Frailty,Cognitive Impairment,and Incident Major Neurocognitive Disorders: Results of the NuAge Cohort Study

Abstract

Background:

Frailty is associated with an increased risk of major neurocognitive disorders (MNCD).

Objective:

This study aims to compare the Fried physical model and the CARE deficit accumulation model for their association with incident major neurocognitive disorders (MNCD), and to examine how the addition of cognitive impairment to these frailty models impacts the incidence in community-dwelling older adults.

Methods:

A subset of community dwellers (n = 1,259) who participated in the “Quebec Longitudinal Study on Nutrition and Successful Aging” (NuAge) were selected in this Elderly population-based observational cohort study with 3 years of follow-up. Fried and CARE frailty stratifications into robust, pre-frail and frail groups were performed using the NuAge baseline assessment. Incident MNCD (i.e., Modified Mini Mental State (3MS) score < 79/100 and Instrumental Activity Daily Living (IADL) score < 6/8) were collected each year over a 3-year follow-up period.

Results:

A greater association with incident MNCD of the CARE frail state was observed with an increased predictive value when combined with cognitive impairment in comparison to Fried’s one, the highest incidences being observed using the robust state as the reference. Results with the Fried frail state were more heterogenous, with no association with the frail state alone, whereas cognitive impairment alone showed the highest significant incidence.

Conclusion:

The association of the CARE frail state with cognitive impairment increased the predictive value of MNCD, suggesting that the CARE frailty model may be of clinical interest when screening MCND in the elderly population.

Keywords

Aging Alzheimer’s disease cognitive impairment cohort study community-dwellers frailty

1 INTRODUCTION

Major neurocognitive disorders (MNCD) are highly present in older populations of high-income countries [1, 2]. For example, in Canada, their overall prevalence ranges from 1% for those aged 65-69 to 25% for those 85 and older [3]. MNCD have negative physical, psychological, social, and economic impacts, not only for people living with them, but also for their caregivers and society at large [1, 2]. There are still no effective treatments to delay, halt or reverse the development of MNCD in the brain, but primary prevention is a promising approach to delay their onset [4]. An operative primary prevention strategy is based on two successive steps: first, screening individuals at risk of MNCD and, second, changing their lifestyle activities [4–6]. Numerous risk factors of MNCD are modifiable, such as cardiovascular risk factors, which depend on lifestyle choices related to physical activity and/or dietary measures [5, 6]. Many of the chronic morbidities that accumulate with age are also modifiable risk factors of MNCD [5–7]. Their co-existence may result in interactions that may increase negative impacts on the brain and accelerate the onset of MNCD [6, 7]. Thus, effective primary prevention of MNCD must not only consider risk factors alone, but also their cumulative effects, as well as the latter’s potential adverse effects on the incidence of MNCD.

A significant adverse consequence of chronic morbidity accumulation is frailty [8, 9]. Frailty is associated with an increased risk of MNCD and, therefore, recognized as a key prevention target [8–11]. The association between frailty and the occurrence of MNCD has been confirmed in a recent review and meta-analysis [12]. However, this last study underscored uncertainty about this association. Firstly, not all epidemiological studies reported a positive association. Secondly, most studies used Fried’s criteria, which defines frailty as a strictly physical impairment, whereas psychological, cognitive, and social domains also contribute to frailty and should be considered in its definition [8–10]. Rockwood’s deficit accumulation model, on the other hand, considers this multidimensional aspect of frailty by counting deficits in various domains [11]. Up to now, only one study which used this model to define frailty showed a positive association with the incidence of MNCD [13]. Thirdly, no study has yet compared the Fried and Rockwood frailty models for their association with incident MNCD. Therefore, the best frailty model to use for the prevention of MNCD remains to be determined. Finally, although cognitive impairment is a strong risk factor for MNCD, its place in frailty models is typically absent or limited [8–11]. Indeed, Fried’s frailty model does not include it, while Rockwood’s frailty model includes ‘poor cognition’ a possible deficit [11, 14, 15]. To counter these issues, the concept of ‘cognitive frailty’ has been created [16, 17]. It is defined as the simultaneous existence of both physical frailty and cognitive impairment in older adults without MNCD. To the best of our knowledge, no study to date has examined the respective and combined effects of both components of cognitive frailty on the occurrence of MNCD.

Primary prevention requires individuals’ active participation and patient empowerment [18, 19]. The World Health Organization recommends the use of self-administered questionnaires to encourage people’s agency in improving their own health [20]. Among the interventions dedicated to promote a healthy lifestyle, those incorporating electronic health (e-health) tools receive increasing attention for their potential [21, 22]. CARE is a health application (app) developed over the COVID-19 period in Quebec (Canada) in 2021. CARE is composed of two parts: first, a frailty assessment performed via a series of self-administered questionnaires and second, customized recommendations that promote preventive health measures through lifestyle changes. The frailty assessment portion of the CARE app is based on Rockwood’s deficit accumulation model and like Fried’s frailty model, stratifies frailty into three states (i.e., robust, pre-frail and frail) [11].

The NuAge database merges all information collected in the “Quebec Longitudinal Study on Nutrition and Successful Aging” (NuAge), which is a Quebec population-based observational cohort study over a 3-year period [23]. All items of CARE and Fried frailty assessments were collected at the baseline assessment of the NuAge study. Information on incident MNCD was also collected over the follow-up period. Numerous research questions are still pending about the association of different frailty models and incident MNCD. For example, is CARE’s deficit accumulation model of frailty effectively associated with incident MNCD? Between the Fried and CARE models, which model best predicts MNCD? How does the combination of cognitive impairment with frailty influence the incidence of MNCD? We hypothesized that both CARE and Fried’s frail states could be associated with the occurrence of MNCD in the NuAge study cohort and that the combination of the frail state with cognitive impairment could increase the strength of this association, especially with the CARE model, given its multidomain definition of frailty. Consequently, the objective of the present study was to examine and compare the CARE and Fried frail states for their association with incident MNCD and the effect of their combination with cognitive impairment on this incidence.

