Abstract
Background
Frailty and intrinsic capacity (IC) are distinct but interrelated constructs. Uncertainty remains regarding how they are related and interact to influence health outcomes. We aim to understand the relationship between frailty and IC by identifying subgroups based on frailty criteria and IC domains and studying one-year outcomes.
Methods
We studied 200 independent community-dwelling older adults (mean age 67.9±7.9 years, Modified Barthel Index (MBI) score 99±2.6). Frailty was defined by modified Fried criteria. Scores (range: 0–2) were assigned to individual IC domains (cognition, psychological, locomotion, and vitality) to yield a total IC score of 8. To identify subgroups, two-step cluster analysis was performed with age, frailty and IC domains. Cluster associations with one-year outcomes (frailty, muscle strength (grip strength, repeated chair stand test), physical performance (gait speed, Short Physical Performance Battery), function (MBI) and quality-of-life (EuroQol (EQ)-5D)) were examined using multiple linear regression adjusted for age, gender and education.
Results
Three distinct clusters were identified — Cluster 1: High IC/Robust (N=74, 37%); Cluster 2: Intermediate IC/Prefrail (N=73, 36.5%); and Cluster 3: Low IC/Prefrail-Frail (53, 26.5%). Comparing between clusters, IC domains, cognition, depressive symptoms, nutrition, strength and physical performance were least impaired in Cluster 1, intermediate in Cluster 2 and most impaired in Cluster 3. At one year, the proportion transitioning to frailty or remaining frail was highest in Cluster 3 compared to Cluster 2 and Cluster 1 (39% vs 6.9% vs 2.8%, P<0.001). Compared to Cluster 1, Cluster 3 experienced greatest declines in grip strength (β=−4.1, P<.001), MBI (β=−1.24, P=0.045) and EQ-5D utility scores (β=−0.053, P=0.005), with Cluster 2 intermediate between Cluster 1 and Cluster 3.
Conclusions
Amongst independent community-dwelling older adults, IC is complementary to frailty measures through better risk-profiling of one-year outcomes amongst prefrail individuals into intermediate and high-risk groups. The intermediate group merits follow-up to ascertain longer-term prognosis.
Key words: Intrinsic capacity, frailty, outcomes, healthy ageing, cluster analysis
Background
The World Health Organisation's (WHO) proposal for a Decade of Healthy Ageing 2020–2030 (1) shifts the discourse from the mere absence of disease to a paradigm of healthy ageing promoting functional ability, determined by intrinsic capacity (IC) — the combination of all the individual's physical and mental capacities. IC comprises cognition, locomotion, psychological, sensory and vitality domains. Empirical evidence supports the structure of IC and its predictive validity for care dependence (2).
On the other hand, frailty is a well-established construct characterized by a decline in physiologic functions, rendering individuals more vulnerable to stressors and adverse outcomes, including mortality and dependency (3). While similarly transcending the disease-centred approach to health in ageing, a key difference between IC and frailty is the conceptual underpinning of both constructs. Deficits and abnormalities define frailty, whereas reserves and residual capacities define IC. However, instead of regarding IC as the mere opposite of frailty, both concepts might be complementary. For example, reductions in IC may better characterize frailty in relatively healthy older persons, and the measurement of IC along the trajectory of frailty may provide more information to better define care plans tailored to the individual (4). Currently, few studies have explored the relationship between frailty and IC in detail. With frailty adopted in many clinical settings, understanding this relationship will clarify the additional role of IC measures in clinical practice.
Therefore, we aim to understand the relationship between frailty and IC in healthy community-dwelling older adults by identifying subgroups based on frailty criteria and IC domains and studying longitudinal outcomes. We hypothesise that frailty and IC are complementary concepts that improve risk profiling of cross-sectional characteristics and one-year outcomes.
Methods
Study design and population
Baseline and longitudinal analysis of the “Longitudinal Assessment of Biomarkers for characterization of early Sarcopenia and predicting frailty and functional decline in community-dwelling Asian older adults Study” (Geri-LABS), a prospective cohort study involving cognitively intact and functionally independent adults aged 50 years and older residing within the community (5). Inclusion criteria included (1) individuals aged between 50 to 99 years old; (2) community-dwelling; and (3) functionally independent in both activities of daily living (ADLs) and instrumental ADLs (IADLs). We excluded individuals with a history of dementia, evidence of cognitive impairment or the inability to walk at least 4.5 metres independently. We also excluded residents of sheltered or nursing homes.
