**1.**Characteristicsoftheincludedstudiesinthemeta-analysis.

#### *3.2. Quality of the Studies*

The assessment of the methodological quality of the studies included in the quantitative analysis is summarised in Table 2.


**Table 2.** The score obtained by the studies on the quality scale.

TS, Total score; Q1, Were the target population and the observation period well defined?; Q2, Diagnostic criteria; Q3, Method of case ascertainment; Q4, Administration of measurement protocol; Q5, Catchment Area; Q6, Prevalence measure. M.S. mean score; Qi stands for a quality rank.

When estimating a study quality, the mean score was 0.833 (range: 0.7–1). Three studies used bioelectrical impedance analysis (BIA) [33,36], three DXA [19,38,39] and only one study used a self-administered questionnaire [34]. Furthermore, three studies used the Pittsburg Sleep Quality Index (PSQI) [19,34,36], and the other three assessed the sleep duration/quality with self-reports. In addition, two studies [34,37] did not specify the case measurement system. The funnel plots suggest the presence of significant publication bias (Figure 2).

**Figure 2.** Funnel plot of the meta-analysis of the published studies. Each plotted point represents the.standard error (SE) and the prevalence. (**a**) Sleep well, (**b**) sleep poorly.

#### *3.3. Meta-Analysis*

The overall results of the prevalence of sarcopenia in the studies included in our meta-analysis revealed a high prevalence (Figure 3). When the methodological quality of the studies was considered in the results (Table 2), the prevalence was decreased (Total = 19,677 participants, 2858 cases; 0.144, 95% CI (0.100–0.189); Q = 41.90, *p* = 0.0000) but maintaining a high heterogeneity (*I* <sup>2</sup> = 88).

**Figure 3.** Overall prevalence of studies included in the analysis (method: random effects).

Results by Sleep Categories

The effects of the prevalence of sarcopenia sorted by categories are shown in Table 3. The prevalence values are grouped by sleep quality categories (sleep well and sleep poorly).


**Table 3.** Prevalence. Pooled results and CIs for three categories by transformation method and model.

Transf., transformation; HCI, higher CI; LCI, lower CI; SW, sleep well; SP, sleep poorly.

People who sleep well had lower values than those who sleep poorly and the prevalence in all the sub-analyses was high, suggesting a prevalence of sarcopenia independently of the category. However, the OR value was not significant (OR 0.81; 95% CI (0.41–1.60); Q = 34.04; *p* = 0.0000; test for overall effect, Z = 0.12, *p* = 0.91) when analysing the relationship between sarcopenia and sleep quality. Nonetheless, when the relationship between sleep quality and sarcopenia was analyzed after excluding the Buchmann et al. [19] and Ida et al. [34] studies from the analysis due to high heterogeneity, the sleep quality was associated with sarcopenia (OR 0.76; 95% CI (0.70–0.83); Q = 1.446; *p* = 0.695; test for overall effect, Z = 6.01, *p* < 0.00001). Likewise, the subjects who self-reported fewer sleeping hours showed a higher prevalence of sarcopenia.

Due to the high heterogeneity of the studies included in the meta-analysis, a gender analysis of prevalence was performed (Figure 4). Only four studies provided sex-dependent data. Non-significant associations for men (OR 1.61; 95% CI (0.82–3.16); Q = 11.80; *p* = 0.0189) or women (OR 0.77; 95% CI (0.29–2.03); Q = 21.35; *p* = 0.0003) were observed. However, the heterogeneity still showed high value in all the sub-analyses that were performed (including the quality of the studies and without

**Figure 4.** Prevalence of sarcopenia according to the sex of the participants. MSP, men sleep poorly; MSW, men sleep well; SWP, women sleep poorly; and WSW, women sleep well.

#### **4. Discussion**

The main finding of this research is that those subjects having inadequate sleep show a higher prevalence of sarcopenia values than those who reported adequate sleep. In addition, our results revealed a high prevalence of sarcopenia in older adults.

