Addictive Behaviors 37 (2012) 1224–1229

Contents lists available at SciVerse ScienceDirect

Addictive Behaviors

Avoidance coping strategies, alexithymia and alcohol abuse: A mediation analysis

Giovanna Coriale a,⁎, Elena Bilotta b, Luigi Leone b, Fernando Cosimi a, Raffaella Porrari a,Francesca De Rosa a, Mauro Ceccanti a

a Center for Alcohol Abuse (Centro Riferimento Alcologico Regione Lazio—CRARL), Department of Clinical Medicine, Sapienza University of Rome,Viale dell'Università 37‐00185 Rome, Italyb Department of Developmental and Social Psychology, Sapienza University of Rome, Via dei Marsi, 78‐00185 Rome, Italy

⁎ Corresponding author at: Via Demetriade 68, 03286236159; fax: +39 0649972093.

E-mail address: gcoriale@tin.it (G. Coriale).

0306-4603/$ – see front matter © 2012 Elsevier Ltd. Alldoi:10.1016/j.addbeh.2012.05.018

a b s t r a c t

a r t i c l e i n f o

Keywords:

AlexithymiaAvoidance coping strategiesAlcohol abuseStructural equation modelingMediation

Alexithymia and avoidance coping strategies are both associated with alcohol abuse, but their effects havebeen seldom studied simultaneously. The present study investigated the interplay between alexithymiaand avoidance coping strategies in predicting the severity of alcohol abuse in an alcohol-dependent sample.The TAS-20 and COPE-NVI questionnaires were administered to 110 alcoholic inpatients enrolled into a re-covery program at the Center for Alcohol Abuse of the Department of Clinical Medicine, Sapienza Universityof Rome, Italy. The alcohol abuse index consisted of the mean alcohol units consumed by participants anddays of abstinence before being enrolled into the recovery program. Results showed that alexithymic alco-holics consumed significantly more alcohol and were less abstinent than non-alexithymic alcoholics. Con-cerning the relationship among alexithymia, coping strategies and alcohol abuse, data showed thatalexithymia completely mediated the effects of avoidance coping strategies on alcohol abuse, suggestingthat avoidance strategies have therefore an indirect effect on alcohol abuse among alcoholics. Theoreticaland clinical implications of the results are discussed.

© 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Alexithymia refers to difficulties identifying and describing sub-jective feelings, distinguishing between feelings and the bodily sensa-tions of emotional arousal, fantasizing and using an internal orientedcognitive style, and a tendency to focus on real events and somaticsymptoms (Nemiah, 1977; Nemiah & Sifneos, 1970). Although initial-ly described in the context of psychosomatic illnesses, alexithymiccharacteristics may be observed in patients with a wide range ofboth medical conditions (Lauriola, Panno, Tomai, Ricciardi, &Potenza, 2011; Lumley, Neely, & Burger, 2007) and psychiatric disor-ders (Salminen, Saarijärvi, Äärelä, Toikka, & Kauhanen, 1999; Wise,Mann, & Hill, 1990). In the last three decades, the alexithymic concepthas gained much popularity in cognitive psychology, where it hasbeen redefined as a deficit in cognitive processing and regulation ofemotions (De Rick & Vanheule, 2007; Taylor, Bagby, & Parker,1997). Alexithymics appear to be less able to recognize their emo-tional signals and regulate emotions effectively (Lindsay & Ciarrochi,2009; Taylor et al., 1997). They are more likely to misinterpret theiremotions as being symptoms of illness (Taylor & Bagby, 2004) andcan undertake a wide range of maladaptive regulation strategies suchas eating and substance-related disorders (Taylor, 2000; Taylor et al.,1997). Several studies have found a relationship between alexithymia

0178 Rome, Italy. Tel.: +39

rights reserved.

and substance abuse (e.g., Cleland, Magura, Foote, Rosenblum, &Kosanke, 2005; De Rick & Vanheule, 2006; Lindsay & Ciarrochi, 2009).Studies have reported the prevalence rates for alexithymia among re-cently abstinent drug- and alcohol-dependent patients to be 41.7 to50.4% (Haviland, Hendryx, Shaw, & Henry, 1994; Haviland, Shaw,Cummings, & MacMurray, 1988; Taylor, Bagby, & Parker, 1990). Theserates are higher than the levels of alexithymia found in non-clinicalsamples (4 to 18%) and psychiatric comparison group samples (12.5to 33%; Taylor, 2000).

When studies on alcohol-dependent inpatients are specifically con-sidered, the range of alcohol-dependent individuals identified asalexithymic goes from 48 to 78% (Evren et al., 2008; Loas, Fremaux,Otmani, Lecercle, & Delahousse, 1997; Rybakowski, Ziólkowski,Zasadzka, & Brzezinski, 1988; Sauvage & Loas, 2006; Taylor et al., 1990;Uzun, Ates, Cansever, &Ozsahin, 2003). In general, the literature suggestsan association between alcohol-related disorders and alexithymia (DeRick & Vanheule, 2006; Kauhanen, Julkunen, & Salonen, 1992;Stasiewicz et al., 2012; Taylor et al., 1997), together with an adverse im-pact of alexithymia on the treatment of alcohol abuse (Loas et al., 1997;Ziólkowski, Gruss, & Rybakowski, 1995).1 On the onehand, some authorssuggest that alexithymia is a personality trait that places individuals at agreater risk of alcohol-related disorders (Uzun et al., 2003; Ziólkowski etal., 1995); on the other hand, others consider it a consequence of alcohol

1 Nevertheless, recent evidence suggests a much more limited relationship to drink-ing outcomes among treated alcoholics (Stasiewicz et al., 2012).