We obtained informed written consent from each participant. This study was approved by the Domain Specific Review Board (DSRB) of the National Healthcare Group (NHG).
Operationalising frailty and intrinsic capacity
Frailty and IC domains are presented in Table 1. Frailty was defined using the modified Fried criteria (6), comprising: (1) Body mass index <18.5; (2) Handgrip strength below gender-specific cut-offs consistent with the Asian Working Group for Sarcopenia (AWGS) criteria (males >28 kg, females >18 kg) (7) using the North Coast Exacta™ Hydraulic Hand Dynamometer; (North Coast Medical, Inc., Morgan Hill, CA, USA); (3) Usual gait speed <1.0 m/s on the 3-metre walk test (7); (4) Low physical activity defined using the quintile cut-off of ≤29 on the Frenchay Activities Index (8) and (5) Fatigue endorsed on either of two questions from the Center for Epidemiologic Studies-Depression Scale (CES-D) modified to assess fatigue, viz. a positive response to both the following: (a) “I felt that everything I did was an effort” and (b) “I could not get going” (9). Participants were classified as frail if 3 to 5 criteria were present, prefrail if 1–2 criteria present and robust if none were present.
Table 1.
Cohort and cluster baseline characteristics
| Overall (N=200) | Cluster 1 (N=74) High IC, Robust | Cluster 2 (N=73) Intermediate IC, Prefrail | Cluster 3 (N=53) Low IC, Prefrail/Frail | P | |
|---|---|---|---|---|---|
| Demographics | |||||
| Age in years, mean (SD) | 67.9 (7.9) | 64.7 (6.8)* | 68.1 (7.5)** | 72.2 (7.7)*** | <0.001 |
| Female gender, N (%) | 137 (68.5) | 58 (78.4) | 45 (61.6) | 34 (64.2) | 0.067 |
| Chinese ethnicity, N (%) | 184 (92.0) | 67 (90.5) | 67 (91.8) | 50 (94.3) | 0.99 |
| Years of education, mean (SD) | 9.0 (4.9) | 9.7 (5.1) | 9.5 (4.4)** | 74 (4 9)*** | 0.013 |
| Comorbidities, N (%) | |||||
| Hypertension | 96 (48.0) | 28 (37.8) | 33 (45.2) | 35 (66.0) | 0.006 |
| DM | 43 (21.5) | 16 (21.6) | 11 (15.1) | 16 (30.2) | 0.13 |
| Hyperlipidemia | 132 (66.0) | 47 (63.5) | 50 (68.5) | 35 (66.0) | 0.82 |
| IHD | 4 (2.0) | 0 | 2 (2.7) | 2(3.8) | 0.28 |
| Stroke | 5 (2.5) | 0 | 1 (1.4) | 4 (7.6) | 0.02 |
| Malignancy | 12 (6.0) | 2 (2.7) | 6 (8.2) | 4 (7.6) | 0.32 |
| Inflammatory disease | 2 (1.0) | 0 | 1 (1.4) | 1 (1.9) | 0.53 |
| Sarcopenia, N (%) | |||||
| Non-sarcopenia | 142 (71.0) | 74 (100) | 48 (65.8) | 20 (37.7) | <0.001 |
| Sarcopenia | 41 (20.5) | 0 | 24 (32.9) | 18 (32.1) | |
| Severe sarcopenia | 17 (8.5) | 0 | 1 (1.4) | 16 (30.2) | |
| Cognition | |||||
| CMMSE, mean (SD) | 26.2 (1.7) | 26.3 (1.7) | 26.8 (1.1)** | 25.1 (1.9)*** | <0.001 |
| Depressive symptoms | |||||
| GDS, mean (SD) | 1.1 (2.0) | 0.55 (0.86) | 0.68 (1.2)** | 2.5 (3.1)*** | <0.001 |
| Nutrition | |||||
| MNA, mean (SD) | 27.0 (1.9) | 27.4 (1.4) | 27.2 (1.9)** | 26.1 (2.3)*** | 0.0005 |
| Muscle strength | |||||
| Grip strength, kg, mean (SD) | 24.0 (7.6) | 25.1 (7.5) | 23.3 (7.9)** | 17 4 (4 4)*** | 0.0004 |
| 5-times repeated chair stand test, sec, mean (SD) | 10.4 (2.6) | 9.1 (1.6)* | 10.5 (2.1)** | 12.2 (3.2)*** | <0.001 |
| Physical performance | |||||
| Gait speed, m/s, mean (SD) | 1.14 (0.21) | 1.23 (0.15) | 1.16 (0.19)** | 0.98 (0.23)*** | <0.001 |
| SPPB total score, mean (SD) | 11.3 (1.4) | 11.9 (0.54) | 11.5 (0.67)** | 10.2 (2.1)*** | <0.001 |
| Activity | |||||
| FAI, mean (SD) | 32.1 (5.1) | 34.9 (2.9)* | 30.8 (5.4) | 29.9 (5.4)*** | <0.001 |
| Function | |||||
| MBI, mean (SD) | 99.0 (2.6) | 99.3 (2.1) | 99.0 (2.5) | 98.4 (3.4) | 0.18 |
| Quality of life | |||||
| EQ-5D utility scores, mean (SD) | 0.95 (0.08) | 0.98 (0.06) | 0.96 (0.07)** | 0.92 (0.11)*** | 0.0009 |
