the relationships between complicated grief, depression, and alexithymia according to the...

6
Research report The relationships between complicated grief, depression, and alexithymia according to the seriousness of complicated grief in the Japanese general population Minako Deno a, , Mitsunori Miyashita b , Daisuke Fujisawa c , Satomi Nakajima d , Masaya Ito d, e a Correspondence Division, Musashino University, Japan b Department of Palliative Nursing, Health Sciences, Tohoku University Graduate School of Medicine, Japan c Psycho-oncology Division, National Cancer Center East, Japan d National Institute of Mental Health, National Center of Neurology and Psychiatry, Japan e Japan Society for the Promotion of Science, Japan article info abstract Article history: Received 31 May 2011 Accepted 28 June 2011 Available online 29 July 2011 Background: The present research investigated whether the relationship between alexithymia and complicated grief was different from the relationship between alexithymia and general depressive symptom according to the seriousness of complicated grief in the Japanese general population. Methods: In the Japanese general population sample, 948 participants between 40 and 79 years old (effective response rate, 48.0%) completed a cross-sectional anonymous questionnaire about alexithymia, depression, and complicated grief. To compare the high risk (n = 243) and low risk (n = 705) of complicated grief groups, simultaneous analysis of two groups with standard maximum likelihood estimation was performed and six hypothesized models were veried. Results: The model (RMSEA = 0.047, AIC = 71.520) that showed that the path coefcients of the latent variable of alexithymia to the observed variables were equal and that the path coefcient of alexithymia to psychological distress was equal was adopted. The contribution ratios from alexithymia to complicated grief were apparently smaller (24%) than those to depression (3738%). Conclusions: Our findings showed that alexithymia scarcely contributed to complicated grief compared to depression and that the contribution ratio in the high risk group was lower than that in the low risk group. The contribution of the latent variable of psychological distress to complicated grief and depression was lower in the high risk group than in the low risk group. The lack of a correlation between alexithymia and complicated grief might indicate that there are different mechanisms underlying the symptoms of alexithymia and complicated grief. © 2011 Elsevier B.V. All rights reserved. Keywords: Complicated grief Alexithymia Depression Bereavement Emotion 1. Introduction Maciejewski et al. (2007) empirically found that natural psychological responses to grief, such as disbelief, yearning, anger, and depression, peek within 6 months following the death of signicant others. Most people who experience bereavement overcome their natural emotional response to bereavement. However, integration of the loss does not occur and acute grief becomes more intensive and continuous in the form of complicated grief (Shear and Shair, 2005). Individuals with complicated grief experience a constella- tion of symptoms that often include preoccupation with the lost person, anger about the death, and avoidance of reminders Journal of Affective Disorders 135 (2011) 122127 Corresponding author at: Correspondence Division, Musashino Univer- sity, 3-40-10 Sekimae, Musashino-shi, Tokyo 180-0014, Japan. Tel./fax: + 81 422 52 7218. E-mail address: [email protected] (M. Deno). 0165-0327/$ see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2011.06.037 Contents lists available at ScienceDirect Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Upload: minako-deno

Post on 05-Sep-2016

213 views

Category:

Documents


1 download

TRANSCRIPT

Journal of Affective Disorders 135 (2011) 122–127

Contents lists available at ScienceDirect

Journal of Affective Disorders

j ourna l homepage: www.e lsev ie r.com/ locate / j ad

Research report

The relationships between complicated grief, depression, and alexithymiaaccording to the seriousness of complicated grief in the Japanesegeneral population

Minako Deno a,⁎, Mitsunori Miyashita b, Daisuke Fujisawa c, Satomi Nakajima d, Masaya Ito d,e

a Correspondence Division, Musashino University, Japanb Department of Palliative Nursing, Health Sciences, Tohoku University Graduate School of Medicine, Japanc Psycho-oncology Division, National Cancer Center East, Japand National Institute of Mental Health, National Center of Neurology and Psychiatry, Japane Japan Society for the Promotion of Science, Japan

a r t i c l e i n f o

⁎ Corresponding author at: Correspondence Divisiosity, 3-40-10 Sekimae, Musashino-shi, Tokyo 180-001422 52 7218.

