Background
The number of children and adolescents diagnosed with bipolar disorder has increased in recent years [
1,
2]. A meta-analysis of 12 international epidemiological studies reported that the overall occurrence of pediatric bipolar disorder was 1.8% [
3], and in an epidemiological sample of more than 10,000 adolescents, the prevalence of bipolar disorder was reported to be 2.9% [
4].
There is a significant reduction in quality of life (QoL) in youth with bipolar disorder [
5‐
7], with patients experiencing social and attention problems, delinquent and aggressive behavior, and a poor ability to maintain stable relationships and perform successfully at work or school [
8‐
12]. Improving the QoL of patients with bipolar disorder is an important aspect of a successful treatment outcome. Although studies have examined the effect of treatment of mania on QoL in bipolar disorder [
13,
14], health-related QoL in bipolar depression has not been studied extensively.
A recent 8-week, double-blind, placebo-controlled study demonstrated that the combination of olanzapine and fluoxetine (OFC) was significantly more efficacious than placebo for the acute treatment of bipolar depression in children and adolescents (aged 10–17 years) [
15]. In this study [
15], the mean improvement in children’s depression rating scale-revised (CDRS-R) [
16] total score was significantly greater for OFC-treated patients than for placebo-treated patients starting at week 1 and for all subsequent visits up to week 8, and the rates of and times to response and remission were also significantly greater for OFC- than placebo-treated patients. The most common treatment-emergent adverse events in the OFC group in this study were somnolence, weight gain, and increased appetite. Based on this study, OFC was the first medication approved by the US Food and Drug Administration (USFDA) for the treatment of depressive episodes associated with bipolar I disorder in children and adolescents [
17].
Antipsychotic monotherapy treatment has been shown to be effective for treating children and adolescents with bipolar disorder [
18,
19] including improvement in manic symptoms in youth with bipolar disorder. However, corresponding improvements in QoL have been mixed. For example, in a 4-week study of manic youth (aged 10–17 years) with bipolar disorder, there were no significant differences between aripiprazole and placebo (at week 4) in QoL as measured by the change in total score on the Pediatric QoL Enjoyment and Satisfaction Questionnaire [
20]. However, another study of bipolar adolescents (mean age 15 years) with manic or mixed episodes reported that significant improvements in health-related QoL (child health questionnaire), especially in psychosocial domains, were observed after 28 days of treatment with quetiapine [
21]. These improvements were not significantly related either to changes in mania or depressive symptoms.
Olanzapine monotherapy is approved by the USFDA for the treatment of schizophrenia and manic or mixed episodes of bipolar I disorder in adolescents, based on a study of patients with schizophrenia [
22] and on a study in patients with bipolar I mania [
23]. Treatment with olanzapine was also effective on multiple domains of psychosocial functioning compared with placebo when examining health-related QoL in manic adolescents with bipolar disorder [
24]. Compared with the placebo group, patients in the olanzapine group showed significantly greater improvement in the psychosocial summary score from baseline to endpoint and in the mean change scores of behavior, family activities, and mental health subscales of the child health questionnaire-parental form 50 [
24].
The selective serotonin reuptake inhibitor (SSRI) fluoxetine is an antidepressant that is approved in children and adolescents for the treatment of major depressive disorder [
25,
26] and for the treatment of obsessive–compulsive disorder [
27]. Fluoxetine has been shown to improve global health, functioning, and QoL in depressed adolescents [
28]. However, there is a paucity of studies assessing changes in QoL in children or adolescents with bipolar depression treated with either an SSRI or an antipsychotic.
This paper focuses on data from a previously published three-country, multi-site, double-blind, placebo-controlled trial [
15] demonstrating the efficacy of OFC compared with placebo during 8 weeks of treatment of depressive episodes associated with bipolar I disorder in children and adolescents. The present analysis evaluates mental health outcomes from that study to determine whether OFC was superior to placebo in improving QoL.
Discussion
Results from the present analysis indicated significant impairment in quality of life (QoL) among patients in an acute episode of pediatric bipolar depression. Although this may not be particularly surprising, it serves as an important reminder of the vulnerability of this population and the need for treatment during this phase of the illness, which may sometimes be overlooked relative to the more dramatic manic phase. The mean baseline KINDL-R domain scores in the pediatric population with bipolar depression reported here (KINDL-R patient-rated scale range 32.3 [self-esteem] to 55.5 [family]; KINDL-R parent-rated scale range: 29.9 [self-esteem] to 55.5 [physical well-being]) were very low in comparison to those of a normative school-aged reference population (range 66.6 [self-esteem] to 84.0 [family]) [
32] but were also lower than those from another study evaluating a pediatric population diagnosed with bipolar disorder in any phase of the illness (range 44.9 [self-esteem] to 62.4 [emotional well-being]) [
6]. The baseline scores were also much lower than those reported in a previous study with a population of children and adolescents with epilepsy from the United Kingdom (range 59.1 [friends] to 81.7 [family] [
39]. This suggests that QoL in children/adolescents with bipolar depression is significantly impaired, a finding also supported by analyses of Freeman et al. [
6]. Relative to the normative data (and according to the SDs from the normative data [
6]), the population in the present study on average was at least 2 SDs worse than normal on their KINDL-R total score, emotional well-being and Friends subscale scores and was at least 1 or more SDs worse than normal on KINDL-R subscales of Self-esteem, Family, and School and on the Disease module.
