7.1. Phenomenology of Child and Adolescent Mania

Unfortunately, there are no large epidemiological studies examining the prevalence and incidence of PBD as defined by current nosology. There are no biological tests that can confirm or refute the diagnosis of PBD. Therefore, making or eschewing this diagnosis in children and adolescents is based on clinical presentation. Pediatric BD can present as a very complicated clinical picture, often involving symptoms that are different from that of adult bipolar disorder. Recent data indicate the possibility of very early onset (i.e., between 4 and 7 years of age) (Wozniak et al., 1995; Geller et al., 1998b; Findling et al., 2001). In one community school survey of older adolescents (aged 14-18 years), the lifetime prevalence of bipolar I (full mania, full depression) and bipolar II (hypomania, full depression) disorder was 1% (Lewinsohn et al., 1995). An additional 5.7% of this sample was reported to have experienced a distinct period of abnormally persistent elated or irritable mood without meeting full criteria for bipolar disorder (Lewinsohn et al., 1995).

Despite the dearth of epidemiologically ascertained data, pediatric bipolar illness is commonly seen in clinical settings and pediatric participants with mania have been reported to have severely impaired psychosocial functioning, as well as high rates of psychopathology and psychiatric hospitalization (Pavuluri et al., 2005).

Wozniak et al. (1995a) and Biederman et al (1996) reported severely impaired psychosocial functioning, as well as increased rates of psychopathology and psychiatric hospitalization. Children and adolescents with BD had little or no inter-episode recovery (Findling et al., 2001) in symptoms or in functioning (Wozniak et al., 1995a; Geller et al., 2004). Psychosis is a common phenomenon reported in up to 60% of cases (Geller et al., 2002). Pediatric BD severely impairs a child’s developmental and emotional growth. It is associated with an alarming suicide rate, school failure, aggression, risk-taking behaviors such as sexual promiscuity, and substance abuse (Carlson and Kelly, 1998; Geller et al, 2004). Tragically, pediatric BD is frequently misdiagnosed by clinicians, resulting in inadequate management (e.g., inappropriate stimulant medication) and a worsening of the disorder (DelBello et al, 2001). Appropriate first line treatments such as Li⁺ need to be tested for safety and efficacy in the form of clinical trials to address this serious public health concern.

Lithium may be prescribed as a tablet or capsule (lithium carbonate) or as a liquid (lithium citrate). This protocol will study the lithium in the form of lithium carbonate. Lithium carbonate is a white, light alkaline powder with the molecular formula Li₂CO₃. Li₂CO₃ has a molecular weight of 73.89.

Although the specific mechanism of action is not known, lithium has been shown to have numerous and complex effects on intra-neuronal functioning (Quiroz et al., 2004). As lithium’s ionic radius is similar to that of magnesium, lithium inhibits some enzymes through competition for this common cofactor. Lithium also significantly inhibits several phosphomonoesterases that are magnesium-dependent. Lithium inhibits phosphoglucomutase and glycogen synthase kinase-3. Furthermore, lithium has been reported to affect cyclic adenosine monophosphate-mediated intra-cellular signaling. Unfortunately, the relationship between these biochemical actions and lithium’s therapeutic effects has yet to be clearly elucidated.

7.2. Efficacy of Lithium

Since the discovery of Li⁺ as a treatment for mania in adults nearly half-century ago, there have been numerous studies establishing its efficacy in adult BD (Prien et al., 1972; Segal et al., 1998; Sachs et al., 2002; Tohen et al., 2002). In contrast to adult studies, there are no double blind, controlled studies examining the efficacy of Li⁺ in pediatric acute mania. The most recent study was a maintenance study using a double dummy design comparing Li⁺ with divalproex sodium (DVPX), yielding the preliminary evidence that Li⁺ is a promising treatment and is comparable to DVPX (Findling et al., 2005). The dearth of information on Li⁺ and the current state of science highlight the need to urgently obtain child-specific data on Li⁺. Children are not just smaller adults, and it may be inappropriate to extrapolate psychopharmacological knowledge from adult studies and to presume that the data from these studies apply to participants (Wiznitzer and Findling, 2003). Just as the antidepressant effects of tricyclic antidepressants are age-dependent in the treatment of major depression (Findling et al 2004), unless otherwise demonstrated, one cannot assume that Li⁺ is either efficacious or safe in pediatric BD. There has been a recent increase in the usage of unproven psychotropic medications in children (Zito et al., 2000) with an assumption that Li⁺ may be ineffective in youth despite preliminary evidence to the contrary.