E-mail address: [email protected] (M. D

0165-0327/$ – see front matter © 2011 Elsevier B.V.doi:10.1016/j.jad.2011.06.037

a b s t r a c t

Article history:Received 31 May 2011Accepted 28 June 2011Available online 29 July 2011

Background: The present research investigated whether the relationship between alexithymiaand complicated grief was different from the relationship between alexithymia and generaldepressive symptom according to the seriousness of complicated grief in the Japanese generalpopulation.Methods: In the Japanese general population sample, 948 participants between 40 and 79 yearsold (effective response rate, 48.0%) completed a cross-sectional anonymous questionnaireabout alexithymia, depression, and complicated grief. To compare the high risk (n=243) andlow risk (n=705) of complicated grief groups, simultaneous analysis of two groups withstandard maximum likelihood estimation was performed and six hypothesized models wereverified.Results: Themodel (RMSEA=0.047, AIC=71.520) that showed that the path coefficients of thelatent variable of alexithymia to the observed variables were equal and that the path coefficientof alexithymia to psychological distress was equal was adopted. The contribution ratios fromalexithymia to complicated grief were apparently smaller (2–4%) than those to depression(37–38%).Conclusions: Our findings showed that alexithymia scarcely contributed to complicated griefcompared to depression and that the contribution ratio in the high risk group was lower thanthat in the low risk group. The contribution of the latent variable of psychological distress tocomplicated grief and depression was lower in the high risk group than in the low risk group.The lack of a correlation between alexithymia and complicated grief might indicate that thereare different mechanisms underlying the symptoms of alexithymia and complicated grief.

© 2011 Elsevier B.V. All rights reserved.

Keywords:Complicated griefAlexithymiaDepressionBereavementEmotion

1. Introduction

Maciejewski et al. (2007) empirically found that naturalpsychological responses to grief, such as disbelief, yearning,

n, Musashino Univer-4, Japan. Tel./fax: +81

eno).

All rights reserved.

anger, and depression, peek within 6 months following thedeath of significant others. Most people who experiencebereavement overcome their natural emotional response tobereavement. However, integration of the loss does not occurand acute grief becomesmore intensive and continuous in theform of complicated grief (Shear and Shair, 2005).

Individuals with complicated grief experience a constella-tion of symptoms thatoften includepreoccupationwith the lostperson, anger about the death, and avoidance of reminders

Alexithymia

DifficultyIdentifying

Feelings

DifficultyDescribingFeelings

ExternallyOrientedThinking

ComplicatedGrief

GeneralDepressiveSymptoms

ig. 1. Hypothesis model for the relationship between alexithymia andmplicated grief.

123M. Deno et al. / Journal of Affective Disorders 135 (2011) 122–127

of the loss (Ogrodniczuk et al., 2005). Although it is widelyrecognized that symptoms of normal grief are similar to thoseof depression, there are distinctions between depression andcomplicated grief in phenomenology, psychophysiology, andresponses to treatment (Lichtenthal et al., 2004). Previousresearch found that complicated grief does not respondwell toproven efficacious treatments for depression (Reynolds et al.,1999; Zygmont et al., 1998). The prevalence rates of compli-cated grief reported by epidemiological studies on the non-clinical population were 4.2% (Middleton et al., 1996), 24.6%(Chiu et al., 2009), and 2.4% (Fujisawa et al., 2010).

To improve the treatment of complicated grief, it is neces-sary to investigate the characteristics that effectively buffercomplicated grief. Alexithymia, a patient characteristic thatinfluences psychological distress in depression, has been foundto influence a patient's response to psychotherapy (McCallumet al., 2003; Taylor, 2000). Ogrodniczuk, et al. (2005) found thatthe negative influence of alexithymia on general symptoms, forexample, anxiety, depression, and interpersonal distress, wasbuffered by psychotherapy. Taylor et al. (1997) found thefollowing three core features of alexithymia: (1) difficulty inidentifying feelings, (2) difficulty in communicating feelings,and (3) externally oriented thinking. This three-factor charac-terization of alexithymia has become the standard for describingthe construct (Ogrodniczuk et al., 2005). Of note, alexithymia isnot a psychiatric disorder, but rather a characterization ofthinking, feeling, and relating processes among patients with awide range of psychiatric diagnoses.

Therehasbeena sizeable amountof theoretical andempiricalwork on the relationships between alexithymia and generaldepressive symptoms (Lipsanen et al., 2004; Ogrodniczuk et al.,2005). Statistically, Lipsanen et al. (2004) demonstrated thatdepression and alexithymia are highly correlated but distinct.Parker et al. (1991) have summarized the possible causes ofoverlap between alexithymia and depression as follows. First,the manifestation of alexithymic features might be a transitoryreaction (secondary alexithymia) evoked by stressful situationsand the accompanying depression and anxiety. Second, second-ary alexithymia is a defensive response to the acute depressionthat typically accompanies stressful situations. Finally, it is aresponse to overall changes in the quality of life, and notdepression per se, that is associated with the manifestation ofsecondary alexithymia. Furthermore, in some patients, second-ary alexithymia may become permanent and indistinguishablefrom primary (i.e., trait) alexithymia (Parker et al., 1991).

