The hallmark manifestation of Alzheimer’s—memory loss—is one thing. When family members hear the diagnosis from their neurologist, they largely know that memories will fade and that they might be ripped away forever. What is less expected is that the slate isn’t merely wiped clean, it’s redrawn and a picture of someone very different can appear. This can present as uncharacteristic depression, anxiety, or irritability early in the disease before progressing to delusions, sleep disturbances, or agitation.

There are currently no approved treatments specifically for agitation in patients with Alzheimer’s, and its presence can be one of the most challenging and distressing aspects of the disease: The Alzheimer’s Association says behavioral symptoms typically are a determining factor in residential care placement.

Hoping to fill this void is AVP-923. Backed by Avanir Pharmaceuticals, recent Phase II data presented at the 2014 American Neurological Association Meeting showed that AVP-923 demonstrated a clinically meaningful and statistically significant improvement in Alzheimer’s patient’s agitation. Additionally, patients saw benefits as early as the first week, which were sustained over five and 10-week endpoints.

“Results from this Phase II study are quite promising and this pharmacological approach may help address a significant unmet need for hundreds of thousands of Alzheimer’s patients with agitation,” said Joao Siffert, MD, Senior Vice President, R&D & Chief Medical Officer of Avanir Pharmaceuticals, in an interview with Practical Neurology™ magazine. “We intend to meet with the FDA and EMA to discuss the next phase of the clinical program to support regulatory approval in the foreseeable future.”


AVP-923 is a combination of two well-characterized compounds: the active CNS ingredient dextromethorphan hydrobromide (an uncompetitive NMDA receptor antagonist, sigma-1 receptor agonist and inhibitor of the serotonin transporter [SERT] and norepinephrine transporter [NET]); plus low-dose quinidine sulfate (a CYP2D6 enzyme inhibitor), which increases the bioavailability of dextromethorphan. Avanir says AVP-923 is known to have certain cardiovascular risks and drug-drug interactions and, accordingly, excluded these patients from the study.


The 10-week randomized, double-blind, placebo-controlled, multicenter Phase II study evaluated the efficacy, safety and tolerability of AVP-923 for the treatment of agitation in Alzheimer’s patients.

Further, researchers chose sequential parallel comparison design (SPCD) over the standard parallel design. “The SPCD was used to help reduce the impact of a placebo response, which is a relatively common finding in clinical trials of neuropsychiatric disorders,” Dr. Siffert said. “SPCD has also been shown to help increase the statistical power in most instances where it has been utilized.”

When using SPCD, a study is conducted in two treatment stages, with the efficacy analysis including a) all subjects in stage 1 and, b) stage 2 data from stage 1 placebo non-responders. The relevant data from the two stages are pooled to compute an overall p-value.1 For AVP-923 this consisted of two consecutive double-blind treatment stages, each of 5-week duration.

A total of 220 Alzheimer’s patients in the US were enrolled. Eligible patients were initially randomized in a 3:4 ratio to receive either AVP-923 (dose escalated from DM 20mg/Q 10mg to DM 30mg/Q 10mg) or placebo. At the end of week 5, patients who initially received placebo were stratified according to their response to treatment and subsequently re-randomized in a 1:1 ratio to receive either AVP-923 or placebo for the remainder of the study (an additional 5 weeks of treatment). Patients who initially received AVP-923 continued to receive AVP-923 DM 30mg/Q 10mg for the remainder of the study.

The main efficacy measure was the agitation/aggression domain of the Neuropsychiatric Inventory (NPI). Dr. Siffert said the NPI was selected because it has been validated in patients with dementia and used in dozens of clinical trials where it showed to be sensitive to change over time and in response to treatment. “The NPI measures multiple neuropsychiatric domains and provides a comprehensive overview of the medication effects across dimensions, including irritability/ lability and aberrant motor behaviors which, in addition to agitation/aggression, are also important features of agitation,” he said.

The NPI is comprised of 12 domains. The score for each domain is the product of frequency (on a four-point scale) x severity (on a three-point scale). The maximum score on any single sub-domain is therefore 12. A 30 percent reduction in the agitation/aggression domain of the NPI is considered to be a clinically meaningful improvement in a patient’s symptoms.

The primary endpoint followed a standard analysis of SPCD by combining the change from baseline to week 5 (stage 1: full analysis population) and change from week 5 to week 10 (stage 2) on the NPI agitation/aggression domain (patients who were considered “non-responders” to placebo during the initial 5 weeks). Secondary outcome measures included global assessments of disease severity, other neuropsychiatric symptoms, cognition, activities of daily living, quality of life and caregiver strain.


At the study baseline, the NPI agitation/aggression domain was 7.0 and 7.1 for patients in the AVP-923 and placebo groups, respectively.

At the end of stage 1, scores for the AVP-923 treated patients had reduced by 3.3 (SD=2.98), vs. 1.7 (SD=3.10) for placebo (p<0.001), amounting to a 47 percent reduction and a Standard Effect Size (SES)=0.505. In stage 2, where only placebo non-responders were included in the primary analysis, a reduction of 2.0 (SD=3.19) was observed in patients treated with AVP-923 vs. 0.8 for patients on placebo (p=0.021), corresponding to a 26 percent reduction for AVP-923 vs. a 6.7 reduction for placebo and a SES=0.340.

In addition to the improvement in agitation, “clinically and statistically significant improvements were observed in global measures such as CGIC-agitation, PGIC, and measures of caregiver burden—caregiver strain and caregiver distress,” Dr. Siffert said. The total NPI score and several individual NPI domains also improved with treatment. He added there was no evidence of cognitive impairment on the MMSE (p=0.053 in favor of AVP-923) and no evidence of somnolence/ sedation. Two secondary endpoints, the Alzheimer’s Disease Cooperative Study-Activities of Daily Living (ADCSADL) Inventory and Quality of Life-AD (QoL-AD) Measure (patient and caregiver) showed no difference between treatment groups.

  1. Baer L., Ivanova A. When should the sequential parallel comparison design be used in clinical trials? Clinical Investigation, Vol. 3, No. 9, Pages 823-833.
  2. Lyketsos CG, Carrillo MC, Ryan JM, Khachaturian AS, Trzepacz P, Amatniek J, Cedarbaum J, Brashear R, Miller DS. Neuropsychiatric symptoms in Alzheimer’s disease. Alzheimers Dement. 2011 Sep;7(5):532-9.