Li⁺ is considered to be a first line mood stabilizer for acute mania, hypomania, depression, and prevention of relapse in adults, and in youth as recommended by the recently published treatment guidelines for pediatric BD (Kowatch et al., 2005). Like other psychotropic agents used to treat pediatric mania, it is yet to be adequately tested using a double-blind, randomized placebo-controlled study design and an adequate sample size. It is also vital to determine the safety profile of Li⁺ given its complex profile of protective and possibly pernicious effects.

7.3. Efficacy Studies of Lithium in Pediatric Bipolar Disorder

There also have been more studies on the use of lithium in bipolar children and adolescents than any other mood stabilizer in this participant population. However, the majority of these studies were carried out in large mixed samples, with a range of diagnoses including bipolar, ADHD, and conduct disorder without controls (DeLong and Aldershof 1987) or with variable assessment protocols in small samples without placebo control groups. Yet, all of these studies suggest a beneficial effect for lithium in many of these child and adolescent participants with bipolar disorder. One study compared lithium, divalproex sodium, and carbamazepine in the acute treatment of youth with bipolar I or II mixed or manic episodes (Kowatch et al., 2000) and found that all three agents showed large effect sizes. More than a third of those treated with lithium showed more than a 50% reduction in their baseline Young Mania Rating Scale (YMRS) scores.

There have been six published controlled trials of Li⁺ in bipolar children and adolescents. Of these six studies, four (Gram and Rafaelsen 1972, Lena 1979, McKnew et al 1981, DeLong and Nieman 1983) used a cross-over design and one used a discontinuation design (Kafantaris et al., 2004). The average number of participants in each of these studies was relatively small (Kafantaris et al. had 40 randomized), and response rates ranged from 33-80%. In the only well controlled, prospective study, which utilized current diagnostic criteria for BD, Geller and colleagues administered Li⁺ in a double-blinded and placebo-controlled fashion to 25 adolescents with a bipolar spectrum disorder (some had depression with a family history of bipolar disorder) and a secondary substance dependency disorder (most had alcohol and marijuana dependence) (Geller et al 1998a). In this study, the adolescent’s diagnosis of BD preceded their substance abuse by several years. After 6 weeks of treatment, those participants treated with Li⁺ showed a significant decrease in their substance abuse and a significant improvement in their global assessment of functioning. In the placebo group, 8.3% of participants were judged to be responders. This study was the first placebo-controlled study to clearly demonstrate the efficacy of lithium carbonate in the treatment of bipolar adolescents with co-morbid substance abuse. The small sample size and co-morbidity limit the generalizability of this study to participants with acute mania.

7.4. Neurocognition and Lithium

7.5. Lithium Maintenance Therapy

Prior to this study, there had been no long-term trials of lithium in children and adolescents. Controlled studies of lithium in adults with bipolar disorder support a prophylactic effect for prevention of mania or depression.

Several studies have documented the relationship between lithium serum concentrations and the prevention or delay of symptom relapse during maintenance therapy. A recent meta-analysis evaluated the benefits and tolerability of long-term lithium therapy for the prevention or delay of symptom relapse in adult bipolar participants (Geddes et al., 2004). Lithium was more effective than placebo in preventing new episodes of a mood disturbance. Unfortunately, there have been no placebo-controlled studies that have evaluated the durability of lithium efficacy over the long-term.

7.6. Adjunctive Treatments with Lithium

For many individuals with bipolar disorder, monotherapy with lithium may not be sufficient (Bauer et al., 1999; Keck, 2004; Kowatch et al., 2005). Although lithium has been found to be effective for the acute treatment of manic and depressive episodes in adult BD, at least one-third of individuals have less than an adequate response (McElroy and Keck, 2000).