However, few researches have been conducted on therelationship between complicated grief and alexithymia, andnone have been conducted in Japan. For example, Ogrodniczuket al. (2005) found that alexithymia (except for externallyoriented thinking) did not correlate with complicated grief, forexample, intrusion, pathological grief, and avoidance. Aquestion related to the difference between complicated griefand general depression symptom is as follows: how differentis the relationship between the extent of complicated griefand alexithymia from the relationship between depressivesymptoms and alexithymia according to the seriousness ofcomplicated grief? The investigation of complicated grief'srelationship with alexithymia, which is related to feeling andexpressing emotion, will be helpful in understanding andtreating complicated grief. In addition, as there are bothdistinctions and similarities between complicated grief and

depressive symptom, it may be necessary to simultaneouslyand separately explore how extent complicated grief relates togeneral depressive symptoms.

The aim of the current study was to investigate whetherthe relationship between alexithymia and complicated griefdiffers from the relationship between alexithymia andgeneral depressive symptom according to the seriousness ofcomplicated grief. The hypothesis model of the relationshipsbetween alexithymia, depression symptom, and complicatedgrief is shown in Fig. 1.

2. Methods

2.1. Procedures

A cross-sectional anonymous questionnaire was adminis-tered to a sample of the general Japanese population. Fourtarget areas (Tokyo, Miyagi, Shizuoka and Hiroshima pre-fectures) were identified to obtain a wide geographic distribu-tion for the nationwide sample. The four areas included anurbanprefecture (Tokyo) andmixedurban–rural areas (Miyagi,Shizuoka and Hiroshima).

Initially, 5000 subjects aged 40–79 yearswere identified by astratified two-stage randomsamplingmethodof residents of thefour areas. Fifty census tracts were randomly selected for eacharea and then 25 individuals were selected within each censustract, thus identifying 1250 individuals for each area. Question-naires were mailed to potential participants in June 2009, andreminder postcards were sent 2 weeks later. The ethical andscientific validity of this studywas confirmedby the institutionalreview boards of graduate medicine in Tokyo University.

2.2. Measures

Alexithymia was assessed by the Japanese version of theTronto Alexithymia Scale-20 (TAS-20) (Bagby et al., 1994;Komaki et al., 2003). This scale is a self-report questionnairethat is constructed by the following three factors: “difficultyidentifying feelings”, “difficulty describing feelings”, and “ex-ternally oriented thinking.” Twenty items are asked, using afive-point Likert scale (1: absolutely disagree, 2: somewhatdisagree, 3: neutral, 4: somewhat agree, and 5: absolutelyagree). Content and concurrent validity and reliability of thisscale were confirmed.

Complicated grief was assessed by the Brief GriefQuestionnaire (BGQ) (Shear et al., 2006). The BGQ is a five-item, self-report questionnaire that inquires about difficultyaccepting death, interference of grief in their life, difficulty of

Fco

124 M. Deno et al. / Journal of Affective Disorders 135 (2011) 122–127

images or thoughts of death, avoidance of things related tothe deceased, and feeling cut off or distant from other people.The answers are rated as 0, not at all; 1, somewhat; or 2, a lot.A past report suggests that a total score on the BGQ of 8 orhigher is indicative of complicated grief; between 5 and 7 ofprobable complicated grief; and less than 5 of no complicatedgrief (Shear et al., 2006). A diagnosis of complicated briefshould not be given within six month after bereavement(Prigerson et al., 2009); therefore, those who had experi-enced bereavement within the past six months wereexcluded. In addition, those who had experienced bereave-ment with their children were also excluded because griefover children's death has been consistently reported to beprolonged, and the diagnostic reliability of complicated griefamong this population has been questioned (Dyregrov et al.,2003; Stroebe et al., 2007).

General depressive symptomwas assessed by K6 (Furukawaet al., 2003; Kessler et al., 2002). K6 is a six-item self-assessmentscale that measures the extent of decreasing self-esteem andfeelings of depression and hopeless. The answers are rated on afive-point Likert scale. Content and concurrent validity andreliability of this scale were confirmed.

Demographic data were obtained including age, gender,time of recent bereavement, relationship with the deceased,cause and place of death of the deceased, and days spent withthe deceased during the last week of the end-of-life period,whether the bereavement was expected or not, and whetherthe participant had been a caregiver of the deceased or not.

2.3. Participants

Of the 5000 questionnaires that were distributed, 44 wereundeliverable and 1975 were returned (Response rate: 39.9%).Of these, 117 were excluded due to missing data. Of theremaining questionnaires, 792 were excluded because therespondents had not experienced bereavementwithin the pastten years, 114 because of bereavement within the past sixmonths, and 4 because of bereavement with their children.Thus, 948 responses were analyzed (effective response rate,48.0%).

