Heterogeneity of Cerebrospinal Fluid Biomarkers Profiles in Individuals with Distinct Levels of Cognitive Decline: A Cross-Sectional Study

Abstract

Background:

Decreased cerebrospinal fluid (CSF) concentrations of the amyloid-β (Aβ), along with increased total (T-tau) and phosphorylated tau protein (P-tau), are widely accepted as core biomarkers of Alzheimer’s disease (AD) pathology. Nonetheless, there are a few remaining caveats that still preclude the full incorporation of AD biomarkers into clinical practice.

Objective:

To determine the frequency of clinical-biological mismatches in a clinical sample of older adults with varying degrees of cognitive impairment.

Methods:

204 participants were enrolled for a cross-sectional assessment and allocated into diagnostic groups: probable AD (n = 60, 29.4%); MCI (n = 84, 41.2%); or normal cognition (NC, n = 60, 29.4%). CSF concentrations of Aβ₄₂, T-tau, and ¹⁸¹Thr-P-tau were determined, and Aβ₄₂/P-tau ratio below 9.53 was used as a proxy of AD pathology. The AT(N) classification was further used as a framework to ascertain the biological evidence of AD.

Results:

The majority (73.7%) of patients in the AD group had the Aβ₄₂/P-tau ratio below the cut-off score for AD, as opposed to a smaller proportion in the MCI (42.9%) and NC (23.3%) groups. In the latter, 21 subjects (35%) were classified as A+, 28 (46.7%) as T+, and 23 (38.3%) as N + . In the AD group, 66.7%of the cases were classified as A+, 78.3%as T+, and 80%as N+.

Conclusion:

Analysis of CSF biomarkers was able to discriminate between AD, MCI, and NC. However, clinical-biological mismatches were observed in a non-negligible proportion of cases.

Keywords

Alzheimer’s disease biomarkers cerebrospinal fluid dementia mild cognitive impairment

INTRODUCTION

In 2011, the National Institute on Aging and the Alzheimer’s Association (NIA-AA) revised the diagnostic criteria for Alzheimer’s disease (AD) and included biomarkers as important tools to aid in the clinical diagnosis [1]. Further, the NIA-AA proposed a new diagnostic framework for research purposes, the AT(N) classification. The biomarkers of AD were divided into three main categories: markers of amyloid pathology (A), phosphorylated tau pathology (T), and neurodegeneration (N). The presence of amyloid pathology (A+) would define the AD continuum, and the association of A + with the presence of phosphorylated tau pathology (T+) would define the diagnosis of AD. The presence of neurodegeneration (N+) would complete the characterization of the disease, presenting as a marker of severity in all stages [2]. This classification framework defines AD using the neuropathological changes based on biomarkers and consolidates the “amyloid cascade hypothesis” [3, 4]. Patients having significant changes in biomarkers tend to present a more pronounced course of cognitive decline, as well as a worse response to treatment with acetylcholinesterase inhibitors [5]. Increment in mortality rate has also been reported [6]. In addition, potential effective treatments for AD rely on early detection of the disease with the use of these biomarkers [7].

The three core cerebrospinal fluid (CSF) biomarkers of AD have been studied for nearly 20 years. These biomarkers have a strong correlation with the disease underlying pathology with a good diagnostic accuracy, with a high specificity and sensitivity for AD diagnosis [8, 9]. Considering only the core CSF biomarkers, the biomarker of amyloid accumulation (A) is the 42 aminoacid-long amyloid-β peptide (Aβ₄₂), whereas the biomarker of tau pathology (T) is ¹⁸¹Thr-phosphorylated-tau protein (P-tau), and the biomarker of neurodegeneration (N) is total tau protein (T-tau) [10]. Reduced Aβ₄₂ and elevated T-tau and P-tau are detected in the CSF of AD patients and are defined as specific markers of disease pathology. Although not a universally available tool, lumbar puncture for CSF collection is generally a safe procedure with low incidence of complications [11, 12].

The identification of the CSF pathological signature of AD [13 –16] has been considered a critical strategy to confirm the clinical diagnosis of AD. The definition of a pathological signature of AD increases the power of diagnosis based on CSF biomarkers. In addition, there is a growing interest in the development of studies with focus on early stages of the disease before the onset of symptoms. The CSF biomarkers are quite specific for identifying an individual at preclinical AD [13]. The preclinical stage of AD represents a “silent” phase when neuropathological changes are still at the stage at which the person exhibits normal cognition [17]. The analysis of CSF biomarkers during preclinical stages of AD can improve the diagnostic accuracy and help in proposing timely interventions.

However, it has become evident that the relation between clinical symptoms and disease pathology varies [10]. There are multiple neuropathological profiles of CSF biomarkers considering the different clinical phenotypes. The clinical course of AD, from mild cognitive impairment (MCI) to dementia, reflects a multifaceted progressive disease, and can therefore be considered a result of several pathological changes that affect the brain [18, 19]. Studies have shown multiple neuropathological changes associated with the clinical diagnosis of AD, suggesting that the timeline of pathological changes responsible for disease progression is still not completely known [20 –22]. In addition, the CSF biomarkers have a limited use in assessing disease progression because of the relative stability in clinical AD [23, 24]. Accordingly, levels of CSF Aβ₄₂ are markedly reduced in preclinical stages but remain relatively stable throughout the rest of the process [25].

When in disagreement with the clinical diagnosis, the profiles of the CSF biomarkers, at the time of the analysis, represent a challenge to understand their role in elucidating the disease etiology. Identifying subjects with normal cognition but at-risk for AD [26], who present evidence of the pathology in CSF (e.g., at least the presence of A+) is a critical focus of preclinical AD studies nowadays. On the other hand, the mismatch between the clinical diagnosis of probable AD and the profile of CSF biomarkers may suggest other degenerative conditions that mimic AD clinical presentation. Noteworthy, the absence of amyloid (A-) would represent the exclusion of AD pathology.

This study aimed to report findings of pathological patterns of CSF biomarkers of AD that were in disagreement with the clinical diagnosis at the time of analysis, in a cohort of patients attending a memory clinic in a university hospital in the city of São Paulo, Brazil. We also classified our sample using the AT(N) conceptual framework, correlating the different groups with the presence of AD pathology in CSF. Breaking down aspects related to mismatches and disagreements between clinical diagnosis and biomarker changes would represent a necessary step towards including biomarkers in clinical routine, since a significant part of patients clinically assessed can show neurobiological changes different from expected.

MATERIALS AND METHODS

Participants and clinical evaluation

The recruitment of participants for this study is part of an ongoing project that includes systematic assessment and longitudinal follow-up of subjects with various degrees of cognitive impairment, from subjective memory complaints to dementia, attending a psychogeriatric service in a tertiary, university-based hospital in São Paulo, Brazil (Institute of Psychiatry, Faculty of Medicine, University of São Paulo). These patients are also treated whenever required. All subjects were recruited from the hospital catchment area through spontaneous demand or referral from other services from 2007 to 2019. The local Ethics Committee approved the study, and it was conducted under the tenets of the Declaration of Helsinki [27], and all subjects submitted to lumbar puncture for CSF collection and analysis signed informed consent authorizing the procedure.

All participants were interviewed and evaluated by a multidisciplinary team of psychiatrists, neurologists, geriatricians, neuropsychologists, and occupational therapists. Clinical history and general and neurological examinations were performed before the cognitive assessment. Trained neuropsychologists performed a comprehensive neuropsychological and functional assessment that included the following tests: Fuld Object Memory Evaluation (FOME) [28, 29], Rivermead Behavioral Memory Test (RBMT) [30, 31], tasks A and B of the Trail Making Test (TMT) [32], Revised Wechsler Adult Intelligence Scale Vocabulary and Block Design subtests [33], and Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) [34].

