Lithium Plus An Atypical Antipsychotic Was More Effective Than Valproate Plus An Atypical Antipsychotic In One Study, But Not Another
Evaristo Nieto et al. of Spain presented a poster about the naturalistic study of the efficacy of acute treatment of manic inpatients with lithium and valproate at the 2014 meeting of the International College of Neuropsychopharmacology. In the lithium group, all patients were treated with lithium and oral antipsychotics (N=85). In the valproate group, all were treated with valproate and oral antipsychotics (N=92). Outcome was measured using scores on scales for mania and for general functioning (the YMRS and the CGI-S). The atypical antipsychotic was typically olanzepine or risperidone.
Nieto et al. found that the mean change in CGI scores from baseline to the day of discharge was significantly higher in the lithium group (-2.84 versus -2.6), and concluded that, “Although it is used in more severe cases, treatment of manic inpatients with lithium associated with antipsychotics is more effective than treatment with valproate associated with antipsychotics.”
However, W.M. Bank et al. came to the opposite conclusion in a Korean study. Bank et al. “compared the 1-year rehospitalization rates of first-episode bipolar manic patients?who were discharged while being treated with lithium or valproate in combination with an?atypical antipsychotic….The rehospitalization rate was 17.3% during the 1-year follow-up period.”
Bank et al. found significantly higher rates of rehospitalization in the lithium (23.1%) compared to the?valproate (13.3%) group using the Kaplan-Meier formula for estimations.
Lithium is known for protecting neurons by inducing neurotrophic factors and inhibiting cell death factors. In a new study, other mood-stabilizing drugs had similar neuroprotective and neurotrophic effects on cultured neurons from the hippocampus.
At the 2014 meeting of the International Society for Bipolar Disorders, CH Lee et al. presented evidence that lithium, carbamazepine, valproic acid, and lamotrigine all increase the outgrowth of dendrites from these cultured neurons. Therapeutic levels of these drugs increased the production of proteins like brain-derived neurotrophic factor (BDNF), postsynaptic density protein-95 (PSD-95), neurolignin 1 (NLG 1), beta-neurexin, and synaptophysin. However, so far only lithium has been shown to increase the volume of the human hippocampus as measured with MRI.
A 2012 study by Kemp et al. in the journal Bipolar Disorders found that lamotrigine added to combination treatment with lithium and valproate was no more effective than placebo in patients with rapid cycling bipolar disorder. Only 14% (19 out of 133) of rapid cycling patients stabilized upon initial treatment with the open combination of lithium and valproate, a startlingly low rate. In the next phase of the study, 49 patients who were not stabilized were given adjunctive treatment with either lamotrigine (n=23) or placebo (n=26) on a double-blind basis, but no significant difference was observed.
Editor’s Note: This study has two pieces of not-so-good news. The first is that it was so difficult to stabilize these patients with rapid cycling bipolar disorder. The second is that the add-on of lamotrigine, which is highly effective in the prevention of depressions in bipolar disorder, was in this case no more effective than placebo.
This study again demonstrates that rapid cycling bipolar disorder is difficult to treat, and even the use of three proven mood stabilizers in combination is not always effective. Many doctors would recommend an atypical antipsychotic as the next clinical option.
In pregnant women, exposure of the fetus to the anticonvulsant valproate (VPA or Depakote) is associated with a variety of serious problems that include congenital malformations, developmental delay, and autism.
The major congenital malformations that can result from valproate exposure include spina bifida, which results in lifelong paralysis of the child’s lower limbs.
Developmental delay resulting from valproate exposure can cause an average loss of 9 IQ points compared to children exposed to other anticonvulsant drugs in utero. The effects appear to be in part dose-related and dependent on the intensity of combination treatment with other agents. These deficits were originally seen in children at 3 years of age and were shown to persist in six-year-olds according to an article by Meador et al. this year in Lancet Neurology.
Now in addition, fetal exposure to valproate has been liked to autism and related disorders in an 11-year longitudinal study published this year in the Journal of Neurology, Neurosurgery and Psychiatry. A diagnosis of a developmental disorder occurred in 17% of children whose mothers were on valproate as opposed to 2% whose mothers were on carbamazepine and 7% whose mothers were on lamotrigine.
Neurologists are increasingly recommending that all women of childbearing age who are on a treatment regimen including valproate be treated with folic acid and vitamins B6 and B12, in the hopes that these might mitigate valproate’s effects on the fetus in the case of an unplanned pregnancy. The effectiveness of these vitamins has not been directly demonstrated. However, the study by Meador et al. did show that children of mothers who took prenatal folic acid supplements had IQs on average 7 point higher than children whose mothers did not. The benefit was seen only when mothers were already taking folic acid when they became pregnant and was not observed in children of mothers who began taking it after the first trimester.
