درمان منطقی - احساسی و شناختی رفتاری (REBT/CBT) در مقابل دارودرمانی در برابر REBT/CBT به علاوه دارودرمانی در درمان اختلال افسردگی اساسی در جوانان؛ کارآزمایی بالینی تصادفی

Major depressive disorder is a highly prevalent and debilitating condition in youth, so developing efficient treatments is a priority for mental health professionals. Psychotherapy (i.e., cognitive behavioral therapy/CBT), pharmacotherapy (i.e., SSRI medication), and their combination have been shown to be effective in treating youth depression; however, the results are still mixed and there are few studies engaging multi-level analyses (i.e., subjective, cognitive, and biological). Therefore, the aims of this randomized control study (RCT) were both theoretical - integrating psychological and biological markers of depression in a multi-level outcome analysis - and practical – testing the generalizability of previous results on depressed Romanian youth population. Eighty-eight (N=88) depressed Romanian youths were randomly allocated to one of the three treatment arms: group Rational Emotive Behavior Therapy (REBT)/CBT (i.e., a form of CBT), pharmacotherapy (i.e., sertraline), and group REBT/CBT plus pharmacotherapy. The results showed that all outcomes (i.e., subjective, cognitive, and biological) significantly change from pre to post-treatment under all treatment conditions at a similar rate and there were no significant differences among conditions at post-test. In case of categorical analysis of the clinical response rate, we found a non-significant trend favoring group REBT/CBT therapy. Results of analyses concerning outcome interrelations are discussed.

1.1. The problem Major depressive disorder in youth is one of the most prevalent and debilitating psychiatric disorders for this age group (Costello et al., 2003), with prevalence rates ranging from 2.8% in children under 13 and 5.6% in adolescents (Costello et al., 2006) and total incidence rate ranging from 5% in children to 20% in adolescents, similar to the adult incidence rate (Rohde et al., 2013). Depression in youth is associated with an increased risk for other psychiatric disorders (e.g., Costello et al., 2003) and/or difficulties in social functioning and school performance (National Institute for Health and Clinical Excellence/NICE, 2005), with a higher risk of reoccurrence in adulthood (Harrington et al., 1990). Although there are evidence-based treatments for depression in youth, both pharmacological and psychotherapeutic, at least between one third and one half of depressed youth still do not respond to treatment (Bridge et al., 2007 and Weisz et al., 2006), and almost half of treated youths experience recurrence within 4 years (Curry et al., 2011). 1.2. The treatment of depression in youth; a brief analysis Following the principles of evidence-based practice, several major randomized control trials (RCTs) have shown that both psychosocial treatments, particularly cognitive behavioral therapy (CBT), and pharmacological treatments (i.e., medication), particularly selective serotonin reuptake inhibitors (SSRI), are effective, both separately and in combination (e.g., Bridge et al., 2007, Emslie et al., 1997 and Vitiello, 2009) for treating youth depression. For instance, one large-scale RCT, the TADS study (Treatment for Adolescents With Depression Study Team, 2004), compared CBT, pharmacotherapy, and their combination in treating major depressive disorder in adolescents and found that combining fluoxetine with CBT appears superior to the other treatments alone (i.e., on the Clinician׳s Global Index; CGI) at the end of the acute treatment (TADS team, 2007) and it is more effective in preventing suicidality in the long term (TADS team, 2007). Another large-scale RCT, the TORDIA study (Emslie et al., 2010), compared switching medication, and the combination of medication plus CBT in treating SSRI-resistant adolescent depression (i.e, major depressive disorder and dysthymia). The results confirmed that the combination is more effective than medication alone in reducing depressive symptoms. However, other reviews have found only medium results for CBT in treating youth depression (Weisz et al., 2006) and the combination of CBT and medication has also yielded mixed results (Dubicka et al., 2010), with some data showing that, at least for moderate to severe major depression in adolescents, it is not clear that combining CBT with medication adds significantly to treatment efficacy (see the ADAPT study; Goodyer et al., 2008). Therefore, conducting more RCTs investigating pharmacotherapy, CBT, and their combination in treating youth depression is extremely important, as many clinical aspects (e.g., multi-level