Biopsychosocial frailty and the incident risk of cardio- and cerebrovascular diseases: an 8-year follow-up of the italian longitudinal study on aging

Abstract

Biopsychosocial frailty, a new construct combining modifiable physical, psychological, and social domains, was associated with increased risk of dementia, mild cognitive impairment, and their subtypes. The aim of the present study was to estimate the role of biopsychosocial frailty on the incident risk of cardio- and cerebrovascular disease, i.e., myocardial infarction, peripheral artery disease (PAD), and definite stroke. From the eight municipalities of the population-based Italian Longitudinal Study on Aging, we examined 2410 older individuals in a longitudinal analysis with an 8-year follow-up. Biopsychosocial frailty operationalization was based on the results of a previous comprehensive geriatric assessment and the presence of physical frailty. Generalized estimating equation models assessed the independent contribution at three different follow-ups of baseline individuals with biopsychosocial frailty on occurring new events of myocardial infarction, PAD, and stroke. The prevalence of biopsychosocial frailty in older individuals with myocardial infarction was 10.9%, in older subjects with PAD was 8.9%, and in older individuals with definite stroke was 23.7%. After 8 years of follow-up, participants with biopsychosocial frailty showed an increased odds ratio (OR) of definite stroke [OR 2.67; 95% confidence interval 1.11–6.40]. No statistically significant associations between biopsychosocial frailty and myocardial infarction or PAD were observed. Over an 8-year follow-up, in a large, older cohort, a biopsychosocial frailty phenotype was associated with incident cerebrovascular disease, in particular with definite stroke. This complex phenotype of frailty has several modifiable components and could be an optimal target for potential interventions and prevention strategies.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11357-026-02135-9.

Introduction

After over two decades of clinical research, although a consensus based on a conceptual framework of frailty exists, a shared and unifying operational definition has not completely emerged. At present, over 70 different operational definitions are employed [1]. The two most recognized frailty models are the physical frailty phenotype, a unidimensional definition [2, 3], including items as unintentional weight loss, poor physical activity, slow gait, weakness, and exhaustion, and the multidimensional frailty model, an age-related deficit accumulation index built on the presence of signs, symptoms, diseases, and disabilities [4]. More recently, complex and multidimensional frailty phenotypes included widely defined and validated constructs such as nutritional frailty [5], oral frailty [6], cognitive frailty [7], social frailty [8], psychological frailty, together with the recently proposed depressive frailty phenotype [1], and biopsychosocial frailty [9].

The original conceptual landscape of frailty has evolved into a complex, multidimensional biopsychosocial syndrome including physical, cognitive, social, and psychological domains [1, 10]. The biopsychosocial frailty construct, according to the age-related deficit accumulation model and the comprehensive geriatric assessment (CGA) [11], is a status of increased biological aging, encompassing molecular, cellular, and intercellular age-related changes [12], considering the integral functioning of older individuals, and covering frailty factors in physical, social, and psychological dimensions. All frailty complex phenotypes include specific and age-associated health deficits, all related to adverse health-related outcomes, i.e., disability, institutionalization, more frequent hospitalization, mortality, and dementia [13, 14].

In particular, for late-life cognitive disorders and dementia, the 2017 Lancet Commission on dementia [15] and the 2024 update [16] reviewed risk factors associated with dementia, identifying, among others, factors like physical inactivity and little social contact in later life. These modifiable dementia risk factors may be encompassed in this recently operationalized biopsychosocial frailty phenotype [9]. The Italian Longitudinal Study on Aging (ILSA) suggested that participants with the biopsychosocial frailty phenotype showed an increased risk of all-cause dementia, in particular vascular dementia (VaD), over 3.5- and 7-year follow-ups [9]. In a recent meta-analysis and systematic review of six longitudinal population-based studies (219,616 subjects), the biopsychosocial frailty phenotype predicted a 41% higher risk of late-life cognitive decline and 53% higher risk of all-cause dementia, and also indicated an 11% higher risk of Alzheimer’s disease (AD) [14]. Moreover, in a recent study from the population-based Italian PRoject on the Epidemiology of Alzheimer’s disease (IPREA), the biopsychosocial frailty phenotype was associated with the mild cognitive impairment (MCI) and particularly with the non-amnestic MCI single and multiple domains subtypes [17]. Finally, other very recent findings from the IPREA showed that the biopsychosocial frailty construct may be associated with all-cause dementia, probable and possible VaD, and probable Alzheimer’s disease (AD) [18].

