Impact of harmonization of collection tubes on Alzheimer's disease diagnosis

Alzheimer’s & Dementia - (2013) 1–7 Research Article Impact of harmonization of collection tubes on Alzheimer’s disease diagnosis Sylvain Lehmanna,*, Susanna Schraenb, Isabelle Quadrioc, Claire Paquetd,e, St
ephanie Bomboisf, Constance Delabya,g, Aline Doreyh, Julien Dumurgierd, Christophe Hirtza, Pierre Krolak-Salmonh, Jean-Louis Laplanchei, Olivier Moreaudj, Katell Peoc’hi, Olivier Rouaudk, Bernard Sablonniereb, Eric Thouvenotl, Jacques Touchonl, Olivier Vercruyssef, Jacques Hugond,e, Audrey Gabellel, Florence Pasquierf, Armand Perret-Liaudetc a CHU de Montpellier, IRB, INSERM-UM1 1040, Montpellier, France Universit
e Lille Nord de France, INSERM U837, Lille University Hospital, Center of Biology and Pathology, Lille, France c Service de Neurobiologie, Hospices Civils de Lyon, Universit
e Lyon 1–CNRS UMR5292–INSERM U1028, Lyon, France d CMRR Paris Nord Ile de France, Lariboisiere-Fernand Widal Hospital, APHP, University 7-Denis Diderot, Paris, France e INSERM U942, Paris, France f Universit
e Lille Nord de France, EA1046, DISTALZ, Memory Center, CHU 59000, Lille, France g Paris 7, Facult
e de M
edecine Xavier Bichat, Paris, France h CMRR Lyon, H^opital des Charpennes, HCL Lyon, Lyon, France i Laboratoire de Biochimie, Lariboisiere-Fernand Widal Hospital, APHP, University Paris 7-Denis Diderot, University Paris Descartes, Paris, France j CMRR, CHU de Grenoble, Grenoble, France k CMRR, CHU Dijon, Dijon, France l CMRR, CHU de Montpellier, Montpellier, France b Abstract Objective: The objective of this study was to analyze differences in biomarker outcomes before and after harmonization of cerebrospinal fluid (CSF) collection tubes in Alzheimer’s disease (AD) diagnosis. Methods: We analyzed data from French memory centers that switched from different CSF collection tubes to a common one. A total of 1966 patients were included in the study. CSF concentrations of b-amyloid 1–42 (Ab42), total tau, and phosphorylated tau (p-tau181) were measured in each center using the same commercial enzyme-linked immunoabsorbent assay (ELISA) kits. The diagnostic value of CSF biomarkers according to the type of tube used was then assessed using different cutoffs. Results: The predictive value of Ab42 was highly affected by the type of collection tube used. The optimal cutoff value for p-tau181 appeared not to be affected by the type of collection tube whereas that of total tau was slightly changed. New optimal cutoff values were then computed. Conclusions: In a routine clinical environment, the selection of the collection tube and biomarker cutoff value makes a major difference in AD biological diagnosis. The use of a common collection tube among different centers will reduce the risk of misdiagnosis and incorrect patient stratification. Ó 2013 The Alzheimer’s Association. All rights reserved. Keywords: Cerebrospinal fluid; Biomarkers; Alzheimer’s disease; Diagnosis; Preanalytics 1. Introduction Intense research efforts have been made to develop biomarkers for the central pathogenic processes in Alzheimer’s *Corresponding author. Tel.: 133-4-67-33-73-23; Fax: 133-4-67-3369-21. E-mail address: s-lehmann@chu-montpellier.fr disease (AD) that can be used as diagnostic tools, especially in an early stage or atypical forms of the disease. As the concept of a neurodegenerative continuum throughout the course of AD [1] emerges, it seems clear that the amyloid markers have a major diagnostic role, especially in young patients and those with atypical clinical presentations. It is also well known that biochemical changes in the brain are reflected in cerebrospinal fluid (CSF). Numerous studies have shown 1552-5260/$ - see front matter Ó 2013 The Alzheimer’s Association. All rights reserved. http://dx.doi.org/10.1016/j.jalz.2013.06.008 2 S. Lehmann et al. / Alzheimer’s & Dementia - (2013) 1–7 Fig. 1. Boxplots showing median values and quartiles for (A) Ab42, (B) tau, (C) p-tau181, and (D) IATI in AD and NAD patient samples collected in the initial tubes (tubes A–D) and the new common collection tube (tube S). AD, Alzheimer’s disease; NAD, non-AD patients; IATI, Innotest amyloid tau index; Ab42, bamyloid 1–42; p-tau181, tau phosphorylated at threonine 181. that AD