SD: standard deviation; DM: diabetes mellitus; IHD: ischemic heart disease; CMMSE: modified version of the Chinese Mini-Mental State Examination; GDS: Geriatric Depression Scale; MNA: Mini-Nutritional Assessment; SPPB: Short Physical Performance Battery; FAI; Frenchay Activities Index; ADL: Activities of Daily Living; EQ-5D: EuroQoL-5D; Bonferroni-corrected post-hoc comparisons: *Cluster 1 versus Cluster 2, P<0.05; **Cluster 2 versus Cluster 3, P<0.05; ***Cluster 1 versus Cluster 3, P<0.05
We operationalised IC using four domains, each comprising two items. As IC is considered a “positive” construct, domain items were scored one if they were unimpaired, for a total score of 8, as follows: (1) Cognition domain: CMMSE above age and education-adjusted cut-offs (10) and if subjects did not endorse the subjective memory complaint item on the 15-item Geriatric Depression Scale (11); (2) Psychological domain: GDS total score ≤4 and if participants did not endorse the modified CES-D items on fatigue; (3) Locomotion domain: unimpaired five-times repeated chair stand test (RCST) <12 seconds and 3-metre usual gait speed ≥1.0 m/s (7); and (4) Vitality domain: unimpaired handgrip strength and Mini-Nutritional Assessment (MNA) scores ≥24 (12). We omitted the sensory domain of IC as data was unavailable, an approach similar to a previous study (13).
Data collection
We collected baseline demographic information, including age, gender, ethnicity, educational level and medical comorbidities. We also determined sarcopenia status using the Asian Working Group consensus criteria (7). Frailty status and IC were determined at baseline, operationalised using the criteria above. Outcome measures included frailty status at one year, muscle strength (handgrip strength, RCST), physical performance (gait speed, Short Physical Performance Battery (SPPB) (14)), physical function (Modified Barthel Index (MBI) (15)) and quality-of-life (EuroQoL (EQ)-5D (16)) measures were also collected at baseline and one year.
Statistical analysis
We performed two-step cluster analysis with age, IC (total and domain scores) and frailty categories as classification variables. Compared to traditional clustering techniques (hierarchical and k-means clustering), the two-step procedure is preferable as both categorical and continuous variables can be analysed (17). Number of clusters was determined automatically by Bayesian Information Criterion (BIC) as clustering criterion. The smallest BIC value accompanied by the highest values of BIC change indicates a good model fit. Log-likelihood was used as a distance measure (18). Cluster validity was assessed by the average silhouette coefficient, a measure of cluster cohesion and separation ranging from −1 to +1 (19). Average silhouette coefficient of >0.5 indicates good model fit, 0.2 to 0.5 fair and <0.2 poor fit (20).
We then determined cluster characteristics based on IC scores (total and domain) and frailty categories (robust, prefrail and frail). Differences in demographic characteristics (age, gender, ethnicity, education level), comorbidities and IC domains were also examined across clusters. Outcomes of frailty (proportion transiting to frailty or remaining frail) across clusters were determined at one year. Chi-square test was used for categorical variables and one-way analysis of variance (ANOVA) for continuous variables with Bonferroni-corrected post-hoc comparisons. To examine associations between clusters and one-year outcomes, multiple linear regression was performed for continuous variables (one-year change in handgrip strength, gait speed, RCST, SPPB total score, ADL function and EQ-5D utility scores), adjusted for baseline scores, age, gender and education level. Two-sided tests with a significance level of P<0.05 were applied. Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 23.0 (Armonk, NY).