With respect to treatment differences, although there were statistically significantly greater gains in quality of life in some QoL subscales for the OFC-treated patients relative to the placebo-treated patients, these differences were relatively modest in this 8-week study and varied somewhat by reporter (parent or child). Based on patients’ self-reports, there were greater improvements on the KINDL-R Self-esteem subscale score in the OFC-treated patients than the placebo-treated patients. Based on parents’ reports about their offspring, there were greater improvements on the KINDL-R subscale scores of Self-esteem, emotional Well-being, and Family. Across these few subscales, the difference in improvement between treatment groups was about 7 points. Although it is difficult to ascertain whether this difference between OFC and placebo is clinically significant, use of the rating anchors for the subscale items can provide some context for the findings. For instance, on the patient-rated Self-esteem subscale for the OFC-treated patients, the average scores were indicative of a change from feeling self-pride “never to seldom” at baseline to “seldom to sometimes” at endpoint, whereas the placebo-treated patients’ average results were indicative of a change from self-pride “never to seldom” at baseline to “seldom” at endpoint. Although subtle, this difference suggests clinically meaningful movement. However, because of the lack of adjustment for multiple comparison and as shown by the lack of statistical significance when controlling for CDRS-R total score, these findings were not statistically robust. Indeed, the effect sizes were small even in those subscales that showed significant improvement with treatment versus placebo. Nevertheless, visual comparison of the findings in this study relative to those of a normal sample (Fig.
1) also suggest that after 8 weeks of treatment, patients’ QoL scores were moving in the direction of “normal” but still below those of their healthy peers.
The KINDL-R questionnaire provides the ability to obtain a self-assessment in children and adolescents ranging from 8 to 12 years (Kid-KINDL), and 13 to 16 years (Kiddo-KINDL), and an external assessment of health-related QoL in children and adolescents ranging from 8 to 16 years (KINDL-R parent-rated). The parents are asked to complete the KINDL-R questionnaire with judgments from their own point of view of their children’s QoL. Because the children/adolescents and parents complete the questionnaires independently of one another, it is interesting to note the differences in perspectives. Patients tended to rate their QoL better at baseline than the parents did. However, parents almost universally rated better baseline to endpoint improvements than the patients, regardless of assigned treatment group. Previous studies also support the importance of obtaining the perspective of both the child and the parent when reporting on studies related to QoL in children [
40,
41]. Limitations in insight observed in children may be the result of experiencing an acute depressive episode but may also be a function of the patients’ age and developing cognitive capabilities. In addition, it is possible that parents may be more prone to the placebo effect and would tend to report their belief that the blinded study medication is helping. Given the subjective nature of the KINDL-R, self-reporting by this young patient group, in particular for adolescents, might be a more valid approach to measuring QoL than a parent report. Previous studies have reported that adults (aged 45–85 years) with bipolar disorder are dissatisfied with their QoL even when they are in a state of remission, and in patients with bipolar disorder and schizophrenia in remission, there was a negative association between insight and physical domain [
42]. In addition, in patients with depressive disorder, a high level of self-stigma was associated with poor QoL [
43]. Including parent-reported measures in studies of bipolar disorder has been shown to add value to studies of treatment outcome by complementing clinician report measures and representing the parent’s perspectives and providing a more comprehensive picture of the child’s functioning [
44]. Interestingly, a recent paper [
45] that examined QoL using KINDL-R in 530 healthy children in Germany found that the perception of QoL has increased in both children and parent reports over the past 10 years. The authors noted that the largest increases occurred in self-esteem, physical well-being, and family, and speculated that this may be due to changes in the social and environmental life of the children. The study also found that QoL decreases with increasing age especially in girls, which may be attributed to increasing pressure in school and declining leisure time.