Cognitive assessment involved the use of the Brazilian version of the Cambridge Examination for Mental Disorders of the Elderly (CAMDEX) interview [35, 36]. The CAMDEX schedule contains the Cambridge Cognitive Test (CAMCOG), a brief cognitive assessment battery that evaluates several neuropsychological domains. For this study, depressive symptoms were ruled out through the 21-item Hamilton Depression Rating Scale [37], with euthymic status considered for scores < 8.

All participants underwent blood tests to detect possible reversible causes of dementia. Therefore, a complete blood count, blood tests for liver enzymes, vitamin B12, folic acid, thyroid functions, kidney functions, liver enzymes, and serology for HIV and syphilis were performed. All participants were also submitted to a brain magnetic resonance (MRI) study.

Inclusion and exclusion criteria

Patients with a history or current neurological and/or psychiatric comorbidities (including major depression), uncompensated systemic diseases, and that had recent introduction or dose adjustment of medications that potentially interfere with cognitive performance were excluded and referred for treatment. We also excluded patients with severe cerebrovascular disease as defined by MRI scans with extensive white matter hyperintensities on T2/FLAIR sequences, such as Fazekas visual rating scale≥2 [38]. Patients with primary psychiatric disorders and those with insufficient personal data were also excluded from the research. All subjects whose data are reported in this study are outpatients at our memory clinic, most of whom are still under regular follow-up, and have been consecutively enrolled to the cohort in the past years. In addition to standard clinical procedures, all participants (patients and controls) are submitted to a comprehensive protocol for determination of CSF and blood-based biomarkers, collection of neuroimaging data and neuropsychological profiling. In this regard, most patients display a good compliance with all procedures, including lumbar puncture (LP) per se. Of course, there is a small, but inevitable, number of patients who have relative contraindications against LP (e.g., use of anticoagulants), in addition to cases where the procedure itself is unsuccessful (e.g., orthopedic conditions affecting the spinal cord rendering collection more laborious, or pain/anxiety reactions during the procedure itself, leading to its interruption).

From an initial sample of 272 individuals attending our outpatient service that had their CSF collected through lumbar puncture during the period of 2007 to 2016, we enrolled 204 subjects for this study. We also included subjects under 65 years of age. Our outpatient service is also dedicated to investigating subject cognitive complaints in younger subjects with a family history of dementia. The main objective is to diagnose early onset dementias. There is also an understanding that neuropathological changes associated with AD are present many years before onset of symptoms.

The Mayo Clinic criteria were used to account for the diagnosis of MCI [39] and its amnestic/non-amnestic subtypes [40]; as well as the DSM-V diagnostic definition of Mild Neurocognitive Disorder [41]. The criteria for MCI diagnosis [39] involve 1) the presence of cognitive complaint (preferably corroborated by an informant); 2) objective cognitive impairment demonstrated by validated neuropsychological testing, defined as a performance below –1.5 standard deviations according to validated neuropsychological tests addressing distinct cognitive domains, corrected for age and educational level; 3) minimal impact on activities of daily living and, therefore, relatively maintained independence. Probable AD cases were diagnosed according to the NIA-AA criteria [42, 43]. Clinical diagnoses were established at consensus meetings that included a multidisciplinary team of physicians (psychiatrists, a geriatrician, and a neurologist), neuropsychologists, physical therapists, speech pathologists, occupational therapists, and gerontologists, taking into account all clinical and laboratorial information.

CSF analysis and biomarkers of AD pathology

All CSF samples were obtained by lumbar puncture. In the L3/L4 or L4/L5 intervertebral space, with a 23-gauge needle and using polypropylene tubes, in the morning period without fasting, CSF samples (12–15 mL) were collected. They were then centrifuged at 3200×g for 10 min at 4°C, split into 0.5 mL aliquots in cryotubes (Sarstedt), and immediately frozen and stored at 28°C until analysis, without being thawed and refrozen. CSF concentrations of the 42 amino acid-long Aβ₄₂, T-tau, and P-tau were determined in duplicates with the INNO-BIA AlzBio3 assay (Fujiorebio), a multiplex microsphere-based xMAP platform that allows the simultaneous analysis of the three biomarkers. After prewetting the filter plate with a wash buffer, a suspension of microsphere carrying the corresponding capturing antibodies (AT120, AT270, and 4D7A3 for T-tau, P-tau, and Aβ₄₂, respectively) was added to the plate. A mixture of biotinylated detection monoclonal antibodies, designed to detect specifically one of the capturing antibodies (HT7 for T-tau and P-tau and 3D6 for Aβ₄₂), and 75 mL of CSF or standards were added to the plate and incubated overnight in the dark. Next, the plate was washed, and a detection conjugate (phycoerythrin-labeled streptavidin) was added and incubated for 1 h at room temperature. The plate was washed and after the addition of a reading solution (phosphate buffer saline) the assay was analyzed on a Luminex 100IS platform (Luminex, Austin, TX, USA). Standard curves were constructed for each biomarker using a sigmoidal curve fitting method, and the mean fluorescence values for the duplicate CSF samples were used to determine the concentration of Aβ₄₂, T-tau, and P-tau [8].

In our laboratory, we use the previously internally qualified cut-off values that discriminate AD patients from older adults with normal cognition (NC). In a previous published study [8], we determined the CSF concentrations of Aβ₄₂, T-tau, and P-tau and the cutoff scores that best discriminated cognitively unimpaired subjects from patients with AD, as well as the values that predicted the conversion from amnestic MCI to dementia in a cohort of older adults. We further performed sensitivity (SE), specificity (SP), and area under the receiver operating characteristic (ROC) curve analyses using CSF biomarker values to discriminate AD from controls. Cutoff values were generated in these analyses for the differential diagnosis between AD and other diagnostic groups (MCI and non-AD cognitive impairment/dementia). The power of these biomarkers and respective cutoff values to predict the risk of progression from MCI to AD in the longitudinal arm of the study was also evaluated. Kaplan-Meier survival curves and time-dependent Cox regressions were carried out to assess the best biomarker combinations and predictors of progression to AD in MCI subjects. The study generated the following cutoff and SE/SP values: Aβ₄₂: 416.0 pg/mL (SE: 83%, SP: 70%); P-tau: 36.1 pg/mL (SE: 83%, SP: 49%), and T-tau: 76.7 pg/mL (SE: 82%, SP: 67%). The Aβ₄₂/P-tau and Aβ₄₂/T-tau ratios were also previously calculated, and the cut-off values are, respectively, 9.53 (SE: 88%, SP: 78%) and 4.13 (SE: 80%, SP: 80%). The combination of the CSF concentration values of the Aβ₄₂ peptide < 416.0 pg/mL and the Aβ₄₂/P-tau ratio < 9.53 was a good predictor of conversion to dementia in 2 years [8]. These parameters also discriminate patients with AD from those with non-AD pathology, as well as patients with MCI who progress to dementia from those who could remain stable during follow-up. As our previous study found similar values of SE and SP for all parameters, in this study we opted to select the Aβ₄₂/P-tau ratio as the main parameter. The Aβ₄₂/P-tau ratio has good SE and SP values to discriminate AD patients from normal controls. This index was conceived to improve the power of diagnosis of CSF biomarkers [44, 45].

