U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Findling R, Taylor-Zapata P, Anand R, et al.; Eunice Kennedy Shriver National Institute of Child Health and Human Development. A Randomized, Double-Blind, Placebo Controlled Study of the Efficacy of Lithium for the Treatment of Pediatric Mania Followed by an Open Label Long-Term Safety Period, Double-Blind, Placebo-Controlled Discontinuation Phase, and Open Label Restabilization Period (COLT2) [Internet]. Bethesda (MD): National Institute of Child Health and Human Development (US); 2015 Dec 1.

Cover of A Randomized, Double-Blind, Placebo Controlled Study of the Efficacy of Lithium for the Treatment of Pediatric Mania Followed by an Open Label Long-Term Safety Period, Double-Blind, Placebo-Controlled Discontinuation Phase, and Open Label Restabilization Period (COLT2)

A Randomized, Double-Blind, Placebo Controlled Study of the Efficacy of Lithium for the Treatment of Pediatric Mania Followed by an Open Label Long-Term Safety Period, Double-Blind, Placebo-Controlled Discontinuation Phase, and Open Label Restabilization Period (COLT2) [Internet].

Show details

11EFFICACY EVALUATION

As noted in Section 10.1.1, the rate of accrual of subjects to the study was much lower than expected. The total sample size of 81 represents 36% of the planned sample size of 225. Interim analyses showed that there was sufficient power to detect significant differences between treatment groups for the primary analysis in the Efficacy Phase.

11.1. Data Sets Analyzed

The safety population for a given study phase consists of all participants who received at least one dose of study medication during that phase. Safety participants were analyzed according to treatment actually received. All enrolled participants have received study medication and are included in the safety population.

All safety participants were evaluated for efficacy if they had at least one post-baseline efficacy assessment for a given phase. In the efficacy population, all participants were analyzed according to the treatment to which they were randomized, regardless of treatment actually received (intention-to-treat principle).

In addition, the Efficacy Phase and the Discontinuation Phase analyses were performed on the per protocol population, which includes the efficacy population without the following major protocol infractions:

  1. Drug non-compliance defined as missed > 40% of the medication doses in one week during the phase
  2. Prohibited medication taken during the phase
  3. Dose exceeded 40 mg/kg/day without a waiver (with the exception of participants who weighed less than 23 kg and may have received up to 900 mg/day)
  4. Received incorrect randomized study medication

As discussed in Section 10.2, a second definition of major protocol deviations, with less stringent requirements, was also analyzed. This new definition included:

  1. Drug non-compliance defined as missed > 40% of the medication doses in two consecutive weeks during the phase
  2. Prohibited medication taken during the phase
  3. Dose exceeded 50 mg/kg/day without a waiver (with the exception of participants who weighed less than 23 kg who may receive up to 900 mg/day)
  4. Received incorrect randomized study medication

11.2. Analysis of Efficacy

The following assessment scales and subscales were analyzed as a part of primary, secondary, tertiary, and descriptive efficacy endpoints as described in the efficacy analysis subsections for each of the 4 phases; detailed definitions are presented in the study protocol (Appendix 16.1.1):

  • ADHD Rating Scale-IV:
    • Total score
    • Inattention subscale
    • Hyperactivity-Impulsivity subscale
  • Brief Psychiatric Rating Scale-For Children (BPRS-C):
    • Total score
    • Behavior Problems subscale
    • Depression subscale
    • Thinking Disturbance subscale
    • Psychomotor excitation subscale
    • Withdrawal subscale
    • Anxiety subscale
    • Organicity subscale
  • Caregiver Strain Questionnaire (CSQ)
    • Total Score
    • Objective Strain
    • Subjective Strain
    • Internalized Subjective Strain
    • Externalized Subjective Strain
  • Child Mania Rating Scale-For Parents (CMRS-P): Total score
  • Children Depression Rating Scale-Revised (CDRS-R): Total score
  • Children’s Global Assessment Scale (CGAS): Single score
  • Clinical Global Impressions Scale-Severity (CGI-S):Single scores for each of three domains:
    • Mania
    • Depression
    • Overall illness
  • Clinical Global Impressions Scale-Improvement (CGI-I): Single scores for each of three domains:
    • Mania
    • Depression
    • Overall illness
  • Family Environmental Scale (FES): Ten domain Subscales:
    • Cohesion subscale
    • Expressiveness subscale
    • Conflict subscale
    • Independence subscale
    • Achievement subscale
    • Intellectual-Cultural subscale
    • Active-Recreational subscale
    • Moral-Religious subscale
    • Organization subscale
    • Control subscale
  • Parent General Behavior Inventory-10 Item Mania Scale (PGBI-10M): Total score
  • Irritability, Depression and Anxiety (IDA): Outward irritability score
  • Nisonger Child Behavior Rating Form (NCBRF-TIQ)
    • Positive Social subscale
    • Conduct Problem subscale
    • Hyperactive subscale
    • Overly Sensitive subscale
    • Oppositional Behavior subscale
    • Inattentive subscale
    • Withdrawn/Dysphoric subscale
  • Pediatric Anxiety Rating Scale (PARS):
    • Total score: sum of responses to all 7 scale items
    • Subscale score
  • Social Adjustment Inventory for Children and Adolescents (SA/CA)
    • Academic reading/English
    • Academic arithmetic
    • School behavior problems subscale;
    • Spare time activities subscale;
    • Spare time problems subscale;
    • Activity with peers subscale;
    • Peer problems subscale;
    • Boy-girl relationships (for participants ≤12 years old) subscale;
    • Problems with opposite sex (for participants ≤12 years old) subscale;
    • Sibling relationships subscale;
    • Problems with sibling(s) subscale;
    • Relationship with mother subscale;
    • Relationship with father subscale;
    • Problems with parents subscale
  • Young Mania Rating Scale (YMRS):
    • Summary: sum of responses to all 11 inventory items as rated by clinician
    • Parent: sum of responses to all 11 inventory items as rated by parent
    • Child: sum of responses to all 11 inventory items as rated by child

11.2.1. Efficacy Phase

11.2.1.1. Demographics and Other Baseline Characteristics

In the Efficacy Phase, participants were randomized to lithium and placebo at a ratio of 2:1 respectively. The actual distribution of treatment among the participants in the Efficacy Phase efficacy population was 1.9:1 (lithium : placebo). Fifty three (65%) participants were randomized to the lithium treatment arm and 28 (35%) participants to the placebo treatment arm.

Baseline demographics and disease characteristics are summarized for the efficacy population of the Efficacy Phase in Table 11-1 and Tables 14.1.3.1, 14.1.4, 14.1.5.1 and 14.1.5.2. These tables include information on participant demographics, disease summary, and medical and surgical history. The baseline visit for the Efficacy Phase is defined as the first visit (at randomization) of the Efficacy Phase.

Table 11-1. Efficacy Phase Demographic Characteristics (Efficacy Population).

Table 11-1

Efficacy Phase Demographic Characteristics (Efficacy Population).

There were 30/81 (37%) participants that were <10 years of age at randomization to the Efficacy Phase, 17/53 (32%) in the lithium group and 13/28 (46%) in the placebo group. The median age of participants treated with lithium was 11.3 years (range 7.3-17.3 years) and for participants treated with placebo it was 10.6 years (range 7.0 – 16.7 years) at randomization to the Efficacy Phase (at baseline), Kruskal-Wallis p=0.68. There is no difference in gender (χ2 p=0.30), ethnicity (χ2 p=0.69) or race (χ2 p=0.75) distributions between treatment arms.

Length of time in the Efficacy Phase is shown in Table 11-2 for all participants entering the phase by randomized treatment arm. Median time in the phase was 7.9 weeks (range 1.3-9.6) in both treatment groups for all participants. Length of time in the study phase was also 7.9 weeks for each age stratum.

Table 11-2. Length of Time in the Efficacy Phase.

Table 11-2

Length of Time in the Efficacy Phase.

The disease history of participants is summarized in Table 14.1.4. All 81 participants had a definite physician diagnosis for manic or mixed episodes of Bipolar I disorder. Thirteen (16%) participants were diagnosed within a year of enrollment (n=9, 17% lithium and n=4, 14% placebo). Ten (12%) participants displayed psychotic features (n=6, 11% lithium and n=4, 14% placebo), Twenty (25%) participants reported rapid cycling (n=13, 25% lithium and n=7, 25% placebo) and 2 (3%) participants showed a seasonal pattern (both lithium-treated participants). Episodes were classified as manic in 41 (51%) of the 81participants of which, 23 (43%) were lithium treated and 18 (64%) were treated with placebo. Episodes were classified as mixed in 40 (49%) of the 81 participants of which, 30 (57%) were lithium treated and 10 (36%) were placebo treated. History of Axis 1 psychiatric disorders at the beginning of Screening Phase is listed for the Efficacy Phase population in Table 11-3; fifty two (64%) participants were diagnosed with ADHD.

Table 11-3. Axis 1 Psychiatric Disorders (Efficacy Phase: Efficacy Population).

Table 11-3

Axis 1 Psychiatric Disorders (Efficacy Phase: Efficacy Population).

Medical history was collected for all participants during screening. A summary of the medical history by MedDRA system organ class and preferred term for the 81 participants randomized to the Efficacy Phase is presented in Table 14.1.5.1. Twenty-four (30%) participants had a history of renal or urinary disorders. This included 15/53 (28%) lithium participants and 9/28 (32%) placebo participants. Other body systems with medical conditions in this population included: respiratory disorders in 38/81 (47%) participants (47% of lithium and 46% of placebo); infections and infestations in 34/81 (42%) participants (34% of lithium and 57% of placebo); nervous system disorders in 27/81 (33%) participants (36% of lithium and 29% of placebo); and skin disorders in 24/81 (30%) participants (28% of lithium and 32% of placebo).

Twenty-two (22/81 [27%] Efficacy Phase participants reported a history of surgery, 14 (26%) lithium participants and 8 (29%) placebo participants. Most common procedure was tonsillectomy in 6 participants. Prior surgical history is summarized in Table 14.1.5.2.