Therefore, these findings from population-based studies [9, 17, 18] and pooled evidence from a large meta-analysis [14] suggested that this phenotype of frailty, which has several modifiable components, could be an optimal target for potential interventions and prevention strategies targeting the different dimensions of frailty to delay cognitive decline and dementia. Major noncommunicable age-related diseases such as cancer, coronary heart disease, dementia, diabetes mellitus, and stroke may share common risk factors [19]. Their prevention implies not only the treatment of modifiable risk factors, but also biological ageing and its mechanisms, and the degree of frailty is considered a multidimensional phenotype [20]. In the large cohort of the ILSA, over an 8-year follow-up, the objective of the present study was to estimate the role of biopsychosocial frailty on the incident risk of cardio- and cerebrovascular disease, i.e., myocardial infarction, peripheral artery disease (PAD), and definite stroke.

Material and methods

Setting

In the ILSA, a sample of 5632 individuals aged 65–84 years, independent or institutionalized, was randomly selected from the electoral rolls of eight Italian municipalities, after stratification for age and sex [21]. The data of the present study have been obtained three wave periods: during the first survey study between March 1992 and June 1993, the second survey study between September 1995 and October 1996, and the third survey study between March 2000 and September 2001. The study project was approved by the institutional review board of the eight municipalities of the ILSA and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting cohort studies (Supplementary Table S1). Informed consent was obtained from each participant and/or their relatives before enrollment.

Participants, clinical examination, and laboratory analyses

Eligibility criteria for the participants to the present study included the presence in the three waves of data on the tools used for diagnosis and assessment of the domains of the biopsychosocial frailty construct, socio-demographic and clinical characteristics, and the cardio- and cerebrovascular outcomes. Cases of coronary artery disease (CAD, myocardial infarction and angina pectoris), congestive heart failure, type-2 diabetes mellitus, hypertension, and cerebrovascular disease (definite stroke and possible cerebrovascular events) were identified with a two-phase procedure, using clinical criteria described in detail elsewhere [21–23]. Based on self-reports, smoking habits were categorized as “ever” or “never,” and the variable “pack-years cigarettes” (number of years smoked × [number of cigarettes smoked per day/20 cigarettes per pack]) was used to represent the total smoking exposure. Body mass index was calculated as weight/height² (kg/m²). Functional status was assessed using the activities of daily living (ADL) scale [24]. The ability of home management was assessed using the Instrumental Activities of Daily Living (IADL) scale [25]. Mini-Mental State Examination (MMSE) was used to evaluate global cognitive function [26]. Depressive symptoms were investigated using the Italian version of the GDS-30 [27]. The Charlson comorbidity index, a weighted index that takes into account the number and the seriousness of comorbid disease, was calculated [28]. Serum albumin was measured by electrophoresis. Serum concentrations of apolipoproteins B and A-I were determined by the nephelometric method using a Behring Nephelometer 100 analyzer (Behring, Marburg, Germany).

Definitions of frailty: physical and biopsychosocial phenotypes

A phenotype of physical frailty was operationalized slightly modifying the Cardiovascular Health Study criteria [2], and identified by the presence of three or more frailty components: (1) weight loss, as unintentional weight loss > 5 kg in the last year; (2) self-reported exhaustion, identified by one question from the GDS-30 scale: “Do you feel full of energy?”; (3) weakness, as inability of standing on one leg without gait and balance disorders; (4) slowness, supported by the answers to the survey questionnaire dedicated to identify congestive heart failure symptoms: “Do you ever have to stop for shortness of breath while walking quickly on the flat or while slightly walking uphill?” and/or “Do you have to walk slower than people your own age due to shortness of breath?” and/or “Do you ever have to stop and breathe when walking at a normal pace on level ground?”; (5) low level of physical activity, using the physical activity questionnaire: inactive or light physical activity. To develop a biopsychosocial construct, the items of GDS-30 and IADL scales, instruments usually used in the CGA [10], went through a qualitative judgment process, as reported elsewhere [9]. A latent trait was identified by latent class analysis among some IADL items (1, 6, and 8) and GDS-30 items (3, 10, 12, 23, 28, and 29), cohabitation status, and presence of physical frailty [9]. We considered the biopsychosocial frail older individuals, as previously defined physical frail individuals with at least one of the two GDS-30 items impaired (item 3 “Do you feel that your life is empty?” or item 10 “Do you often feel helpless?”) [9].

