Multi-trajectories of intrinsic capacity and their effect on higher-level functional capacity, life satisfaction, and self-esteem in community-dwelling older adults: the NILS-LSA

Highlights

• •Declining intrinsic capacity across domains challenges functional independence, life satisfaction, and self-esteem in older adults.
• •Four distinct patterns of intrinsic capacity decline were identified in Japanese community-dwelling older adults.
• •“Vision and cognitive decline” and “comprehensive deterioration” groups showed higher risk of decline in outcomes versus “healthy aging” group.
• •Findings support multidimensional interventions for specific IC decline patterns, prioritizing both physical and psychological well-being.

Abstract

Objectives

Variability in intrinsic capacity (IC) changes among community-dwelling older adults and their effect on health outcomes remain understudied. We examined the variability in IC trajectories and their impact on higher-level functional capacity (HLFC), life satisfaction, and self-esteem.

Design

Longitudinal study.

Setting

Data from the second to seventh waves (2000–2012) of the National Institute for Longevity Sciences–Longitudinal Study of Aging project.

Participants

934 community dwellers (aged ≥60).

Measurements

We used group-based multi-trajectory modeling to obtain IC trajectories across six domains: cognition, locomotion, vitality, vision, hearing, and psychological well-being. We employed multivariable regression to investigate the associations between IC trajectories and a decline in HLFC (assessed using the Tokyo Metropolitan Institute of Gerontology Index of Competence [TMIG-IC]; baseline TMIG-IC - follow-up TMIG-IC ≥ 2; logistic regression model), life satisfaction (assessed using the Life Satisfaction Index-K [LSI-K]; linear mixed model), and self-esteem (assessed using the Rosenberg Self-Esteem Scale [RSES]; linear mixed model).

Results

We identified four IC trajectories: the “healthy aging group” (63.7%), the “hearing decline group” (15.1%), the “vision and cognitive decline group” (12.7%), and the “comprehensive deterioration group” (8.5%). Compared to the healthy aging group, the vision and cognitive decline group and the comprehensive deterioration group displayed a significantly greater risk of a decline in the TMIG-IC score (multivariable-adjusted odds ratio [aOR], 95% confidence interval [CI] = 2.05 [1.11, 3.79], 2.74 [1.41, 5.30], respectively), the LSI-K score (multivariable-adjusted β [standard error] = −0.46 [0.08], −0.52 [0.10], respectively), and the RSES score (multivariable-adjusted β [standard error] = −0.85 [0.16], −0.66 [0.20], respectively). The “hearing decline group” did not show a significantly increased risk for these outcomes.

Conclusion

Older adults with different IC trajectories may differ in HLFC, life satisfaction, and self-esteem. Public health officials should be aware of this and provide targeted interventions.

1. Introduction

As populations age globally, there is a need for health promotion strategies focused on maintaining physical and mental functioning alongside preventing and treating age-related conditions. In 2017, the World Health Organization (WHO) introduced the concept of “intrinsic capacity” (IC), referring to the physical and mental capacities individuals retain as they age across six key domains: cognition, locomotion, vitality (nutritional status), vision, hearing, and psychological well-being (depressive symptoms) [1]. Rather than solely focusing on deficits, the concept of IC encourages a strength-based perspective that identifies and nurtures the residual capabilities of older individuals. Promoting IC can optimize functional autonomy, well-being, and quality of life, even in the presence of age-related changes or chronic conditions.

To date, several studies have investigated the trajectory (i.e., the nature of change) of overall IC [[2], [3], [4], [5], [6], [7]] and revealed variations in IC decline across different populations. However, the multi-trajectories of specific IC domains are rarely discussed. Only one previous study identified four subtypes of IC decline (i.e., robust with mild decline, hearing loss with cognitive decline, physio-cognitive decline with depression, and severe IC decline) among older adults in Taiwan, indicating that distinct subtypes of domain-specific IC decline have varying impacts on quality of life, fall risk, and functional limitations [8]. Nevertheless, the trajectories of individual IC domains are also hypothesized to differ across countries, which underscores the need for further research examining the multi-domain trajectories of IC in diverse populations.

Declining IC among older adults is linked to a broad spectrum of detrimental outcomes [9]. While the association between impairments in IC domains—such as vision [10] and vitality [11]—and greater decline in higher-level functional capacity (HLFC) is well-established, the relationship between IC and mental health in older populations warrants further investigation. Evidence suggests that life satisfaction, functional capacity [12,13], cognitive ability [14], and depression [14] are closely intertwined. Additionally, self-esteem can change in accordance with alterations in one’s capacities, such as locomotion [15], vision, hearing, and communication [16]. These findings collectively suggest that different trajectories of IC domains may have varying impacts on the mental health of older adults. However, this issue has not been adequately explored.