Women of childbearing age should avoid valproate and if this is not possible, they should carefully protect themselves against an unwanted pregnancy. Women with bipolar disorder are 3.9 times more likely to have unplanned pregnancies than women of similar age in the general population. These data suggest the importance of careful education about birth control in patients with bipolar illness so that pregnancies can be planned for periods of good health and so that appropriate pharmacological measures can be taken.
A 30-year observational study published by Andrew Leon and colleagues in the American Journal of Psychiatry has found that anticonvulsants used in epilepsy and for bipolar depression (carbamazepine, lamotrigine, and valproate) do not increase suicidal behavior in bipolar patients.
Editor’s Note: The FDA gave a warning in 2009 that these anticonvulsants were associated with suicidal ideation. This was based on studies of a mixed group of psychiatry and neurological patients in acute placebo-controlled studies, where suicidal ideation is typically a reason for exclusion from the study. Leon et al. used more powerful longitudinal methods to compare the risk of suicidal ideation in individuals taking and not taking anticonvulsants and found no such increase in suicidal behavior.
This is like the FDA warning for antidepressants and suicide, which was based on data from placebo-controlled clinical trials in acute depression (where suicidal patients are excluded). When investigators used the same longitudinal methods as Leon et al. in the anticonvulsant study, they found that antidepressants actually reduced suicidal behavior by 30%.
The bottom line is that the use of anticonvulsants for bipolar disorder should not be discouraged based on the FDA warning about suicidal ideation in mixed neurological and psychiatric patients. In bipolar patients, anticonvulsants do not increase the risk of suicidal behaviors, i.e. suicidal acts or completed suicides.
When Added to Valproate, Memantine Increased HDLs (“Good” Cholesterol) But Did Not Enhance Effectiveness of Treatment
R.B. Lu and S.Y. Lee reported in a poster at the 5th Biennial Conference of the International Society for Bipolar Disorders in 2012 that adding memantine (Namenda) to treatment with valproate (Depakote) was associated with increases in high-density lipoproteins (HDLs) or “good” cholesterol in bipolar II depressed patients. However, the combination was no more clinically effective than valproate alone in treating the patients’ bipolar illness, as had been hoped.
Editor’s Note: These data on memantine’s failure to improve patients’ bipolar illness when used as an adjunct to valproate contrast with those of Amit Anand et al., who reported in 2012 that memantine was a partially successful adjunctive treatment when added to ongoing treatment with lamotrigine. This combination of lamotrigine plus memantine was associated with faster and more robust antidepressant effects than the combination of lamotrigine plus placebo in patients with bipolar depression. This effect was significant in the first four weeks of the study as the dose of memantine was slowly increased from 5mg/day to 20mg/day, but not over the last four weeks of treatment at 20 mg/day.
The data of Anand et al. makes theoretical sense. Since lamotrigine inhibits the release of glutamate and memantine inhibits the actions of glutamate at the NMDA receptor, the two together might produce additive decrements in glutamatergic actions through two different mechanisms. In contrast, valproate is more closely associated with increases in GABAergic mechanisms, and this may explain why its effects on bipolar disorder were not improved by the addition of memantine.
At another symposium at the annual meeting of the American Academy of Child and Adolescent Psychiatry, Bob Kowatch of Ohio State University discussed a controlled trial of valproate, risperidone, and placebo in children 3 to 7 years of age (average age 5.5) with a diagnosis of bipolar I disorder and a Young Mania Rating Scale score (YMRS) greater than 20 at baseline. All of the children were severely ill with an average Clinical Global Assessment of Severity (CGAS) score of 44. Seventy-six percent had comorbid attention deficit hyperactivity disorder (ADHD) and 15% had an anxiety disorder. Valproate doses started at 10mg/kg and were increased after 4 days to achieve blood levels of 80 to 100µg/ml. The average dose of valproate was 300mg/day and the average blood level was 88 µg/ml. Risperidone was started at 0.25mg and increased as needed. The average dose of risperidone was 0.5mg per day.
On the main outcome measure of decrease in the YMRS score risperidone was substantially more effective than placebo, while valproate showed only marginal nonsignificant effects. However on the Clinical Global Impressions (CGI) scale for improvement in illness, risperidone showed 87% response, valproate 75% response, and placebo no response. In terms of 50% reduction in the YMRS score, this endpoint was achieved in 88% on risperidone, 50% valproate, and 15% on placebo.
Weight gain was mild on valproate and substantially more on risperidone. Risperidone was also associated with increases in insulin and prolactin.
The effect size (the size of the change the drug brought about in this study, which is calculated by dividing the mean difference between the experimental group and the control group by the standard deviation) for risperidone was extraordinarily large (3.58); very large for valproate (1.66), and moderate for placebo (0.56). The odds of getting well were 5 times greater than placebo for risperidone and 1.9 times greater than placebo for valproate.