2.4. Statistical analysis

The presence of complicated grief was defined using theabovementioned cutoff score, according to previous research(Shear et al., 2006). Using PASW version 18 (Poler Engineeringand Consulting, 2009), ANOVA was conducted to investigatedifferences in complicated grief by demographic variables andfactors related to bereavement.

To investigate whether the relationship between alexithy-mia and complicated grief differs from the relationshipbetween alexithymia and depressive symptom according tothe seriousness of complicated grief, a simultaneous analysis oftwo groups was performed with AMOS version 18 (Arbuckle,1983–2009), with standard maximum likelihood estimation.Six hypothesized models were compared. The models were asfollows: model 0: all parameters are different between twogroups; model 1: the path coefficients of the latent variable ofalexithymia to observed variables are equal; model 2: the pathcoefficients of the latent variable of psychological distress toobserved variables are equal; model 3: in addition to model 1,

the path coefficients of the latent variable of psychologicaldistress to observed variables are equal; model 4: in addition tomodel 1, the path coefficient of alexithymia to psychologicaldistress is equal; and model 5: in addition to model 3, the pathcoefficient of alexithymia to psychological distress is equal.

Indices of fit of the model to the data were evaluated withseveral statistics: goodness of fit index (GFI) of .95 or greater,adjusted goodness of fit index (AGFI) of .90 or greater, thecomparative fit index (CFI) of .95 or greater, the root meansquare error of approximation (RMSEA) less than or equal to.05, a chi-squared thatwasnot significant (pN .05), andAkaike'sInformation Criterion (AIC).

3. Results

According to previous research, the participants weredivided to three groups: high risk of complicated grief(n=25), probable complicated grief (n=218), and low riskof complicated grief (n=705). It might be appropriate to com-pare three groups; however, the high risk group of complicatedgrief with the group of probable complicated grief werecombined and regarded as the high risk group (a score of 5–10) because the number of participants in the high risk group ofcomplicated grief was insufficient for conducting simultaneousanalysis. In the present results, we compared two groups: thehigh risk of complicated grief (n=243) and low risk ofcomplicated grief (n=705).

The results of the ANOVA that investigated differences bydemographic variables and factors related to bereavement areshown in Table 1. Females experienced higher complicatedgrief than males (t (918)=−2.675, pb .01). Of the factorsrelated with bereavement, the relationship with the deceased(F (4, 943)=22.300, pb .001), primary caregiver (t (929)=6.229, pb .001), cause of death (F (3, 941)=13.072, pb .001),place of death (F (4, 943)=3.653, pb .01), expected death(t (866)=−3.369, pb .001), and days spent with the deceasedduring the end-of-life period (F (3, 864)=13.001, pb .001)were significant. Age and time since bereavement were notstatistically significant. As for the relationship with thedeceased, spouses of the deceased experienced the highestcomplicated grief compared to the other groups. Parents andsiblings of the deceased experienced higher complicated griefthan parents-in-law and others. There were no significantdifferences betweenparents and siblings or betweenparent-in-laws and others. Primary caregivers scored higher on compli-cated grief than thosewhowere not primary caregivers. Cancerand cardiac disease causedmore serious complicated grief thanother causes of death. As for the place of death, those who losttheir significant person in a care facility experienced highercomplicated grief than others. In addition, the group ofunexpected death experienced a higher complicated griefscore than expected death. Among the groups of days spentwith the deceased, the groups of everyday and 4–6 days perweek scored higher on complicated grief than the groups of 1–3 days per week and none.

Pearson's correlation coefficients between variables areshown in Table 2. The results showed that the correlationcoefficients between complicated grief and general depressivesymptoms were significant (the high risk group; r=0.211, thelow risk group; r=0.190, the same order shall apply herein-after). To avoidmodeling the subscales of complicated grief and

Table 1The relationships of demographic variables/factors related to bereavementand complicated grief (n=948).

n Mean SD F or t

GenderMale 387 2.76 2.05 t (918)=−2.675⁎⁎

Female 533 3.14 2.19

Age40–49 215 2.90 2.24 F (3, 926)=0.67650–59 340 2.98 2.0760–69 363 3.04 2.1870–79 12 3.74 2.39

Relationship with the deceasedSpouse 62 4.85 1.76 F (4, 943)=22.300⁎⁎⁎

Parent(s) 458 3.11 2.19Parent(s)-in-law 245 2.41 1.86Sibling(s) 88 3.65 2.15Others 95 2.32 2.14