Utilizing longitudinal data spanning up to 11.4 years, we aimed to investigate the multi-trajectories of IC domains and their associations with changes in HLFC, life satisfaction, and self-esteem among a community-dwelling Japanese population. We hypothesized that distinct trajectories of IC domains would emerge, and these trajectories would differentially impact the aforementioned outcomes.

2. Material and methods

2.1. Study cohort

The National Institute for Longevity Sciences–Longitudinal Study of Aging (NILS-LSA) is a Japanese population-based prospective cohort study that explores variables related to normal aging and age-related diseases. Participants were age- and sex-stratified and selected through random sampling from Obu and Higashiura in Aichi Prefecture, Japan. The details of the NILS–LSA have been previously documented [17]. The initial investigation, conducted from November 1997 to April 2000, involved 2,267 participants aged 40–79. Subsequently, participants were followed up with every two years. If participants aged ≤79 years could not attend follow-up sessions, they were replaced by randomly recruited new participants and matched for age and sex. Additionally, participants aged 40 were recruited annually. The NILS–LSA adheres strictly to the principles outlined in the Declaration of Helsinki. The Committee on the Ethics of Human Research at the National Center for Geriatrics and Gerontology approved the study protocol (No. 1665-2), and all participants provided written informed consent for data collection and analysis before taking part.

In this study, participants for extracting IC trajectories were selected from the second (April 2000 to May 2002) to seventh (July 2010 to July 2012) waves of the NILS–LSA because all the variables used to summarize IC trajectories were only available in these waves. We included only data from participants aged 60 and older, as cognitive function assessments were exclusively conducted among individuals aged 60 and above (n = 2,210 out of a total of 3,635 participants). Furthermore, to ensure the robustness of our examination of IC trajectories, we further narrowed the analysis scope to include only participants who had been involved in three or more surveys (n = 1,308). Subsequently, individuals with missing data in any IC subdomain assessments were excluded (n = 79), as were those who had dementia and cancer at the time of their first survey participation (n = 87), and individuals with missing data for other covariates (n = 208). Finally, we used data from 934 participants (cumulative participation: 4149 times; aged 60–80 at the baseline survey; male: 49.5%; followed up for up to 11.4 years) for IC trajectory analysis.

After extracting the IC trajectories, we linked the data from the second to seventh waves regarding HLFC, life satisfaction, and self-esteem to these 934 participants. We ensured that each participant contributed data from at least two surveys (one for baseline and one for follow-up) on HLFC, life satisfaction, and self-esteem (median participation: 5 times; median follow-up duration: 8 years).

2.2. Assessment of intrinsic capacity

Referring to the WHO integrated care for older people (ICOPE) guidelines, we categorized IC into two groups (0 or 1 point) across various domains, including cognition, locomotion, vitality, vision, hearing, and psychological well-being [18]. As in our previous study [19], we assessed these six domains in the following ways. We evaluated cognitive function utilizing seven items from the Mini-Mental State Examination (MMSE) [20] (Japanese version [21,22]): four pertaining to time and place orientation and three concerning memory. We assigned a score of 0 to participants showing impairment in any of the seven MMSE items. We scored locomotion as 0 if the gait speed was less than 1 m/s. We assigned the following participants a score of 0 for vitality: individuals who lost ≥5% of their weight over two years or had a lack of appetite, assessed using the Center for Epidemiologic Studies Depression (CES-D) Scale (individuals who answered “one day or more” for the item “I did not feel like eating; my appetite was poor [during the past week]”) [23,24]. In terms of vision, we assigned participants who rated their visual acuity (with corrective lenses such as glasses or contact lenses) as “poor” or “very poor” a score of 0. Hearing capacity was evaluated by measuring air-conduction pure-tone thresholds for both ears, without the use of hearing aids, using diagnostic audiometers in a soundproof booth. We assigned a score of 0 if the average threshold level, at frequencies of 0.5, 1, 2, and 4 kHz, exceeded 35 dB in the better ear. Finally, the CES-D Scale was utilized to assess depressive symptoms within the domain of psychological well-being; we assigned 0 points to individuals with a score of ≥16, indicating the presence of significant depressive symptoms.

For each of the aforementioned domains, participants who were not assigned 0 points received 1 point, resulting in a total score range of 0–6 across the six domains.

2.3. Measurement of HLFC, life satisfaction, and self-esteem

Higher-level functional capacity was assessed using the Tokyo Metropolitan Institute of Gerontology Index of Competence (TMIG-IC; 13 items, possible total score range: 0–13) [25], which is widely used among community-dwelling older adults [25,26]. It is a multidimensional index of competence with the following three subscales (possible score range): instrumental activities of daily living (IADL) (0–5), effectance (intellectual activity) (0–4), and social role (0 − 4). A higher score reflects greater functional capability. We defined participants with a decrease of ≥2 points in the total TMIG-IC score from the baseline to the follow-up survey as experiencing a decline in HLFC [27]. Additionally, we considered a decrease of ≥1 point for IADL, ≥2 points for effectance, and ≥2 points for social role to indicate a decline in each of these subscales [28].

