Cognitive Performance and Cerebrospinal Fluid Markers in Preclinical Alzheimer’s Disease: Results from the Gothenburg H70 Birth Cohort Studies

Abstract

Background:

We have previously shown that older adults with preclinical Alzheimer’s disease (AD) pathology in cerebrospinal fluid (CSF) had slightly worse performance in Mini-Mental State Examination (MMSE) than participants without preclinical AD pathology.

Objective:

We therefore aimed to compare performance on neurocognitive tests in a population-based sample of 70-year-olds with and without CSF AD pathology.

Methods:

The sample was derived from the population-based Gothenburg H70 Birth Cohort Studies in Sweden. Participants (n = 316, 70 years old) underwent comprehensive cognitive examinations, and CSF Aβ-₄₂, Aβ-₄₀, T-tau, and P-tau concentrations were measured. Participants were classified according to the ATN system, and according to their Clinical Dementia Rating (CDR) score. Cognitive performance was examined in the CSF amyloid, tau, and neurodegeneration (ATN) categories.

Results:

Among participants with CDR 0 (n = 259), those with amyloid (A+) and/or tau pathology (T+, N+) showed similar performance on most cognitive tests compared to participants with A-T-N-. Participants with A-T-N+ performed worse in memory (Supra span (p = 0.003), object Delayed (p = 0.042) and Immediate recall (p = 0.033)). Among participants with CDR 0.5 (n = 57), those with amyloid pathology (A+) scored worse in category fluency (p = 0.003).

Conclusion:

Cognitively normal participants with amyloid and/or tau pathology performed similarly to those without any biomarker evidence of preclinical AD in most cognitive domains, with the exception of slightly poorer memory performance in A-T-N+. Our study suggests that preclinical AD biomarkers are altered before cognitive decline.

Keywords

Biomarkers in early Alzheimer’s disease cerebrospinal fluid cognition population-based

INTRODUCTION

Alzheimer’s disease (AD) pathology starts about 10–20 years before the onset of cognitive symptoms. According to the “amyloid cascade hypothesis” the starting event is the deposition of aggregated amyloid-β (Aβ) into plaques [1]. These Aβ plaques then start a cascade that results in neuronal and synaptic degeneration and dementia. Many longitudinal population-based studies, investigating different cognitive domains, have shown that the trajectory of cognitive decline precedes onset of a clinical AD diagnosis by up to a decade or more [2 –5]. However, cognitive symptoms appear late during the AD process, and are preceded by changes in biomarkers decades earlier. Preclinical AD pathology characterized by low Aβ₄₂, high P-tau, and high T-tau levels is very common in cognitively unimpaired older adults [6].

Many studies applying the different criteria of pre-clinical AD (using CSF and amyloid PET) have shown that high amyloid burden as measured on amyloid PET and in CSF [7 –11] and/or tau pathology in CSF [12 –14] and/or neurodegenerations markers [15 –18] are related to cognitive decline, mild cognitive impairment (MCI) or dementia. However, many of these studies are done in clinical samples [10 , 18]. Population-based studies [7 , 14–17] have yielded disparate results, some showing a relationship between amyloid positivity on PET and subtle cognitive changes in otherwise cognitively unimpaired individuals [7 , 19], while others demonstrate no such relationship [20, 21]. Population-based studies with CSF data are rare, but there are reports of increased risk of developing cognitive decline in cognitively normal participants with positive AD biomarkers compared to those without [9 , 17]. Using CSF, we have previously shown that cognitively healthy participants with amyloid and tau pathology had slightly worse performance on the Mini-Mental State Examination (MMSE) than participants without underlying preclinical AD pathology [6]. It is, however, less clear if this group shows subtle cognitive decline in other domains. Therefore, we sought to investigate if cognitive performance differs between individuals with and without underlying AD pathology in a population-based sample of 70-year-olds with Clinical Dementia Rating (CDR) 0. Cognitive performance was also compared by AD pathology status in a group of 70-year-olds with established cognitive decline, operationalized as CDR 0.5.

METHODS

The sample was systematically obtained from the Swedish population registry, as previously described [22], and derived from the 2014–2016 examinations of the Gothenburg H70 Gothenburg Birth Cohort Studies in Sweden (the H70 studies), and included people living in private households and in residential care.

Every 70-year-old living in Gothenburg, Sweden, born 1944 on predetermined birth dates was eligible to participate in the examinations during 2014 to 2016. 1,203 participants took part (response rate 72.2%), and 430 (35.8%) consented to a lumbar puncture (LP). Contraindications such as immune modulated therapy, anticoagulant therapy, and cancer therapy were present in 108, leaving 322 (26.8%) who underwent LP. A CDR score was assigned to every participant and those with dementia (n = 6) were excluded. 259 had a CDR score of 0 and 57 had a CDR score of 0.5 [6].

Standard protocol approvals, registrations, and patient consents

All participants and/or their close relatives provided written informed consent. The study was approved by the Regional Ethical Review Board in Gothenburg [6].

Assessments

Participants were examined at the Neuropsychia-tric memory clinic at Sahlgrenska University Hospital in Gothenburg or in their homes. The neuropsychiatric examinations were performed by experienced psychiatric research nurses, and comprised ratings of psychiatric symptoms and signs, tests of mental functioning, including assessments of episodic memory (short-term, long-term), aphasia, apraxia, agnosia, executive functioning, and personality changes [6 , 22–25]. Key informant interviews were performed by a psychologist and research nurses as described previously [6, 22]. Additional cognitive assessments were performed by a psychologist, research nurse, psychiatrist, or medical doctor using a neuropsychological test battery, including the following cognitive tests: 1) memory (immediate and Delayed recall (12 object memory tests), Word memory (10 word memory list), Supra span (10 word memory list (BUSII)), Thurstone’s picture memory test; 2) language (semantic fluency animals, phonetic fluency controlled oral word association FAS); 3) executive function (Figure logic (SRB2), Digit span backwards); 4) visuospatial (Block design, (SRB3)); and 5) mental speed (Psif) [22 , 25–27]. Global cognitive status was assessed by MMSE and the assignment of a CDR score.