APOE status

We had data from APOE status from a part of the subjects (n = 139). We investigated the frequency and distribution of apolipoprotein E alleles (ɛ2, ɛ3, and ɛ4) within the three diagnostic groups to verify their possible correlations with CSF biomarkers concentrations. Genomic DNA was isolated from whole blood from each subject and the APOE genotyping was performed using the TaqMan® 5’-exonuclease allelic discrimination assay30 obtained from Applied Biosystems (Foster City, CA, USA) with primers and probes sets from inventoried assays. This methodology uses two PCR assays to screen for single nucleotide polymorphisms (rs429358, rs7412) within the exon 4 of APOE gene. Results from the individual assays were used to determine the ultimate APOE genotype [46].

Statistical analyses

Sociodemographic characteristics (age, school years, and gender), CAMCOG scores, APOE status, CSF biomarker concentrations at the time of the neuropsychological assessment, and clinical diagnosis were analyzed using one-way analysis of variance (ANOVA) followed by Pearson’s correlation coefficient. Post-hoc Bonferroni test was also applied to determine the statistical significance of the differences between variables. Kolmogorov-Smirnov tests were done to assess the normality of the distribution for the continuous variable, and then we carried out parametric statistical tests for all analyses. Statistical analyses were carried out with the software SPSS (Statistical Package for Social Sciences) v.14 for Windows (Chicago, Ill., USA), and statistical significance was set at α= 0.05 in all analyses.

RESULTS

Participants and sociodemographic characteristics

From 204 subjects, 60 (29.4%) presented with normal cognition (NC), 84 (41.2%) were diagnosed with mild cognitive impairment (MCI), and 60 (29.4%) were diagnosed with dementia due to AD. Thirty subjects were under 65 years. The age of participants ranged from 49 to 90 years. CAMCOG mean scores differed between groups with statistical significance. We had data on APOE status from 139 subjects. Most of them did not carry the ɛ4 allele (n = 99, 71.2%). No differences were found between diagnostic groups as for the presence of the ɛ4 allele (χ² = 0.245; p = 0.884). Table 1 summarizes sociodemographic data, CAMCOG mean scores, and APOE status. Supplementary Table 2 summarizes other additional data.

Table 1

Sociodemographic characteristics and APOE status for all diagnostic groups

	Diagnosis
	NC (n = 60)	MCI (n = 84)	AD (n = 60)	p
Gender W (%)	46 (35.4%)	52 (40.0%)	32 (24.6%)	0.030
Age mean (SD)	71.0 (6.3)	72.6 (6.1)	72.6 (8.1)	0.090
Years of education mean (SD)	11.8 (5.6)	9.7 (5.5)	8.5 (5.0)	0.270
CAMCOG	92.2 (9.1)	84.5 (13.5)	71.9 (17.6)	< 0.001
APOE ɛ4 (n = 139)
ɛ2/ɛ2 or ɛ2/ɛ3 or ɛ3/ɛ3	33 (71.7%)	45 (72.6%)	21 (67.7%)	0.884
ɛ2/ɛ4 or ɛ3/ɛ4 or ɛ4/ɛ4	13 (28.3%)	17 (27.4%)	10 (32.3%)

NC, normal cognition; MCI, mild cognitive impairment; AD, Alzheimer’s disease; W, women; CAMCOG, Cambridge Cognitive Examination; ɛ2/ɛ2 or ɛ2/ɛ3 or ɛ3/ɛ3, no APOE ɛ4 allele present; ɛ2/ɛ4 or ɛ3/ɛ4 or ɛ4/ɛ4, presence of at least one APOE ɛ4 allele. Values are presented as means and standard deviations (SD).

CSF biomarkers concentration for each diagnostic group

We used ANOVA followed by Bonferroni tests to determine the concentrations of biomarkers for each diagnostic group to verify the discriminative power of the CSF Aβ₄₂/P-tau ratio in our sample. The Aβ₄₂/P-tau ratio was able to differentiate all three groups (Table 2 and Fig. 1).

Table 2

Concentrations of cerebrospinal fluid biomarkers for each diagnostic group

	Diagnosis
	NC	MCI	AD	p
Aβ₄₂	488.7 (±22.6)	431.0 (±18.1)	374.9 (±16.9)	< 0.001	AD < NC
P-tau	41.6 (±3.4)	48.8 (±3.1)	69.7 (±15.5)	< 0.001	AD > NC = MCI
T-tau	81.2 (±6.2)	100.1 (±7.2)	148.2 (±13.1)	< 0.001	AD > MCI = NC
Aβ₄₂/P-tau	15.7 (±1.1)	11.9 (±0.8)	8.1 (±0.9)	< 0.001	AD < MCI < NC

Aβ₄₂, amyloid-β peptide; P-tau, ¹⁸¹Thr-phosphorylated-tau; T-tau, total tau; NC, normal cognition; MCI, mild cognitive impairment; AD, Alzheimer’s disease. Biomarker concentrations are given in pg/mL. Values are presented as means and standard deviations.

Fig. 1

AD related biomarkers across all diagnostic groups. Aβ₄₂, amyloid-β peptide; P-tau, ¹⁸¹Thr-phosphorylated-tau; T-tau, total tau; NC, normal cognition; MCI, mild cognitive impairment; AD, Alzheimer’s disease. The letters a, b, and c represent the statistical differences found between groups. p < 0.001 throughout.

There were no significant differences with respect to the presence of CSF Aβ₄₂/P-tau ratio < 9.53 for gender, age groups, years of education, or APOE status. Individuals aged over 70 years did not present a higher frequency of the Aβ₄₂/P-tau ratio < 9.53 in the CSF when compared with those under 70 years old (χ² = 1.049; p = 0.305). The APOE status did not affect the proportion of CSF Aβ₄₂/P-tau ratio < 9.53. Even with aggregated data, the presence of the allele ɛ4 of APOE, in this sample, did not affect the presence of the CSF Aβ₄₂/P-tau ratio < 9.53 (χ² = 3.36; p = 0.067).

Presence of AD biological markers across diagnostic groups

We also determined the number of subjects with the CSF Aβ₄₂/P-tau ratio < 9.53 in each of the diagnostic groups (Fig. 2). Considering the AD continuum, we also determined the presence of altered concentrations of each biomarker independently, amyloid pathology (A+), tau pathology (T+), and neurodegeneration (N+).

Fig. 2

Presence of AD pathology as represented by Aβ₄₂/P-tau < 9.53 across the three diagnostic groups. Aβ₄₂, amyloid- β peptide; P-tau, ¹⁸¹Thr-phosphorylated-tau; T-tau, total tau; NC, normal cognition; MCI, mild cognitive impairment; AD, Alzhei-mer’s disease. p < 0.001 in comparison between groups.

Normal cognition group

In the NC group, 14 subjects (23.3%; p < 0.001) presented with CSF Aβ₄₂/P-tau ratio < 9.53 and 22 (36.7%) were classified as A + (Figs. 2 3). Despite having no symptoms, 38.3%of patients from the NC group (n = 23) were classified as N + , showing signs of neurodegeneration (Fig. 3). We compared age, years of education and CAMCOG scores of NC subjects with signs of pathology (A + or T + or N +) with NC subjects with no sign of any pathology (A- and T- and N-) and found no statistically significant differences between age and years of education (mean age 71.2 and 70.7, respectively, p = 0.77; mean years of education 11.2 and 13, respectively, p = 0.21). Mean total CAMCOG scores were different between these two groups, the former presenting with lower scores than the latter (90.58 and 95.05, respectively, p = 0.049).

