures of depression in cross0sectional designs %i.e., 4ithin but not across time) A$12B, the relative stability among individualsJ differences in ale+ithymia from baseline to follo40up may be attributable to individual differences in depression severity at either

baseline or follo40up. To this end, 4e built a regression model in 4hich follo40up ale+ithymia scores served as the criterion variable. The predictor variables, in order of entry 4ere; %1) a NblockJ containing baseline and follo40up depression scores, and %2) baseline ale+ithymia scores. Sho4ing that variance in follo40up ale+ithymia scores can be predicted by baseline ale+ithymia scores beyond the effects due to baseline and follo40up depression scores demonstrates that the relative stability of ale+ithymia scores cannot be accounted for in terms of depression severity. The third issue 4as

also e+amined using hierarchical regression analysis. Hhile absolute change in both depression and ale+ithymia scores might be adequately detected using paired t tests, it is unclear from this type of analysis to 4hat e+tent change in ale+ithymia scores can be accounted for by changes in depression severity. To assess for this potential effect, 4e built a regression model in 4hich change scores in ale+ithymia %i.e., baseline TAS02 scores minus follo40up
Page 5
25$ Psychother Psychosom 2 1?. ;25(2) Iuminet9Bagby9Taylor TAS02 scores) served as the criterion variable, and change scores

in depression %i.e., baseline HRSD scores minus follo40up HRSD scores) served as the predictor variable. All analyses 4ere performed using the Hindo4s version of the Statistical Package for Social Sciences %SPSS) 1 A35B. Results Gender Differences 7ender differences 4ere e+amined for ale+ithymia, severity of depression and duration of current depressive episode, total number of previous depressive episodes, age at baseline, marital status, years of education, and socioeconomic status. Oone of these comparisons 4ere found to be significant %p ! . 5)? therefore, all the results for the men and

4omen 4ere considered together. Absolute Change Scores and Relative Stability Oe+t, 4e assessed absolute change scores for depres0 sion severity and ale+ithymia. The mean HRSD score at baseline 4as 2 .$. %SD K 3..)) and at follo40up 4as ,.5) %SD K $.3 ), reflecting a substantial decrease in depres0 sion severity from baseline to follo40up? mean change score K 11.31, t%(5) K 1 .5(, p ! . 1. There 4as also a significant decrease in TAS02 scores. The mean TAS02 score at baseline 4as 51.,1 %SD K 1).5 ) and at follo40up 4as ().$ %SD K 1).5 )? mean change score K 5.11, t%(5) K 2.)1, p ! . 1.

/ffect siDes %CohenJs d) 4ere calculated for the TAS02 and HRSD mean change scores. -or the TAS02 , the change in mean scores from baseline to fol0 lo40up produced an effect siDe of ..$? for the HRSD, the effect siDe 4as more than ( times greater %CohenJs d K 3.1(). To assess evidence of relative stability, 4e computed correlations bet4een TAS02 scores at baseline and fol0 lo40up and HRSD scores at baseline and at follo40up. Positive and significant correlations 4ere found bet4een baseline and follo40up TAS02 scores %r K .)(, p . 1? intraclass correlation r K .)3, p ! . 1) and bet4een baseline

and follo40up HRSD scores %r K .5 , p ! . 1? intraclass correlation r K .3$, p ! . 5). Ho40 ever, 4hile the TAS02 4as not correlated significantly 4ith the HRSD at baseline %r K . ,, p K OS, e+plaining appro+imately 1E of the variance), a positive correlation 4as observed at follo40up %r K .3$, p ! . 1, e+plaining appro+imately 1(E of the variance), suggesting that any relative stability for TAS02 scores could be attributed to shared variance 4ith the relative stability in HRSD scores. Estimating the Relative Stability of Alexithymia Controlling for Absolute Change As indicated above, 4e used

hierarchical regression analyses to e+amine the e+tent to 4hich stability in ale+i0 thymia scores may be accounted for by individual differ0 ences in depression severity. The criterion variable 4as follo40up TAS02 scores, and the predictor variables 4ere HRSD scores at baseline and follo40up and TAS02 scores at baseline. A statistically significant amount of variance in follo40 up TAS02 scores 4as predicted from the block including HRSD scores at baseline and follo40up AR 2 K .15, -%2, (2) K 3.5$, p ! . 5, K .3,B. TAS02 scores at follo40up, ho4ever, 4ere predicted by TAS02 scores at baseline

beyond the variance e+plained by depression severity AR chg K .33, - chg %1, (1) K 2).(1, p ! . 1, K .5,B. These results provide strong evidence for relative stability in ale+ithymia. Modelling Change in Alexithymia and De ression -or this analysis change scores for the TAS02 from baseline to follo40up served as the criterion variable and change scores in HRSD from baseline to follo40up served as the predictor variable. Only 3E of the variance in change scores for the TAS02 4as accounted for by the variance in change scores for the HRSD, 4hich 4as statis0 tically nonsignificant A-%1, (3) K 2.5

