Progression from Subjective Cognitive Decline to Mild Cognitive Impairment or Dementia: The Role of Baseline Cognitive Performance

Abstract

Background:

Older adults with subjective cognitive decline (SCD) are at an increased risk of progression to mild cognitive impairment (MCI) or dementia. However, few have examined the specific cognitive tests that are associated with progression.

Objective:

This study examined performance on 18 neuropsychological tests among participants with SCD who later progressed to MCI or dementia.

Methods:

We included 131 participants from the Czech Brain Aging Study that had SCD at baseline. They completed a comprehensive neuropsychological battery including cognitive tests from the Uniform Data Set 2.0 enriched by the verbal memory test Rey Auditory Verbal Learning Test (RAVLT) and Rey-Osterrieth Complex Figure Test (ROCFT).

Results:

Fifty-five participants progressed: 53% to non-amnestic MCI (naMCI), 44% to amnestic MCI (aMCI), and 4% to dementia. Scoring one SD below the mean at baseline on the RAVLT 1 and RAVLT 1–5 was associated with 133% (RAVLT 1; HR: 2.33 [1.50, 3.62]) and 122% (RAVLT 1–5; HR: 2.22 [1.55, 3.16]) greater risk of progression to MCI or dementia over 3.84 years on average. Worse performance on the RAVLT 5, RAVLT 1–5, RAVLT 30, and ROCFT–Recall was associated with progression to aMCI whereas worse performance on the RAVLT 1, TMT B, and Boston Naming Test was associated with progression to naMCI.

Conclusion:

At baseline, lower verbal memory performance was most strongly associated with progression to aMCI whereas lower executive or language performance was most strongly associated with progression to naMCI.

Keywords

Cognitive aging immediate recall subjective memory complaints verbal memory

INTRODUCTION

Subjective cognitive decline (SCD) is a clinical condition describing those with self-experienced (or informant reported) persistent decline in cognitive capacity in comparison with a previously normal status, but with no substantial impairment on standardized cognitive tests that would warrant a diagnosis of mild cognitive impairment (MCI) [1]. Annual progression rates in patients presenting with subjective memory complaints to MCI and to dementia are higher than in older adults without cognitive complaints, though these rates substantially differ by community or clinic setting [2]. Multiple causes of SCD have been identified and may include Alzheimer’s disease (AD) or other neurodegenerative diseases [3, 4], underlying vascular pathology [5, 6], and psychiatric symptomatology [7]. In some cases, SCD may reverse pending adequate treatment of underlying issues (e.g., treatment of depression, addressing unintended effects of medications) [8].

SCD can be broadly categorized as reversible SCD (e.g., cognitive complaints that remit), stable SCD with no further cognitive decline, and SCD with subsequent progression to MCI or dementia [8]. Early identification of progression from SCD to amnestic MCI (aMCI) versus non-amnestic MCI (naMCI) is of great public health importance. Literature suggests that the incidence of AD dementia is greater among adults with aMCI (single or multiple domain) than adults with naMCI (single or multiple domain) [9, 10], while adults with naMCI may progress more often to other forms of dementia (e.g., vascular dementia, frontotemporal dementia, dementia with Lewy bodies) [11]. Moreover, patterns of neuropsychiatric symptoms differ by MCI subtype. Patients with aMCI are more likely to exhibit a greater number of neuropsychiatric symptoms and specifically affective symptoms, while patients with naMCI are more likely to experience hallucinations and sleep disturbances [12, 13]. These specific cognitive and neuropsychiatric phenotypes may differentially affect the trajectory toward dementia in patients with MCI. Thus identifying progression to aMCI or naMCI should be a clinical priority to reduce caregiver burden through the early deployment of services and supports.

Several factors have already been associated with progression to MCI or dementia among adults with SCD: anxiety symptomatology and cognitive complaints [14], a memory clinic-based sample [15], positive AD biomarkers such as amyloid-β (Aβ)-positive scans or cerebrospinal fluid Aβ₄₂, p-tau, and medial temporal lobe degeneration [16 –18], white matter hyperintensities [19], worse baseline cognitive performance [20], older age [21, 22], and lower cognitive reserve [23 –25]. However, less is known about the specific cognitive tests that may predict progression, and whether tests can correctly identify aMCI from naMCI. Therefore, the goal of this study was to determine the cognitive tests most associated with progression to MCI or dementia among participants with SCD. Based on prior literature suggesting that verbal memory performance (particularly delayed recall memory) is associated with progression from SCD to MCI or dementia [21, 26] and that the Rey Auditory Verbal Learning Test (RAVLT) is associated with progression from MCI to AD dementia [27 –29], we hypothesized that performance on the RAVLT delayed recall would be most associated with progression from SCD to MCI or dementia. We provide exploratory analyses to examine whether specific cognitive tests were associated with progression from SCD to aMCI or naMCI.

MATERIALS AND METHODS

Participants

A total of 131 SCD participants were recruited and followed prospectively with at least one follow-up examination at the Memory Clinic in Motol University Hospital in Prague, Czech Republic in the Czech Brain Aging Study (CBAS) [30] (n = 205 participants with SCD were excluded due to no follow-up assessments beyond baseline). All individuals were referred to the clinic by general practitioners, neurologists, psychiatrists, or geriatricians based on memory complaints reported by themselves or their informants. Participants underwent standard clinical and laboratory evaluations, brain magnetic resonance imaging, and comprehensive neuropsychological examination at baseline, and were followed prospectively with clinical and neuropsychological evaluations in order to detect progression to MCI or dementia. After the initial visit, participants were examined one year later, then regularly on a biannual basis. Extra visits were available if significant changes were reported by the patient or by the informant.

Definition of SCD

SCD was defined as having normal cognitive performance on a neuropsychological battery (i.e., no single score more than 1.5 standard deviations (SD) or no multiple scores more than 1 SD below age- and education-adjusted means), self-reported persistent cognitive complaints with onset within the last five years, which motivated the consultation and was unrelated to an acute event, and had a Clinical Dementia Rating (CDR) [31] global score of≤0.5 [1]. Those who did not progress to MCI or dementia were labeled “stable SCD” and those who did progress to MCI or dementia were labeled “progressors.”