Dementia was diagnosed according to the DSM-III-R criteria as previously used in the Gothenburg Birth Cohort studies for over 30 years. Education (defined as years of education) and history of stroke and TIA information was acquired from self-reports and close informants. The participants also underwent comprehensive somatic examinations [6, 22].

APOE ɛ4 genotyping

The SNPs rs7412 and rs429358 in APOE (gene map locus 19q13.2) were genotyped, using KASPar® PCR SNP genotyping system (LGC Genomics, Hoddesdon, Herts, UK). Genotype-data for these two SNPs were used to define ɛ2, ɛ3, and ɛ4 alleles [6]. Data on APOE genotype was lacking for 5 individuals.

CSF sampling and biomarker analyses

As previously described [6], LPs to collect CSF samples were performed in the morning, in the L3/L4 or L4/L5 inter-space. The first 10 mL of CSF were collected in a polypropylene tube and immediately transported to the laboratory and centrifuged at 1,800 g at 20°C for 10 min. The supernatant was gently mixed to avoid possible gradient effects, aliquoted in polypropylene tubes and stored at –70°C [6].

CSF total tau and tau phosphorylated at threonine 181 (P-tau) concentrations were measured using a sandwich enzyme-linked immunosorbent assay (ELISA) (INNOTEST^® htau Ag and PHOSPHO_TAU (181P), Fujirebio (formerly Innogenetics) [28, 29]. CSF Aβ-₄₂ concentration was measured using a sandwich ELISA (INNOTEST^® β-amy-loid_1–42), specifically constructed to measure Aβ starting at amino acid 1 and ending at amino acid 42 [6, 30]. All assays are included in the panel of routine clinical analyses at the Mölndal Clinical Neurochemistry laboratory [6]. Analytical runs had to pass quality control criteria for the calibrators and internal quality control samples had to be approved. The following CSF cut-offs were used to define AD biomarker pathology [31]: CSF Aβ₄₂ levels ≤530 pg/mL (the A criterion), P-tau levels of ≥80 pg/mL (the T criterion), and CSF T-tau levels ≥350 pg/mL (the N criterion) [6, 32].

Statistical analyses

Differences in means of cognitive test scores and the sample characteristics age and education were tested with unpaired Student’s T-tests. Satterthwaite’s T-test was employed when variances were not equal according to Levene’s test. Differences in proportions for the variables sex, stroke, depression, and APOE ɛ4 carriership were tested with Fischer’s exact test. We also investigated the correlation between P-tau and T-tau using Spearman correlation. Further, correction for multiple testing was performed at the 5% level, based on 12 hypotheses (cognitive test variables). The Benjamini-Hochberg method was used since the variables were positively correlated.

Statistical power was determined using STATA. For a two-tailed t-test of differences between the samples, we used an alpha value of 0.05 and a power value of 80% to determine the required sample size using standard deviation and mean difference from the observed sample.

A two-tailed level of significance was used for all analyses (p < 0.05). Statistical analyses were carried out using SPSS for Windows (v. 25, SPSS, Chicago, IL) and STATA (v.14).

RESULTS

Characteristics of the 259 participants with CDR 0 and the 57 participants with CDR 0.5 who participated in the LP are shown in Table 1. In the CDR 0 group, mean age was 70.55 (SD 0.26) years, 129 (49.8%) were female, and 87 (34.1%) had the APOE ɛ4 allele (Table 1). In the CDR 0.5 group, mean age was 70.57 (0.23), 23 (40.4%) were female, and 27 (48.2%) were APOE ɛ4 carriers. The participants with CDR 0.5 had lower education (11.33 years versus 13.08 years, p = 0.002, Cohen’s D = 0.462) and had more often had stroke (12.3% versus 3.5%, p = 0.013, Cohen’s D = 0.749) than participants with CDR 0. They also had lower MMSE score (27.61 versus 29.25 (p < 0.001, Cohen’s D = 1.306) (Table 1). Among participants with CDR 0, a number of factors (e.g., education) were similar in those with and without CSF pathology, but the APOE ɛ4 allele was, as previously published, more prevalent among those with pathological CSF AD biomarkers (Supplementary Tables 1 and 2) [6].

Table 1

Characteristics of the study participants of the Gothenburg Birth Cohort Studies with Clinical Dementia Rating (CDR) 0 or Clinical Dementia Rating 0.5 and cerebrospinal fluid data on preclinical Alzheimer’s disease

Characteristics	CDR0	CDR0.5	p
	(n = 259)	(n = 57)
Age, mean (SD), y	70.55 (0.26)	70.57 (0.23)	0.726
MMSE score, mean (SD)	29.25 (0.95)	27.61 (1.50)	<0.001
Education, mean (SD), y	13.08 (3.86)	11.33 (3.72)	0.002
Women, n (%)	129 (49.8)	23 (40.4)	0.241
Stroke, n (%)	9 (3.5)	7 (12.3)	0.013^*
Any depression n (%)	21 (8.1)	8 (14.3)	0.198
APOEɛ4-positive n (%)	87 (34.1)	27 (48.2)	0.065

^*Not significant at the 5% level after Benjamini-Hochberg correction.

Correlation between T-tau and P-tau

There was a correlation between T-tau and P-tau (Spearman’s rho = 0.932, p < 0.0001).

We then compared cognitive test performance between participants with CDR 0 and CDR 0.5. Better performance was observed in all cognitive domains for the CDR 0 group compared to the CDR 0.5 group (Supplementary Table 3).

Pathology and cognitive tests in 70-year-olds with CDR 0

Among participants with CDR 0, 60 (23.2%) had amyloid pathology, 18 (6.9%) had P-tau pathology, and 87 (33.6%) had T-tau pathology, as previously published [6].

Test scores of general cognitive function (MMSE), memory, language, executive function, visuospatial function, and mental speed were similar among participants with and without AD biomarker pathology for all three analyzed biomarkers (Table 2).

