Journal of Neural Transmission (2022) 129:231–237 https://doi.org/10.1007/s00702-022-02474-9 NEUROLOGY AND PRECLINICAL NEUROLOGICAL STUDIES - ORIGINAL ARTICLE Blood amyloid and tau biomarkers as predictors of cerebrospinal fluid profiles Constance Delaby1,2 · Daniel Alcolea2 · Christophe Hirtz1 · Jérôme Vialaret1 · Jana Kindermans1 · Lisa Morichon1 · Juan Fortea2 · Olivia Belbin2 · Audrey Gabelle3 · Kaj Blennow4,5 · Henrik Zetterberg4,5,6,7,8 · Alberto Lleó2 · Sylvain Lehmann1 Received: 30 December 2021 / Accepted: 5 February 2022 / Published online: 15 February 2022 © The Author(s) 2022 Abstract Introduction Blood biomarkers represent a major advance for improving the management, diagnosis, and monitoring of Alzheimer's disease (AD). However, their context of use in relation to routine cerebrospinal fluid (CSF) analysis for the quantification of amyloid peptides and tau proteins remains to be determined. Methods We studied in two independent cohorts, the performance of blood biomarkers in detecting “nonpathological” (A−/ T−/N−), amyloid (A+) or neurodegenerative (T+ /N+) CSF profiles. Results Plasma Aβ1–42/Aβ1–40 ratio and phosphorylated tau (p-tau(181)) were independent and complementary predictors of the different CSF profile and in particular of the nonpathological (A−/T−/N−) profile with a sensitivity and specificity close to 85%. These performances and the corresponding biomarker thresholds were significantly different from those related to AD detection. Conclusion The use of blood biomarkers to identify patients who may benefit from secondary CSF testing represents an attractive stratification strategy in the clinical management of patients visiting memory clinics. This could reduce the need for lumbar puncture and foreshadow the use of blood testing on larger populations. Keywords Blood · Biomarkers · Clinical management · Lumbar puncture · CSF Introduction 2011). Thus, the identification of AD processes years before the onset of symptoms recently triggered a paradigm shift in Detection of Alzheimer's disease (AD) with high sensitivity which AD could be viewed as a biological rather than clini- and specificity is key for the management of patients. Diag- cal entity (Jack et al. 2018). The importance of biomarkers nosis can include detection of amyloid and tau biomarkers in was also emphasized when defining the unbiased "A/T/N" the cerebrospinal fluid (CSF), which is one of international classification system (Jack et al. 2016). These evolutions guidelines’ criteria (Dubois et al. 2014; McKhann et al. are coupled with the prospect of introducing treatments that would modify the trajectory of the disease by delaying its Constance Delaby and Daniel Alcolea share first authorship. 4 * Sylvain Lehmann Department of Psychiatry and Neurochemistry, Institute s-lehmann@chu-montpellier.fr of Neuroscience and Physiology, The Sahlgrenska Academy at the University of Gothenburg, Gothenburg, Sweden 1 Laboratoire de Biochimie Protéomique Clinique 5 Clinical Neurochemistry Laboratory, Sahlgrenska University (LBPC‑PPC), Univ Montpellier, CHU Montpellier, Hospital, Mölndal, Sweden INM INSERM, Hôpital St Eloi, IRMB 80 av A Fliche, 6 34295 Montpellier, France Department of Neurodegenerative Disease, UCL Institute 2 of Neurology, Queen Square, London, UK Hospital de la Santa Creu i Sant Pau, Biomedical Research 7 Institute Sant Pau, Universitat Autònoma de Barcelona, UK Dementia Research Institute at UCL, London, UK Barcelona, Spain 8 Hong Kong Center for Neurodegenerative Diseases, 3 CMRR, Univ Montpellier, CHU Montpellier, INM INSERM, Hong Kong, China Montpellier, France 13 Vol.:(0123456789) 232 C. Delaby et al. clinical expression. However, the use of CSF to detect AD at Table 1  Demography, CSF biomarker values and AT(N) classifica- an early stage in a large population remains difficult because tion of the cohort of Montpellier and Barcelona of the invasive nature of lumbar puncture (LP). Blood bio- Variable The Montpellier The Barcelona markers are in this context, of particular interest. The pos- cohort cohort sibility of detecting amyloid peptides and tau proteins in Mean SD Mean