analyses of the outcomes) related to treatment response (e.g., efficacy/effectiveness) and its mechanisms of change still remain to be clarified. In this sense, investigating new CBT approaches would be a promising direction given the fact that current treatments still elicit only a moderate success rate, with between one third and one half of youths still not responding to treatment (Bridge et al., 2007 and Weisz et al., 2006). For instance, REBT, rational emotive behavior therapy/REBT – a form of CBT – which assumes that core irrational beliefs generate distorted automatic thoughts that further generate dysfunctional consequences (e.g., dysfunctional feelings, maladaptive behaviors) – has been found (see David et al., 2008) effective in treating major depressive disorder in adult population when compared to pharmacotherapy (being included in the NICE Guidelines; see NICE, 2010), but no such results have yet been reported for a youth population. Another important issue is related to reported outcomes. First, for instance, when a psychosocial intervention is used, investigators typically focus on psychological outcomes (e.g., subjective, cognitive, behavioral) and often ignore biological outcomes. When a pharmacotherapy intervention is used, typically the focus is on the clinical symptoms and biological outcomes (e.g., platelet serotonin reuptake, Axelson et al., 2005; dopamine, norepinephrine, platelet serotonin, Goodnick et al., 1995), and key psychological outcomes and mechanisms of change (e.g., cognitions) are ignored. Second, most of the studies investigating cognitive factors have been conducted on adult population. Secondary analyses conducted within the TADS study have shown that, for example, clinical improvement is mediated by changes in perfectionism (Jacobs et al., 2009), but there are still very few studies employing measures of distorted thinking with adolescents receiving CBT. Third, although research focusing on biological markers of depression (i.e., biological factors) has found evidence for the role of monoamines in the onset of depression, particularly serotonin (5-HT) and norepinephrine (NE), few RCTs including a psychotherapy arm have included biological parameters as outcomes or if they did, they were mainly related to adult rather than to adolescent depression. For instance, changes in serotonin uptake have been correlated to improvements in depressive symptoms following SSRI treatment (e.g., Axelson et al., 2005) and platelet serotonin is related to symptom severity in adult patients treated with citalopram (Fišar et al., 2008). Studies investigating SSRI treatment response correlates of depressive symptoms have found that platelet serotonin levels decrease after SSRI treatment (e.g., Goodnick et al., 1995 and Maurer-Spurej et al., 2003), while the results for plasmatic serotonin remain unclear, with some studies reporting a downward trend with fluoxetine treatment (e.g., Alvarez et al., 1999), while other studies report increases in serotonin levels following fluoxetine treatment (Blardi et al., 2002). With regard to norepinephrine, studies found that administering sertraline to healthy subjects leads to a decrease in plasma norepinephrine compared to placebo (Shores et al., 2001) and the trend appears to be downward for other antidepressants or electroconvulsive therapy (Owens, 1996). 1.3. Overview of the present study Given that youth depression is a complex disorder, measuring multi-level outcomes in a RCT involving psychotherapy, pharmacotherapy, and their combination would bring relevant information about treatment effects. Therefore, the current RCT aimed to: (1) examine the efficacy of group REBT/CBT, pharmacotherapy, and their combination for depression in youth; this is necessary due to the mixed nature of the previous results; (2) bring new innovations in the field by engaging in the same design: (a) a multi-level analysis of the outcomes (e.g., subjective, cognitive, and biological) and (b) a new and potentially more efficacious CBT strategy (i.e., rational emotive behavior therapy/REBT); this is fundamental because although the standard CBT strategies seem to work, they still miss a large segment of patients; and (3) investigate the generalizability and stability of the current results on a new population (i.e., Romanian); this is necessary as most of the previous studies were conducted on English-speaking populations. In order to provide a more cost-effective intervention, we used a group format for this study, given the fact that group CBT has been shown to be effective in treating youth depression (David-Ferdon and Kaslow, 2008), yielding similar results to individual psychotherapy (Weisz et al., 2006).

