At the 139th annual meeting of the American Neurological Association, held October 12-14 in Baltimore, researchers from Avanir Pharmaceuticals in Aliso Viejo, California, presented positive results from two studies of a drug in Alzheimer’s disease. AVP-923 is approved in the United States and European Union for the treatment of pseudobulbar affect (PBA). In this strange phenomenon, people with certain primary neurologic diseases frequently burst into uncontrollable, extended spells of laughter or are overcome by crying. PBA itself is rare in Alzheimer’s disease, but agitation is common. In a new Phase 2 trial, AVP-923, marketed as Nuedexta, reduced agitation scores in patients with Alzheimer’s. In a separate open-label study, the drug dampened the emotional outbursts of laughing and crying that sometimes afflict patients with dementia.

Scientists do not know exactly how AVP-923 works for agitation. It is a stable combination of two old drugs: dextromethorphan hydrobromide, the active ingredient in cough medicines such as Robitussin, and the arrhythmia drug quinidine sulfate. The former is an uncompetitive antagonist of NMDA-type glutamate receptors and an agonist of the sigma-1 receptor, a protein chaperone that resides in the endoplasmic reticulum membranes. The latter blocks cytochrome P450 2D6, a liver enzyme that metabolizes dextromethorphan hydrobromide, increasing its plasma concentration (Cruz, 2013). 

A previous Phase 3 trial reported that AVP-923 improved pseudobulbar affect in people with amyotrophic lateral sclerosis (ALS) and multiple sclerosis (MS) (see Pioro et al., 2010). Trials are ongoing for neuropathic pain and for levodopa-induced movements in Parkinson’s, as well as for autism, migraine, and depression.

To see if AVP-923 treated agitation in Alzheimer’s, researchers at 44 sites, including Jeffrey Cummings of the Cleveland Clinic, Luo Ruvo Center for Brain Health in Las Vegas, enrolled 220 adults aged 50 to 90, all of whom had probable AD with clinical agitation. The researchers used a sequential, parallel comparison design developed for indications that are particularly prone to placebo effects (see Ivanova et al., 2011). These trials have two stages. Scientists first randomize subjects to receive drug or placebo. After some time, they separate the placebo group into those who improve and those who did not, and re-randomize them to receive either drug or placebo. This allows researchers to compare the two treatments among people who exhibit a minimal placebo response.

The researchers divided the double-blind trial into two five-week periods. For the first, 93 people were randomized to receive AVP-923 twice daily; 127 got placebo. Because this drug combination has known side effects, doses started at 20 mg dextromethorphan hydrobromide and 10 mg quinidine sulfate, and then rose to 30 mg and 10 mg, respectively. For the second stage, people already on the drug stayed on it, but the placebo group was re-randomized such that 15 of the 30 who had had a strong placebo response were switched to the drug and 15 stayed on placebo. Of the 89 placebo participants whose scores stayed stable, 44 now got the drug and 45 stayed on placebo. This brought the total number of people who received the drug to 152. 

The agitation/aggression domain of the neuropsychiatric inventory (NPI) served as the primary outcome measure; secondary outcomes included the other subdomains and the total NPI score. As an exploratory outcome, the researchers measured change in the Alzheimer’s Disease Assessment Scale-Cognition (ADAS-Cog).

At the end of the first stage, AVP-923 improved the NPI agitation score by 3.3 points, compared with 1.7 points in the placebo group. At the end of the second stage, the researchers assessed the 44 people in the initial placebo group who had demonstrated no strong placebo effect and began taking AVP-923 after five weeks. In them, the drug reduced agitation and aggression. Scores fell by 2 points, compared with 0.8 points for the 45 people still on placebo. Overall, NPI total and subdomain scores also improved more for the drug treatment group. In both stages, caregivers reported that their own stress abated, and both they and clinicians reported that the patient was less agitated.

Adverse events occurred in 61 percent of people taking AVP-923, versus 43 percent on placebo. They included falls, diarrhea, urinary-tract infections, dizziness, agitation, contusion, back pain, and peripheral edema, and are in line with the drug’s known safety profile, said co-author Joao Siffert of Avanir. Serious adverse events occurred in 12 people on the drug and six on placebo. Twenty-six people dropped out of the trial; 17 from the AVP-923 group, nine from placebo. Eight and four, respectively, stopped because of adverse events. No one died during the trial.