SD plasma has recently shaken the field of neurodegenerative diseases detection. Many research groups, including ours, Age (years) 68.3 10.5 67.6 12.3 are evaluating the diagnostic value of these biomarkers for Sex (M%)* 60.3 – 39.4 – the accurate detection of AD, through cross sectional or lon- CSF biomarkers gitudinal studies using retrospective samples (Alcolea et al. Aβ1–40 (pg/mL) 15,940 6875 12,078 3838 2021; Lewczuk et al. 2018; Brickman et al. 2021). Aβ1–42 (pg/mL) 826 377 906 446 Tau (pg/mL) 461 320 467 326 However, one context of use (COU) that has not yet been p-tau(181) (pg/mL) 66 43 73 63 directly addressed is in relation to CSF testing performed ATN* in a routine clinical setting. The question is whether blood A−/T−/N− 38.1% – 31.8% – biomarkers can be used to decide the need for further CSF A−/T−/N+ 4.8% – 3.9% – analysis. Thus, the objective here is to evaluate the perfor- A−/T+/N− 1.6% – 0.6% – mance of blood biomarkers in detecting a “nonpathological” A−/T+/N+ 4.8% – 2.6% – (A−/T−/N−) CSF profile, rather than a specific pathological A+/T−/N− 4.8% – 25.8% – profile as seen in AD or brain injury. Note that the notion A+/T−/N+ 3.2% – 0,6% – of “nonpathological” does not refer here to the globality of A+/T+/N− 4.8% – 3.9% – the CSF analysis which includes many other biochemical, A+/T+/N+ 38.1% – 31.1% – immunological or microbiological analyses, but is restricted to the results of the amyloid and tau biomarkers. The results SD standard deviation of this study suggest that blood markers can predict the pres- * Significant difference ence of nonpathological CSF profile and could thus be deci- sive in whether or not to perform a LP. had also mixed dementia, normal pressure hydrocephalus and Parkinson disease with cognitive signs or subjective Methods cognitive impairment. All participants gave their written informed consent to Participants participating in clinical research on CSF and plasma bio- markers, and protocols at both centers were approved by the The Barcelona cohort included 150 participants from the respective Ethics Committees. Sant Pau Initiative on Neurodegeneration (SPIN cohort) (Alcolea et al. 2019) evaluated at the Sant Pau Memory Unit Primary outcomes (Barcelona, Spain) between November 2013 and October 2019. Participants in this cohort mostly were patients with a CSF Aβ1–42, Aβ1–40, total tau and phosphorylated tau 181 diagnosis of AD, Dementia with Lewy bodies (DLB), fron- [p-tau(181)] were measured using Fujirebio Lumipulse or totemporal lobar degeneration-related syndromes (FTLD), Innotest assay as described (Lehmann et al. 2020). The cut- and mild cognitive impairment (MCI), or cognitively nor- off values were initially obtained from groups of patients mal controls. The two cohorts differ in the distribution of clinically diagnosed with AD (whose clinical diagnoses A/T/N profiles based on CSF biomarkers, with the Barce- were made blind to biomarker results) and, for the Barce- lona cohort having a much higher percentage of isolated lona cohort, from amyloid-PET positive and amyloid-PET amyloid-positive patients (A+/N−/T−) than the Montpel- negative participants (Alcolea et al. 2019, PMID 31464088) lier cohort (Table 1). All participants had received neuro- or, for Montpellier, from control population of the memory logical and neuropsychological evaluation and provided clinic with various etiology (Lehmann et al. 2013, 2018). CSF and plasma samples. The Montpellier cohort included Based on these data, a nonpathological CSF profile corre- 161 patients recruited from September 2009 to June 2017 sponding the (A−/T−/N−) situation is defined as having a (Lehmann et al. 2020). All patients underwent a thorough value of the Aβ1–42/Aβ1–40 ratio (A) above the cutoff and clinical examination including biological laboratory tests, values of tau (N) and p-tau(181) (T) below the pathologi- neuropsychological assessments, and brain imaging. In addi- cal cutoffs. We also identified amyloid (A+/A−) and tau- tion to AD, DLB, FTLD and MCI, patient from this cohort neurodegeneration (N+T+/N−T−) CSF profiles. 