The association between sarcopenic obesity and malnutrition in community-dwelling older adults

Abstract

Background

The increase in fat tissue and the decrease in muscle mass with advancing age have prompted researchers to explore the coexistence of sarcopenia and obesity, i.e. sarcopenic obesity (SO). SO may lead to malnutrition due to poor diet quality, while malnutrition may contribute to SO by causing further muscle loss and metabolic imbalances.

Objectives

The aims were to investigate: (i) the prevalence of SO in community-dwelling older adults, (ii) the diagnostic ability of two different malnutrition methods, and (iii) the association between SO and malnutrition.

Methods

Community-dwelling older adults (≥65 years) were invited to participate. SO assessment was conducted based on the ESPEN/EASO consensus criteria. Malnutrition was evaluated based on both the Global Leadership Initiative on Malnutrition (GLIM) criteria and the Mini-Nutritional Assessment (MNA).

Results

Five hundred and ninety older adults (69.3% women, mean age: 74.31 ± 6.55 years) were included in the study. The overall prevalence of SO was 5.9% (n = 35). The prevalence of malnutrition was 23.9% according to the GLIM criteria, while it was 3.1% according to MNA. The agreement between the two measurements was ĸ = 0.32. There was no association between SO and malnutrition based on either GLIM (P: .06, OR: 1.971, 95% CI: 0.966–4.024) or MNA (P: .948, OR: 1.934, 95% CI: 0.119–7.306).

Conclusions

Even though the agreement for diagnosing malnutrition between GLIM criteria and MNA was fair, the number of participants diagnosed with malnutrition by GLIM criteria was almost eight times higher than MNA. No association was established between SO and malnutrition defined by GLIM or MNA.

Clinical trial number

NCT05122104.

Key Points

• The overall prevalence of sarcopenic obesity (SO) was 5.9%.

• The prevalence of malnutrition was 23.9% and 3.1% according to the Global Leadership Initiative on Malnutrition (GLIM) criteria and Mini-Nutritional Assessment (MNA), respectively.

• There was no association between SO and malnutrition defined by the GLIM or MNA.

Introduction

Despite appearing as opposite ends of the nutritional spectrum, malnutrition and obesity often coexist in older adults [1]. Earlier studies have suggested a significant association between obesity and malnutrition, due to the consumption of calorie-dense but nutrient-poor foods [1–3]. This is a paradox where individuals may consume excessive calories while still lacking essential nutrients [1–3]. Both malnutrition and obesity are linked to severe adverse outcomes such as disability due to decreased functioning, prolonged hospitalisation, impaired quality of life and even increased mortality rate [4–6].

The addition of sarcopenia to obesity, termed sarcopenic obesity (SO)—a phenomenon defined as ‘the coexistence of obesity, characterized by excess fat mass, and sarcopenia’—further complicates and exacerbates these issues [7]. To elaborate, the addition of obesity to sarcopenia worsens muscle loss, and amplifies inflammation, insulin resistance and oxidative stress, leading to heightened disability, frailty and mortality risks compared to obesity or sarcopenia alone [8]. When malnutrition is superimposed on these conditions, it further accelerates muscle catabolism and systemic inflammation, leading to even more severe consequences [9].

Fortunately, both SO and malnutrition exhibit reversibility when identified in their early stages underlying the importance of timely recognition and intervention [10, 11]. Therefore, experts in the field took many initiatives to properly diagnose, assess and treat these syndromes. In 2022, The European Society for Clinical Nutrition and Metabolism (ESPEN) and the European Association for the Study of Obesity (EASO) proposed a diagram to diagnose SO better [7]. Before the publication of ESPEN/EASO criteria, SO was diagnosed if a patient had sarcopenia and obesity together. The diagnosis of sarcopenia and obesity was conducted separately as two different body composition problems, and they were based on criteria defined by independent organisations, rather than accepting SO as one entity [8].