Statistical analysis

The prevalence estimates of myocardial infarction, PAD, and definite stroke in biopsychosocial frail older individuals were based on baseline data collected in the ILSA first survey. A poststratification adjustment by age and gender was used to make the sample consistent with the population represented. Prevalence estimates and 95% confidence interval (CI) in the time frame of the first survey were evaluated according to the procedures of complex surveys supported by Stata package (The SVYSET command and the SVY: prefix) on the basis of the characteristics of the ILSA sampling. Except for pack-years cigarettes (Mann–Whitney U test), we used parametric statistical tests (Student’s t-test for independent samples and Pearson χ² test). Generalized estimating equation (GEE) models were used to assess the independent contribution at three different follow-ups (1° survey: March 1992 since to June 1993; 2° survey: September 1995 since to October 1996; 3° survey: March 2000 since to September 2001) of baseline individuals with biopsychosocial frailty on occurring new events of myocardial infarction, PAD, and stroke. GEE models were modeled separately, excluding participants with myocardial infarction, PAD, and definite stroke, respectively, and in overall, all-cause mortality. In establishing indicators in the statistical analyses, participants without biopsychosocial frailty served as the reference group. Partially adjusted odds ratios (ORs) (adjustment for 65–69, 70–74, 75–79, and 80–84 years of age categories and sex), and GDS-30 (coded 0 for GDS-30 < 9, coded 0 for GDS-30 ≤9) were estimated for each outcome. Fully adjusted GEE models for covariates were also fit using baseline covariates which could be predictive of myocardial infarction, PAD, and definite stroke: age categories (coded 1 for 65–69 years, coded 2 for 70–74 years, coded 3 for 75–79 years, and coded 4 for 80–84 years), sex (coded 0 for women and 1 for men), pack-years (pack-years cigarettes coded 0 for pack-years cigarettes = 0 (never smoking) and 1 for pack-years cigarettes ≥ 0.5), GDS-30 (coded 0 for GDS-30 < 9, coded 0 for GDS-30≤ 9), IADL, MMSE, Charlson comorbidity index, serum albumin levels, apolipoprotein B/apolipoprotein A-I ratio, and drugs (anticoagulants (dicumarol), antiaggregants (aspirin, ticlopidine), and lipid-lowering compounds (statins and fibrates)). All analyses have been performed using Stata statistical software (StataCorp. 2023, Stata Statistical Software: Release 18, College Station, TX: StataCorp LLC).

Results

Sociodemographic, nutritional-related, and clinical characteristics

Figure 1 shows sample size and attrition, according to the biopsychosocial frailty construct. Differences in mean age (72.9 ± 5.5 vs. 75.3 ± 5.7 years, P 0.01, evaluated by separate variance t-test) and sex (Pearson χ² 34.03, P 0.01) were observed between 2410 participants (1312 [53.4%] men and 1098 women [45.6%]) and 3222 nonparticipants (1502 [46.6%] men and 1720 [53.4%] women). Sociodemographic and clinical characteristics of participants with and without biopsychosocial frailty are shown in Table 1. Among older individuals screened, those with biopsychosocial frailty were older (77.0 ± 4.8 vs. 72.5 ± 5.5 years, P < 0.001), women were more numerous than men, less educated, more depressed, and cognitively impaired, and more multimorbid than older individuals without biopsychosocial frailty (Table 1).

Fig. 1.