1225G. Coriale et al. / Addictive Behaviors 37 (2012) 1224–1229

use (De Rick & Vanheule, 2006; Haviland et al., 1994; Taylor et al., 1997).Although empirical support can be found for both views, the trait para-digm seems to describemore persuasively the state of affairs in the alco-holic population (De Rick & Vanheule, 2006; Taylor et al., 1997; Uzun etal., 2003).

The reasons explaining the prevalence of alexithymics in the alcohol-ic population are not clear yet. Some researchers have argued thatalexithymic individuals use alcohol as a coping mechanism for stress orto improve interpersonal functioning (e.g., verbal and emotional func-tions; Kauhanen et al., 1992; Rybakowski et al., 1988), since alexithymicpeople often feel uncomfortable in social situations (Uzun et al., 2003;Wise, Mann, & Shay, 1992). It has also been suggested that alexithymicsconsume alcohol in an effort to cope with negative emotional states(Stasiewicz et al., 2012; Thorberg et al., 2011). The rewarding experiencewith alcohol may eventually lead to the development of dependence(Rybakowski et al., 1988; Uzun et al., 2003).

Coping skills have found to be significant predictors of alcohol con-sumption (Bussey Rask et al., 2006). In particular, avoidance copingstrategies aremore closely linked to drinking behavior thanmore activestrategies both among social drinkers (McKee, Hinson, Wall, & Spriel,1998) and alcoholics (Hasking & Oei, 2007). Coping skills may also pre-dict the outcome of the treatment of alcoholics (Bussey Rask et al.,2006). A study in a sample of treated alcoholics found that use of activecoping strategies increased the odds of remaining abstinent comparedto use of avoidance coping strategies (Moser & Annis, 1996). Anotherstudy found that increasing coping skills during treatment significantlypredicted abstinence among alcohol abusers (Litt, Kadden, Cooney, &Kabela, 2003).

The association of alexithymia and alcohol consumption and abusehas been reasonably established in non-clinical samples (Bruce, Curren,& Williams, 2012; Kauhanen et al., 1992; Rybakowski et al., 1988).Among alcoholics, however, only a handful of studies have investigatedhow much alexithymia is connected to alcohol consumption in termsof actual alcohol units consumed (Stasiewicz et al., 2012). These re-searchers found that higher scores of alexithymia were associated withfewer percent days abstinence and greater alcohol dependence severity(Stasiewicz et al., 2012).

The present paper aimed first to investigate alcohol consumptionamong alcohol abusers who were also alexithymics, comparing them tonon-alexithymics. Getting to know the differences between them couldbe interesting for the clinical consequences of alcohol abuse treatment.We expected alexithymic participants to consume more alcohol and beless abstinent than non-alexithymic participants.Wewould also estimatehowmany alcohol units could be traced back to increases in the levels ofalexithymia.

Secondly, this paper aimed to analyze the relationship betweenalexithymia, avoidance coping skills and alcohol abuse. Research studieshave proposed that alexithymic individuals use alcohol to cope withstress or negative emotional states, or to improve interpersonal func-tioning (Evren et al., 2008; Kauhanen et al., 1992; Loas et al., 1997;Sauvage & Loas, 2006; Thorberg et al., 2011; Uzun et al., 2003). On theother hand, avoidance coping strategies seem to be closely linked todrinking behavior (Chung, Langenbucher, Labouvie, Pandina, & Moos,2001; Williams & Clark, 1998).

However, alexithymia and coping strategies may be to some ex-tent linked, which renders it difficult to interpret their separate link-ages with alcohol abuse. Given the relationship between alexithymiaand alcohol abuse (e.g., Evren et al., 2008; Uzun et al., 2003) and therelationship between avoidance coping strategies and alcohol abuse(e.g., Carver, Scheier, & Weintraub, 1989; Chung et al., 2001;Williams & Clark, 1998), we explored if alexithymia or avoidance cop-ing strategies serves as amediator of the other variable effect on alcoholabuse. Recently published research reported that coping strategies me-diated the effects of alexithymia on alcohol consumption (Bruce et al.,2012). However, this evidence was found in a social drinkers sampleand therefore the specificmediation pattern recovered could be specific

for non-addicted alcohol consumption. Instead, another recent paperrelying on an alcoholic sample did not find a significant correlation be-tween alexithymia and coping styles (Stasiewicz et al., 2012). Thus, itappears worthwhile to investigate the associations and different pat-terns of mediation involving alcohol consumption, alexithymia andavoidance coping in an alcoholic sample.

2. Material and methods

2.1. Participants

One-hundred and ten participants (76.4% men and 23.6% women)affected by current alcoholism were enrolled into the study (meanage 44.3; s.d. 9.7). Participants were recruited from an inpatient alco-hol program at the Center for Alcohol Abuse (Centro RiferimentoAlcologico Regione Lazio—CRARL), Department of Clinical Medicine,Sapienza University of Rome, Italy. Participants were eligible for thestudy if they met DSM-IV criteria (APA, 2000) for current alcohol de-pendence. The diagnoses were performed by the Center's psychiatrist.Individuals were excluded if they met criteria for a current drug usedisorder other than nicotine. At the onset of withdrawal, patients sys-tematically received a withdrawal substitution treatment to minimizewithdrawal symptoms (Lejoyeux, Solomon, & Ades, 1998). This med-ication was progressively decreased during detox.