The Life Satisfaction Index-K (LSI-K; 9 items, possible total score range: 0–9) [29,30] was employed to measure life satisfaction. The LSI-K includes three subscales: satisfaction with life (4 items; e.g., “As you look back on your life, are you satisfied?”; possible score range: 0–4), psychological stability (3 items; e.g., “Do you think that living is extremely hard?”; possible score range: 0–3), and evaluation of aging (2 items; e.g., “Do you think you have become less useful as you have gotten older?”; possible score range: 0–2). A higher score indicates greater life satisfaction.

To estimate self-esteem, the Rosenberg Self-Esteem Scale (RSES) was utilized, consisting of 10 items with a possible total score range of 10–40 [[31], [32], [33]]. Additionally, the RSES was divided into two subscales: self-liking (5 items; e.g., “On the whole, I am satisfied with myself.”; possible score range: 5–20) and self-competence (5 items; e.g., “I feel that I have a number of good qualities”; possible score range: 5–20) [34,35]. A higher score indicates higher self-esteem.

2.4. Other measurements

Basic data regarding age, medical history of stroke, hypertension, heart disease, dyslipidemia, diabetes, and osteoporosis (yes/no), along with smoking habits (current, former, never), alcohol consumption (current, former, never), years of education, marital status (never married, married, living apart, divorced, bereaved), and living arrangements (alone, with others) were gathered through self-report questionnaires.

Physical activity over 24 h was evaluated through the metabolic equivalent of task scores (METs-hours/day) acquired from participant interviews. Trained interviewers used a semi-quantitative assessment method [36].

Height and weight were measured to the nearest 0.1 cm and 0.1 kg, respectively. Body mass index (BMI) was calculated by dividing weight (kg) by height squared (m²) in the baseline survey. Participants wore light clothing and no shoes, and measurements were taken during a fasting state around 9–10 a.m. Appendicular skeletal muscle mass (kg) was assessed using dual-energy X-ray absorptiometry (DXA; QDR-4500; Hologic, Bedford, MA, USA). Relative appendicular skeletal muscle (kg/m²) was determined by dividing appendicular skeletal muscle mass (kg) by height squared (m²).

2.5. Statistical analysis

We used group-based multi-trajectory modeling [37] to derive the IC trajectories, wherein we assigned each IC domain a score of 0 or 1 and used the number of survey visits as the time variable. Determination of the number of groups and the shape of each domain within them involved factors such as the Bayesian information criterion value, group membership probability, group membership posterior probability, and interpretability.

We evaluated mean differences for continuous variables using the Kruskal-Wallis rank sum test, while we examined proportions in disparities for categorical variables via Pearson’s chi-square test or Fisher’s exact test. In cases where we noted distribution differences among multiple groups, we applied Dunn’s test to the continuous variables, with P-values adjusted using the Bonferroni method. We employed Fisher’s exact test for categorical variables, with P-values adjusted using the Benjamini-Hochberg method for every pairwise comparison.

We estimated cross-sectional associations between IC trajectories and total scores and subscales of the TMIG-IC, LSI-K, and RSES at baseline and the last follow-up using general linear models. We adjusted these models for baseline information on age (in years), sex, medical history (stroke, hypertension, heart disease, dyslipidemia, diabetes, osteoporosis; categorized as yes or no for each item), smoking habits (current or not), alcohol consumption (current or not), total physical activity (METs-hours/day), years of education (≤9, 10–12, and ≥13 years), marital status (married or not), and living arrangements (alone, with others). Additionally, we investigated longitudinal associations between IC trajectories and the declines in the TMIG-IC total score and subscale score using logistic regression models. We explored the longitudinal associations between IC trajectories and score changes (the slopes) using linear mixed models for the LSI-K and RSES, with at least two datasets from baseline to the last follow-up survey. We adjusted all logistic regression models and linear mixed models for the same variables used in the general linear models, and we further adjusted them for follow-up duration (in years) and the corresponding score at baseline.

The statistical analyses outlined were two-sided, with the precise p-value reported. We deemed a P-value < 0.05 to be significant. We conducted group-based multi-trajectory modeling using Statistical Analysis System version 9.3 (SAS Institute, Inc). We conducted other analyses using R Statistical Software (version 4.3.2; R Core Team, 2023).