In 2016, the Global Leadership Initiative on Malnutrition (GLIM) criteria were introduced for diagnosing malnutrition in clinical settings by the major global clinical nutrition societies [12]. Before the GLIM criteria were introduced, malnutrition was mostly defined based on generic tools such as Mini-Nutritional Assessment (MNA) and Nutrition Risk Screening 2002 (NRS). However, neither one of these current tools nor the diagnostic criteria for malnutrition have gained widespread global acceptance [13]. Therefore, the GLIM criteria were developed to establish a common language for the diagnosis of malnutrition worldwide, aiming to be utilised in all settings necessitating malnutrition assessment [14]. Subsequently, recent studies started to report that MNA might be less effective in identifying malnutrition in older adults than GLIM [15]. It was speculated that the success of GLIM originated from its comprehensiveness since it also included inflammatory status and reduction of muscle mass [16]. Nevertheless, the GLIM criteria have been mostly used to evaluate hospitalised patients, and its accuracy on outpatients is not yet fully understood.

Considering the magnitude of adverse events associated with both malnutrition and SO, further efforts in research are crucial to deepen our understanding of these phenomena and their co-occurrence using newly proposed diagnostic criteria [17]. Moreover, while existing studies have primarily focused on hospitalised or patient populations, there is a significant gap in research that evaluates these conditions in the community, where the majority of older adults reside [18]. Therefore, our aims with this study were to investigate: (i) the prevalence of SO in community-dwelling older adults based on newly defined ESPEN-EASO criteria, (ii) the diagnostic ability of GLIM criteria and MNA in terms of malnutrition, and (iii) the association between SO and malnutrition based on different diagnostic tools.

Methods

This was a cross-sectional study conducted in a single centre in Istanbul-Türkiye. A local ethical committee approved the study (date: 28.10.2021; number: 315). The procedures carried out in this study complied with the 1964 Helsinki Declaration and its subsequent amendments. The study was registered to ClinicalTrials.gov (Clinical trial number: NCT05122104).

Participants

This study was carried out in the geriatric outpatient clinic of a hospital between November 2021 and February 2024. The inclusion criteria were as follows: (i) community-dwelling older adults (≥65 years old), (ii) independently mobile (including assistive devices), and (iii) native Turkish speakers. The exclusion criteria were as follows: (i) infectious diseases in the acute or subacute stages, (ii) active malignancy, (iii) steroid use, (iv) unstable cardiovascular diseases, (v) physically dependent individuals (immobilised patients, etc.), and (vii) <21 points on Mini-Mental State Examination. The flowchart of the study is presented in Figure 1.

Outcome measurements

A sociodemographic form was filled out by the researchers based on the information received from the patients, their family members and medical files. The form included questions such as age, gender and comorbidities. Afterwards, the researchers conducted measurements for weight, height and anthropometric data.

SO assessment was carried out in accordance with the ESPEN/EASO criteria [7]. Body composition analysis such as weight, body fat percentage, skeletal muscle mass, etc., was calculated with TANITA© TBF 300 bio-impedance analysis. According to this consensus, the assessment of SO is structured in two levels: screening and diagnosis. At the screening level, body mass index (BMI) or waist circumference and adverse factors associated with sarcopenia are evaluated. The cut-off for BMI was ≥30 kg/m², whereas ≥102 cm for males and ≥88 cm for females were taken as cut-off points for the waist circumference [19]. A SARC-F score of >4 points and a gait speed of <0.8 m/s were classified as adverse events [20].

The diagnostic level included two steps: (i) muscle strength changes, and (ii) body composition changes [7]. A handgrip strength of <30 kg for males and <20 kg for females was set as the cut-off point for muscle strength criteria [21]. A fat percentage of >40.7% for females and > 27.3% for males as well as a reduced muscle mass defined by skeletal muscle mass/weight (SMM/W) ratio with the cut-offs <37% for males and 27.6% for females have been adopted for the second step of the diagnostic level [22, 23].

Malnutrition was evaluated based on both MNA and GLIM criteria. The Turkish version of MNA was used for the assessment [24]. MNA score was categorised as follows: a score of ≥24 points indicated normal nutrition, 17–23.5 points indicated a risk of malnutrition, and <17 points indicated malnutrition [25]. GLIM-based assessment of malnutrition was performed as described in a previous study [12]. The first phenotypic criteria were confirmed if the individual had lost >5% of the body weight within the past 6 months or >10% beyond 6 months. For the second criterion, if the BMI was <20 kg/m² for individuals aged <70 years, or <22 kg/m² for those aged ≥70 years, the BMI was categorised as low [12]. The normative value for appendicular lean mass index (kg/m²) was decided as 9.2 and 7.4 kg/m² for males and females, respectively [26].