Fig. 3

Occurrence of altered concentrations of the three core CSF biomarkers across diagnostic groups according to the AT(N) classification. A + , amyloid-positive cases (low CSF Aβ₄₂), NC < AD (p < 0.001), NC < MCI (p < 0.001), MCI = AD (p > 0.05); T + , (phosphorylated)-tau positive cases (high CSF ¹⁸¹Thr-P-tau), NC < MCI<AD (p < 0.001); N + , neurodegeneration-positive cases (high CSF total tau), NC < MCI < AD (p < 0.001); NC, normal cognition; MCI, mild cognitive impairment; AD, Alzheimer’s disease.

Mild cognitive impairment

In the MCI group, 36 subjects (42.9%) presented with the CSF Aβ₄₂/P-tau ratio < 9.53. The number of individuals with MCI classified as A + was 46 (54.8%). It was significantly differently from the NC group (35%; p = 0.019) and from the probable AD group (73.3%; p = 0.022). There statistically differences between MCI and NC groups in respect to the N classification (p < 0.001).

Twenty-two subjects were classified as amnestic MCI (aMCI) and 17 as non-amnestic MCI (naMCI). In 45 individuals (53.5%) the subtype of MCI was not determined. There were no differences in CSF biomarkers concentrations between aMCI and naMCI (Table 3).

Table 3

Biomarker concentrations in MCI subtypes

	aMCI (n = 22)	naMCI (n = 17)	not defined (n = 45)	F	p
Aβ₄₂	457.9 (197.6)	422.2 (128.9)	422.2 (169.9)	0.281	0.755
T-tau	89.4 (50.8)	101.0 (58.7)	103.9 (73.9)	0.298	0.742
P-tau	41.4 (19.7)	44.5 (23.5)	53.8 (33.4)	1.485	0.232
Aβ₄₂/P-tau	13.3 (8.0)	12.5 (7.5)	11.0 (7.8)	0.615	0.542

Aβ₄₂, amyloid-β peptide; P-tau, ¹⁸¹Thr-phosphorylated-tau; T-tau, total tau; aMCI, amnestic mild cognitive impairment; naMCI, non-amnestic mild cognitive impairment. Biomarker concentrations are given in pg/mL. Values are presented as means and standard deviations.

Alzheimer’s disease group

In 73.3%of the cases (n = 44; p < 0.001), individuals clinically diagnosed with dementia due to AD presented an Aβ₄₂/P-tau ratio < 9.53. The proportion of subjects in the AD diagnosis group classified as A + was 66.7%. It was not significantly superior to MCI subjects and to NC subjects (p < 0.001).

A higher proportion of subjects in the AD diagnosis group were classified as T + and N+, with signs of tau pathology (78.3%) and neurodegeneration (80%), respectively, when compared with the A + classification (66.7%; p < 0.001). The comparison between AD and NC and MCI groups for T + and N + showed significant differences (MCI group (T+): p = 0.008; (N)+: p = 0.004, and NC group (T+): p = 0.012; (N+): p < 0.001). In addition, 20.0%of subjects with a diagnosis of AD did not present with any signs of neurodegeneration (Fig. 3).

AT(N) classification

We applied the AT(N) classification in the three diagnostic groups (NC, MCI, and AD) based on the concentrations of each CSF biomarker and correlated with the presence of AD pathology as defined here by the CSF Aβ₄₂/P-tau ratio < 9.53. The presence of all three biomarkers from the AT (N) classification in each diagnostic group were summarized in a graphic representation. Supplementary Figure 1 shows all possible phenotypes found when assessing CSF biomarkers in a group of subjects with different diagnostic profiles.

DISCUSSION

In this study we highlight that a substantial proportion of subjects clinically diagnosed as with probable AD, MCI, or unimpaired cognition may display discordant values for CSF AD-related biomarkers. These clinical biological mismatches do occur despite a rigorous diagnostic procedure, established at a university-based memory clinic. We further classified our sample using the AT(N) conceptual framework proposed in the 2018 NIA-AA research recommendations [3].

Not surprisingly, in more than half (57.1%) of the individuals with MCI, the presence of the AD pathology, defined here as the CSF Aβ₄₂/P-tau ratio < 9.53, was not observed. This is probably due to the biological heterogeneity of cases of MCI when classified according to clinical criteria [46]. According to the AT(N) system, 54.8%of the subjects (n = 46) in the MCI group were classified as A + which indicates the occurrence of other causes to justify the cognitive changes. On the other hand, in a substantial number of subjects with normal cognition (38.3%), the presence of neurodegenerative mechanisms (N+) was detected, not necessarily the AD pathology. This phenomenon suggests that these individuals would be at risk of progression to subsequent clinical deterioration, possibly non-AD dementia. Additionally, in a small number of subjects (n = 16; 26.7%) diagnosed with AD dementia, the occurrence of AD pathology was not evident. An even higher number of subjects (n = 20; 33.3%) did not show signs of decreased CSF Aβ₄₂. This subgroup possibly has another neurodegenerative pathology mimicking the clinical course of AD dementia. Although other studies from research groups worldwide have presented similar findings regarding CSF and neuroimaging biomarkers of AD before [47 –53], there is a lack of studies in Latin America with this approach. In addition, our study calls for the importance of cautious assessment of biomarkers of AD, indicating that there is a necessary step to take in the validation of these biomarkers for wide clinical use.

As expected, most subjects with cognitive decline present with the AD pathology. However, the subgroup of individuals who have N + and who still maintain normal cognition, could be placed in the preclinical phase of an ongoing non-Alzheimer’s neurodegenerative process. This means that the diagnosis, especially of those in the preclinical or in an early stage of a neurodegenerative disease, remains a challenging procedure for the clinician since, in this context, early interventions can be beneficial for people at risk [54]. An important point is to disclose the protective factors that critically contribute to the postponement of cognitive decline.

In this context, it is worth mentioning a meaningful aspect discussed by McRae-McKee and colleagues [55]. These authors suggested the existence of a “gray zone” within which it is exceedingly difficult to determine the thresholds for dichotomization between the presence of the signature of the AD pathology and the normal biological status. This approach suggests that there is a critical interval that currently requires more data to early separate people at risk of cognitive decline from those at prodromal stages of a clinical syndrome, as well as to distinguish subsequent trajectories of progressive cognitive decline.

Sociodemographics

Age is the main risk factor for the development of AD and is causally related to increasing findings of AD pathology in the brain. However, changes in CSF biomarkers concentrations are known to be present even in cognitively normal older individuals, although generally not related to a neurodegenerative disease [56, 57]. In this sample, no differences were seen between the different age groups, neither for older individuals (> 70 years) when compared with younger ones (< 70 years), for the presence of the CSF Aβ₄₂/P-tau ratio under 9.53, our main parameter of AD pathology in this study. There was a higher proportion of women (65%, p = 0.03) in this sample. However, gender, age, and years of education had no statistical difference between groups (Table 2) neither for the presence of the CSF Aβ₄₂/P-tau ratio < 9.53, a proxy for AD pathology in this study.

The APOE ɛ4 allele, the most important genetic risk factor for late-onset AD, is highly associated with amyloid pathology. With respect to the APOE status in this study, for all diagnostic groups, most participants were non-carriers of the APOE ɛ4 allele, and carriers were evenly distributed across diagnostic groups (Table 2). We found no higher proportion of the CSF Aβ₄₂/P-tau ratio < 9.53 associated with the presence of the allele ɛ4 of APOE (Supplementary Table 2), a result different from expected given that the presence of the allele ɛ4 of APOE increases the risk of development of sporadic AD [58, 59]. These findings deserve further investigation. Although related to the reduction of CSF Aβ₄₂ and the Aβ₄₂ to Aβ₄₀ ratio the APOE ɛ4 allele does not affect the T-tau and P-tau levels. In addition, surprisingly high discordances have been reported for CSF Aβ₄₂, and consequently for the Aβ₄₂ to Aβ₄₀ ratio, between carriers and non-carriers of the APOE ɛ4 allele in different laboratories and is considered an interfering factor for CSF biomarkers interpretation [60].

