Depressive and Biopsychosocial Frailty Phenotypes: Impact on Late-life Cognitive Disorders

1 INTRODUCTION

In older age, frailty, a detrimental transitional status of the aging process, is characterized by declining reserve capacities to maintain homeostasis across multiple physiological systems, manifesting as a functional decline and increased vulnerability to different age-related and subclinical conditions as well as aging stressors, inducing an increased risk of adverse health-related outcomes, including disability, hospitalization, falls, institutionalization, mortality, as well as dementia [1–3].

Frailty is fluctuating and dynamic, potentially reversible, and may occur at any age, although the prevalence increases with aging [1, 2]. However, notwithstanding two decades of intensive clinical research, while there is now a consensus on the conceptual framework of frailty, no consensus has yet emerged as to a shared and unifying operational definition and how frailty should be identified in clinical practice. Indeed, over 70 different operational definitions are used [4, 5]. The two most commonly and widely recognized frailty models are the unidimensional Cardiovascular Health Study (CHS) definition of the physical frailty phenotype [6, 7], based on biologically interconnected symptoms and signs, including weakness, slow gait, poor physical activity, exhaustion, and unintentional weight loss, and the non-specific multidimensional frailty index model of age-related deficit accumulations [8], based on a continuous score of symptoms, signs, disabilities and diseases.

In recent years, the concept of frailty phenotypes has been constantly evolving. Multidimensional frailty phenotypes, such cognitive frailty [9, 10], oral frailty [11, 12], nutritional frailty [13, 14], organ-specific frailty [15], social frailty [16, 17], biopsychosocial frailty [18], and psychological frailty [19, 20], this last including motivational and affective dimensions, i.e., the recently proposed depressive frailty phenotype [21], have been widely defined and validated. Moreover, some subtypes or markers of the physical frailty phenotype, strictly linked to some intrinsic capacity domains, such as mobility [22] or sensorial frailty [23], have also been proposed. All these frailty phenotypes include peculiar health deficits increasing with aging and all related with adverse health-related outcomes. Therefore, the two principal frailty models may not be in conflict but rather reciprocally complementary.

In particular, among multidimensional frailty phenotypes, depressive frailty may be considered a complex phenotype with components resembling those of physical frailty in adults with late-life depression (LLD), representing the clinical manifestation of greater biological aging [21]. The presence of physical frailty components in the context of a depressive disorder may expose older subjects to detrimental trajectories [21], increasing the risk of adverse health-related outcomes, including dementia syndromes. Moreover, biopsychosocial frailty considers the integral functioning of older individuals: it is a broader concept that covers frailty factors in biological/physical, social, and psychological dimensions [18], also defining a status of greater biological aging and including motivational, emotional, and socioeconomic domains. The broader concept of biopsychosocial frailty offers an integrated view and thus probably subsumes depressive frailty. In short, depressive frailty may be viewed as a phenotype within the overall biopsychosocial model.

Frail older people have an increased risk of several adverse health-related outcomes, including cognitive disorders in older age and dementia [3]. In fact, different frailty phenotypes have been related to late-life cognitive disorders, all-cause dementia, mild cognitive impairment (MCI), Alzheimer’s disease (AD), AD neuropathology, vascular dementia (VaD), and non-AD dementias [4] and might be eligible for the earliest intervention, which could bring about a potential reversibility [3, 4]. The strong relationship of frailty with AD and other dementias has been confirmed by cross-sectional findings from the Rush Memory and Aging Project, suggesting that an accumulation-based frailty status might act as a significant moderator of the association between AD neuropathology and dementia, as frailty increases the susceptibility to dementia [24]. Given the associations found among different frailty phenotypes and late-life cognitive disorders, in the present article, we reviewed current evidence of the links existing among depressive/biopsychosocial frailty phenotypes, dementias, and cognitive disorders in late life, also examining the mechanisms underlying these associations and the possible inclusion of these constructs in prevention and timely intervention strategies against dementia syndromes.