Table 2

Cognitive test performance and cerebrospinal fluid biomarker status in 70-year-olds from the Gothenburg Birth Cohort Studies (n = 259) with Clinical Dementia Rating 0

	Amyloid pathology	Amyloid pathology	No amyloid pathology	p¹	T-tau pathology	T-tau pathology	No T-tau pathology	p²	P-tau pathology	P-tau pathology	No P-tau pathology	p³
	N = 60	mean score (SD)	mean score (SD)	N = 87	mean score (SD)	mean score (SD)	N = 18	mean score(SD)	mean score (SD)
MMSE	60	29.20 (0.90)	29.27 (0.96)	0.629	86	29.16 (1.07)	29.30 (0.88)	0.286	18	28.94 (1.06)	29.28 (0.94)	0.154
Memory
Immediate recall	60	8.13 (1.57)	8.23 (1.67)	0.698	86	7.98 (1.65)	8.32 (1.63)	0.114	18	8.33 (1.57)	8.20(1.65)	0.732
Delayed recall	60	7.53 (1.82)	7.75 (1.70)	0.401	86	7.43 (1.70)	7.83 (1.73)	0.078	18	7.94 (1.77)	7.68 (1.72)	0.530
Word memory	59	5.37 (1.83)	5.66 (1.79)	0.281	84	5.54 (1.65)	5.62 (1.88)	0.715	17	5.53 (1.77)	5.60 (1.80)	0.878
Supra span (BUSII)	55	7.71 (1.54)	7.64 (1.49)	0.771	81	7.43 (1.52)	7.77 (1.48)	0.099	17	7.94 (1.20)	7.64 (1.52)	0.420
Thurstone’s picture memory test	56	21.86 (4.96)	23.05 (3.43)	0.098	82	22.60 (4.16)	22.86 (3.72)	0.622	18	22.89 (4.64)	22.76 (3.81)	0.892
Language
Word fluency	60	25.45 (6.51)	24.93 (6.47)	0.586	86	24.29 (6.16)	25.43 (6.60)	0.183	18	27.89 (5.38)	24.84 (6.50)	0.053
FAS	57	42.30 (12.95)	41.81 (14.02)	0.817	81	41.20 (14.23)	42.28(13.55)	0.561	16	41.75 (13.36)	41.94 (13.81)	0.958
Executive function
SRB2	59	20.54 (3.95)	20.52 (4.14)	0.972	85	20.47 (3.90)	20.55 (4.20)	0.879	18	20.83 (3.60)	20.50 (4.13)	0.741
Digit span backwards	58	4.62 (1.06)	4.56 (1.15)	0.715	83	4.55 (1.10)	4.58 (1.14)	0.852	17	4.71 (0.92)	4.56 (1.14)	0.616
Visuospatial
SRB3	57	22.56 (6.78)	21.70 (6.79)	0.402	83	21.10 (7.32)	22.30 (6.48)	0.187	17	22.59 (8.06)	21.85 (6.69)	0.665
Mental speed
Psif	58	29.12 (8.56)	29.90 (7.40)	0.496	85	29.46 (8.13)	29.86 (7.44)	0.696	18	29.94 (5.84)	29.71 (7.80)	0.900

p¹ = difference in mean score for amyloid pathology compared to non-amyloid pathology with T-test. p² = difference in mean score for T-tau pathology compared to non-T-tau pathology with T-test. p³ = difference in mean score for P-tau pathology compared to non-P-tau pathology with T-test.

Pathology and cognitive tests in 70-year-olds with CDR 0.5

We then investigated participants with CDR 0.5 according to their CSF biomarker status, divided into three groups (amyloid pathology, P-tau pathology, and T-tau pathology). Among participants with CDR 0.5 (n = 57), 13 (22.8%) had amyloid pathology, 2 (3.5%) had P-tau pathology, and 18 (31.6%) had T-tau pathology. Participants with amyloid pathology had poorer performance in language (Word fluency) (17.38 versus 22.95, p = 0.003, Cohen’s D = 0.985) than those with no pathology (Table 3). The P-tau group was excluded from analyses because of the small sample size.

Table 3

Cognitive test performance and cerebrospinal fluid biomarker status in 70-year-olds from the Gothenburg Birth Cohort Studies (n = 57) with Clinical Dementia Rating 0.5

	Amyloid pathology	Amyloid pathology	No amyloid pathology	p¹	T-tau pathology	T-tau pathology	No T-tau pathology	p²
	N = 13	mean score (SD)	mean score (SD)	N = 18	mean score (SD)	mean score (SD)
MMSE	13	27.08 (1.38)	27.77 (1.51)	0.142	18	27.33 (1.19)	27.74 (1.62)	0.341
Memory
Immediate recall	12	6.67 (2.23)	6.51 (1.50)	0.778	17	6.06 (1.75)	6.76 (1.60)	0.149
Delayed recall	12	6.67 (1.92)	5.67 (1.89)	0.114	17	5.65 (1.80)	6.00 (1.99)	0.534
Word memory	12	3.58 (1.31)	3.71 (1.57)	0.793	17	3.35 (1.22)	3.84 (1.61)	0.275
Supra span (BUSII)	10	6.90 (1.45)	7.03 (1.70)	0.832	14	7.43 (1.02)	6.83 (1.81)	0.150
Thurstone’s picture memory test	10	20.20 (3.05)	20.07 (4.46)	0.933	16	20.94 (3.96)	19.71 (4.30)	0.339
Language
Word fluency	13	17.38 (5.77)	22.95 (5.54)	0.003	18	20.44 (7.03)	22.22 (5.50)	0.311
FAS	7	27.14 (7.20)	33.03 (10.94)	0.179	15	30.60 (9.61)	32.88 (11.12)	0.499
Executive function
SRB2	12	18.42 (3.55)	18.44 (3.55)	0.985	16	19.56(3.25)	17.95 (3.56)	0.126
Digit span backwards	12	3.42 (0.90)	3.93 (1.01)	0.117	17	3.53 (0.72)	3.95 (1.09)	0.155
Visuospatial
SRB3	11	15.45 (7.46)	17.24 (7.66)	0.493	16	18.38 (7.19)	16.22 (7.75)	0.346
Mental speed
Psif	12	20.58 (5.21)	22.98 (7.26)	0.293	17	23.35 (8.49)	22.03 (6.09)	0.516

p¹ = difference in mean score for amyloid pathology compared to non-amyloid pathology with T-test. p² = difference in mean score for T-tau pathology compared to non-T-tau pathology with T-test. There were two participants with P-tau pathology, and this group was excluded from the analyses due to the small size.

