Is cerebrospinal fluid obtained for diagnostic purpose a good material for biomarker studies in amyotrophic lateral sclerosis?