Definition of MCI

Participants with objective cognitive decline were classified as having MCI based on Petersen’s criteria [32] and included persons with the following: 1) subjectively perceived cognitive decline compared to a previously normal status; 2) neuropsychologically-confirmed objective cognitive impairment ≥1.5 SDs below the mean of age- and education-adjusted norms in at least one of five established cognitive domains (attention/processing speed, executive functioning, language, memory, visuospatial skills) [33]; 3) preservation of independence in functional abilities; and 4) the absence of dementia. Participants with a deficit in memory were labeled as amnestic MCI (aMCI) and participants with a deficit in a non-memory domain(s) were labeled as non-amnestic MCI (naMCI). Participants with deficits in memory and another domain were labeled as aMCI (e.g., multiple domain aMCI). A small proportion of progressors may revert back to SCD or normal functioning [9, 34]; this reversion was not captured in our study, as participants’ data were censored after the date of MCI/dementia diagnosis.

CBAS inclusion and exclusion criteria

CBAS is an ongoing longitudinal memory clinic-based study aimed at detecting early changes associated with pathological brain aging, enrolling non-demented older adults with cognitive difficulties perceived by themselves or by their informants [30]. All diagnoses were based on the relevant clinical history reported by the patient, neurological examination, neuropsychological assessment, and magnetic resonance imaging. Exclusion criteria include a diagnosis of a neurological or other psychiatric disorder (e.g., major depression, bipolar disorder, etc.), a systemic condition potentially causing cognitive impairment, or history of stroke. All participants in this study had signed written informed consent that was approved by a local ethics committee before partaking in study protocols. The progression to MCI or dementia was determined during the regular meetings of neurologists and neuropsychologists, where consensus was established. Although these data were collected for research purposes, participants were provided with the clinical interpretation of all assessments and were referred appropriately to clinical resources.

Neuropsychological and psychometric assessment

The neuropsychological battery included the Czech version of the UDS 2.0 enriched by two additional memory tests and phonemic verbal fluency. Specifically, the battery included the Mini-Mental State Examination (MMSE), tests of attention and working memory (Digit Span Forward and Backward tests; Digit Symbol Modalities subtest; Trail Making Test Part A (TMT A)), tests of executive functioning (TMT Part B (TMT B); Phonemic Verbal Fluency Czech version with letters N, K, P), tests of language (Boston Naming Test 60-item; Semantic Verbal Fluency Animals and Vegetables), and one test of visuospatial skills (Rey–Osterrieth Complex Figure Test (ROCFT) –Copy condition (ROCFT-Copy)) [35 –37]. One strength of the CBAS neuropsychological and psychometric protocol is the availability of normative data from the Czech version of the Uniform Data Set 2.0 (UDS 2.0) [36], which was used in this study.

Three types of memory tests were included: verbal memory with encoding and delayed recall (RAVLT) [38, 39], story learning memory (Logical Memory with immediate and delayed conditions; LM) [36], and visuospatial memory (ROCFT –Recall condition (ROCFT-Recall). The RAVLT is a sensitive test of verbal memory that strongly correlates with progression to AD dementia [27 –29]. First, the patient is read a list of 15 words and is asked to immediately recall as many words as possible (first acquisition trial [RAVLT 1] indexing working memory). The trial is repeated five times, allowing for the calculation of two additional scores. The final acquisition trial (RAVLT 5) measures peak performance of immediate verbal recall memory, while total acquisition (RAVLT 1–5) measures immediate verbal recall plus the effect of learning. After 30 minutes of other cognitive testing, the patient is once again asked to recall the list of 15 words to measure delayed verbal recall (RAVLT 30). The RAVLT delayed verbal recall subscore has previously been associated with progression to AD [27, 28].

Depressive (Geriatric Depression Scale, 15-item version) [40] and anxious (Beck Anxiety Inventory) [41] symptomatology were measured at baseline.

Statistical analysis

Initially, progressors and those with stable SCD were compared on demographic characteristics and baseline cognitive scores using t-tests for differences in means and chi-square tests for differences in frequencies. Subsequently, cognitive test scores were transformed to Z-scores for ease of comparison across individual tests. To allow further direct comparison, the Z-scores for TMT A and B were reversed.

In the main analyses, progression to MCI or de-mentia as a function of cognitive performance was examined using Cox proportional hazards models. Age, sex, and education were included as covariates in the models and the corresponding hazard ratios (HRs) and 95% confidence intervals (CI) were extracted. Interpretation: An HR of 1.50 on the RAVLT 1 would suggest that scoring one SD below the mean on the RAVLT 1 at baseline was associated with a 50% increased risk of progression to MCI or dementia over 3.84 years on average. The proportional hazards assumption was examined with formal statistical tests and plots of the scaled Schoenfeld residuals, and no major divergence was found.

We provide estimates from logistic regressions for progression to aMCI or naMCI as a function of cognitive performance, again controlling for age, sex, and education. Only cognitive tests with p-values < 0.10 from the Cox models were included in the secondary analysis with logistic regression. This p-value cutoff was chosen to reduce the risk of Type I Error. Estimates were exponentiated and reported as odds ratios (OR) with corresponding 95% CI. Interpretation: An OR of 1.50 on the RAVLT 1 would suggest that scoring one SD below the mean on the RAVLT 1 at baseline was associated with a 50% greater odds of an naMCI progression at follow-up. Because MCI diagnostic categories are competing risks for each other (i.e., a diagnosis of aMCI would preclude a diagnosis of naMCI and vice-versa), Fine-Gray competing risks models would be most appropriate for the secondary analyses. However, logistic regression was used in favor of Fine-Gray competing risks models because of the very small sample size of those who progressed (i.e., 24 aMCI and 29 naMCI). In order to capture the effect of competing risks, we provide cumulative incidence functions in a plot (Fig. 1). These functions show the relative probabilities of progressing from SCD to aMCI or naMCI after controlling for competing risks. Analyses were conducted in R3.6.1. (R Core Team, 2013).

Fig. 1

Cumulative incidence functions showing the relative probabilities of progression from SCD to aMCI or naMCI. aMCI, amnestic mild cognitive impairment; naMCI, non-amnestic mild cognitive impairment. The x-axis is time given in years and the y-axis is the probability of progression from subjective cognitive decline to aMCI or naMCI at any given timepoint. The cumulative incidence functions take competing risks into account when presenting probabilities of progression (i.e., the probability of progression to aMCI on year five after controlling for the competing risk of progression to naMCI).

