Defining cases and controls on the basis of biomarkers rather than clinical diagnosis may reduce sample sizes required for genetic studies. The aim of this study was to assess whether characterizing case/control status on the basis of cerebrospinal fluid (CSF) profile would increase power to replica...