11.2.1.2. Measurement of Treatment Adherence

Participants were provided dosing diaries to record their lithium pill intake and were asked to return the dosing diaries and study medication at each visit. Adherence was assessed by comparing the actual number of pills returned and the expected number of pills returned.

Overall compliance was assessed by averaging the weekly visit compliance rates, the overall median compliance rate was 95% with a range from 39% to 106%, see Table 11-4 below.

Table 11-4. Overall Treatment Adherence During the Efficacy Phase (Efficacy Population).

Table 11-4

Overall Treatment Adherence During the Efficacy Phase (Efficacy Population).

In addition, medication compliance was assessed by the review of the lithium trough serum levels. These are shown individually in the Participant Profiles (Section 14.4). A scatter plot of all lithium trough levels by Efficacy Phase randomized treatment group is shown below.

Figure 11-1. Lithium Levels by Treatment Over Time During the Efficacy Phase (Efficacy Population).

Figure 11-1

Lithium Levels by Treatment Over Time During the Efficacy Phase (Efficacy Population).

11.2.1.3. Efficacy Results

11.2.1.3.1. Efficacy Phase: Primary Endpoint

The primary efficacy endpoint for the Efficacy Phase (and for the study) is the change from baseline to the end of phase in the YMRS Summary score, based upon LOCF values.

Results are summarized in Table 11-5A and B below and in Table 14.2.1.1.1. Baseline summary YMRS scores were similar in the 2 treatment groups: mean of 29.5 (SD 5.6) in lithium treated participants and mean of 30.0 (SD 6.0) in placebo treated participants. The change from baseline to End-of-Phase/LOCF was a decline of 11.7 (SD 11.4) points in the lithium group and a decline of 7.8 (SD 9.8) in the placebo group. One participant [▬] was seen at Day 1, Week 1, Week 2 and would not return until for a final visit until Week 14. This participant was early terminated due protocol non-compliance and the data from the Week 14 visit is not included in these analyses.

Table 11-5A. Efficacy Phase - Primary Efficacy Endpoint by Efficacy Phase, Randomized Treatment Group.

Table 11-5A

Efficacy Phase - Primary Efficacy Endpoint by Efficacy Phase, Randomized Treatment Group.

Table 11-5B. Efficacy Phase - Least Squares Means for the Primary Efficacy Endpoint.

Table 11-5B

Efficacy Phase - Least Squares Means for the Primary Efficacy Endpoint.

Linear regression analysis assessed the change from baseline to end of phase score. The change score is the dependent variable, baseline score is included as a covariate, and age strata, gender, weight strata (< 30 kg ≥ 30 kg), site, and treatment group are factors in the model. The Least Squares (LS) Mean from the regression model was −12.9 (SE 3.1) points in the lithium group and −7.3 (SE 3.1) points in the placebo group. The overall decline in YMRS from baseline was 5.5 (SE 2.5) points more in the lithium group than the placebo group. The p-value evaluating the treatment difference is significant at p=0.031 after adjusting for baseline score, age strata, gender, weight strata, and site.

The adjusted standardized effect size (Cohen’s d) and corresponding 95% confidence interval, adjusting for baseline factors in the primary efficacy analysis, was 0.53 (0.06, 0.99). The unadjusted standardized effect size was 0.37 (−0.10, 0.83).

A sensitivity analysis was performed on the change in YMRS summary score from baseline to the summary YMRS score at each visit using a mixed effects repeated measures regression model with no LOCF. This longitudinal multivariate analysis used all available YMRS scores at each time point. Fixed effects included baseline YMRS score, treatment group, visit and age strata and treatment group by visit interaction; random effects included site. An unstructured covariance matrix was used to model the repeated measures within participants.

The results of this analysis found that the overall treatment effect was not significant (p-value 0.435); however, the visit by treatment interaction is significant at p=0.005. An examination of the change from baseline in summary YMRS score at each visit showed significant treatment differences at week 6 (n=62, p=0.012) and week 8 (n=58, p<0.001). Note that Weeks 5 and 7 were changed to a phone call visit in protocol amendment 5 and YMRS score was no longer collected at these visits resulting in reduced sample size (n=29 at week 5 and n=28 at week 7). Figure 11-2 shows that the differences in mean change in YMRS score from baseline between treatment groups begin to appear at week 4. Also, the treatment effect is much stronger in the sub-group of participants that completed the study (LS mean difference −7.6; p<0.001) compared to the analysis using LOCF (LS mean difference −5.5; p=0.031).

Figure 11-2. Efficacy Phase: Change from Baseline in YMRS Score During the Efficacy Phase, by Efficacy Phase Randomized Treatment Group (with LOCF Shown).

Figure 11-2

Efficacy Phase: Change from Baseline in YMRS Score During the Efficacy Phase, by Efficacy Phase Randomized Treatment Group (with LOCF Shown).

A second sensitivity analysis of the primary endpoint was performed on the per-protocol population (see Section 11.1) .using linear regression analysis to assess the change from baseline to end of phase score; see Table 14.2.1.1.2 and Figures 14.2.1.1.3 and 14.2.1.1.4. The change score is the dependent variable, baseline score is included as a covariate, and age strata, gender, weight strata, site, and treatment group are factors in the model.

The results of analysis of the primary endpoint on the per-protocol population showed a mean decrease in YMRS score of 10.9 (SD 12.0) points for participants in the lithium group and a decrease of 7.7 (SD 10.4) points for participants in the placebo group using LOCF methods. Due to the decreased sample size, only 63% of the original population is in the per protocol population, this difference is not statistically significant (p=0.295).

A second revised per protocol population was developed, see Section 11.1. This population included 90% of the original population, excluding only participants with prohibited medications, who received study dose >50 mg/kg/day without a waiver and who were non-compliant for ≥2 consecutive visits. The groups were compared using linear regression analysis described above. The mean (SD) change in YMRS scores from baseline to end-of-phase was −11.4 (11.6) points in the lithium group and −6.6 (9.9) points in the placebo group, p=0.021. See Table 14.2.1.1-3 and Figures 14.2.1.1.5, 14.2.1.1.6 and Figure 11-3 below.

Figure 11-3. Efficacy Phase: Change from Baseline in YMRS Score During the Efficacy Phase by Efficacy Phase Randomized Treatment Group (with LOCF Shown) for the Per-Protocol Population and Revised Per-Protocol Population.

Figure 11-3

Efficacy Phase: Change from Baseline in YMRS Score During the Efficacy Phase by Efficacy Phase Randomized Treatment Group (with LOCF Shown) for the Per-Protocol Population and Revised Per-Protocol Population.

11.2.1.3.2. Efficacy Phase: Secondary Endpoints

Continuous secondary endpoints were analyzed using the same method as the primary endpoint: the change score is the dependent variable, baseline score is a covariate, and age strata, gender, weight strata, site, and treatment group are factors. Endpoints analyzed were YMRS parent score; YMRS child score; Children Depression Rating Scale-Revised (CDRS-R); and Children’s Global Assessment Scale (CGAS). These results are summarized in Table 11-6A below. Tables 14.2.1.2.2, 14.2.1.2.3 and 14.2.1.2.4 include for each assessment scale: N, mean, standard deviation, median, minimum value and maximum value. These results are presented for the baseline visit, each follow-up visit, the LOCF visit, and the change score from baseline to LOCF visit..

Table 11-6A. Efficacy Phase: Secondary Efficacy Endpoints by Efficacy Phase randomized treatment group.

Table 11-6A

Efficacy Phase: Secondary Efficacy Endpoints by Efficacy Phase randomized treatment group.

The parents of children treated with lithium had reported a mean (SD) of 10.9 (SD 12.0) point decline in YMRS score while the parents of children treated with placebo reported a 7.7 (SD 10.3) point decline. The p-value was 0.052 after adjusting for baseline score, age strata, gender, weight strata and site. These results were in the same direction as the primary endpoint YMRS summary score.

The child self report of YMRS had smaller magnitude but in the same direction as the parent and summary scores. The children in the lithium group reported a 7.0 (SD 11.3) point decline while the children in the placebo group reported a 2.6 (SD 11.8) point decline, p-value of 0.277.

The results of the Children Depression Rating Scale-Revised (CDRS-R) showed similar changes in the 2 treatment groups: 5.5 (SD 12.2) point decrease in the lithium group and 6.8 (SD 8.5) point decrease in the placebo group, p-value not significant.

The Children’s Global Assessment Scale (CGAS) also showed no significant differences between treatment arms: mean increase of 9.5 (SD 13.8) in the lithium arm and increase of 8.5 (SD 12.1) points in the placebo arm.

Clinical Global Impressions Scales (CGI-S, CGI-I) were measured on a Likert scale and were analyzed using logistic regression model. Independent factors in the logistic regression model were age strata, gender, weight strata, and treatment group. Success and failures were coded as follows:

CGI-Severity (Mania, Depression and Overall Illness):

  • Success: Baseline = 1 or 2 and less or same severity level at LOCF
    Baseline = 3 and less severe at LOCF
    Baseline = 4, 5, 6 and LOCF=1, 2, 3 and at least a 2 point decrease
  • Failure: Otherwise

CGI-Improvement (Mania, Depression and Overall Illness):

  • Success: Week 8/ET=1,2
  • Failure: Week 8/ET≥3

Clinical Global Impressions Scales (CGI-S, CGI-I) results are presented in Table 11-6B and in Table 14.2.1.2.5. Severity outcomes were not significantly different between the treatment arms, however: success rates (see above) for mania were 11% higher in the lithium arm, success rate for depression were 16% higher in the lithium arm and success rates for overall illness were 19% higher in the lithium arm. The success rate (see above) for the improvement scores were significantly higher in the lithium for mania (51% in the lithium arm vs. 29% in the placebo arm, p=0.044) and for overall illness (47% in the lithium arm vs. 21% in the placebo arm, p=0.026). The depression improvement was not statistically significant, but the lithium arm had a 40% success rate vs. 29% in the placebo arm.