AVP-923’s safety profile is not benign. It is contraindicated for people with heart problems such as prolonged QT interval, atrioventricular block, and others, as well as people with a history of thrombocytopenia, hepatitis, bone-marrow depression, or lupus-like syndrome. People who are overly sensitive to dextromethorphan-containing common cough medicines also should not take AVP-923. Drug interactions with other CNS drugs, such as monoamine oxidase inhibitors (MAOs) and serotonin reuptake inhibitors (SSRIs) are also known (see Cruz, 2013; Schoedel et al., 2014). 

Still, Siffert interpreted the latest trial results to mean that overall, AVP-923 was well-tolerated. The researchers found no evidence of treatment-related cognitive decline on the MMSE or ADAS-cog. This is important, he said, because other drugs aimed at treating agitation have been associated with cognitive decline (see Feb 2014 news story).

The results justify larger Phase 3 trials, claimed Siffert, who said that the company would begin talks with the FDA early next year. “The study shows potential for this drug to be helpful,” agreed Lon Schneider, University of Southern California, Los Angeles, who was not involved in this trial but saw the poster. Schneider noted that the relatively small sample was heterogeneous in age, MMSE score, and both level and type of agitation. Patients also differed on the drugs they were already taking, such as acetylcholinesterase inhibitors, memantine, antidepressants, and antipsychotics. Since quinidine sulfate is a potent inhibitor of cytochrome P450 2D6 and slows the metabolism of several drugs, it may affect other drugs, and that could contribute to efficacy as well as adverse events, he told Alzforum. Schneider noted that larger trials will be needed to discern how well the drug works in which type of patient.

A second AVP-923 study was also presented on a poster at the AAN conference. This study followed 134 patients with AD, vascular or frontotemporal dementia, or dementia with Lewy bodies who took AVP-923. They constituted the dementia cohort, led by Rachelle Doody, Baylor College of Medicine, Houston, of the larger PRISM-II study. Conducted at more than 100 centers in the United States, PRISM-II used no placebo control, but simply evaluated safety, tolerability, and the effect of AVP-923 in 750 people whose PBA accompanied primary conditions such as stroke or a traumatic brain injury. “The purpose of this investigation was to gain more experience with the use of AVP-923 in patients whose PBA was related to disorders that were not well-represented in the trials used for regulatory approval,” Doody wrote to Alzforum.

Participants received the 20 mg/10 mg combination twice daily for 12 weeks, and were assessed for change in the Center for Neurologic Study-Lability Scale (CNS-LS). The Center for Neurologic Study is the not-for-profit organization in La Jolla, California, that originally developed this drug combination. CNS-LS is a questionnaire to measure uncontrollable tearfulness or laughter (see Moore et al., 1997). Other endpoints included how often participants had such episodes, their MMSE score, and clinicians’ and caregivers’ global impression of change.

Patients improved on the CNS-LS. Episodes dropped by half at day 30 compared to baseline, by two-thirds at day 60, and three-quarters of clinicians and caregivers reported that symptoms improved at day 90. Overall, MMSE scores remained steady.

Forty-nine patients had mild to moderate adverse events, such as headache, urinary-tract infection, diarrhea, nausea, falls, dizziness, and sleepiness; a third were considered possibly related to AVP-923. Fourteen people had serious side effects, but Doody and colleagues considered these unrelated to the drug. Seventeen patients dropped out because of adverse events. The authors wrote that the results suggest that AVP-923 treats PBA regardless of the underlying neurological condition.

John (Wes) Ashford of Stanford University Medical Center in California wrote to Alzforum that PBA is uncommon in Alzheimer’s. Ashford suggested that other off-patent drugs could work as well or better than AVP-923 if they were competitively tested against one another. As examples, he mentioned the antidepressants nortriptyline and citalopram, the antipsychotic olanzapine, and the anticonvulsants lamotrigine, valproic acid, and topiramate. The study was funded by Avanir; both Cummings and Doody have consulted for this company.—Gwyneth Dickey Zakaib


  1. Three considerations in interpreting this trial are: the use of the term "agitation" and vague inclusion criteria; the drug combination tested along with the concomitant medications used by participants; and the magnitudes of the clinical outcomes.