3. Results 3.1. Participants Over 200 participants were initially assessed for eligibility, and 88 were randomized under the three treatment conditions. Participants who were initially assessed and were not further recruited in the study were excluded because they did not meet the inclusion and/or exclusion criteria (60%), or refused to participate (40%) for various reasons (e.g., unable to attend all the visits for assessment and/or intervention due to distance etc.). In the medication group, 33 out of 33 participants received intervention, in the group REBT/CBT alone group 26 of 28 received intervention, and 26 out of 27 participants received intervention in the combined treatment group (see Fig. 1). 3.2. Pharmacotherapy In addition to sertraline, some of the participants received other psychotropic medication, namely alprazolam, risperidone, olanzapine, or zolpidem. The additional medication was necessary in some cases in order to address comorbidities (e.g., tics and Tourette syndrome, disruptive behavior). 3.3. Attendance Participants in the three groups attended treatment sessions at an approximately equal rate. In the pharmacotherapy group, five participants did not complete treatment (15.15%), in the group REBT/CBT group five participants did not complete treatment (17.85%), while in the group REBT/CBT plus pharmacotherapy group four participants did not complete treatment (14.81%). 3.4. Adherence to group REBT/CBT protocol Group REBT/CBT was delivered by an experienced psychotherapist (BM) and co-psychotherapist (RM), certified (at supervisory level) by the Albert Ellis Institute/AEI (the original site of REBT). Apart from their professional expertise and qualifications, the therapists were specifically and intensively trained to apply the protocols developed for the study, the training lasting for 14 days. Treatment adherence was assessed using a monitoring and calibrating system. Meetings involving all therapists (including those administering pharmacotherapy) were held every 2 weeks and therapists were provided feedback and guidance. A sample of sessions were analyzed by the second author (DD, a supervisor certified by the AEI), using the REBT competency scale (David, 2007), an instrument adapted after Therapy Adherence and Competency Scale (Young et al., 2006), passing satisfactorily the quality criteria. 3.5. Outcome analyses The means and standard deviations for the pretreatment, midtreatment, and post-treatment main outcomes under the three treatment conditions are displayed in Table 2. As previously mentioned, we considered subjective/symptom-related outcomes (CDI; POMS), cognitive (ATQ), and biological outcomes (serum serotonin and norepinephrine). The results are presented for the entire sample, by providing intent-to-treat scores as post-treatment results. Table 2. Means (and Standard Deviations) of outcome variables (N=88) at pretreatment, mid-treatment (8 weeks), and post-treatment (Intent-to-treat), time-treatment interaction, time effect, pre-post effect sizes, and percentage of participants scoring higher at pretreatment than the average participant at post-treatment. Pre-treatment 8 weeks Post-treatment Time-treatment interaction, P value Time effect (pre-post), P value Pre-post effect size (Cohen׳s d) % Pre-treatment participants higher than the average Post-treatment participant CDI total score REBT/CBT 23.60 (5.82) 21.12 (6.67) 15.92 (6.49) F (2, 82)=0.543; p=0.583 F (1, 82)=71.059; <0.001 1.201 88 Medication 24.78 (6.33) 20.10 (8.55) 15.39 (8.76) 1.179 88 Combined 23.48 (5.14) 21.75 (7.29) 16.80 (8.78) 0.845 79 POMS (distress) REBT/CBT 75.91 (31.07) 61.97 (25.97) 37.44 (25.06) F (2, 82)=1.402; p=0.252 F (1, 82)=55.622; <0.001 1.311 90 Medication 70.45 (36.10) 44.93 (25.86) 42.00 (33.77) 0.791 79 Combined 61.85 (33.98) 51.84 (35.48) 38.30 (34.51) 0.664 76 ATQ REBT/CBT 46.21 (11.52) 41.14 (11.78) 27.78 (9.52) F (2, 82)=1.078; p=0.345 F (1, 82)=104.751; <0.001 1.670 76 Medication 45.51 (11.89) 32.89 (10.52) 31.09 (15.72) 0.991 82 Combined 43.14 (13.32) 36.80 (14.79) 29.46 (12.64) 1.017 84 Serotonin REBT/CBT 248.42 (113.70) – 174.17 (125.63) F (2, 72)=0.693; p=0.503 F (1, 72)=16.856; <0.001 0.599 73 Medication 210.92 (113.18) – 115.42 (117.29) 0.805 79 Combined 274.00 (288.07) – 113.09 (91.37) 0.441 66 Norepinephrine REBT/CBT 358.16 (178.15) – 292.35 (188.43) F (2, 72)=0.797; p=0.454 F (1, 72)=8.928; 0.004 0.347 62 Medication 406.49 (199.25) – 320.52 (140.37) 0.480 69 Combined 450.29 (251.55) – 295.10 (131.58) 0.685 76 Note. Standard deviations are given in parentheses. REBT/CBT=Rational Emotive Behavior Therapy/Cognitive Behavior Therapy; CDI=Child Depression Inventory; POMS (distress)=Profile of Mood States, Distress score; ATQ=Automatic Thoughts Questionnaire. Table options 3.6. Between group analyses 3.6.1. Pretreatment We assessed pretreatment group differences by using one-way analyses of variance (ANOVAs), and we found no significant differences among conditions in terms of CDI total scores, F(2, 85)=0.472, p>0.05; POMS distress scores, F(2, 85)=1.23, p>0.05; ATQ scores, F(2, 85)=0.476, p>0.05; and serotonin, F(2, 85)=0.884, p>0.05 and norepinephrine levels, F(2, 85)=1.317, p>0.05. 