13 Blood amyloid and tau biomarkers as predictors of cerebrospinal fluid profiles 233 Three different approaches were used to measure plasma Results levels of Aβ1–42 and Aβ1–40: “Neurology 3-Plex A” (Q3, both cohorts) and “Neurology 4-plex E Advantage kit” (Q4, We first tested the performance (AUC) of blood biomark- Montpellier cohort) in the Simoa platform (Quanterix) and ers to distinguish a nonpathological (A−/T−/N−) from a an IP-MS approach from Shimadzu (Nakamura et al. 2018) pathological CSF profile represented by the other A/T/N (both cohorts) implemented in Montpellier’s laboratory and situations (Table 1). Plasma Aβ1–40 and Aβ1–42, individually, slightly modified from the original protocol (Alcolea et al. show variable but low accuracy for nonpathological profiles 2021). Levels of p-tau(181) were measured in the Simoa detection (Table 2 and Supp Table 1). In contrast, the Aβ1–42/ platform (Quanterix). Aβ1–40 ratio showed much higher performance regardless of the analytical method used (Q3, Q4 or IP-MS). Statistical analysis In both cohorts, plasma levels of total tau were compa- rable between nonpathological and pathological CSF pro- Statistical analyses were completed with Medcalc (v19.8). files, while elevated plasma p-tau(181) was associated with a The accuracy of the blood-based assays to discriminate pathological CSF profile (Supp Table 1). Plasma p-tau(181) nonpathological amyloid and tau CSF profile was evaluated showed an AUC of 0.865 in Montpellier and 0.773 in Bar- using receiver operating characteristic (ROC) curve analy- celona to discriminate a nonpathological (A−/T−/N−) CSF sis and calculation, using the area under the curve (AUC) profile (Table 1). Pairwise comparison of AUCs confirmed as a measure of diagnostic accuracy. Comparison of ROC the high performance of p-tau(181) when compared to curves to test the statistical significance between assay val- Aβ1–42 and total tau (p < 0.05) in Montpellier cohort. In Bar- ues derived from the method of DeLong et al. (1988) for celona cohort, AUCs obtained for p-tau(181), Aβ1–42(Q3) and the calculation of the standard error of the AUCs. Multiple Aβ1–42(IP-MS) were similar (Table 2) and AUC of p-tau(181) regression was used to examine the relationship between was significantly higher than AUC of total tau (p = 0.022). CSF and blood assays allowing to combine them and evalu- To assess the value of combining biomarkers, we first tested ate if they were independent or not. Logistic regression used the correlation between the different factors and observed to combine independent factors was employed using differ- that Aβ1–42, Aβ1–40, Aβ1–42/Aβ1–40 on the one hand, and ent ways of introducing the factors into the algorithm (Enter/ tau, p-tau(181) on the other hand, were correlated together Forward/Backward/Stepwise) if their p values were < 0.05 (Pearson correlation; p < 0.001). We therefore selected as and removed if p > 0.1. independent variable Aβ1–42/Aβ1–40 and p-tau(181) (with amyloid peptides measured with Q4 and Q3 Quanterix in Montpellier and Barcelona, respectively) that had the best AUCs. Combining biomarkers requires a stepwise approach (Mamtani et al. 2006), however in our case we only have two Table 2  Diagnostic accuracy Non-pathological The Montpellier cohort The Barcelona cohort of plasma biomarkers to CSF (A−/T−/N−) discriminate non pathological CSF (A−/T−/N−) profiles in Blood biomarkers AUC​ SE 95% CI p AUC​ SE 95% CI p the cohort of Montpellier and Aβ1–40(Q3) 0.661 0.069 0.530–0.777 0.020 0.638 0.080 0.490–0.769 0.085 Barcelona Aβ1–40(Q4) 0.672 0.068 0.542–0.785 0.011 – – – – Aβ1–40(IP-MS) 0.526 0.074 0.396–0.653 0.728 0.549 0.051 0.463–0.633 0.329 Aβ1–42(Q3) 0.548 0.077 0.417–0.675 0.534 0.715 0.076 0.571–0.834 0.044 Aβ1–42(Q4) 0.542 0.076 0.410–0.669 0.601 – – – – Aβ1–42(IP-MS) 0.638 0.073 0.507–0.755 0.060 0.656 0.048 0.571–0.733 0.001 Tau 0.612 0.078 0.480–0.733 0.149 0.618 0.081 0.470–0.752 0.495 p-tau(181) 0.865 0.049 0.756––0.938 < 0.0001 0.773 0.039 0.697–0.837 < 0.0001 Aβ1–42/Aβ40(Q3) 0.709 0.068 0.580–0.818 0.002 0.848 0.062 0.719–0.934 < 0.0001 Aβ1–42/Aβ40(Q4) 0.753 0.065 0.627–0.854 < 0.0001 – – – – Aβ1–42/Aβ40(IP-MS) 0.715 0.066 