Sample size and attrition at the different phases of the survey according to the biopsychosocial frailty (BF) construct and myocardial infarction, peripheral artery disease (PAD), and definite stroke. The Italian Longitudinal Study on Aging (1st (1992–1993), 2nd (1995–1996), and 3rd (2000–2001) surveys). CAD, cardiovascular artery disease (angina pectoris and myocardial infarction) and cerebrovascular disease (definite stroke and possible cerebrovascular events)

Table 1.

Demographic and clinical characteristics [mean ± SD or median (25th–75th percentiles) or %] of individuals with and without biopsychosocial frailty. The Italian Longitudinal Study on Aging (1st prevalence survey, 1992–1993)

Variable	Entire cohort (n = 2410)	With biopsychosocial frailty (n = 174)	Without biopsychosocial frailty (n = 2236)	P value
Women (%)	45.5	63.8	44.1	0.001§
Age (years)	72.9 ± 5.5	77.0 ± 4.8	72.5 ± 5.5	0.001†
Education (years)	6.4 ± 4.7	4.3 ± 3.6	6.6 ± 4.7	0.001†
Pack-years	0.7 (0–28.8)	0 (0–15)	1.5 (0–30)	0.001‡
Body mass index	26.9 ± 4.3	27.0 ± 4.7	26.9 ± 4.3	0.950†
Mini Mental State Examination	26.0 ± 3.9	23.2 ± 4.6	26.3 ± 3.8	0.001†
Activities of Daily Living	6.4 ± 1.0	7.4 ± 1.9	6.3 ± 0.8	0.001†
Instrumental Activities of Daily Living	8.6 ± 3.5	11.5 ± 5.2	8.3 ± 3.3	0.001†
Geriatric Depression Scale-30	9.6 ± 6.0	17.4 ± 5.0	9.0 ± 5.6	0.001†
Charlson comorbidity index	1 ± 0.8	1.3 ± 0.8	1.0 ± 0.8	0.001†
Hypertension (%)	68.9	80.5	68.0	0.001§
Type 2 diabetes (%)	12.7	14.9	12.5	0.360§
Atrial fibrillation (%)	27.9	35.1	27.4	0.029§
Congestive heart failure (%)	5.9	12.1	5.4	0.001§
Albumin (mg/dL)	4.2 ± 0.5	4.1 ± 0.5	4.2 ± 0.5	0.001†
Apolipoprotein B/A1	0.9 ± 0.3	0.8 ± 0.3	0.9 ± 0.3	0.010†
Drugs
- Antiaggregant (%)	11.2	13.8	11.0	0.25
- Anticoagulant (%)	0.9	1.2	0.9	0.73
- Lipid lowering (%)	6.0	4.0	6.2	0.25
Cerebrovascular disease				0.001§
- Definite stroke (%)	3.7	10.3	3.2
- Possible cerebrovascular events (%)	0.7	0.6	0.7
Coronary artery disease				0.106
- Myocardial infarction (%)	7.4	9.2	7.3
- Angina pectoris (%)	8.7	10.9	8.5
Peripheral artery disease (%)	4.5	4.6	4.5	0.961§

^†Student’s t-test for independent samples

^§Pearson’s χ² test

^‡Mann–Whitney U test

Estimates of prevalence of the biopsychosocial frailty construct in older adults with cardio- or cerebrovascular diseases

