Drug Treatment for Agitation in Alzheimer’s Disease: A Conundrum


Written By

Stephanie Walters, PharmD
Nicole Brandt, PharmD, MBA, BCPP, CGP

Reviewed By

Marty Eng, Pharm.D., BCPP, CGP, CDP
Jeannie K. Lee, Pharm.D., BCPS, CGP


Cummings JL, Lyketsos CG, Peskind ER, et al. Effect of dextromethorphan-quinidine on agitation in patients with Alzheimer disease dementia. JAMA. 2015: 314: 1233-1254.

Managing behavioral health in persons with dementia is an enormous and growing problem. Today more than 5.3 million Americans are living with Alzheimer’s disease (AD) and that number is expected to increase by 30% by 2025.1 Most patients with AD have challenging neuropsychiatric symptoms such as agitation. These symptoms are distressing and often overwhelming for caregivers often resulting in institutionalization and increased healthcare utilization.1 Several medications are used off-label to treat these symptoms including antidepressants, antipsychotics, anticonvulsants, anxiolytics and cholinesterase inhibitors.2,3 Antipsychotics have a black box warning regarding the risk of increased mortality in this population.4,5 Many regulatory bodies, professional associations, and patient advocacy groups have promoted initiatives to decrease the inappropriate use of antipsychotics and encourage the use of alternatives, particularly behavioral interventions.6-11


The combination of dextromethorphan hydrobromide and quinidine sulfate (Nuedexta) has been recently studied for the off-label treatment of agitation in Alzheimer’s disease. The FDA approved Nuedexta in 2010 for the treatment of pseudobulbar affect, a condition that causes uncontrollable and involuntary emotional lability (e.g. extreme laughing or crying).12  Dextromethorphan is believed to modulate emotional lability and related behaviors by inhibiting serotonin norepinephrine reuptake, N-methyl-D-aspartate (NMDA), and nicotinic receptors in the central nervous system. Quinidine was added to Nuedexta solely to increase the amount of available dextromethorphan by inhibiting cytochrome P450 2D6 (CYP2D6) isoenzyme.13


Case reports and anecdotal off-labeled use of Nuedexta for agitation led to the initiation of a randomized controlled trial by Cummings and colleagues for dementia-related agitation.13,14 This phase 2, 10-week clinical trial was designed to assess the change in agitation symptoms using the Agitation/Aggression section of the Neuropsychiatric Inventory (NPI), a validated tool used in many dementia clinical trials as well as clinical practice. This tool was administered at baseline and follow-up at weeks 1, 3, 5, 6, 8, and 10.  The secondary outcomes included changes in total NPI scores and each NPI domain, including NPI Caregiver Distress scores and a composite of the Agitation/Aggression, Motor Behavior and Irritability/Lability domains.13


The study enrolled 220 individuals with probable AD and at least 1 of the following behaviors: aggressive verbal (e.g. screaming, cursing), aggressive physical (e.g. destroying objects, grabbing, fighting) or nonaggressive physical (e.g. pacing, restlessness) behavior.  Eligible patients had behavioral symptoms that were severe enough to warrant pharmacological treatment, scored 4 or higher (moderately ill) on the Clinical Global Impressions–Severity (CGIS) scale for agitation, and had a Mini-Mental State Examination (MMSE) score between 8 and 28 out of 30.13


All subjects were allowed to continue established acetylcholinesterase inhibitors +/- memantine (> 2 months) therapy as well as established (> 1 month) pharmacotherapy with antidepressants, antipsychotics, short-acting benzodiazepines, and/or non-benzodiazepine hypnotics. Of note, the manuscript does not identify the specific medications used by patients within each of these drug classes or the doses. Within this population certain antidepressants (i.e. citalopram) and antipsychotics (i.e. risperidone) may be more effective for the treatment of the behavioral and psychological symptoms of dementia (BPSD) when compared to alternative treatments; thus, creating a possible note-worthy confounding variable.15,16


The trial had an interesting study design with two stages of randomization. The first stage consisted of blinded randomization to the treatment group (n=93) or placebo (n=127) for a total of five weeks. The second stage, weeks six through ten, involved a second randomization in the placebo group only. Responders (n=30) and nonresponders (n=89) were stratified prior to randomization; responders were those with a CGIS score of 3 (mildly ill) or lower for agitation at the end of stage 1. The stratified groups were randomized 1:1 into active treatment (n=59) or placebo groups (n=60) for the remaining five weeks.  This double randomization strategy increased the number of subjects who received active treatment during the study. Within stages 1 and 2, each of the primary and secondary endpoints were assessed at baseline and at five weeks.  The overall results compared baseline to the 10-week time point. A sequential parallel comparison design (SPCD) was utilized for the primary outcome of agitation for all patients in stage 1 and the placebo nonresponders in stage 2.13