Table 11-6B. Efficacy Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement by Efficacy Phase randomized treatment group.

Table 11-6B

Efficacy Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement by Efficacy Phase randomized treatment group.

Three categorical variables were identified as additional secondary efficacy parameters: Response status, Remission status, YMRS (Summary) percentage improvement category. The Cochran-Mantel-Haenszel (CMH) test was performed to determine if there was a significant difference between the treatment groups for each of these endpoints.

Response status categories were defined as:

  • Response: YMRS (Summary) reduction ≥ 50% and CGI-I (Overall illness) = 1 or 2 at LOCF
  • Partial response: YMRS (Summary) reduction 25-49% and CGI-I (Overall illness) ≤ 3 OR YMRS (Summary) reduction ≥ 50% and CGI-I (Overall illness) = 3 at LOCF
  • Non-response: YMRS (Summary) reduction < 25% or CGI-I (Overall illness) ≥ 4 at LOCF

Remission status was defined as:

  • Remission: YMRS (Summary) ≤12 and CGI-S (Overall illness) ≤ 2 at LOCF
  • No remission: YMRS (Summary) > 12 or CGI-S (Overall illness) > 2 at LOCF

YMRS percentage improvement categories were defined as:

  • YMRS (Summary) reduction ≥ 50% at LOCF
  • YMRS (Summary) reduction = 25-49% at LOCF
  • YMRS (Summary) reduction < 25% at LOCF

Table 11-6C and 14.2.1.2.1 summarize the results of response, remission and improvement. While none of these endpoints were significantly different between the treatment arms, the lithium arm had more favorable rates in all categories with 32% complete response, 26% remission and 38% with reduction in YMRS ≥50% compared to the placebo arm with 21% complete response, 14% remission and 29% reduction in YMRS ≥50%.

Table 11-6C. Efficacy Phase: Secondary Efficacy Endpoints - Response, Remission and Improvement by Efficacy Phase randomized treatment group.

Table 11-6C

Efficacy Phase: Secondary Efficacy Endpoints - Response, Remission and Improvement by Efficacy Phase randomized treatment group.

Discontinuation due to any cause during the Efficacy Phase and discontinuation due to lack of efficacy during the Efficacy Phase were also analyzed with the Cochran-Mantel-Haenszel (CMH) test to determine if there was a significant difference between the treatment groups. These results are reported in Table 11-6D and Table 14.2.1.2.1. No differences were noted between treatment groups in discontinuation for any cause or discontinuation due to lack of efficacy.

Table 11-6D. Efficacy Phase: Secondary Efficacy Endpoints - Discontinuation by Efficacy Phase randomized treatment group.

Table 11-6D

Efficacy Phase: Secondary Efficacy Endpoints - Discontinuation by Efficacy Phase randomized treatment group.

Additionally, Cox proportional hazards regression models were used to analyze time from first dose of study drug to: (1) discontinuation for any reason; and (2) discontinuation due to lack of efficacy. Participants who did not discontinue during the Efficacy Phase were censored at the time of the Week 8 visit for the analysis of time to discontinuation for any reason. Participants who do not discontinue during the Efficacy Phase due to lack of efficacy were censored at the time of the Week 8 visit or at the time of discontinuation for reasons other than lack of efficacy, for the analysis of time to discontinuation for lack of efficacy. Independent factor in the Cox proportional hazards regression models was treatment group. These results are summarized in Tables 14.2.1.2.6 and 14.2.1.2.7.

Figure 11-4 and Figure 14.2.1.2.1 show the Kaplan-Meier estimates for time to discontinuation for any reason. Thirty percent of the lithium treated participants terminated during the Efficacy Phase vs. 25% of the placebo treated participants, this difference was not statistically significant (Cox regression HR=1.25, compared to placebo, p=0.626).

Figure 11-4. Time to Discontinuation for Any Reason During the Efficacy Phase by Efficacy Phase Randomized Treatment Group.

Figure 11-4

Time to Discontinuation for Any Reason During the Efficacy Phase by Efficacy Phase Randomized Treatment Group.

Only four participants discontinued the efficacy phase due to lack of efficacy, 3 (6%) lithium participants and 1 (4%) placebo participant. These results were not significantly different between treatment groups; see Table 11-6D and Table 14.2.1.2.7.

11.2.1.3.3. Efficacy Phase: Tertiary Endpoints

Efficacy Phase tertiary endpoints include results from the Parent General Behavior Inventory-10 Item Mania Scale (PGBI-10M), Child Mania Rating Scale for Parents (CMRS-P), ADHD Rating Scale-IV, Brief Psychiatric Rating Scale for Children (BPRS-C) and Pediatric Anxiety Rating Scale (PARS). The change from baseline to the end of phase scores, based upon LOCF values was evaluated for each measure. Linear regression analysis assessed the change from baseline to end of phase score. The change score is the dependent variable, baseline score is included as a covariate, and age strata, gender, weight strata, site, and treatment group are factors.

Results from the analyses of the tertiary efficacy endpoints from the Efficacy Phase are summarized in Table 11-7 below and Tables 14.2.1.3.1, 14.2.1.3.2, 14.2.1.3.3, 14.2.1.3.4 and 14.2.1.3.5. The Parent General Behavior Inventory-10 Item Mania Scale showed a significant difference between the lithium treated participants and the placebo treated participants. Lithium participants had a mean decrease from baseline score of 7.7 (SD 9.1) while the placebo participants had a mean decrease of 2.1 (SD 8.0) points, p=0.039. In general, the remaining tertiary endpoints while not significant, showed changes in the appropriate directions.

Table 11-7. Efficacy Phase: Tertiary Efficacy Endpoints by Efficacy Phase Randomized Treatment Group.

Table 11-7

Efficacy Phase: Tertiary Efficacy Endpoints by Efficacy Phase Randomized Treatment Group.

11.2.1.3.4. Efficacy Phase: Descriptive Endpoints

Descriptive efficacy endpoints for the Efficacy phase are summarized with descriptive statistics (N, mean, standard deviation, median, minimum and maximum) for baseline, week 8/end-of-phase/LOCF and change from baseline to week 8/end-of-phase/LOCF scores. These measures include Caregiver Strain Questionnaire (CSQ), Family Environmental Scale (FES), Irritability, Depression and Anxiety (IDA), Nisonger Child Behavior Rating Form (NCBRF-TIQ) and Social Adjustment Inventory for Children and Adolescents (SA/CA).

Results from the analyses of the descriptive efficacy endpoints from the Efficacy Phase are summarized in Tables 14.2.1.4.1 through 14.2.1.4.5. The descriptive endpoints were only completed on a fraction of the participants due to Amendment 5 protocol changes (Section 9.9.1.5) to reduce burden on the participants and increase accrual. The results were unremarkable but in general, showed changes in the appropriate directions.

11.2.1.4. Tabulation of Individual Response Data

Tabulations of individual response data can be found in Appendix 16.2.

11.2.1.5. By-Participants Displays

By-participant displays (i.e. “participant profiles”) are presented for the Efficacy Phase in Section 14.4. These profiles include randomized treatment, study drug dose in mg, study drug dose in mg/kg, lithium level, % compliance, YMRS score, CGI overall illness score, creatinine clearance, TSH, adverse events and serious adverse events temporally over the course of the eight week Efficacy Phase.

11.2.1.6. Conclusions – Acute Efficacy Phase (Efficacy)

The data collected in this phase, when considered in their entirety (primary, secondary, and tertiary endpoints), provide consistent evidence that lithium has acute efficacy in the treatment of pediatric participants with acute manic or mixed episodes. The adjusted standardized effect size (Cohen’s d) and corresponding 95% confidence interval was 0.53 (0.06, 0.99). This is similar to the effect size noted for lithium in adults with bipolar illness.

Of note, the responses noted herein are generally specific for manic symptoms. These data suggest that lithium has specific anti-manic effects rather than broader psychotropic effects in this patient population. It is possible that lithium has other acute psychotropic effects. However, due to sample size considerations, these other findings might not have been detectable. Regardless, these data suggest that lithium’s acute psychotropic effects in this population are most robust on symptoms of mania.

11.2.2. Long-Term Effectiveness Phase

Participants eligible to return to the study for the Long-Term Effectiveness (LTE) Phase included lithium treated participants from the Efficacy Phase that were determined to be responders/partial responders and placebo treated participants that were partial responders/non-responders at the end of the Efficacy Phase; they were defined as follows –

A responder was a participant with a reduction ≥50% in YMRS score and a CGI-I of 1-very much improved or 2-much improved.

A partial responder was a participant that had a 25-49% reduction in YMRS score and a CGI-I of ≤3 (very much improved, much improved or minimally improved) or YMRS reduction ≥ 50% and CGI-I=3 (minimally improved).

A non-responder had a <25% reduction in baseline YMRS or a CGI-I ≥4 (no change, minimally worse, much worse or very much worse).

Forty-four participants were enrolled in the Long-Term Effectiveness Phase from 9 sites. Two of these participants [▬] never returned for any LTE Phase visits. These participants were early terminated from the study (termed lost to follow-up) and are not included in the efficacy analysis population for the LTE Phase, leaving 42 participants analyzed in the LTE Phase, efficacy population. At entry into the LTE Phase, 29 (69%) of the participants were already receiving a therapeutic lithium dose and 13 (31%) participants were receiving placebo and tapered to lithium during the LTE Phase. The randomization in the Efficacy Phase was 2:1 (lithium : placebo), the distribution of treatment in the participants in the LTE Phase efficacy population was 2.2:1 (lithium : placebo). Efficacy phase treatment arm was similarly distributed between the two age strata (Table 11-8).

Table 11-8. Long-Term Effectiveness Phase Demographic Characteristics (Efficacy Population).

Table 11-8

Long-Term Effectiveness Phase Demographic Characteristics (Efficacy Population).