Biopsychosocial frailty construct was diagnosed in 174 of 2410 individuals (7.2%). As shown in Table 2, the prevalence rate of biopsychosocial frailty among subjects with myocardial infarction amounted to 9.9% [95% confidence interval (CI) 6.82–14.15]. It increased with class of age [Mantel–Haenszel estimate, odds ratio (OR) 1.78, 95% CI 1.26–2.53], was higher in women than in men (Mantel–Haenszel estimate, OR 0.43, 95% CI 0.20–0.91), and increased with class of age indifferently in men and in women (Mantel–Haenszel estimate, OR 1.44, 95% CI 0.87–2.39 for women and Mantel–Haenszel estimate, OR 2.22, 95% CI 1.33–3.71 for men). The prevalence rate of biopsychosocial frailty among subjects with PAD amounted to 8.90% (95% CI 5.06–15.20). It increased with class of age (Mantel–Haenszel estimate, OR 3.18, 95% CI 1.81–5.61), was not prevalent in men in comparison with women (Mantel–Haenszel estimate, OR 2.22, 95% CI 0.46–10.78), and increased with class of age but not by gender (Mantel–Haenszel estimate, OR 2.40, 95% CI 0.67–8.57 for women and Mantel–Haenszel estimate, OR 3.55, 95% CI 1.86–6.76 for men) (Table 2). Finally, the prevalence rate of biopsychosocial frailty among subjects with definite stroke amounted to 23.72% (95% CI 17.63–31.11). It increased with class of age (Mantel–Haenszel estimate, OR 2.01, 95% CI 1.38–2.94), was prevalent in women than in men (Mantel–Haenszel estimate, OR 0.23, 95% CI 0.09–0.56), and increased with class of age indifferently in men and in women (Mantel–Haenszel estimate, OR 2.27, 95% CI 1.37–3.77 for women and Mantel–Haenszel estimate, OR 1.48, 95%CI 0.77–2.83 for men) (Table 2).

Table 2.

Prevalence of biopsychosocial frailty phenotype among subjects with myocardial infarction, peripheral artery disease, and definite stroke by gender and age. The Italian Longitudinal Study on Aging (1st prevalence survey, 1992–1993)

Biopsychosocial frailty status	Gender	Age
		65–74 years		75–84 years
		%	95% CI	%	95% CI	%	95% CI
Myocardial infarction	Men	3.0	(1.1–7.7)	12.7	(7.4–21.0)
	Women	7.4	(3.0–16.7)	20.5	(12.4–32.1)
	Whole	4.5	(2.3–8.4)	16	(11.2–22.3)	10.9	(8.0–14.7)
Peripheral artery disease	Men	2.0	(0.3–13.8)	23.1	(12.6–38.4)
	Women	-	-	-	-
	Whole	1.5	(0.2–10.2)	19	(10.9–30.9)	8.9	(5.1–15.2)
Definite stroke	Men	10.3	(3.5–26.3)	15.0	(6.3–31.5)
	Women	-	-	-	-
	Whole	13.1	(6.8–23.8)	32.0	(22.4–43.4)	23.7	(17.6–31.1)

CI, confidence interval

Relationships among biopsychosocial frailty and cardio- or cerebrovascular diseases

Over an 8-year follow-up, 80 of 2091 (3.8%) older participants developed myocardial infarction, nine among 144 biopsychosocial frail individuals at baseline (6.3%); 85 of 2301 (3.7%) older participants developed PAD, nine among 166 biopsychosocial frail individuals at baseline (5.4%); 73 of 2304 (3.2%) older participants developed definite stroke, 11 among 155 biopsychosocial frail individuals at baseline (7.1%). Over an 8-year follow-up, participants with biopsychosocial frailty showed an increased OR of definite stroke (OR:2.67, 95% CI 1.11–6.40, in the fully adjusted model) in comparison with participants without biopsychosocial frailty. No statistically significant associations between new events of myocardial infarction (OR 1.59, 95% CI 0.65–3.86) in the fully adjusted model) or PAD and biopsychosocial frailty were observed (OR 2.08, 95% CI 0.93–4.64, in the fully adjusted model) (Table 3).

Table 3.

Odds ratios of myocardial infarction, peripheral artery disease, and definite stroke in biopsychosocial frail individuals. The Italian Longitudinal Study on Aging (1st (1992–1993), 2nd (1995–1996), and 3rd (2000–2001) surveys)

	Odds ratio partially adjusted		Odds ratio fully adjusted
	Odds ratios	(95% confidence interval)*	Odds ratios	(95% confidence interval)**
Myocardial infarction	1.92	(0.86–4.26)	1.59	(0.65–3.86)
Peripheral artery disease	1.64	(0.76–3.54)	2.08	(0.93–4.64)
Definite stroke	2.98	(1.41–6.30)	2.67	(1.11–6.40)

