DONEPEZIL FOR NURSING HOME PATIENTS WITH DEMENTIA: A REINTERPRETATION OF THE EVIDENCE ‐ Response to the Editor

Abstract

The above letter was referred to the authors of the original paper; their reply follows. To the Editor: Patients had to have behavioral problems to be enrolled in this trial. As clearly stated in the paper, both groups exhibited improvements in behavior as measured by changes in the total score of the Neuropsychiatric Inventory—Nursing Home version (NPI-NH) and there were no significant drug-placebo differences on this measure. The categorical analysis of improvement, no change and worsening for each of the assessed behaviors, was a planned secondary analysis. Only agitation was significant in this analysis. We agree with the caveats raised by Drs. Steinman and Kovinsky regarding the statistical robustness of this finding. Nevertheless, agitation was the most prevalent symptom, present in 64% of the patients, and, in our clinical experience, any degree of improvement of this behavior is of clinical importance to patient management in the nursing home setting. Given the consistent findings of beneficial effects of cholinesterase inhibitors (ChEIs) on behavior in outpatients,1–3 the absence of a drug-placebo difference on the NPI-NH total score was surprising to us. For example, a study of donepezil in patients with moderate-to-severe Alzheimer's disease (AD) reported improvements in total score at the 24-week endpoint for the drug-treated group only, with all 12 measured symptoms improving relative to placebo.3 In a 5-month trial of galantamine in patients with mild-to-moderate AD, total scores progressively worsened in the placebo group.2 In attempting to account for the discrepant findings in the nursing home, it is reasonable to suggest that the use of psychoactive medication (in about 60% of these patients) and nonpharmacological methods of behavioral management as employed in skilled nursing homes may have contributed to the observed placebo group improvement. Beneficial effects of donepezil on cognition were evident in this study despite the known limited sensitivity of the Mini-Mental State Examination (MMSE) in detecting decline in untreated patients with advanced dementia. The lack of progressive decline was apparent in the placebo group in this study. Furthermore, an improvement of at least 3 points at any time during the study was observed in 51% of the donepezil-treated patients, compared with 36% of the placebo patients (P < .05). Statistical significance favoring donepezil was also reached on the cognitive subscale of the Clinical Dementia Rating (CDR) at Week 24. This is the only placebo-controlled trial of ChEI in the nursing home setting. The use of ChEIs has consistently been shown to lead to behavior and functional benefits. The differential pattern of findings for the CDR functional subscale (P = .057) and the lack of effect on the physical self-examination scale in these nursing home patients reinforces the importance of appropriate instrument selection in patient cohorts with high comorbidity and limited mobility. The paper is thorough in its presentation of safety and tolerability findings. The importance of weight loss in these patients is not underestimated. Clinically significant weight loss (≥7% baseline weight) occurred in 9% and 6% of donepezil- and placebo-treated patients, respectively. The percentage of patients who discontinued because of weight loss was 1% and 4% of donepezil- and placebo-treated patients, respectively. The mean weight decrease of 3 kg in patients with weight loss as an adverse event was observed in both treatment groups, indicating that this degree of weight loss was not specific to donepezil. We conclude that these data provide further evidence of the stabilization of cognition and overall dementia severity consistently reported with donepezil, an effect that is clinically important in a progressive neurodegenerative disease. Donepezil was safe and generally well tolerated in these nursing home patients, despite their advanced age and high comorbidity and concomitant medication usage.