ATN classification system and cognitive performance in 70-year-olds with CDR 0

Participants with A-T-N+ had lower scores on the memory tests Delayed recall (7.33 versus 7.91, p = 0.042, Cohen’s D = 0.358) and Immediate recall (7.79 versus 8.38, p = 0.033, Cohen’s D = 0.365) (but these findings did not hold true after correction for multiple testing), and Supra span (7.07 versus 7.82, p = 0.003, Cohens’ D = 0.489).

The small group with A+T+N+ (N = 6) scored higher on language (Word fluency) (30.33 versus 25.33, p = 0.003, Cohen’s D = 0.981) than the group without any pathology. Those with A+T+N+ had lower scores on general cognitive function (MMSE) than the group without pathology (28.5 versus 29.28, p = 0.043, Cohen’s D = 0.724) as previously published [6]. This did not hold true after correction for multiple testing.

No other ATN group differed in cognitive test results from the group without pathology (A-T-N- group) (Table 4).

Table 4

Cognitive test performance in 70-year-olds from the Gothenburg Birth Cohort Studies (n = 259) with Clinical Dementia Rating 0 according to the ATN classification system for preclinical Alzheimer’s disease

	Normal CSF (A-T-N-)	A-T-N-	A+T-N-	A+T-N-	p¹	A+T-N+	A+T-N+	p²	A+T+N+	A+T+N+	p³	A-T-N+	A-T-N+	p⁴	A-T+N+	A-T+N+	p⁵
	N = 138	Mean Score (SD)	N = 34	Mean Score (SD)		N = 20	Mean Score (SD)		N = 6	Mean Score (SD)		N = 49	Mean Score (SD)		N = 12	Mean Score (SD)
MMSE	138	29.28(0.90)	34	29.35(0.77)	0.677	20	29.15(0.93)	0.542	6	28.50(1.23)	0.043^* ^†	48	29.25(1.14)	0.841	12	29.17(0.94)	0.672
Memory
Immediate recall	138	8.38(1.67)	34	8.06(1.46)	0.299	20	8.10(1.92)	0.487	6	8.67(0.82)	0.682	48	7.79(1.56)	0.033 ^*	12	8.17(1.85)	0.669
Delayed recall	138	7.91(1.72)	34	7.53(1.75)	0.256	20	7.20(2.04)	0.096	6	8.67 (1.03)	0.285	48	7.33(1.51)	0.042 ^*	12	7.58 (1.98)	0.539
Word memory	137	5.73 (1.86)	33	5.18 (1.88)	0.131	20	5.50 (1.67)	0.601	6	6.00(2.19)	0.730	47	5.55 (1.64)	0.563	11	5.27 (1.56)	0.429
Supra span (BUSII)	132	7.82 (1.42)	32	7.56 (1.74)	0.384	18	7.89 (1.23)	0.841	5	8.00 (1.23)	0.778	46	7.07 (1.64)	0.003	12	7.92 (1.24)	0.817
Thurstone’s picture memory test	124	23.06 (3.59)	32	22.09(4.17)	0.193	18	21.89(5.96)	0.428	6	20.50 (6.32)	0.369	46	22.76 (3.06)	0.621	12	24.08(3.23)	0.341
Language
Word fluency	137	25.33(6.73)	34	25.85(6.11)	0.678	20	23.30(7.24)	0.215	6	30.33(2.58)	0.003	48	23.35(5.54)	0.070	12	26.67(6.07)	0.506
FAS	132	42.27(13.56)	33	42.33(13.69)	0.982	19	42.84(12.64)	0.863	5	40.00(11.09)	0.712	46	40.33(15.32)	0.419	11	42.55(14.71)	0.949
Executive function
SRB2	134	20.56(4.28)	34	20.53(3.87)	0.970	19	20.68(4.57)	0.907	6	20.17(2.48)	0.824	48	20.25(3.78)	0.659	12	21.17(4.11)	0.638
Digit span backwards	136	4.63(1.19)	34	4.41(0.96)	0.333	19	4.95(1.27)	0.273	5	4.80(0.45)	0.743	47	4.34(1.05)	0.146	12	4.67(1.07)	0.907
Visuospatial
SRB3	133	22.11(6.64)	33	23.09(5.78)	0.436	19	21.74(8.39)	0.827	5	22.20(7.33)	0.975	47	20.30(6.62)	0.110	12	22.75(8.66)	0.754
Mental speed
Psif	136	29.90(7.42)	33	29.67(7.63)	0.870	19	27.95(10.41)	0.309	6	29.83(7.99)	0.982	48	29.88(7.95)	0.982	12	30.00(4.86)	0.965

p¹ = difference in mean score for A+T-N- compared to normal CSF with T-test. p² = difference in mean score for A+T-N+compared to normal CSF with T-test. p³ = difference in mean score for A+T+N+ compared to normal CSF with T-test. p⁴ = difference in mean score for A-T-N+ compared to normal CSF with T-test. p⁵ = difference in mean score for A-T+N+ compared to normal CSF with T-test. ^†Previously published in Kern et al. [6]. There were no participants with CDR 0 in the groups A-T+N- and A+T+N-. ^*Not significant at the 5% level after Benjamini-Hochberg correction

There were no individuals with A-T+N- or A+T+N-.

ATN classification system and cognitive performance in 70-year-olds with CDR 0.5

The distribution of the ATN preclinical AD categories in participants with CDR 0.5 was as follows: A-T-N- (n = 34), A+T-N-(n = 5), A+T-N+ (n = 6), A-T-N+ (n = 10), A+T+N+ (n = 2).