^*Odds ratios were adjusted for age (coded 1 for 65–69, coded 2 for 70–74, coded 3 for 75–79, and coded 4 for 80–84), sex (coded 0 for women and 1 for men), and 30-item Geriatric Depression Scale (GDS-30) (coded 0 for GDS- 30 < 9, coded 1 for GDS-30 ≥ = 9) in partially adjusted models

^**ORs were adjusted for age categories (coded 1 for 65–69, coded 2 for 70–74, coded 3 for 75–79, and coded 4 for 80–84), gender (coded 0 for women and 1 for men), education, pack-years (pack-years cigarettes (coded 0 for pack-years cigarettes = 0 (never smoking), and 1 for pack-years cigarettes ≥ 0.5)), Instrumental Activities of Daily Living score, Mini Mental Score Examination score at baseline, Charlson Comorbidity Index score, GDS-30 (coded 0 for GDS-30 < 9, coded 1 for GDS-30 ≥ 9), serum albumin levels in fully  adjusted models, atrial fibrillation (coded 0, without disease; coded 1, with disease), antiaggregant drugs (code 0, no assumption; code 1, aspirin and ticlopidine assumption), lipid-lowering drugs (code 0, no assumption; code 1, fibrates and statins assumption)

Discussion

In a large cohort of Italian older individuals, we identified a biopsychosocial frailty construct, involving item 3 or 10 of the GDS-30 combined with the presence of physical frailty [9]. The prevalence of biopsychosocial frailty in older individuals with myocardial infarction was 10.9%, in older individuals with PAD was 8.9%, and in older individuals with definite stroke was 23.7%. The biopsychosocial frailty phenotype was more prevalent in women only in myocardial infarction and stroke groups, but not in PAD and increased with age. Furthermore, over an 8-year follow-up, biopsychosocial frailty was associated to an increased about two and a half times odds ratio of developing definite stroke events, not myocardial infarction neither PAD.

To the best of our knowledge, this was the first study that examined the impact of a biopsychosocial frailty phenotype on the incident risk of cardio- and cerebrovascular disease, i.e., myocardial infarction, PAD, and definite stroke in a very large longitudinal population-based study. Previous clinical studies have shown that the degree of frailty (assessed principally with the physical frailty phenotype) was related to the risk of various outcomes linked to cardiovascular disease, such as myocardial infarction, stroke, heart failure, or atrial fibrillation [29–32]. Findings from the China Health and Retirement Longitudinal Study showed that there was a combined effect of frailty and loneliness on the risk of cardiovascular diseases, stroke, and cardiac events. The cumulative incidence of cardiovascular diseases, stroke and cardiac events during a 7-year follow-up was significantly higher in individuals with both loneliness and frailty than those without loneliness or frailty alone [33]. Emerging evidence from population-based studies confirmed that both loneliness and frailty were causally associated with increased risk of cardiovascular diseases. In fact, the UK Biobank study followed 479,054 participants for over 7 years and found that loneliness had 1.49-fold and 1.36-fold increased risks of incident acute myocardial infarction and stroke [34]. These findings indicated that social isolation (more than loneliness), similar to other risk factors such as depression, can be regarded as a risk factor for poor prognosis for cardiovascular disease. However, it remains unclear whether these associations are independent of conventional risk factors, i.e., biological, behavioral, socioeconomic, and health-related factors [35]. A systematic review including 23 studies from 16 datasets with 4628 cardiovascular diseases and 3002 stroke events over 3 to 21 years indicated that poor social relationship was associated with a 29% increase in risk of incident cardiovascular diseases and a 32% increase in risk of stroke, and the significant association was independent of age, gender, and other potential conventional risk factors [36]. According to previous studies [9, 14, 17, 18], the new construct of biopsychosocial model of frailty could explain this gap, providing a valid appraisal of the combined effect of frailty with loneliness and depression in older age.

