A 24-week, double-blind, multi-center, randomised parallel group study compared the efficacy and safety of 800 mg bid
cyclandelate with placebo in patients with mild to moderate
dementia of primary degenerative or vascular origin. A total of 196 patients entered the study, 147 patients completed treatment in adherence with the protocol. Primary outcome measures were the cognitive score of the
Alzheimer's Disease Assessment Scale (ADAS-Cog), the subscale Instrumental
Activities of Daily Living of the Nurses' Observation Scale for Geriatric Patients (NOSGER-IADL) and the Clinical Global Impressions of Change (CGI-C). Safety assessments included adverse events, vital signs, ECG and clinical laboratory parameters. The primary efficacy results based on a multi-level responder analysis including ADAS-Cog, NOSGER-IADL and CGI-C failed to demonstrate statistical superiority of
cyclandelate in comparison to placebo. The direction of changes favored
cyclandelate in each of the variables, but the differences to placebo were small and varied considerably between patients and centers. Retrospective exploratory analyses suggested that efficacy of
cyclandelate might be dependent on the severity of the disease. The treatment effects in favor of
cyclandelate were statistically significant in the subgroup of moderately impaired patients (MMSE at baseline <18) for ADAS-Cog (delta = -4.0 points, p = 0.015) and CGI-C (delta = -0.4 points, p = 0.043) but not for NOSGER-IADL (delta = -1.6 points, p = 0.059). When patients were stepwise selected for the severity of the disease according to ADAS-Cog at baseline (>15, >20, >25 points), statistical significance was reached for ADAS-Cog and NOSGER-IADL beginning with the step ADAS-Cog >20 points: delta ADAS-Cog = -3.9 points, p = 0.044; delta NOSGER-IADL = -1.0, p = 0.023. The treatment differences increased further with the step ADAS-Cog >25 points: delta ADAS-Cog = -7.0 points, p = 0.008; delta NOSGER-IADL = -1.7, p = 0.003. Treatment differences in CGI-C increased marginally with the stepwise selection but did not reach statistical significance. The
drug was safe and well tolerated.