Mets_2007_Curr.Med.Res.Opin_23_2485

OBJECTIVE: Despite response variability, cholinesterase inhibitors are recommended in mild to moderate Alzheimer's disease. Dose titration is common; however randomized controlled trials (RCTs) have mainly investigated fixed-dose regimens. We examined practice patterns and outcomes of 6 +/- 1.5-month rivastigmine therapy.
METHODS: Prospective, pharmacoepidemiologic, naturalistic study of 175 evaluable patients with mild to moderate Alzheimer's disease (+ 151 caregivers) from 52 centers in Belgium on 6 +/- 1.5 month (titrated) rivastigmine treatment. MAIN OUTCOME MEASURES: Measured at baseline (enrollment) and follow-up (6 +/- 1.5 months). For patients: Mini-Mental State Exam (MMSE), Activities of Daily Living (ADL), Neuropsychiatric Inventory (NPI), Global Deterioration Scale (GDS) scores; treatment response (improvement, maintenance, or decline less than normative slope). For caregivers: hours/week spent caring; Zarit Caregiver Burden Scale (ZCBS), 12-item version of General Health Questionnaire (GHQ-12), Instrumental Activities of Daily Living (IADL) scores.
RESULTS: Patients' MMSE and NPI scores (p < 0.001) improved from baseline to follow-up, but not ADL and GDS scores. Treatment response was 89.1% of patients for MMSE (including 60.6% with improvement) and 77.7% for NPI (including 57.1% with improvement). Quadratic curves were fitted for the average daily dose and the MMSE and NPI scores; with a trend towards average daily dose of 6.0 +/- 3.0 mg/day. Caregivers' ZCBS (p = 0.036) and GHQ-12 (p = 0.029) scores improved, but not IADL scores and time spent caring.
CONCLUSIONS: Patients' MMSE and ADL scores confirmed the meta-analyses of rivastigmine efficacy trials, while NPI scores exceeded efficacy results. Proportionately more patients responded to (titrated) treatment than in fixed-dose RCTs. Caregivers reported less burden (similar to meta-analysis) and better general health over the study period. Where efficacy and effectiveness results diverge, the benefit is in 'real-world' effectiveness. Large sample, multi-country replications, less sensitive to censoring secondary to missing data and powered to permit advanced modeling, as well as RCTs with adaptive designs to accommodate titration, are needed. The profile of patients most likely to benefit from treatment or most vulnerable to treatment outcome must be studied, as must the impact of physician- and center-related variables on outcomes.