Results
Baseline characteristics
Two-hundred participants were recruited, mean age 67.9±7.9 years. 137 (68.5%) were female and 184 (92%) of Chinese ethnicity. Overall, participants were functionally independent (mean MBI 99.0±2.6), had unimpaired cognitive scores (mean CMMSE 26.2±1.7) and did not have significant depressive symptoms (mean GDS 1.1±2.0). Mean total IC score for the whole cohort was 6.8±1.3, with a majority possessing unimpaired IC domain scores (75%, 86%, 67% and 70% for cognition, psychological, locomotion and vitality domains respectively). According to the modified Fried criteria, 89 (44.5%) were robust, 96 (48.0%) prefrail and 15 (7.5%) frail.
Cluster analysis
Three clusters based on IC, frailty and age were identified, comprising 74 (37%), 73 (36.5%) and 53 (26.5%) in Clusters 1, 2 and 3 respectively. Average silhouette coefficient was 0.4, indicating fair cluster validity. The most important predictors of cluster membership were IC total scores (predictor importance: 1.0) and frailty category (predictor importance: 0.65), followed by locomotion, cognition, vitality, age and IC psychological domain.
In the distribution of IC scores across 3 clusters (Figure 1), Cluster 1 comprised 61 (82%) with full IC scores with none scoring below 6. Cluster 2 comprised 48 (66%) with IC score of 7, followed by 15 (21%) scoring 6. In Cluster 3, 37 (70%) scored 5 and below. In the analysis of frailty category by cluster, 100% of individuals in Cluster 1 were robust. In Cluster 2, 14 (19.2%) were robust, 59 (80.8%) prefrail and none frail. In Cluster 3, 1 (1.9%) individual was robust, 37 (69.8%) prefrail and 15 (28.3%) frail. Therefore, based on the descriptive analysis of frailty and IC characteristics, we labelled the three cluster as: Cluster 1 — High IC and Robust; Cluster 2 — Intermediate IC and Prefrail; and Cluster 3 — Low IC, Prefrail/ Frail.
Figure 1.
Intrinsic capacity and frailty across 3 clusters
A. Distribution of IC scores by cluster; B. Frailty categories by cluster
Cluster characteristics: demographics and comorbidities
Cluster characteristics are presented in Table 1. Participants in Cluster 3 were older (72.2±7.7 years) compared to Cluster 2 (68.1±7.5 years) and Cluster 1 (64.7±6.8 years, P, ANOVA <0.001). Lowest education level was observed in Cluster 3. There were no differences in gender and ethnicity across clusters. In terms of comorbidities, Cluster 3 had the highest proportion of participants with hypertension (P=0.006) and stroke (P=0.02) compared to Clusters 1 and 2. The proportion of participants with sarcopenia and severe sarcopenia was also the highest in Cluster 3, with Cluster 2 intermediate between Clusters 1 and 3 (P<0.001).
Cluster characteristics: IC domains and components
When IC domain scores were analysed by cluster (Figure 2) all individuals in Cluster 1 were unimpaired in psychological, locomotion and vitality domains. Cluster 3 was the poorest performing for all 4 IC domains. Except for cognition, IC domain scores in Cluster 2 were intermediate between Clusters 1 and 3.
Figure 2.
Intrinsic capacity domain scores by cluster
Further analysis of the components making up the IC domains (Table 1) revealed poorest cognitive performance (CMMSE, Cluster 1: 26.3±1.7; Cluster 2: 26.8±1.1; Cluster 3: 25.1±1.9, P<0.001), greater depressive symptoms (GDS, Cluster 1: 0.55±0.86; Cluster 2: 0.68±1.2; Cluster 3: 2.5±3.1, P<0.001) and poorest nutrition (MNA, Cluster 1: 27.4±1.4; Cluster 2: 27.2±1.9; Cluster 3: 26.1±2.3, P=0.0005) in Cluster 3, compared to Clusters 1 and 2. A similar pattern was observed for muscle strength and physical performance measures. Post-hoc analyses did not show significant differences between Clusters 1 and 2 except for RCST, where significant differences were found between all 3 clusters (RCST, Cluster 1: 9.1±1.6; Cluster 2: 10.5±2.1; Cluster 3: 12.2±3.2, P<0.001). Notably, mean RCST amongst individuals in Cluster 3 falls within the impaired range as defined by the AWGS 2019 consensus criteria (7).