2 BIDIRECTIONAL RELATIONSHIP BETWEEN FRAILTY AND LATE-LIFE DEPRESSION: THE DEPRESSIVE FRAILTY PHENOTYPE

In the last decade, clinical research has underlined similarities between frailty and LLD regarding symptomatology, outcomes, and risk factors for both conditions [25]. LLD onset has an age range from 45 to 65 years and includes both early-onset depression (EOD), recurring or continuing after 65 years of age, and late-onset depression (LOD), that first appears after 65 years [26]. LOD accounts for about half of LLD cases [27]. At present, it is still unclear whether LLD should be considered a unique and complex phenotype or else a difference should be made according to the age of onset and its related subtypes (EOD and LOD), also regarding adverse health-related outcomes [27].

Although the concepts of frailty and LLD are strictly linked and partly overlapping, some population-based findings suggested that these two constructs may be disentangled [28, 29]. In fact, a latent class analysis from the Baltimore Epidemiologic Catchment Area Study suggested the separation of frailty and LLD as different latent constructs [28]. However, after classification according to depression criteria as severe, mildly, and not depressed older people, and to frailty criteria as frail and non-frail, the classes were highly overlapping and all subjects with severe LLD were also classified as frail [28]. Findings from the Health and Retirement Study confirmed the presence of both LLD and frailty [29]. Nonetheless, in this study, LLD severity was directly related with frailty severity, regardless of the frailty model used (physical frailty phenotype or frailty index) [29], but after accounting for shared symptoms of LLD and frailty, these correlations were somewhat weakened [29]. Altogether, these findings suggested that LLD and frailty may be two distinct constructs but with some degree of overlap that, in clinical practice, can be difficult to distinguish [30]. Two meta-analyses and systematic reviews investigated the associations between frailty and depression in older age in clinical- and population-based studies [31, 32]. The first meta-analysis showed a robust association between depression and frailty, suggesting that 38.6% of frail people also had significant coexisting depressive symptoms and 40.4% of LLD subjects were also pre-frail or frail [31]. This meta-analysis suggested an interaction between these two conditions, LLD and frailty being associated with a reciprocally increased prevalence and incidence [31], although these findings could be partly explained by the fact that only 3 of the 24 studies included had a categorical diagnosis of LLD. However, another more recent meta-analysis showed that subjects with LLD were more prone to frailty than those without LLD, regardless of study design, region, depression or frailty criteria, and covariance adjustment [32]. In this study, only gender had a significant influence, older men with LLD being at higher risk of frailty than older women with LDD [32]. These meta-analyses thus suggested there was a bidirectional relationship between LLD and frailty [33, 34]. In 2016, the depressive frailty phenotype model was proposed, starting from the hypothesis that frailty characteristics in LLD subjects may be the clinical manifestation of an increased biological aging, and that the coexistence of frailty and a depressive disorder may expose older people to deleterious trajectories [21, 35]. A very recent neuroimaging study, showing that LLD and frail subjects had significant microstructural changes within white matter tracts detected with diffusion tensor imaging as compared to never-depressed and robust older individuals, suggested that an increased neuroinflammatory burden may explain the co-occurrence of these two conditions, as well as the development of a depressive frailty phenotype in older age [36]. There is currently no consensual operationalization serving to describe and identify this frailty phenotype. While several studies explored the association between LLD and frailty [31, 32], little investigation has been made about the existence of an intrinsic vulnerability to emotional stressors with increased age indicating depressive frailty as a possible predictor of dementia and other negative health-related outcomes. However, some subtypes of LLD could be defined by specific combinations of age-related clinical features and labeled as “frail-depressed, physically dominated”, “frail-depressed, cognitively dominated”, and “amnestic depression”, [37]. Patients in this last subgroup had the highest remission rates, while decreased survival at two-year follow-up was observed among patients in both frail-depressed subgroups [37]. Therefore, given the increased mortality risk of LLD patients with physical frailty, a depressive frailty phenotype operationalized with these clinical features may mark an increased risk also of other adverse health-related outcomes, including dementia and late-life cognitive disorders.

From an epidemiological point of view, the depressive frailty phenotype suggested by several studies, describing the co-existing constructs of cognitive frailty and LLD [51–65], is poorly operationalized. However, the 2017 Lancet Commission on dementia [67] and the 2020 update [68] reviewed risk factors associated with dementia, identifying, among others, factors like physical inactivity and little social contact. These modifiable dementia risk factors may be encompassed in a recently operationalized multidimensional frailty construct, the biopsychosocial phenotype [18]. According to the deficit accumulation model and comprehensive geriatric assessment (CGA), the biopsychosocial frailty construct combines physical and psychosocial domains [18, 69, 70]. It could be defined as a status of increased biological aging including, emotional, motivational, and socioeconomic features.