2 METHODS

2.1 Population

The NuAge Database was used for the present study. The collection of NuAge data was reported in detail previously [23]. In summary, the NuAge study recruited community-dwelling men and women aged 67-84 years without cognitive impairment (i.e., Modified Mini Mental State (3MS) score > 79/100) and major physical disabilities (i.e., able to walk 300 meters and climb 10 stairs without rest), who agreed to an annual follow-up over a 5-year period [24]. 1,793 participants were recruited in the study after providing their written agreement between November 2003 and June 2005 in three Quebec cities (Montreal, Laval, and Sherbrooke) and 1,753 (97.8%) agreed to the integration of their data into the NuAge Database and Biobank. Among this subset, 1,526 (85.1% of full set of participants) were followed during a 3-year period. We excluded, in this second subset, participants with missing values for any frail item at baseline and information about MNCD over the 3-year follow-up, leaving the data of 1,259 (70.2% of full set of participants) participants available for use in the present study.

2.2 Study baseline assessment

Participants had a full, standardized face-to-face examination at baseline. For this study, collected information was binarized and recoded using the convention that “0” indicated the absence of a deficit and “1” the presence of a deficit. Age (year), sex (female versus male), measured body weight (kg) and height (cm), reported weight loss (in kg) in the prior 6 months, vitamin D supplementation and the number of prescribed medications taken daily extracted from the list of medications shown by the participants, ability to stand-up five times from a chair without rest, using a walking aid regardless of its type, and the inability to walk outside alone around one’s dwelling were recorded, both extracted from the Functional Autonomy Measurement System (SMAF). Age≥80, male and polypharmacy (i.e., number of medications taken daily≥5) were coded as 1. The body mass index (BMI; kg/m²) was calculated and abnormal BMI combining values below 18.5 kg/m² and above 24.9 kg/m² were coded as 1. Reported weight loss in the previous 6 months was calculated ((weight loss (kg)/ actual weight measured (kg)) x 100) and was coded 1 when it was > 5% of body weight. The intention to lose weight was not available at baseline and could not be considered. Memory impairment, activities of daily living (ADLs), and instrumental activities of daily living (IADLs) abilities were all collected using the SMAF [25]. SMAF assesses functionality in 5 domains including mobility, communication, mental functions, ADLs, and IADLs. Disability was scored on a 5-point scale for each item: 0 (independent), 0.5 (with difficulty), 1 (needs supervision), 2 (needs help), 3 (dependent). All SMAF items used for the CARE scale were recorded as follows: 0 and 0.5 = 0 and 1 to 3 = 1. Feelings of emptiness and low energy were collected using the following 2 questions from the 30-item Geriatric Depression Scale, and each variable was coded as binary variables (yes = 1 and no = 0): “Do you feel that your life is empty?” and “Do you feel full of energy?” [26]. The caregiver state, defined as an individual who regularly cares for another who needs help taking care of his or herself, was derived from the Physical Activity Scale for the Elderly (PASE) questionnaire and coded as 1 if providing care [27]. Social isolation was defined as individuals who lived alone, had no help at home and no contact with another person over the past week, and coded 1. Slow gait speed was defined using the best of two walking trials, measured at a usual pace over a 4-meter distance. Time was measured in seconds and stratified according to sex and then according to height (2 strata per sex). In each stratum, participants in the slowest quintile were considered to have slow gait speed and coded as 1 (men: < 0.98 m/s for height < 168 cm, <1.03 m/s for height≥168 cm; women: <0.91 m/s for height < 155 cm,<0.95 m/s for height≥155 cm). Low muscle strength was determined using a Martin Vigorimeter tested 3 times for each hand. We used the best value of the 3 tests for each hand to calculate the mean hand grip strength. Participants in the lowest quintile by sex-BMI strata (using median BMI) were coded as 1 (men <59.0 kPa for BMI <25.3 kg/m²_, women <45.0 kPa for BMI <24.2 kg/m²). Physical activity levels were assessed using the PASE [27]. A low physical activity level was defined as below the lowest tertile per sex strata (i.e., <69.1 for women and <87.7 for men) and coded 1.

2.3 Frailty scales and indexes

The CARE frailty assessment is composed of 21 items coded as binary variables with 1 indicating a deficit and 0 indicating an absence of deficit [28]. Its score ranges from 0 (no deficits) to 21 (all deficits present). The CARE frailty index (FI) is constructed from these 21 items and is calculated by dividing the number of recorded deficits by the total number of items [11]. The cut-off points for the CARE FI follow the recommended procedure for categorizing a value≥0.24 as frail and a value below 0.05 as robust [11]. An FI score ranging between 0.06 to 0.23 was categorized as pre-frail. Thus, the participants were classified as followed based on their score: robust (0-1), pre-frail (2-4) and frail (≥5).

The Fried frailty assessment consists of five items including unintentional weight loss, exhaustion, low physical activity, weakness, and slowness as defined above [14]. Each Fried item is coded as a binary variable like the CARE frailty scale (i.e., 0 = absence of deficit and 1 = presence of deficit) and any reported weight lost was considered in the actual study (intention not available). The Fried score ranges from 0 (i.e., no item abnormal) to 5 (i.e., all items abnormal). Participants were classified as follows: 0 = robust, 1-2 = prefrail and 3-5 = frail.

2.4 Outcomes

A standardized assessment using the same questions and tests performed at baseline (T1) was repeated each year of the follow-up period (T2, T3 and T4). Cognitive impairment at baseline was defined with a score of <4/6 on the immediate and delayed recalls of the three words (S-MMSE; validated short version) extracted from the Mini-Mental State Examination (MMSE) originally performed in NuAge [24, 29]. Major neurocognitive disorders were defined as the combination of abnormal 3MS scores (i.e., ≤79/100) and IADL scores (i.e., ≤6/8) [24].

2.5 Standard protocol approvals, registrations, and patient consents

Written informed consent for research was obtained for each NuAge participant. The Research Ethics Boards (REB) of the University Institute of Geriatrics of Sherbrooke and the “Institut Universitaire de gériatrie de Montréal” approved the NuAge study. The REB of the CIUSSS-de-l’Estrie-CHUS approved the NuAge Database and Biobank. The REB of the Jewish General Hospital of Montreal (Quebec, Canada) approved the present study.