In terms of etiological criteria, both participant and family member statements were compared with medical records. For the first criterion, patients were queried about any decrease in food intake or assimilation due to gastrointestinal issues [12]. The first question in the MNA (i.e. ‘Has food intake declined over the past 3 months due to loss of appetite, digestive problems, chewing or swallowing difficulties?’) was asked to confirm this question [24]. For the second criterion, acute inflammation factors (such as trauma, major infections and burns) were omitted since these were exclusion criteria, and the study was conducted in an outpatient setting. Instead, chronic inflammation was evaluated through recorded comorbidities: diabetes mellitus, chronic kidney disease, ischemic heart disease, chronic heart failure, severe depression and chronic obstructive pulmonary disease [12]. An indicator of active inflammation in blood tests, namely, serum C-reactive protein (CRP), was also taken into account. In cases where the serum CRP level exceeded 5 mg/dl, it indicated the presence of inflammation, and this criterion was confirmed [27]. Subsequently, a diagnosis of malnutrition was confirmed if at least one phenotypic and one etiological criteria were validated [12].

Statistical analysis

Statistical analyses were performed by using Statistical Package for Social Science version 28.0 (IBM SPSS Statistics, Armonk, NY). Descriptive statistics are presented as means with standard deviations for continuous variables, while numbers and frequencies are provided for categorical and binary variables. The diagnosis of SO and malnutrition was performed with frequency analysis. Differences in these diagnoses between genders were calculated with the chi-square test. Binomial regression was used to analyse the association between age, comorbidities and malnutrition as well as SO and malnutrition. The statistical level of significance was set at P < .05.

Results

This study included 590 older adults (69.3% women, mean age 74.31 ± 6.55 years). Table 1 represents a comprehensive overview of the characteristics of the participants.

The overall prevalence of the SO was 5.9% (n = 35) based on the ESPEN-EASO consensus. SO was present in 7.1% of the females and 3.3% of the males. In terms of malnutrition, the mean MNA score was 24.68 ± 3.6 points, and 25.4% of the population was at risk for malnutrition, whereas 3.1% already had malnutrition. However, when malnutrition was assessed according to the GLIM criteria, malnutrition was present in 141 older adults (23.9%). Among those patients, the severity of malnutrition was medium in 61 older adults (10.3%) and severe in 80 individuals (13.6%). Figure 2 shows the distribution of patients who had SO and malnutrition based on GLIM or MNA. According to the GLIM criteria, 13 older adults (2.2%) were identified with both SO and malnutrition, while the MNA criteria identified only 1 older adult (0.17%) with both conditions. Figure 3 represents the overlap of the patients with malnutrition diagnosed by MNA and GLIM. The agreement (i.e. kappa coefficient) between the two measurements was ĸ = 0.32.

The association between the malnutrition defined by GLIM criteria and the other variables in the study was also investigated. Age was found to be mildly associated with malnutrition (P: <.0001, OR: 1.065, 95% CI: 1.035–1.096), whereas malnutrition was not linked to the comorbidities (P > .05 for all), except asthma/COPD. Asthma/COPD was significantly associated with malnutrition (P: < .0001, OR: 4.181, 95% CI: 2.220–7.873).

Binary logistic regression analysis revealed that there was no association between SO and malnutrition based on GLIM (P: .06, OR: 1.971, 95% CI: 0.966–4.024). There was also no association after the model was adjusted by age and the presence of asthma/COPD (P: .299, OR: 1.500, 95% CI: 0.698–3.220). When malnutrition was defined by MNA, there was again no association between malnutrition and SO (P: .948, OR: 1.934, 95% CI: 0.119–7.306). An adjusted model revealed a nonsignificance as well (P: .854, OR: 0.822, 95% CI: 0.101–6.694).

