Background

The population around the world is ageing rapidly. It is expected that by 2050, the number of people older than 65 years worldwide will increase to two billion [1]. In England, the number of older people reached 18% of the population in 2016 [2]. Regardless of the decline in fertility and birth rates, all countries are experiencing an increase in life expectancy [3]. This increase has resulted in a rising demand for, and pressure on, local services. There is a particular demand for these services within the National Health Service (NHS) in the UK, where budget spending on older people grew from two percent to 16% between 2007 and 2016 [4]. Therefore, the health status of older members of the population is becoming a priority. A healthy dietary intake and active lifestyle are key components for healthy aging and are well recognised as factors associated with a lower risk of developing lifestyle-related diseases [5–7].

The aging process involves changes in body functions including alterations in the sense of smell and taste and capacity to chew and swallow [8]. These changes directly affect appetite and food intake [8], which can have an impact on nutritional status and can lead to malnutrition in the form of undernutrition [8]. There are two forms of malnutrition undernutrition, and being overweight or obese which results from diet-related non-communicable diseases [9, 10]. Cachexia is a another nutritional disorder which can be define as “a complex metabolic syndrome associated with underlying illness and characterized by loss of muscle with or without loss of fat mass” [11]. Malnutrition “undernutrition” is common in older people and generally increases with aging [12, 13]. It is crucial to detect malnutrition in older people because it causes a decline in overall health [14]. It can worsen the deterioration of health and functional status, and increase the risk of dependency and disability [12]. It is associated with undesirable health conditions, including an increased likelihood of morbidity and mortality [15, 16], an increased risk of falls [17], reduced cognitive function, and decreased bone mass in older people living in community dwellings [18, 19]. Furthermore, older malnourished inpatients have a higher number of complications and longer hospital episodes, which result in increased associated healthcare costs [10].

Over the last few decades, many criteria have been proposed for identifying malnutrition and many tools have been implemented. Most of these tools include a combination of the same variables or criteria: weight loss, body mass index (BMI) and signs of eating difficulties [20]. One of the most recently proposed definitions of malnutrition emerged from the Global Leadership Initiative on Malnutrition (GLIM). Here malnutrition is determined using a phenotype and aetiological model, which includes three phenotypes (non-volitional weight loss, low BMI and reduced muscle mass) and two aetiological criteria (reduced food intake and inflammation) [21]. The GLIM has broader criteria because it was developed using previous screening and assessment tools to ensure it is applicable in all countries and transferable to all settings. By using the GLIM criteria malnutrition can be compared in different settings through prevalence, incidence, intervention and outcome [21].

In addition to malnutrition, two other conditions frailty and sarcopenia are widespread in older populations. All three conditions are characterised with a loss of body tissue, which can result in negative outcomes such as an increased risk of falls [22], poor quality of life [23], increased risk of mortality [24, 25]. Malnutrition and frailty both have other similar characteristics in the assessments used, including, functional and cognitive status. Thus there is overlap in the definition and treatment of these conditions [26].

Frailty is associated with ageing and may lead to premature death [27, 28], multi-morbidity, specific long-term conditions and mortality [29, 30]. In addition, frailty has been shown to decrease the quality of life of older people [31]. It can be defined as a ‘state of increased vulnerability to poor resolution of homoeostasis after a stressor event’ [32]. Two of the most widely used models used to identify frailty are the phenotype model, which describes a group of patient characteristics [32], and the cumulative deficit model, which assumes an accumulation of deficits that happen with ageing [33, 34].

The phenotype model describes frailty as a ‘clinical syndrome in which three or more of the following criteria are present: unintentional weight loss (10 lbs in the past year), self-reported exhaustion, weakness (grip strength), slow walking speed and low physical activity’ [35]. This model is used as the definition to diagnose frailty in many studies [36].

The cumulative deficit model was proposed by Rockwood and Mitnitski and was validated in the Canadian study of Health and Aging [36, 37]. This model used an index that considers symptoms, signs, diseases and disabilities as deficits that are combined into an FI to determine frailty [38]. The original cumulative deficit model was based on 92 variables [39]. Subsequent research revealed that it could be reduced to 30 variables, without a loss in predictive validity [40]. The cumulative deficit model calculates the FI by defining the numbers of health deficits present in a person, and then dividing by the number of health deficits considered [38, 41].

In addition to malnutrition and frailty, the older adult population may also experience the related condition sarcopenia. In 2000, the World Health Organisation(WHO) estimated that 600 million people aged 60 or older worldwide were diagnosed with sarcopenia [42]. This number is predicted to increase to 1.2 billion by 2025, and to double by the year 2050 [42]. A study suggested that frailty and its relation to sarcopenia, malnutrition and other factors is a network called the ‘geriatric web of frailty’ [43]. It argued that frailty is a complex condition caused by many interacting factors, such as falls, fractures, impaired protein synthesis, malnutrition and sarcopenia [43]. Many researchers have shown that the overlap between frailty and sarcopenia is due to the shared features between them, such as weight loss and low physical capability. This is because both conditions involve a failure in skeletal muscle (low physical function) that is detected through grip strength and gait speed [35, 44].