Normal cognition

For amyloid pathology (A), as represented by low concentration of Aβ₄₂ in CSF, 35%of NC subjects were classified as A+(n = 21) (Fig. 3). This is an interesting finding, since it places these subjects in the AD continuum. A part of them (n = 8) showed only A + without the Aβ₄₂/P-tau ratio < 9.53 meaning that these subjects are at an earlier or premature stage of the pathological process. These data are in conformity with previous studies that found that up to 30%of cognitively normal elderly subjects present with some level of AD pathology, and these numbers are similar to the ones found in studies with PET amyloid imaging [1 , 47–50].

Although not presenting with symptoms, 38.3%of subjects showed signs of neurodegeneration (N+) and were classified as A-T-N+or A-T+N+(Fig. 3). All these subjects had the CSF Aβ₄₂/P-tau ratio > 9.53, confirming the absence of AD pathology. The determination of the mechanisms involved in these findings and the possible outcomes together represent a challenge since neurodegeneration is the aspect that is most closely related to cognitive impairment [1, 48]. A possible explanation to these findings could be a selection bias. The data presented in this study were gathered from a sample attending a specialized group in a tertiary hospital. NC subjects assessed by our professionals frequently have a family history of AD or other dementias, they tend to seek treatment when presenting with memory complaints even without objective cognitive decline, a known risk factor for development of cognitive decline. In addition, although not significant for this sample, the subjects attending our outpatient service tend to have a higher educational level with the impact on cognitive reserve, providing an explanation for the finding of NC individuals with brain changes or pathology related to AD as they tolerate more of these changes than others and maintain little or no cognitive impairment. The selection of populations with different characteristics (e.g., screened at the community, rather than memory clinic) and the use of other diagnostic tools should dismiss this type of selection bias. The patterns of metabolism in functional imaging such as positron emission tomography with [¹⁸F]fluorodeoxyglucose ([¹⁸F]FDG-PET) found in cognitively normal subjects could represent an advantageous tool in discriminating types and location of neuropathological changes [53], therefore increasing the diagnostic power of biomarker assessment. No differences were found between biomarker-positive and biomarker-negative NC in regard to mean age, education, and APOE status. However, global cognitive performance as shown by total CAMCOG scores were slightly different between these two groups, albeit both well above cutoff scores for suspected decline, but nonetheless indicating subtle changes in cognitive status. A further study with other clinical aspects would be of importance to confirm and clarify these findings, as would be the follow up of these subjects with subsequent evaluation of outcomes.

Mild cognitive impairment

Literature shows that a part of subjects diagnosed with MCI will present evidence of AD pathology, defined as prodromal AD. One previous study from our group showed that patients with MCI who convert to AD, as compared to non-converters, had the AD CSF biomarker signature (i.e., low Aβ₄₂, high T-tau, and high P-tau) [46]. In this study we found that 54.8%of the MCI group in our sample presented with signs of AD pathology. This is another interesting finding that goes in conformity with previous studies on the subject. As an example, roughly 60%of MCI patients are amyloid PET positive, and autopsy studies show similar percentages [58]. Another study showed that 55%of subjects diagnosed with MCI had significantly increased PIB retention values compared to controls at baseline, and 48%clinically converted to AD during a 12–36-month follow-up period after their baseline Pittsburgh compound B labeled with carbon-11 ([¹¹C]PIB) PET scan [59]. However, our finding was significantly higher than the proportion of AD pathology found in the aMCI group (37%) in another study with [¹¹C]PIB PET scan [53]. AD pathology is the most frequent pathology found in both amnestic and non-amnestic MCI, and their level of AD pathology is placed between the level in NC and clinical AD, reaffirming these findings. Nonetheless, it would be acceptable to infer that a clinical diagnosis made by a specialized memory clinic would increase the proportion of MCI patients with AD pathology by selection bias, with more subjects with a family history of AD looking for a specialized assessment, and with the ability of the specialized team to single out subjects with a more typical clinical presentation.

Although no differences were found in CSF bio-markers concentrations between subtypes of MCI, the presence of AD pathology in a substantial part of the subjects diagnosed with MCI in our service calls for the importance of the follow up of this population as they present with higher risk of progression to dementia. In addition, the MCI group was the most diverse group in respect to the AT(N) classification. All eight possible combinations were present in this group, showing that the clinical presentation of MCI is a result of a myriad of neuropathological changes (Supplementary Figure 1).

AD dementia

Patients clinically diagnosed with AD can show no pathological changes both in CSF biomarkers and at postmortem examination [61 –63]. Postmortem studies have already demonstrated some heterogeneity on neuropathological changes associated with dementia diagnosis. A study held by Suemoto and colleagues showed the almost one-fourth of cognitively normal individuals met the criteria for a neuropathological diagnosis with AD being the most common diagnosis (57%), and some individuals lack a neuropathological diagnosis to explain the cognitive impairment observed before death [63]. In our study, 26.7%of subjects diagnosed with AD did not present with the CSF Aβ₄₂/P-tau ratio < 9.53 (Fig. 3), despite the presence of symptoms. Our findings call for a question on the accuracy of the clinical diagnosis of AD. Biomarkers should be included in the diagnostic criteria. However, the hallmarks of AD pathology (i.e., amyloid deposition and elevated tau) are also seen in other neurological disorders [1], and it is widely accepted that patients clinically diagnosed with AD but not presenting with the expected biomarker changes may have been given an incorrect clinical diagnosis.

In addition, 20%of subjects diagnosed with AD did not show any signs of neurodegeneration, 6 of them (10%) with no signs of any changes in CSF biomarkers concentrations (Fig. 3). Interestingly, a subset of patients clinically diagnosed as with ‘probable AD’ may not display the expected biomarker signature in the CSF. One study showed that 6.2%of subjects with clinically probable AD did not have any signs of AD pathology [64]. This was the case in one-third of AD patients ranked as amyloid-negative, and one-fifth with no signs of neurodegeneration in the present study. These cases may require pathological examination for an accurate etiological discrimination. These findings could also represent that the clinical manifestations of cognitive decline in some patients with dementias that mimic AD are associated with other mechanisms, and my show the importance of the cognitive reserve in the progression of cognitive decline in the elderly. Other pathologic diagnoses found in previous studies include frontotemporal dementia with ubiquitin-positive inclusions and hippocampal sclerosis, tauopathy, hippocampal sclerosis, limbic-predominant age-related TDP-43 encephalopathy (LATE), limbic Lewy bodies, multisystem atrophy, progressive supranuclear palsy, microscopic infarcts, and dementia lacking distinctive histology [63 –66].

AT(N) classification

There was a higher proportion of subjects with positivity in all three biomarkers (A + T + N +) in the MCI group with AD pathology in CSF biomarkers (prodromal AD), and in the AD group with the expected CSF biomarker changes (Supplementary Figure 1). In the NC group, there was a higher number of subjects without any changes in biomarkers (A-T-N-). The AT(N) classification makes it easier to understand the AD continuum and could represent a significant improvement in the diagnostic procedures when combined with the clinical assessment [3]. One study from our institution used a PET staging system based on the 2018 NIA-AA research framework to compare the proportion of amyloid positivity (A+) and hypometabolism (N+) in cases of mild probable AD, aMCI, and healthy controls, and found interesting insights on the clinical-biological heterogeneity of phenotypes associated with the AD diagnosis [53].