2.6 Statistics

The baseline characteristics of participants were described using means, standard deviations (SD), and frequencies (percentages). Participants were ascribed to one of the three groups based on their frail state (i.e., robust, pre-frail and frail) using CARE and Fried stratification. Comparisons between groups were performed using analysis of variance (ANOVA) with Bonferroni correction and the Chi-square test, as appropriate. Using different referent levels (robust and pre-frail individuals pooled together, robust used alone, pre-frail used alone, respectively), Cox models were used to examine the association of the CARE and Fried frail states as dependent variables (separate models for each frail state), combined or not with cognitive impairment, with the overall incidence of MNCD. p values less than 0.05 were considered statistically significant. All statistical analyses were performed using SPSS (version 26.0; SPSS, Inc., Chicago, IL).

3 RESULTS

The baseline characteristics of the participants, according to their frailty states, are shown in Table 1. Groups differed significantly for all characteristics, except for housebound and social isolation, using the CARE frailty stratification (Tables 1 and 2). Participants were older in the frail group compared to robust and pre-frail groups (p≤0.001) but no significant difference was shown between robust and pre-frail groups. Male were less prevalent in the robust group compared to pre-frail and frail groups (p≤0.001). There was less significant difference between groups with the Fried stratification compared to the CARE frailty stratification. Older participants were present in frail and pre-frail groups compared to the robust group (p≤0.001), and in the frail group compared to the pre-frail group (p≤0.001). No significant difference was shown for sex.

Table 1
Baseline characteristics of the NuAge participants separated by their frailty states using CARE and Fried frailty stratifications (n = 1,259)

Characteristics CARE frailty states^μμ Overall Fried frailty states^*** Overall

Robust Pre-Frail Frail p Robust Pre-Frail Frail p

(n = 125) (n = 854) (n = 280) (n = 562) (n = 610) (n = 87)

CARE items:

Age

Mean±SD (y) 73.0±3.9 73.8±4.0 75.7±4.3 ≤ 0.001 73.2±3.9 74.6±4.2 76.5±3.9 ≤ 0.001

>80, n (%) 6 (4.8) 84 (9.8) 72 (25.7) ≤ 0.001 39 (6.9) 101 (16.6) 22 (25.3) ≤ 0.001

Male, n (%) 34 (27.2) 439 (51.4) 127 (45.4) ≤ 0.001 253 (45.0) 305 (50.0) 42 (48.3) 0.232

Body mass index abnormal^*, n (%) 26 (20.8) 648 (75.9) 251 (89.6) ≤ 0.001 377 (67.1) 473 (77.5) 75 (86.2) ≤ 0.001

No vitamin D supplementation, n (%) 12 (9.6) 168 (19.7) 130 (46.4) ≤ 0.001 138 (24.6) 150 (24.6) 22 (25.3) 0.989

Polypharmacy^†, n (%) 4 (3.2) 364 (42.6) 221 (78.9) ≤ 0.001 223 (39.7) 306 (50.2) 60 (69.0) ≤ 0.001

Unable to stand up from chair five-time, n (%) – 7 (0.8) 12 (4.3) ≤ 0.001 5 (0.9) 11 (1.8) 3 (3.4) 0.135

Use of walking aid^||, n (%) – 15 (1.8) 37 (13.2) ≤ 0.001 4 (0.7) 33 (5.4) 15 (17.2) ≤ 0.001

Housebound^‡, n (%) – 1 (0.1) – 0.789 – 1 (0.2) - 0.587

Memory impairment, n (%) 1 (0.8) 120 (14.1) 123 (43.9) ≤ 0.001 89 (15.8) 131 (21.5) 24 (27.6) 0.007

Feeling of empty life^¶, n (%) 4 (3.2) 123 (14.4) 119 (42.5) ≤ 0.001 93 (16.5) 134 (22.0) 19 (21.8) 0.056

Caregiver^#, n (%) 5 (4.0) 162 (19.0) 91 (32.5) ≤ 0.001 143 (25.4) 106 (17.4) 9 (10.3) ≤ 0.001

Social isolation^α, n (%) 1 (0.8) 16 (1.9) 8 (2.9) 0.359 7 (1.2) 18 (3.0) - 0.044

Need help for, n (%):

Phone use^∞ – 17 (2.0) 27 (9.6) ≤ 0.001 11 (2.0) 27 (4.4) 6 (6.9) 0.014

Transportation^μ – – 13 (4.6) ≤ 0.001 4 (0.7) 8 (1.3) 1 (1.1) 0.594

Medication intake^** – 14 (1.6) 16 (5.7) ≤ 0.001 7 (1.2) 18 (3.0) 5 (5.7) 0.017

Handle finances^†† – 64 (7.5) 78 (27.9) ≤ 0.001 49 (8.7) 80 (13.1) 13 (14.9) 0.032

Bathing 1 (0.8) 4 (0.5) 14 (5.0) ≤ 0.001 3 (0.5) 12 (2.0) 4 (4.6) 0.007

Toileting – 71 (8.3) 102 (36.4) ≤ 0.001 44 (7.8) 103 (16.9) 26 (29.9) ≤ 0.001

Dressing – 2 (0.2) 9 (3.2) ≤ 0.001 1 (0.2) 8 (1.3) 2 (2.3) 0.038

Feeding – 2 (0.2) 4 (1.4) 0.030 1 (0.2) 5 (0.8) - 0.225

Incontinence^||||, n (%) 11 (8.8) 209 (24.5) 151 (53.9) ≤ 0.001 146 (26.0) 191 (31.3) 34 (39.1) 0.017

Fried items:

Weight loss^¶¶, n (%) 1 (0.8) 105 (12.3) 36 (12.9) ≤ 0.001 – 116 (19.0) 26 (29.9) ≤ 0.001

Poor energy^# #, n (%) 10 (8.0) 142 (16.6) 82 (29.3) ≤ 0.001 – 172 (28.2) 62 (71.3) ≤ 0.001

Slowness^αα, n (%) 7 (5.6) 139 (16.3) 84 (30.0) ≤ 0.001 – 161 (26.4) 69 (79.3) ≤ 0.001

Low muscle strength, n (%) 16 (12.8) 141 (16.5) 78 (27.9) ≤ 0.001 – 178 (29.2) 57 (65.5) ≤ 0.001

Low physical activity^∞∞, n (%) 19 (15.2) 143 (16.7) 80 (28.6) ≦ 0.001 – 176 (28.9) 66 (75.9) ≤ 0.001