Primary caregiverYes 457 3.46 2.18 t (929)=6.229⁎⁎⁎

No 487 2.59 2.07

Time since bereavement6–12 months 110 3.04 2.11 F (9, 938)=0.8641–2 years 140 3.26 2.312–3 years 135 2.83 2.133–4 years 117 2.80 2.174–5 years 87 2.98 2.205–6 years 84 2.89 2.136–7 years 96 3.12 2.107–8 years 60 2.88 2.318–9 years 48 2.90 2.049–10 years 71 3.45 2.08

Cause of deathCancer 350 3.42 2.20 F (3, 941)=13.072⁎⁎⁎

Stroke 94 2.99 2.22Cardiac disease 110 3.46 2.08Others 391 2.52 2.04

Place of deathHome 173 2.88 2.13 F (4, 943)=3.653⁎⁎

General hospital 645 3.06 2.15Hospice/PCU 32 3.41 2.34Care facility 62 2.27 2.09Others 36 3.84 2.31

Expected deathExpected 564 2.84 2.13 t (866)=−3.369⁎⁎⁎

Unexpected 304 3.36 2.20

Days spent with the deceased during the end-of-life periodEveryday 215 3.62 2.25 F (3, 864)=13.001⁎⁎⁎

4–6 days/week 93 3.61 2.121–3 days/week 221 2.80 2.00None 339 2.62 2.13

** pb .01, *** pb .001.

125M. Deno et al. / Journal of Affective Disorders 135 (2011) 122–127

general depressive symptoms as factors of alexithymia, anunderlying latent variable labeled “psychological distress”wasassumed. Also, we assumed that “alexithymia” was theunderlying construct between difficulty identifying feelings,difficulty describing feelings, and externally oriented thinkingbecause of the high correlation coefficients between them(r=0.340–0.618 and r=0.415–0.650, respectively).

The fit indices of these models are shown in Table 3. Asshown in Table 3,model 4 (with RMSEA=0.047, AIC=71.520)

yielded values smaller than the values of the other models, andthus was the correct result. Model 4 showed that the pathcoefficients of the latent variable of alexithymia to observedvariables were equal and that the path coefficient of alexithy-mia to psychological distress was equal. The standardized pathcoefficients are shown in Fig. 2. The standardized indirect effectcoefficients (contribution ratio) from alexithymia to compli-cated griefwere 0.200 (4%) in the low risk groupof complicatedgrief and 0.137 (2%) in the high risk of complicated grief group.The standardized indirect effect coefficients (contributionratio) from alexithymia to general depressive symptom were0.608 (37%) in the low riskof complicatedgrief groupand0.616(38%) in the high risk of complicated grief group. Overall, thefinal model accounted for 66% of the variance in the latentvariable of psychological distress.

4. Discussion

The results of the current study have advanced our under-standing of the relationship between alexithymia, complicat-ed grief, and general depressive symptom in the Japanesegeneral population. The findings showed that the relationshipbetween alexithymia and complicated grief differed accord-ing to the seriousness of complicated grief. In other words,alexithymia scarcely contributed to complicated grief, espe-cially in the high risk of complicated grief group, and thecontribution ratio in the high risk groupwas lower than in thelow risk group. In addition, the relationship of complicatedgrief and general depressive symptom with alexithymiadiffered in both groups, as previous research has shown(Ogrodniczuk et al., 2005). The contribution of the latentvariable of psychological distress to complicated grief andgeneral depressive symptoms was lower in the high riskgroup than in the low risk group.

As for demographic variables and factors related to bereave-ment, thepresent results supportedprevious research. First, thecomplicated grief score was significantly related to gender,which is consistentwith previous research (Boelen et al., 2006).However, because some research did not find a genderdifference (Boelen and van den Bout, 2003), further investiga-tion is necessary. Second, the relationship with the deceasedwas an important factor of complicated grief, which supportedthe previous research that indicated the important role playedby close kinship in complicated grief (Gana and K'Delant, 2011;Johnson et al., 2007). The present result suggested that spousesof the deceased experienced the highest complicated grief,whereas, there was no significant difference between parents-in-law and others. Therefore, the impact of the death of sig-nificant otherswas likely influenced by the degree of kinship ordirect blood relationships. Finally, as for primary caregivers anddays spent with the deceased, these variables might be relatedto each other. All of these variables were quantity (not quality)aspects of the relationship with the deceased. Thus, primarycaregivers might be those who were close relatives and mighthave spent more time with the deceased.

Interestingly, alexithymia, which had been found to influ-ence patients' response to psychotherapy (Ogrodniczuk et al.,2005), showeddifferent relations to complicated grief accordingto the seriousness of complicated grief. It suggests that, in thehigh risk group of complicated grief, alexithymia may notdirectly influence the improvement of complicated grief. In

Table 2Correlations between demographic data, alexithymia, depression, and complicated grief for the seriousness of complicated grief.