patients display characteristic CSF changes with decreased levels of b-amyloid 1–42 (Ab42) and elevated levels of total tau (tau) protein and its form that is phosphorylated at threonine 181 (p-tau) [2,3]. CSF biomarkers correlated with the extent of neuropathological lesions and showed good sensitivity and specificity for clinically diagnosed AD versus controls in several monocentric cohorts [4] and for differential diagnosis [5,6]. To optimize and expand the use of biomarkers in clinical practice, workshops on the harmonization and standardization of CSF procedures stress the importance of preanalytical and analytical factors in minimizing variability and optimizing the diagnostic value of CSF biomarkers for patients [7–10]. This has also been emphasized in our recent intersite study of a large representative population of patients coming from French memory clinics [11,12]. However, it is noteworthy that using different collection tubes has been demonstrated as an important confounder and might result in misdiagnosis of AD [13,14]. In this study, we investigated differences in biomarker outcomes before and after harmonization of the collection tubes. 2. Materials and methods Between January 2008 and December 2011, 1966 patients who had a lumbar puncture were recruited from four French clinical and research memory centers (Montpellier, Lille, Lyon, and Paris) specialized in the diagnosis and care management of patients with cognitive disorders. These centers use the same diagnostic procedure and criteria [11]. All patients had a thorough examination, including clinical and neuropsychological evaluations and brain imaging. Patients were classified into two groups: AD (as defined by the National Institute of Neurological and Communicative Disorders and Stroke [NINCDS]-Alzheimer’s Disease and Related Disorders Association [ADRDA] criteria [15]) and non-AD (NAD) patients. NAD diagnoses (i.e., frontotemporal lobar degeneration, semantic dementia, dementia with Lewy bodies and Parkinson’s disease, progressive supranuclear palsy, amyotrophic lateral sclerosis, normal pressure hydrocephalus, and psychiatric disorder) were defined according to international criteria. Patients with mild cognitive impairment, as well as those with AD with a mixed phenotype or those who may correspond to a specific/early form of AD, were excluded from the cohorts (mixed and vascular dementia, corticobasal degeneration, primary progressive aphasia, amyloid angiopathy). CSF was obtained during the routine follow-up of patients with cognitive complaints. It was collected in different polypropylene tubes (Supplemental Table 1) under standardized conditions, preferably between 9:00 a.m. 3 S. Lehmann et al. / Alzheimer’s & Dementia - (2013) 1–7 P < .01 P < .07 P < .25 P < .01 Fig. 2. (A) AUC values (mean 6 SD) for each biomarker when using the initial tubes (tubes A–D) and the new common collection tube (tube S). No significant differences (P ..05, Student’s test) were observed between initial and new tubes for each biomarker. (B) Sensitivity for AD detection using two different cutoff values for population of the Montpellier, Lille, and Lyon center collected in the new tube S. (C) Values of optimal cutoffs (mean 6 SD) for each biomarker when using the initial tubes (tubes A-D) and the new common collection tube (tube S). Significant differences (Student’s test) were observed between initial and new tubes for Ab42 and IATI. AUC, area under the receiver operating characteristic curve; AD, Alzheimer’s disease; Ab42, b-amyloid 1–42; IATI, Innotest amyloid tau index. and 1:00 p.m., to minimize the influence of diurnal variation on CSF Ab42 levels. Each CSF sample was sent to the local laboratory within 4 hours after collection and was centrifuged at 1000g for 10 minutes at 4 C. A small amount of CSF was used for routine analyses, including total cell count, bacteriological exam, total protein, and glucose levels. CSF was aliquoted in polypropylene tubes of 1.5 mL and stored at –80 C until further analysis (variation in biomarker concentrations linked to these storage tubes was known to be minimal, i.e., ,10% [A. PerretLiaudet, personal communication]). CSF Ab42, total tau, and phosphorylated tau (p-tau) were measured using the standardized, commercially