Cognitive impairment: S-MMSE

Mean±SD (score) 5.4±0.8 5.2±1.1 4.8±1.2 ≦ 0.001 5.1±1.1 5.1±1.1 5.1±0.9 0.934

≤4/6 80, n (%) 15 (11.7) 179 (20.9) 91 (32.4) ≦ 0.001 116 (20.6) 147 (24.1) 20 (23.0) 0.364

Characteristics	CARE frailty states^μμ	Overall	Fried frailty states^***	Overall
CARE items:
Age
Mean±SD (y)	73.0±3.9	73.8±4.0	75.7±4.3	≤ 0.001	73.2±3.9	74.6±4.2	76.5±3.9	≤ 0.001
>80, n (%)	6 (4.8)	84 (9.8)	72 (25.7)	≤ 0.001	39 (6.9)	101 (16.6)	22 (25.3)	≤ 0.001
Male, n (%)	34 (27.2)	439 (51.4)	127 (45.4)	≤ 0.001	253 (45.0)	305 (50.0)	42 (48.3)	0.232
Body mass index abnormal^*, n (%)	26 (20.8)	648 (75.9)	251 (89.6)	≤ 0.001	377 (67.1)	473 (77.5)	75 (86.2)	≤ 0.001
No vitamin D supplementation, n (%)	12 (9.6)	168 (19.7)	130 (46.4)	≤ 0.001	138 (24.6)	150 (24.6)	22 (25.3)	0.989
Polypharmacy^†, n (%)	4 (3.2)	364 (42.6)	221 (78.9)	≤ 0.001	223 (39.7)	306 (50.2)	60 (69.0)	≤ 0.001
Unable to stand up from chair five-time, n (%)	–	7 (0.8)	12 (4.3)	≤ 0.001	5 (0.9)	11 (1.8)	3 (3.4)	0.135
Use of walking aid^\|\|, n (%)	–	15 (1.8)	37 (13.2)	≤ 0.001	4 (0.7)	33 (5.4)	15 (17.2)	≤ 0.001
Housebound^‡, n (%)	–	1 (0.1)	–	0.789	–	1 (0.2)	-	0.587
Memory impairment, n (%)	1 (0.8)	120 (14.1)	123 (43.9)	≤ 0.001	89 (15.8)	131 (21.5)	24 (27.6)	0.007
Feeling of empty life^¶, n (%)	4 (3.2)	123 (14.4)	119 (42.5)	≤ 0.001	93 (16.5)	134 (22.0)	19 (21.8)	0.056
Caregiver^#, n (%)	5 (4.0)	162 (19.0)	91 (32.5)	≤ 0.001	143 (25.4)	106 (17.4)	9 (10.3)	≤ 0.001
Social isolation^α, n (%)	1 (0.8)	16 (1.9)	8 (2.9)	0.359	7 (1.2)	18 (3.0)	-	0.044
Need help for, n (%):
Phone use^∞	–	17 (2.0)	27 (9.6)	≤ 0.001	11 (2.0)	27 (4.4)	6 (6.9)	0.014
Transportation^μ	–	–	13 (4.6)	≤ 0.001	4 (0.7)	8 (1.3)	1 (1.1)	0.594
Medication intake^**	–	14 (1.6)	16 (5.7)	≤ 0.001	7 (1.2)	18 (3.0)	5 (5.7)	0.017
Handle finances^††	–	64 (7.5)	78 (27.9)	≤ 0.001	49 (8.7)	80 (13.1)	13 (14.9)	0.032
Bathing	1 (0.8)	4 (0.5)	14 (5.0)	≤ 0.001	3 (0.5)	12 (2.0)	4 (4.6)	0.007
Toileting	–	71 (8.3)	102 (36.4)	≤ 0.001	44 (7.8)	103 (16.9)	26 (29.9)	≤ 0.001
Dressing	–	2 (0.2)	9 (3.2)	≤ 0.001	1 (0.2)	8 (1.3)	2 (2.3)	0.038
Feeding	–	2 (0.2)	4 (1.4)	0.030	1 (0.2)	5 (0.8)	-	0.225
Incontinence^\|\|\|\|, n (%)	11 (8.8)	209 (24.5)	151 (53.9)	≤ 0.001	146 (26.0)	191 (31.3)	34 (39.1)	0.017
Fried items:
Weight loss^¶¶, n (%)	1 (0.8)	105 (12.3)	36 (12.9)	≤ 0.001	–	116 (19.0)	26 (29.9)	≤ 0.001
Poor energy^# #, n (%)	10 (8.0)	142 (16.6)	82 (29.3)	≤ 0.001	–	172 (28.2)	62 (71.3)	≤ 0.001
Slowness^αα, n (%)	7 (5.6)	139 (16.3)	84 (30.0)	≤ 0.001	–	161 (26.4)	69 (79.3)	≤ 0.001
Low muscle strength, n (%)	16 (12.8)	141 (16.5)	78 (27.9)	≤ 0.001	–	178 (29.2)	57 (65.5)	≤ 0.001
Low physical activity^∞∞, n (%)	19 (15.2)	143 (16.7)	80 (28.6)	≦ 0.001	–	176 (28.9)	66 (75.9)	≤ 0.001
Cognitive impairment: S-MMSE
Mean±SD (score)	5.4±0.8	5.2±1.1	4.8±1.2	≦ 0.001	5.1±1.1	5.1±1.1	5.1±0.9	0.934
≤4/6 80, n (%)	15 (11.7)	179 (20.9)	91 (32.4)	≦ 0.001	116 (20.6)	147 (24.1)	20 (23.0)	0.364