The biopsychosocial frailty definition may add a great contribution in the assessment and intervention against dementia. 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 VaD, over 3.5- and 7-year follow-ups [18]. In the ILSA, this construct was operationalized using the CHS physical frailty phenotype plus assessment tools usually used in the CGA, such as items from 30 item-Geriatric Depression Scale (GDS-30) and Instrumental Activities of Daily Living (IADL). In particular, an 11-item instrument was obtained from CGA (3 items from IADL and 6 items from GDS-30), plus cohabitation status, and physical frailty status [18]. The biopsychosocial construct was identified with a latent class analysis involving the items 3 (“Do you feel that your life is empty?”) or 10 (“Do you often feel helpless?”) of the GDS-30 (psychosocial criterion), plus physical frailty (biological criterion) [18]. Furthermore, other longitudinal population-based studies investigated the incidence risk of all-cause dementia [71–74] or AD [75] in older subjects with biopsychosocial frailty. Table 2 shows the principal population-based longitudinal studies linking biopsychosocial frailty with all-cause dementia [18, 71–74] and AD [18, 75]. The prevalence estimates of this construct ranged from 5% [18] to 6.8% [76, 77]. However, the operationalization of the biopsychosocial frailty construct may be a source of variability as regards epidemiological estimates. A recent narrative review article identified possible tools for the evaluation of biopsychosocial frailty features in community-dwelling older adults [78]. The review focused on the biopsychosocial dimensions of frailty included in the SUNFRAIL Checklist, a tool with 9 items; of these, 5 items were in the physical domain, 2 in the neuropsychological domain, and 2 in the socio-economic domain [79, 80]. These biopsychosocial domains (polypharmacy, nutrition, cognitive decline, medical visits, physical activity, falls, loneliness, social support, and economic limitations), should be appropriately selected and integrated in primary care, providing a dataset for the employment of targeted and timely health-related advancement, prevention, and information services [78].

In a recent meta-analysis and systematic review of six longitudinal population-based studies (219,616 subjects), the multiconcept frailty and late-life cognition were investigated. In this study, 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 AD [81]. 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 MCI and particularly with the non-amnestic MCI single and multiple domains subtypes [76]. Biopsychosocial frailty was not associated with either the amnestic MCI single or multiple domains subtypes. Therefore, these results suggested that a modifiable risk factor such as the biopsychosocial frailty phenotype may be associated with cognitive decline in older age, i.e., MCI and non-amnestic MCI [76]. In fact, a greater adverse impact of the biopsychosocial frailty phenotype on dementia risk was observed for non-amnestic MCI compared with amnestic MCI [74]. 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 AD [77]. No statistically significant association between biopsychosocial frailty and probable AD or other dementias was observed. These findings confirmed the association of the biopsychosocial frailty phenotype, interpreted as a modifiable risk factor, with all-cause dementia, VaD, and AD [77]. Therefore, these findings from population-based studies [71–77] and pooled evidence from a large meta-analysis [81] suggested that this model of frailty, that 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.

The identification of shared biological pathways among biopsychosocial and depressive frailty phenotypes and late-life cognitive disorders is fundamental for achieving targeted, more personalized prevention and treatment interventions. However, the mechanisms underlying these interplays are not entirely clear. An improved understanding of how common etiologic mechanisms may contribute to both cognitive dysfunction and different frailty phenotypes will contribute to develop prevention strategies in at-risk groups and to set up targeted clinical interventions for older adults with these syndromes. The goal will be the prevention of disability and premature death.

4.1 Depressive frailty phenotypes: Possible underlying mechanisms

The model of a depressive frailty phenotype stems from the hypothesis that frailty characteristics in LLD subjects may be the clinical manifestation of greater biological aging, and that frailty dimensions in the context of a depressive disorder may expose older people to harmful outcomes [21]. Shared biological substrates that may result in the depressed frailty phenotype are thought to be inflammatory processes, multimorbidity, mitochondrial dysfunction, dopaminergic neurotransmission, and neuroendocrine dysregulation (Fig. 1). Considering the bidirectional relationship between LLD and frailty and the decreased survival associated with the depressed frailty phenotype, the identification of shared biological substrates may have potential implications for the assessment and treatment of LLD [21].