3. Results

3.1. IC trajectories

Supplementary Table S1 depicts the baseline characteristics of all participants. Four distinct trajectories emerged from the analysis, designated as follows: the “healthy aging group” (63.7%), the “hearing decline group” (15.1%), the “vision and cognitive decline group” (12.7%), and the “comprehensive deterioration group” (8.5%) (see Supplementary Table S2 and Fig. 1 for grouping details). For subsequent analysis, the healthy aging group served as the reference group.

Fig. 1.

Trajectories of intrinsic capacity derived by the group-based multi-trajectory modeling (n = 934). In the “vision and cognitive decline group,” the change curves for “locomotion” and “cognition” overlap.

3.2. Baseline characteristics of the participants

At baseline, the mean (standard deviation [SD]) age of all participants was 66.1 (5.9) years, with 49.5% being male. There were differences among the four IC trajectories in terms of mean age (higher in the hearing decline and comprehensive deterioration groups), total physical activity level (lower in the comprehensive deterioration group compared to the healthy aging group), total IC scores (decreasing from healthy aging to comprehensive deterioration group), gender, current smoker status, years of education (lower in the hearing decline and comprehensive deterioration groups), and medical history (higher prevalence of heart disease in the comprehensive deterioration group and dyslipidemia in the vision and cognitive decline and comprehensive deterioration groups). Each IC domain showed significant group differences except for cognition, with the healthy aging group generally showing the least impairment and the comprehensive deterioration group showing the most (Table 1).

Table 1.

Baseline characteristics of participants with different intrinsic capacity trajectories (n = 934).

		Four intrinsic capacity trajectories
		Healthy aging group	Hearing decline group	Vision and cognitive decline group	Comprehensive deterioration group	P-value a
No. of participants	n	603	144	111	76
Age (year)	mean (SD)	64.7 (5.2) b, d	69.8 (5.9) b, c	65.4 (5.9) c, e	71.0 (5.3) d, e	<0.001
Male	%	47.4	59.7	44.1	53.9	0.030
Body mass index (kg/m2)	mean (SD)	23.1 (2.8)	22.8 (2.7)	22.9 (2.7)	22.9 (3.1)	0.829
Relative skeletal muscle mass index (kg/m2)	mean (SD)	6.7 (1.0)	6.7 (0.9)	6.6 (0.9)	6.6 (0.9)	0.862
Current smoker	%	13.6	18.8	17.1	25.0	0.043
Current drinker	%	54.4	58.3	50.5	44.7	0.232
Total physical activities (METs-hours/day)	mean (SD)	32.2 (3.0) d	32.1 (3.5)	32.3 (3.5)	31.3 (3.3) d	0.031
Years of education
≤9	%	28.4 b, d	44.4 b, c	27.9 c, e	47.4 d, e	<0.001
≥10	%	71.6	55.6	72.1	52.6
Married	%	85.7	81.2	82.9	77.6	0.206
Living alone	%	6.8	12.5	11.7	10.5	0.071
Medical history
Stroke	%	3.3	3.5	4.5	3.9	0.871
Hypertension	%	32.3	36.1	33.3	43.4	0.255
Heart disease	%	4.6 d	6.2	6.3	17.1 d	0.002
Dyslipidemia	%	23.4	13.9 c, f	29.7 c	30.3 f	0.008
Diabetes	%	9.5	9.0	7.2	9.2	0.903
Osteoporosis	%	6.8	8.3	8.1	5.3	0.807
Intrinsic capacity (without impairment)
Cognition	%	87.6	87.5	91.0	93.4	0.372
Locomotion	%	98.2 d	98.6 f	100.0 e	80.3 d, e, f	< 0.001
Vitality	%	87.7 d, g	83.3 c, f	55.0 c, g	63.2 d, f	< 0.001
Vision	%	90.7 b, g	81.9 b	76.6 g	81.6	< 0.001
Hearing	%	100.0 b, d	49.3 b, c	100.0 c, e	59.2 d, e	< 0.001
Psychological well-being	%	98.3 d, g	97.9 c, f	55.9 c, g	53.9 d, f	< 0.001
Total intrinsic capacity score	mean (SD)	5.6 (0.6) b, d, g	5.0 (0.9) b, f	4.8 (0.9) g	4.3 (1.1) d, f	< 0.001

^aFor continuous variables, the Kruskal–Wallis rank sum test was used. For categorical variables, the Pearson’s chi-squared test or the Fisher’s exact test was used.

^b,c,d,e,f,gMultiple comparisons were conducted using Dunn’s test for continuous variables (P-values were adjusted using the Bonferroni method), while pairwise comparisons were performed using Fisher’s exact test for categorical variables (P-values were adjusted using the Benjamini-Hochberg method). The same characters indicated significant differences between groups.