SD, standard deviation; CI, confidence interval; S-MMSE, Short-Mini Mental state examination. Overall comparisons between groups performed with an analysis of variance (ANOVA) with Bonferroni correction and a Chi-square test, as appropriate. *>25 or < 18.5 kg/m². ^†Number of drugs daily taken≥5. ^‡Related to bad health condition. ^||Regardless of the type of walking aid. ^¶Answer to the 30-item Geriatric Depression scale “do you feel that your life is empty?” yes. ^#To be an individual who gives regularly care to another who need help taking care of her/himself. ^αLiving alone, had no home help and no contact with another person over the past week. ^∞Answers telephone and dials few memorized numbers OR Does not dial OR Does not use telephone at all. ^μTravels on public transportation when accompanied by another OR Travel limited to taxi or automobile with assistance of another OR does not travel at all. **Need weekly supervision OR takes responsibility if medication is prepared in advance in separate dosage OR Is not capable of dispensing own medication. ^††Manages day-to-day purchases, but needs help with major purchases OR for regular purchases OR Incapable of handling money. ^||||Partially or totally incontinent of bowel or bladder. ^¶¶Reported weight loss in previous 6 months expressed in percentage using the formula (weight loss (kg)/ actual weight measured (kg)) x 100 and present if≥5% of body weight. ^# #Answer to 30-item Geriatric Depression scale “Do you feel full of energy?” no. ^ααSlow gait speed defined using the best of the two trials to walk at usual pace over a 4-meter distance. Time was measured in seconds and stratified according to sex and then according to height (2 strata per sex). In each stratum, participants in the slowest quintile were considered to have slow gait speed and coded 1. ^∞∞Low muscle strength (hand grip) was determined using a Martin Vigorimeter tested 3 times for each hand. The cohort was stratified according to sex and then according to BMI (in quartiles, four strata for each sex) to adjust for the effects of sex and BMI on muscle strength. From each stratum, the highest quintile (20%) of hand grip strength was chosen to defined low muscle strength and coded 1. ^μμScore ranged between 0 (no frailty) and 21(highest frailty) with three level status including robust (0-1), pre-frail (2-4) and frail (≥5). ***Robust (0 positive item), pre-frail (1-2 positive items) and frail (3-5 positive items).

Table 2

p-values of comparisons of baseline characteristics of frailty states using CARE (a) and Fried (b) frailty stratifications (n = 1,259)

Characteristics	CARE frailty states^μμ			Fried frailty states^***
	Robust versus Pre-Frail	Robust versus Frail	Pre-Frail versus Frail	Robust versus Pre-Frail	Robust versus Frail	Pre-Frail versus Frail
CARE items:
Age
Mean±SD (y)	0.133	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001
>80, n (%)	0.069	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	0.046
Male, n (%)	≤ 0.001	0.001	0.079	0.088	0.570	0.763
Body mass index abnormal^*, n (%)	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	0.065
No vitamin D supplementation, n (%)	0.007	≤ 0.001	≤ 0.001	0.989	0.883	0.888
Polypharmacy^†, n (%)	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	0.001
Unable to stand up from chair five-time, n (%)	0.310	0.019	0.019	0.178	0.044	0.306
Use of walking aid^\|\|, n (%)	0.135	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001
Housebound^‡, n (%)	0.702	-	0.567	0.337	-	0.705
Memory impairment, n (%)	≤ 0.001	≤ 0.001	≤ 0.001	0.014	0.007	0.200
Feeling of empty life^¶, n (%)	≤ 0.001	≤ 0.001	≤ 0.001	0.019	0.224	0.978
Caregiver^#, n (%)	≤ 0.001	≤ 0.001	≤ 0.001	0.001	0.002	0.098
Social isolation^α, n (%)	0.391	0.195	0.321	0.044	0.295	0.105
Need help for, n (%):
Phone use^∞	0.112	≤ 0.001	≤ 0.001	0.017	0.007	0.310
Transportation^μ	-	0.014	≤ 0.001	0.308	0.664	0.900
Medication intake^**	0.149	0.006	≤ 0.001	0.044	0.004	0.172
Handle finances^††	0.002	≤ 0.001	≤ 0.001	0.016	0.066	0.639
Bathing	0.627	0.039	≤ 0.001	0.029	0.001	0.125
Toileting	0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	0.003
Dressing	0.588	0.043	≤ 0.001	0.026	0.007	0.469
Feeding	0.588	0.179	0.017	0.124	0.694	0.397
Incontinence^\|\|\|\|, n (%)	≤ 0.001	≤ 0.001	≤ 0.001	0.044	0.011	0.147
Fried items:
Weight loss^¶¶, n (%)	≤ 0.001	≤ 0.001	0.805	≤ 0.001	≤ 0.001	0.019
Poor energy^# #, n (%)	0.013	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001
Slowness^αα, n (%)	0.002	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001
Low muscle strength, n (%)	0.291	0.001	≤ 0.001	≤ 0.001	≤ 0.001	≤ 0.001
Low physical activity^∞∞, n (%)	0.664	0.004	≦ 0.001	≦ 0.001	≦ 0.001	≦ 0.001
Cognitive impairment: S-MMSE
Mean±SD (score)	0.014	≦ 0.001	≦ 0.001	1.000	1.000	1.000
≤4/6 80, n (%)	0.010	≦ 0.001	≦ 0.001	0.156	0.617	0.821

SD, standard deviation; CI, confidence interval; S-MMSE, Short-Mini Mental state examination. Comparisons between groups performed using analysis of variance (ANOVA) with Bonferroni correction and the Chi-square test, as appropriate. *>25 or < 18.5 kg/m². ^†Number of drugs daily taken≥5. ^‡Related to bad health condition. ^||Regardless of the type of walking aid. ^¶Answer to the 30-item Geriatric Depression scale “do you feel that your life is empty?” yes. ^#To be an individual who gives regularly care to another who need help taking care of her/himself. ^αLiving alone, had no home help and no contact with another person over the past week. ^∞Answers telephone and dials few memorized numbers OR Does not dial OR Does not use telephone at all. ^μTravels on public transportation when accompanied by another OR Travel limited to taxi or automobile with assistance of another OR does not travel at all. **Need weekly supervision OR takes responsibility if medication is prepared in advance in separate dosage OR Is not capable of dispensing own medication. ^††Manages day-to-day purchases, but needs help with major purchases OR for regular purchases OR Incapable of handling money. ^||||Partially or totally incontinent of bowel or bladder. ^¶¶Reported weight loss in previous 6 months expressed in percentage using the formula (weight loss (kg)/ actual weight measured (kg)) x 100 and present if≥5% of body weight. ^# #Answer to 30-item Geriatric Depression scale “Do you feel full of energy?” no. ^ααSlow gait speed defined using the best of the two trials to walk at usual pace over a 4-meter distance. Time was measured in seconds and stratified according to sex and then according to height (2 strata per sex). In each stratum, participants in the slowest quintile were considered to have slow gait speed and coded 1. ^∞∞Low muscle strength (hand grip) was determined using a Martin Vigorimeter tested 3 times for each hand. The cohort was stratified according to sex and then according to BMI (in quartiles, four strata for each sex) to adjust for the effects of sex and BMI on muscle strength. From each stratum, the highest quintile (20%) of hand grip strength was chosen to defined low muscle strength and coded 1. ^μμScore ranged between 0 (no frailty) and 21(highest frailty) with three level status including robust (0-1), pre-frail (2-4) and frail (≥5). ***Robust (0 positive item), pre-frail (1-2 positive items) and frail (3-5 positive items).