7.7. Ensuring Safety During Treatment with Li⁺

Common side effects of Li⁺ are thirst, enuresis, nausea, vomiting, upset stomach, acne, and tremors. While these adverse events are not medically serious, they can decrease adherence. Appropriate formulation and dosing for lithium need to be established to minimize these side effects. Participants need to be monitored carefully for potential hypothyroidism, renal dysfunction and cardiac conduction defects. Two major concerns are risk for teratogenicity in pregnancy and lithium’s narrow therapeutic index that can lead to Li⁺ related toxicity. Drug monitoring is useful for lithium given the inter-individual variability in dose response, narrow intra-individual concentration index, broad inter-individual distribution of sensitivity and tolerance, and lithium’s dependence on stable kidney function for its successful and safe use. In adults, the therapeutic range is suggested to lie between 0.3 and 1.3 mEq/L, with 1.5 mEq/L representing the lower limit of risk for intoxication (Amdisen, 1980). In the pediatric population, side effects are reported to have been observed at doses of 26 to 52 mg/kg per day and at serum levels of 0.65 to 1.37 mEq/L (Hagino et al., 1995). Higher dose and higher serum levels seem to be directly linked to side effects, and may be especially prevalent during the first week of treatment. Therefore, it is critical to monitor serum levels of Li⁺ as well as to dose the participants appropriately. Given the high risk for suicide, it is critical to ensure suicide precautions are in place with regular checks to prevent completed suicide and self-injurious behavior. It is also equally important that participants are carefully monitored in order to prevent overdosing on Li⁺ either by intention or by accident to avoid Li⁺ related toxicity.

7.8. BPCA Prioritization

The Best Pharmaceuticals for Children Act (BPCA) mandates the National Institutes of Health (NIH) to prioritize therapeutic areas in critical need for pediatric labeling, sponsor pediatric clinical trials, and submit these data to FDA for consideration for labeling changes. This study was conducted under the auspices of BPCA and hence in accordance with Section 409I of the Public Health Service Act; as such, the results from this research will be submitted to the FDA for review and used in negotiated labeling changes.

The NICHD prioritized lithium as a drug that needed further study in pediatric populations for treating bipolar disorder in its 2003 Federal Register notice (FR 68:13 January 21, 2003). FDA then issued a Written Request (Appendix 16.1.13) to satisfy the pediatric labeling gap for lithium. The labeling gap was for dosing, pharmacokinetic (PK) parameters, effectiveness, and safety of this drug in children. Two (COLT1 and COLT2) of the three clinical studies as outlined in the WR were funded by NIH and were conducted under IND 73,844. The first study NICHD-2005-07-1 (COLT1), titled Pediatric Pharmacokinetic and Tolerability Study of Lithium for the Treatment of Pediatric Mania followed by an Open Label Long Term Safety Period, Discontinuation Phase, and Restabilization Period was completed in 2008 (Findling et al., 2010; Findling et al., 2011; Findling et al., 2013). An initial Clinical Study Report (CSR) was submitted on February 16, 2010 (IND 73,844 / A0017), and a revised CSR was submitted November 12, 2015 (IND 73,844 / A0033). The dosing and tolerability information obtained from the first study were used in the design of the second study. This is the CSR for the second study (COLT2).

7.9. Developing an Evidence-Based Dosing Strategy for Lithium in the COLT study

A key goal of the Collaborative Lithium Trial 1 (COLT1) was to develop an evidence-based dosing strategy for lithium that could be employed in the acute randomized controlled trial described herein (Findling et al., 2008). There were 3 treatment arms in COLT1. The first 2 arms (Arm 1 and Arm 2) compared the safety and tolerability of a starting lithium dose of 600 mg versus 900 mg per day. Arm 3 included a starting dose of 900 mg per day. Whereas participants in Arms 1 and 2 had their lithium doses increased, based on effectiveness and tolerability, once weekly, participants in the third Arm (Arm 3) of this trial had their lithium dosing increased twice weekly. Thus the key purpose of Arm 3 was to determine whether dosing increments could occur more frequently than once weekly.

A key parameter considered in the COLT1 study was safety. Table 7-1 is a listing of participants who withdrew from COLT1 due to adverse events.

Table 7-1

Participants Who Withdrew From COLT1 Due to Adverse Events.

As can be seen, all three of the dosing paradigms employed in COLT1 were generally well-tolerated.