0.587–0.822 0.001 0.661 0.048 0.577–0.739 0.0008 Logistic regression 0.904 0.040 0.804–0.964 < 0.0001 0.882 0.050 0.759–0.956 < 0.0001 Aβ1–40, Aβ1–42, p-tau(181) Biomarkers were quantified using either Quanterix technology (Q3 and Q4) or Shimadzu approach (IP- MS). Significant differences are indicated by bolded p (threshold 0.05) 13 234 C. Delaby et al. A B (A-/T-/N-) (A-/T-/N-) Montpellier Barcelona (AUC: 0.753) (AUC: 0.848) (AUC: 0.865) (AUC: 0.773) (AUC: 0.904) (AUC: 0.882) C 100 D 100 80 80 60 60 Sensitivity Sensitivity (A+) (T+/N+) 40 40 Montpellier Montpellier (AUC: 0.726) (AUC: 0.747) 20 20 (AUC: 0.820) (AUC: 0.881) (AUC: 0.841) (AUC: 0.920) 0 0 0 20 40 60 80 100 0 20 40 60 80 100 100-Specificity 100-Specificity Fig. 1  Receiver operating characteristic (ROC) curves for plasma (orange) or ratios (pink) to discriminate CSF profiles. Blue line biomarkers to discriminate non pathological (A−/T−/N−) (A and B), corresponds to the ROC curve yielded by a logistic regression that amyloid (A+) (C) or neurodegenerative (T+/N+) (D) CSF profiles. included all three plasma markers and ratios Lines indicate areas under the curve (AUC) for individual biomarker factors and we tested different logistic regression approaches comparison was not significant between p-tau(181) and (see “Methods”) which all resulted in the same algorithm Aβ1–42/Aβ1–40 (measured with Quanterix Q3). However, confirming the independence and statistical relevance of the AUC of logistic regression was higher than that of the two selected biomarkers. As illustrated (Table 2 and p-tau(181) alone (p = 0.002). In this cohort, the highest Fig. 1A), AUCs obtained for Aβ1–42/Aβ1–40 (both meas- Youden index was obtained when performing the logistic ured with Quanterix Q4), p-tau(181) and logistic regression regression, reaching a sensitivity of 85.0% and a specific- combining these three parameters were very close and pair- ity of 86.7% for the detection of nonpathological amyloid wise comparison of the different AUCs was significant only and tau CSF profiles. in Montpellier cohort between Aβ1–42/Aβ1–40 and logistic We also studied the performance of biomarkers to identify regression (Supp Table 3). two specific pathological situations with amyloid (A+, Sup The highest Youden index in this context was obtained Table 1, Fig. 1C) or tau-neurodegeneration (T+/N+, Sup when performing the logistic regression, reaching a sen- Table 2, Fig. 1D) profiles. p-tau(181) and logistic regression sitivity of 85.2% and a specificity of 83.6% for the detec- were again the most discriminant. The performance of the tion of nonpathological amyloid and tau CSF profiles. Aβ1–42/Aβ1–40 ratio was comparatively lower, especially for In the Barcelona cohort (Table 1 and Fig. 1B), pairwise the T+/N+ profile. 13 Blood amyloid and tau biomarkers as predictors of cerebrospinal fluid profiles 235 Discussion et al. 2020; Bayoumy et al. 2021), can be expected to be even more effective. An important stage in the management of patients consulting With performances close to 85% sensitivity and speci- for cognitive complaints is the decision to perform or not a ficity for the detection of nonpathological (A−/T−/N−) LP, which likely provides early indicators of neurodegenera- CSF profiles, one can really consider the results of blood tive diseases. Indeed, CSF analysis provides indirect signs biomarkers, which would then condition the subsequent to a broader range of diagnoses than AD since amyloid and need for a LP. Importantly, the biomarker cutoff decision tau biomarkers might also be altered in different pathologi- points will likely be different than for AD detection. Such an cal situations such as non-Alzheimer's neurodegenerative approach could help clinicians in the decision to add other diseases like DLB, FTLD and Creutzfeldt-Jakob disease diagnostic tests (such as imaging), depending on the clinical (Gabelle et al. 2011; Bousiges et al. 2018; Bibl et al. 2008; evaluation of the patient. Depending on the prevalence of Lehmann et al. 2019) as well as brain damage (Alosco et al. AD as well as that of other diseases modifying CSF amyloid 2018), normal