Blood Biomarkers Improve The Diagnostic Accuracy Of Alzheimer’s Disease As Compared With Current Diagnostic Standard In the Primary Care Setting
Sebastian Palmqvist, Pontus Tideman, Joel B. Braunstein, Tim West, Kristopher M. Kirmess, Matthew R. Meyer, Ruben Smith, Ayesha Fawad, Erik Stomrud, Shorena Janelidze, Oskar Hansson- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
The first and final level of care for most individuals with cognitive impairment is primary care. The aim was to examine the diagnostic accuracy of plasma phospho‐tau217/non‐phospho‐Tau217 (p‐tau217 Ratio) and Aβ42/40 Ratio for Alzheimer’s disease (AD) as compared to usual care by primary care physicians (PCPs).
Method
In the BioFINDER‐Primary Care study, we recruited patients examined due to cognitive complaints at 25 primary care centers in Sweden (n = 307). After care‐as‐usual (CT/MRI, cognitive testing, and clinical assessment), the PCPs documented the most likely underlying etiology and proposed a treatment plan. Plasma biomarkers included p‐tau217 Ratio, Aβ42/Aβ40 Ratio, and a predefined algorithm combining both (C2N mass‐spectrometry). The outcome was AD pathology defined using FDA‐approved cerebrospinal fluid Aβ42/p‐tau181 (Lumipulse®) or visual read of 18F‐Flutemetamol Aβ‐PET.
Result
The mean age was 76±6.3 years, 48% were women, and 49% had AD pathology. Twenty‐five percent had subjective cognitive decline, 47% mild cognitive impairment, and 28% dementia. Using AD pathology as the outcome, the AUCs were 0.80 (95%CI 0.75‐0.86) for Aβ42/Aβ40 Ratio, 0.91 (0.88‐0.94) for pTau217 Ratio, and 0.94 (0.92‐0.97) for the algorithm. In those with available PCP questionnaires (n = 265), the PCPs correctly classified AD in 54% (50‐59%) of the cases, compared with 77% (73‐81%) for Aβ42/40 Ratio, 85% (82‐88%) for p‐tau217 Ratio, and 87% (84‐90%) for the algorithm. The mean certainty of the PCPs’ diagnosis (1‐10, not at all to completely certain) was 4.7 (5.1 in those with AD pathology who were correctly diagnosed as AD and 4.6 in the missed AD cases). The PCPs’ proposed treatment, which in 87% followed the treatment guidelines, led to 50% of true AD cases not receiving symptomatic treatment and 30% of non‐AD cases incorrectly receiving symptomatic AD treatment (DLB patients excluded).
Conclusion
We found that it is difficult for PCPs to identify AD among patients with cognitive impairment in primary care, which leads to diagnostic uncertainty and incorrect treatment. The use of a blood test for AD pathology has great potential for improving the diagnostic accuracy and treatment of AD in primary care. At AAIC, a larger dataset will be presented including the effects of co‐morbidities and medications.