available InnotestÒ sandwich enzyme-linked immunoabsorbent assay (ELISA) according to manufacturer’s procedures (Innogenetics, Ghent, Belgium). CSF samples were simultaneously analyzed for all three biomarkers but with different batches of kits over the years. The intra-assay variability based on replicates was less than 5% for the three biomarkers. In each laboratory, the interassay variability was estimated using internal quality controls that were included in all series. In Mont- pellier, it was estimated to be 10.1%, 13.2%, and 14.4% for Ab42, tau, and p-tau, respectively. The other centers found similar ranges of variability as recently published [11]. The latter study was the opportunity to standardize our protocols and minimize preanalytical and analytical confounding factors [7]. In addition, the laboratories performed two external quality controls: one run by a working group of the Soci
et
e Française de Biologie Clinique (French Society of Clinical Biology) and the other by the Alzheimer’s Association [16]. A total of 1149 samples (616 in AD patients, 533 in NAD patients) were collected in initial tubes (tubes A, B, C, and D in Montpellier, Lille, Lyon, and Paris, respectively), and 817 (487 AD, 330 NAD) samples were collected using a second common tube (tube S) selected for its low adsorption for Ab42 (Supplemental Table 1) [17]. CSF tau, p-tau181, and Ab42 concentrations were measured. In France, this retrospective study that is based on routine biological analyses is not considered as biomedical research and does not require informed consent. However, the different centers were authorized to handle personal data by the Commission 4 S. Lehmann et al. / Alzheimer’s & Dementia - (2013) 1–7 Nationale de l’Informatique et des Libert
es (French Data Protection Authority). Statistical analyses were performed using MedCalc software (version 11.3). Receiver operating characteristic (ROC) curves were used to represent the sensitivity and specificity for AD at different cutoff values. Analyses were stratified by center and type of collection tube used. 3. Results Population demographics data and CSF biomarker and Ab42/tau index (Innotest amyloid tau index [IATI]) [18] values obtained from the different centers using the different collection tubes are presented in Supplemental Table 2. Biomarker concentrations in AD and NAD patient samples collected before (tubes A–D) and after harmonization of the collection tubes (tube S) are presented in Figure 1 and summarized in Supplemental Table 2. The change of collection tube was characterized by a general increase in Ab42 and IATI values for AD and NAD populations. The effect on tau was smaller, whereas p-tau values remained comparable. The characteristics of the collection tubes, which have already been documented [17], are summarized in Supplemental Table 1. The area under the ROC curve (AUC) values for the different biomarkers were computed in each center and then grouped together before and after harmonization of the collection tubes (Fig. 2A). Differences in AUC values were found between biomarkers, but their individual values were not significantly different before and after harmonization of the collection tubes (P . .05). The optimal cutoffs for each biomarker were defined using the highest Youden index, which was selected to maximize sensitivity and specificity on the basis of ROC curve analyses. The use of different cutoffs had a major effect on sensitivity and specificity for AD detection, with an increase in the Ab42 sensitivity from 28/42% to 68/80% when using optimal cutoffs for tube S (Fig. 2B and Supplemental Table 1). Overall, the optimal cutoff values before and after harmonization of the collection tubes were significantly different for Ab42 and IATI (Fig. 2C and Supplemental Table 1). 