We read with interest the important paper by Chang et al. (Reference Chang2021) which explored the relationship between baseline depressive symptoms and incident frailty in older adults in Taiwan. The authors undertook a stratified analysis looking at the relationship between these two conditions in baseline robust and prefrailty groups in an attempt to address the important issue of overlapping criteria between depressive symptoms and frailty.
Overall, 2,717 participants (mean age 67.5 ± 0.1 years, 50.3% female) were analyzed from the Healthy Aging Longitudinal Study in Taiwan, a nationwide survey. Depressive symptoms were measured using the 20-item Center for Epidemiological Studies – Depression (CES-D) scale, which has good psychometric properties in older adults (Cosco et al., Reference Cosco2020), with a total score of 60, and frailty status was measured using the Fried frailty phenotype. The authors reported that the adjusted hazard ratio of becoming frail in those with depressive symptoms was 2.6 (95% CI: 1.6, 4.2) during the average 5.9-year follow-up period. Of the individual five domains of the Fried frailty phenotype, only the development of unintentional weight loss and exhaustion demonstrated a significant association with depressive symptoms, with adjusted hazard ratios of 2.1 (95% CI: 1.2, 3.9) and 4.8 (95% CI: 2.5, 9.3), respectively. There were no associations between depressive symptoms and the development of poor muscle strength, slowness, or lower levels of physical activity reported. Subgroup analyses findings were fairly consistent with the overall results despite limited sample sizes.
The strengths of this study include the sample that represents a diverse socio-demographic population in Taiwan achieved through systematic random sampling. This helps to further our confidence that the association between depressive symptoms and frailty likely remains relevant across racial and ethnic groups, which is similar to what has been found in western populations (Feng et al., Reference Feng2017). Further, the follow-up period, of almost 6 years, strengthens the effect estimation and solidifies the reported relationship between depressive symptoms and frailty as the limited number of longitudinal studies to date have had a shorter period (Soysal et al., Reference Soysal2017). A particularly novel aspect of this study was the subgroup analyses which attempted to mitigate the confounding effects of frailty on depressive symptoms and vice versa due to their overlapping criteria. From these analyses, the authors concluded that the reported association between frailty and depressive symptoms was in part due to their overlapping criteria.
While this paper was novel, some limitations may affect the validity of this study’s results and require addressing in future studies. First, the sample size was small in the stratified analysis and there was only one participant with depressive symptoms but no frailty symptoms at baseline. Clearly, this also introduces some bias into the results. Also of concern, but not uncommon, is the lower follow-up rates in those with depressive symptoms, coupled with the fact that those with baseline depressive symptoms were more likely to drop out of the study because of a depression-related decline in health and function. Attrition is a challenging bias to address in secondary analysis research. Yet, this is in part explained by the fact the population of interest is more likely to drop out. Therefore, when undertaking such research in those with depressive symptoms and long-term conditions, participant burden should be minimized, and additional efforts made to try and acquire complete data analysis. Lastly, the CES-D was used to assess depressive symptoms; while this has reasonable psychometric properties (Cosco et al., Reference Cosco2020), it cannot be used to determine a clinical diagnosis of depression itself and the results cannot be translated to this population with comorbid frailty.
To provide broader context to this study, it is imperative to explain both these conditions often cooccur and have a close comorbidity. A previous systematic review found that the prevalence of depression in a sample of people with frailty was 38.6% (95% CI: 30.07–47.10), while the prevalence of frailty in people with depression was 40.4% (95% CI: 27.00–55.30). The review also found that the odds of people with frailty having comorbid depression compared to nonfrailty control groups were 4.42 (95% CI: 2.66–7.35) and the odds of people with depression having frailty compared to nondepressed control groups were 4.07 (95% CI: 1.93–8.55) (Soysal et al., Reference Soysal2017). Lastly, this review found the odds ratio (OR) for incident frailty in older adults with depression was 3.72 (95% CI: 1.95–7.08) and the OR for incident depression in older adults with frailty was 1.90 (95% CI: 1.55–2.32; Soysal et al., Reference Soysal2017).