Cox models using the CARE frailty stratification revealed that frail state, cognitive impairment, and their combination were significantly associated with incident MNCD, regardless of the referent group used (Fig. 1). The highest Hazard ratio (HR) was observed in the combination of frail state and cognitive impairment using the robust group as the referent group (HR = 6.53, CI = [3.49-12.23] with p≤0.001). Results were more heterogenous using the Fried frailty stratification (Fig. 2). Fried’s frail state was not significantly associated with incident MNCD when using robust and pre-frail participants pooled together and pre-frail participants respectively as the referent groups. Cognitive impairment and its combination with Fried’s frail state were consistently associated with incident MNCD, the highest HR being observed with cognitive impairment when using robust participants as the referent group (HR = 8.68, CI = [1.13-18.23] with p≤0.001).

Fig. 1

Cox regression showing the association of CARE frail state, abnormal S-MMSE and combination of both with incident major neurocognitive disorders in NuAge participants. a) Robust and prefrail participants pooled together and used as the referent group. b) Robust participants used as the referent group. c) Prefrail participants used as the referent group (n = 1,259). S-MMSE, Short-Mini Mental state examination; CARE frail state, CARE score ranged between 0 (no frailty) and 21 (highest frailty) with frail state score ≥5; CI, confident interval.

Fig. 2

Cox regression showing the association of Fried frail state, abnormal S-MMSE and combination of both with incident major neurocognitive disorders in NuAge participants. a) Robust and prefrail participants pooled together and used as the referent group. b) Robust participants used as the referent group. c) Prefrail participants used as the referent group (n = 1,259). S-MMSE, Short-Mini Mental state examination; Fried frail state, Fried score ranged between 0 (no frailty) and 5 (highest frailty) with frail state score ≥3. CI, confident interval.

4 DISCUSSION

Our findings show a greater association of the CARE frail state with incident MNCD and an increased predictive value when combined with cognitive impairment in comparison to the Fried frail state, the highest incidences being observed when using the robust state as the referent group. In addition, an overall significant incidence with MNCD was reported with the CARE frail state, regardless of the type of association examined. The results for the Fried frail state were more heterogenous, with no association demonstrated between the frail state and MNCD incidence with either all groups pooled together or the pre-frail group as the referent. Furthermore, cognitive impairment alone showed the highest incidence.

Few studies to date have examined the association of frailty with cognitive decline or incident MNCD using the deficit accumulation model of frailty [12]. Two studies showed a significant incidence with cognitive decline and one with MNCD [13, 29, 30]. Our study confirms these previous results, demonstrating that the deficit accumulation model, like Fried’s physical impairment model, is a good method for quantifying frailty and its association with the occurrence of MNCD. The CARE frailty stratification is a multidomain deficit model that assesses physical, psychological, and social deficits. Like the Rockwood model, the weight of cognitive impairment in the calculation of frailty is limited to one item, which is a memory complaint reported by relatives [11, 31]. Given that the NuAge participants were free of MNCD at baseline, the CARE frail state includes individuals that may report such cognitive complaints, which is the first pre-MNCD stage, which has a weaker association with incident MNCD compared to the second stage. This second stage corresponds to minor neurocognitive disorders (i.e., mild cognitive impairment; MCI), characterized by low cognitive performance without consequence in daily living activities [32]. In our study, the MCI stage was defined as an abnormal S-MMSE score in individuals free of MNCD. Both frailty stratification models demonstrated a significant association with incident MNCD, which is consistent with the findings of previous studies [12]. In addition, the highest HR [8.68] was shown with an abnormal S-MMSE score while using robust participants as the referent, highlighting that cognitive deficit is the best domain to consider for the prediction of MNCD. This finding may be explained by the fact that S-MMSE specifically assesses episodic memory impairment, which is the short-term memory characteristic in people with MNCD such as Alzheimer’s disease [29]. Thus, it is not surprising that a test assessing the performance of this memory may detect the early onset of MNCD.

Interestingly, we reported that the combination of cognitive impairment with the CARE frail state, unlike the Fried frail state, increased the predictive value of incident MNCD. To the best of our knowledge, this is the first time that such a result is reported. This result bolsters the case for using a deficit accumulation model to quantify frailty when the ultimate goal is to predict MNCD. The value of the deficit accumulation model, when compared to a model limited to the physical domain, may be attributed to its multidomain assessment criteria [11]. Adverse consequences of MNCD are not limited to the cognitive domain: they include physical, psychological, and social domains [4–6]. It has been shown that the pathological process in the brain which leads to MNCD starts early, before the first clinical symptoms [4–6, 32]. Thus, it might be suggested that early symptoms of MNCD may be present in different cognitive domains, explaining the greater ability of the deficit accumulation model to assess the first symptoms of MNCD, before their diagnosis, when compared to Fried’s physical model of frailty. In addition, our results showed that the combination of cognitive impairment with Fried’s frail state were a greater predictor of MNCD compared to the frail state alone, regardless of the referent group used. This result is consistent with the literature showing that combining physical impairment and cognitive impairment is useful for the prediction of MNCD [33, 34]. For example, slow walking speed and subjective cognitive complaint are the two clinical characteristics defining motoric cognitive risk syndrome in individuals free of MNCD [35]. They are independently associated with incident dementia, but the risk for MNCD is greater when they are combined [36].

