By Lucy Piper, medwireNews reporter
medwireNews: Blood-based markers based on amyloid-β and phosphorylated (p)-tau217 ratios have high diagnostic accuracy for predicting Alzheimer’s disease (AD) in patients in primary and secondary care settings, suggests research from Sweden.
The amyloid probability score (APS) 2 combines the ratio of plasma p-tau217 to non-p-tau217 with the ratio of plasma amyloid-β 42 and amyloid-β 40, explained Oskar Hansson, reporting the findings at the Alzheimer’s Association International Conference 2024, held in Philadelphia, Pennsylvania, USA.
In a single batch analysis, this biomarker accurately identified AD pathology (cerebrospinal fluid (CSF) amyloid-β and tau positivity), based on a predefined cutoff of 90% specificity, in 91.5% of primary care patients and 87.7% of secondary care patients, with positive and negative predictive values ranging from 87.9 to 91.4% and 87.4 to 91.7%, respectively.
A similar level of accuracy was seen when the ratio of plasma p-tau217 to non-p-tau217 was measured alone.
The findings are based on data for 1213 patients – 515 from the BioFINDER Primary Care study and 698 from the Memory Clinics of Skåne University Hospital or Ängelholm Hospital. The participants (52% men) were aged a mean of 74.2 years and 23% had subjective cognitive decline, 44% had mild cognitive impairment, and 33% had dementia. The rate of amyloid positivity was 53.6% in the primary care cohort and 58.0% in the secondary care cohort and AD pathology was present in 50% of patients from each group.
Most of the patients had multiple comorbidities, including 16.9 to 26.2% with chronic kidney disease.
In addition to the single batch analysis, which was conducted in 307 and 300 of the primary and secondary care patients, respectively, the researchers also did prospective analyses where the patients had measurements taken every 2 weeks over a 1-year period for the primary care patients (n= 208) and a 4-year period for the secondary care patients (n=398).
APS2 identified patients with AD pathology with an accuracy of 88.9% in primary care patients and 91.0% in secondary care patients. The positive predictive values were a corresponding 87.8% and 91.3% and the negative predictive values were 90.0% and 90.6%. Again, measurement of the p-tau217 ratio alone showed similar levels of accuracy.
Hansson highlighted the value of using these biomarkers versus not using them in the standard-of-care diagnostic workup.
In an analysis comparing the use of APS2 and the p-tau217 ratio to identify AD pathology compared with primary care practitioner and dementia specialist diagnosis without their use, the results showed that APS2 and the p-tau217 ratio were 89.1% to 91.8% accurate in primary and secondary patients. This compared with accuracies of 71.4% for dementia specialists and 58.0% for primary care practitioners, the latter of which Hansson said was “not significantly different to chance.”
He stressed that “this is not an indication that primary care practitioners are not good at their job it is because they do not have access to good tools today to diagnose Alzheimer’s disease.”
The respective positive predictive values were 86.0% to 91.7% versus 65.6% and 56.8%, while the negative predictive values were 90.1% to 96.9% versus 80.7% and 58.8%.
When the team made comparisons for identifying clinical AD (full clinical assessment and CSF/PET confirmation) in patients with mild cognitive impairment or mild dementia, the findings were similar. APS2 and the p-tau217 ratio were 90.5% to 91.4% accurate, whereas the accuracy was 73.4% for dementia specialists and 61.3% for primary care practitioners who did not use biomarker information.
Hansson noted that the accuracy of the blood-based biomarkers was consistent across the different cognitive groups, and when they used a two-cutoff approach that has been used in cerebrospinal fluid tests.
It is based on 95% sensitivity and specificity, with results between the two termed “intermediate risk,” which would indicate the need for further CSF or PET testing.
The accuracy of APS2 and the p-tau217 ratio for identifying AD pathology across the primary and secondary care patients ranged from 91.0% to 93.5%, the positive predictive values from 92.0% to 96.2%, and the negative predictive values from 90.1% to 92.5%.
Hansson pointed out how low the intermediate values were, at 4–13% in primary care patients and 6–11% in secondary care patients.
He concluded that the findings show “that there is a great need for these biomarkers and we can be super optimistic that we now have these tools.”
He added, however, that more evaluation is needed to see how the use of these biomarkers affects clinical care and changes the treatment of patients, and more guidance “is imperative for their successful integration into routine clinical practice.”
The study was simultaneously published in JAMA.
medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2024 Springer Healthcare Ltd, part of the Springer Nature Group
AAIC24; Philadelphia, Pennsylvania, USA: July 28–Aug 1
JAMA 2024; doi:10.1001/jama.2024.13855