3.3. Associations between the IC trajectories and HLFC

The mean (SD) follow-up duration for all participants was 7.2 (2.4) years. Table 2 and Supplementary Table S3 show the associations between the IC trajectories and HLFC, as well as the associations between IC trajectories and the subscales of HLFC, respectively. At both baseline and the final follow-up, individuals in the vision and cognitive decline group and those in the comprehensive deterioration group exhibited significantly lower TMIG-IC total scores and scores on the social role subscale compared to participants in the healthy aging group. Moreover, individuals in the comprehensive deterioration group demonstrated lower scores on the IADL subscale at both baseline and follow-up and a lower score on the effectance subscale at follow-up. Similarly, participants in the vision and cognitive decline group showed a reduced score on the effectance subscale at both baseline and follow-up.

Table 2.

Association between intrinsic capacity trajectories and Tokyo Metropolitan Institute of Gerontology Index of Competence (TMIG-IC) (n = 934).

Baseline a,c	Estimate	Std. Error	P-value	Follow-up a,c	Estimate	Std. Error	P-value	Change during follow-up b,d	OR	95% CI	P-value
Total score [Healthy aging group (mean ± SE): 12.57 ± 0.16]				Total score [Healthy aging group (mean ± SE): 12.34 ± 0.19]				Total score decreased ≥2
Hearing decline group	−0.07	0.11	0.528	Hearing decline group	−0.09	0.13	0.476	Hearing decline group	1.32	(0.72, 2.41)	0.367
Vision and cognitive decline group	−0.54	0.11	<0.001	Vision and cognitive decline group	−0.69	0.13	<0.001	Vision and cognitive decline group	2.05	(1.11, 3.79)	0.022
Comprehensive deterioration group	−0.53	0.14	<0.001	Comprehensive deterioration group	−1.07	0.16	<0.001	Comprehensive deterioration group	2.74	(1.41, 5.30)	0.003

^aAnalyzed using the general linear model.

^bAnalyzed using the logistic regression model.

^cAdjusted for baseline information on age (years), sex, medical history (stroke, hypertension, heart disease, dyslipidemia, diabetes, osteoporosis; yes or no for each item), smoking habits (current or not), alcohol consumption (current or not), total physical activity (METs-hours/day), years of education (≤9, 10–12, and ≥13 years), marital status (married or not), and living arrangements (alone, with others).

^dAdjusted for baseline information on age (years), sex, medical history (stroke, hypertension, heart disease, dyslipidemia, diabetes, osteoporosis; yes or no for each item), smoking habits (current or not), alcohol consumption (current or not), total physical activity (METs-hours/day), years of education (≤9, 10–12, and ≥13 years), marital status (married or not), living arrangements (alone, with others), follow-up duration (years), and corresponding score at baseline.

However, upon examining changes over follow-up periods, it was evident that participants in the vision and cognitive decline group faced a greater risk of decline in the TMIG-IC total score. Likewise, those in the comprehensive deterioration group were more prone to experiencing declines in the TMIG-IC total score and the effectance and social role scores.

3.4. Associations between the IC trajectories and life satisfaction

As for life satisfaction, individuals in the vision and cognitive decline group and the comprehensive deterioration group displayed lower total scores on the LSI-K and lower scores across all three subscales than those in the healthy aging group, both at baseline and during the last follow-up assessment. Additionally, they experienced greater score declines throughout the follow-up period (Table 3). Moreover, apart from declines in psychological stability, participants in the comprehensive deterioration group experienced the greatest declines in scores (both total and subscale) among the four trajectories. Supplementary Table S4 portrays the association between the IC trajectories and the subscales of the LSI-K.

Table 3.

Association between intrinsic capacity trajectories and Life Satisfaction Index-K (LSI-K) (n = 934).

Baseline a,c	Estimate	Std. Error	P-value	Follow-up a,c	Estimate	Std. Error	P-value	Change during follow-up b,d	Estimate	Std. Error	P-value
Total score [Healthy aging group (mean ± SE): 4.87 ± 0.27]				Total score [Healthy aging group (mean ± SE): 4.84 ± 0.30]				Total score
Hearing decline group	−0.05	0.18	0.772	Hearing decline group	0.32	0.20	0.111	Hearing decline group	−0.06	0.07	0.395
Vision and cognitive decline group	−1.96	0.19	<0.001	Vision and cognitive decline group	−2.22	0.21	<0.001	Vision and cognitive decline group	−0.46	0.08	<0.001
Comprehensive deterioration group	−2.25	0.24	<0.001	Comprehensive deterioration group	−2.49	0.26	<0.001	Comprehensive deterioration group	−0.52	0.10	<0.001

^aAnalyzed using the general linear model.

^bAnalyzed using the linear mixed model.

^cAdjusted for baseline information on age (years), sex, medical history (stroke, hypertension, heart disease, dyslipidemia, diabetes, osteoporosis; yes or no for each item), smoking habits (current or not), alcohol consumption (current or not), total physical activity (METs-hours/day), years of education (≤9, 10–12, and ≥13 years), marital status (married or not), and living arrangements (alone, with others).