In the present study, over an 8-year follow-up, biopsychosocial frailty was not associated to an increased risk of developing PAD. While the impact of frailty on mortality and some disease-specific health status in cardiovascular disease has been well described [37], its impact on PAD has been less well-studied. PAD results in skeletal muscle apoptosis and atrophy through ischemia-mediated tissue injury, leading to exercise intolerance, exertional limb symptoms, and progressive functional decline. Frailty probably may further exacerbate this functional decline. In a cross-sectional study of 216 patients with symptomatic PAD, frail subjects reported more health status impairment than non-frail individuals [38]. However, health status was examined in this study with a single-domain instrument capturing health status related to physical functioning and thus substantially overlapping with frailty itself [38]. On the other hand, the role of claudication in frail older subjects on functional impairment and quality of life and the impact of frailty on health status outcomes with guideline recommended therapies are unknown. These unanswered questions should be prioritized given the substantial overlap in functional limitations between PAD and frailty and the potential to improve frailty with PAD-specific treatments [37].

In the present study, the biopsychosocial frailty phenotype was more prevalent in women only in myocardial infarction and stroke groups, but not in older subjects with PAD and increased with age. This observation may reflect biological/gender vulnerability, greater psychosocial burden, or differences in access to healthcare. Concerning gender vulnerability, the present findings confirmed those from a population-based cohort (SLAS-1) of 2387 community-dwelling Singaporean Chinese older adults with physical, mental, and social frailty that were more likely to be older, female, single, divorced or widowed, smokers, at risk of malnutrition, underweight, and to have low total cholesterol, but less likely to be drinkers [39]. These multidimensional and biopsychosocial frail subjects were also more likely to have stroke, asthma/chronic obstructive pulmonary disease, chronic kidney disease, visual impairment, and neurological and psychiatric disorders [39].

The risks for frailty range from subcellular damage to social determinants. The precise mechanisms linking biopsychosocial frailty phenotype with risk of cardiovascular disease are not fully understood, but several potential explanations could be proposed. Chronic low-grade inflammation, insulin resistance that leads to impaired muscle protein breakdown, and vitamin D deficiency had also been implicated in both cardiovascular disease and frailty, as well as several nuclear factors including telomere shortening and lamin A/C reduction [40]. The conceptual change of frailty, from a unidimensional physical to a complex multidimensional biopsychosocial phenotype, expanded the field of frailty, including social and behavioral domains with critical interaction between different frailty models. Environmental exposures—including physical exercise, psychosocial factors, and diet—may play a role in the frailty pathophysiology [20]. The present findings also suggested that efforts to reduce the risk of stroke could benefit from taking both physical and psychological components of frailty, particularly loneliness and social isolation, into account [41]. Therefore, the present findings suggesting that biopsychosocial frailty was significantly associated with stroke risk, but not with myocardial infarction or PAD, could be partly explained by potential pathophysiological and behavioral mechanisms underlying this cerebrovascular predominance—such as depression, social isolation, physical inactivity, or atrial fibrillation. Understanding the degree of frailty with this biopsychosocial construct may provide important background information that is relevant to clinical studies of cardio- and cerebrovascular disease. For example, a high degree of biopsychosocial frailty may increase the risk of adverse outcomes from cardio- and cerebrovascular disease, including myocardial infarction and stroke [33, 34, 36]. This is not unexpected given that preclinical studies in aging mice have shown that high frailty index scores predict declines in cardiac function at the organ, cellular, and even subcellular levels [42].

Notwithstanding the importance of the multidimensional biopsychosocial frailty phenotype in predicting health- and cognitive-related adverse outcomes in older age, at present, this model is poorly operationalized [43]. The present construct derived from a latent class analysis obtained from a 11-item instrument in which we summarized nine items based on the CGA (three items from IADL and six items from the GDS-30), plus cohabitation status and physical frailty status [9]. Therefore, this construct included the concepts of social participation, loneliness, and social contacts (GDS-30), together with the physical/biological phenotype of frailty, so, encompassing biological, psychological, and social dimensions of frailty. For the present biopsychosocial frailty phenotype, we considered the physical frailty phenotype plus items derived from the GDS-30 (concerning social participation and loneliness) [9]. Of course, this construct may be generalized to other cohorts but it strongly depended on the specific assessment tools used (IADL, GDS-30, and cohabitation and physical frailty status). Therefore, the identication of physical frailty/cohabitation status plus a CGA investigating psychosocial domains may be useful for the operationalization of biopsychosocial frailty in clinical practice. However, at present, a series of tools for assessing biopsychosocial frailty domains are under investigation in community-dwelling older adults [44].