People with sarcopenia have specific characteristics, including a low body weight and a deterioration in physical performance [45]. It can be defined as ‘a progressive and generalised skeletal muscle disorder that is associated with an increased likelihood of adverse outcomes including falls, fractures, physical disability and mortality’ [46]. Older adults with sarcopenia have reported poorer general health and physical function scores compared to participants without sarcopenia [45].

The European Working Group on Sarcopenia in Older People (EWGSOP) recommended criteria for diagnosing sarcopenia. These include low muscle mass, low muscle strength and lower physical performance [46]. In order to detect sarcopenia, two parameters are needed pertaining to the amount of muscle and its function. These can be translated into measurable variables: muscle mass, muscle strength, and physical performance [47]. To measure muscle mass, there are many techniques that could be used for assessment. These include computed tomography, magnetic resonance imaging, dual energy X-ray, and bioimpedance analysis [47]. The measurement of muscle strength has fewer validated techniques, and one of the most widely used is handgrip strength because it is easy to use and readily available [48]. Others include knee flexion or extension or peak expiratory flow. There are a number of methods that have been used for physical performance, including the short physical performance battery, gait speed, the timed get-up-and-go test and the stair climb power test [47]. However, the gait speed test is usually used for clinical practice and research [47].

There are different cut-off points when defining sarcopenia that are based on the measurement technique used and the availability of reference standards. The EWGSOP recommends the use of a reference value derived by calculating two standard deviations below the mean from data collected from healthy, young adults rather than other predictive reference population cut-off points [47].

The aim of this study is to explore the overlap between malnutrition, frailty and sarcopenia in an older population, using UK biobank data, by determining the estimated prevalence of these three conditions.

Methods/design of UK Biobank

This is a cross-sectional study using participants recruited to the UK Biobank. The UK Biobank is a population-based prospective longitudinal cohort with information on approximately 500,000 people. Participants were recruited from the NHS aged between 40–69 years. However, the database included participants aged 37 to 73 for the baseline visit. Data were collected for the initial assessment visit between 2007 and 2010. This was termed the baseline assessment that will be used for this study. An invitation was sent by mail to people who were living within a reasonable travelling distance from the assessment centres used for data collection. People were recruited from England, Scotland and Wales.

At the baseline assessment visit, all participants had to be able to give consent before starting the assessment. Then a series of assessment stations where undertaken that included a touchscreen questionnaire and an interview questionnaire about diet, cognitive function, work history and digestive health [49]. In addition to demographic data, the participants were given cognitive function tests, their height, hip and waist measurements were recorded, as well as their bioelectrical impedance measurements and hand-grip strength. The UK Biobank database aims to improve prevention, diagnosis and treatment of a broad number of diseases. This detailed information on participants provides a resource for investigators to conduct research relating to particular diseases [49].

The UK Biobank Ethics and Governance Council (EGC) was founded by the Medical Research Council and Wellcome Trust, allowing them to act independently [50]. The UK Biobank controls all access to and use of the data [50]. However, UK Biobank evaluates all proposals to verify that they would be compliant with the participants’ consent and the Biobank’s Governance Framework and have received the required ethics approval [50]. The researcher and collaborators signed a Material Transfer Agreement (MTA) with the UK Biobank. The MTA was submitted and the payment process completed for the approved application, allowing the data to be released to the approved researcher.

Baseline characteristics

Participant characteristics that will be recorded are detailed in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline participant characteristics.

https://doi.org/10.1371/journal.pone.0278371.t001

At the baseline assessment participants’ medical histories were recorded including a list of comorbidities that is described as 43 long-term conditions similar to the Hanlon study [29]. In addition, we will include other characteristics as shown in Table 1.

Mapping malnutrition to biobank variables.

The GLIM criteria will be applied to variables collected in the UK Biobank database in order to define participants with malnutrition. It includes five criteria, which will be mapped to equivalent variables that are available in the UK Biobank shown in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Variable mapped to malnutrition, frailty and sarcopenia.

https://doi.org/10.1371/journal.pone.0278371.t002

These variables will represent three phenotypic criteria and two etiologic criteria, where patients will be defined as having malnutrition if they have one phenotypic criterion and one aetiological criterion. The three phenotypic criteria are BMI < 22 kg/m² if participants are 70 years or older, or BMI < 20 kg/m² if participants are under 70 years old [21]; weight loss, as reported by the participants’ response to ‘Compared with one year ago, has your weight changed?’; and finally fat free mass index, calculated by dividing fat free mass by height squared (FFMI, kg/m2). The thresholds will be <17 kg/m² for males and < 15 kg/m² for females [20].

The first etiologic criteria, is defined as reduced food intake < 50% reduction of the standard energy requirement for the male group (50–54 years: 2,581 kcal/day, 55–64 years: 2,581 kcal/day, 65–74 years: 2,342 kcal/day) and the female group (50–54 years: 2,103 kcal/day, 55–64 years: 2,079 kcal/day, 65–74 years: 1,912 kcal/day) [51]. The second etiologic criterion is inflammation based on the C-reactive protein (CRP). [11, 52].