, p 1 .12B. These results suggest that changes in TAS02 scores cannot be attributed to changes in HRSD scores. Discussion The present study evaluated both the absolute stability and the relative stability of ale+ithymia in the conte+t of significant changes in the severity of depressive symp0 toms. The results indicate that, although significant changes in absolute levels of depression severity and ale+i0 thymia 4ere observed, ale+ithymia still evidenced a high degree of relative stability. This 4as sho4n first by a posi0 tive and highly significant correlation bet4een ale+ithy0 mia scores at

treatment initiation %baseline) and after 1( 4eeks of aggressive treatment 4ith antidepressant medi0 cation. Stronger evidence for relative stability 4as pro0 vided by the hierarchical regression analyses, indicating that variance in follo40up ale+ithymia scores can be pre0 dicted from ale+ithymia scores at baseline over the effects contributed by baseline and follo40up depression severity %i.e., HRSD scores).
Page 6
Stability of Ale+ithymia Psychother Psychosom 2 1?. ;25(2) 25, Hhile several previous studies sho4ed evidence of absolute stability in ale+ithymia A3, 1), 2 , 21B, the

present results indicated absolute changes in ale+ithymia, 4hich could suggest that ale+ithymia is a state reaction. The magnitude of absolute change in ale+ithymia scores 4as small, ho4ever, in comparison 4ith the magnitude of absolute change in depression severity. Fn contrast, in the earlier study of patients 4ith major depression 4ho sho4ed absolute stability in ale+ithymia at the end of a 10year period A1)B, the magnitude of change in depression severity 4as small, perhaps because the majority of pa0 tients 4ere treated less aggressively 4ith only lo40dosage antidepressant medication

and9or supportive discussions 4ith a mental health professional once or t4ice a month AP.6. Salminen, personal commun., August 3, 2 B. Fn the present study, absolute change in ale+ithymia 4as not une+pected as a large decrease in depression severity 4as anticipated in response to intensive pharmacotherapy. Although our findings need to be replicated 4ith larger samples of depressed patients, it seems that absolute sta0 bility of ale+ithymia is unlikely in the presence of large changes in depression severity. Fn previous studies 4ith depressed patients, university students, and eating disor0

der patients, changes in psychological distress 4ere rela0 tively small, and the degree of ale+ithymia did not change significantly A1), 21, 3 B. The present study, by sho4ing clear and substantial changes in symptomatology, pro0 vides a more rigorous test of the stability of the ale+ithy0 mia construct than do these previous studies. /ven in the conte+t of the large change in depressive symptoms, the relative stability of ale+ithymia 4as demonstrated, indi0 cating its status as a stable personality trait and not a state0dependent variable. Hhereas Honkalampi et al. A1,B concluded that ale+i0

thymia does not demonstrate stability in depressed pa0 tients, they failed to assess relative stability and did not e+amine statistically 4hether changes in TAS02 scores 4ere correlated significantly 4ith changes in depression scores. One reason for their failure to assess relative stabil0 ity 4as their decision to create discrete ale+ithymic and nonale+ithymic groups based on a TAS02 cutoff score. As 4ith other personality constructs, ale+ithymia is best conceptualiDed as a dimensional rather than categorical variable, especially 4hen e+amining its association 4ith symptom severity. Moreover,

evaluating 4hether ale+i0 thymia is stable is best done by assessing if an individualJs score on a measure of this construct relative to othersJ scores remains the same over a prescribed interval, even in the conte+t of acute symptom change. Fn our opinion, this is preferable to assessing 4hether an individualJs score moves from one discrete category to another in rela0 tion to symptom change. One une+pected finding in the present study 4as the absence of a significant correlation bet4een the measures of ale+ithymia and depression at baseline, although the t4o measures did correlate

significantly at follo40up. This finding can probably be e+plained by the truncated distri0 bution of scores for the HRSD at baseline, 4hich ranged from 1) to 2,. This distribution 4as the result of the inclusion criterion that the patients 4ith major depression needed to score greater than 1) to be diagnosed as clinical0 ly depressed. At follo40up, no constraint 4as imposed on HRSD scores, and the resultant distribution 4as much broader than it 4as at baseline. The narro4er range of scores at baseline compared to the range of scores at fol0 lo40up most likely constrained the upper limit of

the cor0 relation coefficient at baseline. Although the present study sho4s strong evidence for relative stability of the ale+ithymia construct, this does not necessarily mean that ale+ithymic characteristics can0 not be modified by specific therapeutic interventions. /+perienced clinicians recommend individual and9or group psychotherapies that focus on educating patients about their emotions and on developing imaginative ca0 pacities as 4ell as an ability to identify and describe sub0 jective feelings A1, 3), 3.B. Fndeed, in a recent preliminary study of postmyocardial infarction patients,