2.2. Measures

Alexithymia was measured using the Toronto Alexithymia Scale inits validated Italian version (TAS-20; Bressi et al., 1996). The TAS-20consists of 20 items rated on a five-point Likert scale (range from 1“strongly disagree” to 5 “strongly agree”). The total scores of thescale range from 20 to 100, with a score≤51 being considered acut-off point for classifying individuals as non-alexithymic and≥61for classifying individuals as alexithymic (Bagby, Parker, & Taylor,1994; Taylor et al., 1997). The range between 52 and 60 classifies in-dividuals as borderline. The TAS-20 is composed of three factors. Thefirst two factors (Difficulties Identifying Feelings—DIF and DifficultiesDescribing Feelings—DDF) refer to emotional awareness and expres-sion and might be considered “affect-related” factors. The third factor(Externally-Oriented Thinking—EOT) refers to a specific tendency todeal with superficial themes and to avoid emotional thinking, andcan be considered a more “cognitive” factor (Franz et al., 2008). Weused here the total scale score (α=.79).

Coping strategies were assessed using the Coping Orientation toProblems Experienced in its validated Italian version (COPE-NVI; Sicaet al., 2008;α=.86). The COPE-NVI scale consists of 60 items indicatinghow often people undertake a specific coping process when facing dif-ficult or stressful situations (range from 1 “I usually don't do this atall” to 4 “I usually do this a lot”). The items refer to five large indepen-dent dimensions, composed of fifteen subscales assessing a broadrange of coping strategies: social support (composed of comprehensionand information seeking, and emotional expression); avoidance strate-gies (including denial, behavioral and mental disengagement, drug andalcohol abuse); positive attitude (composed of acceptance, positive re-interpretation and restraint); problem solving (such as suppression ofcompeting activities, planning and activities); turning to religion (com-posed of religion and absence of humour; Sica et al., 2008).

2.3. Procedure

From January to September 2010, the alcoholic patients underwenttwo semi-structured interviews to assess lifetime alcohol consumption.The Lifetime Drinking History (LDH; Skinner & Sheu, 1982) and TimeLine Follow Back (TLFB; Sobell & Sobell, 1992) were used to assess alco-hol consumption from the first year of regular drinking to the presentand specific amounts of alcohol use in the last month before enrolling

1226 G. Coriale et al. / Addictive Behaviors 37 (2012) 1224–1229

into the rehabilitation program, respectively. Patients had an average of17.2 (s.d. 10.9) years of at-risk drinking (up to three drinks a day formen and up to two a day for women). The average standard drinksdaily (1 standard drink=12 g of pure ethanol) was 14.7 (s.d. 10.2) ofdrinks drunk per day in the last month.

Participants were asked to complete the questionnaire aboutalexithymia and coping strategies only after theirwithdrawal symptomsdisappeared (according to a set of specified withdrawal symptoms, suchas elevated blood pressure, tachycardia, tremor and sweating, amongothers) and after medication had been discontinued.

2.4. Data analyses

Amultivariate analysis of variance was conducted to explore the im-pact of alexithymia levels on several alcohol consumption variables. A re-gression was computed to investigate how much the alcoholics drankcontingently on the severity of their alexithymia. Structural equationmodels were applied to test the relationships among alexithymia, avoid-ance coping strategies and alcohol abuse, specifying two alternative me-diation processes.

3. Results

Among 110 alcoholics, 34.5% were alexithymics, 33.6% border-line and 31.8% non-alexithymics. We intended to investigate the ef-fects of alexithymic types on four alcohol consumption indicators:mean of alcohol units consumed, days of abstinence before enroll-ing into the program, age of the first drink, and total years of alco-hol abuse. We ran a MANOVA on these dependent variables tomaintain low type I error rates. Results revealed a significant multi-variate effect (Wilk's Λ=.741, F (8, 208)=4.20, pb .001), whichenabled us to safely conduct univariate tests on each dependentvariable. Alcohol consumption varied significantly across thegroups (F (2,107)=8.1; pb .001). In particular, alexithymic partici-pants differed from non-alexithymic ones, but not from the borderlinegroup in the mean of alcohol units consumed per day, as shown inTable 1. Similar differences were found when investigating days of absti-nence before being enrolled into the recovery program (F (2,107)=11.5;pb .001). Alexithymic and borderline individuals abstained for shorter pe-riods than non-alexithymic alcoholics (Table 1). No difference was foundamong the three groups for age offirst drinking and years of alcohol abuse(Fsb1). We decided to investigate the amount of alcohol consumptionlinked with increasing levels of alexithymia. Thus, we ran a bivariate re-gression model where alcohol units consumed were regressed on thestandardized TAS-20 total score. The computed regression coefficient of2.84 (se=.94, pb .01) indicated that an increase in one standard devia-tion of the TAS-20 scorewas associatedwith about a 2.84 increase of alco-hol units consumed per day. The 90% confidence interval around thisestimate shows that in the Italian alcoholic population, variations inalexithymia are associated with an increase of alcohol consumptionfrom a minimum of 1.29 to a maximum of 4.39 of alcohol units per day.