Editors note: These data in very young children (aged 3 to 7) resemble other controlled data in the literature about the treatment of older children and adolescents, indicating a superiority of atypical antipsychotics over placebo and a greater magnitude of effect achieved with atypicals than with valproate. Based on these new data and the Federal Drug Administration (FDA) approval of several atypical antipsychotics for children with bipolar illness from ages 10 to 17, Dr. Kowatch recommended a new treatment algorithm for childhood onset bipolar disorder. Read more
Researcher C. Lewis reported in two posters presented at the Ninth International Conference on Bipolar Disorder (ICBD) in 2011 that in a study of patients treated with valproate, some increases in ammonia levels occurred. This condition, hyperamonaemia, was identified in 31 patients among those treated between 2005 and 2009 at the Cleveland Clinic in Cleveland, Ohio. High levels of ammonia are associated with a flapping tremor and, in some cases, encephalopathy with confusion, psychiatric symptoms, and motor incoordination.
The recommended management for hyperamonaemia is discontinuation of valproate and use of lactulose, a synthetic sugar that can lower ammonia levels. These approaches were not always used. Another option for patients who require valproate treatment is to supplement the drug with carnitine, which is available as a nutritional supplement. Lewis reported success in three such cases.
Editor’s Note: Patients on valproate presenting with a gross flapping tremor of the hands, confusion, or motor imbalance should be tested for hyperamonaemia and treated accordingly.
According to two studies in 2011, lithium is more effective in treating episodes of bipolar disorder than valproate, while the drugs may be equally effective in reducing suicidality among bipolar patients.
The first study, of more than 4000 Danish patients with bipolar disorder, found that patients taking lithium had fewer hospital visits and were less likely to need new medications than those taking valproate. Patients taking lithium had fewer admissions to a hospital for any type of episode. In addition, patients taking valproate had a higher rate of switching to or adding on treatment with antidepressants, antipsychotics, or anticonvulsants than those taking lithium. The study included up to 12 years of follow-up with the patients and is the largest study with the longest period of follow-up of patients taking valproate or lithium to date. Results were published by Lars Kessing et al. in the British Journal of Psychiatry in July 2011.
In the other study, a randomized controlled trial of 100 patients with bipolar disorder who had attempted suicide at least once in the past, Maria Oquendo of Columbia University found that there were no significant differences in number of suicide attempts, hospitalizations for suicide attempts, or time to a new attempt between patients taking lithium and those taking valproate over a follow-up period of 2.5 years.
Forty-five suicide events, which included attempts, hospitalizations, and changes to medication in response to suicide plans, were experienced by 35 patients (16 who were taking lithium and 19 who were taking valproate). Eighteen suicide attempts were made by 6 patients taking lithium and 8 taking valproate. There were no suicide completions during this study, which was published in Volume 168 of the American Journal of Psychiatry in 2011.
Editor’s Note: Suicide attempts are much more common than completed suicides. It appears that the second study was not large enough or long enough to detect differences in the rate of completed suicides. Older naturalistic studies suggest that treatment results in low suicide rates and that in patients who stop treatment with lithium, the rate of suicide attempts and completion increases dramatically. This is another reason for good responders to treatment regimens that include lithium to continue taking their medications.
Shannon Stepan, Eric Peselow and Nunzio Pomara from Maimonides Medical Center in Brooklyn, NY, have analyzed naturalistic observations of long-term maintenance treatment of bipolar disorder with valproic acid, lithium, and carbamazepine and found that patients on lithium went much longer before experiencing an episode of mania or depression than patients taking carbamazepine or valproate.
The team followed 225 outpatients for up to 124 months, or until they had a manic or depressive episode or dropped out of the study during a well phase. Ninety-eight patients took lithium, 78 took valproate, and 50 took carbamazepine. Fifty-two percent of the participants dropped out of the study during a well phase.
One hundred three patients (45.8%) had either a manic or depressive episode during the study. This included 36.7% of the patients taking lithium, 55% of patients taking valproate, and 50% of patients taking carbamazepine. Median time until a first episode was 45 months for the entire sample, 36 months for those patients on valproate, 42 months for those on carbamazepine, and 81 months for those on lithium. A statistical analysis known as a Cox regression model indicated that patients taking valproate had a significantly higher risk of having a manic or depressive episode than those taking lithium.
Editor’s note: These naturalistic data are highly consistent with a number of more controlled clinical studies. In particular, the BALANCE study by Geddes et al. (2010) reported that lithium was superior to valproate on most outcome measures in a two-year randomized study, and that the combination of lithium and valproate was significantly better than valproate alone. Read more