Drug treatments for Alzheimer’s disease. I. A comparative analysis of clinical trials
Alzheimer's disease is a devastating disorder of the elderly, which is expected to become a major public health problem in Canada over the next two decades. Recent projections indicate that the number of cases of Alzheimer's disease in Canada will increase from 160,000 to more than 380,000 by the year 2020. These projections are based on the assumption that the incidence of Alzheimer's disease will remain unchanged.
Significant research efforts in Canada and around the world are seeking to determine the cause of AD and to find a cure. Until this has been achieved, however, there remain thousands of Canadian elderly suffering from AD. A major research focus in AD, at the present time, is the development of therapies that could alter the course of the disease in those already afflicted.
In this report we review those published clinical trials that we believe have the methodological integrity to provide the best evidence on the efficacy of donepezil, metrifonate, rivastigmine, selegiline, vitamin E, lecithin, linopirdine, propentofylline and gingko biloba for the treatment of Alzheimer's disease.
Twenty-seven randomized clinical trials were retrieved from the literature and found to meet appropriate methodological standards.
We conclude that for selegiline, vitamin E, lecithin, linopirdine, and propentofylline the published data do not provide support for efficacy.
Based on the evidence we reviewed, it is our conclusion that donepezil, metrifonate and rivastigmine, however, all provide statistically significant modest benefit on cognitive performance and global functioning to the elderly with probable AD who are eligible for inclusion in clinical trials. The magnitude of the effect is similar for all of the medications. The results from the trials of gingko biloba are promising but the effects are smaller than those from the above mentioned therapies.
Although all of these medications appear to be well tolerated, in terms of the occurrence of adverse events, dropout rates are sometimes high and may have resulted in overestimation of apparent treatment effects.
It is important to note that this report is based on the results of published trials only and as a result may be subject to publication bias. In particular, if a bias exists such that trials showing no effect are less likely to be published then our findings may overestimate efficacy through the selective inclusion of positive trials. The exclusion of trials that were not published due to poor methodology, however, would not have resulted in a bias as it is unlikely that such trials would have met our standards of methodogical rigour.
In a companion report (Drug Treatments for Alzheimer's Disease. II. A Review of Outcome Measures in Clinical Trials), we review the psychometric properties of the primary and secondary outcome measures used in these trials and we remain concerned about the wide variety of scales used that do not have adequate psychometric assessment. Although a great number of the scales used have evidence of reliability and validity, responsiveness to change has generally not been adequately assessed. For this reason, the clinical significance of the treatment: placebo differences remain unclear.
The ability to carry out a comparative analysis of therapies for AD depends not only on the comparability of design, duration, and outcome measures used but also on the methods of reporting the results of the trials. There was no consistent method of reporting the results of the AD trials even when emanating from the same group of investigators. Although journals do have different guidelines for authors, we recommend that reports of clinical trials of therapies for AD follow a standardized format for presenting results, which would enhance the ability to compare results across studies.