Objective: To determine if the treatment effects of donepezil vs placebo, reported in moderate to severe Alzheimer’s disease (AD) (Feldman et al. Neurology 2001), vary as a function of baseline dementia severity, as measured by the standardized Mini-Mental State Examination (sMMSE).
Design: 290 patients with moderate to severe AD (sMMSE score 5-17; Functional Assessment Staging score of 6 or less) were randomized to treatment in this 24-week, double-blind, placebo-controlled trial.
Methods: Patients received donepezil, 5 mg/day for the first 28 days and 10 mg/day thereafter per clinician’s judgment (n=144), or placebo (n=146). Outcome measures included global (Clinician’s Interview-Based Impression of Change with caregiver input; CIBIC-plus), cognitive (sMMSE and Severe Impairment Battery; SIB), functional (Disability Assessment for Dementia; DAD), and behavioral assessments (12-item Neuropsychiatric Inventory; NPI). Statistical methods included ANOVAs and ANCOVAs, performed on the Week 24 change from baseline scores for each measure (last observation carried forward; LOCF), with terms for treatment (donepezil or placebo), dementia severity (baseline sMMSE score), and severity x treatment interaction. In one set of analyses (ANCOVA), baseline sMMSE score was treated as a continuous variable ranging from 5 to 17. In another (ANOVA), sMMSE scores were dichotomized at the median value (12) into severe (5-12) and moderate AD (13-17) subgroups.
Results: There was no imbalance in baseline patient demographic characteristics or assessment scores between treatment groups. Mean sMMSE scores (± SD) at baseline were 11.9 ± 3.56 for the donepezil group and 12.1 ± 3.75 for the placebo group. Mean sMMSE scores (± SD) for the severe AD subgroup were 9.0 ± 2.19 for donepezil and 8.9 ± 2.19 for placebo, and for the moderate AD subgroup were 15.0 ± 1.52 and 15.4 ± 1.53, respectively. Treatment effects were significant for each outcome measure (P <0.01) at Week 24 LOCF, for both analyses (continuous and discrete). There was no treatment x severity interaction for any measure at Week 24 LOCF for both continuous and discrete analyses (Table 1).
Table 1. Interaction of treatment effect by baseline sMMSE score
| Efficacy measure | Treatment x severity P-value | |
| Continuous (ANCOVA) | Discrete (ANOVA) | |
| CIBIC | 0.5785 | 0.1779 |
| sMMSE | 0.8540 | 0.8772 |
| SIB | 0.2531 | 0.1007 |
| NPI 12-item total | 0.9879 | 0.4845 |
| DAD | 0.5898 | 0.7295 |
Conclusions: The treatment effects of donepezil were consistent across the full range of dementia severity in patients with moderate to severe AD and were independent of baseline disease severity in analyses using either continuous or discrete variables. These data suggest that donepezil is efficacious across the spectrum of more advanced AD.
Back to PB Tuesday Poster Sessions
Back to The Eleventh International Congress