As mentioned before, research relying on clinical samples tradition-ally analyzed separately the effects of alexithymia and avoidance copingstrategies, demonstrating that both are correlated with the frequencyand quantity of alcohol consumed (e.g., Chung et al., 2001; Evren et

Table 1Means and standard deviations for alcohol units consumed and days of abstinence beforeenrolling into the rehabilitation program (N=110).

Average of alcohol units per day Days of abstinence

Alexithymic 15.8 (10.5)a .7 (2.2)a

Borderline 18.4 (10.5)a 1.8 (3.0)a

Non‐alexithymic 7.2 (7.2)b 5.0 (5.9)b

Note: different letters indicate significant differences (LSD post-hoc test) pb .05.

al., 2008; Sauvage & Loas, 2006; Uzun et al., 2003). Nevertheless, theseassociationswere computedwithout taking into account the possible as-sociation between the twopredictors. Our data detects a significant over-lap between avoidance coping strategies and alexithymia (Table 2). Weobserved the expected positive correlation between alexithymia and al-cohol abuse, and between avoidance coping strategies and alcohol abuse.This pattern of associations suggests that the two potential predictors ofalcohol abuse (i.e., alexithymia and avoidance coping strategies) are notindependent andmakes the zero order associations reported in the liter-aturewith alcohol consumption (e.g., Kauhanen et al., 1992; Rybakowskiet al., 1988) likely to be spurious and potentially difficult to interpret.

Data obtained in a social drinkers sample modeled these associa-tions as a mediationmodel, reporting that coping fully mediated the ef-fects of alexithymia on alcohol consumption (Bruce et al., 2012). Weinvestigated in our alcoholics sample two alternativemediationmodelsto account for the web of associations among the three variables. First,we tested whether avoidance coping mediated the distal effects ofalexithymia on alcohol consumption (Bruce et al., 2012). Then, we com-pared thismodel to an alternative one specifying alexithymia as theme-diator channelling the effects of avoidance coping strategies on alcoholconsumption. Isolating themediationmechanism that explains best thedata would be useful for both its theoretical and clinical consequences.

Following recent recommendations (Iacobucci, Saldanha, & Deng,2007; Zhao, Lynch, & Chen, 2010), we opted for a structural equationmodeling (SEM) approach for testing mediation. SEM compares favor-ably with more usual regression based approaches (Baron & Kenny,1986) to mediation because it takes into account measurement error,computes directly the extent and significance of the indirect effect,and allows the comparison of the fit of alternativemodels and data con-figurations (Iacobucci et al., 2007), among other advantages. Inmodel 1(Fig. 1, panel A),we specified amodelwhere alexithymia is relatedwithavoidance coping, which in turn is associated with alcohol abuse. Thismodel specifies a full mediation (Baron & Kenny, 1986) or, adoptingthe terminology of Zhao et al. (2010, p. 204), “indirect only mediation”(indicating that no direct effect remains once indirect effects are consid-ered). The latent factors for avoidance coping strategies and alexithymiawere defined by three random item-parcels each (e.g., Bandalos &Finney, 2001), while a latent factor for alcohol consumption was de-fined by the two indicators referring to the total alcohol units drank inthe past 30 days and the number of days of abstinence before being en-rolled into the recovery program (reverse coded). To limit estimationproblems owing to normality violations, all models were estimatedwith Robust Maximum Likelihood Estimation. Therefore, models wereestimated relying on the covariance and the asymptotic Covariancema-trices (Jöreskog & Sörbom, 1996). The LISREL8 (Jöreskog & Sörbom,1996) software was used to estimate the model.

As reported in Fig. 1, factor loadings and regression weights were allsignificant, but the model fitted unsatisfactorily (χ2 (18)=40.56,pb .01; RMSEA=.11 (90%CI: .066–.16; NNFI=.90; CFI=.94). Althoughalexithymia was significantly related to the supposed mediator (avoid-ance coping) and the mediator was associated with alcohol consump-tion, the architecture of the model failed to adequately reproduce theobserved data. This suggests that inverting the role of mediator and dis-tal predictor could better represent the data. Model 2 (Fig. 1, panel B)therefore specified that the association of avoidance coping strategies

Table 2Correlations of alexithymia, avoidance coping strategies and alcohol abuse (N=110).

1 2 3

Alexithymia 1 .41⁎⁎ .49⁎⁎

Avoidance coping strategies – 1 .20⁎⁎

Alcohol abuse – – 1

Note: Items about drug and alcohol abuse were taken out from the “avoidance copingstrategies” scale.⁎⁎pb .001.

Fig. 1.Mediation models on avoidance coping strategies, alexithymia, and alcohol abuse. Note: standardized estimates; all parameter estimates are significant (pb .01). Items aboutdrug and alcohol abuse were taken out from the “avoidance coping strategies” scale.

2 The chi square difference test is scaled following the Satorra and Bentler procedure.It is appropriate when using the robust Satorra–Bentler estimation (Satorra & Bentler,2001).