The results showed that a higher prevalence of sarcopenia values from those who do not sleep adequately were almost twice the value of the grouped prevalence, according to the model and the transformation that were used in the analysis. In line with our findings, Chien et al. [36], observed a significant association between sleep duration and the prevalence of sarcopenia on a sample of 488 adults (224 men and 264 women) from Taiwan, even though the assessment of sarcopenia was performed by electrical bioimpedance (BIA). Moreover, in the only study carried out in Europe, focused on the German subjects [19], similar results to those described above were observed, with the addition of the association between the sleep length and the quantity of muscle mass and recommending longitudinal studies to better understand the potential association. Similarly, Hu et al. [38] observed a relationship between sleep hours and sarcopenia in a Chinese cohort (*n* = 920, 95 cases). However, in this case, a U-shaped association in the prevalence of sarcopenia was obtained, in which older adults with short or long sleep length obtained higher values compared to those with normal sleep duration.

One plausible explanation to this findings is that the participants with an inefficient sleep may have differences in hormonal regulation (anabolic and catabolic balance), with elevated levels of cortisol (catabolic hormone promoting protein degradation), and low levels of IGF-1 (anabolic hormones promoting protein synthesis), developing a positive balance towards muscle degradation and, therefore, favoring the loss of muscle mass [19,40]. Likewise, Buchmann et al. [19] also observed elevated c-reactive protein (CRP) values. CRP is a pro-inflammatory cytokine and has been proposed as a possible cause of muscular atrophy [41] and also associated with sleep deprivation in high concentrations [42]. The sleep restriction generates hormonal imbalances and pro-inflammatory effects, favoring the loss of muscle mass with age. This could be one reason for the higher prevalence of sarcopenia values of sarcopenia in people with inadequate sleep. We must consider that the losses of muscle strength and muscle mass are associated and, therefore, related to a decrease in the functional capacity and quality of life [43]. Further studies to determine the effects of sleep deprivation in patients diagnosed with sarcopenia are necessary.

Interestingly, our results suggest a higher prevalence of sarcopenia in men compared to women (men = 0.19, 95% CI (0.14–0.25); women = 0.15, 95% CI (0.09–0.22)). These results are in line with previous studies in which the prevalence of sarcopenia can occur at earlier ages, as shown by Kwon et al. [39]. They observed a prevalence of sarcopenia of 14.3% in a group of 16,148 Koreans (44.1 ± 0.2 years), being higher in men (18.7%) than in women (9.7%). This can be explained by the fact that men had a higher muscle mass compared to women, but also a larger magnitude in muscle decrease was observed in men versus women as age increased [14]. On the contrary, in previous studies where the prevalence of sarcopenia was identified at different age intervals, a lower prevalence in men compared to women was observed [44]. Although age is the main causal effect of sarcopenia, the prediction ratio of increased sarcopenia based on age is difficult to verify, due to the multitude of factors that could have an influence in the prevalence values [45]. Nonetheless, it is estimated that the prevalence of clinically significant sarcopenia ranges from 8.8% in elderly women to 17.5% in elderly men, but it should be noted that these values may be higher or lower depending on the environmental factors [46]. In our study, only four articles considered gender as a sarcopenia-modifying variable, with very different methodologies and difficult interpretation. Therefore, the effect of sex on the prevalence of sarcopenia is unclear and more studies are needed to determine this interaction.