OPEN IN VIEWER

Fig. 1

Depressive frailty phenotype should be considered a complex phenotype deriving from an association between late-life depression (LLD) and physical frailty, with a difference according to the age of LLD onset (early-onset depression – EOD and late-onset depression – LOD), also regarding adverse health-related outcomes. Multimorbidity, brain dopamine activity, chronic inflammatory state, increased biological aging, age-related mitochondrial function changes, and neuroendocrine dysregulation could be the underlying links between LLD and physical frailty. ADL, activities of daily living; TNFα, tumor necrosis factor alpha; IL-6, interleukin 6; ATP, adenosine triphosphate; IGF-1, insulin-like growth factor-1; HPA, hypothalamic-pituitary-adrenal axis

4.8 Biopsychosocial frailty phenotypes and dementia: Possible underlying mechanisms

The physical/biological perspective of frailty does not completely describe the susceptibility of older adults at risk of developing late-life cognitive disorders. Instead, the concept of the biopsychosocial frailty phenotype may capture important aspects in terms of assessment and targeted intervention in older age. Although poorly operationalized, the multidimensional biopsychosocial frailty model, with its polymorphic biological changes underpinning the frailty condition, may predict health- and cognitive-related adverse outcomes in older age [3]. Among adverse health-related outcomes, functional disability, increased institutionalization, dementia, and mortality are noteworthy [18]. Other poor health-related outcomes are influenced by social determinants such as engagement and social support, social isolation and loneliness and a poor social network, together with clinical aspects such as hypertension and immune system dysfunction [131, 132], cognitive decline and dementia [133, 134], psychological dysregulation, and a shorter life expectancy [135].

According to the biopsychosocial construct derived from the ILSA data [18], this frailty phenotype impacted dementia in a multifactorial way (Fig. 2). Several mediators or possible pathways implicated in the process may explain the physical frailty-cognition links: from hormonal to inflammatory processes, together with nutritional, vascular, neuropathological, and metabolic influences [4]. To a large extent, family functionality and neighborhood showed an independent association with frailty and pre-frailty. Dysfunctional family engagements can negatively influence adaptive responses to sources of stress, showing the essential impact of social support [136]. Also, quality of life in older age is influenced by family relationships, because of their impact on the individual’s ability to perform tasks [136]. Recognizing dysfunctional dynamics of family among frail and pre-frail older adults is essential to make it possible to carry out social interventions aimed at delaying or preventing the progression of frailty syndromes [137]. Older adults’ perception of their life and quality of life during the occurrence of frailty and pre-frailty [138, 139] is influenced by psychological, social, and environmental components in a bidirectional way. Also, compliance to drug and rehabilitation treatments is influenced by negative self-assessments of health [140].

OPEN IN VIEWER

Fig. 2

The biopsychosocial frailty phenotype with coexistent biological/physical and psychosocial dimensions has been associated with late-life cognitive disorders: common pathways/substrates and underlying mechanisms are depicted in the scheme.

The physical, social, and psychological dimensions of the biopsychosocial frailty phenotype are interrelated rather than independent [141], and this construct extensively reflects the multidimensional and dynamic nature of frailty. In fact, there are different potential mechanisms linking biopsychosocial frailty and late-life cognitive decline and dementia. Firstly, social stress, through immune functioning and hormonal and inflammatory processes, may lead to the development of cognitive dysfunction. Then, social frailty, independently linked with a decline in physical function [16], may induce cognitive decline through physical frailty. Although social withdrawal may precede the manifestations of dementia, several studies suggested an increased risk of developing incident AD with loneliness, but not with social isolation [142–144]. Several pathways may explain the association of loneliness with dementia, including the triggering of neural responses directly influencing the development of subsequent neurodegenerative conditions or the association with low physical activity levels, substance abuse and poor nutrition, unhealthy behaviors that negatively affect cognition directly or via an increased cardiometabolic risk [145].