1227G. Coriale et al. / Addictive Behaviors 37 (2012) 1224–1229

with alcohol abuse would be fully mediated by alexithymia. Differentlyput, avoidance coping strategies have a significant indirect effect on alco-hol abuse among alcoholics. The fullmediationmodel fitted satisfactorily(Hu & Bentler, 1999): χ2 (18)=21.45, p=.26; RMSEA=.044 (90% CI:.00–.10), RMSEA test of close fit (RMSEAb .05), p=.51; NNFI=.99;CFI=.99. Note how strikingly absolute (chi square and RMSEA) and rel-ativefit indices (NNFI and CFI) improved over those computed formodel1. Avoidance coping was related to alexithymia (.56, pb .01), which wasin turn related to alcohol consumption (.68, pb .01). Turning to media-tion, the crucial parameters to consider are estimates of the indirect ef-fects. Indeed, significance of indirect effects is the crucial requirementfor mediation to be established (Iacobucci et al., 2007; Zhao et al.,2010). The indirect effect of avoidance coping through alexithymia on al-cohol consumption turned out to be significant (.38, pb .01), lending fairsupport to the mediation hypothesis.

To test whether direct effects of avoidance coping are detectablealong with the discussed significant indirect effect, a model includingboth direct and indirect paths from avoidance coping to alcohol con-sumption was fitted. Such amodel (partial mediation) is less restrainedcompared with the total mediation model. Therefore, if the partial me-diation model fails to outperform the total mediationmodel in terms offit, the latter—more restrained and parsimonious model—would bepreferable.

Releasing the direct effect of avoidance coping on alcohol con-sumption yielded an adequate model: χ2 (17)=20.68, p=.24;RMSEA=.047 (90% CI: .00–.11), RMSEA test of close fit (RMSEAb .05),p=.49; NNFI=.98; CFI=.99. Nevertheless, adding the direct effectdid not significantly improve the fit of the model, as indicated by a

scaled χ2 (Satorra & Bentler, 2001) difference test2: Δχ2 (1)=.69,p=.44. The non‐significant difference test indicated that the hypoth-esis of no significant direct effects of avoidance coping on alcohol con-sumption should be retained. Inspection of parameter estimates ofthe direct effect confirmed that indeed no association (−.08, ns) ofavoidance coping strategies with alcohol consumption was detectableonce alexithymia was taken into account. This pattern of results fur-ther supports a full mediation hypothesis.

4. Discussion

In the present study, we first described the consumption of alco-hol among alcoholics who were alexithymic, borderline, and non-alexithymic. Our data showed alcoholics who were also alexithymicconsumed twice the amount of alcohol than non-alexithymics andwere also less abstinent before being enrolled into the rehabilita-tion program. Borderline patients showed the same patterns ofalexithymic participants, thus suggesting that there is no differencebetween them in terms of alcohol abuse. This result is particularlyinteresting for its clinical implications, suggesting that borderline-alexithymic patients should be treated as proper alexithymic in aclinical setting, as long as the gravity of the alcohol addiction is

1228 G. Coriale et al. / Addictive Behaviors 37 (2012) 1224–1229

concerned. Even if the cut-offs of the TAS-20 distinguish the threelevels of alexithymia (Bagby, Parker, et al., 1994; Bagby, Taylor, &Parker, 1994), our data basically indicate that the alexithymic cate-gory should be merged with the borderline one when dealing withalcohol abuse. This is also in line with the recent new directions ofthe literature about the structure of the alexithymia construct, con-cerning its categorical vs. dimensional nature (e.g., Parker, Keefer,Taylor, & Bagby, 2008). In fact, the field of research about psychopa-thology still accommodates both views of alexithymia as a categor-ical and dimensional variable; thus, the construct of alexithymia isaccepted as an affective deficit disorder as well as a normally dis-tributed continuous personality trait (Zackheim, 2007). There aremany studies analyzing data both ways, using the TAS-20 as a con-tinuous variable and applying the cut-offs suggested by the litera-ture (e.g., Grabe et al., 2006; Lauriola et al., 2011; Parker, Taylor, &Bagby, 1993). Taking a dimensional view, we tried to estimate theeffect of increases in alexithymia on alcohol consumption, findingan impressive association in terms of alcohol units. The increase of1 standard deviation of the TAS-20 was related to a consumptionof alcohol units (2.84 alcohol units consumed per day) that is con-sidered at risk for women and very close to risk for men. Thiscould suggest that, in some instances, high levels of alexithymia it-self may be sufficient to signal a possible alcohol abuse.

