The effect of funding sources on donepezil randomised controlled trial outcome: a meta-analysis

OBJECTIVE To investigate whether there is a difference in the treatment effect of donepezil on cognition in Alzheimer disease between industry-funded and independent randomised controlled trials. DESIGN Fixed effects meta-analysis of standardised effects of donepezil on cognition as measured by the Mini Mental State Examination and the Alzheimer's Disease Assessment Scale-cognitive subscale. DATA SOURCES Studies included in the meta-analyses reported in the National Institute for Health and Care Excellence (NICE) technical appraisal 217 updated with new studies through a PubMed search. ELIGIBILITY CRITERIA Inclusion criteria were double-blind, placebo-controlled trials of any length comparing patients diagnosed with probable Alzheimer disease (according to the NINCDS-ADRDA/DSM-III/IV criteria) taking any dosage of donepezil. Studies of combination therapies (eg, donepezil and memantine) were excluded, as were studies that enrolled patients with a diagnosis of Alzheimer disease associated with other disorders (eg, Parkinson's disease and Down's syndrome). RESULTS Our search strategy identified 14 relevant trials (4 independent) with suitable data. Trials sponsored by pharmaceutical companies reported a larger effect of donepezil on standardised cognitive tests than trials published by independent research groups (standardised mean difference (SMD)=0.46, 95% CI 0.37 to 0.55 vs SMD=0.33, 95% CI 0.18 to 0.48, respectively). This difference remained when only data representing change up to 12 weeks from baseline were analysed (industry SMD=0.44, 95% CI 0.34 to 0.53 vs independent SMD=0.35, 95% CI 0.18 to 0.52). Analysis revealed that the effect of funding as a moderator variable of study heterogeneity was not statistically significant at either time point. CONCLUSIONS The effect size of donepezil on cognition is larger in industry-funded than independent trials and this is not explained by the longer duration of industry-funded trials. The lack of a statistically significant moderator effect may indicate that the differences are due to chance, but may also result from lack of power.