RESULTS

See Table 1 for demographic characteristics and baseline cognitive performance among progressors and those with stable SCD. The average time of follow-up among all participants was 3.84 years (Range: 0.45 to 13.86 years), and the average time of follow-up among SCD progressors was 3.54 years (Range: 0.78 to 13.86 years). Among the 131 participants with SCD, 42% (n = 55) progressed to a clinical level of impairment. Of those who progressed, 53% (n = 29) were diagnosed with naMCI, 44% (n = 24) were diagnosed with aMCI, and 4% (n = 2) were diagnosed with dementia. The majority of participants with SCD were diagnosed at their second visit (55%), while 25% were diagnosed at their third visit, 13% at their fourth visit, 5% at their fifth visit, and 2% at their sixth visit. Progressors did not differ on sex, education, depressive symptomatology, or anxious symptomatology from those with stable SCD, but they were older on average at baseline (71.00 years versus 64.04 years; p < 0.001) (Table 1). At baseline, progressors performed worse on all memory scores (RAVLT, ROCFT-Recall, LM), both measures of executive functioning (TMT B, Phonemic Verbal Fluency), all measures of language (Boston Naming Test, Semantic Verbal Fluency Animals, Semantic Verbal Fluency Vegetables), the measure of visuospatial skills (ROCFT-Copy), and one measure of processing speed (Digit Symbol Modalities).

Table 1

Demographic characteristics and cognitive performance of participants who progressed and those with stable subjective cognitive decline

	Baseline	Follow-up
	SCD	Stable SCD	Progressors
	(n = 131)	(n = 76)	(n = 55)
	Mean±SD	Mean±SD	Mean±SD
Demographics
Years of follow-up	3.84±2.62	4.06±2.68	3.54±2.52
Gender (male/female)	48/83	24/52	24/31
Age	67.01±8.71	64.04±7.16	71.00±9.06^***
Education	15.89±2.98	16.12±3.02	15.56±2.92
GDS-15	3.43±3.00	3.30±3.11	3.60±2.88
BAI	9.05±6.72	8.22±6.36	10.30±7.12
Progressor diagnosis			% (n)
Amnestic MCI	-	-	44% (24)
Non-amnestic MCI	-	-	53% (29)
Dementia	-	-	4% (2)
Cognitive performance (baseline)
MMSE	28.85±1.16	28.81±1.22	28.91±1.09
RAVLT 1	6.39±2.03	7.12±2.07	5.40±1.50^***
RAVLT 5	12.47±1.97	13.04±1.63	11.70±2.13^***
RAVLT 1–5	50.89±8.66	54.12±7.74	46.49±7.91^***
RAVLT 30	10.36±2.96	11.06±2.34	9.42±3.44^**
ROCFT-Recall	17.10±6.61	19.54±5.67	13.59±6.33^***
ROCFT-Copy	30.74±5.87	32.25±2.64	28.61±8.14^***
LM-I	15.58±3.54	16.18±3.77	14.72±3.03^*
LM-D	13.71±3.88	14.53±3.52	12.53±4.10^**
TMT A	38.99±12.30	37.88±11.32	40.58±13.57
TMT B	89.12±28.31	81.32±20.04	99.84±34.19^***
F-Digit Span-SC	9.33±2.15	9.33±2.11	9.34±2.24
B-Digit Span-SC	6.98±1.93	7.11±1.85	6.81±2.05
Digit Symbol Modalities	43.88±10.13	46.34±9.71	40.29±9.73^**
BNT 60	57.08±3.71	57.74±2.67	56.18±4.67^*
P-VF	46.33±11.01	48.04±10.93	43.98±10.79^*
S-VF-A	25.38±5.76	26.38±5.52	23.55±5.83^*
S-VF-V	13.26±3.75	13.88±3.66	12.12±3.71^*

Statistically significant differences between stable subjective cognitive decline (SCD) and progressors: ^***p < 0.001, ^**p < 0.01, ^*p < 0.05. SCD, subjective cognitive decline; MCI, mild cognitive impairment; GDS, Geriatric Depression Scale; BAI, Beck Anxiety Inventory; MMSE, Mini-Mental State Examination, total score; RAVLT 1, Rey Auditory Verbal Learning Test, trial 1 recall; RAVLT 5, Rey Auditory Verbal Learning Test, trial 5 recall; RAVLT 1–5, Rey Auditory Verbal Learning Test, sum of trials 1 to 5; RAVLT 30, Rey Auditory Verbal Learning Test, recall after 30 minutes; ROCFT-Recall, Rey–Osterrieth Complex Figure Test Recall condition –after 3 minutes; ROCFT-Copy, Rey–Osterrieth Complex Figure Test Copy condition; LM-I, Logical Memory Immediate Recall from the Uniform Data Set; LM-D, Logical Memory Delayed Recall from the Uniform Data Set; TMT A, Trail Making Test Part A, given in seconds; TMT B, Trail Making Test Part B, given in seconds; F-Digit Span-SC, Forward Digit Span –score; B-Digit Span-SC, Backward Digit Span –score; Digit Symbol Modalities, Digit Symbol Score from the WAIS-R; BNT, Boston Naming Test 60-item; P-VF, Phonemic Verbal Fluency; S-VF-A, Semantic Verbal Fluency –Animals; S-VF-V, Semantic Verbal Fluency –Vegetables.

Cognitive tests associated with progression

See Table 2 for adjusted HRs, 95% CIs, and corresponding p-values. Figure 2 presents the adjusted HRs graphically in a forest plot. After controlling for age, sex, and education, scoring one SD below the mean on subcomponents from the RAVLT (RAVLT 1, RAVLT 5, RAVLT 1-5, but not RAVLT 30), ROCFT-Recall, TMT A and B, and the Boston Naming Test were significantly associated with a greater risk of progression to MCI or dementia. Noticeably, the RAVLT 1 (HR: 2.33, [1.50, 3.62], p < 0.001) and RAVLT 1–5 (HR: 2.22, [1.55, 3.16], p < 0.001) were more strongly associated with progression than the TMT B (HR: 1.50, [1.10, 2.05], p = 0.01), RAVLT 5 (HR: 1.46, [1.10, 1.94], p = 0.009), ROCFT-R (HR: 1.45, [1.06, 1.97], p = 0.02), TMT A (HR: 1.40, [1.03, 1.90], p = 0.03), and Boston Naming Test (HR: 1.28, [1.01, 1.60], p = 0.04). Worse performance on global cognition (MMSE), Logical Memory (LM-Immediate, LM-Delayed), delayed recall in the RAVLT (RAVLT 30), Digit Span, Digit Symbol Modalities, visuospatial skills (ROCFT-Copy), and all measures of verbal fluency (phonemic, semantic-animals, semantic-vegetables) were not associated with a greater risk of progression.