In summary, a direct association between sleep duration and prevalence of sarcopenia were confirmed in all the studies included in the quantitative data analysis. However, the interaction of gender and sleep duration/quality is not entirely clear. Hu et al. [38] observed that the prevalence of sarcopenia due to sleep deprivation was more pronounced in women. Similar results were described by Chien et al. [36] and Ida et al. [34]. However, Buchmann et al. [19] reported poor associations between sleep deprivation and the prevalence of sarcopenia in women. This discrepancy could be justified based on the ethnicity of the participants [2] or the age range difference between the studies. This higher and more evident prevalence in women could be associated with the negative effects of menopause. Thus, a decrease in the estrogen levels during menopause could play a potential role in decreasing the muscle mass after the fifth decade of life [13]. In addition, muscle mass seems to play an important role in osteoporosis in women, since muscle contractions involve a mechanical load on the bone that could promote the rate of bone regeneration [47]. Therefore, it could be stated that there is a close link between muscle strength, muscle mass, and bone tissue [48]; and that menopausal women are a sensitive population for the prevalence of sarcopenia, although to determine the gender role of this prevalence more studies would be needed. In addition, physical activity and programmed exercise should be considered, as it could play a relevant role in the prevalence of sarcopenia by improving sleep quality [49].

Finally, the results of this review should be interpreted with caution, since several limitations could be influencing them. For example, the high heterogeneity shown in the analyses could not be corrected by means of the rescaled bias scale. Another point to consider is the origin of the studied population. In five of the analyzed studies, the subjects cohort was from Asia, while only one single study was performed using a European population and the way of measuring the cut-off point and the provenance could be biasing the results [2], resulting in different tendencies between men and women. Other limitations are the way in which the sarcopenia [6] is conceptualized, resulting in prevalence variations due to the different techniques developed to measure sarcopenia, as well as the classification of the pathologic incidence [50] and that sleep quality was only self-reported throughout questionnaires and not objectively monitored; and the low number of articles included for the data for quantitative analysis (low security measure).

#### **5. Conclusions**

The main conclusion is the observed association between sleep duration/quality and the prevalence of sarcopenia. In addition, this prevalence seems to be higher in men than in women. These results could have a practical application for the public health since it can help us to consider sleep quality as a risk factor, as well as the need to incorporate therapies in order to improve the sleep quality and to reduce the negative effects of age-associated sarcopenia.

**Author Contributions:** Conceptualization, J.Á.R.-A., L.A. and R.R.-F.; methodology, R.R.-F.; software, J.Á.R.-A., A.M.-R. and L.M.M.-A.; formal analysis, J.Á.R.-A. and D.J.R.-C.; investigation, J.Á.R.-A., R.R.-F., L.A., L.M.M.-A., A.M.-R. and D.J.R.-C.; writing—original draft preparation, J.Á.R.-A.; writing—review and editing, R.R.-F., L.A., L.M.M.-A., A.M.-R. and D.J.R.-C.; supervision, J.Á.R.-A., R.R.-F., L.A., L.M.M.-A., A.M.-R. and D.J.R.-C.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors thank to G. Sanz for proofreading in English writing. No sources of funding were used in the preparation of this article.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Impact of Psychological Distress and Sleep Quality on Balance Confidence, Muscle Strength, and Functional Balance in Community-Dwelling Middle-Aged and Older People**

### **Raquel Fábrega-Cuadros, Agustín Aibar-Almazán \* , Antonio Martínez-Amat and Fidel Hita-Contreras**

Department of Health Sciences, Faculty of Health Sciences, University of Jaén, 23071 Jaén, Spain; rfabrega@ujaen.es (R.F.-C.); amamat@ujaen.es (A.M.-A.); fhita@ujaen.es (F.H.-C.)

**\*** Correspondence: aaibar@ujaen.es; Tel.: +34-953-213-408

Received: 19 August 2020; Accepted: 21 September 2020; Published: 22 September 2020

