Comorbid alcohol dependence and anxiety disorder: Effect of concurrent mood and personality disorders on treatment outcomes

Alcohol dependent individuals who also were diagnosed with an anxiety disorder were treated with one of two cognitive-behavioral treatments. Treatment ALC, consisting of 6 sessions of alcohol-specific treatment was compared to treatment ALCANX consisting of 6 sessions of alcohol-specific treatment followed by 4 anxiety-specific sessions. At the end-of-treatment, no significant differences on measures of alcohol consumption or psychiatric (including anxiety) symptomatology were found. Both groups improved to a significant degree. Post-hoc analyses, based on the high degree of other, especially mood and personality disorder, comorbidity in addition to the alcohol use/anxiety disorder indicated that individuals who had the most comorbidity did the poorest on alcohol and psychiatric measures. Implications for the cognitive-behavioral treatment of concurrent alcohol dependence and anxiety disorder are discussed.


Introduction
Rates of psychiatric comorbidity between alcohol use and anxiety disorders in both clinical and community populations is well established [1,2]. Such comorbidity may be associated with predict poorer alcohol treatment outcomes and higher rates of relapse [3]. Despite the robust association between anxiety and alcohol problems relatively little treatment research has been reported [4][5][6][7][8]. Recently, Toneatto and Calderwood (2015) reviewed this literature reported that alcohol dependent individuals also diagnosed with an anxiety disorder who received alcohol-specific treatment consisting of 6 sessions of was compared to individuals who received 6 sessions of alcohol-specific treatment followed by 4 anxiety-specific sessions. At the end-oftreatment no significant differences were found on both measures of alcohol consumption or psychiatric (including anxiety) symptoms. Both groups improved to a significant degree up to the 10-month follow-up. Despite the lack of clinical impact of the two treatments, the role of additional psychiatric comorbidity (in addition to alcohol and anxiety) may be an important variable in therapeutic response. In that study, almost a third of the sample also had a mood disorder; unpublished data also revealed high rates of personality disorders as well. Mood disorders [5,[9][10][11] and personality disorders [12,13] among alcohol use disorders are also common. Whether treating concurrent psychiatric comorbidity in primary addiction problems, whether pharmacologically or psychotherapeutically, continue to be an important clinical issue. [4, 5,9] have examined this issue and have concluded that much more research is required before firm conclusions can be made but that in general, treatment comorbidity among those with alcohol use disorders may improve addiction-related outcomes and quality of life.
In light of the complex comorbidity associated with alcohol use disorders, the impact of mood and personality disorders, in addition to the primary anxiety and alcohol use disorders that defined the inclusion criteria for [14], on both addiction and psychiatric outcomes, post-treatment and at a 10-month follow-up were investigated.

Method Sample
Participants meeting DSM-IV (American Psychiatric Association, 1994) diagnostic criteria for concurrent anxiety disorder and alcohol dependence were primarily recruited from the Intake Service of the Center for Addiction and Mental Health (CAMH), a University of Toronto teaching hospital and the largest addiction and mental health facility in Canada providing both ambulatory and outpatient services for addiction and mental health. The study was approved by the CAMH Research Ethics Board.

Design
Although complete details of the study are presented in Toneatto and Calderwood 2015, the core details are presented here. Participants who met the inclusion criteria were randomly assigned to 6 sessions of alcohol-only treatment (ALC) or to an enhanced treatment consisting of 6 alcohol-only sessions followed by 4 sessions of anxiety coping skills training (ALCANX). Treatment was administered on an individual basis.

Assessment
Frequency and quantity of daily alcohol consumption at baseline was assessed using the 60-day Timeline Follow Back Interview [15,16]. The Symptom Checklist-90 (SCL90), a 90-item self-report rating scale that measures distress on 9 psychiatric dimensions corresponding to several DSM-IV Axis I disorders [17] assessed past-week psychiatric functioning. The Global Severity Index (GSI) of the SCL90 provides an overall index of psychiatric distress ranging from 0, 'not at all' to 4, 'extremely'. The Beck Anxiety Inventory [18], a 21-item validated [19] was used to assess the severity of anxiety symptoms in the past week on a 4-point rating scale ranging from 0, 'not at all' to 4, 'I could barely stand it'. Participants were re-administered the TLFB (past two weeks), SCL90 (past week) and the BAI (past week) at post-treatment and 10 months post-treatment. The follow-up assessment consisted of the TLFB, the BAI, and the SCL90.

