Premium
Using cerebrospinal fluid amyloid‐beta (1‐42) in the memory clinic: Concordance with PET and use of biomarker ratios across immunoassays
Author(s) -
Willemse Eline A.J.,
Tijms Betty M.,
Van Berckel Bart N.M.,
Gan Sara,
Le Bastard Nathalie,
Latham Jessica,
Radwan Rachel R.,
van Der Flier Wiesje,
Scheltens Philip,
Teunissen Charlotte E.
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.045128
Subject(s) - concordance , biomarker , medicine , cerebrospinal fluid , dementia , memory clinic , cohort , amyloid (mycology) , gold standard (test) , youden's j statistic , cognitive impairment , surrogate endpoint , receiver operating characteristic , gastroenterology , pathology , nuclear medicine , oncology , disease , chemistry , biochemistry
Background Shifting to automated cerebrospinal fluid biomarker platforms urges re‐evaluation of biomarker interpretation in the memory clinic setting. We aimed to evaluate whether different ratios of biomarkers, pTau:Aβ 1– 42 and Aβ 1– 42 :Aβ 1– 40 , improved the discrimination of amyloid PET positivity in a retrospective cohort and whether this was dependent on the platform used. Method Cerebrospinal fluid samples of individuals in the biobank of the Amsterdam Dementia Cohort were selected based on availability of visual amyloid PET reads (n=288; n=133 florbetaben, n=86 Pittsburgh compound‐B, n=64 flutemetamol, n=5 florbetapir; 179 (62%) amyloid PET positive; age 63±7 years; 131(45%) female; MMSE 24±4; diagnoses subjective cognitive decline (n=58), mild cognitive impairment (n=42), Alzheimer’s disease (AD; n=145), or other dementias (n=43)). Aβ 1– 42 , Aβ 1– 40 (Lumipulse only), pTau and tTau were measured using both Elecsys and Lumipulse automated platforms. Sensitivity and specificity percentages were determined based on optimal Youden’s index in receiver operating curve analyses with amyloid PET result as gold standard. Result Both pTau:Aβ 1– 42 and Aβ 1– 42 :Aβ 1– 40 ratios showed higher concordance with amyloid PET compared to Aβ 1– 42 alone. Sensitivity and specificity percentages for Elecsys were 96% and 89% for pTau:Aβ 1– 42 (cut‐point 0.020), 96% and 80% for Aβ 1– 42 :Aβ 1– 40 (cut‐point 0.091), and 91% and 75% for Aβ 1– 42 (cut‐point 1089 pg/ml). For Lumipulse, sensitivity and specificity percentages were 97% and 91% for pTau:Aβ 1– 42 (cut‐point 0.072), 99% and 83% for Aβ 1– 42 :Aβ 1– 40 (cut‐point 0.071), 91% and 73% for Aβ 1– 42 (cut‐point 714 pg/ml). Areas under the curves were higher for pTau:Aβ 1– 42 than Aβ 1– 42 :Aβ 1– 40 using Elecsys (0.949 and 0.924, respectively, p‐value = 0.014) but comparable for both ratios using Lumipulse (0.956 and 0.942, respectively, p‐value = 0.20). Conclusion We recommend the pTau:Aβ 1– 42 ratio as optimal biomarker for use in memory clinic setting using the Elecsys platform, since it showed the strongest concordance with PET amyloid. For Lumipulse assays, pTau:Aβ 1– 42 and Aβ 1– 42 :Aβ 1– 40 perform equally well.