Editorial: We may need large trials to find treatments for neurodegenerative diseases

Abstract

The article by Brookmeyer and Abdalla in this issue of the journal calculated that clinical trials to prevent Alzheimer’s disease would require tens of thousands of patients, depending on the ages that patients were enrolled and whether they had any signs of the disease. The sample sizes ranged from 4000 to 84,000 patients. The authors comment that these are larger than sample sizes for Alzheimer’s disease trials to date. We note these sample sizes might not be larger than clinical trials for heart disease, and would be much smaller than the 1954 trial testing the polio vaccine. Although the evidence is limited, there is very good reason to believe that treating neurodegenerative diseases early might be much more effective than initiating treatment only after symptomatic disease has emerged. By the time a patient is diagnosed with a neurodegenerative disease such as Alzheimer’s, Parkinson’s or amyotrophic lateral sclerosis (ALS), irreparable neuronal damage may already have occurred, making it unlikely that a disease modifying treatment could do more than slow or stop further progression. Given the prevalence of Alzheimer’s disease and its personal and societal costs, the expense of conducting large trials to test treatments to prevent dementia in patients at high risk, or even in normal elders, is justified by benefits of reducing the disease’s prevalence and morbidity. In rare diseases, such as ALS with an incidence of only two per 100,000 (compared to several hundred per 100,000 in Alzheimer’s disease), a trial in asymptomatic people would only be feasible in the population at a very high lifetime risk of disease, that is, those at genetic risk for ALS. An early intervention study may be feasible in the non-genetic population, but would require significant effort to reduce the long diagnostic delay, with a median of 10 months from symptom onset. But the benefits of trials for asymptomatic or early disease would radically change our approach to treating the disease. And the potential impact on diseases that are currently untreatable would be profound and would well justify the complexity and cost of even a large trial. This article is valuable because it gives us a method for calculating the sample sizes required for trials to delay or prevent these neurodegenerative diseases. The fact that these sample sizes are larger than we might have hoped increases the value of these results as they inform us of what we need to do.