With nearly 90% of subjects completing the study, Nuedexta demonstrated a positive effect with statistically significant and clinically meaningful improvements in agitation scores at almost every time point (except weeks 6 and 8). A change in the NPI agitation domain of at least 30% was considered clinically important. Based on the SPCD, the active treatment-only group showed an improvement in agitation by 50.7% at 10 weeks compared to a 26.4% improvement observed in the placebo-only group. A clinically relevant change was further supported by improvements in both patient impression and caregiver distress as assessed by validated tools. Adverse effects were reported by 62.7% in the treatment group and 43.3% of the placebo group.  The most common adverse effects in the treatment group included falls, diarrhea, urinary tract infections, and dizziness. 13


While the study findings were positive and Nuedexta was generally well tolerated, there are several risks that must be considered before using this medication to treat agitation in patients with AD. This combination product relies on a drug-drug interaction to increase the dextromethorphan concentrations through CYP2D6 inhibition.  Polypharmacy is commonplace in this population and several drug classes rely on CYP2D6 metabolism, including antiarrhythmics, beta-blockers, neuroleptics, tricyclic antidepressants (TCAs), and selective serotonin reuptake inhibitors (SSRIs).  Moreover, intrinsic CYP2D6 activity varies widely.  Genetic polymorphisms in CYP2D6 results in sufficient, poor, intermediate or ultra-metabolism. Ultra-metabolism is most commonly seen in individuals of Western European decent. 17 Nuedexta may have a diminished effect or be ineffective in ultra-metabolizers. No data was provided regarding CYP2D6 metabolism status from this trial and pharmacogenomic testing is not widely available in practice yet.13


Table 1 shows known substrates and inhibitors of the CYP2D6 enzyme.  These medications can potentially interact with quinidine.17 Several of these medications are also used in the treatment of BPSD. For example, desmethylcitalopram is an active metabolite of citalopram and escitalopram, an SSRI that has been shown to improve neuropsychiatric symptoms of dementia.15 Likewise, haloperidol, a medication with questionable efficacy for BPSD, also interacts with quinidine. Many people with AD require pharmacotherapy for cardiovascular diseases. Beta-blockers are a substrate of CYP2D6 and the risk of bradycardia and hypotension is increased if given with quinidine.  Furthermore, the risk of bradycardia may also be compounded by possible use of acetylcholinesterase inhibitors (e.g. donepezil).  Thus, close monitoring will be required.


Table 1: Drugs with CYP2D6 Activity


CYP2D6 Substrates




Desipramine Desmethylcitalopram

Dextromethorphan Dihydrocodeine


Ethylmorphine Flecainide



Imipramine Indoramin


Methoxyamphetamine Methoxyphenamine








Perphenazine Propafenone Propranolol








CYP2D6 Inhibitors








In addition to many potential cytochrome P450 interactions, dextromethorphan use increases the risk of serotonin syndrome. Not only are many antidepressants substrates of CYP2D6, thereby increasing their bioavailability when used concurrently with quinidine, but these medications can also increase the serotonergic burden.  Patients on antidepressants and other serotonergic agents (e.g. tramadol) should be closely monitored for the serotonin syndrome during initiation and titration of Nuedexta. Signs and symptoms of serotonin syndrome can include autonomic hyperactivity (e.g. vomiting, diarrhea, high blood pressure), neuromuscular symptoms (e.g. clonus, tremor, rigidity), as well as mental status changes (e.g. increased anxiety, disorientation, delirium) — which may be difficult to distinguish from worsening AD symptoms.


A patient-centered, systematic, and evidence-based approach should be considered when addressing BPSD.18,19  Nonpharmacological options should be utilized first and be tailored to the patient, caregiver, and specific neuropsychiatric issues. Any pharmacologic treatment carries possible risk.  Unfortunately, for the treatment of BPSD and agitation, the risk often outweighs the benefits. Given that Nuedexta has many, many potential drug interactions (both pharmacodynamic and pharmacokinetic), meticulous patient assessment is required before it is considered.  This should include a comprehensive medication review, an analysis of the potential drug-drug interactions (specifically CYP2D6), an assessment of the risk for developing serotonin syndrome, and a careful review of the nonpharmacological and pharmacological modalities that have been used in the past.  A frank and open discussion with the caregiver(s) regarding realistic expectations and potential risks should follow.


If nonpharmacological strategies have been tried and failed, would you consider a trial of dextromethoraphan + quinidine? What factors would you consider when starting this medication? Would you favor other pharmacotherapeutic options first?