pressure hydrocephalus (Manniche et al. and tau concentrations in a cohort, the reduction in the need 2020) and cerebral amyloid angiopathy (Renard et al. 2012) for LP could be well over 50%, significantly reducing the 19 . This is illustrated in our cohorts by the fact that non-AD cost of management of these patients and limiting invasive patients represent 40–50% of the CSF pathological profiles. and unnecessary medical procedures. In conclusion, blood In this study, we evaluated the performance of blood bio- amyloid and tau biomarkers perform well in detecting non- markers for the detection of “nonpathological” (A−/T−/N−), pathological amyloid and tau CSF patterns. The importance amyloid (A+) or neurodegenerative (T+/N+) CSF profiles. and value of this “prediction” are linked to the exclusion of The main interest of the detection of a nonpathological pathologies that vary these biomarkers (AD but not only) (A−/T−/N−) profile lies in the fact that this information can and to the decision to perform a LP or not. Blood biomarkers be taken into account in the decision to perform a LP or not. can therefore represent the first step in the patient's manage- Using different analytical approaches and in two independent ment strategy, to determine whether or not other diagnostic cohorts, we show that plasma p-tau(181) and Aβ1–42/Aβ1–40 examinations by more invasive or more expensive means ratio achieve the best performance to detect nonpathologi- are necessary. cal CSF profiles. Interestingly, the amyloid ratio performed better in the Barcelona cohort, which may be explained by Supplementary Information The online version contains supplemen- tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00702-0​ 22-0​ 2474-9. the fact that this cohort has a significantly higher percent- age of positive amyloid profiles (Table 1). These plasma Author contributions SL and CD had full access to all of the data in biomarkers are also those identified as the best predictors the study and take responsibility for the integrity of the data and the of AD (Palmqvist et al. 2021), but here, the context of use accuracy of the data analysis. Concept and design: SL, CD, AL, DA. is different, and their performance are event higher than Acquisition, analysis, or interpretation of data: all authors. Drafting of the manuscript: SL and CD. Critical revision of the manuscript for for discriminating AD from non-AD. Other differences are important intellectual content: all authors. noted such as the fact that Aβ1–42 detection by IP-MS out- performed other Aβ1–42 detections, but this was not the case Funding S. Lehmann is supported by the “Claude Pompidou Foun- when considering the Aβ1–42/Aβ1–40 ratio, thus differently dation”, H2020 Marie Skłodowska-Curie Actions MIRIADE project than when AD is the performance criterion (Janelidze et al. and the European Metrology Programme for Innovation and Research Neuromet2 project. H. Zetterberg is a Wallenberg Scholar supported by 2021). As mentioned above, this could be partly explained grants from the Swedish Research Council (#2018-02532), the H2020 by the fact that diseases other than AD showed pathological European Research Council (#681712), Swedish State Support for amyloid or tau profiles. Thus, combining plasma amyloid Clinical Research (#ALFGBG-720931), the Alzheimer Drug Discov- and p-tau(181) slightly increased the performance, there- ery Foundation (ADDF), USA (#201809-2016862), the AD Strategic Fund and the Alzheimer's Association (#ADSF-21-831376-C, #ADSF- fore suggesting their complementarity. This was confirmed 21–831381-C and #ADSF-21-831377-C), the Olav Thon Foundation, when the blood biomarker performance criteria were based the Erling-Persson Family Foundation (Richard and Susan Smith Fam- on the detection of amyloid (A+) or neurodegenerative (T+/ ily Foundation), Stiftelsen för Gamla Tjänarinnor, Hjärnfonden, Swe- N+) CSF profiles (Fig. 1). Strikingly, blood p-tau(181) out- den (#FO2019-0228), the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant performs the amyloid ratio for the detection of an amyloid agreement No 860197 (MIRIADE), and