^dAdjusted for baseline information on age (years), sex, medical history (stroke, hypertension, heart disease, dyslipidemia, diabetes, osteoporosis; yes or no for each item), smoking habits (current or not), alcohol consumption (current or not), total physical activity (METs-hours/day), years of education (≤9, 10–12, and ≥13 years), marital status (married or not), living arrangements (alone, with others), follow-up duration (years), and corresponding score at baseline.

3.5. Associations between the IC trajectories and self-esteem

Table 4 and Supplementary Table S5 show the associations between the IC trajectories, self-esteem, and self-esteem subscales. Again, participants in the vision and cognitive decline group and the comprehensive deterioration group had lower total RSES scores and scores on the two subscales compared to those in the healthy aging group, both at baseline and final follow-up. Further, they demonstrated more significant declines in all scores throughout the follow-up duration. However, participants in the vision and cognitive decline group exhibited the greatest score decline among all IC trajectories.

Table 4.

Association between intrinsic capacity trajectories and Rosenberg Self-Esteem Scale (RSES) (n = 934).

Baseline a,c	Estimate	Std. Error	P-value	Follow-up a,c	Estimate	Std. Error	P-value	Change during follow-up b,d	Estimate	Std. Error	P-value
Total score [Healthy aging group (mean ± SE): 29.32 ± 0.61]				Total score [Healthy aging group (mean ± SE): 29.46 ± 0.62]				Total score
Hearing decline group	−0.27	0.41	0.507	Hearing decline group	−0.33	0.42	0.437	Hearing decline group	−0.07	0.15	0.623
Vision and cognitive decline group	−3.45	0.43	<0.001	Vision and cognitive decline group	−3.73	0.44	<0.001	Vision and cognitive decline group	−0.85	0.16	<0.001
Comprehensive deterioration group	−4.27	0.54	<0.001	Comprehensive deterioration group	−4.65	0.55	<0.001	Comprehensive deterioration group	−0.66	0.20	0.001

^aAnalyzed using the general linear model.

^bAnalyzed using the linear mixed model.

^cAdjusted for baseline information on age (years), sex, medical history (stroke, hypertension, heart disease, dyslipidemia, diabetes, osteoporosis; yes or no for each item), smoking habits (current or not), alcohol consumption (current or not), total physical activity (METs-hours/day), years of education (≤9, 10–12, and ≥13 years), marital status (married or not), and living arrangements (alone, with others).

^dAdjusted for baseline information on age (years), sex, medical history (stroke, hypertension, heart disease, dyslipidemia, diabetes, osteoporosis; yes or no for each item), smoking habits (current or not), alcohol consumption (current or not), total physical activity (METs-hours/day), years of education (≤9, 10–12, and ≥13 years), marital status (married or not), living arrangements (alone, with others), follow-up duration (years), and corresponding score at baseline.

4. Discussion

In this study, we analyzed different IC trajectories and their impacts on changes in HLFC, life satisfaction, and self-esteem using longitudinal cohort data spanning up to 11.4 years among older individuals in the community. We identified four distinct IC trajectories: the healthy aging group, the hearing decline group, the vision and cognitive decline group, and the comprehensive deterioration group. Participants in the vision and cognitive decline group and the comprehensive deterioration group showed lower levels of HLFC, life satisfaction, and self-esteem than those in the healthy aging group, both at baseline and follow-up. They also faced a higher risk of declining HLFC and experiencing greater reductions in life satisfaction and self-esteem during follow-up. To the best of our knowledge, this study is the first to explore the associations between IC trajectories and changes in HLFC, life satisfaction, and self-esteem.

Some IC trajectories identified in the present study are similar to those derived in previous studies. Three studies conducted on older Chinese people [[2], [3], [4]] and two on Mexican middle-aged and older adults [5,7] identified three IC trajectories, two of which are akin to the healthy aging group and comprehensive deterioration group in our analysis. Another study conducted in China identified four trajectories over three years; among them, one is similar to our healthy aging group [38]. However, these studies only used a total IC score to summarize trajectories, making identifying different trends in domain changes impossible. One study investigated IC domain changes in Chinese community-dwelling older adults, but the number of participants was very small (n = 196), and they were only followed up only once (2 years later) (6); thus, information on long-term IC changes could not be obtained. The only study that discussed IC domain changes in a long-term setting involved Taiwan’s community-dwelling older adults [8]; four trajectories were identified and characterized as “robust with mild decline” (similar to the healthy aging group in our study), “hearing loss with cognitive decline,” “physio-cognitive decline with depression,” and “severe IC decline” (similar to the comprehensive deterioration group in our study), which is partially consistent with our findings.

