TORONTO – Hallucinations – especially auditory hallucinations – triple the risk that a risperidone-controlled Alzheimer’s patient with psychoses will relapse if the drug is withdrawn.
“I think the clinical impact we see here is that for patients with hallucinations, and particularly auditory hallucinations, antipsychotic discontinuation should be done very, very cautiously because they do have a very high risk of relapse,” Anjali Patel, DO, said at the Alzheimer’s Association International Conference 2016. “Close monitoring will be necessary and antipsychotic medications promptly reinstated if relapse occurs.”
The findings come from a responder analysis of an open-label study of risperidone (Risperdal) use in Alzheimer’s patients who express neuropsychiatric symptoms. The primary results of the multicenter Antipsychotic Discontinuation in Alzheimer’s Disease (ADAD) trial were published in 2012 (N Engl J Med. 2012;367:1497-507). The 48-week study administered open-label, flexible-dose risperidone for 16 weeks to 180 patients with Alzheimer’s dementia, with agitation and/or aggression. The patients in this study had a mean age of 79 years at baseline. Most (62%) were taking a cholinesterase inhibitor, and many took memantine (35%). Patients commonly used anxiolytics (17%) and antidepressants (24%). The mean Neuropsychiatric Inventory (NPI) score at baseline was 36. Patients were moderately impaired, with a mean Mini Mental State Exam score of 14.
At 16 weeks, patients who had not responded left the study. After 16 weeks of open-label treatment, 110 patients who had responded well continued the dosing schedule for 32 weeks, continued risperidone for 16 weeks and then went on placebo for 16 weeks, or were switched to placebo for 32 weeks. Discontinuation of risperidone was associated with a two- to four-fold increased risk of relapse over 16-32 weeks.
Dr. Patel of Columbia University, New York, presented the preplanned post-hoc analysis that examined the association between the 12 NPI symptom domains and the likelihood of relapse at week 32. In a univariate analysis, only hallucinations posted a significant association with discontinuation of risperidone. Hallucinations of any severity at baseline were present in 43 patients (39%). The relapse rates were similar among patients without baseline hallucinations (35%) and those with mild baseline hallucinations (37%). But 78% of those with severe hallucinations relapsed when risperidone was withdrawn.
Baseline hallucinations remained a strong predictor of relapse in a multivariate model as well, with a risk ratio of 2.96 for relapse among patients who had severe baseline hallucinations, compared with those with mild or no hallucinations. Age, gender, race, and nursing home placement had no significant impact on relapse rate.
Of the 17 patients with any baseline hallucinations who were switched to placebo, 13 (77%) relapsed, compared with 38% of the patients with hallucinations who continued risperidone (risk ratio, 1.98).
The risk for relapse was particularly high when the hallucinations were primarily auditory, Dr. Patel said. “In fact, visual hallucinations were not predictive of relapse.”
Among the 11 patients with severe baseline hallucinations, 10 relapsed when risperidone was withdrawn (91%; RR, 2.88), compared with 57% of patients with severe hallucinations who stayed on the drug (RR, 1.59).
The ADAD trial was funded by the National Institutes of Health and the Department of Veterans Affairs. Dr. Patel had no financial disclosures for the subanalysis.
On Twitter @alz_gal