Limitations of the study and future perspectives

It is worth mentioning that in this subgroup of subjects with cognition considered normal, the neuropsychological assessment would not have captured subtle cognitive changes. This fact is not uncommon among individuals with a high level of education, since they tend to perform very well tasks required in the routine assessment, and a more complex approach should be adapted to their cognitive reserve [68]. Thus, some individuals from the NC group perhaps should be moved to the MCI group. Additionally, in 45 individuals the cognitive classification of MCI was not determined. On the other hand, subjects with exceptionally low cognitive reserve would represent a source of bias since the threshold of detection of cases would be artificially different. One study showed that the pathological burden needed to cause dementia in subjects with fewer years of education is significantly lower, even reflecting in the positivity of biomarkers [69].

Another limitation of our study is the lack of thorough determination of MCI subtypes. In addition, there were no differences in CSF biomarkers concentrations between aMCI and naMCI, and this finding calls for further investigation. Finally, a topic for further assessment would be the reason why subjects with NC and evidence of neurodegeneration (n = 23, 38.3%) remain cognitively preserved. Our findings show that 15 of these subjects did not present with signs of decreased CSF Aβ₄₂ or any other sign of AD pathology. The further assessment of the cognitive reserve of these subjects, assessed by years of education, CAMCOG results, and a different Neuropsychological assessment strategy could shed some light on this matter. The combination of imaging and CSF biomarkers with the AT(N) classification would represent another step forward in developing better diagnostic procedures.

CONCLUSION

It is acceptable that the clinical diagnosis of MCI yields a heterogeneous group of patients with mixed pathologies, where the presence of positive AD-related biomarkers may warrant the actual prediction of conversion. Likewise, cognitively unimpaired elders presenting with positive AD-biomarkers are presumably at higher risk of incident cognitive decline in the forthcoming years, and therefore must be regarded as possible cases of preclinical AD. Finally, subjects clinically diagnosed as probable AD showed no signs of the expected CSF pathology. These cases may in fact represent other forms of dementia related to non-AD pathology, although sharing a similar clinical phenotype (i.e., sporadic, late-onset, progressive, amnestic impairment), that can be clinically undistinguishable from AD.

Taking together, data on the association between cognition and the AT(N) classification reflect the need to improve the diagnosis of cognitive decline. This requires especially targeting individuals who have a clinical trajectory compatible with AD dementia, but who do not meet the neurobiological criteria of the pathology continuum. It becomes clear with recent developments that the stages of AD pathology have a dynamic development, rather than an organized or simultaneous one.

Looking into the multifactorial nature of AD, and the overlapping pathology with other forms of dementia, it is important to integrate the core CSF biomarkers with pivotal clinical aspects, leading to an improvement of their diagnostic power. However, our findings reinforce the need to combine different biomarkers (CSF, neuroimaging, plasma) with a careful clinical follow-up to guarantee a precise and timely diagnosis.

Footnotes

ACKNOWLEDGMENTS

This work was supported by: Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Grants No. 429079/2018-4 and 301629/2018-8; Na-tional Institute for Biomarker Research in Neuro-psychiatry (InBion), Grants Nos. 14/50873-3, 2016/01302-9 (FAPESP) and 465412/2014-9 (CNPq); Associação Beneficente Alzira Denise Hertzog da Silva (ABADHS).

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Jack

, Albert

, Knopman

, McKhann

, Sperling

, Carrillo

, Thies

, Phelps

(2011) Introduction to the recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 257–262.

Jack

, Bennett

, Blennow

, Carrillo

, Feldman

, Frisoni

, Hampel

, Jagust

, Johnson

, Knopman

, Petersen

, Scheltens

, Sperling

, Dubois

(2016) A/T/N: An unbiased descriptive classification scheme for Alzheimer disease biomarkers. Neurology 87, 539–547.

Jack

, Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

Contributors (2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562.

Hardy

, Higgins

(1992) Alzheimer’s disease: The amyloid cascade hypothesis. Science 256, 184–185.

Querfurth

, LaFerla

(2010) Alzheimer’s disease. N Engl J Med 362, 329–344.

Wallin

, Blennow

, Zetterberg

, Londos

, Minthon

, Hansson

(2010) CSF biomarkers predict a more malignant outcome in Alzheimer disease. Neurology 74, 1531–1537.

Cummings

(2019) The role of biomarkers in Alzheimer’s disease drug development. Adv Exp Med Biol 1118, 29–61.

Forlenza

, Radanovic

, Talib

, Aprahamian

, Diniz

, Zetterberg

, Gattaz

(2015) Cerebrospinal fluid biomarkers in Alzheimer’s disease: Diagnostic accuracy and prediction of dementia. Alzheimers Dement (Amst) 1, 455–463.

Hansson

, Zetterberg

, Buchhave

, Londos

, Blennow

, Minthon

(2006) Association between CSF biomarkers and incipient Alzheimer’s disease in patients with mild cognitive impairment: A follow-up study. Lancet Neurol 5, 228–234.

10.

Dhiman

, Blennow

, Zetterberg

, Martins

, Gupta

(2019) Cerebrospinal fluid biomarkers for understanding multiple aspects of Alzheimer’s disease pathogenesis. Cel Mol Life Sci 76, 1833–1863.

11.

Duits

, Martinez-Lage

, Paquet

, Engelborghs

, Lleó

, Hausner

, Molinuevo

, Stomrud

, Farotti

, Ramakers

IHGB

, Tsolaki

, Skarsgård

, Åstrand

, Wallin

, Vyhnalek

, Holmber-Clausen

, Forlenza

, Ghezzi

, Ingelsson

, Hoff

, Roks

, de Mendonça

, Papma

, Izagirre

, Taga

, Struyfs

, Alcolea

, Frölich

, Balasa

, Minthon

, Twisk

JWR

, Persson

, Zetterberg

, van der Flier

, Teunissen

, Scheltens

, Blennow

(2016) Performance and complications of lumbar puncture in memory clinics: Results of the multicenter lumbar puncture feasibility study. Alzheimers Dement 12, 154–163.

12.

Costerus

, Brouwer

, van de Beek

(2018) Technological advances and changing indications for lumbar puncture in neurological disorders. Lancet Neurol 17, 268–278.

13.

Dubois

, Feldman

, Jacova

, Hampel

, Molinuevo

, Blennow

, DeKosky

, Gauthier

, Selkoe

, Bateman

, Cappa

, Crutch

, Engelborghs

, Frisoni

, Fox

, Galasko

, Habert

, Jicha

, Nordberg

, Pasquier

, Rabinovici

, Robert

, Rowe

, Salloway

, Sarazin

, Epelbaum

, de Souza

, Vellas

, Visser

, Schneider

, Stern

, Scheltens

, Cummings

(2014) Advancing research diagnostic criteria for Alzheimer’s disease: The IWG-2 criteria. Lancet Neurol 13, 614–629.

14.

Bateman

, Xiong

, Benzinger

, Fagan

, Goate

, Fox

, Marcus

, Cairns

, Xie

, Blazey

, Holtzman

, Santacruz

, Buckles

, Oliver

, Moulder

, Aisen

, Ghetti

, Klunk

, McDade

, Martins

, Masters

, Mayeux

, Ringman

, Rossor

, Schofield

, Sperling

, Salloway

, Morris

; Dominantly Inherited Alzheimer Network (2012) Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N Engl J Med 367, 795–804.