the UK Dementia Research profile in CSF. This confirms the value of detecting phos- Institute at UCL. phorylated tau proteins in blood. In this work we quanti- fied blood p-tau(181) but other phosphorylated isoforms, Declarations such as p-tau(217) or p-tau(231) which have shown better diagnostic performance (Brickman et al. 2021; Barthelemy Conflict of interest Dr. Lleó has served as a consultant or at advisory boards for Fujirebio-Europe, Roche, Biogen, Zambon and Nutricia. In addition, Dr. Lleó has a patent WO2019175379 A1 Markers of synap- 13 236 C. Delaby et al. topathy in neurodegenerative disease issued. Dr. Alcolea participated identification. Alzheimers Res Ther 12(1):26. https://​doi.​org/​10.​ in advisory boards from Fujirebio-Europe and Roche Diagnostics and 1186/​s13195-​020-​00596-4 received speaker honoraria from Fujirebio-Europe, Roche Diagnostics, Bayoumy S, Verberk IMW, den Dulk B, Hussainali Z, Zwan M, van der Nutricia, Krka Farmacéutica S.L., Zambon S.A.U. and Esteve Pharma- Flier WM, Ashton NJ, Zetterberg H, Blennow K, Vanbrabant J, ceuticals S.A. In addition, Dr. Alcolea has a patent WO2019175379 Stoops E, Vanmechelen E, Dage JL, Teunissen CE (2021) Clinical A1 Markers of synaptopathy in neurodegenerative disease issued. Dr. and analytical comparison of six Simoa assays for plasma P-tau Lehmann has served as a consultant or at advisory boards for Fujire- isoforms P-tau181, P-tau217, and P-tau231. Alzheimers Res Ther bio-Europe, Roche Diagnostics, Shimadzu and Euroimmun. Dr. Zet- 13(1):198. https://​doi.​org/​10.​1186/​s13195-​021-​00939-9 terberg has served at scientific advisory boards and/or as a consult- Bibl M, Lewczuk P, Esselmann H, Mollenhauer B, Klafki HW, Welge ant for Alector, Eisai, Denali, Roche Diagnostics, Wave, Samumed, V, Wolf S, Trenkwalder C, Otto M, Kornhuber J, Wiltfang J Siemens Healthineers, Pinteon Therapeutics, Nervgen, AZTherapies, (2008) CSF amyloid-beta 1–38 and 1–42 in FTD and AD: bio- CogRx and Red Abbey Labs, has given lectures in symposia sponsored marker performance critically depends on the detergent accessible by Cellectricon, Fujirebio, Alzecure and Biogen, and is a co-founder of fraction. Proteom Clin Appl 2(10–11):1548–1556. https://d​ oi.o​ rg/​ Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part 10.​1002/​prca.​20080​0006 of the GU Ventures Incubator Program. All other authors declare that Bousiges O, Bombois S, Schraen S, Wallon D, Quillard MM, Gabelle they have no competing interests. A, Lehmann S, Paquet C, Amar-Bouaziz E, Magnin E, Miguet- Alfonsi C, Delbeuck X, Lavaux T, Anthony P, Philippi N, Blanc Open Access This article is licensed under a Creative Commons Attri- F (2018) Cerebrospinal fluid Alzheimer biomarkers can be useful bution 4.0 International License, which permits use, sharing, adapta- for discriminating dementia with Lewy bodies from Alzheimer’s tion, distribution and reproduction in any medium or format, as long disease at the prodromal stage. J Neurol Neurosurg Psychiatry as you give appropriate credit to the original author(s) and the source, 89(5):467–475. https://​doi.​org/​10.​1136/​jnnp-​2017-​316385 provide a link to the Creative Commons licence, and indicate if changes Brickman AM, Manly JJ, Honig LS, Sanchez D, Reyes-Dumeyer D, were made. The images or other third party material in this article are Lantigua RA, Lao PJ, Stern Y, Vonsattel JP, Teich AF, Airey included in the article's Creative Commons licence, unless indicated DC, Proctor NK, Dage JL, Mayeux R (2021) Plasma p-tau181, otherwise in a credit line to the material. If material is not included in p-tau217, and other blood-based Alzheimer’s disease biomarkers the article's Creative Commons licence and your intended use is not in a multi-ethnic, community study. Alzheimers Dement. https://​ permitted by statutory regulation or exceeds the permitted use, you will doi.​org/​10.​1002/​alz.