11.2.2.1. Demographics and Other Baseline Characteristics

Baseline demographics and disease characteristics are summarized for the Efficacy Population of the LTE Phase in Table 11-8 and Table 14.1.3.2. LTE Phase participants had a median age of 10.3 years (range 7.2-17.2 years) at enrollment to the LTE Phase, with 28 (67%) in the 7-11 years age stratum and 14 (33%) in the 12-17 years age stratum. Fifty two percent were female and 48% were male; 14% were of Hispanic ethnicity; 64% reported white race, 26% were African-American, 2% were Asian, 2% were of multiple race and race was not reported in 1% of the participants.

Length of time in the Long-Term Effectiveness Phase is shown in Table 11-9 for participants in the efficacy population by LTE phase age strata, Efficacy Phase treatment group and by treatment group/age strata. Median time in the phase was 23.6 weeks, range 2.1-25.9 weeks for all participants in the efficacy population. The older age stratum (children 12-17 years at the beginning of the phase) remained in the phase for shorter periods, median of 14.6 weeks (range 4.6-25.4) compared to a median of 23.7 weeks (range 2.1-25.9) for children 7-11 years of age. There was little difference in time in the phase by Efficacy Phase treatment arm: 23.7 weeks (range 2.1-25.4) for the lithium group and 23.6 (range 2.3-25.9) for the placebo group.

Table 11-9. Length of Time in the Long-Term Effectiveness Phase (Efficacy Population).

Table 11-9

Length of Time in the Long-Term Effectiveness Phase (Efficacy Population).

11.2.2.2. Measurement of Treatment Adherence

Overall adherence during the Long-Term Effectiveness Phase is reported in Table 11-10 below. Adherence was assessed by averaging the weekly visit adherence rates. The overall median adherence rate was 93% with a range from 50% to 102%.

Table 11-10. Overall Treatment Adherence during the Long-Term Effectiveness Phase (Efficacy Population).

Table 11-10

Overall Treatment Adherence during the Long-Term Effectiveness Phase (Efficacy Population).

In addition, medication adherence was assessed by the review of the lithium trough serum levels. These are shown individually in the Participant Profiles (Section 14.4). A scatter plot of all lithium levels is shown in Figure 11-5 below.

Figure 11-5. Lithium Levels Over Time During the Long-Term Effectiveness Phase (Efficacy Population).

Figure 11-5

Lithium Levels Over Time During the Long-Term Effectiveness Phase (Efficacy Population).

11.2.2.3. Efficacy Results

11.2.2.3.1. Long-Term Effectiveness Phase: Primary Endpoint

The primary efficacy endpoint for the LTE Phase is the change from baseline to the end of phase in the YMRS Summary score, based upon LOCF values. Descriptive summaries are provided for the Efficacy population by both Efficacy Phase randomized treatment group in Table 14.2.2.1B and by LTE Phase age strata in Table 14.2.2.1A.

Primary endpoint results are summarized in Table 14.2.2.1A. The mean baseline score, measured at the last (Week 8) visit of the Efficacy Phase, was 14.9 (SD 9.0). Mean week 24/LOCF summary YMRS was 8.8 (SD 8.8) with the mean change from baseline to week 24/LOCF of −6.1 (SD 11.2). Participants in the 7-11 age stratum had a mean change score of −6.9 (SD 12.8) and the participants in the 12-17 age stratum had a mean change score of −4.5 (SD 7.3).

As the criteria for enrollment to the LTE Phase was different for participants randomized to lithium vs. participants randomized to placebo, Table 11-11: Long-Term Effectiveness Phase: Primary Efficacy Endpoint by Efficacy Phase randomized treatment group and Table 14.2.2.1B present the summary YMRS by Efficacy Phase randomized treatment group. Participants randomized to lithium and eligible to enter the LTE Phase (responders or partial responders in the Efficacy Phase) had a mean baseline summary YMRS of 10.6 (SD 6.4), a week 24/LOCF YMRS of 9.0 (SD 8.6) and a change score of −1.6 (SD 8.4). Participants randomized to placebo treatment during the Efficacy Phase and eligible for enrollment in the LTE Phase (partial responders or non-responders during the Efficacy Phase) had a mean baseline summary YMRS of 24.6 (SD 5.8), a week 24/LOCF mean YMRS of 8.5 (SD 9.5) and a change from baseline to week 24/LOCF of −16.1 (SD 10.6). Placebo non-responders and partial responders showed improvement when tapered to lithium treatment.

Table 11-11. Long-Term Effectiveness Phase: Primary Efficacy Endpoint by Efficacy Phase randomized treatment group.

Table 11-11

Long-Term Effectiveness Phase: Primary Efficacy Endpoint by Efficacy Phase randomized treatment group.

Figure 11-6 and Figure 14.2.2.1.1 show each weekly YMRS summary score by randomized Efficacy Phase treatment arm. This longitudinal analysis uses all available YMRS scores at each time point.

Figure 11-6. Long-Term Effectiveness Phase: YMRS Score Over Time by Efficacy Phase Randomized Treatment Group (with LOCF Shown).

Figure 11-6

Long-Term Effectiveness Phase: YMRS Score Over Time by Efficacy Phase Randomized Treatment Group (with LOCF Shown).

The placebo participants enter the phase with a higher YMRS score (mean of 24.6, SD of 5.8 for placebo and mean of 10.6, SD 6.4 for lithium). As the placebo participants are tapered to lithium, their YMRS scores decrease and at the 10 week visit, lithium and placebo treated participants had similar YMRS scores, with a mean of 9.6 (SD 9.4) and 9.4 (SD 7.7) respectively. LOCF was also similar for the two groups with a mean of 9.0 (SD 8.6) for lithium participants and 8.5 (SD 9.5) for placebo participants.

The baseline visit for the LTE Phase is the last Efficacy Phase visit. The change in summary YMRS score from the baseline visit to each visit during the LTE Phase is shown in Figure 11-7 and Figure 14.2.2.1.2. Eight (62%) of the 13 participants originally treated with placebo completed all 24 weeks of the LTE Phase and had a mean decrease in YMRS score of 21.6 (SD 6.9) points. Twenty-nine participants entered the LTE Phase already receiving lithium and 15 (52%) of these participants completed all 24 weeks. These 15 participants had a mean decrease in YMRS of 3.7 (SD 7.0) from baseline score.

Figure 11-7. Long-Term Effectiveness Phase: Change from Baseline in YMRS Score Over Time by Efficacy Phase Randomized Treatment Group (with LOCF Shown).

Figure 11-7

Long-Term Effectiveness Phase: Change from Baseline in YMRS Score Over Time by Efficacy Phase Randomized Treatment Group (with LOCF Shown).

11.2.2.3.2. Long-Term Effectiveness Phase: Secondary Endpoints

Descriptive summaries of baseline, week 24/LOCF and of change from baseline to end of phase/LOCF values are included in Section 14 for all continuous secondary endpoints by Efficacy Phase randomized treatment group and by LTE Phase age strata.

Table 11-12A below includes YMRS parent score, YMRS child score, Children Depression Rating Scale-Revised (CDRS-R) and Children’s Global Assessment Scale (CGAS). These results are also summarized by Efficacy Phase randomized treatment arm in Tables 14.2.2.2.2B, 14.2.2.2.3B and 14.2.2.2.4B and by age strata in Tables 14.2.2.2.2A, 14.2.2.2.3A and 14.2.2.2.4A.

Table 11-12A. Long-Term Effectiveness Phase: Secondary Efficacy Endpoints by Efficacy Phase Randomized Treatment Group.

Table 11-12A

Long-Term Effectiveness Phase: Secondary Efficacy Endpoints by Efficacy Phase Randomized Treatment Group.

The parents of children treated with lithium during the Efficacy Phase reported a mean (SD) of 10.0 (SD 6.3) YMRS at baseline and 8.5 (SD 8.2) at week 24/LOCF, with a mean (SD) change of −1.4 (SD 8.0). The parents of children treated with placebo during the Efficacy Phase reported YMRS of 23.3 (SD 7.1) at baseline, 7.5 (SD 9.3) at week 24/LOCF and a change of −16.8 (SD 10.4) points. The children treated with lithium self-reported YMRS scores of 6.1 (SD 5.5) at baseline, 5.4 ± 6.6 at week 24/LOCF and a change of −0.7 (SD 8.7) points. Placebo treated children entered the LTE Phase with a reported YMRS of 15.8 (SD 7.6), a week 24/LOCF YMRS of 5.2 (SD 6.7) and a change score of −10.5 (SD 9.0) points.

The results of the Children Depression Rating Scale-Revised (CDRS-R) showed similar results as above, the participants treated with lithium during the Efficacy Phase remained stable (change score of −0.8 (SD 8.7)) and the participants tapering from placebo to lithium during the LTE Phase had a 10.9 (SD 8.3) point decrease in CDRS-R score from baseline to week 24/LOCF.

The Children’s Global Assessment Scale (CGAS) also showed the same trends, with a 2.5 (SD 18.2) increase in score for Efficacy Phase lithium participants and a 16.4 (SD 14.9) increase in the Efficacy Phase placebo participants at week 24/LOCF.

Clinical Global Impressions Scales (CGI-S, CGI-I) were measured on a Likert scale and used the following definition of success and failures.:

CGI-Severity (Mania, Depression and Overall Illness):

  • Success: Baseline = 1 or 2 and less or same severity level at LOCF
    Baseline = 3 and less severe at LOCF
    Baseline = 4, 5, 6 and LOCF=1, 2, 3 and at least a 2 point decrease
  • Failure: Otherwise

CGI-Improvement (Mania, Depression and Overall Illness):

  • Success: Baseline CGI-S = 1 or 2 and LOCF CGI-I = 1,2,3,4
    Baseline CGI-S = 3 and LOCF CGI-I = 1,2,3
    Baseline CGI-S = 4, 5, 6 and LOCF CGI-I = 1,2
  • Failure: Otherwise

Clinical Global Impressions Scales (CGI-S, CGI-I) results are presented in Table 11-12B and Table 14.2.2.2.5B by Efficacy Phase randomized treatment group and in Table 14.2.2.2.5A by age strata. Overall, participants entering the LTE Phase after placebo treatment, showed success rates in CGI severity of 85% for mania, 92% for depression and 69% for over all illness. At week 24/LOCF success rates for the CGI-improvement was 88% for mania (90% for participants treated with lithium in the Efficacy Phase and 85% for placebo participants), 86% for depression (83% lithium and 92% placebo) and 79% for overall illness (79% lithium and 77% placebo).

