Adjunctive Valproic Acid in Management-Refractory Hyperactive Delirium: A Case Series and Rationale

Delirium is the most commonly encountered psychiatric diagnosis in the general hospital setting and significantly increases patients’ hospital stay, morbidity, and mortality.¹ It causes considerable distress to patients and their families and is associated with later cognitive decline. Delirium is mediated by several mechanisms, including neuronal aging, inflammation, oxidative stress, and dysregulation in cellular signaling and secondary messenger systems, all of which lead to neurotransmitter imbalance. Deficiency in acetylcholine and melatonin, excessive dopaminergic and glutamatergic activity, and alterations in γ-aminobutyric acid (GABA), serotonin, and histamine have all been implicated in its development.¹

Although antipsychotics are often considered a first-line pharmacological treatment, there may be multiple contraindications (e.g., akathisia, movement disorders, QTc prolongation).¹ Valproic acid (VPA) is one potential alternative, alone or as an adjunct, in the management of hyperactive and mixed-type delirium. VPA is available intravenously and orally, which allows for greater flexibility in administration, especially in agitated, delirious patients in need of rapid behavioral control.

VPA’s mode of action involves many of the neurotransmitters involved in the development of delirium. VPA has antiglutamatergic and N-methyl-d-aspartic acid (NMDA) receptor antagonistic activity.^2,3 It potentiates GABA effects, induces expression of melatonin receptor,⁴ and decreases dopamine in the striatum⁵ and basolateral complex of amygdala⁶ while increasing dopaminergic activity in the hippocampal and prefrontal areas.⁷ In addition, VPA increases acetylcholine efflux in rat hippocampus.⁷ Finally, VPA has anti-inflammatory and antioxidant properties²; exerts transcriptional effects through epigenetic modulation of gene expression,² including histone deacetylase inhibition; and has beneficial effects on the kynurenine pathway.⁸

Currently, the data on VPA’s efficacy in delirium are limited to only two case series comprising a total of eight patients.^9,10 In addition, there are limited data on VPA’s usefulness in the management of agitation in patients with dementia,¹¹ traumatic brain injury (TBI),¹² alcohol withdrawal, and corticosteroid-induced mania.¹³ Finally, emerging data support the use of VPA for neuroprotection in acute central nervous system injuries because of its anti-inflammatory, antiapoptotic, and neurotrophic effects.¹⁴

Methods

We conducted a retrospective chart analysis of hyperactive delirium patients treated by the psychosomatic medicine service at our institution over a period of 13 months using VPA as an adjunct in cases refractory to conventional antipsychotic-based therapy. Identified patients were diagnosed with delirium by a psychiatrist using DSM-IV-TR criteria.¹⁵ A hyperactive motor subtype of delirium was identified according to Liptzin criteria.¹⁶ Patients younger than 18 years of age and those in active alcohol withdrawal at the time of the initial consult were excluded. Electronic medical records were reviewed for patient demographics, clinical presentation, medications, and mental status examinations.

Serum aspartate aminotransferase levels, alanine aminotransferase levels, hemoglobin levels, platelet counts, and international normalized ratio values before and after initiation of VPA were extracted. VPA levels were recorded if available.

The electronic medical records of the primary medical, nursing, and consulting psychiatry teams were reviewed to determine whether patients met DSM-IV-TR criteria for delirium on a daily basis and to characterize the degree of their agitation.

The primary outcome was resolution of delirium; secondary outcomes included resolution of agitation associated with delirium and VPA-related adverse effects.

The study protocol was approved by the Stanford University Institutional Review Board (IRB #25647).

Results

Fifteen patients with 16 total episodes of hyperactive delirium meeting DSM-IV-TR and Liptzin criteria for a hyperactive motor subtype were identified¹⁶ (see Table 1 for brief descriptions of these cases). Fourteen patients were male, and the average age was 51.8 years (SD = 15.7, range = 25–87). Most patients were treated in the intensive care unit (ICU). Of note, one patient had dementia (patient #7), one patient had newly acquired TBI (patient #9), four patients were treated with high-dose corticosteroids (patient #1 for immunosuppression after heart transplantation; patient #5 for intracranial swelling; patient #12 for immunosuppression after lung transplantation; patient #13/14 for myasthenia gravis), and one patient exhibited residual alcohol withdrawal after being weaned off dexmedetomidine. Based on a chart review, delirium started an average of 5.5 days prior to VPA initiation (SD = 4.6, range = 1–17). The etiologies of delirium included postoperative status, hypoxia, infection, acute renal injury/uremia, intracranial pathology, opiate use, benzodiazepine use, corticosteroid use, and, most frequently, a combination of these factors.

In most cases, prior to a psychiatric consult, the primary team had tried multiple medications in an attempt to control agitation associated with hyperactive delirium, including various antipsychotic and benzodiazepine agents, opiates, dexmedetomidine, and propofol. The primary teams also attempted to address the etiology of each patient’s delirious state (e.g., treatment for infections, reduction of offending medications, dialysis). One subject (patient #9) was treated solely with VPA and received no antipsychotics because of a prolonged baseline QTc. All other subjects failed to adequately respond to antipsychotics, at which point treatment with VPA was initiated by the psychiatry consult team.

After initiation of VPA, complete resolution of delirium according to DSM-IV-TR criteria was documented in 13 cases. The average time from VPA initiation to delirium resolution was 6.2 days (SD = 8.0, range = 1–30). Two patients resolved within 1 day, four resolved within 2 days, two resolved within 3 days, one resolved within 4 days, and four resolved after 9 days or longer (9, 11, 11, and 30 days). Excluding the outlier of 30 days, the average number of days to delirium resolution was 4.2 days (SD = 3.8). The median time to delirium resolution was 3 days.