Table 2

Cox proportional hazard models associating neuropsychological test Z-scores with progression from subjective cognitive decline to mild cognitive impairment or dementia

	Adjusted hazard ratio (Age, Sex, Education)	p
Neuropsychological tests
MMSE^∧	1.05 [0.77, 1.43]	0.75
RAVLT 1	2.33 [1.50, 3.62]	< 0.001^***†
RAVLT 5	1.46 [1.10, 1.94]	0.009^**
RAVLT 1–5	2.22 [1.55, 3.16]	< 0.001^***†
RAVLT 30	1.30 [0.97, 1.74]	0.07
ROCFT-Recall	1.45 [1.06, 1.97]	0.02^*
ROCFT-Copy	1.06 [0.86, 1.29]	0.60
LM-I^∧	1.16 [0.80, 1.68]	0.43
LM-D^∧	1.05 [0.74, 1.49]	0.80
TMT A^∧	1.40 [1.03, 1.90]	0.03^*
TMT B^∧	1.50 [1.10, 2.05]	0.01^**
F-Digit Span-SC^∧	1.08 [0.82, 1.43]	0.58
B-Digit Span-SC^∧	1.05 [0.80, 1.37]	0.75
Digit Symbol Modalities^∧	1.42 [1.00, 2.01]	0.05
BNT 60^∧	1.28 [1.01, 1.60]	0.04^*
P-VF	1.25 [0.90, 1.76]	0.19
S-VF-A^∧	1.13 [0.78, 1.65]	0.52
S-VF-V^∧	1.15 [0.76, 1.74]	0.50

Brackets contain 95% confidence intervals. ^∧indicates that this test is part of the Uniform Data Set (UDS) 2.0. ^***p < 0.001, ^**p < 0.01, ^*p < 0.05. ^†indicates statistical significance after Holm-Bonferroni correction. MMSE, Mini-Mental State Examination, total score; RAVLT 1, Rey Auditory Verbal Learning Test, trial 1 recall; RAVLT 5, Rey Auditory Verbal Learning Test, trial 5 recall; RAVLT 1-5, Rey Auditory Verbal Learning Test, sum of trials 1 to 5; RAVLT 30, Rey Auditory Verbal Learning Test, recall after 30 minutes; ROCFT-Recall, Rey–Osterrieth Complex Figure Test Recall condition –after 3 minutes; ROCFT-Copy, Rey–Osterrieth Complex Figure Test Copy condition; LM-I, Logical Memory Immediate Recall from the Uniform Data Set; LM-D, Logical Memory Delayed Recall from the Uniform Data Set; TMT A, Trail Making Test Part A, given in seconds; TMT B, Trail Making Test Part B, given in seconds; F-Digit Span-SC, Forward Digit Span –score; B-Digit Span-SC, Backward Digit Span –score; Digit Symbol Modalities, Digit Symbol Score from the WAIS-R; BNT, Boston Naming Test 60-item; P-VF, Phonemic Verbal Fluency; S-VF-A, Semantic Verbal Fluency –Animals; S-VF-V, Semantic Verbal Fluency –Vegetables.

Fig. 2

Forest plot of the adjusted hazard ratios for 18 neuropsychological test Z-scores associated with progression to mild cognitive impairment or dementia. MMSE, Mini-Mental State Examination, total score; RAVLT 1, Rey Auditory Verbal Learning Test, trial 1 recall; RAVLT 5, Rey Auditory Verbal Learning Test, trial 5 recall; RAVLT 1–5, Rey Auditory Verbal Learning Test, sum of trials 1 to 5; RAVLT 30, Rey Auditory Verbal Learning Test, recall after 30 minutes; ROCFT-Recall, Rey–Osterrieth Complex Figure Test Recall condition –after 3 minutes; ROCFT-Copy, Rey–Osterrieth Complex Figure Test Copy condition; LM-I, Logical Memory Immediate Recall from the Uniform Data Set; LM-D, Logical Memory Delayed Recall from the Uniform Data Set; TMT A, Trail Making Test Part A, given in seconds; TMT B, Trail Making Test Part B, given in seconds; F-Digit Span-SC, Forward Digit Span –score; B-Digit Span-SC, Backward Digit Span –score; Digit Symbol Modalities, Digit Symbol Score from the WAIS-R; BNT, Boston Naming Test 60-item; P-VF, Phonemic Verbal Fluency; S-VF-A, Semantic Verbal Fluency –Animals; S-VF-V, Semantic Verbal Fluency –Vegetables.

Exploratory tests associated with aMCI or naMCI

Figure 1 shows the cumulative incidence functions, which describe the relative probabilities of progressing from SCD to aMCI or naMCI after controlling for competing risks. See Table 3 for the adjusted ORs and corresponding p-values. Scoring one SD below the mean at baseline on the ROCFT-R (OR: 2.66, [1.39, 5.67], p = 0.006), RAVLT 5 (OR = 2.35, [1.37, 4.33], p = 0.003), RAVLT 1–5 (OR = 2.30, [1.25, 4.58], p = 0.01), and RAVLT 30 (OR = 1.96, [1.14, 3.63], p = 0.02) was positively associated with progressing from SCD to aMCI. Worse performance on the Boston Naming Test (OR = 0.45, [0.22, 0.82], p = 0.02) and TMT B (OR = 0.58, [0.30, 1.06], p = 0.09) were negatively associated with aMCI progression, though the relationship with the TMT B was not statistically significant. The RAVLT 1 (OR = 2.29, [1.28, 4.43], p = 0.008), TMT B (OR = 1.98, [1.17, 3.53], p = 0.01), and Boston Naming Test (OR = 1.93, [1.24, 3.09], p = 0.004) were positively associated with progressing from SCD to naMCI.