We applied the dimensional perspective in testing possible media-tionmechanisms that could explain the relationship among alexithymia,avoidance coping strategies and alcohol abuse. Mediation structures arealways an intriguing possibility when, as it happens with the variablesstudied herein, significant intercorrelations exist among the constructs(e.g., Bruce et al., 2012; Carver et al., 1989; Evren et al., 2008; Uzun etal., 2003).Wefirst tested amodelwhere avoidance coping strategiesme-diated the relationship between alexithymia and alcohol abuse. This pat-tern of mediation has been recently reported in a non‐clinical sample(Bruce et al., 2012), but this model did not adequately reproduce the ob-served data in our clinical sample. Then, we tested an alternative modelwhere alexithymiamediated the relationship between avoidance copingstrategies and alcohol abuse, which yielded a satisfactory fit. In otherwords, avoidance coping strategies had an indirect effect on alcoholabuse, which operated trough the individual's alexithymia level. Thisparticular result underlies the importance of a clinical focus on copingstrategies when dealing with alcohol abusers who are also alexithymic.In a therapeutic setting, working on the patient's coping strategies (asstated by the cognitive-behavioral coping-skills therapy for alcohol de-pendence; Longabaugh & Morgenstern, 1999), may help, on the onehand, indirectly improve abstinence (as reported in the literature; Littet al., 2003). However, on the other hand, it might also help to changehis/her alexithymic levels. This would be particularly true if we consideralexithymia as a dimensional variable and not a categorical one (Parkeret al., 2008). The role of alexithymia as a mediator between avoidancecoping strategies and alcohol abuse is also to some extent an encourag-ing result, since alexithymia has been negatively associated with thera-peutic outcomes by predicting less favourable treatment outcomes(Grabe et al., 2008; McCallum, Piper, Ogrodniczuk, & Joyce, 2003), espe-ciallywhen the patients are alcoholics (Loas et al., 1997; Ziólkowski et al.,1995). Starting from the results of this study, we could expect that after atreatment based on learningmore active coping strategies, patient levelsof alexithymia might decrease. Future studies should take into accountalexithymic levels before and after treatment on a sample of alcoholic pa-tients in a therapeutic program focused on learning new, more activeand adaptive coping skills, in order to test the possible improvement inalexithymic levels over time.

The results must be considered in light of the cross-sectional na-ture of the study. Therefore, no claim on causality and on specific di-rection of the effects can be made. Prospective studies andexperimental designs could be envisioned to test for the most in-triguing possibilities that might be suggested from the mediationmodel presented herein.

An interesting issue to be investigated further concerns the differentrole of alexithymia in social drinkers and alcoholics.Wemight speculatethat when individuals are not addicted, avoidance coping strategieswould mediate the relationship between alexithymia and alcohol con-sumption, as reported recently (Bruce et al., 2012). In our clinical sam-ple this model needed to be rejected on empirical grounds. It appearsthat before developing a dependency, alexithymia shows and indirectassociation with alcohol consumption through the use of coping strate-gies (Bruce et al., 2012). This relationship would change when alcoholconsumption turns into alcohol dependence; coping strategies couldthen be in some way blocked in the avoidance style, influencingalexithymia levels and, in turn, alcohol consumption in a sort of viciouscircle. Of course, these kinds of theoretical implications should be inves-tigated in the future to test their plausibility, ideally employing prospec-tive designs.

5. Conclusions

The present study gives additional information on the relationshipamong avoidance coping strategies, alexithymia and alcohol abuse, pro-viding some insight into the treatment of alexithymia among the alco-holic population. In particular, our study suggests that when dealingwith alcoholics, a focus on learning more adaptive coping strategiesmight help patients to drink less and bemore abstinent, eventually im-proving patients’ alexithymia levels.

Role of funding sourcesNone.

ContributorsGiovanna Coriale and Mauro Ceccanti conceived and designed the study. Fernando

Cosimi, Raffaella Porrari, and Francesca De Rosa revised the questionnaire and collect-ed the data. Elena Bilotta and Luigi Leone analyzed the data and drafted the manu-script. All authors gave final approval to the manuscript.

Conflict of interestNone of any author.

AcknowledgmentsWe are thankful to Mauro Giacomantonio for his critical feedback on earlier ver-

sions of the manuscript. The authors also thank all the students that assisted in thedata collection.

References

American Psychiatric Association (2000). Diagnostic and statistical manual of mentaldisorders, fourth edition, text revision (DSM-IV TR). Washington, DC: American Psy-chiatric Association.

Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (1994). The twenty-item TorontoAlexithymia Scale-I. Item selection and cross-validation of the factor structure.Journal of Psychosomatic Research, 38, 23–32.

Bagby, R. M., Taylor, G. J., & Parker, J. D. A. (1994). The twenty-item TorontoAlexithymia Scale: II Convergent, discriminant, and concurrent validity. Journal ofPsychosomatic Research, 38, 33–40.

Bandalos, D. L., & Finney, S. J. (2001). Item parceling issues in structural equationmodeling. In G. A. Marcoulides, & R. E. Schumacker (Eds.), New developments andtechniques in structural equation modeling (pp. 269–296). Mahwah, NJ: Erlbaum.

Baron, R. M., & Kenny, D. A. (1986). Moderator mediator variables distinction in socialpsychological research: Conceptual, strategic, and statistical considerations. Jour-nal of Personality and Social Psychology, 51, 1173–1182.

Bressi, C., Taylor, G. J., Parker, J. D. A., Bressi, G., Brambilla, V., Aguglia, E., et al. (1996).Cross validation of the factor structure of the 20-item Toronto Alexithymia Scale:An Italian multicenter study. Journal of Psychosomatic Research, 41, 551–559.

Bruce, G., Curren, C., & Williams, L. (2012). Alexithymia and alcohol consumption: Themediating effects of drinking motives. Addictive Behaviors, 37, 350–352.

Bussey Rask, M., Jørgensen, T., Pinnerup Jensen, J., Bech Jørgensen, K., Madsen, M.,Nielsen, B., et al. (2006). Influence of pretreatment coping strategies on the out-come of outpatient treatment of Danish alcohol abusers. European Addiction Re-search, 12, 83–90.

Carver, C. S., Scheier, M. F., &Weintraub, J. K. (1989). Assessing coping strategies: A the-oretically based approach. Journal of Personality and Social Psychology, 56, 267–283.