Table 11-12B. Long-Term Effectiveness Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement by Efficacy Phase Randomized Treatment Group.

Table 11-12B

Long-Term Effectiveness Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement by Efficacy Phase Randomized Treatment Group.

Three categorical variables were identified as additional secondary efficacy parameters for the LTE Phase: Response status, Remission status, YMRS (Summary) percentage improvement category. The definition of response was revised from the Efficacy Phase definition to use pre-treatment Efficacy Phase day 1 as the baseline summary YMRS value and the CGI-I (overall illness) at LOCF was modified to the CGI-S (overall illness) at LOCF. See below:

  • Response: YMRS reduction ≥ 50% from Efficacy Phase baseline and CGI-S (Overall illness) = 1 or 2 at LOCF
  • Partial response: YMRS reduction 25-49% % from Efficacy Phase baseline and CGI-S (Overall illness) ≤ 3 at LOCF OR
    YMRS reduction ≥ 50% and CGI-S (Overall illness) = 3 at LOCF
  • Non-response: YMRS reduction < 25% % from Efficacy Phase baseline or CGI-S (Overall illness) ≥ 4 at LOCF

Definition of remission is the same as defined in the Efficacy Phase. Improvement categories are defined as in the Efficacy Phase but then baseline YMRS value is the pre-treatment Efficacy Phase day 1 value.

Tables 11-12C and 14.2.2.2.1B summarize the results of response, remission, and improvement analyses. Response, remission and improvement status was similar in both participants treated with lithium in the Efficacy and LTE Phases and participants treated with placebo in the Efficacy and tapering to lithium in the LTE phase. Complete response rates were 66% in the lithium group and 69% in the placebo group, remission rates are 62% and 69% respectively and improvement (≥50% decrease in YMRS) rates are 86% and 77% respectively at week 24/LOCF..

Table 11-12C. Long-Term Effectiveness Phase: Secondary Efficacy Endpoints – Response, Remission and Improvement by Efficacy Phase Randomized Treatment Group.

Table 11-12C

Long-Term Effectiveness Phase: Secondary Efficacy Endpoints – Response, Remission and Improvement by Efficacy Phase Randomized Treatment Group.

Discontinuation due to any cause during the LTE Phase and discontinuation due to lack of efficacy during the LTE Phase were evaluated. These results are reported by Efficacy Phase randomized treatment group in Table 11-12D, Table 14.2.2.2.1B, and Table 14.2.2.2.6B.

Table 11-12D. Long-Term Effectiveness Phase: Secondary Efficacy Endpoints – Discontinuation by Efficacy Phase Randomized Treatment Group.

Table 11-12D

Long-Term Effectiveness Phase: Secondary Efficacy Endpoints – Discontinuation by Efficacy Phase Randomized Treatment Group.

Nineteen participants (48% lithium and 38% placebo) did not complete the LTE phase, and there was little difference in the time to discontinuation during the LTE Phase for participants entering the phase on lithium compared to participants entering the phase on placebo (Figure 11-8: Time to Discontinuation for Any Reason During the Long-Term Effectiveness Phase by Efficacy Phase Randomized Treatment Group).

Figure 11-8. Time to Discontinuation for Any Reason During the Long-Term Effectiveness Phase by Efficacy Phase Randomized Treatment Group.

Figure 11-8

Time to Discontinuation for Any Reason During the Long-Term Effectiveness Phase by Efficacy Phase Randomized Treatment Group.

Only one participant discontinued due to lack of efficacy.

Discontinuation for any reason during the LTE phase is also analyzed by age stratum. 61% (17/28) of participants in the 7-11 year age stratum and 43% (6/14) of participants in the 12-17 year age stratum completed the LTE Phase. Time to discontinuation by age strata is shown in Table 11-12E, Figure 11-9, Figure 14.2.2.2A and Table 14.2.2.2.6A.

Table 11-12E. Long-Term Effectiveness Phase: Secondary Efficacy Endpoints – Discontinuation by Age Strata.

Table 11-12E

Long-Term Effectiveness Phase: Secondary Efficacy Endpoints – Discontinuation by Age Strata.

Figure 11-9. Time to Discontinuation for Any Reason During the Long-Term Effectiveness Phase by Age Strata.

Figure 11-9

Time to Discontinuation for Any Reason During the Long-Term Effectiveness Phase by Age Strata.

11.2.2.3.3. Long-Term Effectiveness Phase: Tertiary Endpoints

Long-Term Effectiveness Phase tertiary endpoints include results from the Parent General Behavior Inventory-10 Item Mania Scale (PGBI-10M), ADHD Rating Scale-IV, Brief Psychiatric Rating Scale for Children (BPRS-C) and Pediatric Anxiety Rating Scale (PARS). The change from baseline to the end of phase scores, based upon LOCF values were evaluated for each measure.

Results from the analyses of the tertiary efficacy endpoints from the LTE Phase are summarized in Table 11-13 below and Tables 14.2.2.3.1, 14.2.2.3.2, 14.2.2.3.3 and 14.2.2.3.4. The Parent General Behavior Inventory-10 Item Mania Scale showed a change of −1.1 (SD 6.7) in participants treated with lithium during the Efficacy Phase, while participants treated with placebo and titrated to lithium had a mean of −7.5 (SD 5.3). Similar results are seen with ADHD Total Score: mean (SD) change in lithium participants of −4.0 (SD 13.6) and −11.6 (SD 8.8) in placebo participants; BPRS Total Score with changes of 0.7 (SD 14.8) in lithium and −10.5 (SD 17.8) in placebo participants; and in PARS Total Score with a decrease of 1.7 (SD 5.5) in lithium and a decrease of 4.5 (SD 5.2) in placebo participants.

Table 11-13. Long-Term Effectiveness Phase:Tertiary Efficacy Endpoints.

Table 11-13

Long-Term Effectiveness Phase:Tertiary Efficacy Endpoints.

11.2.2.3.4. Long-Term Effectiveness Phase: Descriptive Endpoints

Descriptive efficacy endpoints for the LTE phase are summarized with descriptive statistics (N, mean, standard deviation, median, minimum, and maximum) for baseline, Week 24/End-of-Phase/LOCF and change from baseline to Week 24/End-of-Phase/LOCF scores. These measures include Family Environmental Scale (FES), Caregiver Strain Questionnaire (CSQ), Irritability, Depression and Anxiety (IDA), Nisonger Child Behavior Rating Form (NCBRF-TIQ) and Social Adjustment Inventory for Children and Adolescents (SA/CA).

Results from the analyses of the descriptive efficacy endpoints from the LTE Phase are summarized in Tables 14.2.2.4.1 through 14.2.2.4.5. The descriptive endpoints were only completed on a fraction of the participants due to Amendment 5 protocol changes (Section 9.9.1) to reduce burden on the participants and increase accrual. The results showed no unexpected findings and in general, showed changes in the appropriate directions.

11.2.2.4. By-Participants Displays

By-participant displays (i.e. “patient profiles”) are presented for the LTE Phase in Section 14.4. These profiles include lithium and placebo doses in mg, study drug dose in mg/kg, lithium level, % compliance, YMRS score, CGI overall illness score, creatinine clearance, TSH, adverse events and serious adverse events temporally over the course of the eight week Efficacy Phase and the 24 week LTE Phase.

11.2.2.5. Conclusions – Long-Term Effectiveness Phase (Efficacy)

Overall, participants who were previously treated with lithium during the acute double blind efficacy phase had continued benefit in their manic symptoms. This observation provides evidence to support lithium as a maintenance treatment in this patient population.

Also, participants who were placebo-treated in the acute double blind efficacy phase who entered this open-label phase symptomatic had salutary effects, as measured by reductions in YMRS scores. These results further suggest that lithium has efficacy that is superior to placebo in the treatment of pediatric manic/mixed episodes.

11.2.3. Discontinuation Phase

To be eligible for the Discontinuation Phase the participant must have completed at least 6 of the last 8 consecutive weeks of the LTE Phase with a complete remission of psychotic features and a remission of mood symptoms (YMRS < 10 and CDRS-R < 35). The analyses of this phase combines 21 participants from the COLT1 study and 10 participants from the COLT2 study for a total of 31 participants. See Section 10.1.4 for a discussion of participant disposition.

11.2.3.1. Demographics and Other Baseline Characteristics

The Discontinuation Phase baseline demographics are summarized for the Efficacy Population (N=31) is presented in Table 11-14 and Tables 14.1.3.3. These tables include information on participant demographics by randomized treatment. The baseline visit of the Discontinuation Phase was defined as the last visit of the LTE Phase.

Table 11-14. Discontinuation Phase Demographic Characteristics (Efficacy Population).

Table 11-14

Discontinuation Phase Demographic Characteristics (Efficacy Population).

Lithium participants had a median age of 11.2 years (range 8.2-14.3 years) at randomization to the Discontinuation Phase. Placebo participants’ median age was 12.2 years (range 9.5 – 17.1 years); Kruskal-Wallis p=0.068. There was also no difference in gender (χ2 p=0.69), ethnicity or race (χ2 p=0.28) distributions between treatment arms. Baseline characteristics were comparable between participants enrolled in the COLT1 study and participants who were enrolled in COLT2 study for age, gender, and race (data not shown).

The Discontinuation Phase was a 28-week-long, double-blind phase where participants were randomized to receive either continued treatment with lithium or tapered to placebo. Length of time in the Discontinuation Phase by randomized treatment arm is shown in Table 11-15. Median number of weeks in the phase was 27.4 for participants randomized to lithium treatment and 6.0 weeks for participants randomized to placebo treatment.

Table 11-15. Length of Time in the Discontinuation Phase.