None of the participants with CDR 0.5 had A-T+N+, A-T+N-, or A+T+N- pathology. Due to the small number of participants, we chose not to further analyze according to ATN groups as findings may be spurious.

Power analysis

We performed power analyses on several of the results in the tables. The statistical power in most of our subgroups was satisfactory to good (60–80%).

DISCUSSION

This study investigated neurocognitive test performance in healthy older individuals from the general population with and without underlying CSF biomarker signs of AD pathology. Our study contributes to the sparse literature on population-based CSF data in healthy elderly. We found, besides subtle differences, that cognitively healthy older adults with AD pathology performed almost similar in all cognitive tests compared to healthy older adults without AD pathology. Our finding is in line with previous studies using PET [19–21 , 33]. One of these studies, performed in cognitively unimpaired older individuals from the community (mean age 74.4 years), showed that participants with underlying amyloid pathology were similar to participants without amyloid pathology regarding performance on a number of different cognitive tests examining several different cognitive domains [20]. Another study examined episodic memory in cognitively normal older people with amyloid-β positivity on PET in people aged 60 years and older from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the Berkeley Aging Cohort (BAC) [33] and could not detect any differences in test performance compared to participants without amyloid-β positivity in the ADNI sample, but in the BAC sample there was an association between elevated amyloid and worse episodic memory [33]. In contrast to our study, some population-based studies in older people report cognitive differences between healthy individuals with amyloid pathology on PET and CSF compared to healthy subjects without preclinical AD [7 , 18]. A few of these studies are based on volunteers [12, 18]. One possible explanation is that participants in some of these studies [11 , 16] had a higher mean age which may be related to worse cognitive performance.

Another possible explanation for the lack of differences in our study may be that most participants with preclinical AD were examined early in the disease process, before cognitive symptoms appear, a suggestion which is strengthened by the high mean test scores in the different cognitive tests. Our study is therefore in line with Jack’s hypothetical model of the development of AD [31], where CSF biomarkers are altered long before the onset of cognitive symptoms. Another possible explanation for the lack of differences is that cognitively normal participants with CDR 0 and both amyloid and tau pathology have a higher cognitive reserve which may have allowed them to display AD pathology without manifesting any cognitive impairment. In line with this theory, one study in cognitively normal older adults with amyloid pathology showed that older adults with high cognitive reserve had normal test results on a variety of cognitive tests [34]. However, participants with and without AD pathology in our study were similar regarding a number of factors, such as education, which may indicate similar premorbid function. Further, it is possible that there are ceiling effects due to the generally high cognitive performance in our sample of healthy 70-year-olds. One study showed that amyloid-related memory impairment could be detected with a difficult face-name associative memory test in a sample of older adults with CDR 0, who performed otherwise normally on less demanding neuropsychological tests [35].

Although cognitively healthy participants with and without AD pathology performed similarly on most tests, we found subtle differences on three memory tests. Participants with CDR 0 and A-T-N+ performed worse on memory tests (Immediate recall, Delayed recall, and Supra span). The finding of slightly worse memory performance in participants with T-tau pathology is in line with findings from a community based study of older adults enrolled in St. Louis, where cognitively normal adults with suspected non-AD pathophysiology (SNAP), who only had T-tau pathology, performed worse on episodic memory tests and global cognition (MMSE) [12]. Participants displaying A-T-N+ have previously been classified in the NIA-AA criteria as SNAP, reflecting that this is a heterogeneous group who may have a variety of other health related brain conditions (e.g., stroke, age-related tauopathies, Lewy body dementia) that could contribute to neurodegeneration [36]. A report from the Mayo Clinic Study of Aging showed that neurodegenerative markers, such as CSF T-tau and neurofilament light protein, increased the risk for progression from cognitively normal to MCI in the general population, independent of amyloid pathology status [37]. This might reflect that individuals with underlying neurodegeneration may have higher risk for conversion, not only to AD, but also to non-AD dementia. We did not find any statistically significant differences in cognitive performance on T-tau pathology (A+T-N+, A-T-N+, Table 2), despite seeing subtle differences in the A-T-N+ subgroup and Supra span memory test. This may be due to the small sample size of these groups.

Although participants with and without preclinical AD performed similarly on a number of cognitive tests, we unexpectedly found that the A+T+N+ group performed better than the A-T-N- group in Word fluency. However, there were only 6 persons in the A+T+N+ group, making it difficult to draw any conclusions based on this finding.

As expected, participants with CDR 0.5 performed worse on all cognitive tests compared to participants with CDR 0 [38 –40]. In this group, amyloid positive participants performed worse in category fluency tests than participants with CDR 0.5 and no amyloid pathology. Participants with CDR 0.5 and amyloid positivity could be expected to be on the AD-pathway, while the A-T-N- CDR 0.5 group could be more heterogeneous, and may include people with various types of brain damage or dementias unrelated to AD. The finding that amyloid positive participants with CDR 0.5 performed worse than their peers without such pathology suggests that they are closer to a conversion to dementia. This is in line with a report from the Mayo Clinic Study of Aging where amyloid positive participants with MCI had a higher risk for AD dementia than amyloid-negative participants [7].

Those with amyloid pathology had lower scores on the category fluency test (Word fluency) than those without such pathology among participants with CDR 0.5. However, the two groups scored similarly on the phonemic fluency test (FAS). This is in line with studies reporting that patients with AD perform worse in category fluency than in letter fluency tests [41 –43]. Impairment in category fluency may be seen as early as in the MCI stage [44 –46], although MCI patients have also been reported to perform low in both category and letter fluency [47]. It is noteworthy, and rather surprising, that there were no associations between amyloid positivity and memory performance in those with CDR 0.5.