Chung, T., Langenbucher, J., Labouvie, E., Pandina, R. J., & Moos, R. H. (2001). Changes inalcoholic patients’ coping responses predict 12-month treatment outcomes. Jour-nal of Consulting and Clinical Psychology, 69, 92–100.

1229G. Coriale et al. / Addictive Behaviors 37 (2012) 1224–1229

Cleland, C., Magura, S., Foote, J., Rosenblum, A., & Kosanke, N. (2005). Psychometricproperties of the Toronto Alexithymia Scale (TAS-20) for substance users. Journalof Psychosomatic Research, 58, 299–306.

De Rick, A., & Vanheule, S. (2006). The relationship between perceived parenting, adultattachment style and alexithymia in alcoholic inpatients. Addictive Behaviors, 31,1265–1270.

De Rick, A., & Vanheule, S. (2007). Alexithymia and DSM-IV personality disorder traitsin alcoholic inpatients: A study of the relation between both constructs. Personalityand Individual Differences, 43, 119–129.

Evren, C., Kose, S., Sayar, K., Ozcelik, B., Borckardt, J. P., Elhai, J. D., et al. (2008).Alexithymia and temperament and character model of personality in alcohol de-pendent Turkish men. Psychiatry and Clinical Neurosciences, 62, 371–378.

Franz, M., Popp, K., Schaefer, R., Sitte, W., Schneider, C., Hardt, J., et al. (2008).Alexithymia in the German general population. Social Psychiatry and Psychiatric Ep-idemiology, 43, 54–62.

Grabe, H. J., Frommer, J., Ankerhold, A., Ulrich, C., Groger, R., Franke, G. H., et al. (2008).Alexithymia and outcome in psychotherapy. Psychotherapy and Psychosomatics, 77,189–194.

Grabe, H. J., Ruhrmann, S., Ettelt, S., Muller, A., Buhtz, F., Hocrein, A., et al. (2006).Alexithymia in obsessive–compulsive disorder: Results from a family study. Psy-cotherapy and Psychosomatics, 75, 312–318.

Hasking, P. A., & Oei, T. P. S. (2007). Alcohol expectancies, self-efficacy and coping in analcohol-dependent sample. Addictive Behaviors, 32, 99–113.

Haviland, M. G., Hendryx, M. S., Shaw, D. G., & Henry, J. P. (1994). Alexithymia inwomen and men hospitalized for psychoactive substance dependence. Comprehen-sive Psychiatry, 35, 124–128.

Haviland, M. G., Shaw, D. G., Cummings, M. A., & MacMurray, J. P. (1988). Alexithymia:Subscale and relationship to depression. Psychotherapy and Psychosomatics, 50,164–170.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structureanalyses: Conventional criteria versus new alternatives. Structural Equation Model-ing, 6, 1–55.

Iacobucci, D., Saldanha, N., & Deng, X. (2007). A meditation on mediation: evidencethat structural equations models perform better than regressions. Journal of Con-sumer Psychology, 17, 139–153.

Jöreskog, K. G., & Sörbom, D. (1996). LISREL 8: User's reference guide. Chicago, IL:Scientific Software International, Inc.

Kauhanen, J., Julkunen, J., & Salonen, J. T. (1992). Coping with inner feelings and stress:Heavy alcohol use in the context of alexithymia. Behavioral Medicine, 18, 121–126.

Lauriola, M., Panno, A., Tomai, M., Ricciardi, V., & Potenza, A. E. (2011). Is alexithymia relatedto colon cancer? A survey of patients undergoing a screening colonoscopy examination.Journal of Clinical Psychology in Medical Settings. http://dx.doi.org/10.1007/s10880-011-9267-y.

Lejoyeux, M., Solomon, J., & Ades, J. (1998). Benzodiazepine treatment foralcohol-dependent patients. Alcohol, 33, 563–575.

Lindsay, J., & Ciarrochi, J. (2009). Substance abusers report being more alexithymicthan others but do not show emotional processing deficits on a performance mea-sure of alexithymia. Addiction Research and Theory, 17, 315–321.

Litt, M. D., Kadden, R. M., Cooney, N. L., & Kabela, E. (2003). Coping skills and treatmentoutcomes in cognitive-behavioral and interactional group therapy for alcoholism.Journal of Consulting and Clinical Psychology, 71, 118–128.

Loas, G., Fremaux, D., Otmani, O., Lecercle, C., & Delahousse, J. (1997). Is alexithymia anegative factor for maintaining abstinence? A follow-up study. Comprehensive Psy-chiatry, 38, 296–299.

Longabaugh, R., & Morgenstern, J. (1999). Cognitive-behavioral therapy for alcohol de-pendence: Current status and future directions. Alcohol Research & Health, 23,78–85.

Lumley, M. A., Neely, L. C., & Burger, A. J. (2007). The assessment of alexithymia in med-ical settings: Implications for understanding and treating health problems. Journalof Personality Assessment, 89, 230–246.

McCallum, M., Piper, W. E., Ogrodniczuk, J. S., & Joyce, A. S. (2003). Relationships amongpsychological mindedness, alexithymia and outcome in four forms of short-termpsychotherapy. Psychology and Psychotherapy, 76, 133–144.