    The term "agitation" doesn’t fairly describe the participants in this trial (and a few others).  Here, "agitation" was defined as poorly organized and purposeless psychomotor activity characterized by verbally or physically aggressive behavior, or non-aggressive agitated behavior, that interfered with daily activities, occurred at an intermittent or constant frequency over at least one week, and for which medication was judged by a clinician to be indicated.  Participants also had to be scored at least a four, "moderately ill," on a seven-point Clinical Global Impression-Severity rating.  The inclusion criteria allowed for outpatients, assisted living, and long-term care patients, a broad range for cognitive severity, i.e., MMSE scores from eight to 28, and few restrictions on the use of other psychotropic drugs.  For example, 74 percent were taking cholinesterase inhibitors, 50 percent memantine, 56 percent antidepressants, 21 percent antipsychotics, and 8 percent benzodiazepines.

    The drug combination, dextromethorphan and quinidine, was tested.  Dextromethorphan, the antitussive ingredient in over-the-counter Robitussin DM and other cough syrups, is rapidly metabolized by the liver cytochrome enzyme CYP2D6 to an active metabolite dextrorphan.  Both parent and metabolite have sigma-1 agonist and uncompetitive NMDA antagonist activities, among other actions. Quinidine, a strong inhibitor of CYP2D6, increases the bioavailability of dextromethorphan; and also increases the bioavailability of several other drugs metabolized by CYP2D6, including some antidepressants, antipsychotics, and, in principle, donepezil.  Quinidine itself is associated with fatal cardiac arrhythmias and is anticholinergic, although very low doses of 10 mg were used. 

    Finally, the clinical rating scales effects favoring the dextromethorphan/quinidine combination were fairly modest and of the same magnitude as effects observed with antipsychotics and the antidepressant citalopram in similar trials and patients. 

    In sum, the "agitated" participants in this trial varied broadly from fairly mild to severe agitation, non-aggressive to verbally and physically aggressive,  very mild to quite severe dementia, and in the use of other drugs intended to quell these "agitated" behaviors and that might adversely interact with either dextromethorphan and quinidine.  This was not a simple trial comparing a drug with a placebo; yet it is not large enough for statisticians to dissect out the possible interactive effects and to convincingly determine which patients may have responded.  The Phase 3 trial design will pose interesting challenges.

  2. In response to Lon Schneider: 

    Agitation is one of the most disabling features of AD and other dementias, compromising the quality of life of both patients and care providers. It is a common cause of institutionalization in the United States and is associated with accelerated mortality. Although non-pharmacologic interventions are preferable to control agitation, many patients do not respond to such interventions and many families or professional caregivers find it difficult to implement them consistently. Pharmacologic treatment options are often required. While somewhat effective, existing medications have substantial side effects and none are approved for treatment of agitation with dementia. New solutions are needed, and the field has seen a welcome recent resurgence of interest in better treatments for neuropsychiatric symptoms of neurodegenerative disorders.

    The clinical trial of dextromethorphan/quinidine (DM/Q: Nuedexta™) for the treatment of agitation in Alzheimer’s disease is an important study in several ways:

    • Use of a novel agent to treat agitation.
    • Innovative use of Sequential Parallel Comparison Design (SPCD).
    • Significant treatment effect.
    • Acceptable adverse event profile.

    The results of the trial support the idea that DM/Q may be an effective treatment for a specific type and severity of agitation in patients with mild-moderate AD. As Dr. Schneider points out, trials involve choices and answer very specific questions constrained by the decisions made in sample definition and trial design.

    In this trial, the definition of agitation reflects other trials in this research area. It is similar to the consensus definition recently published by the International Psychogeriatric Association (IPA; Cummings et al., 2014). The patients identified for the trial are those recognized by most practitioners as “agitated” and in need of pharmacologic treatment if not responsive to non-pharmacologic interventions. The population was further defined by requiring a specific score on the clinical global impression rating. This insured that the degree of agitation was sufficient to allow detection of a treatment response with the study sample size. Caregiver strain scores indicated that the caregivers of patients in the study had severe strain, suggesting that they perceived the agitation as substantial. Inclusion of patients with different types or severities of agitation might have resulted in different outcomes. These alternate populations may be included in future trials.