One-year outcomes
One hundred and ninety-five (97.5%) participants completed one-year follow up. At one year, 20 (39%) in Cluster 3 transited to frailty or remained frail, compared to 5 (6.9%) in Cluster 2 and 2 (2.8%) in Cluster 1 (P<0.001) (Figure 3). In multiple linear regression analyses adjusted for age, gender, education level and baseline scores (Table 2), Cluster 3 was associated with significant declines in handgrip strength (β=−4.1, 95% CI −5.67- −2.52), gait speed (β=−0.08, 95% CI −0.16- −0.007), and RCST (β=1.4, 95% CI 0.27–2.53), using Cluster 1 as the reference group. Handgrip strength also showed significant decline in Cluster 2 (β=−1.5, 95% CI −2.8- −0.12), but the magnitude of decline is smaller compared to Cluster 3. For ADL function, Cluster 3 showed significant one-year decline (MBI, β=−1.2, 95% CI −2.5- −0.03). Quality-of-life also declined significantly in both Clusters 2 and 3, but at a greater magnitude in Cluster 3 (EQ-5D, Cluster 2: β=−0.034, 95% CI −0.07- −0.002; Cluster 3: β=−0.053, 95% CI −0.09- −0.02).
Figure 3.
Proportion of individuals within clusters transiting to frailty or remaining frail at one year
Table 2.
Multiple linear regression models examining the association between clusters and one-year change scores
| One-year change scores | Model 1* | Model 2** | ||||||
|---|---|---|---|---|---|---|---|---|
| Cluster 2 | Cluster 3 | Cluster 2 | Cluster 3 | |||||
| β (95% CI) | P | β (95% CI) | P | β (95% CI) | P | β (95% CI) | P | |
| Strength and physical performance | ||||||||
| Grip strength | −0.057 (−1.58–1.69) | 0.95 | −2.07 (−3.89- −0.25) | 0.026 | −1.46 (−2.80- −0.12) | 0.033 | −4.1 (−5.67- −2.52) | <0.001 |
| Gait speed | −0.032 (−0.09–0.03) | 0.3 | −0.14 (−0.22- −0.067) | <0.001 | −0.067 (−0.066–0.053) | 0.82 | −0.082 (−0.16- −0.007) | 0.033 |
| 5-times repeated chair stand | 0.37 (−0.53–1.27) | 0.42 | 1.67 (0.57–2.760 | 0.003 | 0.35 (−0.55–1.26) | 0.44 | 1.40 (0.27–2.53) | 0.015 |
| SPPB total score | −0.32 (−0.69–0.05) | 0.089 | −0.64 (−1.1- −0.17) | 0.007 | −0.28 (−0.64–0.074) | 0.12 | −0.42 (−0.88–0.036) | 0.071 |
| Function | ||||||||
| MBI | −0.28 (−1.32–0.77) | 0.60 | −1.51 (−2.66- −0.35) | 0.011 | −0.87 (−1.1–0.67) | 0.48 | −1.24 (−2.45- −0.03) | 0.045 |
| Quality of life | ||||||||
| EQ-5D utility scores | −0.025 (−0.056–0.006) | 0.11 | −0.042 (−0.077- −0.007) | 0.018 | −0.034 (−0.065- −0.002) | 0.035 | −0.053 (−0.091- −0.016) | 0.005 |
*Model 1: adjusted for baseline scores (reference group: Cluster 1); **Model 2: adjusted for baseline scores, age, gender and education level (reference group: Cluster 1); CMMSE: Modified version of the Chinese Mini-Mental State Examination; GDS: Geriatric Depression Scale; MNA: Mini-Nutritional Assessment; SPPB: Short Physical Performance Battery; FAI: Frenchay Activities index; MBI: Modified Barthel Index; EQ-5D: EuroQoL-5D
Discussion
This study applied cluster analysis to explore the relationship between frailty and IC in a cohort of relatively healthy, independent community-dwelling older adults. Three clusters were identified: high IC and robust (Cluster 1); intermediate IC and prefrail (Cluster 2); and low IC and prefrail/frail (Cluster 3). Amongst the three clusters, Cluster 3 is associated with the highest risk of poorest outcomes at one year, including the highest proportion transiting to frailty or remaining frail, declines in gait speed, handgrip strength and repeated chair stand time, decreased ADL function and quality-of-life at one year, with cluster 2 intermediate between clusters 1 and 3. Taken together, our results supports the complementarity between IC and frailty assessments and suggests that the integration of information about an individual's reserves and vulnerability allows better risk-profiling of longitudinal outcomes amongst prefrail individuals into intermediate and high-risk groups.
