Annals of Neurology | 2019
Medicare for All?
Abstract
We thank Prins et al for applying the methods of our article on “Plasma Amyloid as Prescreener for the Earliest Alzheimer Pathological Changes” to their research. Designing prescreening tools for amyloid positivity will ultimately facilitate recruitment of cognitively normal participants for Alzheimer disease clinical trials. Current studies comparing the performance of bloodbased amyloid assays, as well as efforts to compare effects of preanalytical sample handling procedures, show that accurate blood-based amyloid measurement is highly complex. Nevertheless, Prins and colleagues did not provide information on the plasma type, preanalytical sample handling procedures, and method of plasma amyloid measurement. Also, essential details on the lumbar puncture procedure and cerebrospinal fluid (CSF) measurement method were lacking, as were details on the demographics of the included population, such as recruitment method, medical history, prevalence of amyloid positivity, and distribution of APOEε4 carriership. None of these is a trivial issue, and yet they are the main variables on which their statistical analysis focused. The missing details make a head-to-head comparison between the study results and identification of causes of discrepancies very difficult, if not impossible. Therefore, we are not able to place their results in a wider context. In our study, we observed that plasma amyloid was useful for prescreening amyloid positivity in a 2-step diagnostics approach, with application of a blood test first and subsequently funneling a smaller number of individuals toward amyloid testing with established tests such as positron emission tomography or CSF. Prins et al observed low sensitivity in their cohort, raising the question as to whether our results could be explained by overfitting. We acknowledge that cross-validation in external cohorts is essential. Of note, our results do not stand alone; multiple recent studies using highly sensitive immunoassays for blood-based amyloid beta quantification showed capability to discriminate between amyloidpositive and amyloid-negative cognitively normal individuals, with or without subjective cognitive complaints, providing independent support for our findings. Moreover, in our article we not only investigated the association of plasma amyloid and CSF amyloid, but showed that low plasma amyloid levels are related to a higher risk of subsequent clinical disease progression to mild cognitive impairment or dementia. Investigating two types of statistical outcomes (ie, CSF amyloid positivity and change in clinical syndrome during follow-up) provides converging evidence for the diagnostic potential of the plasma amyloid biomarker. We therefore conclude that plasma amyloid measurement is promising for prescreening Alzheimer pathological changes in cognitively normal individuals with subjective cognitive decline.