    A broad range of patients were allowed in the trial—as Dr. Schneider points out—with a cognitive impairment varying from mild to severe and concomitant medications being allowed. This reflects the “real world” context of AD patients with agitation, increases the practical applicability of the results, and enhances the ecological validity of the trial. A frequently voiced criticism of trials is that the patients included are rarely seen by practitioners and the generalizability of the trial is uncertain. The DM/Q trial partially avoided this pitfall.

    The adverse-event profile observed in the trial was acceptable, especially when compared to the side effects observed with the agents most commonly used now for agitation, antipsychotics. These latter agents are associated with cognitive impairment, extrapyramidal symptoms, weight gain, metabolic syndrome, stroke, and an elevated risk of death.

    The design of the trial may prove instructive to the field. Placebo responses are common in trials of drugs for the treatment of neuropsychiatric symptoms in neurodegenerative disorders, and a variety of strategies are evolving to assist in differentiating drug and placebo responses. The DM/Q trial is the first application of SPCD in dementia. In this design, there is a standard drug-placebo comparison in stage 1. In stage 2, placebo non-responders are re-randomized to drug or placebo. The placebo response in stage 1 tends to be much higher than in stage 2 and the second stage of the trial may allow a clearer view of the pharmacological effect of the agent with limited placebo responses. In the DM/Q trial a robust response was seen in stage 1 and again in stage 2; in other trials and with other agents the second stage of the trial may be particularly informative.

    A significant reduction in agitation was observed with DM/Q. The primary outcome—the Neuropsychiatric Inventory (NPI) agitation scale—was decreased based on the primary analysis (SPCD; Chen et al. 2012) with p = 0.00008; similar results were seen when each five-week, parallel group stage was assessed separately: stage 1 p<0.001 (standardized effect size [SES] = 0.505) and stage 2 p = 0.021 (SES=0.340). There was a convergence among instruments with significant drug-placebo differences observed in the global impression of change (p<0.001) and a measure of caregiver burden (p<0.05).

    The DM/Q agitation trial is a step forward in trial conduct and efficacy demonstration for an agent with acceptable safety and representing a novel mechanism of action that differs from existing agents. Further investigations of the agent are warranted to improve understanding and replicate the finding of this novel approach.

    —Response by the DM/Q Agitation Study Steering Committee


    . Agitation in cognitive disorders: International Psychogeriatric Association provisional consensus clinical and research definition. Int Psychogeriatr. 2015 Jan;27(1):7-17. Epub 2014 Oct 14 PubMed.

Make a Comment

To make a comment you must login or register.


Therapeutics Citations

  1. AVP-923
  2. Valproate

News Citations

  1. Citalopram Calms Agitation in Alzheimer’s, but Carries Risks

Paper Citations

  1. . Nuedexta for the treatment of pseudobulbar affect: a condition of involuntary crying or laughing. P T. 2013 Jun;38(6):325-8. PubMed.
  2. . Dextromethorphan plus ultra low-dose quinidine reduces pseudobulbar affect. Ann Neurol. 2010 Nov;68(5):693-702. PubMed.
  3. . Optimality, sample size, and power calculations for the sequential parallel comparison design. Stat Med. 2011 Oct 15;30(23):2793-803. Epub 2011 Jul 29 PubMed.
  4. . Evaluating the safety and efficacy of dextromethorphan/quinidine in the treatment of pseudobulbar affect. Neuropsychiatr Dis Treat. 2014;10:1161-74. Epub 2014 Jun 26 PubMed.
  5. . A self report measure of affective lability. J Neurol Neurosurg Psychiatry. 1997 Jul;63(1):89-93. PubMed.

External Citations


Further Reading


  1. . Mechanism of action of dextromethorphan/quinidine:comparison with ketamine. CNS Spectr. 2013 Oct;18(5):225-7. PubMed.
  2. . Pseudobulbar affect: prevalence and management. Ther Clin Risk Manag. 2013;9:483-9. Epub 2013 Nov 29 PubMed.
  3. . A naturalistic on-off-on trial of dextromethorphan/quinidine for agitation associated with cerebellar injury. Psychosomatics. 2012 Sep-Oct;53(5):470-3. Epub 2012 Mar 27 PubMed.