The distinct trajectories of IC decline identified in our study may reflect different underlying biological mechanisms of aging. The healthy aging group likely represents successful biological aging with well-preserved homeostatic mechanisms and cellular repair processes. The hearing decline group’s relatively isolated sensory impairment may primarily reflect age-related cochlear degeneration and changes in the auditory nervous system, for which noise is the most-studied and best-documented environmental factor [39]. The post-analysis of noise exposure patterns across groups revealed that participants in the comprehensive deterioration group and hearing decline group had significantly higher proportions of current or past exposure to occupational or residential noise compared to those in the vision and cognitive decline group and healthy aging group (50.0%, 44.8%, 29.4%, and 29.3%, respectively). This differential exposure to noise environments in work or living settings suggests a potential environmental contribution to the observed trajectory patterns. The vision and cognitive decline group’s pattern might stem from shared pathophysiological mechanisms between visual and cognitive impairment, potentially involving microvascular changes, neurodegeneration, or inflammatory processes that affect both sensory and cognitive neural circuits [40]. The comprehensive deterioration group likely represents accelerated biological aging affecting multiple systems simultaneously, possibly due to increased oxidative stress, chronic inflammation, mitochondrial dysfunction, and accumulated cellular damage—hallmarks of accelerated aging that can impact multiple organ systems concurrently. Additionally, the presence of comorbidities in this group, such as cardiovascular conditions, may create cascading effects that accelerate functional decline across multiple domains through shared pathophysiological pathways. However, more research is needed to fully elucidate these biological mechanisms and their interactions.

Increasing evidence has indicated that IC impairment is associated with a higher risk of IADL decline [2,8,38,[41], [42], [43]]. However, the association between IC trajectories and life satisfaction has rarely been investigated. Older adults with visual impairment reported a lower level of life satisfaction than their sighted peers [[44], [45], [46]]. Meanwhile, a systematic review suggests that visual impairment is associated with cognitive impairment (which is also positively associated with a decline in life satisfaction [47]) in the older adult population [48]. This may explain why participants in the vision and cognitive decline group showed a higher decline in life satisfaction during follow-up. Moreover, impaired psychological well-being (i.e., the presence of depressive symptoms [49]) and impaired vitality (i.e., being malnourished [50, 51]) since the baseline may have also contributed to the decline in life satisfaction in the vision and cognitive decline group. Furthermore, participants in the comprehensive deterioration group exhibited the largest decline in life satisfaction, which may be due to the additional adverse effects of locomotion impairment [52,53] and hearing loss [54].

We also observed that compared with participants in the healthy aging group, participants in the vision and cognitive decline group, as well as the comprehensive deterioration group, were more likely to experience a decline in self-esteem. While vision [16,55,56] and hearing [16] impairments have been reported as risk factors in declining self-esteem in older individuals, cognitive function alone might not strongly influence self-esteem [57]. The causal relationships between physical function (or physical activity) and self-esteem [[58], [59], [60], [61]], as well as between depressive symptoms and self-esteem [[62], [63], [64]], are controversial. Additionally, in our findings, the decline in self-esteem in the vision and cognitive decline group was larger than in the comprehensive deterioration group. We assume the more significant vision-based decline in the vision and cognitive decline group may be a major contributing factor. Visual functioning is more closely associated with social relationships [65], which interact with self-esteem across the lifespan [66]. Moreover, individuals in the vision and cognitive decline group tended to be younger than those in the comprehensive deterioration group, suggesting that vision decline may have a stronger effect in the younger population [67].

Notably, we found no effects of the hearing decline group on any outcome compared to the healthy aging group, although hearing loss has been associated with a decline in IADL [68,69] and decreased self-esteem [70] in previous studies. Because hearing aids may improve IADL [71] and quality of life (including self-esteem) [72] among older adults, we checked the usage of hearing aids among participants with hearing decline in each group at the last follow-up survey. We found that the percentage of frequent/occasional use of hearing aids among people who are hard of hearing was much higher in the hearing decline group (39/139 = 28.1%) than in the healthy aging group (3/53 = 5.7%). Our findings indicate that following declines in IC, proficient self-management of health conditions—coupled with assistive environmental facilities and equipment—can help preserve functional abilities in the older adult population.