Very few studies have examined the relationship between IC and frailty and their mutual influence on clinical outcomes. In a study by Gutiérrez-Robledo et al. (21), multiple correspondence analysis revealed that frailty and prefrailty clustered with low IC and robustness with optimal IC, depending on the measure of IC used. However, important differences between the present and aforementioned study include the different tools used to operationalise frailty and IC, and the reporting of longitudinal outcomes in our study. Another approach to examine the relationship between IC and frailty is to study the effect of co-existent frailty and low IC on clinical outcomes. Declines in IC was associated with incremental risks of increased dependence and mortality in individuals who were frail and prefrail, compared to those who were robust (22). Regardless of the approach, both studies affirmed the value of measuring IC on top of characterizing frailty status to refine functional assessment and predict outcomes in older adults.
In our cluster analysis, total IC scores rank as the most important predictor of cluster membership, followed by frailty status and individual IC domains, underscoring the salience of the overall IC construct. However, consensus on a standardised measure of IC is still lacking. There is also a lack of agreement on how scores from the different domains should be computed to yield a representative composite index of IC (23). In our study, we operationalised IC using a mix of self-reported and performance-based tools that can be applied in clinical practice. Our method of estimating a composite IC score by ascribing scores to IC domains and summing the domain scores to yield a total IC score can be easily computed and is consistent with previous studies (13, 24). However, further validation studies on the operationalisation and scoring of IC are still required.
Our results demonstrate how adding IC measures to frailty status value-adds to profiling. Firstly, the addition of IC to frailty measures improved detection in the highest risk group to identify suitable targets for early intervention. Using the Fried frailty phenotype alone, only 7.5% of the cohort were classified as frail, compared with 26.5% in Cluster 3 (Prefrail/ frail with low IC). Importantly, prefrail individuals comprise 86% of non-robust individuals in our cohort, and constitute a significant proportion of community-dwelling older adults in frailty prevalence studies (25). Using IC, prefrail individuals can be further stratified into lower and higher risk groups (analogous to Clusters 2 and 3 respectively in our study), with potential targeted intervention based on impairments in IC domains.
While Cluster 3 (prefrail/frail with low IC) was associated with the poorest outcomes at one year, outcomes for Cluster 2 (prefrail individuals with intermediate levels of IC) were intermediate between Clusters 1 and 3. This is exemplified by significant declines in handgrip strength and quality-of-life at one year, though they experienced smaller declines compared to Cluster 3, indicating that Cluster 2 remains an at-risk group that warrants further attention and intervention.
The strengths of the present study include a well-characterized community-based population and comprehensive evaluation of frailty and IC in an Asian cohort, the operationalization of a measure of IC that can be readily applied in the clinical setting and the availability of longitudinal data to study associations with adverse outcomes. Nonetheless, a key limitation of this study is the lack of the sensory domain of IC. However, based on current data, total IC score and frailty status rank higher as clustering predictors compared to individual IC domains, which may have less impact on defining cluster membership. The generalisability of the present study is also limited by the study cohort, comprising mostly participants of Chinese ethnicity who were community-dwelling, cognitively intact and independent. The findings of this study may not be applicable to other settings with older adults who are more frail or of other ethnicities. Lastly, varying measures of IC across different studies, particularly in the vitality and psychological domains, limit the comparability of our findings with other studies.