Table 11-15

Length of Time in the Discontinuation Phase.

11.2.3.2. Measurement of Treatment Adherence

COLT1 site ▬ participants (n=9) are not included in the adherence calculations due to a problem with how the expected pill counts were entered. Overall compliance was assessed by averaging the weekly visit compliance rates. The overall median compliance rate was 94% with a range from 80% to 111% as shown in Table 11-16.

Table 11-16. Overall Treatment Adherence during the Discontinuation Phase (Efficacy Population).

Table 11-16

Overall Treatment Adherence during the Discontinuation Phase (Efficacy Population).

In addition, medication compliance was assessed through review of the lithium trough serum levels. Individual lithium trough serum levels are listed in Appendix 16.2. A scatter plot of all lithium levels by treatment arm is shown below in Figure 11-10. Four participants from COLT1 randomized to placebo, received the incorrect treatment (lithium); see Section 10.2.1.2.

Figure 11-10. Lithium Levels by Treatment Over Time During the Discontinuation Phase (Efficacy Population).

Figure 11-10

Lithium Levels by Treatment Over Time During the Discontinuation Phase (Efficacy Population).

11.2.3.3. Efficacy Results

The discontinuation phase included 10 participants from COLT2 and 21 participants from COLT1, for a total of 31 participants. Accrual was estimated to be higher with an expected N=60 from COLT2 for a total of 81 participants.

11.2.3.3.1. Discontinuation Phase: Primary Endpoint

The primary efficacy endpoint for the Discontinuation Phase is the relative risk of discontinuation for any reason, comparing lithium to placebo.

Time from randomization to discontinuation for any reason was analyzed using Cox proportional hazards regression models. Participants who did not discontinue during the Discontinuation Phase were censored at the time of the Week 28 visit. The Cox proportional hazards regression model included treatment group as the independent factor and was stratified by age strata. These results are summarized in Table 14.2.3.1.1 and Figure 14.2.3.1.1.

Figure 11-11 shows the Kaplan-Meier estimates for time to discontinuation for any reason. Although 43% of the participants randomized to lithium vs. 71% of the participants randomized to placebo did not complete the Discontinuation Phase, this difference was not statistically significant, Cox regression HR=0.54 (compared to placebo) 95% CI 0.20 – 1.49, p=0.235.

Figure 11-11. Primary Efficacy Endpoint: Time to Discontinuation for Any Reason During the Discontinuation Phase by Discontinuation Phase Randomized Treatment Group (Efficacy Population).

Figure 11-11

Primary Efficacy Endpoint: Time to Discontinuation for Any Reason During the Discontinuation Phase by Discontinuation Phase Randomized Treatment Group (Efficacy Population).

The results of Cox regression analysis of the primary endpoint on the per-protocol population, (Table 14.2.3.1.2) show that risk of discontinuation is significantly lower in the lithium arm with HR=0.27 (compared to placebo), 95% CI 0.08 – 0.87, and p=0.029. This analysis excluded 8 participants; 2 participants from COLT2 with compliance < 60% for at least 1 week (and 1 also receiving prohibited medication) and 6 participants from COLT1 (3 with randomized treatment errors and 2 who received a dose >40 mg/kg/day and 1 with both randomized treatment error and dose >40 mg/kg/day). A second revised per protocol population was developed (see Section 10.2 for details). This population excluded the COLT2 participant receiving prohibited medication and the 4 COLT1 participants with randomized treatment errors. The results (Table 14.2.3.1.3) are no longer statistically significant with HR=0.41 (compared to placebo), 95% CI 0.14 – 1.15, and p=0.090. The Kaplan-Meier estimates for these analyses are shown in Figure 11-2.

Figure 11-12. Time to Discontinuation for Any Reason During the Discontinuation Phase by Discontinuation Phase Randomized Treatment Group (Per-Protocol and Revised Per-Protocol Populations).

Figure 11-12

Time to Discontinuation for Any Reason During the Discontinuation Phase by Discontinuation Phase Randomized Treatment Group (Per-Protocol and Revised Per-Protocol Populations).

An additional sensitivity analysis was performed on the primary endpoint in Table 14.2.3.1.4. Three major protocol deviations from the COLT1 study included participants randomized to placebo but actually received lithium treatment for the entire phase (Appendix 16.4.2.3). The sensitivity analysis is grouped by treatment received (not treatment assigned). All 31 participants are included, with the 3 placebo participants who received lithium grouped with the lithium participants. The results of Cox regression analysis of the primary endpoint by treatment received show risk of discontinuation is significantly lower in the lithium arm with HR=0.28 (compared to placebo), 95% CI 0.10 – 0.78, and p=0.015.

Figure 11-13. Discontinuation Phase: Time to Discontinuation for Any Reason During the Discontinuation Phase by Treatment Received Group (Efficacy Population).

Figure 11-13

Discontinuation Phase: Time to Discontinuation for Any Reason During the Discontinuation Phase by Treatment Received Group (Efficacy Population).

Figure 11-14 shows survival estimates of time to discontinuation for any reason by study and by age strata. Time to discontinuation was similar in COLT1 and COLT2 and for children age 7-11 years and 12-17 years.

Figure 11-14. Time to Discontinuation for Any Reason During the Discontinuation Phase by Study and by Age Strata (Efficacy Population).

Figure 11-14

Time to Discontinuation for Any Reason During the Discontinuation Phase by Study and by Age Strata (Efficacy Population).

11.2.3.3.2. Discontinuation Phase: Secondary Endpoints

Time to discontinuation for lack of efficacy, Table 14.2.3.2.1, was identified as the key secondary efficacy endpoint for the Discontinuation Phase. Only one participant ▬, randomized to lithium] discontinued due to lack of efficacy (mania) during the discontinuation phase, the participant left the study at 25.3 weeks.

Recurrence of mood episodes was also considered as a secondary efficacy variable for the Discontinuation Phase. Participants were considered to have a recurrence of mood episodes if, when compared to their clinical status at entry into the Discontinuation Phase, they have any of the following:

1)

A ≥ 10 point increase from their respective baseline rating in YMRS or CDRS-R scores for ≥ 2 consecutive visits,

2)

Persistent moderate symptoms (as defined in protocol Section 8.3.2) of depression and/or mania for 2 consecutive visits, including interim visits, separated by at least 14 days,

3)

Development of syndromal manic episode, mixed episode, or syndromal depressive episode.

Thirty-six percent of lithium participants experienced a recurrence of mood episode compared to 65% of placebo participants (p=0.156) as shown in Table 11-17 and Table 14.2.3.2.3.

Table 11-17. Discontinuation Phase: Secondary Endpoint - Recurrence of Mood Episodes by Discontinuation Phase Randomized Treatment Group.

Table 11-17

Discontinuation Phase: Secondary Endpoint - Recurrence of Mood Episodes by Discontinuation Phase Randomized Treatment Group.

Time to intervention for mood episode is a secondary endpoint and is shown in Figure 11-15 and Table 14.2.3.2.4. Participants with recurrence of mood episodes are considered to have an intervention if they received concomitant psychotherapy, concomitant psychotropic medications or if they leave the Discontinuation Phase at the time of the recurrence of mood episode. Participants are considered failures at the time of the intervention and other participants are censored when they leave the Phase. One participant [▬] had a recurrence of mood episodes at week 2 but had no interventions and completed all 28 weeks of the Discontinuation Phase. This participant was censored at week 28 for this analysis. Results are not statistically significant between treatment groups, Cox regression HR=0.44 (compared to placebo) 95% CI 0.14 – 1.40, p=0.166.

Figure 11-15. Discontinuation Phase: Secondary Endpoint - Time to Intervention for Mood Episode During the Discontinuation Phase by Discontinuation Phase Randomized Treatment Group.

Figure 11-15

Discontinuation Phase: Secondary Endpoint - Time to Intervention for Mood Episode During the Discontinuation Phase by Discontinuation Phase Randomized Treatment Group.

Continuous secondary endpoints were analyzed using the change score as the dependent variable, baseline score as a covariate, and age strata, gender, study (COLT1 vs. COLT2) and treatment group as factors. Endpoints analyzed were YMRS total score; YMRS parent score; YMRS child score; Children Depression Rating Scale-Revised (CDRS-R); and Children’s Global Assessment Scale (CGAS). These results are summarized in Table 11-18A below.

Table 11-18A. Discontinuation Phase: Secondary Efficacy Endpoints by Discontinuation Phase Randomized Treatment Group.

Table 11-18A

Discontinuation Phase: Secondary Efficacy Endpoints by Discontinuation Phase Randomized Treatment Group.

The baseline (last visit of the LTE Phase) YMRS summary score (mean (SD) was 4.6 (SD 3.3) in participants randomized to lithium and 5.9 (SD 2.2) in participants randomized to placebo. YMRS summary score increased 5.1 (SD 10.4) points in the lithium group and 9.0 (SD 9.6) in placebo group, p=0.127. Results are similar for Parent and Child YMRS score. Parents of children randomized to lithium reported a 5.1 (SD 10.3) point increase and parents of children randomized to placebo reported a 9.1 (SD 10.0) increase in YMRS score over the course of the phase, p=0.183 and children reported a 2.7 (SD 8.5) increase when randomized to lithium and a 4.9 (SD 9.2) point increase when randomized to placebo, p=0.251 after adjusting for baseline score, age strata, gender and study. See Table 14.2.3.2.5.

Baseline scores for the results of the Children Depression Rating Scale-Revised (CDRS-R) are 19.4 (2.9) in the lithium group and 22.6 (SD 5.7) in the placebo group, with changes at week 28/LOCF of 2.4 ± 6.1 and 4.0 ± 10.3 in each group respectively, p=0.257. See Table 14.2.3.2.6.

The Children’s Global Assessment Scale (CGAS), Table 11-18A below and Table 14.2.3.2.7, a mean (SD) decrease of 1.6 (SD 20.9) in the lithium arm and decrease of 7.6 (SD 12.9) points in the placebo arm from baseline to LOCF. These results were not statistically significant, p=0.123.