Subjective and objective social support provide interpersonal buffering that may be an important psychosocial resource to cope with stressors. Biopsychosocial frailty led to various health-related outcomes later in life [146]. Objective social isolation may induce incident dementia through cardiovascular pathways and an increasing risk of hypertension [147] and coronary heart disease [148]. Moreover, growing evidence linked sleep disturbances [149], cigarette smoking [150, 151], increased alcohol consumption [152], and excessive television viewing [153, 154] to cognitive decline. All these risk factors were linearly increased with a greater objective social isolation.

Genome-wide assessments in the UK Biobank also suggested a relation between personality traits and social isolation [155]. Among AD risk factors, the neuroticism score, reflecting emotional vulnerability to stress, showed the strongest effect size for loneliness and lack of social support. Higher late-life neuroticism levels have been related to increased risks of developing MCI [156] and dementia [157, 158]. Specific personality traits, through social buffering of stress, can affect individual susceptibility to stressors, while social support can reduce physiological stress responses [159]. Similarly, the “cognitive reserve” hypothesis suggests that intellectual enrichment may provide a cognitive buffer to deal with CNS injuries [160]. For the same reason, rich social networks may contribute to better cognitive function, allowing easier access to health information [161]. In fact, an association was found among different factors of social interaction and several aspects related to cognitive load, i.e., education levels, socioeconomic status, computer use, and sensory impairment [162]. Therefore, shared biological pathways/substrates linking biopsychosocial frailty phenotypes and late-life cognitive disorders are hypothesized to be inflammatory and cardiometabolic processes, multimorbidity, specific personality traits, sensorial impairment, lack of subjective and objective social support, and loneliness (Fig. 2).

In the last decade, the concept of multidimensional frailty phenotypes has been constantly evolving and different constructs have been widely defined and validated [9–23]. In particular, depressive frailty may be defined as a complex phenotype resembling the components of physical frailty in LLD older adults [21]. Moreover, biopsychosocial frailty is a broader concept covering frailty factors in biological/physical, social, and psychological dimensions [18, 69], and that probably subsumes depressive frailty, that may be viewed as a phenotype within the overall biopsychosocial model.

Nowadays, the precision medicine initiative suggests that old age psychiatrists, gerodontologists, or old age nutritionists, after screening with general frailty approaches, could use specific frailty phenotypes to more precisely predict prevention strategies and effective interventions. Clearly, to confirm the true utility of different frailty phenotypes in predicting adverse health-related outcomes, in future studies, we need to operationalize and test them in different settings, using novel vulnerability indicators (e.g., oral frailty or social/biopsychosocial frailty) [11, 12, 16, 17], or including among existing frailty constructs other associated health indicators, i.e., physical frailty plus cognitive impairment for cognitive frailty [9, 10], physical frailty plus nutritional imbalance for nutritional frailty [13, 14], and physical frailty plus LLD for depressive frailty [21]. Therefore, physical frailty may be the basal component in the construct of several frailty phenotypes. But other phenotypes, including social and psychological frailty, only correspond to social and psychological domains. Therefore, a common component is required to establish a unifying link among multidimensional frailty phenotypes.

The depressive frailty phenotype, suggested by several studies describing the coexisting constructs of cognitive frailty and LLD [51–65], with prevalence estimates ranging from 3.5% [61] to 91% [64], is still poorly operationalized. Therefore, at present, there is a lack of studies with depressive frailty phenotypes operationalized with physical frailty plus LLD or plus one of its subtypes (EOD or LOD) [26], or also with the presence of significant depressive symptoms evaluated with specific scales, i.e., the GDS-30 or the Center for Epidemiologic Studies-Depression scale [163]. However, a higher progression of the depressive frailty phenotype towards cognitive dysfunction in older age was suggested by the existence of TOI [41–43], defined by the co-occurrence of physical, cognitive, and emotional decline as a surrogate frailty marker, and by findings suggesting that the cognitive frailty phenotype may be a risk factor for LLD and vice versa [51–65] (Table 1). Therefore, a construct of depressive frailty linking physical frailty to a categorical (LLD, EOD, or LOD) or dimensional (GDS-30 > /=10) diagnosis of LLD, two modifiable risk factors for cognitive dysfunction, may be a good predictor of dementia and late-life cognitive decline.