Discussion

This study was the first to explore the relationship between sarcopenic obesity, as defined by the ESPEN/EASO criteria, and malnutrition diagnosed using the GLIM criteria in community-dwelling older adults. Our findings regarding the first aim of this study (i.e. investigating the prevalence of SO based on newly defined ESPEN-EASO criteria) revealed that the SO prevalence was 5.9% in community-dwelling older adults of Türkiye. Secondly, the prevalence of malnutrition was 3.1% according to MNA, whereas it was 23.9% according to the GLIM criteria. The coexistence of SO and malnutrition was observed in 2.2% of the population based on the GLIM criteria, whereas only one older adult (0.17%) exhibited both conditions according to the MNA criteria. Thirdly, there was no association between SO and malnutrition based on both diagnostic tools.

Gao et al. [28] reported that the prevalence of SO was 11% globally in one of the recent meta-analyses. In another meta-analysis, the SO was present in 9% of the population and this was the case for both males and females [29]. There were expected discrepancies between the current study and these studies, as the prevalence rates in the latter were not based on the newly defined ESPEN-EASO criteria. However, the prevalence rate defined by bio-impedance analysis in Gao et al.’s [28] study was 7% (with a range of 5%–9%), which was similar to our finding. On the other hand, a previous study [30] also reported a different prevalence rate for SO (16.4%) based on ESPEN-EASO criteria in community-dwelling older adults of Türkiye. The difference between Ozturk et al.’s study [30] and our study could have originated from the adoption of different cut-off values for SMM/W to define reduced muscle mass. Ozturk et al. [30] defined reduced muscle mass using higher cut-off values compared to ours, which, in turn, led to a higher prevalence rate. Similarly to our findings, the prevalence rate of SO defined by ESPEN-EASO criteria was 5.8% in a very recent study with a large cohort from the Netherlands [31].

The authors believe the most prominent finding of this study was that even though the agreement between GLIM criteria and MNA was fair [32], the number of participants diagnosed with malnutrition by GLIM criteria was almost eight times higher than those diagnosed with MNA. Furthermore, half of these older adults have been suffering from severe malnutrition based on GLIM criteria and most of them were not even identified as having malnutrition according to the MNA. Although the prevalence rates of malnutrition in older adults largely vary in the current literature due to the variations in both diagnostic tools and populations [5], our findings were almost identical to the results of other studies. Cereda et al. [33] reported in a meta-analysis that the prevalence of malnutrition and the risk of malnutrition assessed with MNA were 3.1% and 26.5% in community-dwelling older adults, respectively. Similarly, Murawiak et al. [34] reported a malnutrition rate of 3.3% and a risk of malnutrition rate of 28%. Moreover, the prevalence of malnutrition was 23.4% according to the GLIM criteria in the community-dwelling older adults in the SarcoPhAge study [35]. We also found that the prevalence of malnutrition defined by GLIM criteria was 24.5% in our earlier work [36]. Not so surprisingly, these findings perfectly match our results.

There were also contradictory findings in the literature. For example, contrary to our findings, the prevalence of malnutrition based on GLIM was reported to be 37.7% in Polish older adults [37]. The difference between the two studies may have originated from the methodology. The reduced muscle mass was calculated as the ratio of skeletal muscle mass and square of height (i.e. SMM/height²) in the Polish study [37], whereas our study adopted the SMM/W formula. These differing measurement methods may lead to variations in the classification of muscle loss and subsequently malnutrition, thus contributing to the observed discrepancies in prevalence rates. Another reason could be the difference between the two populations (i.e. cultural and dietary differences). Lastly, the mean number of diseases in the current study was 2, whereas it was 3 for the Polish study [37]. Therefore, it is rational to suggest that the differences in disease burden may contribute to the observed discrepancy in malnutrition prevalence between the studies. Nevertheless, it is important to highlight two shared findings of the current study and the Polish study [37]. Firstly, both studies reported a ‘fair’ agreement between the MNA and GLIM tools. Secondly, despite this fair agreement, both studies revealed a vast difference in malnutrition prevalence based on GLIM and MNA [37].

Our findings suggest that SO was linked with malnutrition defined by neither GLIM nor MNA. However, this does not conclusively prove that there is no association between SO and malnutrition. Considering the coexistence of SO and malnutrition in the current study (2.2%), this finding should be carefully interpreted. The lack of significance may be attributed to the relatively small sample size, with only 35 older adults with SO in the current study, which could limit the statistical power needed to detect an association. Another possible explanation is that current malnutrition assessment tools may not be fully applicable to obese individuals, as they often use a low BMI as a diagnostic criterion, which may not effectively capture malnutrition in those with excess fat mass [9, 38]. Additionally, cut-off values for weight loss in these tools may not be suitable for individuals with obesity, as they were based on studies of those with normal BMI [9, 39]. While alternative tools may be needed to assess malnutrition in this population, GLIM among the available tools seems to offer more reliable results for detecting malnutrition in older adults with SO [12].