1 2 3 4 5 6 7 8 9 Mean SD

1 Age 1.000 −.054 −.014 −.049 −.012 .024 .046 −.069 −.0292 Sex −.073 1.000 −.060 .175⁎ −.123 −.049 −.125 −.103 .1063 Primary caregiver −.119⁎⁎ −.191⁎⁎⁎ 1.000 .185⁎⁎ .098 .000 −.019 .149⁎ .0024 Expected death .007 .055 .041 1.000 −.044 .051 −.014 .036 .150⁎

Alexithymia5 Difficulty identifying feelings .040 −.003 .059 −.009 1.000 .650⁎⁎⁎ .415⁎⁎⁎ .582⁎⁎⁎ .132⁎ 16.24 5.636 Difficulty describing feelings .005 −.063 .056 −.048 .618⁎⁎⁎ 1.000 .473⁎⁎⁎ .447⁎⁎⁎ .143⁎ 14.82 2.877 Externally oriented thinking −.023 −.062 .039 .006 .340⁎⁎⁎ .360⁎⁎⁎ 1.000 .301⁎⁎⁎ −.062 24.43 3.19

Psychological distress8 Depression −.034 .040 .033 −.035 .529⁎⁎⁎ .416⁎⁎⁎ .200⁎⁎⁎ 1.000 .211⁎⁎⁎ 10.89 4.219 Complicated grief .021 .053 −.118⁎⁎ .122⁎⁎ .187⁎⁎⁎ .121⁎⁎ .043 .190⁎⁎⁎ 1.000 3.01 2.17

Above = the high risk group of complicated grief (n=243) and below = the low risk group of complicated grief (n=705).⁎ pb .05.

⁎⁎ pb .01.⁎⁎⁎ pb .001.

126 M. Deno et al. / Journal of Affective Disorders 135 (2011) 122–127

addition, thosewhohave ahigher extent of alexithymiamaynotsuffer from complicated grief when they experience bereave-ment with significant others, and those who suffer fromcomplicated grief may not experience clinical increases inalexithymia. Because alexithymia had been found to influencepatient's response to psychotherapy (McCallum et al., 2003;Taylor, 2000), this result was attention-provoking.

The difference between these mechanisms may be due tothe lack of a correlation between complicated grief and alexi-thymia in the high risk group. Lipsanen et al.(2004) found thatthere is a very low distress level that accompanies the alexi-thymia factor based on the minimal correlation betweenalexithymia and GHQ (General Health Questionnaire) distressscores. The items of complicated grief described the dailydistress following the experience of bereavement. This lack of arelationship with alexithymia may reflect the extent ofemotional distress and that there are different mechanismsunderlying the symptomsof alexithymia and complicated grief.Therefore, the extent of alexithymia may not significantlyinfluence complicated grief. In regards to patients' character-istics that relate to complicated grief, previous research has

Table 3The fit indices and AIC of each model.

χ2 df p RMR GFI AGFI CFI RMSEA AIC

Model 0 31.031 8 .000 .328 .987 .952 .978 .055 75.031Model 1 32.766 10 .000 .388 .987 .960 .978 .049 72.766Model 2 33.790 9 .000 .365 .986 .953 .976 .054 75.790Model 3 35.409 11 .000 .420 .985 .960 .976 .048 73.409Model 4 33.520 11 .000 .467 .986 .963 .978 .047 71.520Model 5 35.866 12 .000 .487 .985 .963 .977 .046 71.866

Model 0: all parameters are different between two groups.Model 1: the path coefficients of the latent variable of alexithymia toobserved variables are equal.Model 2: the path coefficients of the latent variable of psychological distressto observed variables are equal.Model 3: in addition to Model 1, the path coefficients of the latent variable opsychological distress to observed variables are equal.Model 4: in addition to Model 1, the path coefficient of alexithymia topsychological distress is equal.Model 5: in addition to Model 3, the path coefficient of alexithymia topsychological distress is equal.

Alexithymia

DifficultyIdentifying

Feelings

DifficultyDescribingFeelings

ExternallyOrientedThinking

.72 .76.86 .87 .43 .46

PsychologicalDistressR2= .66

ComplicatedGrief

GeneralDepressiveSymptoms

.22/.29 .79/.86

.75/.75

Fig. 2. The standardized path coefficients in model 4. Left coefficients: thelow risk group of complicated grief (n=705), Right coefficients: the highrisk group of complicated grief (n=243).

f

found a significant influence of attachment style, especiallyinsecure attachment. Shear and Shair (2005) discussedcomplicated grief in relation to separation anxiety, attachmentstyle and grief process. These other factors of attachment stylemay contribute to research that aims to understand and im-prove complicated grief. Further research is necessary con-cerning the factors that relate to or influence complicated grief.