On the contrary, the operationalization of biopsychosocial frailty has been based on CGA and the deficit accumulation model (i.e., frailty indexes) and widely validated by longitudinal studies [18, 71–75] (Table 2). Furthermore, findings from different population-based studies [18, 71–77] and pooled meta-analytic evidence [81] suggested that the biopsychosocial model of frailty may be an optimal target for potential interventions and prevention strategies in delaying cognitive decline and dementia by targeting the different modifiable components of frailty. In the near future, we need well-designed prospective population-based studies investigating the associations among the depressive and biopsychosocial frailty phenotypes and incident dementia or MCI and its progression to dementia, also addressing confounding sources and potential bias.

6 POTENTIAL INTERVENTIONS AND PREVENTION STRATEGIES TARGETING DIFFERENT FRAILTY DIMENSIONS

At present, considering the findings from epidemiological studies and meta-analyses [18, 71–77, 81], it may be useful to implement public health policy and interventions to reduce the impact of the depressive/biopsychosocial frailty phenotypes on cognitive dysfunction in older age, so preventing or delaying MCI and ultimately dementia. Although there is not yet a consensual operationalization able to describe and identify the depressive frailty phenotype, given the increased mortality risk of physical frail patients with LLD [37], frailty assessment may better modulate the pharmacological and psychological treatment of LLD, and the need for interventions primarily targeting frailty, i.e., the reduction of polypharmacy plus lifestyle interventions [30]. Nonetheless, physical exercise and nutritional advice should be considered together with targeting age-related changes underlying physical frailty and LLD, e.g., low-grade inflammation [30].

Furthermore, multidomain interventions might play an important part in delaying the onset of late-life cognitive disorders and secondary occurrence of disability, institutionalization, hospitalization, and mortality. Physical exercise is a vital component of multidomain interventions, particularly if combined with interventions in other areas. Some randomized clinical trials (RCTs) showed that physical exercise can manage frailty, improving cognitive dysfunction. Recent findings from the Lifestyle Interventions and Independence for Elders (LIFE) trial, involving 1,298 cognitive frailty participants, showed that a 24-month structured physical activity program reduced the severity of cognitive frailty in comparison with a healthy education program, without modifications of this benefit by underlying inflammation [164]. Moreover, other recent findings from the LIFE suggested that a well-structured physical activity program produced a faster 400-m gait speed potentially able to prevent mobility disability among physical frail older individuals with preserved muscle strength in the lower limbs [165]. Another RCT showed that after three months, physical exercise may have greater benefits on cognitive status and depression among physical pre-frail/frail patients with MCI or mild dementia compared to controls [166]. Furthermore, results from another RCT found that a supervised-facility multicomponent exercise program can reverse physical frailty and improve cognitive function, emotional, and social networking in community-dwelling frail older adults [167].

Secondary preventive strategies for the physical frailty dimension of the biopsychosocial frailty phenotype should be suggested, including individualized multidomain interventions targeting physical and nutritional domains that may delay MCI onset and the progression to overt dementia [3]. In addition, some primary intervention strategies may reduce the impact of the psychosocial dimension on the biopsychosocial frailty phenotype, including social skills improvement, social support enhancement, increased opportunities for social contact, and the management of maladaptive social cognition [168].

7 CONCLUSION

The present article reviewed evidence suggesting that the CGA-based biopsychosocial frailty construct, combining physical and psychosocial domains [18, 69], defines a status of biological aging including emotional, motivational, and socioeconomic characteristics, with prevalence estimates ranging from 5% [18] to 6.8% [76, 77]. This multidimensional frailty phenotype has been transversally and prospectively associated with all-cause dementia, MCI, non-amnestic MCI, AD, probable AD, AD neuropathology, VaD, probable and possible VaD, and non-AD dementias [18, 71–77, 81]. Therefore, the concept of the biopsychosocial frailty phenotype may capture important aspects of assessment and targeted intervention in older age [3]. From an epidemiological point of view, the depressive frailty phenotypes suggested by several studies describing the co-existing constructs of cognitive frailty and LLD, with prevalence estimates ranging from 3.5% [61] to 91% [64], are still poorly operationalized. In any case, depressive frailty may be viewed as a phenotype within the overall biopsychosocial model and a construct linking physical frailty to a categorical or dimensional diagnosis of LLD, two modifiable risk factors for cognitive dysfunction, may be a good predictor of dementia onset in late life.