Despite this being the first study to investigate multiple trajectories of IC and their effects on HLFC, life satisfaction, and self-esteem among community-dwelling older adults, several limitations should be noted. First, as the binary classification of IC domains was based on the WHO guidelines—which aim to screen for the high-risk population with IC impairment in older adults—the severity of impairment in the IC domains and their impact on outcomes has not been fully examined. Future studies incorporating both categorical and continuous measures could better capture the gradual progression of impairments and potentially identify earlier intervention points. Additionally, age-specific thresholds might better reflect the heterogeneity of aging trajectories. Second, although we performed longitudinal analyses, as IC and HLFC, life satisfaction, and self-esteem all changed during the follow-up duration, a causal relationship could not be established. We acknowledge that the temporal relationship between IC changes and functional/psychological outcomes remains unclear in our study, as all variables were measured simultaneously during follow-up. The relationships between IC and these outcomes may be bidirectional - declining IC could lead to reduced functional capacity and psychological well-being, while poor psychological status and functional limitations might also accelerate IC deterioration. Future studies employing more sophisticated analytical approaches with more frequent assessment points would be valuable for disentangling these complex bidirectional relationships. Additionally, intervention studies targeting specific IC domains could help establish causal relationships by examining whether improving IC leads to better functional and psychological outcomes. Third, we only controlled for baseline characteristics, while many potential confounding factors likely changed during the follow-up period. While our data revealed that lower educational level was associated with poorer IC trajectories (44.4% and 47.4% of participants in the hearing decline and comprehensive deterioration groups had ≤9 years of education, compared to 28.4% and 27.9% in the healthy aging and vision and cognitive decline groups), we were unable to fully explore how these social determinants might mediate the relationship between IC trajectories and our outcomes of interest. Education level and other social determinants such as economic status, access to healthcare, and social support networks could mediate or moderate the impact of IC decline on functional capacity and psychological well-being. A sophisticated path analysis incorporating these social determinants as potential mediators would be valuable for understanding these complex relationships. Additionally, the development of new medical conditions, changes in physical activity levels, or alterations in living arrangements during follow-up could have influenced both IC trajectories and our outcomes. While our dataset includes these time-varying measurements, incorporating them into the analysis presents methodological challenges due to their potential role as both confounders and mediators in the causal pathway between IC and outcomes. Future studies with larger sample sizes might consider using more advanced statistical methods, such as marginal structural models, path analysis, or time-varying effect models, to better account for the mediating effects of social determinants while avoiding potential overadjustment for mediating factors. Fourth, although we adjusted for many confounding factors, we cannot rule out the potential influence of residual confounding factors, such as economic status [73] and personalities [74,75], on the results. Finally, our comparative analysis revealed that excluded participants were generally older, less physically active, and had a higher prevalence of chronic conditions compared to the included participants. They also showed higher proportions of impairment across most IC domains and lower total IC scores (Supplementary Table S6). This suggests that our findings may primarily reflect IC trajectories among relatively healthier and more engaged older adults and may not be generalizable to those with worse health conditions or functional limitations. Future studies with more complete data are needed to better understand IC trajectories across the full spectrum of health status in older adults.

5. Conclusion

By analyzing long-term longitudinal follow-up data, we summarized four different trajectories of IC and found that compared to older adults experiencing healthy aging, those who exhibited visual impairment and cognitive decline during the aging process—as well as those who experienced decline across all six domains of vision, hearing, locomotion, cognition, vitality, and psychological well-being—showed greater declines in HLFC, life satisfaction, and self-esteem. Our findings suggest that when screening for and intervening in IC impairments among community-dwelling older adults, attention should be paid to their psychological and mental health to comprehensively maintain and improve their functional abilities and promote healthy aging in society.

6. Author contributions

Shu Zhang: Conceptualization, Methodology, Software, Formal analysis, Writing- Original draft preparation, Funding acquisition. Chikako Tange: Data curation, Methodology, Software, Formal analysis. Shih-Tsung Huang: Methodology, Software, Formal analysis. Sayaka Kubota: Writing- Reviewing and Editing. Hiroshi Shimokata: Data curation, Writing- Reviewing and Editing. Yukiko Nishita: Data curation, Writing- Reviewing and Editing. Rei Otsuka: Supervision, Project administration, Writing- Reviewing and Editing, Funding acquisition.

CRediT authorship contribution statement

All authors have made substantial contributions to all of the following: [1] the conception and design of the study, or acquisition of data, or analysis and interpretation of data [2], drafting the article or revising it critically for important intellectual content [3], final approval of the version to be submitted.

Declaration of Generative AI and AI-assisted technologies in the writing process

AI-assisted technologies were only used for grammar checking in the writing process.

Funding

This work was partly supported by the HORI Sciences and Arts Foundation, the National Center for Geriatrics and Gerontology, Japan (grant number 21–18, 24–10), Japan Agency for Medical Research and Development (AMED) (grant number 24dk0110050h0001), and the Food Science Institute Foundation. The funding sources had no role in the study design, data collection, analysis, interpretation, writing of the report, or the decision to submit the paper.

Data availability

The data of NILS-LSA analyzed in the current study are not publicly available for privacy reasons but are available from the corresponding author upon reasonable request.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is Supplementary data to this article: