Predictors of treatment nonresponse in bipolar children and adolescents with manic or mixed episodes.

Even though juvenile bipolar disorder (BD) is reported to be more treatment-resistant than adult BD, predictors of nonresponse are not well studied. The aim of this study was to address this issue in a naturalistic sample of bipolar children and adolescents with manic or mixed episodes treated under the condition of routine clinical practice. This study was comprised of 40 patients (19 females and 21 males; mean age, 14.2 years; SD = 3.3; range, 7-18) with a Diagnostic and Statistical Manual of Mental Disorders-fourth edition (DSM-IV) diagnosis of manic (n = 23) or mixed episodes (n = 17). The clinical characteristics of 20 patients considered to be treatment responders, according to the Clinical Global Impression-Improvement (CGI-I) scores, were compared to those of the 20 nonresponders. The effect of predictors on the probability of treatment nonresponse was analyzed using the multiple stepwise logistic regression, backward procedure. Demographic variables (mean age, gender ratio, socioeconomic status), as well as the inpatients-outpatients ratio (75% versus 65%), duration of the follow-up (10.5 +/- 2.5 months versus 9.6 +/- 3.2 months), index episode (manic versus mixed), and rates of pharmacologic hypomania did not differentiate the 2 groups. According to stepwise logistic regression, predictors of nonresponse were the presence of comorbidity with conduct disorder (odd ratio, 3.36; 95% CI, 2.20-4.52), attention deficit hyperactivity disorder (ADHD) (odd ratio, 2.30; CI, 1.24-3.26), and the baseline CGI Severity score (odd ratio, 2.31; CI, 1.33-3.29). It is relevant to point out that patient age at the onset of BD, and at the first visit, and comorbid anxiety disorders did not influence treatment response. Follow-up studies with a larger sample size with BD and/or externalizing disorders appropriately managed with different treatment options and/or combinations are warranted.