4. Discussion To our knowledge, this is the first report that evaluated the effect of the harmonization of collection tubes in routine practice in a large multicenter cohort (which is part of the Paris-North, Lille and Montpellier [PLM] study). With regards to the different preanalytical properties of the tubes (Supplemental Table 1), significant changes in the values of the different biomarkers were observed. Because Ab42 is especially affected by the nature of the tubes [14,17], this harmonization resulted in important changes in Ab42 and IATI mean values in AD and NAD patients (Figure 1 and Supplemental Table 2). Changes in tau values were also observed, but to a lesser extent. These changes paralleled those observed in controlled studies [13] in which “multiplication factors” were computed based on data obtained by measuring the same series of CSF samples collected in different sets of tubes (Supplemental Table 1). However, the use of a multiplication factor to later combine data from different cohorts should be considered with much caution because the variation of concentration might not be linear and might be affected by other factors than the type of collection tube [7– 10]. The fact that the harmonization did not result in a significant change in the AUCs for the different biomarkers (Fig. 2A) suggests that the nature of the tube does not affect the intrinsic diagnostic value of each biomarker. Although the use of a common tube did not result in significant changes in individual biomarker AUCs, a clear change was observed in optimal cutoffs, mainly for Ab42 and IATI (Fig. 2C and Supplemental Table 3). Differences in optimal cutoffs still existed between centers using the same collection tube (Supplemental Table 3), which was most likely due to differences in the populations of the different cohorts. The major variations of sensitivity and specificity for AD detection that resulted from using different cutoff values (Supplemental Table 3) illustrated the importance of adjusting the cutoffs according to the type of collection tube being used. For example, the clinical effect of tube harmonization can be illustrated by computing the number of misdiagnoses linked to the use of former Ab42 cutoffs instead of the optimal ones when using the new tube S (Fig. 2B). This resulted in 204 additional false-negative diagnoses of AD out of 487 (42%). On the other hand, the number of true negative diagnoses increased only by 56 of 330 NAD (17%) patients. When the new collection tube (tube S) was used in the different centers, differences in optimal cutoffs were still observed (which were linked to the clinical cohorts and other confounding factors [7,11]). Nevertheless, following the implementation of the standard operating procedures [8,11] and tube harmonization as described in this manuscript, we are now considering using common cutoff values as follows: 700 ng/L for Ab42, 400 ng/L for total tau, 60 ng/L for p-tau181, and 1 for IATI. In conclusion, our study confirmed the risk of patient misclassification if the type of collection tube is not carefully taken into account. It also showed the interest of harmonizing the collection tubes and defining common cutoffs in a defined preanalytical context to improve biomarker predictive values and reduce bias in AD diagnosis and patient stratification. Acknowledgments This work was supported in part by grants from the Association France Alzheimer and the French Alzheimer Plan (“Plan Alzheimer”). The authors thank Val
erie Macioce for editing the manuscript. S. Lehmann et al. / Alzheimer’s & Dementia - (2013) 1–7 All financial and material support for this research is academic, and the authors declare no potential conflicts. RESEARCH IN CONTEXT 1. Systematic review: CSF biomarkers (Ab42, tau proteins) are being more and more used by clinicians as an aid to the diagnosis of AD. Variability between centers and studies is a major issue that impairs the use of these biomarkers. Preanalytics and, in particular, the type of CSF collection tube used has been identified as a major issue regarding this variability. 2. Interpretation: We first evidenced the “real” clinical effect on the routine diagnosis of AD of the type of collection tube used in different centers. We then report the effect of the homogenization of the preanalytics between the three centers (in particular a modification of the collection tube). This led us to define new guidelines and cutoff values for the different biomarkers to be used within the same preanalytical context. 