Table 3

Logistic regression models associating neuropsychological test Z-scores with progression from subjective cognitive decline to amnestic or non-amnestic MCI

	aMCI progressors Adjusted odds ratios (Age, Sex, Education)	p	naMCI progressors Adjusted odds ratios (Age, Sex, Education)	p
Neuropsychological tests
RAVLT 1	1.51 [0.75, 3.27]	0.27	2.29 [1.28, 4.43]	0.008^**
RAVLT 5	2.35 [1.37, 4.33]	0.003^**^†	0.85 [0.53, 1.32]	0.48
RAVLT 1–5	2.30 [1.25, 4.58]	0.01^*	1.31 [0.81, 2.16]	0.28
RAVLT 30	1.96 [1.14, 3.63]	0.02^*	0.82 [0.51, 1.29]	0.40
ROCFT-Recall	2.66 [1.39, 5.67]	0.006^**^†	1.30 [0.79, 2.13]	0.30
TMT A^∧	0.69 [0.35, 1.24]	0.24	1.26 [0.78, 2.06]	0.34
TMT B^∧	0.58 [0.30, 1.06]	0.09	1.98 [1.17, 3.53]	0.01^*
Digit Symbol Modalities^∧	1.09 [0.57, 2.16]	0.80	1.20 [0.71, 2.05]	0.49
BNT 60^∧	0.45 [0.22, 0.82]	0.02^*	1.93 [1.24, 3.09]	0.004^**^†

Brackets contain 95% confidence intervals. Tests from Cox models with p < 0.10 were included for secondary analysis with logistic regression. Analyses excluded two participants who progressed to dementia. ‘^∧’ indicates that this test is part of the Uniform Data Set (UDS) 2.0. ^***p < 0.001, ^**p < 0.01, ^*p < 0.05. ^†indicates statistical significance after Holm-Bonferroni correction. aMCI, amnestic mild cognitive impairment; naMCI, non-amnestic mild cognitive impairment; RAVLT 1, Rey Auditory Verbal Learning Test, trial 1 recall; RAVLT 5, Rey Auditory Verbal Learning Test, trial 5 recall; RAVLT 1–5, Rey Auditory Verbal Learning Test, sum of trials 1 to 5; RAVLT 30, Rey Auditory Verbal Learning Test, recall after 30 minutes; ROCFT-Recall, Rey–Osterrieth Complex Figure Test Recall condition –after 3 minutes; TMT A, Trail Making Test Part A, given in seconds; TMT B, Trail Making Test Part B, given in seconds; Digit Symbol Modalities, Digit Symbol Score from the WAIS-R; BNT, Boston Naming Test 60-item.

DISCUSSION

Our study found that verbal memory performance at baseline, specifically measured by the RAVLT, was associated with progressing from SCD to MCI or dementia. Particularly, RAVLT 5 (final acquisition trial), RAVLT 1–5 (total acquisition trial), and RAVLT 30 (delayed verbal recall memory) were associated with progression to aMCI whereas the RAVLT 1 (first acquisition trial) was associated with progression to naMCI in exploratory analyses. Additionally, visuospatial memory performance on the ROCFT-Recall was associated with progression to aMCI while poorer executive functioning (TMT B) and language performance (Boston Naming Test) were associated with progression to naMCI in exploratory analyses.

It is well-known that SCD increases the risk of progression to MCI or dementia in comparison to older adults with no cognitive complaints [2]. However, a sizeable proportion of participants first diagnosed with SCD will remain stable or revert to normal functioning without cognitive complaints [8]. In our study, 42% of the participants progressed during the period of observation, and 4 participants reverted back to SCD (7% of the 55 progressors). The long follow-up period, the self-perception of cognitive deficits at baseline, and the site of our study (i.e., memory clinic) [15] likely contributed to the relatively high rate of progression in our study. Although progressors performed worse than those with stable SCD on nearly all tests at baseline in our study (Table 1), few tests were associated with progression to MCI or dementia after controlling for age, sex, and education (Table 2). Given that older age [21, 22] and lower cognitive reserve [23 –25] are associated with progression from SCD to MCI or dementia, these factors likely contributed to the lower baseline cognitive performance in those who progressed from our sample.

Literature suggests that the RAVLT is associated with progression from MCI to AD dementia and may be superior to the ROCFT –Recall (memory trial) [27 –29]. Our study extends these findings to a preclinical stage as the RAVLT was most strongly associated with progression to cognitive impairment out of all the included neuropsychological tests in our sample. This is in line with prior work suggesting that verbal memory may be worse in SCD progressors [21, 26] and in healthy controls who will later progress to SCD [42]. Prior work has also associated the 20 min delayed recall on the Seoul Verbal Learning Test or the 24 h delayed recall on the Five Words Acquisition with being an SCD progressor [21, 26]. Memory tests (e.g., episodic, semantic, verbal, visual) were also among the best predictors in other studies [43 –45]; however, one study included participants with MCI [43] and all examined progression to dementia [43 –45]. Conversely, our findings suggest that only the immediate recall indices (the RAVLT 5 (final acquisition trial), RAVLT 1–5 (total acquisition trial)) were associated with being an SCD progressor. This discrepancy could be explained by the large proportion of participants (53%) progressing to naMCI in our study. When analyzing aMCI and naMCI separately, the 30 min RAVLT (delayed verbal recall memory) was strongly associated with progression from SCD to aMCI but not from SCD to naMCI. These results are in line with a well-established fact that memory is the earliest cognitive hallmark of AD and that its decline precedes the onset of MCI by three to seven years and the onset of dementia by up to 11 years (for review, see [46]). Little work has examined the specific cognitive tests or cognitive domains that associate with progression among participants with SCD [21 , 43], unlike the many studies that document neuropsychological correlates of progression among participants with MCI (see these meta-analyses for relevant studies [29, 47]).

Progression to naMCI was predominantly predicted by non-memory tests. The only score of a memory test which predicted progression to naMCI was RAVLT 1 which is not surprising, as the RAVLT 1 is considered a score of attention and working memory rather than as an indicator of memory functioning [48]. However, other specific tests of attention and working memory (Digit Span, Digit Symbol Modalities, TMT A), were not associated with progression from SCD to naMCI. One hypothesis may be that in the RAVLT 1, participants may become more easily overwhelmed when presented with a large amount of information, which makes it a more challenging task. Future work is needed to further examine the role of immediate verbal recall memory in those who progress from SCD to naMCI and the neuropathological, genetic, or volumetric indices that correlate with this lower performance. Surprisingly, no association was found between LM performance and risk of progression. LM is a widely-used story recall test in the UDS battery and the Preclinical Alzheimer Cognitive Composite [49]. In MCI, LM is considered sensitive to cognitive decline [33, 43]. However, story recall naturally elicits compensatory strategies in comparison to the RAVLT, which uses a list of unrelated words. Utilizing these compensatory strategies may be impaired to a greater extent in those with MCI than in those with SCD, which could explain the null association between LM and progression.