The results that alexithymia was significantly related togeneral depressive symptom and the differing relationships ofalexithymia with complicated grief and general depressivesymptoms were supported by previous research (Lipsanenet al., 2004; Ogrodniczuk et al., 2005). The difference betweenthe relationships suggests that improvements in alexithymiawill influence improvements in general depressive symptom,but not complicated grief. However, Zygmont et al. (1998)reported that improvements in complicated grief were corre-lated with improvements in depressive symptoms. Followingthe suggestion of Zygmont et al. (1998), it is possible thatimprovements in alexithymia are indirectly related to improve-ments in complicated grief through improvements in generaldepressive symptoms. The observed correlation betweendepressive symptoms and complicated grief in the presentresults might indicate a similarity between the constructs.Therefore, although there are distinctions between theseconstructs in phenomenology, psychophysiology, and re-sponses to treatment (Lichtenthal et al., 2004), the correlationbetween them and the relationship between alexithymia anddepressive symptoms might be considered in the treatment of

127M. Deno et al. / Journal of Affective Disorders 135 (2011) 122–127

complicated grief. The current researchwas cross-sectional and,therefore, causality could not be determined and furtherresearch is necessary.

Limitations of the present study are as follows. First, thepresent study used a cross-sectional design. Therefore, alter-native explanations are possible. In daily life, individuals whohave a strong tendency to experience psychological distressmay experience more alexithymic symptoms, resulting ingreater psychological distress. These psychological aspectsmay exacerbate each other in a downward-spiral fashion. Wecannot determine a causal relationship among these psycho-logical variables. It is necessary to conduct further researchwitha longitudinal design to clarify causality. The second limitationis that the results cannot be generalized to a clinical populationbecause the number of participants in the high risk ofcomplicated grief group was insufficient and the data solelydepended on the participants' self-report.

The current study has several limitations that must beconsidered; however, the study may have implications in thetreatment of complicated grief. Further clinical and researchattention will help researchers and clinicians better under-stand patients' complicated grief and, therefore, will helpdesign and evaluate interventions for clinical patients withcomplicated grief.

Role of funding sourceThis study was fully supported by the Grant-in-Aid for Cancer Research

endowed to Mitsunori Miyashita from the Ministry of Health, Labor andWelfare, Japan (MHLW); the MHLW had no further role in study design; inthe collection, analysis and interpretation of data; in the writing of thereport; and in the decision to submit the paper for publication. The authorsexpress gratitude to Koken Ozaki, Ph. D., for advising on the analysis.

Conflicts of interestAll authors declare that they have no conflicts of interest.

References

Arbuckle, J.L., 1983–2009. Amos 18.0.0.Bagby, R.M., Parker, J.D., Taylor, G.J., 1994. The twenty-item Tronto

Alexithymia Scale-I. Item selection and cross-validation of the factorstructure. Journal of Psychosomatic Research 38, 23–32.

Boelen, P.A., van den Bout, J., 2003. Gender differences in traumatic griefsymptom severity after the loss of a spouse. Omega: Journal of Death andDying 46, 183–198.

Boelen, P.A., van den Bout, J., van den Hout, M.A., 2006. Negative cognitionsand avoidance in emotional problems after bereavement: a prospectivestudy. Behaviour Research and Therapy 44, 1657–1672.

Chiu, Y.W., Huang, C.T., Yin, S.M., Huang, Y.C., Chien, C.h., Chuang, H.Y., 2009.Determinants of complicated grief in caregivers who cared for terminalcancer patients. Supportive Care in Cancer 18, 1321–1327.

Dyregrov, K., Nordanger, D., Dyregrov, A., 2003. Predictors of psychosocialdistress after suicide, SIDS and accidents. Death Study 27, 143–165.

Fujisawa, D., Miyashita, M., Nakajima, S., Ito, M., Kato, M., Kim, Y., 2010.Prevalence and determinants of complicated grief in general population.Journal of Affective Disorders 127, 352–358.

Furukawa, T.A., Kessler, R.C., Slade, T., Andrews, G., 2003. The performance ofthe K6 and K10 screening scales for psychological distress in the

Australian National Survey of Mental Health and Well-Being. Psycho-logical Medicine 33, 357–362.

Gana, K., K'Delant, P., 2011. The effects of temperament, character, anddefense mechanisms on grief severity among the elderly. Journal ofAffective Disorders 128, 128–134.