3.6.2. Outcome at 8 weeks (mid-treatment) Mid-treatment outcome (8 weeks) differences were computed using one-way analyses of variance (ANOVAs), and we found no significant differences among the three conditions on the CDI total scores, F(2, 74)=0.317, p>0.05; POMS distress scores, F(2, 75)=2.226, p>0.05; and ATQ scores, F(2, 75)=2.904, p>0.05. 3.6.3. Outcome at post-treatment 3.6.3.1. Continuous analyses Post-treatment outcome differences were computed using one-way analyses of variance (ANOVAs), and we found no significant differences among the three conditions on the CDI total scores, F(2, 84)=0.220, p>0.05; POMS distress scores, F(2, 84)=0.177, p>0.05; ATQ scores, F(2, 84)=0.460, p>0.05; and serotonin, F(2, 74)=2.222, p>0.05; and norepinephrine levels, F(2, 74)=0.268, p>0.05. 3.6.3.2. Categorical analyses Clinical response rate at post-treatment was defined as scores lower than 19 on the CDI (see also Kazdin, 1989), meaning that the patients do not meet the criteria for clinical depression anymore (i.e., the inclusion criteria). There was a 67.85% response rate in the group REBT/CBT alone group, a 60.60% response rate under the medication condition, and a 53.84% response rate in the medication plus group REBT/CBT condition. Even though there was a positive trend favoring group REBT/CBT, the differences were not significant (all ps>0.05). 3.7. Time-treatment interaction and within group analyses We used repeated measures ANOVAs in order to test for time-treatment interaction effects for all outcomes. While there are no differences among treatment groups related to improvement, there are significant differences from pre to mid and post-treatment for all outcomes. Table 2 displays the summary statistics for subjective, cognitive, and biological outcomes, including time-treatment interaction F values, time effect F values, effect sizes (Cohen׳s d), and the percentage of pretreatment participants who would score above the average post-treatment participant (i.e., higher scores indicating more symptoms). This last indicator, signifying a transformation of Cohen׳s d to percentages, was proposed by Coe (2002) and allows for a more clinically-oriented data presentation, beyond significance values and effect sizes. Mean outcomes by treatment groups for the most important outcomes (i.e., CDI, ATQ, and sertraline, and norepinephrine levels) are presented in Fig. 2. Full-size image (43 K) Fig. 2. Mean outcomes for CDI total score, and ATQ at pretreatment, week 8, and post-treatment, and mean outcomes for serum serotonin at pretreatment and post-treatment. Figure options 3.8. Additional analyses Results also showed that changes (delta change pre-post scores) in symptoms (CDI) are significantly related to changes (pre-post) in ATQ scores, r=0.421, p<0.001. Also, pre-mid changes in ATQ were prospectively associated to the pre-post changes in CDI total scores (r=0.220, p=0.026). These results are consistent to REBT/CBT theory. However, changes (pre-post) in symptom scores (CDI) and ATQ are not related to changes (pre-post) in serotonin, r=0.047 and norepinephrine levels, r=0.185, ps>0.05, and, respectively, r=0.056, and r=−0.004. Also, changes (pre-post) in serotonin and norepinephrine levels are not inter-related, r=−0.004, p>0.05. 3.9. Suicidality Suicidal ideation was assessed using item nine from the CDI. None of the participants actually attempted suicide during the trial. Reports of suicidal ideation significantly decreased from pre to post-treatment, F(4, 72)=25.37, p<0.001, but there were no differences across groups, F(2, 72)=0.042, p>0.05. 3.10. Adverse events None of the participants reported major side-effects. Some participants reported mild side-effects, displayed in Table 3. Table 3. Adverse events by treatment group (Total N=88). Adverse events Group REBT/CBT (N=28) Medication (N=33) Medication+Group REBT/CBT (N=27) Headaches 1 2 3 Insomnia 1 1 0 Sleepiness 0 1 1 Low appetite 0 2 1 Weight gain 0 1 0 Fatigue 0 1 0 Dizziness 0 0 1 Sweats 0 0 1