This paper was focused on individual cognitive tests to improve the detectability of progression in patients with SCD. However, the findings should also be interpreted with respect to the tests’ indexed cognitive domains. At baseline, lower verbal memory performance was most strongly associated with progression to aMCI whereas lower executive functioning or language performance was most strongly associated with progression to naMCI. Given that brief cognitive tests have little effectiveness in differentiating SCD from MCI in primary care settings [50], understanding the specific cognitive tests that map onto progression, delineate cognitive phenotypes, or index neurobiological change is increasingly important [20].

Clinical relevance

Our findings support the existence of subtle cognitive abnormalities in patients with SCD who ultimately progress to MCI within 3.54 years on average. The nature of these cognitive abnormalities differs between those who further develop aMCI and naMCI. To the best of our knowledge, this is the first study to examine progression not only to MCI, but also separately to aMCI and naMCI. The RAVLT seems to be more sensitive for aMCI progression in comparison to LM. The RAVLT and LM are among the most commonly used memory tests. In addition, the LM is part of the UDS and therefore widely used by AD centers in the United States.

SCD progressors in our study may be different than “SCD Plus.” SCD Plus can be defined as pre-clinical AD, characterized by subjective decline in memory, age of onset greater than 60 years, cognitive complaints, and biomarker evidence of AD where available (e.g., APOE ɛ4 allele genotype, Aβ-positive scans or cerebrospinal fluid Aβ₄₂, total tau, and phosphorylated tau) [1]. Because a large proportion of progressors were diagnosed with naMCI in our study, it begs questioning whether a non-AD SCD Plus category should be developed. Although first, work is needed to better understand the biological differences between typical SCD Plus patients and SCD progressors who may not move along the AD continuum.

Our findings may be used by clinicians that rely on the National Institute on Aging and Alzheimer’s Association (NIA-AA) criteria for preclinical AD to identify patients at risk of progressing from Stage 2 (e.g., normal cognitive performance with signs of decline from previous functioning) to Stage 3 (e.g., impaired cognitive performance but maintained functional status) [51]. Before this study can be readily used by clinicians who wish to predict progression in their memory clinics with greater accuracy, findings on the importance of immediate and delayed verbal memory should be validated in a larger, more diverse, sample.

Strengths and limitations

A major strength of this study is that it included a unique sample of participants with SCD and a comprehensive cognitive battery encompassing all major cognitive domains. Our study was conducted in the Czech Republic, which may have affected sample generalizability and recruitment. Several factors discern the Czech Republic from the United States, such as the availability of universal healthcare access (i.e., reduced barriers in healthcare access), ethnic and racial homogeneity of our sample (100% White and of European ancestry), lack of language barriers (i.e., all participants were native Czech speakers and completed the neuropsychological battery in Czech), and most participants completed 12 or more years of education (94%).

Our study was limited by the small sample of SCD participants in CBAS with at least one follow-up visit, and by the large number of tests conducted. Due to the small sample size, Fine-Gray competing risks models were not reported in the main text; however, sensitivity analyses revealed similar findings to the logistic models. It is of note that we found at least some differences in most tests, despite non-significant p-values, which may be attributed to a small sample size and the associated bias towards Type II error. Therefore, this study design deserves replication in a larger, well-powered sample. In addition, magnitude of the associations, rather than p-values should be prioritized when interpreting results.

A large number of participants (n = 205) have not yet received a follow-up assessment, which may be due to participant disinterest in reassessment. That is, participants receive feedback after baseline assessment in CBAS, which may have impacted their decision to undergo reassessment and inflate our progression rates (i.e., those with the most concerning and unremitting complaints may be more likely to seek cognitive reassessment). The site of our study (i.e., memory clinic) likely affected the rate of progression in our sample. It may be that the participants with SCD who progressed to aMCI would further progress to dementia within the next few years, which should be an area of future inquiry.

Another limitation is the reversion from MCI back to SCD. Participants were censored after a change in diagnosis was noted, though only 4 participants of 55 progressors ultimately reverted. Future works should examine the cognitive tests that best associate with reversion back to SCD or normal cognitive functioning from MCI. Comparing aMCI and naMCI progressors is a strength of our study but does introduce some degree of circularity given that these diagnostic categories are defined by test-specific patterns of cognitive impairment. Our study design was limited by this circularity in comparison to other studies that examined dementia and included functional impairment (e.g., impairment in activities of daily living) to identify progression. However, because baseline performance is measured and not the impairment at follow-up, our results do suggest a predictive ability of specific tests within each domain (i.e., Boston Naming Test was a significant predictive test, but phonemic or semantic fluency were not). Excluding normal controls without SCD may be viewed as a limitation, though normal controls have a lower risk of progression than those with SCD [2], and our goal was to consider participants with SCD specifically. Our comparison group (i.e., progressors versus stable SCD) allows for an accurate estimation of the effect size in our clinical sample. Additionally, we did not include biomarker data that would provide needed context to the lower baseline cognitive performance and incident progression among a large proportion of participants with SCD. Finally, we could not separate dementia from MCI because of a small number of those who converted to dementia.

Conclusion

In our study, 42% of participants with SCD progressed to aMCI, naMCI, or dementia after 3.54 years on average. Of the 18 neuropsychological tests or subtests examined, verbal memory performance from the RAVLT was most associated with progression to clinical impairment for those with SCD. At baseline, lower verbal memory performance was most associated with progression to aMCI whereas executive dysfunction or language performance was most associated with progression to naMCI.

Footnotes

ACKNOWLEDGMENTS

This work was supported by the Institutional Support of Excellence 2. LF UK Grant No. 6990332, MH CZ –DRO, the European Regional Development Funds No. CZ.02.1.01/0.0/0.0/16_019/0000868 ENOCH grant, and the Czech Ministry of Health grant NV 18-04-00455. This work was also supported in part by funds from the University of South Florida Nexus Initiative (UNI) Award.

Authors’ disclosures available online ().