Understanding depressive symptoms-frailty comorbidity and risk is essential as these are among the most common conditions seen in older adults and both individually associated with a range of deleterious outcomes, which are amplified when they cooccur (Soysal et al., Reference Soysal2017). While there have been multiple theoretical definitions and proposed models for frailty, there remains a lack consensus on one definition, nonetheless it is generally agreed that frailty is a geriatric syndrome whereby the risk of physical and mental decline increases and is associated with disability and mortality (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013; Soysal et al., Reference Soysal2017). Similarly, depressive symptoms are correlated with many of the same outcomes as frailty such as higher disability, cognitive impairment, fractures, and mortality (Lohman et al., Reference Lohman, Dumenci and Mezuk2016; Mezuk et al., Reference Mezuk, Edwards, Lohman, Choi and Lapane2012). A key difference between depressive symptoms in the younger population and later-life depressive symptoms is that medical comorbidities can make differential diagnosis in later-life depressive symptoms from physical illness and indeed frailty symptoms challenging (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013).
As evident through merely reviewing the definitions and outcomes of these conditions, there is considerable overlap/similarity between depressive symptoms and frailty. The research to date has indicated a strong relationship, specifically a bidirectional relationship between depressive symptoms and frailty, and this relationship holds across different study designs, regions, sexes, and confounder variables used, but what has yet to be deciphered is the extent, type, and mechanism underlying this relationship (Chu et al., Reference Chu, Chang, Ho and Lin2019). A study by Mezuk et al. (Reference Mezuk, Lohman, Dumenci and Lapane2013) proposed that although people with depressive symptoms or frailty are distinct subgroups due to the substantial overlap, these syndromes should be considered together to identify predictors and outcomes of health and disability in older adults. A further study suggested that due to the psychological vulnerability that is also apparent in people with frailty, which is usually associated with depressive symptoms, the definition of frailty should incorporate depressive symptoms rather than treating them as a separate but related disorder (Lohman et al., Reference Lohman, Dumenci and Mezuk2016).
What should be evident as a common theme through this commentary and indeed the corresponding paper is that multiple theories exist but still there is uncertainty regarding concrete evidence around directionality regarding depressive symptoms-frailty. While Soysal et al. (Reference Soysal2017) established the prevalence and incidence of those with one of these conditions developing the other, the review was unable to decipher the underlying relationship between these two constructs or their overlapping pathways. A systematic review by Buigues et al. (Reference Buigues, Padilla-Sanchez, Garrido, Navarro-Martinez, Ruiz-Ros and Cauli2015) found that (1) frailty and depression are comorbid geriatric syndromes in a subgroup of older adults; (2) frailty is a risk factor for depressive symptoms due to them sharing the same common symptoms. It has also been suggested that frailty and depressive symptoms are interrelated rather than overlapping syndromes (Lohman et al., Reference Lohman, Dumenci and Mezuk2016). This conclusion was reached as the association between these syndromes holds across the various definitions of frailty and depressive symptoms and is also not fully explained by their shared common symptoms (Lohman et al., Reference Lohman, Dumenci and Mezuk2016). Therefore, where consensus in the literature is reached is that there is a need to disentangle the connection between these conditions and subsequently to explore the mechanisms explaining the relationship.
As mentioned, it remains unclear what factors underlie the risk and overlap between depressive symptoms and frailty (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013; Zhang et al., Reference Zhang2020). However, exploring proposed mechanisms can provide some insight. The first and potentially the most straightforward explanation is that several of the indications/individual diagnostic criteria used to identify frailty and depressive symptoms measure similar concepts. For example, fatigue is a key factor in depressive symptoms and exhaustion is a key component of frailty while weight loss is key in both conditions (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013; Soysal et al., Reference Soysal2017). Second, characteristics of depressive symptoms such as a reduction in social and physical activity, smoking, and a reduced dietary quality increase the likelihood of frailty occurring as well as depression (Firth et al., Reference Firth2020; Zhang et al., Reference Zhang2020). Third, depressive symptoms and frailty share common mechanisms such as inflammation (Köhler et al., Reference Köhler2017; Soysal et al., Reference Soysal2016). Recent meta-analyses demonstrated that both frailty (Soysal et al., Reference Soysal2016) and depression (Köhler et al., Reference Köhler2017) are cross-sectionally associated with similar raised inflammatory patterns, but longitudinal research is unclear in the conditions individually and in depressive symptom-frailty comorbidity.
The results of Chang et al. study are in line with previous literature which identified a relationship between depressive symptoms and frailty and concluded that depression or depressive symptoms may be a risk factor or predictor for incident frailty. Similar to Zhang et al. (Reference Zhang2020), this association held in the prefrailty and robust frailty at baseline groups. Based on our above potential explanations of the shared symptomology and mechanisms between depressive symptoms and frailty, it is not surprising that the authors found a strong association between the weight loss and exhaustion component of the Fried frailty phenotype and depressive symptoms. What was slightly unexpected was no significant association was found between slowness and a low level of physical activity and depressive symptoms as psychomotor slowness and a reduction in physical activity are also characteristics of depressive symptoms (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013). The authors also note this with surprise as this result is not consistent with previous research; however, this may be due to a few factors such as a very limited sample size or unsatisfactory follow-up rates, thus compromising the studies ability to have sufficient power.