Mapping frailty to biobank variables.

Two models will be used to define frailty and compared. Only one model will be used to consider the overlap with malnutrition and sarcopenia. The 36 deficits model, has been constructed and validated, and applied to large data sets previously [39]. It will be matched to equivalent variables recorded in the UK Biobank. These are listed in S1 Table. The cumulative deficit model calculates the FI by defining the numbers of health deficits present in a person and then dividing by the number of health deficits considered [38, 41]. For instance, a participant with three deficits would have an index score of 3, which out of 36 equals a 0.083 FI. The threshold levels will be: low (robust), where the FI is less than or equal to 0.2; medium (pre-frail), where the FI is more than 0.2 and less than or equal to 0.35; and high (frail), where the FI is more than 0.35. These levels were suggested by Kulminski and colleagues and used in studies that looked at frailty, bone health, and overall mortality [53, 54].

The phenotype model, includes five criteria that will also be mapped with available variables in the UK Biobank. The first four criteria are derived from the self-reported variables entered by participants including: weight loss, exhaustion, slow walking pace and low physical activity. Hand grip strength is a measured variable See Table 2. These criteria were previously used in two studies based on the UK Biobank database [29, 55]. Participants will be classified as robust (met none of the frailty criteria), pre-frail (met one or two of the criteria) or frail (met three or more of the criteria) according to cut-offs described by Fried and colleagues [35]. This is done to ensure consistency with the existing literature, refer to Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Mapping the phenotype model.

https://doi.org/10.1371/journal.pone.0278371.t003

Mapping sarcopenia.

The EWGSOP standard definition of Sarcopenia has three criteria that can be mapped to variables in the biobank. Sarcopenia is diagnosed if the person has low muscle mass in addition to criterion two or three. The first is muscle mass measured by bioelectrical impedance and the second is muscle strength, measured by handgrip strength (see Table 2). Finally, physical performance as assessed by the participants’ self-reported answers to the validated International Physical Activity Questionnaire [56].

Discussion

Research describing the three conditions- malnutrition, frailty and sarcopenia- and their overlap in the population is currently limited. Few researchers have studied the effect of two of the three conditions in older adults [26, 57], but the three conditions are rarely studied together [43]. In a group of 92 selected nursing home residents aged 75 and above all three conditions were present in 7%. Twenty-nine (32%) residents were neither malnourished, sarcopenic nor frail. In addition, 5% of the residents were malnourished and simultaneously sarcopenic and frailty coexisted with sarcopenia in 10% of the subjects. In a group of 92 residents, selected from nursing home aged 75 and above, were measured for the overlap between malnutrition, frailty and sarcopenia. It showed all three conditions were present in 7% of residents, while 32% of the participants were free from malnutrition, frailty, and sarcopenia. In addition, 5% of residents were malnourished with sarcopenia in the nursing home and frailty coexisted with sarcopenia in 10% of subjects [58]. These three conditions affect mobility and increase the risk of disability, which has been shown to shorten lifespans [29, 45, 59]. One study reported an indicator called healthy life years (HLY) and measured it across 14 European countries between 1995 and 2003. The results showed that HLY indicator was stagnated in UK males, and decreased in UK females aged 65 years and above [60, 61]. This highlights that early identification of these conditions in older adults is vital so treatment and interventions can be initiated in this age group [62]. However, the evidence behind these conditions is still developing and an evidence gap still exists. Thus, further studies are needed in order to describe then understand the overlap between malnutrition, frailty and sarcopenia.

This proposed research aims to investigate the estimated prevalence of three conditions—malnutrition, frailty and sarcopenia—in people aged 50 years and above who are living in community dwellings using the UK Biobank database. In order to conduct this study, many variables need to be mapped with the available variables in the UK Biobank database. It will be challenging, as not all variables in the diagnostic criteria for malnutrition, frailty and sarcopenia are in the database. For example, the GLIM criteria will be used to identify malnourished older people within the UK Biobank participants. In the phenotypic criteria, we will map weight loss using the participants’ answers to the question ‘compared with one year ago, has your weight changed?’, and not by using the amount of weight recorded. For frailty, the 36 deficits will be matched with the UK Biobank database. However, not all 36 deficits variables were available in the database. For example, the deficit ‘best mobility and transfer problems’ was mapped with an attendance/disability/mobility allowance variable in the database, which has the best-matched information. Although we have made as much effort as possible to match equivalent variables there will be some variables where matches are not as specific as they could be. However, the sarcopenia EWGSOP diagnostic criteria was matched with similar variables in the database with no issue. In addition to these factors the biobank data is not representative of the UK population as a whole [49]. However, the findings of this study will be standardised using data from a second source so the results will be more generalisable.

Other studies have used biobank data and published results mapping definitions of sarcopenia and frailty to available biobank variables [29, 55, 63] and these studies have been considered in this study design.

Declarations

Supporting information

Acknowledgments

I would express my deep gratitude to the participants who gave their time and consent for participating in the UK Biobank database.