15.

Blennow

, Wallin

, Chong

(1995) Cerebrospinal fluid ‘neuronal thread protein’ comes from serum by passage over the blood-brain barrier. Neurodegeneration 4, 187–193.

16.

Motter

, Vigo-Pelfrey

, Kholodenko

, Barbour

, Johnson-Wood

, Galasko

, Chang

, Miller

, Clark

, Green

(1995) Reduction of beta-amyloid peptide42 in the cerebrospinal fluid of patients with Alzheimer’s disease. Ann Neurol 38, 643–648.

17.

Sperling

, Aisen

, Beckett

, Bennett

, Craft

, Fagan

, Iwatsubo

, Jack

Jr , Kaye

, Montine

, Park

, Reiman

, Rowe

, Siemers

, Stern

, Yaffe

, Carrillo

, Thies

, Morrison-Bogorad

, Wagster

, Phelps

(2011) Toward defining the preclinical stages of Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 280–292.

18.

Shaw

, Arias

, Blennow

, Galasko

, Molinuevo

, Salloway

, Schindler

, Carrillo

, Hendrix

, Ross

, Illes

, Ramus

, Fifer

(2018) Appropriate use criteria for lumbar puncture and cerebrospinal fluid testing in the diagnosis of Alzheimer’s disease. Alzheimers Dement 14, 1505–1521.

19.

Masters

, Bateman

, Blennow

, Rowe

, Sperling

, Cummings

(2015) Alzheimer’s disease. Nat Rev Dis Primers 1, 15056.

20.

Pais

, Martinez

, Ribeiro

, Loureiro

, Fernandez

, Valiengo

, Canineu

, Stella

, Talib

, Radanovic

, Forlenza

(2020) Early diagnosis and treatment of Alzheimer’s disease: New definitions and challenges. Braz J Psychiatry 42, 431–441.

21.

Drachman

(2014) The amyloid hypothesis, time to move on: Amyloid is the downstream result, not cause, of Alzheimer’s disease. Alzheimers Dement 10, 372–380.

22.

Knopman

, Jack

, Wiste

, Weigand

, Vemuri

, Lowe

, Kantarci

, Gunter

, Senjem

, Mielke

, Roberts

, Boeve

, Petersen

(2013) Brain injury biomarkers are not dependent on β-amyloid in normal elderly. Ann Neurol 73, 472–480.

23.

Perrin

, Fagan

, Holtzman

(2009) Multimodal techniques for diagnosis and prognosis of Alzheimer’s disease. Nature 461, 916–922.

24.

Zhou

, Teramukai

, Yoshimura

, Fukushima

(2009) Validity of cerebrospinal fluid biomarkers as endpoints in early-phase clinical trials for Alzheimer’s disease. J Alzheimers Dis 18, 89–102.

25.

Jack

, Lowe

, Weigand

, Wiste

, Senjem

, Knopman

, Shiung

, Gunter

, Boeve

, Kemp

, Weiner

, Petersen

, Alzheimer’s Disease Neuroimaging Initiative (2009) Serial PIB and MRI in normal, mild cognitive impairment and Alzheimer’s disease: Implications for sequence of pathological events in Alzheimer’s disease. Brain 132, 1355–1365.

26.

Dubois

(2018) The emergence of a new conceptual framework for Alzheimer’s disease. J Alzheimers Dis 62, 1059–1066.

27.

World Medical Association (2013) World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects. JAMA 310, 2191–2194.

28.

Avila

, Lopes

, Nakano

, Bottino

(2016) Normative data of Fuld Object Memory Evaluation test for Brazilian elderly population. Arq Neuropsiquiatr 74, 138–144.

29.

Fuld

(1980) Guaranteed stimulus-processing in the evaluation of memory and learning. Cortex 16, 255–271.

30.

Oliveira

, Schmidt

(1999) Teste Comportamental de Memória de Rivermead: Manual da Versão Traduzida e Adaptada para o Brasil [Rivermead Behavioural Memory Test: Manual for the Version Translated and Adapted for Brazil]. Cognição, Rio de Janeiro.

31.

Wilson

, Cockburn

, Baddeley

, Hiorns

(1989) The development and validation of a test battery for detecting and monitoring everyday memory problems. J Clin Exp Neuropsychol 11, 855–870.

32.

Army Individual Test Battery (1944) Manual of directions and scoring. War Department, Adjunct General’s Office, Washington, DC.

33.

Wechsler

(1981) The psychometric tradition: Developing the Wechsler Adult Intelligence Scale. Cont Educ Psychol 6, 82–85.

34.

Jorm

, Jacomb

(1989) The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): Socio-demographic correlates, reliability, validity, and some norms. Psychol Med 19, 1015–1022.

35.

Sampaio

, Lourenço

(2009) CAMDEX-R and dementia diagnosis in Brazil: Translation and transcultural adaptation of the informant interview. Acta Med Port 22, 571–578.

36.

Roth

, Tym

, Mountjoy

, Huppert

, Hendrie

, Verma

, Goddard

(1986) CAMDEX: A standardized instrument for the diagnosis of mental disorder in the elderly with special reference to the early detection of dementia. Br J Psychiatry 149, 698–709.

37.

Hamilton

(1960) A rating scale for depression. J Neurol Neurosurg Psychiatry 23, 56–62.

38.

Fazekas

, Chawluk

, Alavi

, Hurtig

, Zimmerman

(1987) MR signal abnormalities at 1.5 T in Alzheimer’s dementia and normal aging. AJR Am J Roentgenol 149, 351–356.

39.

Petersen

, Smith

, Waring

, Ivnik

, Tangalos

, Kokmen

(1999) Mild cognitive impairment: Clinical characterization and outcome. Arch Neurol 56, 303–308.

40.

Petersen

(2004) Mild cognitive impairment as a diagnostic entity. J Intern Med 256, 183–194.

41.

American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders - 5th ed (DSM-5). American Psychiatric Association, Washington, DC.

42.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

43.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

, Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

44.

Molinuevo

, Blennow

, Dubois

, Engelborghs

, Lewczuk

, Perret-Liaudet

, Teunissen

, Parnetti

(2014) The clinical use of cerebrospinal fluid biomarker testing for Alzheimer’s disease diagnosis: A consensus paper from the Alzheimer’s Biomarkers Standardization Initiative. Alzheimers Dement 10, 808–817.

45.

Sierra-Rio

, Balasa

, Olives

, Antonell

, Iranzo

, Castellví

, Bosch

, Grau-Rivera

, Fernandez-Villullas

, Rami

, Lladó

, Sánchez-Valle

, Molinuevo

(2016) Cerebrospinal fluid biomarkers predict clinical evolution in patients with subjective cognitive decline and mild cognitive impairment. Neurodegener Dis 16, 69–76.

46.

Forlenza

, Diniz

, Talib

, Radanovic

, Yassuda

, Ojopi

, Gattaz

(2010) Clinical and biological predictors of Alzheimer’s disease in patients with amnestic mild cognitive impairment. Braz J Psychiatry 32, 216–222.

47.

Shaw

, Vanderstichele

, Knapik-Czajka

, Clark

, Aisen

, Petersen

, Blennow

, Soares

, Simon

, Lewczuk

, Dean

, Siemers

, Potter

, Lee

, Trojanowski

, Alzheimer’s Disease Neuroimaging Initiative (2009) Cerebrospinal fluid biomarker signature in Alzheimer’s disease neuroimaging initiative subjects. Ann Neurol 65, 403–413.