Nonetheless, despite the above uncertainties and need for future research to better understand this comorbidity, it is well established both depressive symptoms and frailty have substantial health and social implications for individuals, as well as increasing the risk of further comorbidity such as hip fracture and osteoporosis which can have a devastating impact (Stubbs et al., Reference Stubbs, Brefka, Dallmeier, Stubbs, Vancampfort and Denkinger2016). Hip fractures in particular have a significant medical burden for both the patient and healthcare services and a negative impact on quality of life and disproportionately effect and negatively impact these populations (Stubbs et al., Reference Stubbs2020). Therefore, future research is needed to explore the extent and intricacies of the association and the mechanisms between depressive symptoms and frailty with adequately powered studies and people with diagnosed depression. Once this has been understood, this will potentially allow the early identification, prevention, or at least the delay of the transition from depressive symptoms to frailty or vice versa. This prevention may be in the form of interventions such as physical rehabilitation, one of the management approaches in mental health, or through stratification of care to identify those particularly at risk of developing both syndromes. As discussed above, detangling this relationship will not be simple or easy; however, a first step might be to delve into the overlapping symptomology of these constructs and how their absence and presence affects the correlation between depressive symptoms and frailty.
You have
Access
We read with interest the important paper by Chang et al. (Reference Chang2021) which explored the relationship between baseline depressive symptoms and incident frailty in older adults in Taiwan. The authors undertook a stratified analysis looking at the relationship between these two conditions in baseline robust and prefrailty groups in an attempt to address the important issue of overlapping criteria between depressive symptoms and frailty.
Overall, 2,717 participants (mean age 67.5 ± 0.1 years, 50.3% female) were analyzed from the Healthy Aging Longitudinal Study in Taiwan, a nationwide survey. Depressive symptoms were measured using the 20-item Center for Epidemiological Studies – Depression (CES-D) scale, which has good psychometric properties in older adults (Cosco et al., Reference Cosco2020), with a total score of 60, and frailty status was measured using the Fried frailty phenotype. The authors reported that the adjusted hazard ratio of becoming frail in those with depressive symptoms was 2.6 (95% CI: 1.6, 4.2) during the average 5.9-year follow-up period. Of the individual five domains of the Fried frailty phenotype, only the development of unintentional weight loss and exhaustion demonstrated a significant association with depressive symptoms, with adjusted hazard ratios of 2.1 (95% CI: 1.2, 3.9) and 4.8 (95% CI: 2.5, 9.3), respectively. There were no associations between depressive symptoms and the development of poor muscle strength, slowness, or lower levels of physical activity reported. Subgroup analyses findings were fairly consistent with the overall results despite limited sample sizes.
The strengths of this study include the sample that represents a diverse socio-demographic population in Taiwan achieved through systematic random sampling. This helps to further our confidence that the association between depressive symptoms and frailty likely remains relevant across racial and ethnic groups, which is similar to what has been found in western populations (Feng et al., Reference Feng2017). Further, the follow-up period, of almost 6 years, strengthens the effect estimation and solidifies the reported relationship between depressive symptoms and frailty as the limited number of longitudinal studies to date have had a shorter period (Soysal et al., Reference Soysal2017). A particularly novel aspect of this study was the subgroup analyses which attempted to mitigate the confounding effects of frailty on depressive symptoms and vice versa due to their overlapping criteria. From these analyses, the authors concluded that the reported association between frailty and depressive symptoms was in part due to their overlapping criteria.
While this paper was novel, some limitations may affect the validity of this study’s results and require addressing in future studies. First, the sample size was small in the stratified analysis and there was only one participant with depressive symptoms but no frailty symptoms at baseline. Clearly, this also introduces some bias into the results. Also of concern, but not uncommon, is the lower follow-up rates in those with depressive symptoms, coupled with the fact that those with baseline depressive symptoms were more likely to drop out of the study because of a depression-related decline in health and function. Attrition is a challenging bias to address in secondary analysis research. Yet, this is in part explained by the fact the population of interest is more likely to drop out. Therefore, when undertaking such research in those with depressive symptoms and long-term conditions, participant burden should be minimized, and additional efforts made to try and acquire complete data analysis. Lastly, the CES-D was used to assess depressive symptoms; while this has reasonable psychometric properties (Cosco et al., Reference Cosco2020), it cannot be used to determine a clinical diagnosis of depression itself and the results cannot be translated to this population with comorbid frailty.