Johnson, J.G., Zhang, B., Greer, J.A., Prigerson, H.G., 2007. Parental control,partner dependency, and complicated grief among widowed adults inthe community. The Journal of Nervous and Mental Disease 195, 26–30.

Kessler, R.C., Andrews, G., Colpe, L.J., Hiripi, E., Mroczek, D.K., Normand, S.L.T.,Walters, E.E., Zaslavsky, A.M., 2002. Short screening scales to monitorpopulation prevalences and trends in non-specific psychological distress.Psychological Medicine 32, 959–976.

Komaki, G., Maeda, M., Arimura, T., Nakata, A., Shinoda, H., Ogata, I., Shimura,M., Kawamura, N., Kudo, C., 2003. The reliability and factorial validity ofthe Japanese version of the 20-item Tronto Alexithymia Scale (TAS-20).Japanese Journal of Psychosomatic Medicine 43, 839–846.

Lichtenthal, W.G., Cruess, D.G., Prigerson, H.G., 2004. A case for establishingcomplicated grief as a distinct mental disorder in DSM-V. ClinicalPsychology Review 24, 637–662.

Lipsanen, T., Saarijarvi, S., Lauerma, H., 2004. Exploring the relations betweendepression, somatization, dissociation and alexithymia—overlapping orindependent constructs? Psychopathology 37, 200–206.

Maciejewski, P.K., Zhang, B., Block, S.D., Prigerson, H.G., 2007. An empiricalexamination of the stage theory of grief. Journal of the American MedicalAssociation 297, 716–723.

McCallum, M., Piper, W.E., Ogrodniczuk, J.S., Joyce, A.S., 2003. Relationshipsamong psychological mindedness, alexithymia and outcome in fourforms of short-term psychotherapy. Psychology and Psychotherapy 76,133–144.

Middleton, W., Burnett, P., Raphael, B., Martinek, N., 1996. The bereavementresponse: a cluster analysis. The British Journal of Psychiatry 169, 167–171.

Ogrodniczuk, J.S., Piper,W.E., Joyce,A.S., 2005. Thenegative effectof alexithymiaon the outcomeof group therapy for complicatedgrief:what rolemight thetherapist play? Comprehensive Psychiatry 46, 206–213.

Parker, J.D.A., Bagby, R.M., Taylor, G.J., 1991. Alexithymia and depression:distinct or overlapping constructs? Comprehensive Psychiatry32, 387–394.

Poler Engineering and Consulting, 2009. PASW Statistics 18.Prigerson, H.G., Horowitz, M.J., Jacobs, S.C., Parkes, C.M., Aslan, M., Goodkin,

K., Raphael, B., Marwit, S.J., Wortman, C., Neimeyer, R.A., Bonanno, G.,Block, S.D., Kissane, D., Boelen, P., Maercker, A., Litz, B.T., Johnson, J.G.,First, M.B., Maciejewski, P.K., 2009. Prolonged grief disorder: psycho-metric validation of criteria proposed for DSM-V and ICD-11. PLoSMedicine 6, e1000121.

Reynolds, C.F.I., Miller, M.D., Pasternak, R.E., Frank, E., Perel, J.M., Cornes, C.,Houck, P.R., Mazumdar, S., Dew, M.A., Kupfer, D.J., 1999. Treatment ofbereavement-related major depressive episodes in later life: a controlledstudy of acute and continuation treatment with nortriptyline andinterpersonal psychotherapy. The American Journal of Psychiatry 156,202–208.

Shear, K., Shair, H., 2005. Attachment, loss, and complicated grief.Developmental Psychobiology 47, 253–267.

Shear,K.M., Jackson,C.T., Essock, S.M., Donahue, S.A., Felton, C.J., 2006. Screeningfor complicated grief among Project Liberty service recipients 18 monthsafter September 11, 2001. Psychiatric Services 57, 1291–1297.

Stroebe, M., Schut, H., Stroebe, W., 2007. Health outcomes of bereavement.Lancet 370, 1960–1973.

Taylor, G.J., 2000. Recent developments in alexithymia theory and research.Canadian Journal of Psychiatry 45, 134–142.

Taylor, G.J., Bagby, R.M., Parker, J.D.A., 1997. Disorders of Affect Regulation:Alexithymia in Medical and Psychiatric Illness. Cambridge UniversityPress, New York.

Zygmont, M., Prigerson, H.G., Houck, P.R., Miller, M.D., Shear, M.K., Jacobs, S.,Reynolds III, C.F., 1998. A post hoc comparison of paroxetine andnortriptyline for symptoms of traumatic grief. The Journal of ClinicalPsychiatry 59, 241–245.