Treatment
The ALC treatment consisted of 6 sessions of cognitive-behavioral treatment for alcohol problems and consisted of motivational, behavioral, cognitive and affect-based sessions designed to reduce drinking or achieve abstinence (cf. Toneatto and Calderwood, 2015 for additional details). Group ALC did not receive anxiety-specific treatment. The first 6 sessions of the ALCANX treatment were identical to those of group ALC. The final four sessions consisted of a functional analysis of the anxiety and the development of cognitive and behavioral coping skills to manage anxiety and reduce avoidance behavior. Both treatments included reading assignments, handouts and behavioral exercises. Treatment was administered by two doctoral-level clinical psychologists (average of 11 years post-graduate therapy experience) and 2 masters level therapists (average of 3 years post-graduate experience) experienced in treating concurrent disorders.

Data analysis
Mood disorders [MOOD] were diagnosed in 47.5% of the sample and personality disorders [PD] in 59% at assessment. Based on these patterns of psychopathology, concurrent MOOD or PD disorder were identified as independent variables in addition to the primary anxiety and alcohol use disorder defining entry into the study. Thus, four types of comorbid alcohol-dependent subjects were formed: anxiety disorder [ANX, n = 11], anxiety and personality disorder [ANX-PD, n = 18], anxiety and mood disorder [ANX-MOOD, n = 13], anxiety, mood and personality disorder [ANX-MOOD-PD, n = 13]. Separate 2 MOOD (presence, absence) X 2 PD (presence, absence) multivariate analyses of variance (MANOVA) were conducted on baseline and end-of-treatment alcohol consumption and psychiatric symptomatology. The end-of-treatment quantity and frequency data were square root-transformed to correct non-normal distributions but only untransformed values are reported. Roy-Bargman stepdown F tests were used to assess the impact of the independent variables on the dependent variables [19]. MANOVA assumptions of linearity, multicollinearity, normality and homogeneity of variance-covariance matrices were met. Within group mean comparisons were tested with one-tailed paired t-tests. Associations between categorical variables were tested with Pearson chi-square.

Subjects
Fifty-five (67%) of the 82 participants who completed the baseline assessment also completed treatment (34 in group ALC and 21 in group ALCANX). Treatment completion was defined as attendance at > 3 sessions of the ALC and > 8 of the ALCANX treatments. Of the 27 non-completers, 21 failed to complete the required number of sessions, 5 subjects decided that treatment was unsuitable for them shortly after beginning treatment and one subject was terminated by the therapist. Significant differences were found between completers and non-completers.

Psychometric measures
A significant PD main effect on measures of baseline psychiatric symptomatology and urges to drink was found, Wilk's Λ = 0.84, F(3, 49) = 3.07, p < .05. Only baseline Global Severity Index (GSI) scores significantly different between groups as evaluated by the stepdown F-test, F(1, 49) = 7.42, p < .01. Participants diagnosed with a personality disorder had a higher GSI, M = 1.5, SD = 0.7, compared to those without an Axis II diagnosis, M = 1.3, SD = 0.5. Groups ANX-MD and ANX-PD showed significant reductions in BAI scores over treatment, t (12) = 6.38, p < .001; t (16) = -4.15, p < .001, respectively. Groups ANX-MD, ANX-PD and ANX also showed significant reductions on the GSI, t (9) = 4.83, p < .001; t (15) = 4.88, p < .0005, and t (9) = 2.95, p < .01, respectively. Table 3 Discussion Treatment outcome was correlated with the severity of psychopathology. Although participants with alcohol and anxiety disorders were explicitly recruited for this study, a significant proportion (44/55 or 80%) of the sample was diagnosed with an additional concurrent mood or personality disorder, is consistent with other research. For example [19][20][21] found Cluster C personality disorders and mood disorders, especially major depressive episodes and dysthymic disorder, to be commonly diagnosed in anxiety patients. [22] Presenting data from a cross-sectional sample of treatment-seeking anxious patients, found 50% to have an additional mood or anxiety disorder. Axis II disorders have also been shown to be consistently prevalent in alcohol abusers and to predict poorer treatment response [21,23,24]. however, suggest that the mere presence of a personality disorder is not an indication that it has a causal role in negatively influencing treatment since such individuals may differ on several other clinically relevant variables which can also impact on treatment response and outcome. For [24] reported that anxiety and personality disorder are often associated with greater severity of depression than those without a personality disorder and that depression may mediate treatment outcomes. Similarly, [25] found that patients with both mood and anxiety disorders were associated with greater psychosocial impairment than participants with either mood or anxiety disorders. [26] found that depression was the primary reason treated anxious alcoholics relapsed following treatment. [27] have shown that the aggregation of comorbid alcoholism, anxiety and depression in families may reflect the influence of genetic variables.
Other evidence suggests, however, that personality disorders may also predict positive therapeutic outcomes. [28] found that personality disorders were associated with poorer outcomes following interpersonal therapy but not cognitive therapy (e.g., dependency traits may aid in the commitment of the client to treatment). [29] argue that structured treatments, such as cognitive-behavioral therapy, may be preferred to insight-oriented therapy in treating Axis II because of the goal-centered and coping-skill focus of such treatments. In that study, [29] found cocaine abusers with borderline characteristics did worse on drug-related outcome measures following behavior therapy than did those with dependent personality features.