​12301 need to obtain permission directly from the copyright holder. To view a DeLong ER, DeLong DM, Clarke-Pearson DL (1988) Comparing the copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. areas under two or more correlated receiver operating characteris- tic curves: a nonparametric approach. Biometrics 44(3):837–845 Dubois B, Feldman HH, Jacova C, Hampel H, Molinuevo JL, Blen- now K, DeKosky ST, Gauthier S, Selkoe D, Bateman R, Cappa References S, Crutch S, Engelborghs S, Frisoni GB, Fox NC, Galasko D, Habert MO, Jicha GA, Nordberg A, Pasquier F, Rabinovici G, Alcolea D, Clarimón J, Carmona-Iragui M, Illán-Gala I, Morenas-Rod- Robert P, Rowe C, Salloway S, Sarazin M, Epelbaum S, de ríguez E, Barroeta I, Ribosa-Nogué R, Sala I, Sánchez-Saudinós Souza LC, Vellas B, Visser PJ, Schneider L, Stern Y, Scheltens MB, Videla L, Subirana A, Benejam B, Valldeneu S, Fernández S, P, Cummings JL (2014) Advancing research diagnostic crite- Estellés T, Altuna M, Santos-Santos M, García-Losada L, Bejanin ria for Alzheimer’s disease: the IWG-2 criteria. Lancet Neu- A, Pegueroles J, Montal V, Vilaplana E, Belbin O, Dols-Icardo O, rol 13(6):614–629. https://​d oi.​o rg/​1 0.​1 016/​S 1474-​4 422(14)​ Sirisi S, Querol-Vilaseca M, Cervera-Carles L, Muñoz L, Núñez 70090-0 R, Torres S, Camacho MV, Carrió I, Giménez S, Delaby C, Rojas- Gabelle A, Roche S, Geny C, Bennys K, Labauge P, Tholance Y, Quad- Garcia R, Turon-Sans J, Pagonabarraga J, Jiménez A, Blesa R, rio I, Tiers L, Gor B, Boulanghien J, Chaulet C, Vighetto A, Croi- Fortea J, Lleó A (2019) The Sant Pau Initiative on Neurodegen- sile B, Krolak-Salmon P, Perret-Liaudet A, Touchon J, Lehmann S eration (SPIN) cohort: a data set for biomarker discovery and (2011) Decreased sAbetaPPbeta, Abeta38, and Abeta40 cerebro- validation in neurodegenerative disorders. Alzheimers Dement spinal fluid levels in frontotemporal dementia. J Alzheimers Dis (N Y) 5:597–609 JAD 26(3):553–563. https://​doi.​org/​10.​3233/​JAD-​2011-​110515 Alcolea D, Delaby C, Munoz L, Torres S, Estelles T, Zhu N, Barroeta Jack CR Jr, Bennett DA, Blennow K, Carrillo MC, Feldman HH, I, Carmona-Iragui M, Illan-Gala I, Santos-Santos MA, Altuna M, Frisoni GB, Hampel H, Jagust WJ, Johnson KA, Knopman DS, Sala I, Sanchez-Saudinos MB, Videla L, Valldeneu S, Subirana A, Petersen RC, Scheltens P, Sperling RA, Dubois B (2016) A/T/N: Pegueroles J, Hirtz C, Vialaret J, Lehmann S, Karikari TK, Ashton An unbiased descriptive classification scheme for Alzheimer dis- NJ, Blennow K, Zetterberg H, Belbin O, Blesa R, Clarimon J, ease biomarkers. Neurology 87(5):539–547. https://​doi.​org/​10.​ Fortea J, Lleo A (2021) Use of plasma biomarkers for AT(N) 1212/​WNL.​00000​00000​002923 classification of neurodegenerative dementias. J Neurol Neurosurg Jack CR, Bennett DA, Blennow K, Carrillo MC, Dunn B, Haeber- Psychiatry. https://​doi.​org/​10.​1136/​jnnp-​2021-​326603 lein SB, Holtzman DM, Jagust W, Jessen F, Karlawish J, Liu E, Alosco ML, Tripodis Y, Fritts NG, Heslegrave A, Baugh CM, Conneely Molinuevo JL, Montine T, Phelps C, Rankin KP, Rowe CC, Schel- S, Mariani M, Martin BM, Frank S, Mez J, Stein TD, Cantu RC, tens P, Siemers E, Snyder HM, Sperling R, Elliott C, Masliah E, McKee AC, Shaw LM, Trojanowski JQ, Blennow K, Zetterberg Ryan L, Silverberg N (2018) NIA-AA research framework: toward H, Stern RA (2018) Cerebrospinal fluid tau, Aβ, and sTREM2 in a biological definition of Alzheimer’s disease. Alzheimers Dement Former National Football League Players: modeling the relation- 14(4):535–562. https://​doi.​org/​10.​1016/j.​jalz.​2018.​02.​018 ship between repetitive head impacts, microglial activation, and Janelidze S, Teunissen CE, Zetterberg H, Allue JA, Sarasa L, Eichen- neurodegeneration. Alzheimers Dement 14(9):1159–1170 laub U, Bittner T, Ovod V, Verberk IMW, Toba K, Nakamura Barthelemy NR, Bateman RJ, Hirtz C, Marin P, Becher F, Sato C, A, Bateman RJ, Blennow K, Hansson O (2021) Head-to-head Gabelle A, Lehmann S (2020) Cerebrospinal fluid phospho-tau comparison of 8 plasma amyloid-beta 42/40 assays in Alzheimer T217 outperforms T181 as a biomarker for the differential diag- disease. JAMA Neurol. https://​doi.​org/​10.​1001/​jaman​eurol.​2021.