GDNF in Parkinson disease: An object lesson in the tyranny of type II

Type II errors may be having a significant impact on drug discovery. This is of particular importance in the clinical neurosciences, where endpoints are often subjective scores of disability rather than unequivocal events such as survival. Here we examine a recently published study [Lang AE, Gill S, Patel NK, et al. Randomized controlled study of intraputamenal glial cell-derived neurotrophic factor infusion in Parkinson disease. Ann Neurol 2006;59:459-66] in an area of immense importance to neuroscience. This small study found no detectable clinical benefit from infused intraputamenal GDNF as a treatment for Parkinson disease. However the standard deviation of the accrued data turned out to be considerably higher than had been anticipated in the power analysis performed prior to the study. In order to determine what impact, if any, this had on the conclusions that could be drawn, the actual data were analyzed by means of both the t-test and the rank-based Somers' D. The study was found to be underpowered and thus incapable of ruling out a large effect of GDNF on Parkin-on disease. It therefore does not contradict the large effects seen in previous open-label studies. (c) 2006 Elsevier B.V. All rights reserved.