References

Jessen

, Amariglio

, Van Boxtel

, Breteler

, Ceccaldi

, Chételat

, Dubois

, Dufouil

, Ellis

, Van Der Flier

(2014) A conceptual framework for research on subjective cognitivedecline in preclinical Alzheimer’s disease. Alzheimers Dement 10, 844–852.

Mitchell

, Beaumont

, Ferguson

, Yadegarfar

, Stubbs

(2014) Risk of dementia and mild cognitive impairment in older people with subjective memory complaints: Meta-analysis. Acta Psychiatr Scandinav 130, 439–451.

Amariglio

, Becker

, Carmasin

, Wadsworth

, Lorius

, Sullivan

, Maye

, Gidicsin

, Pepin

, Sperling

(2012) Subjective cognitive complaints and amyloid burden in cognitively normal older individuals. Neuropsychologia 50, 2880–2886.

Selnes

, Aarsland

, Bjørnerud

, Gjerstad

, Wallin

, Hessen

, Reinvang

, Grambaite

, Auning

, Kjærvik

(2013) Diffusion tensor imaging surpasses cerebrospinal fluid as predictor of cognitive decline and medial temporal lobe atrophy in subjective cognitive impairment and mild cognitive impairment. J Alzheimers Dis 33, 723–736.

Dufouil

, Fuhrer

, Alpérovitch

(2005) Subjective cognitivecomplaints and cognitive decline: Consequence or predictor? Theepidemiology of vascular aging study.. J Am Geriatr Soc 53, 616–621.

Hays

, Zlatar

, Campbell

, Meloy

, Wierenga

(2018) Subjective cognitive decline modifies the relationship between cerebral blood flow and memory function in cognitively normal older adults. J Int Neuropsychol Soc 24, 213–223.

Balash

, Mordechovich

, Shabtai

, Giladi

, Gurevich

, Korczyn

(2013) Subjective memory complaints in elders: Depression, anxiety, or cognitive decline? Acta Neurol Scand 127, 344–350.

Jessen

, Amariglio

, Buckley

, van der Flier

, Han

, Molinuevo

, Rabin

, Rentz

, Rodriguez-Gomez

, Saykin

(2020) The characterisation of subjective cognitive decline. Lancet Neurol 19, 271–278.

Aerts

, Heffernan

, Kochan

, Crawford

, Draper

, Trollor

, Sachdev

, Brodaty

(2017) Effects of MCI subtype and reversion on progression to dementia in a community sample. Neurology 88, 2225–2232.

10.

Oltra-Cucarella

, Ferrer-Cascales

, Alegret

, Gasparini

, Díaz-Ortiz

, Ríos

, Martínez-Nogueras

ÁL

, Onandia

, Pérez-Vicente

, Cabello-Rodríguez

(2018) Risk ofprogression to Alzheimer’s disease for different neuropsychologicalMild Cognitive Impairment subtypes: A hierarchical meta-analysis oflongitudinal studies. Psychol Aging 33, 1007.

11.

Yaffe

, Petersen

, Lindquist

, Kramer

, Miller

(2006) Subtype of mild cognitive impairment and progression to dementia and death. Dement Geriatr Cogn Disord 22, 312–319.

12.

Edwards

, Spira

, Barnes

, Yaffe

(2009) Neuropsychiatric symptoms in mild cognitive impairment: Differences by subtype and progression to dementia. Int J Geriatr Psychiatry 24, 716–722.

13.

Rozzini

, Chilovi

, Conti

, Delrio

, Borroni

, Trabucchi

, Padovani

(2008) Neuropsychiatric symptoms in amnestic and nonamnestic mild cognitive impairment. Dement Geriatr Cogn Disord 25, 32–36.

14.

Desai

, Whitfield

, Said

, John

, Saunders

, Marchant

, Stott

, Charlesworth

(2021) Affective symptoms and risk of progression to mild cognitive impairment or dementia in subjective cognitive decline: A systematic review and meta-analysis. Ageing Res Rev 71, 101419.

15.

Snitz

, Wang

, Cloonan

, Jacobsen

, Chang

C-CH

, Hughes

, Kamboh

, Ganguli

(2018) Risk of progression from subjective cognitive decline to mild cognitive impairment: The role of study setting. Alzheimers Dement 14, 734–742.

16.

Wolfsgruber

, Polcher

, Koppara

, Kleineidam

, Froelich

, Peters

, Huell

, Ruether

, Wiltfang

, Maier

(2017) Cerebrospinal fluid biomarkers and clinical progression in patients with subjective cognitive decline and mild cognitive impairment. J Alzheimers Dis 58, 939–950.

17.

Ebenau

, Timmers

, Wesselman

, Verberk

, Verfaillie

, Slot

, Van Harten

, Teunissen

, Barkhof

, Van Den Bosch

(2020) ATN classification and clinical progression in subjective cognitive decline: The SCIENCe project. Neurology 95, e46–e58.

18.

Scarth

, Rissanen

, Scholten

, Geerlings

(2021) Biomarkers of Alzheimer’s disease and cerebrovascular lesions and clinical progression in patients with subjective cognitive decline: A systematic review. J Alzheimers Dis 83, 1089–1111.

19.

Benedictus

, van Harten

, Leeuwis

, Koene

, Scheltens

, Barkhof

, Prins

, van der Flier

(2015) White matter hyperintensities relate to clinical progression in subjective cognitive decline. Stroke 46, 2661–2664.

20.

Jester

, Andel

, Cechová

, Laczó

, Lerch

, Marková

, Nikolai

, Vyhnálek

, Hort

(2021) Cognitive phenotypes ofolder adults with subjective cognitive decline and amnestic mildcognitive impairment: The Czech Brain Aging Study. J IntNeuropsychol Soc 27, 329–342.

21.

Hong

, Yoon

, Shim

, Kim

S-O

, Kim

, Choi

, Jeong

, Yoon

, Yang

, Lee

J-H

(2015) Predictors of clinical progression of subjective memory impairment in elderly subjects: Data from the Clinical Research Centers for Dementia of South Korea (CREDOS). Dement Geriatr Cogn Disord 40, 158–165.

22.

Mazzeo

, Padiglioni

, Bagnoli

, Carraro

, Piaceri

, Bracco

, Nacmias

, Sorbi

, Bessi

(2020) Assessing the effectiveness of subjective cognitive decline plus criteria in predicting the progression to Alzheimer’s disease: An 11-year follow-up study. Eur J Neurol 27, 894–899.