4ho received once 4eekly group psychotherapy for ( months, there 4as a significant decrease in the mean TAS score, 4hich there0 after remained stable over a 20year follo40up period %r K ..5) A3$B. -urther studies are needed, ho4ever, to verify the e+tent to 4hich psychotherapeutic interventions can enable individuals to reduce or adapt to deficits in the cognitive processing of emotions. Acknowledgements Dr. Olivier Iuminet is postdoctoral researcher at the Belgian Oational -und for Scientific Research and Oorth Atlantic Treaty OrganiDation %OATO) research fello4ship. His contribution to this

study 4as supported by grant -R-C 2.(5().,. and grant -ORS 1.5.12(. from the Belgian Oational -und for Scientific Research, and made 4hile he 4as an in0resident Post0Doctoral -ello4 at the Centre for Addiction and Mental Health, Toronto, Canada.
Page 7
2) Psychother Psychosom 2 1?. ;25(2) Iuminet9Bagby9Taylor References 1 Taylor 7P, Bagby RM, Parker PDA; Disorders of Affect Regulation; Ale+ithymia in Medical and Psychiatric Fllness. Cambridge, Cambridge Cniversity Press, 1,,.. 2 Taylor 7P; Recent developments in ale+ithy0 mia theory and research. Can P Psychiatry 2 ?(5;13(1(2. 3

Salminen P6, SaarijGrvi S, GrelG /, Tammin0 en T; Ale+ithymia state or traitR One0year fol0 lo40up study of general hospital psychiatric consultation out0patients. P Psychosom Res 1,,(?3$;)$1)$5. ( Hise TO, Mann IS, Mitchell PD, Hryvniak M, Hill B; Secondary ale+ithymia; An empirical validation. Compr Psychiatry 1,, ?31;2$( 2$$. 5 Haviland M7, Sha4 D7, Cummings MA, MacMurray PP; Ale+ithymia; Subscales and re0 lationship to depression. Psychother Psycho0 som 1,$$?5 ;1)(1. . ) Taylor 7P, Ryan D, Bagby RM; To4ard the development of a ne4 self0report ale+ithymia scale. Psychother

Psychosom 1,$5?((;1,1 1,,. . Bagby RM, Parker PDA, Taylor 7P; The T4en0 ty0Ftem Toronto Ale+ithymia Scale. F. Ftem se0 lection and cross0validation of the factor struc0 ture. P Psychosom Res 1,,(?3$;2332. $ Bagby RM, Taylor 7P, Ryan D; Toronto Ale+i0 thymia Scale; Relationship 4ith personality and psychopathology measures. Psychother Psychosom 1,$)?(5;2 .215. , Deary FP, Scott S, Hilson PA; Oeuroticism, ale+ithymia and medically une+plained symp0 toms. Pers Fndivid Diff 1,,.?22;5515)(. 1 Honkalampi 6, Hintikka P, Tanskanen A, Ieh0 tonen P, LiinamGki H; Depression is strongly associated 4ith

ale+ithymia in the general pop0 ulation. P Psychosom Res 2 ?($;,,1 (. 11 Honkalamp 6, Saarinen P, Hintikka P, Lirtan0 en L, LiinamGki H; -actors associated 4ith ale+ithymia in patients suffering from depres0 sion. Psychother Psychosom 1,,,?)$;2. 2.5. 12 Hendry+ MS, Haviland M7, Sha4 D7; Di0 mensions of ale+ithymia and their relation0 ships to an+iety and depression. P Pers Assess 1,,1?5);22.23.. 13 Parker PDA, Bagby RM, Taylor 7P; Ale+ithy0 mia and depression; Distinct or overlapping constructsR Compr Psychiatry 1,,1?32;3$. 3,(. 1( Bach M, Bach D, de S4aan M; Fndependency of ale+ithymia and

somatiDation. Psychoso0 matics 1,,)?3.;(51(5$. 15 Derogatis IR, Melisaratos O; The brief symp0 tom inventory; An introductory report. Psychol Med 1,$3?13;5,5) 5. 1) SaarijGrvi S, Salminen P6, Tikka P, Toikka T; Ale+ithymia and depression; A one0year fol0 lo40up study in outpatients 4ith major depres0 sion %abstract). P Psychosom Res 2 ?($;23$. 1. Beck AT, Steer RA; Beck Depression Fnventory Manual. San Antonio, Psychological Corpora0 tion, 1,$.. 1$ Hamilton M; Development of a rating scale for primary depressive illness. Br P Soc Clin Psy0 chol 1,).?);2.$2,). 1, Honkalampi 6, Hintikka P,