48.

Savva

, Wharton

, Ince

, Forster

, Matthews

, Brayne

, Medical Research Council Cognitive Function and Ageing Study (2009) Age, neuropathology, and dementia. N Engl J Med 360, 2302–2309.

49.

Mintun

, Larossa

, Sheline

, Dence

, Lee

, Mach

, Klunk

, Mathis

, DeKosky

, Morris

(2006) [¹¹C]PIB in a nondemented population: Potential antecedent marker of Alzheimer disease. Neurology 67, 446–452.

50.

Knopman

, Parisi

, Salviati

, Floriach-Robert

, Boeve

, Ivnik

, Smith

, Dickson

, Johnson

, Petersen

, McDonald

, Braak

, Petersen

(2003) Neuropathology of cognitively normal elderly. J Neuropathol Exp Neurol 62, 1087–1095.

51.

Mattsson

, Insel

, Donohue

, Landau

, Jagust

, Shaw

, Trojanowski

, Zetterberg

, Blennow

, Weiner

, Alzheimer’s Disease Neuroimaging Initiative (2015) Independent information from cerebrospinal fluid amyloid-β and florbetapir imaging in Alzheimer’s disease. Brain 138, 772–783.

52.

Sala

, Nordberg

, Rodriguez-Vieitez

, Alzheimer’s Disease Neuroimaging Initiative (2020) Longitudinal pathways of cerebrospinal fluid and positron emission tomography biomarkers of amyloid-β positivity. Mol Psychiatry, doi: 10.1038/s41380-020-00950-w

53.

Coutinho

, Busatto

, de Gobbi Porto

, de Paula Faria

, Ono

, Garcez

, Squarzoni

, de Souza Duran

, de Oliveira

, Tres

, Brucki

, Forlenza

, Nitrini

, Buchpiguel

(2020) Brain PET amyloid and neurodegeneration biomarkers in the context of the 2018 NIA-AA research framework: An individual approach exploring clinical-biomarker mismatches and sociodemographic parameters. Eur J Nucl Med Mol Imaging 47, 2666–2680.

54.

Boccardi

, Gallo

, Yasui

, Vineis

, Padovani

, Mosimann

, Giannakopoulos

, Gold

, Dubois

, Jack

, Winblad

, Frisoni

, Albanese

, Geneva Task Force for the Roadmap of Alzheimer’s Biomarkers (2017) The biomarker-based diagnosis of Alzheimer’s disease. 2-lessons from oncology. Neurobiol Aging 52, 141–152.

55.

McRae-McKee

, Udeh-Momoh

, Price

, Bajaj

, de Jager

, Scott

, Hadjichrysanthou

, McNaughton

, Bracoud

, Ahmadi-Abhari

, de Wolf

, Anderson

, Middleton

, Alzheimer’s Disease Neuroimaging Initiative (2019) Perspective: Clinical relevance of the dichotomous classification of Alzheimer’s disease biomarkers: Should there be a “gray zone”? Alzheimers Dement 15, 1348–1356.

56.

Paternicò

, Galluzzi

, Drago

, Bocchio-Chiavetto

, Zanardini

, Pedrini

, Baronio

, Amicucci

, Frisoni

(2012) Cerebrospinal fluid markers for Alzheimer’s disease in a cognitively healthy cohort of young and old adults. Alzheimers Dement 8, 520–527.

57.

Lobo

, Launer

, Fratiglioni

, Andersen

, Di Carlo

, Breteler

, Copeland

, Dartigues

, Jagger

, Martinez-Lage

, Soininen

, Hofman

(2000) Prevalence of dementia and major subtypes in Europe: A collaborative study of population-based cohorts. Neurologic Diseases in the Elderly Research Group. Neurology 54, S4–S9.

58.

Rohn

(2013) Proteolytic cleavage of apolipoprotein E4 as the keystone for the heightened risk associated with Alzheimer’s disease. Int J Mol Sci 14, 14908–14922.

59.

Eisenstein

(2011) Genetics: Finding risk factors. Nature 475, S20–S22.

60.

Vogelgsang

, Vukovich

, Wedekind

, Wiltfang

(2019) Higher level of mismatch in APOEɛ4 carriers for amyloid-beta peptide Alzheimer’s disease biomarkers in cerebrospinal fluid. ASN Neuro 11, 1759091419845524.

61.

Jack

, Knopman

, Jagust

, Petersen

, Weiner

, Aisen

, Shaw

, Vemuri

, Wiste

, Weigand

, Lesnick

, Pankratz

, Donohue

, Trojanowski

(2013) Tracking pathophysiological processes in Alzheimer’s disease: An updated hypothetical model of dynamic biomarkers. Lancet Neurol 12, 207–216.

62.

Okello

, Koivunen

, Edison

, Archer

, Turkheimer

, Någren

, Bullock

, Walker

, Kennedy

, Fox

, Rossor

, Rinne

, Brooks

(2009) Conversion of amyloid positive and negative MCI to AD over 3 years: An 11C-PIB PET study. Neurology 73, 754–760.

63.

Suemoto

, Ferretti-Rebustini

, Rodriguez

, Leite

, Soterio

, Brucki

, Spera

, Cippiciani

, Farfel

, Chiavegatto Filho

, Naslavsky

, Zatz

, Pasqualucci

, Jacob-Filho

, Nitrini

, Grinberg

(2017) Neuropathological diagnoses and clinical correlates in older adults in Brazil: A cross-sectional study. PLoS Med 14, e1002267.

64.

Schneider

, Arvanitakis

, Leurgans

, Bennett

(2009) The neuropathology of probable Alzheimer disease and mild cognitive impairment. Ann Neurol 66, 200–208.

65.

Nelson

, Dickson

, Trojanowski

, Jack

, Boyle

, Arfanakis

, Rademakers

, Alafuzoff

, Attems

, Brayne

, Coyle-Gilchrist

ITS

, Chui

, Fardo

, Flanagan

, Halliday

, Hokkanen

SRK

, Hunter

, Jicha

, Katsumata

, Kawas

, Keene

, Kovacs

, Kukull

, Levey

, Makkinejad

, Montine

, Murayama

, Murray

, Nag

, Rissman

, Seeley

, Sperling

, White III

, Yu

, Schneider

(2019) Limbic-predominant age-related TDP-43 encephalopathy (LATE): Consensus working group report. Brain 142, 1503–1527.

66.

Schott

, Warren

(2012) Alzheimer’s disease: Mimics and chameleons. Pract Neurol 12, 358–366.

67.

Ala

, Beh

, Frey

2nd (2000) Pure hippocampal sclerosis: A rare cause of dementia mimicking Alzheimer’s disease. Neurology 54, 843–848.

68.

Barulli

, Stern

(2013) Efficiency, capacity, compensation, maintenance, plasticity: Emerging concepts in cognitive reserve. Trends Cogn Sci 17, 502–509.

69.

Busatto

, de Gobbi Porto

, Faria

, Squarzoni

, Coutinho

, Garcez

, Rosa

PGP

, da Costa

, Carvalho

, Torralbo

, de Almeida Hernandes

, Ono

, Brucki

SMD

, Nitrini

, Buchpiguel

, Souza Duran

, Forlenza

(2020) In vivo imaging evidence of poor cognitive resilience to Alzheimer’s disease pathology in subjects with very low cognitive reserve from a low-middle income environment. Alzheimers Dement (Amst) 12, e12122.