Strength and limitations

The strengths of this study include the comprehensively examined population-based sample, which was systematically selected based on birth year. In contrast to many other studies that use convenience samples or volunteers, this sample is likely to be representative of the general population of 70-year-olds in Sweden. A relatively large proportion agreed to LP, yielding a relatively large CSF sample. However, due to contraindications, participants taking anticoagulant treatment were excluded from LP, which may have led to a selection bias with an overrepresentation of healthier participants. Another limitation is that many ATN groups were small, and this may give rise to spurious findings and low statistical power. It is possible that in these smaller subsamples there might have been subtle cognitive differences between the groups that were not detected either due to low power in small groups or that the cognitive tests were not challenging enough. However, the statistical power (60–80%) in some of our subgroups was adequate for the T-test to detect subtle differences in means. Thus performing T-tests in some of our subgroups was meaningful in relation to sample size. In addition, there may be some false positive results due to multiple testing. However, multiple testing introduces a decrease in statistical power: this may shift the possibility of false positive findings to a risk of false negative results. It might therefore be advisable to interpret adjustments for multiple tests with caution. One way to handle this problem, which we have followed, is to state the number of comparisons made and to emphasize that the findings should be considered only suggestive until further confirmed [48]. In addition to the p-values, we added a Benjamini-Hochberg correction so the reader could interpret the results in different ways. As expected, some significances did not hold true after this correction.

In our study, P-tau and T-tau were highly correlated, as reported previously [17]. Thus, these pathologies might not necessarily be regarded as separate entities, and may affect each other.

Further, the ATN classifications are dependent on the cut points chosen. We have therefore used cut points which have been used in previous studies [49]. Another limitation in the study is the lack of serial cognitive data which would be necessary in order to establish the trajectory of cognitive decline.

Lastly, this study examined 70-year-old Swedish citizens, and results cannot be generalized to other populations.

In conclusion, this study showed that cognitively normal 70-year-olds with pathological AD markers performed similarly to their peers without biomarker-defined preclinical AD in most cognitive domains. The study supports the hypothesis that preclinical AD biomarkers are altered before onset of cognitive decline.

Footnotes

ACKNOWLEDGMENTS

IS was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALF 716681), the Swedish Research Council (2012-5041, 2015-02830, 2019-01096, 2013-8717, 2017-00639)

Swedish Research Council for Health, Working Life and Welfare (2013-1202, 2018-00471, AGECAP 2013-2300, 2013-2496), Konung Gustaf V:s och Drottning Victorias Frimurarestiftelse, Hjärnfonden, Alzheimerfonden, Eivind och Elsa K:son Sylvans stiftelse

SK is supported by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG-81392, ALF GBG-771071), the Swedish Research Council (2019-02075), the Alzheimerfonden (AF-842471, AF-737641), Stiftelsen Demensfonden, Stiftelsen Hjalmar Svenssons Forskningsfond, Stiftelsen Wilhelm och Martina Lundgrens vetenskapsfond. HZ is a Wallenberg Scholar supported by grants from the Swedish Research Council (#2018-02532), the European Research Council (#681712), the Swedish Brain Foundation (FO2019-0228), and Swedish State Support for Clinical Research (#ALFGBG-720931). KB holds the Torsten Söderberg Professorship in Medicine at the Royal Swedish Academy of Sciences, and is supported by the Swedish Research Council (#2017-00915), the Swedish Alzheimer Foundation (#AF-742881), Hjärnfonden, Sweden (#FO2017-0243), the Alzheimer Drug Discovery foundation (ADDF), USA (#RDAPB-201809-2016615), and European Union Joint Program for Neurodegenerative Disorders (JPND2019-466-236)The study was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALF 716681, ALFGBG-715986). The Swedish Research Council 2015-02830, 2013-8717, Swedish Research Council for Health, Working Life and Welfare (2013-1202, 2018-00471, 2013-2300, 2013-2496, 2013-0475), Hjärnfonden, Konung Gustaf V:s och Drottning Victorias Frimurarestiftelse, Alzheimerfonden, Eivind och Elsa K:son Sylvans stiftelse.

Authors’ disclosures available online ().

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

References

Blennow

, de Leon

, Zetterberg

(2006) Alzheimer’s disease. Lancet 368, 387–403.

Rajan

, Wilson

, Weuve

, Barnes

, Evans

(2015) Cognitive impairment 18 years before clinical diagnosis of Alzheimer disease dementia. Neurology 85, 898–904.

Amieva

, Jacqmin-Gadda

, Orgogozo

, Le Carret

, Helmer

, Letenneur

, Barberger-Gateau

, Fabrigoule

, Dartigues

(2005) The 9 year cognitive decline before dementia of the Alzheimer type: A prospective population-based study. Brain 128, 1093–1101.

Amieva

, Le Goff

, Millet

, Orgogozo

, Peres

, Barberger-Gateau

, Jacqmin-Gadda

, Dartigues

(2008) Prodromal Alzheimer’s disease: Successive emergence of the clinical symptoms. Ann Neurol 64, 492–498.

Ribbe

, Kern

, Borjesson Hansson

, Ostling

, Zetterberg

, Blennow

, Skoog

(2019) Amyloid beta42 and total tau levels in cerebrospinal fluid associate with survival in an 85-year-old population-based cohort followed until death. Dement Geriatr Cogn Disord 47, 114–124.

Kern

, Zetterberg

, Kern

, Zettergren

, Waern

, Hoglund

, Andreasson

, Wetterberg

, Borjesson-Hanson

, Blennow

, Skoog

(2018) Prevalence of preclinical Alzheimer disease: Comparison of current classification systems. Neurology 90, e1682–e1691.

Roberts

, Aakre

, Kremers

, Vassilaki

, Knopman

, Mielke

, Alhurani

, Geda

, Machulda

, Coloma

, Schauble

, Lowe

, Jack

Jr. , Petersen

(2018) Prevalence and outcomes of amyloid positivity among persons without dementia in a longitudinal, population-based setting. JAMA Neurol 75, 970–979.

Morris

, Roe

, Grant

, Head

, Storandt

, Goate

, Fagan

, Holtzman

, Mintun

(2009) Pittsburgh compound B imaging and prediction of progression from cognitive normality to symptomatic Alzheimer disease. Arch Neurol 66, 1469–1475.