Saarinen P, Ieh0 tonen P, LiinamGki H; Fs ale+ithymia a perma0 nent feature in depressed patientsR Results from a )0month follo40up study. Psychother Psychosom 2 ?),;3 33 $. 2 Pinard I, Oegrete PC, Annable I, Audet O; Ale+ithymia in substance abusers; Persistence and correlates of variance. Am P Addic 1,,)?5; 323,. 21 MartT UneD0Sa UncheD -, Ato07arcia M, Co U rcoles Adam /, Huedo Medina TB, Selva /spaVa PP; Stability in ale+ithymia levels; A longitudinal analysis on various emotional ans4ers. Pers Fndivid Diff 1,,$?2(;.)...2. 22 Santor DA, Bagby RM, Poffe RT; /valuating stability and

change in personality and depres0 sion. P Pers Soc Psychol 1,,.?.3;135(13)2. 23 Hirschfeld MA, 6lerman 7I, Clayton PP, 6ell0 er MB, McDonald0Scott P, Iarkin BH; Assess0 ing personality; /ffects of the depressive state on trait measurement. Am P Psychiatry 1,$3? 1( ;),5),,. 2( Costa PT Pr, McCrae RR, Sonderman AB, Bar0 bano H/, Iebo4itD B, Iarson DM, Iebo4itD B; Cross0sectional studies of personality in a na0 tional sample. 2. Stability in neuroticism, e+0 traversion and openness. Psychol Aging 1,$)? 1;1((1(,. 25 Ahrens S, Deffner 7; /mpirical study of ale+i0 thymia; Methodology and results.

Am P Psy0 chother 1,$)?( ;(3 ((.. 2) Porcelli P, Ieoci C, 7uerra L, Taylor 7P, Bag0 by RM; A longitudinal study of ale+ithymia and psychological distress in inflammatory bo4el disease. P Psychosom Res 1,,)?(1;5), 5.3. 2. Sigmond AS, Snaith RP; The Hospital An+iety and Depression Scale. Acta Psychiatr Scand 1,$3?).;3)13. . 2$ Pandey R, Mandal M6, Taylor 7P, Parker PDA; Cross0cultural ale+ithymia; Development and validation of a Hindi translation of the 2 0 item Toronto Ale+ithymia Scale. P Clin Psy0 chol 1,,)?52;1.31.). 2, Iinden H, Hen -, Paulhus DI; Measuring ale+ithymia; Reliability,

validity, and preva0 lence. Adv Pers Assess 1,,5?1 ;51,5. 3 Schmidt C, Pi4any A, Treasure P; A controlled study of ale+ithymia in eating disorders. Compr Psychiatry 1,,3?3(;5(5$. 31 Bagby RM, Poffe RT, Parker PDA, 6alemba L, Harkness 6I; Major depression and the five0 factor model of personality. P Personal Disord 1,,5?,;22(23(. 32 -irst MB, SpitDer RI, 7ibbon M, Hilliams PBH; Structured Clinical Fntervie4 for DSM0 FL A+is F Disorders, Patient /dition %SCFD0P), version 2. Oe4 Work, Oe4 Work State Psy0 chiatric Fnstitute, Biometrics Research, 1,,5. 33 Bagby RM, Taylor 7P, Parker PDA; The

T4en0 ty0Ftem Toronto Ale+ithymia Scale. FF. Conver0 gent, discriminant, and concurrent validity. P Psychosom Res 1,,(?3$;33( . 3( Snaith RP; Present use of the Hamilton De0 pression Rating Scale; Observations on method of assessment in research of depressive disor0 ders. Br P Psychiatry 1,,)?1)$;5,(5,.. 35 Oorusis; Statistical Package for the Social Sciences %SPSS). Basic and Professional Statis0 tics 1 . Chicago, SPSS, 1,,,. 3) 6rystal H; Ale+ithymia and psychotherapy. Am P Psychother 1,.,?33;1.31. 3. S4iller HF; Ale+ithymia; Treatment utiliDing combined individual and group psycho0

therapy. Fnt P 7roup Psychother 1,$$?3$;(. )1. 3$ Beresnevaite M; /+ploring the benefits of group psychotherapy in reducing ale+ithymia in coronary heart disease patients; A prelimina0 ry study. Psychother Psychosom 2 ?),;11. 122.