McKee, S. A., Hinson, R. E., Wall, A., & Spriel, P. (1998). Alcohol outcome expectanciesand coping styles as predictors of alcohol use in young adults. Addictive Behaviors,23, 17–22.

Moser, A. E., & Annis, H. M. (1996). The role of coping in relapse crisis outcome: A pro-spective study of treated alcoholics. Addiction, 91, 1101–1113.

Nemiah, J. C. (1977). Alexithymia: Theoretical considerations. Psychotherapy and Psy-chosomatics, 28, 199–206.

Nemiah, J. C., & Sifneos, P. E. (1970). Psychosomatic illness: A problem in communica-tion. Psychotherapy and Psychosomatics, 18, 154–160.

Parker, J. D. A., Keefer, K. V., Taylor, G. J., & Bagby, R. M. (2008). Latent structure ofalexithymia construct: A taxometric investigation. Psychological Assessment, 20,385–396.

Parker, J. D. A., Taylor, G. J., & Bagby, R. M. (1993). Alexithymia and the processing ofemotional stimuli: An experimental study. New Trends Experimental Clinical Psychi-atry, 9, 9–14.

Rybakowski, J., Ziólkowski, M., Zasadzka, T., & Brzezinski, R. (1988). High prevalence ofalexithymia in male patients with alcohol dependence. Drug and Alcohol Depen-dence, 21, 133–139.

Salminen, J. K., Saarijärvi, S., Äärelä, E., Toikka, T., & Kauhanen, J. (1999). Prevalence ofalexithymia and its association with sociodemographic variables in the generalpopulation of Finland. Journal of Psychosomatic Research, 46, 75–82.

Satorra, A., & Bentler, P. M. (2001). A scaled difference chi-square test statistic for mo-ment structure analysis. Psychometrika, 66, 507–514.

Sauvage, L., & Loas, G. (2006). Criterion validity of Bermond–Vorst AlexithymiaQuestionnaire-20 form B: A study of 63 alcoholic subjects. Psychological Reports,98, 234–236.

Sica, C., Magni, C., Ghisi, M., Altoè, G., Sighinolfi, C., Chiri, L. R., et al. (2008). Coping ori-entation problems experienced—Nuova versione italiana (COPE-NVI): Unostrumento per la misura degli stili di coping. Psicoterapia Cognitiva e Com-portamentale, 14, 27–53.

Skinner, H. A., & Sheu, W. J. (1982). Reliability of alcohol use indices: The lifetimedrinking history and the MAST. Journal of Studies on Alcohol, 43, 1157–1170.

Sobell, L. C., & Sobell, M. B. (1992). Timeline follow-back: A technique for assessingself-reported alcohol consumption. In R. Litten, & J. Allen (Eds.), Measuring alcoholconsumption (pp. 41–72). Totowa, NJ: Humana Press.

Stasiewicz, P. R., Bradizza, C. M., Gudleski, G. D., Coffey, S. F., Schlauch, R. C., Bailey, S. T.,et al. (2012). The relationship of alexithymia to emotional dysregulation within analcohol dependent treatment sample. Addictive Behaviors, 37, 469–476.

Taylor, G. J. (2000). Recent developments in alexithymia theory and research. CanadianJournal of Psychiatry, 45, 134–142.

Taylor, G. J., & Bagby, R. M. (2004). New Trends in alexithymia research. Psychotherapyand Psychosomatics, 73, 68–77.

Taylor, G. J., Bagby, R. M., & Parker, J. D. A. (1990). A preliminary investigation ofalexithymia in men with psychoactive substance dependence. The American Jour-nal of Psychiatry, 147, 1228–1230.

Taylor, G. J., Bagby, R. M., & Parker, J. D. A. (1997). Disorders of affect regulationalexithymia in medical and psychiatric illness. Cambridge: Cambridge UniversityPress.

Thorberg, F. A., Young, R. McD, Sullivan, K. A., Lyvers, M., Hurst, C. P., Connor, J. P., et al.(2011). Alexithymia in alcohol dependent patients is partially mediated by alcoholexpectancy. Drug and Alcohol Dependence, 116, 238–241.

Uzun, A. Z., Ates, A., Cansever, A., & Ozsahin, A. (2003). Alexithymia in male alcoholics:Study in a Turkish sample. Comprehensive Psychiatry, 44(4), 349–352.

Williams, A., & Clark, D. (1998). Alcohol consumption in university students: The roleand reasons for drinking, coping strategies, expectancies, and personality traits.Addictive Behaviors, 23, 371–378.

Wise, T. N., Mann, L. S., & Hill, B. (1990). Alexithymia and depressed mood in the psy-chiatric patient. Psychotherapy and Psychosomatics, 54, 26–31.

Wise, T. N., Mann, L. S., & Shay, L. (1992). Alexithymia and the five-factor model of per-sonality. Comprehensive Psychiatry, 33, 147–151.

Zackheim, L. (2007). Alexithymia: The expanding realm of research. Journal of Psycho-somatic Research, 63, 345–347.

Zhao, X., Lynch, J. G., Jr., & Chen, Q. (2010). Reconsidering Baron and Kenny: Myths andtruths about mediation analysis. Journal of Consumer Research, 37, 197–206.

Ziólkowski, M., Gruss, T., & Rybakowski, J. K. (1995). Does alexithymia in male alco-holics constitute a negative factor for maintaining abstinence? Psychotherapy andPsychosomatics, 63, 169–173.