Abstract

There is increasing interest in the value of CSF biomarkers to predict those individuals with mild cognitive impairment who will progress to dementia. However, lumbar puncture is not routine in these patients and biomarker assays are not universally available. To change clinical practice there must be very good evidence that biomarkers are helpful over and above clinical impression. Here we discuss the merits of CSF biomarkers compared with clinicians using simple bedside cognitive tests. Although every biomarker has a superior positive predictive value, most have inferior negative predictive values. When predicting the progression of mild cognitive impairment, the overall misclassification rate by clinicians using bedside cognitive tests is approximately 38% but this could be reduced to approximately 30% by using Aβ1–42 and to 24% with phosphorylated tau or total tau (or a combination of CSF biomarkers). Clinicians and patients together should decide whether this is sufficient to warrant the additional burden of a lumbar puncture, and the cost of the test, or whether further studies are needed before useful and clinically practical conclusions can be reached.