Significant improvements were observed in terms of severity of depression, as assessed by the CDRS-R. Item analysis of the CDRS-R indicated significantly greater improvement for the OFC group than the placebo group on 7 of 17 items on the CDRS-R, including 4 items that overlap conceptually with domains assessed by the KINDL-R. These 4 items were self-esteem, difficulty having fun, impaired schoolwork, and social withdrawal. This might suggest that much of the CDRS-R total score improvement was due to items that are more socially oriented. This conceptual overlap could also explain why none of the KINDL-R scores were significantly different between the OFC and placebo groups when the CDRS-R total (change from baseline) was included in the baseline adjustment of the KINDL-R analyses. The change in CDRS-R total score was placed into the KINDL-R statistical model as an explanatory variable. Although not shown, the analyses showed that the CDRS-R had a strong relationship with the KINDL-R results. Nevertheless, OFC improved depression in these pediatric patients with bipolar depression as noted not only by the significant improvement in item 11 (depressed feelings) of the CDRS-R, but also by the significant improvement in the bipolar depression rating scale and the clinical global impressions scale-bipolar version severity of depression as noted in the primary publication [
15], although these 2 scales have not been validated in children and adolescents with bipolar depression. Of note, on the CDRS-R items (rated on a 0–5 scale) that were significantly improved for OFC compared to placebo, the between-group difference in mean change from baseline ranged from about 0.4–0.8 points. For example, for sleep disturbance the between-group difference in improvement was 0.8 points, with endpoint score being nearly 1 for OFC-treated patients (no difficulty or occasional difficulty sleeping) versus 2 at endpoint for placebo (frequent difficulty sleeping).
Few treatment options have proven to be effective for treating the depressive phase of bipolar disorder in children and adolescents. In addition to pharmacotherapy, it is important to consider the use of promising psychosocial interventions such as child- and family-focused cognitive behavioral therapy, dialectical behavioral therapy, interpersonal and social rhythm therapy, multifamily psychoeducation group psychotherapy, and family-focused treatment [
46]. Psychosocial interventions have yielded positive results in combination with pharmacotherapy and may enhance or help maintain improvements in QoL. For example, West et al. [
47] found that children and adolescents with bipolar disorder who were maintained on treatment and a child- and family-focused cognitive-behavioral therapy program for 3 years after the initial intervention showed significant long-term improvement in symptoms and psychosocial functioning relative to the control group that received medication and standard psychotherapy. Hlastala et al. [
48] found that a dozen adolescents with bipolar disorder who received medication and also participated in 16–18 sessions of Interpersonal and Social Rhythm Therapy over a period of 20 weeks also showed substantial improvement in global functioning as well as on measures of psychiatric symptoms. Fifty-eight adolescents with either bipolar I, II, or not otherwise specified were assigned to either family focused therapy and pharmacotherapy or enhanced care and pharmacotherapy for up to 2 years [
49]. Although recovery rates from the index episode and time to recurrence of depression were not different between groups, patients in the family focused therapy group showed faster recovery from baseline depressive symptoms, spent fewer weeks in depressive episodes and had a more favorable trajectory of depressive symptoms for 2 years.
The depressive phase in pediatric bipolar disorder is associated with various negative outcomes, such as suicidality, problem behaviors and hopelessness, and significant impairment in QoL [
50], thus highlighting the urgency for needed intervention during the depressive phase. Changes from baseline to week 8 in the CDRS-R total score were significantly greater for OFC-treated compared with placebo-treated patients and significant between-group differences were also seen starting from week 1 and all subsequent visits up to week 8 [
15]. Given the improvement in depressive symptoms [
15] and improvement in some aspects of the QoL as shown with the KINDL-R, this suggests that OFC may be a treatment option for children and adolescents with bipolar depression; however, these results must be balanced against the safety results. Safety findings in the primary study were consistent with those observed in adults treated with OFC or adolescents treated with olanzapine, with the exception of a greater increase in QTc interval that was observed in this study [
15]. Somnolence, weight gain, and increased appetite were the most common treatment-emergent adverse events reported in the OFC group, and weight gain was significantly greater for OFC- than placebo-treated patients [
15].
There are potential limitations to the present analyses that need to be considered. Assessment of QoL was a secondary objective of the study [
15], and findings were not adjusted for multiple comparisons. Therefore, the findings are not confirmatory and should be interpreted with caution. An active comparator was not used in this study so results should be interpreted accordingly. In addition, because the study did not include healthy controls, QoL scores have been shown relative to a pre-existing normative population assessed as part of the development and validation of the KINDL-R [
32]. Although the mean ages for the 2 populations align well, differences in QoL scores between the populations could be due at least in part to cultural or geographic differences. Also, because the data for the healthy population were published over a decade before the results of the current pediatric bipolar depression study, the current analysis may be less reflective of the true differences between the present pediatric bipolar sample and a healthy pediatric sample today. Finally, the duration of this study was 8 weeks, which is relatively short when assessing QoL. It is unknown whether a longer study might have resulted in greater or lesser differences between drug and placebo on patient- or parent-rated QoL. Because there is a scarcity of QoL studies in children and adolescents with bipolar depression, it is difficult to put the present findings into context or compare outcomes.
Authors’ contributions
DJW, MPD, JL, DND, and HCD participated in interpreting data, reviewing references, and drafting/revising the manuscript. JL additionally executed the statistical analysis. All authors read and approved the final manuscript.