We observed a distribution of frailty which was different between the CARE and Fried frail scales. There were more vigorous participants according to the Fried scale compared to the CARE scale (44.6% versus 9.9%) and, thus, the proportion of participants with frail state was higher with the CARE scale compared to the Fried scale (22.2% versus 6.9%). A similar distribution has been previously reported in a previous study which used the full set of NuAge participants [37]. Furthermore, the distribution of frail states showed in our study looks like it observed in the database of the Survey of Health, Aging and Retirement in Europe which was composed of 11,015 community-dwelling men and women aged 60+ recruited in 11 countries of the European Union [38]. The difference between CARE and Fried distribution of frail states may be explained by the fact that the CARE and Fried scales use a different health deficit count. The CARE items cover physical, psychological, and social domains, whereas the Fried items are limited to the physical domain. This difference between the CARE and Fried scales may explained that only a significant association between CARE frail state and incident MNCD was found. It has been reported that physical impairment was associated with an increased risk of MNCD [28, 33, 35]. But this association was mainly reported with slow gait speed and not with the muscle strength and other Fried items. Thus, this non-conclusive result with the Fried frail scale is consistent with the published literature on this topic [38]. The greater ability of the CARE scale to detect an increased risk of incident MNCD compared to the Fried scale may be explained by the fact that the CARE scale assesses deficits (i.e., symptoms, signs, diseases, and disabilities) in more than one domain including physical, psychological, and social domains compared to the Fried scale, which is limited to the physical domain.

Strengths of the present study include the large sample size of the NuAge cohort, exhaustive assessment, and the 3-year duration of prospective and observational follow-up. However, there are limitations to consider. First, the incidence of MNCD in our study was low (2.6%) in comparison to previous studies. It suggests that incidence might have been underestimated because of the criteria used to define MNCD. A diagnosis of MNCD is usually based on interdisciplinary evaluation and more exhaustive information, including a comprehensive neuropsychological assessment and brain imaging. Second, the incidence of MCND may depend on their etiologies, Alzheimer disease being the most prevalence MCND in the older population [39]. In the NuAge study we have no information about MCND etiologies. Third, although we were able to control for multiple characteristics likely to modify the association, residual confounders might still be present.

4.1 Conclusion

In the NuAge participants, we reported that both the CARE and Fried frailty models are associated with incident MNCD, and that the CARE frail state (based on the Rockwood model) was superior to predict MNCD. Furthermore, the association of the CARE frail state with cognitive impairment increased the predictive value of MNCD. All these results suggest that the CARE frailty model may be of clinical interest when screening MCND in the elderly population.

Footnotes

ACKNOWLEDGMENTS

The authors gratefully acknowledge the voluntary participation of all study participants, the NuAge team for its assistance and members of the iTeQ scientific council including Dr. Cyrille P Launay, Dr. Daniel Benatar, Dr. Alain Ptito, Dr. Pittie Chou, MD; and Dr. Frédéric Prate.

FUNDING

The NuAge Study was funded by the Canadian Institutes of Health Research (CIHR; MOP-62842). The NuAge Database and Biobank are supported by the Fonds de Recherche du Québec (FRQ; 2020-VICO-279753), the Quebec Network for Research on Aging (a thematic network funded by the FRQ-Santé) and the Merck-Frosst Chair (funded by La Fondation de l’Université de Sherbrooke).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

Access to the NuAge Database can be obtained by contacting the NuAge team via their website (https://nuage.recherche.usherbrooke.ca/en/) or at NuAge-cdrv@usherbrooke.ca.

References

Lobo

, Lobo

, De-la-Cámara

(2019) Dementia care inhigh-income countries. Curr Opin Psychiatry 32, 465–470.

Grande

, Qiu

, Fratiglioni

(2020) Prevention of dementia in an ageing world: Evidence and biological rationale. Ageing Res Rev 64, 101045.

Public Health Agency of Canada (2019) A Dementia Strategy for Canada: Together We Aspire. https://www.canada.ca/en/public-health/services/publications/diseases-conditions/dementia-strategy.html

Stephen

, Barbera

, Peters

, Ee

, Zheng

, Lehtisalo

, Kulmala

, Håkansson

, Chowdhary

, Dua

, Solomon

, Anstey

, Kivipelto

(2021) Development of the First WHO Guidelines for Risk Reduction of Cognitive Decline and Dementia: Lessons learned and future directions. Front Neurol 12, 763573.

Solomon

, Stephen

, Altomare

, Carrera

, Frisoni

, Kulmala

, Molinuevo

, Nilsson

, Ngandu

, Ribaldi

, Vellas

, Scheltens

, Kivipelto

; European Task Force for Brain Health Services (2021) Multidomain interventions: State-of-the-art and future directions for protocols to implement precision dementia risk reduction. A user manual for Brain Health Services-part 4 of 6. Alzheimers Res Ther 13, 171.

Ranson

, Rittman

, Hayat

, Brayne

, Jessen

, Blennow

, van Duijn

, Barkhof

, Tang

, Mummery

, Stephan

BCM

, Altomare

, Frisoni

, Ribaldi

, Molinuevo

, Scheltens

, Llewellyn

; European Task Force for Brain Health Services (2021) Modifiable risk factors for dementia and dementia risk profiling. A user manual for Brain Health Services-part 2 of 6. Alzheimers Res Ther 13, 169.

Hallam

, Rees

, Petersen

, Cooper

, Avgerinou

, Walters

(2021) How are people with mild cognitive impairment or subjective memory complaints managed in primary care? A systematic review. Fam Pract 38, 669–683.

Church

, Rogers

, Rockwood

, Theou

(2020) A scoping review of the Clinical Frailty Scale. BMC Geriatr 20, 393.

Huang

, Lam

(2021) Review of frailty measurement of older people: Evaluation of the conceptualization, included domains, psychometric properties, and applicability. Aging Med (Milton) 4, 272–291.

10.

Dent

, Kowal

, Hoogendijk

(2016) Frailty measurement in research and clinical practice: A review. Eur J Intern Med 31, 3–10.

11.

Searle

, Mitnitski

, Gahbauer

, Gill

, Rockwood

(2008) A standard procedure for creating a frailty index. BMC Geriatr 8, 24.

12.

Chu

, Chang

, Ho

(2021) Adverse health effects of frailty: Systematic review and meta-analysis of middle-aged and older adults with implications for evidence-based practice. Worldviews Evid Based Nurs 18, 282–289.

13.

Song

, Mitnitski

, Rockwood

(2011) Nontraditional risk factors combine to predict Alzheimer disease and dementia. Neurology 77, 227–234.

14.

Fried

, Tangen

, Walston

, Newman

, Hirsch

, Gottdiener

, Seeman

, Tracy

, Kop

, Burke

, McBurnie

; Cardiovascular Health Study Collaborative Research Group (2011) Frailty in older adults: Evidence for a phenotype. J Gerontol A Biol Sci Med Sci 56, M146–156.