All study patients, including three patients whose delirium did not resolve completely after VPA initiation, experienced marked improvement in agitation and behavioral management as reflected by documented mini-mental state evaluations; reports of decreased impulsivity, restlessness, and aggression; an ability to discontinue restrains; and a decreased need for sedation. Three patients whose terminal delirium did not resolve despite decreased agitation and improved comfort were transitioned to comfort care, transferred out of the ICU, and eventually passed away because of their primary medical condition. The average dose of standing VPA used on days 1 through 4, when most patients experienced delirium resolution, was 1133–1258 mg per 24 hours (SD = 480–625), administered in two to three divided daily doses.

Meropenem may decrease VPA serum level and efficacy by up to 92%,¹⁷ and indeed, this was observed in three of our patients. In these three cases, a dramatic normalization of the VPA level and effectiveness was observed 24 hours after discontinuation of meropenem.

Regarding adverse effects, two patients developed thrombocytopenia, one of whom experienced gastrointestinal bleeding (patient #12). Of note, the same patient experienced gastrointestinal bleeding 10 days after discontinuation of VPA; therefore, the correlation between VPA and gastrointestinal bleeding remains unclear. There were no statistical differences between the averages for aspartate aminotransferase, alanine aminotransferase, hemoglobin, platelets, and international normalized ratio prior to VPA and following VPA initiation. There were no other adverse effects attributed to VPA.

Discussion

This case series adds to the limited data available and suggests potential benefits of VPA in the treatment of refractory hyperactive delirium. More importantly, the adverse effects were minimal, and overall the agent was well tolerated. Of note, almost one-half of the patients had dementia, TBI, presence of corticosteroids, or residual alcohol withdrawal, conditions in which VPA has been shown to be beneficial.^11–13

As an open-label, nonrandomized case series, our study has some limitations, including the lack of a control group; the heterogeneity of the study group; the concomitant use of multiple pharmacological agents; the retrospective nature of data collection; and the lack of validated scales to measure delirium severity. However, the patients in this study represented some of the sickest patients seen by the psychosomatic medicine service at our institution. All had failed multiple pharmacological interventions prior to VPA initiation. Arguably, the use of polypharmacy prior to VPA initiation, including benzodiazepines and opiates, could have contributed to the severity of delirium. However, introduction of VPA allowed patients to taper off their use of these sedative medications, all of which have themselves been associated with the worsening of delirium. It is also possible that in some cases VPA addressed withdrawal from benzodiazepines, which made weaning off these and other GABA-ergic agents (such as propofol) challenging before the introduction of VPA. As a result, all patients experienced some degree of improvement in behavior and cognition, with 13 patients experiencing complete resolution of delirium.

There are important considerations for the use of VPA. Patients should be selected carefully and monitored for possible adverse effects. VPA significantly increases the risk of neural tube defects and neuropsychiatric deficits in the fetus and thus should not be used in pregnant patients. Thrombocytopenia, macrocytic anemia, and leucopenia have all been described in patients using VPA.³ Thus, we recommend daily monitoring of blood cell counts, especially platelets, in critically ill patients on VPA. The hypothetical risk of bleeding can be increased further in patients on warfarin, as VPA raises warfarin levels via CYP2C9 and CYP2C19 inhibition and protein binding displacement.¹⁸

VPA-associated pancreatitis is thought to be a rare idiosyncratic reaction to VPA use,³ thus administering physicians should monitor for this possibility. Patients might present with acute abdomen, nausea and vomiting, elevated white blood cell count, fever, and an increase in the base deficit. These symptoms should prompt urgent measurement of lipase and amylase and, if indicated, discontinuation of VPA.

The use of VPA may be associated with minor elevation of liver transaminase enzymes and lactate dehydrogenase, usually in the first months of treatment in a dose-dependent fashion. On the other hand, serious VPA-induced hepatotoxicity, which may lead to fulminant hepatic failure, has been reported in 1 out of 20,000 patients.³ This risk may be lower, however, in patients treated with VPA for delirium, as VPA's use in delirium is usually short-lived. Similarly, VPA-induced hyperammonemic encephalopathy (VHE) is rare,¹⁹ although asymptomatic elevations of ammonia may be more common. VHE is mediated by the effects of VPA on the urea cycle and impaired excretion of ammonia.¹⁹ Patients with urea cycle disorders, mitochondrial disorders, and carnitine deficiencies are at higher risk. Thus, it is important to check ammonia levels and consider VHE in any patient whose mental status worsens while on VPA or who appears to be transitioning from a hyperactive to a hypoactive state. If there is concern for VHE, VPA should be stopped and supplementation with l-carnitine should be strongly considered. Of importance, the risk of encephalopathy might be greater in patients who are taking a combination of VPA and other psychotropic medications (e.g., lamotrigine, zotepine, quetiapine), highlighting the need to monitor drug-drug interactions.

Conclusions

We observed a significant improvement in symptoms of hyperactive delirium with the adjunct use of VPA in patients with failed or contraindicated conventional treatment. We suggest that with proper monitoring of carefully selected patients, VPA can be a promising adjunct for treatment of hyperactive delirium, given its actions on multiple pathways implicated in the pathophysiology of delerium. Of course, randomized control trials are needed to determine the true efficacy and side effects of VPA for the treatment of delirium in this critically ill population.