Skoog

, Davidsson

, Aevarsson

, Vanderstichele

, Vanmechelen

, Blennow

(2003) Cerebrospinal fluid beta-amyloid 42 is reduced before the onset of sporadic dementia: A population-based study in 85-year-olds. Dement Geriatr Cogn Disord 15, 169–176.

10.

Doraiswamy

, Sperling

, Johnson

, Reiman

, Wong

, Sabbagh

, Sadowsky

, Fleisher

, Carpenter

, Joshi

, Lu

, Grundman

, Mintun

, Skovronsky

, Pontecorvo

(2014) Florbetapir F 18 amyloid PET and 36-month cognitive decline: A prospective multicenter study. Mol Psychiatry 19, 1044–1051.

11.

Lim

, Pietrzak

, Ellis

, Jaeger

, Harrington

, Ashwood

, Szoeke

, Martins

, Bush

, Masters

, Rowe

, Villemagne

, Ames

, Darby

, Maruff

(2013) Rapid decline in episodic memory in healthy older adults with high amyloid-beta. J Alzheimers Dis 33, 675–679.

12.

Vos

, Xiong

, Visser

, Jasielec

, Hassenstab

, Grant

, Cairns

, Morris

, Holtzman

, Fagan

(2013) Preclinical Alzheimer’s disease and its outcome: A longitudinal cohort study. Lancet Neurol 12, 957–965.

13.

van Harten

, Smits

, Teunissen

, Visser

, Koene

, Blankenstein

, Scheltens

, van der Flier

(2013) Preclinical AD predicts decline in memory and executive functions in subjective complaints. Neurology 81, 1409–1416.

14.

Fagan

, Roe

, Xiong

, Mintun

, Morris

, Holtzman

(2007) Cerebrospinal fluid tau/beta-amyloid(42) ratio as a prediction of cognitive decline in nondemented older adults. Arch Neurol 64, 343–349.

15.

Knopman

, Jack

Jr. , Wiste

, Weigand

, Vemuri

, Lowe

, Kantarci

, Gunter

, Senjem

, Ivnik

, Roberts

, Boeve

, Petersen

(2012) Short-term clinical outcomes for stages of NIA-AA preclinical Alzheimer disease. Neurology 78, 1576–1582.

16.

Petersen

, Wiste

, Weigand

, Rocca

, Roberts

, Mielke

, Lowe

, Knopman

, Pankratz

, Machulda

, Geda

, Jack

Jr. (2016) Association of elevated amyloid levels with cognition and biomarkers in cognitively normal people from the community. JAMA Neurol 73, 85–92.

17.

Vos

SJB

, Gordon

, Su

, Visser

, Holtzman

, Morris

, Fagan

, Benzinger

TLS

(2016) NIA-AA staging of preclinical Alzheimer disease: Discordance and concordance of CSF and imaging biomarkers. Neurobiol Aging 44, 1–8.

18.

Burnham

, Bourgeat

, Dore

, Savage

, Brown

, Laws

, Maruff

, Salvado

, Ames

, Martins

, Masters

, Rowe

, Villemagne

(2016) Clinical and cognitive trajectories in cognitively healthy elderly individuals with suspected non-Alzheimer’s disease pathophysiology (SNAP) or Alzheimer’s disease pathology: A longitudinal study. Lancet Neurol 15, 1044–1053.

19.

Villemagne

, Pike

, Chetelat

, Ellis

, Mulligan

, Bourgeat

, Ackermann

, Jones

, Szoeke

, Salvado

, Martins

, O’Keefe

, Mathis

, Klunk

, Ames

, Masters

, Rowe

(2011) Longitudinal assessment of Abeta and cognition in aging and Alzheimer disease. Ann Neurol 69, 181–192.

20.

Aizenstein

, Nebes

, Saxton

, Price

, Mathis

, Tsopelas

, Ziolko

, James

, Snitz

, Houck

, Bi

, Cohen

, Lopresti

, DeKosky

, Halligan

, Klunk

(2008) Frequent amyloid deposition without significant cognitive impairment among the elderly. Arch Neurol 65, 1509–1517.

21.

Jack

Jr. , Lowe

, Senjem

, Weigand

, Kemp

, Shiung

, Knopman

, Boeve

, Klunk

, Mathis

, Petersen

(2008) 11C PiB and structural MRI provide complementary information in imaging of Alzheimer’s disease and amnestic mild cognitive impairment. Brain 131, 665–680.

22.

Rydberg Sterner

, Ahlner

, Blennow

, Dahlin-Ivanoff

, Falk

, Havstam Johansson

, Hoff

, Holm

, Horder

, Jacobsson

, Johansson

, Kern

, Machado

, Mellqvist Fassberg

, Nilsson

, Ribbe

, Rothenberg

, Ryden

, Sadeghi

, Sacuiu

, Samuelsson

, Sigstrom

, Skoog

, Thorvaldsson

, Waern

, Westman

, Wetterberg

, Zetterberg

, Zettergren

, Ostling

, Skoog

(2019) The Gothenburg H70 Birth cohort study 2014-16: Design, methods and study population. Eur J Epidemiol 34, 191–209.

23.

Kern

, Kern

, Blennow

, Zetterberg

, Waern

, Guo

, Borjesson-Hanson

, Skoog

, Ostling

(2016) Calcium supplementation and risk of dementia in women with cerebrovascular disease. Neurology 87, 1674–1680.

24.

Skoog

, Borjesson-Hanson

, Kern

, Johansson

, Falk

, Sigstrom

, Ostling

(2017) Decreasing prevalence of dementia in 85-year olds examined 22 years apart: The influence of education and stroke. Sci Rep 7, 6136.

25.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

26.

Thorvaldsson

, Karlsson

, Skoog

, Johansson

(2017) Better cognition in new birth cohorts of 70 year olds, but greater decline thereafter. J Gerontol B Psychol Sci Soc Sci 72, 16–24.

