Author Affiliation: Department of Radiology, The Alexander Family Professor of Alzheimer's Disease Research, Mayo Clinic, Rochester, Minnesota.
We appreciate the opportunity to respond to the critique of our paper “Evidence for ordering of Alzheimer disease biomarkers.”1 In their critique, Glodzik et al stated, “Cross-sectional data presenting percentages of abnormal values across 3 static diagnostic categories cannot be substituted for observing within subject temporal trajectories of changes.” While we explicitly stated that additional longitudinal data are needed, we believe the logic of our approach is sound. If 1 biomarker crosses the threshold from normal to abnormal earlier than another as the disease progresses, it follows that the earlier changing biomarker should be abnormal more often in a sample of subjects that span the cognitive continuum than a biomarker that crosses the normal-to-abnormal threshold later. Furthermore, comparing how often a biomarker is abnormal across multiple subjects is very analogous to the classic approach taken by Braak and Braak.2 Through autopsy studies of individuals spanning all ages, Braak and Braak developed a scheme to stage the topographic spread of Alzheimer disease pathologies. A time element is inherent in their staging. The major difference from our approach is that in the Braak and Braak approach, subjects were arrayed on the x-axis by age; in our approach, subjects were arrayed on the x-axis by disease severity—both age and disease severity increment by time. We find it odd that in the same letter critiquing our cross-sectional approach, Glodzik et al later supported their “tau-comes-first” position with references to Braak autopsy data.2 Autopsies surely represent the ultimate form of cross-sectional data.
Jack CR. Ordering of Alzheimer Disease Biomarkers—Reply. Arch Neurol. 2012;69(3):414-415. doi:10.1001/archneurol.2011.2908