3. Future directions: An important question is the need to share a common collection tube and/or to define “correction” factors between centers. More importantly, we need to find a way (algorithm, etc.) to use (adapt) cutoff values of CSF biomarkers to obtain more homogeneous and reliable results between centers to help in AD diagnosis and patient inclusion in clinical studies. References [1] Dubois B, Feldman HH, Jacova C, Cummings JL, Dekosky ST, Barberger-Gateau P, et al. Revising the definition of Alzheimer’s disease: a new lexicon. Lancet Neurol 2010;9:1118–27. [2] Blennow K, de Leon MJ, Zetterberg H. Alzheimer’s disease. The Lancet 2006;368:387–403. [3] Bateman RJ, Xiong C, Benzinger TL, Fagan AM, Goate A, Fox NC, et al. Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N Engl J Med 2012;367:795–804. [4] Johansson P, Mattsson N, Hansson O, Wallin A, Johansson JO, Andreasson U, et al. Cerebrospinal fluid biomarkers for Alzheimer’s disease: diagnostic performance in a homogeneous mono-center population. J Alzheimers Dis 2011;24:537–46. 5 [5] Bibl M, Mollenhauer B, Lewczuk P, Esselmann H, Wolf S, Otto M, et al. Cerebrospinal fluid tau, p-tau 181 and amyloid-beta(38/40/42) in frontotemporal dementias and primary progressive aphasias. Dement Geriatr Cogn Disord 2010;31:37–44. [6] Gabelle A, Roche S, Geny C, Bennys K, Labauge P, Tholance Y, et al. Decreased sAPPb, Ab38, and Ab40 cerebrospinal fluid levels in frontotemporal dementia. J Alzheimers Dis 2011;26:553–63. [7] Bjerke M, Portelius E, Minthon L, Wallin A, Anckars€ater H, Anckars€ater R, et al. Confounding factors influencing amyloid beta concentration in cerebrospinal fluid. Int J Alzheimers Dis 2010; 2010:986310. [8] Del Campo M, Mollenhauer B, Bertolotto A, Engelborghs S, Hampel H, Simonsen AH, et al. Recommendations to standardize preanalytical confounding factors in Alzheimer’s and Parkinson’s disease cerebrospinal fluid biomarkers: an update. Biomark Med 2012; 6:419–30. [9] Teunissen CE, Verwey NA, Kester MI, van Uffelen K, Blankenstein MA. Standardization of assay procedures for analysis of the CSF biomarkers amyloid beta(1-42), tau, and phosphorylated tau in Alzheimer’s disease: report of an international workshop. Int J Alzheimers Dis 2010;2010:635053. [10] Shaw LM, Vanderstichele H, Knapik-Czajka M, Figurski M, Coart E, Blennow K, et al. Qualification of the analytical and clinical performance of CSF biomarker analyses in ADNI. Acta Neuropathol 2011;121:597–609. [11] Dumurgier J, Vercruysse O, Paquet C, Bombois S, Chaulet C, Laplanche JL, et al. Intersite variability of CSF Alzheimer’s disease biomarkers in clinical setting. Alzheimers Dement 2012; 9:406–13. [12] Gabelle A, Dumurgier J, Vercruysse O, Paquet C, Bombois S, Laplanche JL, et al. Impact of the 2008-2012 French Alzheimer Plan on the use of cerebrospinal fluid biomarkers in research memory center: the PLM Study. J Alzheimers Dis 2013;34: 297–305. [13] Perret-Liaudet A, Pelpel M, Tholance Y, Dumont B, Vanderstichele H, Zorzi W, et al. Risk of Alzheimer’s disease biological misdiagnosis linked to cerebrospinal collection tubes. J Alzheimers Dis 2012; 31:13–20. [14] Lewczuk P, Beck G, Esselmann H, Bruckmoser R, Zimmermann R, Fiszer M, et al. Effect of sample collection tubes on cerebrospinal fluid concentrations of tau proteins and amyloid beta peptides. Clin Chem 2006;52:332–4. [15] McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA work group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 1984;34:939–44. [16] Mattsson N, Andreasson U, Persson S, Arai H, Batish SD, Bernardini S, et al. The Alzheimer’s Association external quality control program for cerebrospinal fluid biomarkers. Alzheimers Dement 2011;7:386–395.e36. [17] Perret-Liaudet A, Pelpel M, Tholance Y, Dumont B, Vanderstichele H, Zorzi W, et al. Cerebrospinal fluid collection tubes: a critical issue for Alzheimer disease diagnosis. Clin Chem 2012;58:787–9. [18] Hulstaert F, Blennow K, Ivanoiu A, Schoonderwaldt HC, Riemenschneider M, De Deyn PP, et al. Improved discrimination of AD patients using beta-amyloid(1-42) and tau levels in CSF. Neurology 1999;52:1555–62. 