23.

Jia

, Li

, Cao

(2021) Subjective cognitive decline, cognitive reserve indicators, and the incidence of dementia. J Am Med Direct Assoc 22, 1449–1455.e1444.

24.

Mazzeo

, Padiglioni

, Bagnoli

, Bracco

, Nacmias

, Sorbi

, Bessi

(2019) The dual role of cognitive reserve in subjective cognitive decline and mild cognitive impairment: A 7-year follow-up study. J Neurology 266, 487–497.

25.

Yang

, Chen

, Sheng

, Xie

, Li

, Hu

, Sun

, Han

(2020) Cognitive reserve, brain reserve, APOE ɛ4, and cognitionin individuals with subjective cognitive decline in the SILCODEstudy. J Alzheimers Dis 76, 249–260.

26.

Bessi

, Mazzeo

, Padiglioni

, Piccini

, Nacmias

, Sorbi

, Bracco

(2018) From subjective cognitive decline to Alzheimer’s disease: The predictive role of neuropsychological assessment, personality traits, and cognitive reserve. A 7-year follow-up study. J Alzheimers Dis 63, 1523–1535.

27.

, Chen

, Zhao

, Li

, Guo

(2020) Short-term delayed recall of auditory verbal learning test provides equivalent value to long-term delayed recall in predicting MCI clinical outcomes: A longitudinal follow-up study. Applied Neuropsychol Adult 27, 73–81.

28.

Zhao

, Guo

, Liang

, Chen

, Zhou

, Ding

, Hong

(2015) Auditory verbal learning test is superior to rey-osterrieth complex figure memory for predicting mild cognitive impairment to Alzheimer’s disease. Curr Alzheimer Res 12, 520–526.

29.

Belleville

, Fouquet

, Hudon

, Zomahoun

HTV

, Croteau

(2017) Neuropsychological measures that predict progression from mild cognitive impairment to Alzheimer’s type dementia in older adults: A systematic review and meta-analysis. Neuropsychol Rev 27, 328–353.

30.

Sheardova

, Vyhnalek

, Nedelska

, Laczo

, Andel

, Marciniak

, Cerman

, Lerch

, Hort

(2019) Czech Brain Aging Study (CBAS): Prospective multicentre cohort study on risk and protective factors for dementia in the Czech Republic. BMJ Open 9, e030379.

31.

Hughes

, Berg

, Danziger

, Coben

, Martin

(1982) A new clinical scale for the staging of dementia. Br J Psychiatry 140, 566–572.

32.

Petersen

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

33.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

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

34.

Koepsell

, Monsell

(2012) Reversion from mild cognitive impairment to normal or near-normal cognition: Risk factors and prognosis. Neurology 79, 1591–1598.

35.

Drozdová

, Štěpánková

, Lukavský

, Bezdíček

, Kopeček

(2015) Normativní studie testuReyovy-Osterriethovy komplexní figury v populaci českýchseniorringu. Čáá. Neurol Neurochir 111, 542–549.

36.

Nikolai

, Stepankova

, Kopecek

, Sulc

, Vyhnalek

, Bezdicek

(2018) The uniform data set, Czech version: Normative data in older adults from an international perspective. J Alzheimers Dis 61, 1233–1240.

37.

Nikolai

, Štěpánková

, Michalec

, Bezdíček

, Horáková

, Marková

, Kopeček

(2015) Testy verbální fluence, česká normativnístudie pro osoby vyššího věku. Čáá. Neurol Neurochir 111, 292–299.

38.

Bezdicek

, Stepankova

, Moták

, Axelrod

, Woodard

, Preiss

, Nikolai

, Růžička

, Poreh

(2014) Czech version of Rey Auditory Verbal Learning test: Normative data. Aging Neuropsychol Cogn 21, 693–721.

39.

Rey

(1958) L’examen clinique en psychologie.

40.

Sheikh

, Yesavage

(1986) Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. Clin Gerontol 5, 165–173.

41.

Beck

, Epstein

, Brown

, Steer

(1988) An inventory for measuring clinical anxiety: Psychometric properties. J Consult Clin Psychol 56, 893.

42.

Koppara

, Wagner

, Lange

, Ernst

, Wiese

, König

, Brettschneider

, Riedel-Heller

, Luppa

, Weyerer

(2015) Cognitive performance before and after the onset of subjective cognitive decline in old age. Alzheimers Dement (Amst) 1, 194–205.

43.

Silva

, Guerreiro

, Maroco

, Santana

, Rodrigues

, Marques

, de Mendonça

(2012) Comparison of four verbal memory tests forthe diagnosis and predictive value of mild cognitive impairment. Dement Geriatr Cogn Disord Extra 2, 120–131.

44.

Guarch

, Marcos

, Salamero

, Gastó

, Blesa

(2008) Mild cognitive impairment: A risk indicator of later dementia, or a preclinical phase of the disease? Int J Geriatr Psychiatry 23, 257–265.

45.

Blackwell

, Sahakian

, Vesey

, Semple

, Robbins

, Hodges

(2004) Detecting dementia: Novel neuropsychological markers of preclinical Alzheimer’s disease. Dement Geriatr Cogn Disord 17, 42–48.

46.

Karr

, Graham

, Hofer

, Muniz-Terrera

(2018) When does cognitive decline begin? A systematic review of change point studies on accelerated decline in cognitive and neurological outcomes preceding mild cognitive impairment, dementia, and death. Psychol Aging 33, 195.

47.

Prado

, Watt

, Treeby

, Crowe

(2019) Performance on neuropsychological assessment and progression to dementia: A meta-analysis. Psychol Aging 34, 954.

48.

Lezak

, Howieson

, Bigler

, Tranel

(2012) Neuropsychological assessment. Oxford University Press, USA.

49.

Donohue

, Sperling

, Salmon

, Rentz

, Raman

, Thomas

, Weiner

, Aisen

(2014) The preclinical Alzheimer cognitive composite: Measuring amyloid-related decline. JAMA Neurol 71, 961–970.

50.

Petrazzuoli

, Vestberg

, Midlöv

, Thulesius

, Stomrud

, Palmqvist

(2020) Brief cognitive tests used in primary care cannot accurately differentiate mild cognitive impairment from subjective cognitive decline. J Alzheimers Dis 75, 1191–1201.

51.

Jack

Jr , Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

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