These findings also don’t prove that SO and malnutrition have a causal relationship since the literature knowledge on this matter is inconclusive. Our previous work [40] suggested that SO and malnutrition defined by MNA were significantly associated with each other in community-dwelling older adults. On the other hand, Liu et al. [41], Perna et al. [42] and the current study found no association between SO and malnutrition. One possible explanation for these discrepancies is the use of different diagnostic criteria for SO across studies, making direct comparisons difficult. It is also noteworthy to point out that the older adults with SO in the latter studies may have benefited from the ‘obesity paradox’, where excess fat mass may offer some protective effect against malnutrition [42]. Nonetheless, we recommend future research to define SO based on ESPEN/EASO criteria and malnutrition based on GLIM criteria, thereby establishing a common basis for the evidence.

In addition, the overlap of MNA-defined malnutrition and SO was present in only one participant, whereas GLIM criteria detected 13 patients with malnutrition in participants with SO. Similarly to our findings, Murawiak et al. [34] reported that none of the SO patients was diagnosed with malnutrition based on MNA. The differences in the prevalence rates defined by these two tools originate from the use of divergent conceptual frameworks to diagnose malnutrition [43]. The MNA uses mostly subjective approaches to define malnutrition in obese older adults such as reduced food intake [44]. On the other hand, GLIM employs the combination of subjective and objective malnutrition criteria such as reduced muscle mass and inflammation as well as reduced food intake. Subsequently, this results in an increased likelihood of malnutrition detection while once more underlining the importance of using GLIM criteria for diagnosing malnutrition until a more effective approach is developed. Our findings regarding GLIM being able to detect more patients with malnutrition (i.e. eight times higher) than MNA support this notion as well.

The biggest strength of our study was the use of newly defined diagnostic criteria to define SO and malnutrition. We have evaluated the patients based on the latest consensus reports ESPEN/EASO and GLIM criteria for SO and malnutrition, respectively. Most importantly, this study was the first to investigate the association between SO defined by ESPEN/EASO criteria and malnutrition diagnosed by GLIM criteria. Another key strength of this study lies in its focus on evaluating the applicability and diagnostic utility of the GLIM criteria for malnutrition among community-dwelling older adults. While previous research has predominantly assessed the performance of the GLIM criteria in hospitalised settings, this study addresses a significant gap by investigating its effectiveness in outpatient or community-dwelling settings. Notably, our results cannot be generalised to all older adults since our study only includes relatively healthy, community-dwelling older adults without cognitive problems and active malignancy. Also, the number of participants is somewhat limited in this study since it was conducted in a single center. Further research should aim at larger cohorts in multiple sites.

Conclusion

Conclusively, we have found that the SO prevalence was 5.9% based on the newly defined ESPEN/EASO criteria in community-dwelling older adults of Türkiye. The number of participants classified as ‘having malnutrition’ was eight times higher when the diagnosis was based on GLIM compared to MNA. Thus, GLIM criteria were superior in diagnosing malnutrition compared to MNA. The coexistence of SO and malnutrition was observed in 2.2% of the population based on the GLIM criteria, whereas only one older adult (0.17%) was identified with both conditions according to the MNA criteria. Nonetheless, no association was detected between SO and malnutrition defined by GLIM or MNA. We suggest future research include GLIM criteria as a diagnostic tool for malnutrition and further investigate the association between SO and malnutrition.

Acknowledgements:

The authors would like to thank the participants for participating in this research.

Declaration of Conflicts of Interest:

None declared.

Declaration of Sources of Funding:

None declared.

Research Data Transparency and Availability:

There are no linked data sets for this paper. The data are confidential since the participants of this study were informed upon admission to the hospital that the data would remain confidential and would not be shared with third parties.

Declaration of Generative AI and AI-Assisted Technologies in the Writing Process:

The authors only used AI to correct the grammar in some sentences.

References