A comment on "Cognitive and functional effects of Ginkgo biloba in neurodegenerative disorders: Systematic review and meta-analysis"

Abstract

Omar and Ghani (2006) concluded that Ginkgo biloba extract had small and clinically meaningless effects on cognition, behavior and functional ability in neurodegenerative disorders based on a systematic review and meta-analysis. Unfortunately, study selection did not correspond to the authors’ own search strategy, they extracted incorrect data from selected studies and applied statistical methods inappropriately. The authors claim that cognition was assessed by means of the Alzheimer’s Disease Assessment Scale – Cognitive Subscale (ADAS-Cog) in ten trials and performed a meta-analysis of test results. In fact, the ADAS-cog was not used in six of these 10 trials. The values the authors used for meta-analysis cannot be derived from the published studies. Studies demonstrating statistically significant superiority are presented with non-significant result in the forest plot. The authors did not realize that results from some trials were published in more than one paper and handle these publications as separate trials. Finally, the analysis uses studies encompassing patients with heterogenous diagnoses and too large a range in disease severity. Given the interaction of diagnosis, disease severity, and scale the analysis is distorted by ceiling and floor effects and time-dependent effect sizes. Consequently, the meta-analyses presented arrived at incorrect results, that served as basis for the authors’ interpretation and conclusions. Most systematic reviews and meta-analyses concluded that the standardized Ginkgo biloba extract EGb 761 demonstrated benefits, mainly on cognitive deficits and neuropsychiatric symptoms, in patients with mild neurocognitive disorder or mild to moderate dementia, and that the drug was safe and well tolerated.