To provide broader context to this study, it is imperative to explain both these conditions often cooccur and have a close comorbidity. A previous systematic review found that the prevalence of depression in a sample of people with frailty was 38.6% (95% CI: 30.07–47.10), while the prevalence of frailty in people with depression was 40.4% (95% CI: 27.00–55.30). The review also found that the odds of people with frailty having comorbid depression compared to nonfrailty control groups were 4.42 (95% CI: 2.66–7.35) and the odds of people with depression having frailty compared to nondepressed control groups were 4.07 (95% CI: 1.93–8.55) (Soysal et al., Reference Soysal2017). Lastly, this review found the odds ratio (OR) for incident frailty in older adults with depression was 3.72 (95% CI: 1.95–7.08) and the OR for incident depression in older adults with frailty was 1.90 (95% CI: 1.55–2.32; Soysal et al., Reference Soysal2017).
Understanding depressive symptoms-frailty comorbidity and risk is essential as these are among the most common conditions seen in older adults and both individually associated with a range of deleterious outcomes, which are amplified when they cooccur (Soysal et al., Reference Soysal2017). While there have been multiple theoretical definitions and proposed models for frailty, there remains a lack consensus on one definition, nonetheless it is generally agreed that frailty is a geriatric syndrome whereby the risk of physical and mental decline increases and is associated with disability and mortality (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013; Soysal et al., Reference Soysal2017). Similarly, depressive symptoms are correlated with many of the same outcomes as frailty such as higher disability, cognitive impairment, fractures, and mortality (Lohman et al., Reference Lohman, Dumenci and Mezuk2016; Mezuk et al., Reference Mezuk, Edwards, Lohman, Choi and Lapane2012). A key difference between depressive symptoms in the younger population and later-life depressive symptoms is that medical comorbidities can make differential diagnosis in later-life depressive symptoms from physical illness and indeed frailty symptoms challenging (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013).
As evident through merely reviewing the definitions and outcomes of these conditions, there is considerable overlap/similarity between depressive symptoms and frailty. The research to date has indicated a strong relationship, specifically a bidirectional relationship between depressive symptoms and frailty, and this relationship holds across different study designs, regions, sexes, and confounder variables used, but what has yet to be deciphered is the extent, type, and mechanism underlying this relationship (Chu et al., Reference Chu, Chang, Ho and Lin2019). A study by Mezuk et al. (Reference Mezuk, Lohman, Dumenci and Lapane2013) proposed that although people with depressive symptoms or frailty are distinct subgroups due to the substantial overlap, these syndromes should be considered together to identify predictors and outcomes of health and disability in older adults. A further study suggested that due to the psychological vulnerability that is also apparent in people with frailty, which is usually associated with depressive symptoms, the definition of frailty should incorporate depressive symptoms rather than treating them as a separate but related disorder (Lohman et al., Reference Lohman, Dumenci and Mezuk2016).
What should be evident as a common theme through this commentary and indeed the corresponding paper is that multiple theories exist but still there is uncertainty regarding concrete evidence around directionality regarding depressive symptoms-frailty. While Soysal et al. (Reference Soysal2017) established the prevalence and incidence of those with one of these conditions developing the other, the review was unable to decipher the underlying relationship between these two constructs or their overlapping pathways. A systematic review by Buigues et al. (Reference Buigues, Padilla-Sanchez, Garrido, Navarro-Martinez, Ruiz-Ros and Cauli2015) found that (1) frailty and depression are comorbid geriatric syndromes in a subgroup of older adults; (2) frailty is a risk factor for depressive symptoms due to them sharing the same common symptoms. It has also been suggested that frailty and depressive symptoms are interrelated rather than overlapping syndromes (Lohman et al., Reference Lohman, Dumenci and Mezuk2016). This conclusion was reached as the association between these syndromes holds across the various definitions of frailty and depressive symptoms and is also not fully explained by their shared common symptoms (Lohman et al., Reference Lohman, Dumenci and Mezuk2016). Therefore, where consensus in the literature is reached is that there is a need to disentangle the connection between these conditions and subsequently to explore the mechanisms explaining the relationship.