**Abstract:** The objective was to evaluate the associations of psychological distress and sleep quality with balance confidence, muscle strength, and functional balance among community-dwelling middle-aged and older people. An analytical cross-sectional study was conducted (*n* = 304). Balance confidence (Activities-specific Balance Confidence scale, ABC), muscle strength (hand grip dynamometer), and functional balance (Timed Up-and-Go test) were assessed. Psychological distress and sleep quality were evaluated by the Hospital Anxiety and Depression Scale and the Pittsburgh Sleep Quality Index, respectively. Age, sex, physical activity level, nutritional status, and fatigue were included as possible confounders. Multivariate linear and logistic regressions were performed. Higher values of anxiety (OR = 1.10), fatigue (OR = 1.04), and older age (OR = 1.08) were associated with an increased risk of falling (ABC < 67%). Greater muscle strength was associated with male sex and improved nutritional status (adjusted R<sup>2</sup> = 0.39). On the other hand, being older and using sleeping medication were linked to poorer functional balance (adjusted R<sup>2</sup> = 0.115). In conclusion, greater anxiety levels and the use of sleep medication were linked to a high risk of falling and poorer functional balance, respectively. No associations were found between muscle strength and sleep quality, anxiety, or depression.

**Keywords:** fall risk; balance; muscle strength; anxiety; depression; sleep quality

#### **1. Introduction**

Aging brings with it a series of changes that can affect the mobility and independence of people [1]. These changes affect the mood and attitude towards their environment, and this depends largely on the degree of acceptance of aging since it contributes to the feeling of happiness and satisfaction with life, whose lack can cause feelings of loneliness and sadness [2].

Certain disorders associated with this process, such as anxiety and/or depression, are psychological indicators of a decrease in quality of life [3]. Specifically, the prevalence of depression in the geriatric population worldwide is 7%, and its incidence increases with age [4]. Conversely, the prevalence of anxiety in people over 60 years old ranges between 0.7% and 18.6%, values far below those of younger adults [5].

Sleep quality is a key contributor to good health, and its importance among the older population cannot be overstated, given that sleep disorders and the difficulty to fall asleep become more common with age [6]. It has been shown that although the need to sleep remains the same throughout an individual's life, the ability to get enough sleep does in fact decrease with age. This brings about several adverse health outcomes such as reduced physical function, depression, increased risk of falls, and mortality [7].

Falls represent a major health care problem among older people and are linked to increased morbidity, mortality, and health costs [8]. Around 30% of older people living in the community experience a fall each year [9]. Fall risk factors have been studied in detail and include demographic, environmental, and health-related factors [10]. Balance confidence is one of the most important psychological factors linked to falls and the deterioration of balance, and its decrease can lead to diminished independence and participation in activities of daily living, thus creating a vicious circle that affects the quality of life and creates more isolated and dependent people [11]. On the other hand, the impaired functional balance has been shown to be one of the most important predictors of falls [12].

Muscle strength also declines with age more sharply than muscle mass [13]. It has been reported that muscle loss in older women decreases 3.7% per decade, however, strength decreases 15% per decade until age 70 when the loss accelerates considerably [14]. Moreover, in 2018 the European Working Group on Sarcopenia (EWGSOP2) listed low strength as a primary indicator of probable sarcopenia [15]. A decrease in muscle strength contributes to an elevated prevalence of falls and the loss of functional capacity and is a major cause of disability, mortality, and other adverse health outcomes [16].

Not many studies have examined the impact of psychological distress and sleep quality on balance confidence and function, and muscle strength in older people, which, in many cases, have shown inconclusive results and have focused on sleep duration or insomnia. Based on all of the above, the objective of this study was to evaluate the associations of psychological distress and sleep quality with the risk of falling according to balance confidence, functional balance, and muscle strength among community-dwelling middle-aged and older individuals.

#### **2. Experimental Section**

#### *2.1. Study Design and Participants*

An analytical cross-sectional study was conducted, to which end 315 community-dwelling middle-aged and older people were initially contacted and 304 finally took part. Recruitment was performed by contacting several senior centers from the Eastern Andalusia region. Prior to the beginning of the study, all participants provided their written informed consent. The research was approved by the Research Ethics Committee of the University of Jaén, Spain (NOV.18/2.TES), and was conducted in accordance with the Declaration of Helsinki, good clinical practices, and all applicable laws and regulations.

