What conclusions, if any, can we make from the recent Archives of Neurology report on “Antioxidants for Alzheimer Disease”?
Galasko et al’s recently published their results from a randomized controlled clinical trial evaluating whether certain antioxidant supplements affected cerebrospinal fluid (CSF) biomarkers in subjects with mild to moderate Alzheimer’s Disease (Galasko et al 2012).
There was no index of compliance….this study was…. underpowered…..more subjects should have been enrolled, especially to compensate for subjects who failed to complete the 16-week study,” notes Dr. Evans.
The authors reported that treatment for 16 weeks with either Coenzyme Q (CoQ; 400 mg, tid), or 800 IU/d of vitamin E (α-tocopherol) plus 500 mg/d of vitamin C plus 900 mg/d of α-lipoic acid (E/C/ALA), or placebo did not influence CSF markers related to amyloid or tau pathology, but that F2-isoprostanes were reduced (~19%) in the E/C/ALA group compared to placebo. In addition, the authors reported that scores declined on the Mini-Mental State Examination in the E/C/ALA group.
Dr. Evans points out some of the many significant flaws including:
- There was no index of compliance with the specified intervention reported, so it is unclear as to how the investigators determined whether the subjects consumed the correct daily dose of the respective interventions.
- The source, quality, and type of supplements used is not adequately reported. For example, no information is provided as to whether the vitamin E was d- or d-,l-α-tocopherol, or something else. The various isomers of alpha tocopherol exhibit various degrees of absorption, bioavailability, and biological activities. With regard to CoQ, there are huge differences in absorption and bioavailability, depending on whether the material is crystalline (poor absorption/bioavailability) or crystalline-free (higher absorption/bioavailability). The type and source of α-lipoic acid is not specified. If α-lipoic acid was not prepared using a solvent-free process, the possibility exists that a 900 mg daily dose might not be very well tolerated.
- There is an obvious disparity in the number of subjects in the treatment groups vs. the placebo who were on medications for AD. For example,100% of subjects in the placebo group were on an acetylcholinesterase inhibitor vs 88% in the E/C/ALA group, and only 84% in the CoQ group.
- The small number of subjects in the treatment arms (CoQ, n = 20; E/C/ALA, n = 24; placebo, n = 22) is a major flaw. The methods section indicated that power calculations for CSF biomarkers were based on published data on serial CSF levels of tau protein. With 25 subjects in each treatment arm, there would be 80% power to detect a 40% or more change in CSF tau level (α = 0.05). Clearly this study was markedly underpowered (i.e. more subjects should have been enrolled, especially to compensate for subjects who failed to complete the 16-week study). Furthermore, the goal of observing a 40% change in the CSF tau level appears overly aggressive for this study design.
- The authors failed to point out that E/C/ALA treatment reduced the level CSF tau by ~20% and phosphorylated tau (P-tau) by ~9% vs changes in the placebo group by -10.5% and 0.9%, respectively. In addition, CoQ intervention reduced the level CSF Aβ42 peptide by ~8% vs a reduction of 3% in the placebo group. Although these changes were not indicated in Table 2 as statistically significant, they are suggestive of a treatment effect and might have reached statistical significance if the group sizes were larger. At the very least, these results provide a clear rationale for a larger study evaluating the benefits of antioxidant intervention for AD.
Despite all the media hype and anti-vitamin hysteria surrounding this Arch of Neurol report, Dr. Evan’s scientific rigor clearly exposes multiple scientific flaws and also points to some unreported, yet beneficial observations discovered in the antioxidant therapy group.
Galasko, MD et al. Antioxidants for Alzheimer Disease:A Randomized Clinical Trial With Cerebrospinal Fluid Biomarker Measures.Arch Neurol. Published online March 19, 2012. doi:10.1001/archneurol.2012.85
Reported prepared by Joseph L. Evans, PhD of P & N Development Ventures LLC