As mentioned, it remains unclear what factors underlie the risk and overlap between depressive symptoms and frailty (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013; Zhang et al., Reference Zhang2020). However, exploring proposed mechanisms can provide some insight. The first and potentially the most straightforward explanation is that several of the indications/individual diagnostic criteria used to identify frailty and depressive symptoms measure similar concepts. For example, fatigue is a key factor in depressive symptoms and exhaustion is a key component of frailty while weight loss is key in both conditions (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013; Soysal et al., Reference Soysal2017). Second, characteristics of depressive symptoms such as a reduction in social and physical activity, smoking, and a reduced dietary quality increase the likelihood of frailty occurring as well as depression (Firth et al., Reference Firth2020; Zhang et al., Reference Zhang2020). Third, depressive symptoms and frailty share common mechanisms such as inflammation (Köhler et al., Reference Köhler2017; Soysal et al., Reference Soysal2016). Recent meta-analyses demonstrated that both frailty (Soysal et al., Reference Soysal2016) and depression (Köhler et al., Reference Köhler2017) are cross-sectionally associated with similar raised inflammatory patterns, but longitudinal research is unclear in the conditions individually and in depressive symptom-frailty comorbidity.
The results of Chang et al. study are in line with previous literature which identified a relationship between depressive symptoms and frailty and concluded that depression or depressive symptoms may be a risk factor or predictor for incident frailty. Similar to Zhang et al. (Reference Zhang2020), this association held in the prefrailty and robust frailty at baseline groups. Based on our above potential explanations of the shared symptomology and mechanisms between depressive symptoms and frailty, it is not surprising that the authors found a strong association between the weight loss and exhaustion component of the Fried frailty phenotype and depressive symptoms. What was slightly unexpected was no significant association was found between slowness and a low level of physical activity and depressive symptoms as psychomotor slowness and a reduction in physical activity are also characteristics of depressive symptoms (Mezuk et al., Reference Mezuk, Lohman, Dumenci and Lapane2013). The authors also note this with surprise as this result is not consistent with previous research; however, this may be due to a few factors such as a very limited sample size or unsatisfactory follow-up rates, thus compromising the studies ability to have sufficient power.
Nonetheless, despite the above uncertainties and need for future research to better understand this comorbidity, it is well established both depressive symptoms and frailty have substantial health and social implications for individuals, as well as increasing the risk of further comorbidity such as hip fracture and osteoporosis which can have a devastating impact (Stubbs et al., Reference Stubbs, Brefka, Dallmeier, Stubbs, Vancampfort and Denkinger2016). Hip fractures in particular have a significant medical burden for both the patient and healthcare services and a negative impact on quality of life and disproportionately effect and negatively impact these populations (Stubbs et al., Reference Stubbs2020). Therefore, future research is needed to explore the extent and intricacies of the association and the mechanisms between depressive symptoms and frailty with adequately powered studies and people with diagnosed depression. Once this has been understood, this will potentially allow the early identification, prevention, or at least the delay of the transition from depressive symptoms to frailty or vice versa. This prevention may be in the form of interventions such as physical rehabilitation, one of the management approaches in mental health, or through stratification of care to identify those particularly at risk of developing both syndromes. As discussed above, detangling this relationship will not be simple or easy; however, a first step might be to delve into the overlapping symptomology of these constructs and how their absence and presence affects the correlation between depressive symptoms and frailty.
Acknowledgements
Brendon Stubbs is supported by a National Institute for Health Research (NIHR) Advanced Fellowship (301206, 2021–2026). Brendon Stubbs is a lead/co-investigator the following active grants (1) NIHR program grant: Supporting Physical and Activity through Co-production in people with Severe Mental Illness (SPACES, 2021–2027); (2) TB multi-morbidity with the Medical Research Council (GCRF call (2020–2022); (3) Determinants of MLTCs among young adults with mental disorders: a data-linkage study, Guy’s and St Thomas' Charity (2020–2022); (4) Mechanisms underlying the role of gut-microbiota in exercise-induced changes in cognitive function in middle-age, Reta Lila Weston Trust For Medical Research (2021–2024); (5) Improving Outcomes in Mental and Physical Multi-morbidity and Developing Research Capacity (IMPACT) in South Asia, NIHR Global Research program grant (2017–2022). Brendon Stubbs has published a book on exercise and mental illness and is on the Editorial board of Ageing Research Reviews, Mental Health and Physical Activity, The Journal of Evidence Based Medicine and The Brazilian Journal of Psychiatry. Brendon Stubbs has received honorarium for advisory work from ASICS Europe BV and ParachuteBH for work unrelated to this paper. The views expressed are those of the author(s) and not necessarily those of mentioned above, the NHS, the NIHR, the Department of Health and Social Care, the MRC or GSTT.