The strengths of the present study were the population-based setting, the long follow-up period, and the number of older subjects included, notwithstanding a relatively small number of those with biopsychosocial frailty and myocardial infarction, PAD, or stroke. However, given the population-based nature of the study, we had a relatively small sample size of older subjects with the biopsychosocial frailty phenotype and, consequently, the new events of cardio- and cerebrovascular disease. A further limitation was the lack of pharmacological records about all anticoagulant drugs, also because they were rarely administered in older subjects with atrial fibrillation in the 90 s, especially in older adults over 75 years of age. However, special consideration for older adults is warranted for effective stroke prophylaxis. Moreover, the analysis was not adjusted for the participants’ setting (e.g., community-dwelling vs. institutionalized) introducing a possible bias. Finally, the present findings were also based on a retrofitted assessment of physical frailty because this was not initially part of the study design and could have influenced the negative findings regarding myocardial infarction and PAD.

Conclusions

The present findings suggested over an 8-year follow-up, in a large, older Italian cohort, a biopsychosocial frailty phenotype was associated with incident cerebrovascular disease, in particular with definite stroke, rather than myocardial infarction and PAD. Frailty is a complex age-related clinical status characterized by a decline in the physiological capacity of various organ systems. Their degree of frailty makes frail individuals more susceptible to cardiovascular disease and mortality than are their fitter counterparts. In according with our previous study conducted in the same large cohort of Italian older individuals without cognitive impairment at baseline, a biopsychosocial frailty model was also a short- and long-term predictor of overall dementia, particularly VaD [9]. Therefore, understanding the degree of frailty may be critical, as frailty predicts brain vascular multimorbidity (VaD and stroke) and mortality better than do traditional risk factors for cardiovascular disease. The three principal dimensions of frailty (physical, social, and psychological) may be interrelated rather than independent [14, 37], and the new biopsychosocial concept may reflect the multidimensional nature of frailty. Finally, given that several components of this new construct are potentially modifiable, it may be prudent to implement public health policy and intervention to reduce the impact of the biopsychosocial frailty phenotype and to aid in the prevention of brain vascular multimorbidity, particularly stroke. While secondary preventive strategies for the physical frailty domain of the biopsychosocial frailty phenotype should be suggested, with an individualized multidomain interventions targeting physical and nutritional domains [43, 44], some primary intervention strategies have been identified to reduce the impact of the psychosocial domain of the biopsychosocial frailty phenotype that included the improvement of social skills, the enhancement of social support, the increase of opportunities for social contact, and addressing maladaptive social cognition [45, 46].

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

VS conceived and designed the study; VS, FP, and ML interpreted the data and wrote the manuscript; VS performed the statistical analysis and takes responsibility for the integrity of the data, the accuracy of data analysis, and had full access to all of the data of the study; ES, ML, CC, VD, DC, CC, MR, CG, SG, RS, AD, and FP assisted in literature search, interpretation of data, and manuscript preparation; ES, ML, CC, VD, and FP assisted in study design and data interpretation; ES and AD critically revised the manuscript. All authors approved the final draft submitted.

Funding

This paper was supported by the Project “Development of an ensemble learning-based, multi-dimensional sensory impairment score to predict cognitive impairment in an elderly cohort of Southern Italy” funded by the European Union – Next Generation EU – NRRP M6C2 – Investment 2.1 Enhancement and Strengthening of NHS biomedical research (Grant Agreement PNRR-MAD-2022-12376656).

Data Availability

The datasets analyzed during the current study are available from the corresponding Authors (F.P. and V.S.) upon reasonable request.

Declarations

Ethics approval

The study project was approved by the institutional review board of the 8 municipalities of the Italian Longitudinal Study on Aging.

Informed consent

Informed consent was obtained from each participant and/or their relatives before enrollment.

Conflict of interest

The authors declare no competing interests.