15.

Theou

, Brothers

, Peña

, Mitnitski

, Rockwood

(2014) Identifying common characteristics of frailty across seven scales. J Am Geriatr Soc 62, 901–906.

16.

Kelaiditi

, Cesari

, Canevelli

, van

Kan GA

, Ousset

, Gillette-Guyonnet

, Ritz

, Duveau

, Soto

, Provencher

, Nourhashemi

, Salvà

, Robert

, Andrieu

, Rolland

, Touchon

, Fitten

, Vellas

, IANA/IAGG (2013) Cognitive frailty: Rational and definition from an (I.A.N.A./I.A.G.G.) international consensus group. J Nutr Health Aging 17, 726–734.

17.

Facal

, Burgo

, Spuch

, Gaspar

, Campos-Magdaleno

(2021) Cognitive frailty: An update. Front Psychol 12, 813398.

18.

Jull

, Köpke

, Smith

, Carley

, Finderup

, Rahn

, Boland

, Dunn

, Dwyer

, Kasper

, Kienlin

, Légaré

, Lewis

, Lyddiatt

, Rutherford

, Zhao

, Rader

, Graham

, Stacey

(2021) Decision coaching for people making healthcare decisions. Cochrane Database Syst Rev 11, CD013385.

19.

Marques

MDC

, Pires

, Perdigão

, Sousa

, Fonseca

, Pinho

, Lopes

(2021) Patient-centered care for patients withcardiometabolic diseases: An integrative review.. J Pers Med 11, 1289

20.

World Health Organization (1996) Health interview surveys: Towards international harmonization of methods and instruments. WHO Regional Publications – European Series No. 58.

21.

Andrews

, Kemp

, Sunderland

, Von Korff

, Ustun

(2009) Normative data for the 12 item WHO Disability Assessment Schedule 2.0. PLoS One 4, e8343.

22.

Ustün

, Chatterji

, Kostanjsek

, Rehm

, Kennedy

, Epping-Jordan

, Saxena

, von Korff

, Pull

; WHO/NIH Joint Project (2010) Developing the World Health Organization Disability Assessment Schedule 2.0. Bull World Health Organ 88, 815–823.

23.

Gaudreau

, Morais

, Shatenstein

, Gray-Donald

, Khalil

, Dionne

, Ferland

, Fülöp

, Jacques

, Kergoat

, Tessier

, Wagner

, Payette

(2007) Nutrition as a determinant of successful aging: Description of the Quebec longitudinal study Nuage and results from cross-sectional pilot studies. Rejuvenation Res 10, 377–386.

24.

Teng

, Chui

(1987) The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry 48, 314–318.

25.

Hebert

, Carrier

, Bilodeau

(1988) The Functional Autonomy Measurement System (SMAF): Description and validation of an instrument for the measurement of handicaps. Age Ageing 17, 293–302.

26.

Yesavage

, Brink

, Rose

, Lum

, Huang

, Adey

, Leirer

(1982) Development and validation of a geriatric depression screening scale: A preliminary report. J Psychiatr Res 17, 37–49.

27.

Washburn

, McAuley

, Katula

, Mihalko

, Boileau

(1999) The physical activity scale for the elderly (PASE): Evidence for validity. J Clin Epidemiol 52, 643–651.

28.

Beauchet

, Matskiv

, Launay

, Gaudreau

, Benatar

, Ptito

, Chou

, Allali

, Prate

, Vaillant-Ciszewicz

, Guerin

(2022) CARE frailty e-health scale: Association with incident adverse health outcomes and comparison with the Cardiovascular Health Study frailty scale in the NuAge cohort. Maturitas 162, 37–43.

29.

Haubois

, de Decker

, Annweiler

, Launay

, Allali

, Herrmann

, Beauchet

(2013) Derivation and validation of a Short Form of the Mini-Mental State Examination for the screening of dementia in older adults with a memory complaint. Eur J Neurol 20, 588–590.

30.

Rockwood

, Abeysundera

, Mitnitski

(2007) How should we grade frailty in nursing home patients? J Am Med Dir Assoc 8, 595–603.

31.

Mitnitski

, Fallah

, Rockwood

(2011) A multistate model of cognitive dynamics in relation to frailty in older adults. Ann Epidemiol 21, 507–516.

32.

Chen

, Liang

, Li

, Yang

, Wang

, Shi

(2021) Diagnosis and treatment for mild cognitive impairment: A systematic review of clinical practice guidelines and consensus statements. Front Neurol 12, 719849.

33.

Beauchet

, Annweiler

, Callisaya

, De Cock

, Helbostad

, Kressig

, Srikanth

, Steinmetz

, Blumen

, Verghese

, Allali

(2016) Poor gait performance and prediction of dementia: Results from a meta-analysis. J Am Med Dir Assoc 17, 482–490.

34.

Wang

, Wang

, Hu

, Qu

, Wang

, Shen

, Xu

, Dong

, Tan

, Yu

(2021) Association of subjective cognitive decline with risk of cognitive impairment and dementia: A systematic review and meta-analysis of prospective longitudinal studies. J Prev Alzheimers Dis 8, 277–285.

35.

Verghese

, Wang

, Lipton

, Holtzer

(2013) Motoric cognitive risk syndrome and the risk of dementia. J Gerontol A Biol Sci Med Sci 68, 412–418.

36.

Beauchet

, Sekhon

, Cooper-Brown

, Launay

, Gaudreau

, Morais

, Allali

(2021) Motoric cognitive risk syndrome and incidentdementia in older adults from the Québec NuAge cohort. AgeAgeing 50, 969–973.

37.

Macklai

, Spagnoli

, Junod

, Santos-Eggimann

(2013) Prospective association of the SHARE-operationalized frailty phenotype with adverse health outcomes: Evidence from 60+ community-dwelling Europeans living in 11 countries. BMC Geriatr 13, 3.

38.

Kueper

, Speechley

, Lingum

, Montero-Odasso

(2017) Motor function and incident dementia: A systematic review and meta-analysis. Age Ageing 46, 729–738.

39.

Goodman

, Lochner

, Thambisetty

, Wingo

, Posner

, Ling

(2017) Prevalence of dementia subtypes in United States Medicare fee-for-service beneficiaries, 2011-2013. Alzheimers Dement 13, 28–37.