MOOD-ANX-PD MOOD-ANX ANX-PD ANX
The present research found that alcohol-dependent anxious subjects with additional mood and Axis II comorbidity were drinking more heavily at baseline and achieved poorer drinking outcomes at posttreatment. Participants who had either, or neither, a mood or Axis II disorder in addition to the alcohol and anxiety problem, showed comparable rates of improvement. Thus, specific psychiatric disorders, per se, were not predictors of negative outcomes. These results support the hypothesis that severity of psychopathology may be more important than type of psychopathology [30,31]. In this study, the presence of multiple psychiatric disorders in type MD-ANX-PD was correlated with a significantly lower GAF score, heavier drinking, and higher GSI score at baseline suggesting that this subgroup indeed had more severe psychopathology. Participants who had either, or neither, a mood or Axis II disorder in addition to the alcohol and anxiety problem, had relatively higher GAF scores, less heavy drinking and lower GSI.
Individuals who are depressed may lack effective behavioral (e.g., withdrawn), cognitive (e.g., helpless) and affective (e.g., anhedonic) regulation skills [32]. The comorbid presence of personality dysfunction, characterized by severe cognitive-perceptual aberrations which may distort responses to interpersonal events, emotions and cognitions [33] and proneness to impulsive actions which can produce, or exacerbate, psychiatric distress [34], may further accentuate deficits in coping effectiveness. The presence of these symptoms, in individuals who are also anxious and alcohol dependent, may interfere with the optimal response to the brief intervention provided in this study.
The small sample sizes in the post-hoc analyses prevented examination of differences in treatment response by each of the four types as a function of the ALC or ALCANX treatment. Participants with the most severe psychopathology (i.e., type ANX-MD-PD) demonstrated little change over the course of treatment (combining results from the ALC and ALCANX treatments). Thus, it is not clear whether the inclusion of a longer anxiety-specific component [35,[36][37][38] or interventions directed at other concurrent Axis I (i.e., depression, [26]) or Axis II (i.e., development of client-therapist relationship, interpersonal dysfunction, poor insight, therapeutic ambivalence; [28]) psychopathology would produce better outcomes. The results from this study do suggest that a primarily alcohol-focused treatment may be inadequate for more severe comorbid disorders but may be relatively effective for less severe comorbid psychopathology as found for types ANX, ANX-PD, and ANX-MD.
The results of the current study are weakened by the small sample size, the lack of sensitive measures of the mediators of change, and the lack of control over the use of other psychoactive substances (including prescribed medications) [39][40][41]. The longer term effects of the treatment needs to be assessed in order to determine whether the benefits of ALCANX treatment, for example, are delayed. The determination of psychiatric diagnoses very early in treatment, although made by experienced clinicians using a structured psychiatric interview, occurred in the context of ongoing alcohol use. Consequently, the risk of confusing organic effects of alcohol ingestion for true psychiatric symptoms is considerable. The diagnostic assessment should ideally occur when the individual is substance-free, a difficult goal when the participants are attending an outpatient treatment program which permits non-abstinence.
In addition, more reliable measure of adherence to the treatment protocol would be desirable through the rating of randomly audiotaped therapy sessions. Future research should include theoretically meaningful measures of therapeutic change (e.g., anxiety sensitivity, impulse control) and the measurement of a wider variety of clinicallyrelevant variables reflective of the psychiatric heterogeneity of comorbid substance abusers.