5.e1 S. Lehmann et al. / Alzheimer’s & Dementia - (2013) 1–7 Supplemental Table 1 List and characteristics of collection tubes Tube Center Manufacturer Catalog numbers Ab42 ! factor to S tube Tau ! factor to S tube P-tau ! factor to S tube A B C D S Montpellier Lille Lyon Paris Montpellier Lille Lyon Greiner Becton Dickinson VWR CML Sarstedt 18 82 81 Falcon 35 2097 216.0154 TC10PCS 62.610.201* 1.56 (1.40–1.62) 1.72 (1.54–1.92) 1.65 (1.44–1.72) NP NP 1.03 (0.99–1.06) 1.04 (0.97–1.09) 1.01 (0.98–1.02) NP NP 0.95 (0.89–0.97) 1.00 (0.98–1.02) 0.94 (0.84–1.05) NP NP Abbreviation: NP, not performed. NOTE. Collection tubes used in the different centers and multiplication factors obtained by measuring the same series of CSF samples collected in two different sets of tubes. *The same tubes can be ordered in sterile, individual packaging (catalog number 62.610.018). Supplemental Table 2 Demographic and biological characteristics of the different populations Summary Montpellier Statistics table Age Sex (%M) Ab42 tau p-tau MMSE N AD Tube A n 5 129 Age Sex (%M) Ab42 tau p-tau MMSE Age Sex (%M) Ab42 tau p-tau MMSE Age Sex (%M) Ab42 tau p-tau MMSE N NAD Tube A n 5 215 N AD Tube S n 5 142 N NAD Tube S n 5 147 Lille Mean 69.7 47% 505 611 85.9 21.9 Mean 64.1 53% 706 291 44.8 20.7 Mean 71.1 49% 654 702 86 20.7 Mean 63.4 55% 999 310 38.4 21.1 SD 8.8 224 327 40.2 5.5 SD 13.6 266 233 23.7 7.1 SD 10.1 256 727 37.7 7.4 SD 13.6 373 241 18.6 6.9 N AD Tube B n 5 143 N NAD Tube B n 5 128 N AD Tube S n 5 73 N NAD Tube S n 5 51 Lyon Mean 68.3 37% 338 608 98.1 18.1 Mean 67.3 51% 494 273 52.6 21.3 Mean 67 47% 603 778 101.5 19.6 Mean 64.6 59% 974 284 46.5 21.2 SD 9.0 162 336 46.9 6.5 SD 10.7 192 197 28.7 5.5 SD 9.5 245 364 41.2 6.3 SD 10.7 355 149 15.8 6.0 N AD Tube C n 5 226 N NAD Tube C n 5 163 N AD Tube S n 5 272 N NAD Tube S n 5 141 Paris Mean 69.7 45% 388 714 96.6 18.9 Mean 66.2 53% 549 301 47.4 21.1 Mean 71.8 43% 567 761 93.4 20.6 Mean 69.7 53% 808 373 52.3 22 SD 9.9 178 476 42.4 5.5 SD 10.3 232 265 29.0 5.7 SD 9.4 N AD Tube D n 5 118 N NAD Tube D n 5 53 Mean 73.6 38% 440 598 99 19.4 Mean 62.1 49% 686 253 48.6 23 SD 8.8 188.8826 295 40.4 5.6 SD 13.1 243 226 23.1 5.3 243 437 36.5 5.5 SD 9.8 317 281 27.5 6.6 Abbreviations: M, male; CSF, cerebrospinal fluid; MMSE, Mini-Mental State Examination; AD, Alzheimer’s disease; NAD, non-AD patients; Ab42, b-amyloid 1–42; p-tau, phosphorylated tau. NOTE. Means and standard deviations of demographics (age, sex), CSF biomarker levels (Ab42, tau and p-tau181), and MMSE scores for patients with AD or NAD diagnosis in the different centers (Montpellier, Lille, Lyon, Paris) after CSF collection with the different collection tubes (A, B, C, D, S; see Supplemental Table 1). 5.e2 S. Lehmann et al. / Alzheimer’s & Dementia - (2013) 1–7 Supplemental Table 3 Performance of the biomarkers in the different populations Mpt-A Ab42 (ng/L) Sensitivity Specificity Optimal cutoff Tau (ng/L) Sensitivity Specificity Optimal cutoff p-tau (ng/L) Sensitivity Specificity Optimal cutoff IATI Sensitivity Specificity Optimal cutoff Mpt-A 500 700 66% 86% 80% 49% 572 350 400 74% 68% 80% 85% .351 60 74% 85% .63 0.8 1 80% 84% 83% 73% 0.861 Lille-B Lille-B 500 700 87% 94% 50% 12% 380 350 400 73% 65% 84% 91% .352 60 77% 82% .59 0.8 1 88% 92% 71% 52% 0.598 Lyon-C Lyon-C Paris-D Paris-D 500 82% 56% 700 92% 27% 500 81% 74% 700 92% 42% 400 75% 85% 350 81% 87% 475 350 81% 79% 494 400 73% 91% .385 60 81% 82% .56 .282 60 90% 87% .63 0.8 1 85% 92% 67% 55% 0.711 0.8 1 87% 91% 83% 74% 0.743 Mpt-S Mpt-S 500 700 27% 68% 90% 76% 831 350 500 88% 61% 74% 88% .386 60 80% 93% .55 0.8 1 67% 87% 86% 81% 1.028 Lille-S Lille-S 500 600 32% 60% 90% 86% 746 350 400 88% 85% 79% 83% .372 60 85% 86% .62 0.8 1 82% 90% 90% 86% 0.921 Lyon-S Lyon-S 500 46% 81% 700 81% 58% 669 350 400 92% 88% 63% 71% .416 60 88% 75% .58 0.8 1 86% 91% 76% 74% 0.979 Abbreviations: AD, Alzheimer’s disease; IATI, Innotest amyloid tau index; Mpt, Montpellier; p-tau; phosphorylated tau; Ab42, b-amyloid 1–42; p-tau181, tau phosphorylated at threonine 181. NOTE. For each biomarker, in each cohort (Montpellier, Lille, Lyon, Paris with tubes A–D and S), the sensitivity and specificity for AD detection were calculated at different cutoffs: 500 and 700 ng/L for Ab42, 350 and 400 ng/L for tau, and 60 ng/L for p-tau181. These cutoff values were selected based on the mean of the optimal cutoffs observed when using the former (A–D) and new (S) tubes. The optimal cutoffs for each population were defined using the highest Youden index selected to maximize sensitivity and specificity based on receiver operating characteristic curve analyses.