​ nosis of Alzheimer’s disease and PET amyloid-positive patient 3180 13 Blood amyloid and tau biomarkers as predictors of cerebrospinal fluid profiles 237 Lehmann S, Schraen S, Quadrio I, Paquet C, Bombois S, Delaby C, Manniche C, Simonsen AH, Hasselbalch SG, Andreasson U, Zetterberg Dorey A, Dumurgier J, Hirtz C, Krolak-Salmon P, Laplanche JL, H, Blennow K, Høgh P, Juhler M, Hejl AM (2020) Cerebrospinal Moreaud O, Peoc’h K, Rouaud O, Sablonniere B, Thouvenot E, fluid biomarkers to differentiate idiopathic normal pressure hydro- Touchon J, Vercruysse O, Hugon J, Gabelle A, Pasquier F, Perret- cephalus from subcortical ischemic vascular disease. J Alzheimers Liaudet A (2013) Impact of harmonization of collection tubes on Dis JAD 75(3):937–947. https://​doi.​org/​10.​3233/​jad-​200036 Alzheimer’s disease diagnosis. Alzheimers Dement. https://​doi.​ McKhann GM, Knopman DS, Chertkow H, Hyman BT, Jack CR Jr, org/​10.​1016/j.​jalz.​2013.​06.​008 Kawas CH, Klunk WE, Koroshetz WJ, Manly JJ, Mayeux R, Mohs Lehmann S, Delaby C, Boursier G, Catteau C, Ginestet N, Tiers L, RC, Morris JC, Rossor MN, Scheltens P, Carrillo MC, Thies B, Maceski A, Navucet S, Paquet C, Dumurgier J, Vanmechelen E, Weintraub S, Phelps CH (2011) The diagnosis of dementia due to Vanderstichele H, Gabelle A (2018) Relevance of Abeta42/40 Alzheimer’s disease: recommendations from the National Insti- ratio for detection of alzheimer disease pathology in clinical rou- tute on Aging-Alzheimer’s Association workgroups on diagnostic tine: the PLMR scale. Front Aging Neurosci 10:138. https://​doi.​ guidelines for Alzheimer’s disease. Alzheimers Dement 7(3):263– org/​10.​3389/​fnagi.​2018.​00138 269. https://​doi.​org/​10.​1016/j.​jalz.​2011.​03.​005 Lehmann S, Dumurgier J, Ayrignac X, Marelli C, Alcolea D, Ormae- Nakamura A, Kaneko N, Villemagne VL, Kato T, Doecke J, Dore V, chea JF, Thouvenot E, Delaby C, Hirtz C, Vialaret J, Ginestet Fowler C, Li QX, Martins R, Rowe C, Tomita T, Matsuzaki K, N, Bouaziz-Amar E, Laplanche JL, Labauge P, Paquet C, Lleo Ishii K, Arahata Y, Iwamoto S, Ito K, Tanaka K, Masters CL, A, Gabelle A, Alzheimer’s Disease Neuroimaging I (2020) Cer- Yanagisawa K (2018) High performance plasma amyloid-beta ebrospinal fluid A beta 1–40 peptides increase in Alzheimer’s biomarkers for Alzheimer’s disease. Nature 554(7691):249–254. disease and are highly correlated with phospho-tau in control https://​doi.​org/​10.​1038/​natur​e25456 individuals. Alzheimers Res Ther 12(1):123. https://​doi.​org/​10.​ Palmqvist S, Tideman P, Cullen N, Zetterberg H, Blennow K, Alzhei- 1186/​s13195-​020-​00696-1 mer’s Disease Neuroimaging I, Dage JL, Stomrud E, Janelidze Lehmann S, Paquet C, Malaplate-Armand C, Magnin E, Schraen S, S, Mattsson-Carlgren N, Hansson O (2021) Prediction of future Quillard-Muraine M, Bousiges O, Delaby C, Dumurgier J, Hugon Alzheimer’s disease dementia using plasma phospho-tau com- J, Sablonniere B, Blanc F, Wallon D, Gabelle A, Laplanche JL, bined with other accessible measures. Nat Med 27(6):1034–1042. Bouaziz-Amar E, Peoc’h K (2019) Diagnosis associated with Tau https://​doi.​org/​10.​1038/​s41591-​021-​01348-z higher than 1200pg/mL: insights from the clinical and laboratory Renard D, Castelnovo G, Wacongne A, Le Floch A, Thouvenot E, practice. Clin Chim Acta Int J Clin Chem 495:451–456. https://​ Mas J, Gabelle A, Labauge P, Lehmann S (2012) Interest of CSF doi.​org/​10.​1016/j.​cca.​2019.​04.​081 biomarker analysis in possible cerebral amyloid angiopathy cases Lewczuk P, Ermann N, Andreasson U, Schultheis C, Podhorna J, defined by the modified Boston criteria. J Neurol 259(11):2429– Spitzer P, Maler JM, Kornhuber J, Blennow K, Zetterberg H 2433. https://​doi.​org/​10.​1007/​s00415-​012-​6520-8 (2018) Plasma neurofilament light as a potential biomarker of neurodegeneration in Alzheimer’s disease. Alzheimers Res Ther Publisher's Note Springer Nature remains neutral with regard to 10(1):71. https://​doi.​org/​10.​1186/​s13195-​018-​0404-9 jurisdictional claims in published maps and institutional affiliations. Mamtani MR, Thakre TP, Kalkonde MY, Amin MA, Kalkonde YV, Amin AP, Kulkarni H (2006) A simple method to combine multiple molecular biomarkers for dichotomous diagnostic classification. BMC Bioinform 7:442. https://​doi.​org/​10.​1186/​ 1471-​2105-7-​442 13