YMRS score over time in the Discontinuation Phase is shown in Figure 14.2.3.2.2. Change in YMRS score over time is shown in Figure 14.2.3.2.3 and below in Figure 11-16. Participants that terminated the phase early had higher LOCF YMRS scores and larger changes in YMRS as is seen in the LOCF values below.

Figure 11-16. Discontinuation Phase: Change from Baseline in YMRS Summary Score Over Time by Discontinuation Phase Randomized Treatment Group (with LOCF Shown).

Figure 11-16

Discontinuation Phase: Change from Baseline in YMRS Summary Score Over Time by Discontinuation Phase Randomized Treatment Group (with LOCF Shown).

Clinical Global Impressions Scales (CGI-S, CGI-I) were measured on a Likert scale and were analyzed using Fisher’s exact test for categorical data. Success and failures used the same definition as described in the Long-Term Effectiveness Phase section.

Clinical Global Impressions Scales (CGI-S, CGI-I) results are presented in Table 11-18B and in Table 14.2.3.2.8. Severity and Improvement outcomes were measured in the same way as defined in the LTE Phase. While for each outcome (mania, depression, overall illness), more participants in the lithium group had better outcome, these results did not achieve statistical significance. Success rates for mania were 28% higher in the lithium arm, success rates for depression were 23% higher in the lithium arm and success rates for overall illness were 28% higher in the lithium arm. The CGI-I scores showed similar results as the severity scores, with mania and overall illness having 26% higher success rate in the lithium group compared to the placebo group and depression having a 14% higher success rate in the placebo group.

Table 11-18B. Discontinuation Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement by Discontinuation Phase Randomized Treatment Group.

Table 11-18B

Discontinuation Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement by Discontinuation Phase Randomized Treatment Group.

Response, remission and improvement status are shown in Table 11-18C below. See the Long-Term Effectiveness Phase for definition of response. While p-values were not significant, response, remission and improvement rates were all higher for participants randomized to lithium during the Discontinuation Phase compared to participants randomized to placebo. Complete response rates were 57% in the lithium group and 29% in the placebo group, remission rates are 57% and 29% respectively and improvement (≥50% decrease in YMRS) rates are 64% and 53% respectively at week 28/LOCF.

Table 11-18C. Discontinuation Phase: Secondary Efficacy Endpoints - Response, Remission and Improvement by Discontinuation Phase Randomized Treatment Group.

Table 11-18C

Discontinuation Phase: Secondary Efficacy Endpoints - Response, Remission and Improvement by Discontinuation Phase Randomized Treatment Group.

11.2.3.3.3. Discontinuation Phase: Tertiary Endpoints

Tertiary endpoints for the Discontinuation Phase included results from the Parent General Behavior Inventory-10 Item Mania Scale (PGBI-10M); ADHD Rating Scale-IV; Brief Psychiatric Rating Scale for Children (BPRS-C); and Pediatric Anxiety Rating Scale (PARS). The change from baseline to the end of phase scores, based upon LOCF values, was evaluated for each measure. Linear regression analysis assessed the change from baseline to end of phase score. The change score is the dependent variable; baseline score is included as a covariate; and age strata, gender, study, and treatment group are factors.

Results from the analyses of the tertiary efficacy endpoints from Discontinuation Phase are summarized in Table 11-19 below and Tables 14.2.3.3.1, 14.2.3.3.2, 14.2.3.3.3, and 14.2.3.3.4. None of the endpoints showed any significant differences, the Parent General Behavior Inventory-10 Item Mania Scale showed a 0.9 (SD 8.0) mean increase from baseline to week 28/LOCF lithium participants and a 5.0 point (SD 8.0) mean increase in the placebo treated participants. The remaining endpoints showed little difference in change scores between lithium and placebo groups.

Table 11-19. Discontinuation Phase: Tertiary Efficacy Endpoints by Discontinuation Phase Randomized Treatment Group.

Table 11-19

Discontinuation Phase: Tertiary Efficacy Endpoints by Discontinuation Phase Randomized Treatment Group.

11.2.3.4. Tabulation of Individual Response Data

Tabluations of individual response data can be found in Appendix 16.2.

11.2.3.5. By-Participants Displays

No by-participant displays or participant profiles were created for the Discontinuation Phase

11.2.3.6. Conclusions – Discontinuation Phase (Efficacy)

These data suggest that lithium has effectiveness in the maintenance treatment of pediatric bipolarity. On the primary efficacy measure, participants treated with lithium had a more durable positive effect.

Of note, on many of the symptomatic measures, superiority for lithium was not found. There may be several reasons for this. First, the investigators were supposed to discontinue participants from this study phase should clinical worsening occurred. As investigators were not directed to wait until participants experienced a complete relapse (in which marked symptomatology would be manifest, the symptomatic changes that were observed in participants who discontinued were expected to be modest. This intentional goal of limiting the degree to which clinical worsening could occur may have also limited to detect between group assessments of symptomatology.

There was another noteworthy reason to explain why between group differences might not have been observed in this phase. There was a relatively modest number of participants randomized during this phase and sample size considerations may have impacted statistical power.

Regardless of the limitations of this phase, the primary measure provides evidence to support lithium as a maintenance treatment in this patient population. In addition, the secondary measures provide supportive evidence for the efficacy of lithium.

11.2.4. Restabilization Phase

11.2.4.1. Demographics and Other Baseline Characteristics

Participants who experienced significant deterioration in clinical status during the Discontinuation Phase were offered 8 weeks of treatment with open label lithium in a Restabilization Phase. The decision to withdraw a participant from the Discontinuation Phase was made at the treating investigator’s discretion based on the clinical status of the participant. Participants who were receiving placebo during the Discontinuation Phase were treated with the dose of lithium at which earlier stabilization had occurred. Thirteen participants were enrolled in the Restabilization Phase: 3 participants from the COLT2 study and 10 participants from the COLT1 study. See Section 10 for a discussion of participant disposition.

The efficacy population for the Restabilization Phase includes all participants who received study medication during this phase. All 13 participants who enrolled in the Restabilization Phase received lithium and had follow-up visits during the phase.

Demographic characteristics are shown in Table 14.1.3.4 and Table 11-20 below. Three participants enrolled in the Restabilization Phase were treated with lithium during the Discontinuation Phase. Participants have a median age of 12.3 years (range 9.6-16.2) at enrollment to the Restabilization Phase. Sixty-two percent were male, 92% reported white race and all were non-Hispanic ethnicity.

Table 11-20. Restabilization Phase Demographic Characteristics (Efficacy Population).

Table 11-20

Restabilization Phase Demographic Characteristics (Efficacy Population).

Length of time in the Restabilization Phase is shown in Table 11-21 for all participants entering the phase by Discontinuation Phase treatment arm, Restabilization Phase age strata, and study. Median time in the phase was 8 weeks, range 2.9-8.4 weeks for all participants. Length of time in the Restabilization was similar for Discontinuation Phase treatment arm groups, age strata groups, and study.

Table 11-21. Length of Time in the Restabilization Phase.

Table 11-21

Length of Time in the Restabilization Phase.

11.2.4.2. Measurement of Treatment Adherence

Overall compliance for the Restabilization Phase (Table 11-22) was assessed by averaging the weekly visit compliance rates. The overall median compliance rate was 98% with a range from 62% to 118%. Compliance from the COLT1 study was 99% (88%-118%) and from COLT2 study was 93% (62%-99%).

Table 11-22. Overall Treatment Adherence during the Restabilization Phase.

Table 11-22

Overall Treatment Adherence during the Restabilization Phase.

In addition, medication adherence was assessed by the review of the lithium trough serum levels. A scatter plot of all lithium levels is shown below.

Figure 11-17. Lithium Levels Over Time During the Restabilization Phase (Efficacy Population).

Figure 11-17

Lithium Levels Over Time During the Restabilization Phase (Efficacy Population).

11.2.4.3. Efficacy Results

11.2.4.3.1. Restabilization Phase: Primary Endpoint

The primary efficacy endpoint for the Restabilization Phase was the change from baseline to the end of phase in the YMRS Summary score, based upon LOCF values. Baseline score is defined as the last visit of the Discontinuation Phase. Participants discontinued the Discontinuation Phase and entered the Restabilization Phase at a median of 5 weeks (range 2 – 11) of Discontinuation Phase treatment. Descriptive summaries are provided for the efficacy population.

Primary endpoint results are summarized in Table 14.2.4.1 and Table 11-23. The mean baseline score, measured when the participant terminated the Discontinuation Phase, was 20.8 (SD 6.6). Mean week 8/LOCF summary YMRS is 11.3 (SD 7.6) with the mean change from baseline to week 8/LOCF of −9.5 (SD 10.9).

Table 11-23. Restabilization Phase: Primary Efficacy Endpoint by Discontinuation Phase Randomized Treatment Group.

Table 11-23

Restabilization Phase: Primary Efficacy Endpoint by Discontinuation Phase Randomized Treatment Group.

Figure 14.2.4.1.1 shows each weekly YMRS summary score. This longitudinal analysis uses all available YMRS scores at each time point.

The baseline visit for the Restabilization Phase is the last Discontinuation Phase visit. The change in summary YMRS score from the baseline visit to each visit during the Restabilization Phase is shown in Figure 14.2.4.1.2 and Figure 11-18. Nine of the 13 participants completed all 8 weeks of the Restabilization Phase and had a mean decrease in YMRS score of 13.9 (SD 7.0) points at 8 weeks.

Figure 11-18. Restabilization Phase: Change from Baseline in YMRS Summary Score During the Restabilization Phase (with LOCF Shown).

Figure 11-18

Restabilization Phase: Change from Baseline in YMRS Summary Score During the Restabilization Phase (with LOCF Shown).

11.2.4.3.2. Restabilization Phase: Secondary Endpoints

Descriptive summaries of baseline, week 8/LOCF, and change from baseline to end of phase/LOCF values are included for all continuous secondary endpoints. This includes YMRS parent score; YMRS child score; Children Depression Rating Scale-Revised (CDRS-R); and Children’s Global Assessment Scale (CGAS). These results are summarized in Table 11-24A below and by Discontinuation Phase randomized treatment arm in Tables 14.2.4.2.2, 14.2.4.2.3 and 14.2.4.2.4.