In the Collaborative Lithium Trial 2 (COLT2) study, the starting dose of lithium was either 600 mg/day or 900 mg/day. Participants weighing less than 30 kg received 600 mg/day as their starting dose. All other participants began lithium therapy on 900 mg/day. These starting daily doses were determined based on the results of the COLT1 study, as well as prior studies done by others (Weller et al., 1986). Those trials demonstrated that initial lithium doses of 30 mg/kg/day were generally well-tolerated.

Another key aspect of lithium treatment is confidence that an appropriate dosing interval is employed. In adults, the immediate release formulation of lithium is oftentimes dosed thrice daily. Thrice daily dosing was also employed in the COLT1 study. Results from the first 30 participants dosed with lithium in COLT1 found that the terminal elimination half-life and allometrically scaled clearance and volume of distribution from population PK analysis for participants were within the range of literature reports from adults. Based on these data, and taking into consideration that lithium is available in 300 mg capsules, the following dosing strategy was developed for the COLT2 study: Participants who were treated with 900 mg or more of lithium per day were dosed thrice daily. Participants who received 600 mg/day of lithium had their treatment administered twice a day. Another key component of lithium treatment is the frequency with which dose increases of lithium can occur. Lithium has a narrow therapeutic index and mania is a serious condition. Thus, it is important to achieve optimal therapeutic levels of lithium both safely and with celerity. In the COLT1 study, participants who participated in Arm 3 were able to have their dose of lithium increased twice weekly (e.g. approximately every 3 to 4 days) until the participant experienced intolerable side effects; the participant’s symptoms substantively abated; or the participant achieved a blood level of 1.4 mEq/L. For the 18 participants who participated in Arm 3 and provided reliable data, 11 (61.1%) had upward dosing adjustments made during the middle of the first week of treatment. However, only 5 (27.8%) had upward dosing increases in the middle of the second week of treatment that were subsequently maintained. Based on this result, participants in the COLT2 study could have their dose of lithium increased during the middle of Week 1 and at the end of each subsequent week in order to more efficiently achieve therapeutic lithium doses.

A final key component in developing a rational dosing strategy for lithium carbonate in this participant population was defining an appropriate therapeutic range for lithium concentrations. Historically, studies of lithium in children have focused on achieving lithium concentrations that have been found to be effective in adults (e.g. 0.8-1.2 mEq/L). This was based on the assumption that the therapeutic range for lithium in the young was the same as the therapeutic range for lithium in adults. Whether or not the therapeutic range for lithium in participants was in fact similar to that described in adults had yet to be tested. As noted in the COLT1 study, participants were allowed to have their dose of lithium increased until either they achieved symptom amelioration; experienced dose-limiting side effects; or had their blood level exceed 1.4 mEq/L. Disposition of participants in COLT1 Phase 1, and their last measured lithium concentration and dose, are listed below in Table 7-2. Of the original 61 participants enrolled in COLT1 Phase 1, three participants were considered unreliable reporters and therefore their data have not been included. Additionally, one participant who withdrew consent was not included below as that participant did not provide any data.

Table 7-2

COLT1 Participant Disposition Plus Last Measured Lithium Concentration and Dose.

As shown above, using a maximum concentration of 1.4 mEq/L, above which further dose increases could not occur, appeared to be associated with an appropriate degree of safety. For that reason, the COLT2 study also used 1.4 mEq/L as the lithium concentration above which dose increases could not occur.

7.10. Structure of the Clinical Study Report

This was a multiphase study and involved the following phases of study for each participant: Efficacy Phase, Long-term Effectiveness Phase, Discontinuation Phase, and Restabilization Phase. See Section 9.1 for a description of the different phases.

For the efficacy analyses (Section 11), several assessment scales and subscales were analyzed as primary, secondary, tertiary, and descriptive efficacy endpoints. We have approached the Efficacy Analysis by splitting the section into the four phases and grouping all the analyses pertaining to each phase. Our rationale for this structure is to enable the reviewer to look at all the efficacy analyses together in order to understand the effect of the drug during each phase.

We have taken a similar approach for the safety analyses (Section 12) based same rationale as the efficacy section; all safety analyses for each phase have been grouped together under four subsections, one for each phase.