Conclusion
Frailty is a well-established concept, with good evidence built over many years for the measurement and predictive validity of the construct. Our results suggest that IC may complement frailty measures in relatively healthy community-dwelling older adults and provide additional information beyond physical frailty for further risk stratification. Measuring IC may be especially relevant in older adults upstream in the trajectory of frailty, particularly in prefrail individuals who comprise a large proportion of community-dwelling older adults. Targeting higher-risk individuals based on IC scores and designing specific interventions for IC domains go beyond current dietary and exercise recommendations for physical frailty. Other advantages of a complementary approach to assessing frailty and IC include the “positive” connotation of the IC construct, given the stigma associated with the label of “frailty” (26). Adopting both IC and frailty measures at an individual and health system level allows the opportunity to pay more attention to the specific needs of older persons. Our findings provide the impetus for further studies to examine the validity of the current method of operationalising IC inclusive of the sensory domain, and longitudinal studies beyond a year to establish associations between IC and frailty with health outcomes.
Acknowledgments
We would like to thank all participants who contributed to this study. This study was supported by the Lee Foundation Grant 2013.
Conflict of interest
The authors have no conflicts of interest to disclose.
Ethical standards
All participants provided written informed consent prior to taking part in the research study.
References
- 1.WHO Decade of Healthy Ageing 2020–2030. https://www.who.int/ageing/decade-of-healthy-ageing. Accessed June 13, 2021.
- 2.Beard JR, Jotheeswaran AT, Cesari M, Araujo de Carvalho I. The structure and predictive value of intrinsic capacity in a longitudinal study of ageing. BMJ Open. 2019;9(11):e026119. doi: 10.1136/bmjopen-2018-026119. 10.1136/bmjopen-2018-026119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Dent E, Morley JE, Cruz-Jentoft AJ, et al. Physical Frailty: ICFSR International Clinical Practice Guidelines for Identification and Management. J Nutr Health Aging. 2019;23(9):771–787. doi: 10.1007/s12603-019-1273-z. 10.1007/s12603-019-1273-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Belloni G, Cesari M. Frailty and Intrinsic Capacity: Two Distinct but Related Constructs. Front Med. 2019;6:133. doi: 10.3389/fmed.2019.00133. 10.3389/fmed.2019.00133 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chew J, Tay L, Lim JP, et al. Serum Myostatin and IGF-1 as Gender-Specific Biomarkers of Frailty and Low Muscle Mass in Community-Dwelling Older Adults. J Nutr Health Aging. 2019;23(10):979–986. doi: 10.1007/s12603-019-1255-1. 10.1007/s12603-019-1255-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146–56. doi: 10.1093/gerona/56.3.m146. 10.1093/gerona/56.3.M146 Mar. [DOI] [PubMed] [Google Scholar]
- 7.Chen L-K, Woo J, Assantachai P, et al. Asian Working Group for Sarcopenia: 2019 Consensus Update on Sarcopenia Diagnosis and Treatment. J Am Med Dir Assoc. 2020;21(3):300–307.e2. doi: 10.1016/j.jamda.2019.12.012. 10.1016/j.jamda.2019.12.012 [DOI] [PubMed] [Google Scholar]
- 8.Holbrook M, Skilbeck CE. An activities index for use with stroke patients. Age Ageing. 1983;12(2):166–170. doi: 10.1093/ageing/12.2.166. 10.1093/ageing/12.2.166 [DOI] [PubMed] [Google Scholar]
- 9.Radloff LS. The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Appl Psychol Meas. 1977;1(3):385–401. 10.1177/014662167700100306 [Google Scholar]
- 10.Sahadevan S, Lim PP, Tan NJ, Chan SP. Diagnostic performance of two mental status tests in the older chinese: influence of education and age on cutoff values. Int J Geriatr Psychiatry. 2000;15(3):234–241. doi: 10.1002/(sici)1099-1166(200003)15:3<234::aid-gps99>3.0.co;2-g. 10.1002/(SICI)1099-1166(200003)15:3234::AID-GPS993.0.CO;2-G [DOI] [PubMed] [Google Scholar]
- 11.Yesavage JA, Brink TL, Rose TL, et al. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res. 1982;17(1):37–49. doi: 10.1016/0022-3956(82)90033-4. 10.1016/0022-3956(82)90033-4 [DOI] [PubMed] [Google Scholar]