27.

Österberg

IDLKE

(1971) Manual till DS-batteriet (Manual for the DS-battery). Psykologiförlaget AB.

28.

Vanmechelen

, Vanderstichele

, Davidsson

, Van Kerschaver

, Van Der Perre

, Sjogren

, Andreasen

, Blennow

(2000) Quantification of tau phosphorylated at threonine 181 in human cerebrospinal fluid: A sandwich ELISA with a synthetic phosphopeptide for standardization. Neurosci Lett 285, 49–52.

29.

Blennow

, Wallin

, Agren

, Spenger

, Siegfried

, Vanmechelen

(1995) Tau protein in cerebrospinal fluid: A biochemical marker for axonal degeneration in Alzheimer disease? Mol Chem Neuropathol 26, 231–245.

30.

Andreasen

, Hesse

, Davidsson

, Minthon

, Wallin

, Winblad

, Vanderstichele

, Vanmechelen

, Blennow

(1999) Cerebrospinal fluid beta-amyloid(1-42) in Alzheimer disease: Differences between early- and late-onset Alzheimer disease and stability during the course of disease. Arch Neurol 56, 673–680.

31.

Jack

Jr. , 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.

32.

Hoglund

, Kern

, Zettergren

, Borjesson-Hansson

, Zetterberg

, Skoog

, Blennow

(2017) Preclinical amyloid pathology biomarker positivity: Effects on tau pathology and neurodegeneration. Transl Psychiatry 7, e995.

33.

Mormino

, Kluth

, Madison

, Rabinovici

, Baker

, Miller

, Koeppe

, Mathis

, Weiner

, Jagust

(2009) Episodic memory loss is related to hippocampal-mediated beta-amyloid deposition in elderly subjects. Brain 132, 1310–1323.

34.

Rentz

, Locascio

, Becker

, Moran

, Eng

, Buckner

, Sperling

, Johnson

(2010) Cognition, reserve, and amyloid deposition in normal aging. Ann Neurol 67, 353–364.

35.

Rentz

, Amariglio

, Becker

, Frey

, Olson

, Frishe

, Carmasin

, Maye

, Johnson

, Sperling

(2011) Face-name associative memory performance is related to amyloid burden in normal elderly. Neuropsychologia 49, 2776–2783.

36.

Vassilaki

, Aakre

, Kremers

, Mielke

, Geda

, Alhurani

, Dutt

, Machulda

, Knopman

, Vemuri

, Coloma

, Schauble

, Lowe

, Jack

Jr. , Petersen

, Roberts

(2019) The association of multimorbidity with preclinical AD stages and SNAP in cognitively unimpaired persons. J Gerontol A Biol Sci Med Sci 74, 877–883.

37.

Kern

, Syrjanen

, Blennow

, Zetterberg

, Skoog

, Waern

, Hagen

, van Harten

, Knopman

, Jack

Jr. , Petersen

, Mielke

(2019) Association of cerebrospinal fluid neurofilament light protein with risk of mild cognitive impairment among individuals without cognitive impairment. JAMA Neurol 76, 187–193.

38.

Zhao

, Lv

, Zhou

, Hong

, Guo

(2012) Short-term delayed recall of auditory verbal learning test is equivalent to long-term delayed recall for identifying amnestic mild cognitive impairment. PLoS One 7, e51157.

39.

Grundman

, Petersen

, Ferris

, Thomas

, Aisen

, Bennett

, Foster

, Jack

Jr. , Galasko

, Doody

, Kaye

, Sano

, Mohs

, Gauthier

, Kim

, Jin

, Schultz

, Schafer

, Mulnard

, van Dyck

, Mintzer

, Zamrini

, Cahn-Weiner

, Thal

(2004) Mild cognitive impairment can be distinguished from Alzheimer disease and normal aging for clinical trials. Arch Neurol 61, 59–66.

40.

Meguro

, Shimada

, Yamaguchi

, Sano

, Inagaki

, Matsushita

, Sekita

, Mori

(2004) Neuropsychosocial features of very mild Alzheimer’s disease (CDR 0.5) and progression to dementia in a community: The Tajiri project. J Geriatr Psychiatry Neurol 17, 183–189.

41.

Cerhan

, Ivnik

, Smith

, Tangalos

, Petersen

, Boeve

(2002) Diagnostic utility of letter fluency, category fluency, and fluency difference scores in Alzheimer’s disease. Clin Neuropsychol 16, 35–42.

42.

Henry

, Crawford

, Phillips

(2004) Verbal fluency performance in dementia of the Alzheimer’s type: A meta-analysis. Neuropsychologia 42, 1212–1222.

43.

Salmon

, Heindel

, Lange

(1999) Differential decline in word generation from phonemic and semantic categories during the course of Alzheimer’s disease: Implications for the integrity of semantic memory. J Int Neuropsychol Soc 5, 692–703.

44.

Teng

, Leone-Friedman

, Lee

, Woo

, Apostolova

, Harrell

, Ringman

, Lu

(2013) Similar verbal fluency patterns in amnestic mild cognitive impairment and Alzheimer’s disease. Arch Clin Neuropsychol 28, 400–410.

45.

Cottingham

, Hawkins

(2010) Verbal fluency deficits co-occur with memory deficits in geriatric patients at risk for dementia: Implications for the concept of mild cognitive impairment. Behav Neurol 22, 73–79.

46.

Lonie

, Herrmann

, Tierney

, Donaghey

, O’Carroll

, Lee

, Ebmeier

(2009) Lexical and semantic fluency discrepancy scores in aMCI and early Alzheimer’s disease. J Neuropsychol 3, 79–92.

47.

Mirandez

, Aprahamian

, Talib

, Forlenza

, Radanovic

(2017) Multiple category verbal fluency in mild cognitive impairment and correlation with CSF biomarkers for Alzheimer’s disease. Int Psychogeriatr 29, 949–958.

48.

Rothman

(1986) Modern epidemiology. Little, Brown, Boston, pp. 147-151.

49.

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.