Table 11-24A. Restabilization Phase: Secondary Efficacy Endpoints.

Table 11-24A

Restabilization Phase: Secondary Efficacy Endpoints.

The parents of children reported a mean (SD) of 20.8 (SD 6.6) YMRS score at baseline and 11.2 (SD 7.6) at week 8/LOCF, with a mean (SD) change of −9.6 (SD 10.9). The children self-reported YMRS scores of 13.2 (SD 9.5) at baseline, 10.2 (SD 7.7) at week 8/LOCF and a change of −2.9 (SD 11.7) points.

The results of the Children Depression Rating Scale-Revised (CDRS-R) were 29.7 (SD 11.0) at baseline score, 26.6 (SD 9.7) score at week 8/LOCF and a decrease of 3.1 (SD 13.2) from baseline to LOCF.

The Children’s Global Assessment Scale (CGAS) reported 56.3 (SD 6.8) at baseline, 65.8 (SD 15.0) at LOCF and a 9.3 (SD 10.9) increase in score from baseline to LOCF.

Clinical Global Impressions Scales (CGI-S and CGI-I) are measured on a Likert scale and used the same definition of success and failures as was described in the Long-Term Effectiveness Phase section.

Clinical Global Impressions Scales (CGI-S and CGI-I) results are presented in Table 11-24B and in Table 14.2.4.2.5. Overall, participants entering the Restabilization Phase after mood relapse in the Discontinuation Phase showed success rates of 62% for mania; 69% for depression; and 54% for overall illness. At Week 8/LOCF success rates for the CGI-improvement of 77% for mania, 69% for depression and 69% for overall illness.

Table 11-24B. Restabilization Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement.

Table 11-24B

Restabilization Phase: Secondary Efficacy Endpoints - CGI Severity and Improvement.

Three categorical variables were identified as additional secondary efficacy parameters for the Restabilization Phase: Response status, Remission status, YMRS (Summary) percentage improvement category. Calculation of these endpoints is described in the Long-Term Effectiveness Phase. Baseline values are from the pre-dose value from Day 1 of the Efficacy Phase (COLT2) or Pharmacokinetic Phase (COLT1).

Table 11-24C and 14.2.4.2.1 summarize the response, remission and improvement status during the Restabilization Phase. Six (46%) of the 13 participants had a complete response, 6 (46%) had remission and 10 (77%) had a ≥50% reduction in YMRS summary score from pre-treatment baseline to week 8/LOCF of the Restabilization Phase.

Table 11-24C. Restabilization Phase: Secondary Efficacy Endpoints - Response, Remission and Improvement.

Table 11-24C

Restabilization Phase: Secondary Efficacy Endpoints - Response, Remission and Improvement.

Discontinuation due to any cause during the Restabilization was evaluated. These results are reported in Table 14.2.4.2.1, Figure 14.2.4.2, Table 11-24D, and Figure 11-19. Four participants discontinued the Restabilization Phase before completing the 8 weeks: 2 had consent withdrawn; 1 was lost to follow-up; and 1 for protocol non-compliance (see Section 10.1 for disposition of participants). No participants discontinued due to lack of efficacy.

Table 11-24D. Secondary Efficacy Endpoints: Discontinuation for Any Reason During the Restabilization Phase.

Table 11-24D

Secondary Efficacy Endpoints: Discontinuation for Any Reason During the Restabilization Phase.

Figure 11-19. Time to Discontinuation for Any Reason During the Restabilization Phase.

Figure 11-19

Time to Discontinuation for Any Reason During the Restabilization Phase.

11.2.4.3.3. Restabilization Phase: Tertiary Endpoints

Restabilization Phase tertiary endpoints included results from the Parent General Behavior Inventory-10 Item Mania Scale (PGBI-10M); ADHD Rating Scale-IV; Brief Psychiatric Rating Scale for Children (BPRS-C); and Pediatric Anxiety Rating Scale (PARS). The change from baseline to the end of phase scores, based upon LOCF values, were evaluated for each measure.

Results from the analyses of the tertiary efficacy endpoints from the Restabilization Phase are summarized in Table 11-25 below and Tables 14.2.4.3.1, 14.2.4.3.2, 14.2.4.3.3 and 14.2.4.3.4. The Parent General Behavior Inventory-10 Item Mania Scale showed a mean change of −3.8 (SD 4.7) in participants from baseline to the last visit of the Restabilization Phase. ADHD Total Score mean decreased by 3.6 (SD 6.0) points over the course of the phase, BPRS-C Total Score mean decreased 6.3 (SD 14.5) points and the PARS Subscale Score mean decreased 0.8 (SD 4.7) points during the Restabilization Phase.

Table 11-25. Restabilization Phase: Tertiary Efficacy Endpoints.

Table 11-25

Restabilization Phase: Tertiary Efficacy Endpoints.

11.2.4.4. Tabulation of Individual Response Data

Tabulation of individual response data can be bound in Appendix 16.2.

11.2.4.5. By-Participants Displays

No by-participant displays or participant profiles were created for the Restabilization Phase.

11.2.4.6. Conclusions – Restabilization Phase (Efficacy)

Although numbers are small in the Restabilization Phase, data suggest that lithium improved YMRS scores of participants experiencing mood relapse during the Discontinuation Phase. This is an important consideration. These data suggests, despite a limited number of participants, that stable participants who discontinue lithium may continue to receive benefit from re-initiating previously effective lithium treatment.

11.2.1. Statistical and Analytical Issues

11.2.1.1. Adjustments for Co-variates

For Efficacy Phase efficacy analyses: primary, secondary, and tertiary continuous endpoints were adjusted for baseline value, age strata, gender, weight strata, and site.

11.2.1.2. Handling of Dropouts or Missing Data

Primary, secondary, and tertiary continuous endpoints were analyzed using last observation carried forward. Time to discontinuation analyses counted dropouts as failures. Time to mood relapse used survival analysis methods that censored dropouts at their last follow-up.

11.2.1.3. Interim Analysis and Data Monitoring

Interim safety analyses were scheduled in the protocol to be submitted to the DMC after enrollment of approximately 10%, 25%, 50% and 75% of participants. Completed interim analyses include:

  • 06/2011: Interim safety analysis.
  • 09/2011: Blinded sample size recalculation.
  • 01/2012: Response to DMC’s recommendation to re-estimate sample size using observed treatment effect.
  • 05/2012: Interim safety analysis
  • 05/2012: Interim futility and inefficacy analysis.
  • 07/2012: Updated interim futility and inefficacy analysis.
  • 09/2013: Final safety analysis

See Section 9.8 for details.

11.2.1.4. Multi-Center Studies

Primary, secondary and tertiary continuous endpoints were adjusted for possible site differences. Data from all small sites (<4 participants per arm) were combined to form a single site. There were 10 sites that enrolled 81 participants in the Efficacy Phase of this trial. Due to the low enrollment and the 2:1 randomization, only 3 of 10 sites enrolled 4 or more participants in each treatment arm. Therefore there were 4 site groups with 24, 15 and 9 participants in the single site groups and 33 participants in the combined site group.

11.2.1.5. Multiple Comparison/Multiplicity

The primary efficacy hypotheses associated with the Efficacy Phase and the Discontinuation Phase (Section 11.2.3.3) were independent hypothesis tests. The Efficacy Phase, after adjustment for interim analyses described above, used a final significance level of 0.04999 and the significance level for the primary efficacy endpoint in the Discontinuation Phase was 0.05.

The secondary and tertiary efficacy endpoints were considered supportive of the primary efficacy endpoint. A significance level of 0.05 was used for these analyses also; no adjustments for multiplicity were made.

11.2.1.6. Use of an “Efficacy Subset” of Participants

The primary efficacy endpoints for the Efficacy Phase and the Discontinuation Phase were analyzed using all safety participants who had data from baseline and at least 1 visit. This includes all 81 participants enrolled in the Efficacy Phase and all 31 participants enrolled in the Discontinuation Phase.

Primary Endpoint sensitivity analyses were performed on the per protocol populations and revised per protocol populations.

11.2.1.7. Active-Control Studies Intended to Show Equivalence

This section does not apply to this report.

11.2.1.8. Examination of Subgroups

Due to small sample size, subgroups were not analyzed.

11.3. Drug Dose, Drug Concentration and Relationship to Response

These analyses were not done.

11.4. Drug-Drug and Drug-Disease Interactions

Potential interactions were not investigated in this study.

11.5. Overall Conclusion – Efficacy

These data provide substantive evidence supporting the observation that lithium has efficacy in the treatment of children and adolescents with bipolar I disorder.

The initial randomized placebo-controlled phase demonstrates that lithium has acute efficacy in the treatment of manic/mixed states in this patient population.

This work also provides convincing evidence that lithium also is effective as a longer-term treatment. Despite the relative small sample sizes, the extant data all point to the fact that lithium has efficacy as a maintenance treatment in pediatric bipolar I disorder.

The acute and longer- term efficacy of lithium in this patient population is further supported by the observation that participants who are treated with placebo subsequently have symptomatic improvements after lithium is initiated. Furthermore, participants who are treated with lithium in a blinded fashion have durable symptomatic response when treated with lithium over an extended period of time.

When considered together, lithium consistently demonstrated both acute- and long-term efficacy in this patient population.

Image ch10f2
Image ch10f3
Image ch10f4
Image ch10f5
Image ch10f6
Image ch10f7
Image ch10f9
Image ch10f10
Image ch10f12
Image ch10f11
Image ch10f13
Image ch10f14
Image ch10f15
Image ch10f16
Image ch10f17
Image ch10f18
Image ch10f19
Image ch10f20
Image ch10f21
Image ch10f22

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (14M)

In this Page

Other titles in this collection

Recent Activity

ClearTurn OffTurn On

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...