- 12.Vellas B, Guigoz Y, Garry PJ, et al. The Mini Nutritional Assessment (MNA) and its use in grading the nutritional state of elderly patients. Nutr Burbank Los Angel Cty Calif. 1999;15(2):116–122. doi: 10.1016/s0899-9007(98)00171-3. 10.1016/S0899-9007(98)00171-3 [DOI] [PubMed] [Google Scholar]
- 13.Giudici KV, de Souto Barreto P, Beard J, et al. Effect of long-term omega-3 supplementation and a lifestyle multidomain intervention on intrinsic capacity among community-dwelling older adults: Secondary analysis of a randomized, placebo-controlled trial (MAPT study) Maturitas. 2020;141:39–45. doi: 10.1016/j.maturitas.2020.06.012. 10.1016/j.maturitas.2020.06.012 [DOI] [PubMed] [Google Scholar]
- 14.Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol. 1994;49(2):M85–94. doi: 10.1093/geronj/49.2.m85. 10.1093/geronj/49.2.M85 [DOI] [PubMed] [Google Scholar]
- 15.Mahoney FI, Barthel DW. Functional evaluation: The Barthel Index. Md State Med J. 1965;14:61–65. PubMed PMID: 14258950. [PubMed] [Google Scholar]
- 16.Abdin E, Subramaniam M, Vaingankar JA, Luo N, Chong SA. Measuring health-related quality of life among adults in Singapore: population norms for the EQ-5D. Qual Life Res Int J Qual Life Asp Treat Care Rehabil. 2013;22(10):2983–2991. doi: 10.1007/s11136-013-0405-x. 10.1007/s11136-013-0405-x [DOI] [PubMed] [Google Scholar]
- 17.Benassi M, Garofalo S, Ambrosini F, et al. Using Two-Step Cluster Analysis and Latent Class Cluster Analysis to Classify the Cognitive Heterogeneity of Cross-Diagnostic Psychiatric Inpatients. Front Psychol. 2020;11:1085. doi: 10.3389/fpsyg.2020.01085. 10.3389/fpsyg.2020.01085 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chiu T, Fang D, Chen J, Wang Y, Jeris C. A robust and scalable clustering algorithm for mixed type attributes in large database environment. In: KDD’ 01.; 2001. doi: 10.1145/502512.502549. [DOI]
- 19.Rousseeuw P. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. 1987. doi: 10.1016/0377-0427(87)90125-7. [DOI]
- 20.Finch CF, Stephan K, Shee AW, et al. Identifying clusters of falls-related hospital admissions to inform population targets for prioritising falls prevention programmes. Inj Prev. 2015;21(4):254–259. doi: 10.1136/injuryprev-2014-041351. 10.1136/injuryprev-2014-041351 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gutiérrez-Robledo LM, García-Chanes RE, González-Bautista E, Rosas-Carrasco O. Validation of Two Intrinsic Capacity Scales and Its Relationship with Frailty and Other Outcomes in Mexican Community-Dwelling Older Adults. J Nutr Health Aging. 2021;25(1):33–40. doi: 10.1007/s12603-020-1555-5. 10.1007/s12603-020-1555-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Prince MJ, Acosta D, Guerra M, et al. Intrinsic capacity as a framework for Integrated Care for Older People (ICOPE); insights from the 10/66 Dementia Research Group cohort studies in Latin America, India and China. medRxiv. September 2019:19006403. doi: 10.1101/19006403. [DOI]
- 23.George PP, Lun P, Ong SP, Lim WS. A Rapid Review of the Measurement of Intrinsic Capacity in Older Adults. J Nutr Health Aging. 2021;25(6):774–782. doi: 10.1007/s12603-021-1622-6. 10.1007/s12603-021-1622-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gutiérrez-Robledo LM, García-Chanes RE, Pérez-Zepeda MU. Allostatic Load as a Biological Substrate to Intrinsic Capacity: A Secondary Analysis of CRELES. J Nutr Health Aging. 2019;23(9):788–795. doi: 10.1007/s12603-019-1251-5. 10.1007/s12603-019-1251-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Collard RM, Boter H, Schoevers RA, Oude Voshaar RC. Prevalence of frailty in community-dwelling older persons: a systematic review. J Am Geriatr Soc. 2012;60(8):1487–1492. doi: 10.1111/j.1532-5415.2012.04054.x. 10.1111/j.1532-5415.2012.04054.x [DOI] [PubMed] [Google Scholar]
- 26.Woo J. Combating frailty and sarcopenia in aging populations: Switching to a more positive paradigm. Aging Med Milton NSW. 2019;2(1):7–10. doi: 10.1002/agm2.12052. 10.1002/agm2.12052 [DOI] [PMC free article] [PubMed] [Google Scholar]