In adolescents with ADHD and SUD, those with co-occurring MDD used drugs and/or alcohol on significantly more days prior to treatment entry, as well as throughout and at the end of treatment compared to those without MDD. Depressed adolescents also had fewer days of abstinence and fewer negative urine drug screens during treatment despite having similar rates of treatment completion and adherence and comparable reduction in ADHD symptom severity compared to non-depressed adolescents. There was no difference, however, in reduction in days of non-nicotine substance use in those with or without MDD during treatment for SUD.
These findings are consistent with most previous studies that reported depression was associated with more severe substance use problems in clinical and community samples (Rohde et al., 1996
; Whitmore et al., 1997
; Riggs et al., 1995
). Our finding that adolescents who had co-occurring MDD but no reported therapy targeted to MDD had poorer SUD outcomes is consistent with findings in adolescents by Dobkin et al. (1998)
but not Crowley et al. (1998)
. However, most prior studies have not reported the association between MDD and SUD treatment outcomes in well-characterized samples of adolescents with MDD such as this one.
The presence of MDD was not associated with ADHD treatment response in this sample. Adolescents with and without MDD began the trial with no difference in severity of ADHD symptoms. Although those without MDD had a larger decrease in symptom severity during the initial weeks, this difference between groups leveled out and does not appear to be clinically meaningful since there was no difference between the groups in symptom improvement or final symptom severity at the end of the trial. Similar to our finding that the presence of MDD was not meaningfully related to ADHD outcomes, the presence of symptoms of depression or anxiety did not appear to impede response to atomoxetine for ADHD in children and adolescents treated with atomoxetine alone or in combination with fluoxetine (Kratochvil at al., 2005
Given the similarity in both ADHD symptom severity at treatment initiation and ADHD outcomes in both groups, it is also unlikely that ADHD symptoms or treatment response were responsible for the differences in non-nicotine substance use outcomes in those with and without MDD. ADHD was similarly unrelated to SUD outcomes in the controlled trial of fluoxetine for MDD in adolescents also receiving CBT for SUD where adolescents with or without ADHD had similar substance treatment as well as depression outcomes (Riggs et al., 2009
Depressed adolescents in this sample tended to be older, female, and fewer were court-mandated to substance treatment compared to those without MDD. Depression and mood disorders have been reported as more likely in substance using adolescent females than males (Grella et al., 2001
; Buckstein et al., 1992
; Deykin et al., 1992
Co-occurring MDD did not appear to impact CBT treatment adherence or study completion. Some prior studies similarly found no association between comorbid disorders and treatment attendance (Rivers et al., 2001
; Rowe et al., 2004
), although an association between internalizing symptoms and better treatment completion was found in an earlier study (Kaminer et al., 1992
). Since attendance at CBT visits and retention rates in the study did not differ in adolescents with and without MDD, adherence to the psychosocial intervention or retention also cannot account for the difference between non-nicotine substance use in the adolescents with and without MDD.
The current treatment literature does not provide a clear answer to whether providing targeted treatments for both MDD and SUD will improve SUD outcomes in adolescents (Riggs et al, 2007
; Cornelius et al., 2009
; Cornelius et al., 2010
; Deas et al., 2000). However, given the finding from this study that MDD is associated with greater severity of SUD and continued higher use throughout and at the end of SUD treatment and preliminary evidence that depression remission is associated with improved SUD outcomes (Riggs et al., 2007
), targeting both for treatment may be helpful. Studies in adults with MDD and SUD suggest that treating both may improve outcomes (Rao and Chen, 2008
). Receiving psychiatric services in addition to SUD treatment was associated with improved outcome (Ray et al., 2005
), and manualized behavioral therapies including CBT have been shown to decrease substance or alcohol use and/or depressive symptoms in adults (Carroll, 2004
; Maude-Griffin et al., 1998
; Brown et al., 1997
There may be alternate factors impacting SUD outcomes for adolescents with MDD, SUD, and ADHD, such as the severity of the disorders, additional burden of other disorders, psychosocial functioning, or quality of life. For example, number of days of heavy alcohol use was associated with lack of remission of depression in adolescents (Cornelius et al., 2009
). However, the association between SUD or MDD severity and outcomes was not evaluated in this sample.
Finally, we also evaluated whether the use of OROS-MPH or placebo might be associated with different outcomes in the adolescents with MDD or the adolescents without MDD. There was no difference in either days of non-nicotine substance use or ADHD symptom severity outcomes over time in either group when OROS-MPH and placebo were compared, consistent with the primary outcomes for the entire sample in the main study (Riggs et al., 2010). While there is some evidence for the efficacy of psychostimulants in reducing symptoms of depression (Candy et al., 2008
), the finding that substance use outcomes do not differ with its use in the adolescents with MDD is consistent with the prior studies that found psychopharmacology for depression was not associated with improvement in substance use, at least in the presence of CBT (Riggs et al., 2007
, Cornelius et al., 2009
, Cornelius et al., 2010
). Other possible benefits of treating ADHD with OROS-MPH, given the poor treatment completion and adherence and worse outcomes reported in adolescents with ADHD and SUD (Grella et al., 2001
, Rowe et al., 2004
, Wise et al., 2001
) were also not apparent.
Since pharmacological and psychosocial treatments specifically targeted to MDD were not allowed in this study, we could not evaluate the impact of combining treatments for MDD and SUD, although the effects of such combined treatment may have been masked by possible antidepressant effects of the SUD-focused CBT here as well. Since there was also no measure of depression symptom change during the trial, it is also not clear whether a difference in depression response or remission, even absent treatment targeted to depression, would have helped close the gap in drug use outcomes between those with and without MDD, as findings from a prior study suggest (Riggs et al., 2007
). The age of onset and timing of onset of MDD in relation to the SUD and measures of other concurrent psychiatric disorders were not available. Finally, with 38 patients with MDD, options for questions and analyses were limited. For example, a subgroup analysis for adolescents with MDD or estimating the group by time interaction would have been more reliable with a larger sample. Further validation is needed to provide support for the generalizability of these findings.
Outcomes with SUD treatments in adolescents remain modest, and co-occurring conditions are the rule rather than the exception. Preliminary results from this study, which includes thorough diagnostic assessment of MDD with the K-SADS-E, extend previous research by suggesting that the presence of co-occurring MDD in the context of ADHD and SUD is associated with greater severity of SUD and continued higher use in the absence of treatment specifically targeted to MDD. It is also associated with specific clinical characteristics (older age, female gender, fewer court referrals). Adequately powered trials evaluating treatments focused on MDD in the context of SUD and other frequently present, co-occurring psychiatric disorders are needed. In the interim, mechanisms for assessing and treating these disorders are essential in SUD treatment settings since their presence appears to interfere with maximizing SUD outcomes.