Community-dwelling ambulatory adults aged 50 years and older, able to understand and complete the required questionnaires and willing to give written informed consent to participate in the study were included in the protocol. Exclusion criteria were: conditions that limit physical activity, chronic and/or severe medical disease or any neuropsychiatric disorder that could influence their responses to the questionnaires.

#### *2.2. Study Parameters*

#### 2.2.1. Balance Confidence

The Activities-specific Balance Confidence scale (ABC) was used to assess balance confidence in the performance of activities of daily living [17]. This is a 16-item questionnaire that quantifies the level of confidence in performing a specific task without losing balance or becoming unsteady [18]. Each item score ranges from 0–100%, and the total score is obtained by summing the ratings (0–1600) and then dividing by 16. A higher percentage indicates a greater level of balance confidence. A score of <67% has been identified as a reliable means of predicting a future fall [19]. This cut-off was used to identify which participants were at high risk of falling.

#### 2.2.2. Muscle Strength

Muscle strength was assessed with an analog dynamometer (TKK 5001, Grip-A, Takei, Tokyo, Japan). Participants were required to apply their maximum handgrip strength three times with the dominant hand, each separated by 30 s. The maximal measured effort was regarded as their handgrip strength [20].

#### 2.2.3. Functional Balance

The Timed Up-and-Go (TUG) test [21] is a simple and valid method for predicting changes in functional balance in older adults [22]. It is a sensitive and specific measure for identifying community-dwelling adults who are at risk of falls [23]. In the TUG test, individuals rise from a seated position on a chair, walk three meters, turn around, return, and sit down again. The time required to complete this task was recorded.

#### 2.2.4. Sleep Quality

Sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI) [24,25]. It comprises 19 self-rated questions and 5 more (only used for clinical purposes) to be completed by bedmates or roommates. The 19 items (ranged from 0–3) generate a total score and seven different domains or subscales (subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbance, use of sleeping medication, and daytime dysfunction). Higher scores indicate poorer subjective sleep quality.

#### 2.2.5. Psychological Distress

The Hospital Anxiety and Depression Scale (HADS) is a self-administered questionnaire widely used to assess psychological distress in the general population [26,27]. This questionnaire contains 14 items, 7 related to anxiety symptoms, and 7 to depressive symptoms. Each item ranges from 0–3, and the total scores for both anxiety and depression range from 0 to 21, with higher scores indicating more severe symptoms.

#### 2.2.6. Fatigue Severity

In order to assess fatigue severity during the last 7 days, the Fatigue Severity Scale was used [28]. This questionnaire consists of 9 items (rated from 1–7) and produces a total score where larger values imply greater fatigue.

#### 2.2.7. Nutritional Status

The Mini Nutritional Assessment survey (MNA) was used to evaluate nutritional status [29,30]. It has 18 questions that include anthropometric measures, health status, dietary patterns, and subjective assessments of nutritional and health status. Higher scores indicate better nutritional status.

#### 2.2.8. Physical Activity Level

Physical activity level was assessed with the International Physical Activity Questionnaire-Short Form (IPAQ-SF) [31]. It consists of seven items that measure physical activity within three intensity levels (walking, moderate, and vigorous) during an average week. Physical activity was evaluated by combining the activity score of both moderate and vigorous-intensity activity (min/day) for each work and recreational activity domain. Responses were converted to Metabolic Equivalent Task minutes per week (MET-min / week) according to the scoring protocol.

#### *2.3. Sample Size Calculation*

For sample size calculation, at least 20 subjects per variable are required in the linear regression model [32], while a minimum of 10 subjects per variable was needed in the logistic regression model [33]. Since 15 possible predicting variables were considered (7 domains plus the total score of the PSQI, anxiety, depression, as well as physical activity level, nutritional status, fatigue, sex, and age as possible confounders), over 300 subjects were required for the purposes of our analysis. The final number of participants was 304.

#### *2.4. Statistical Analysis*

Continuous variables were described using means and standard deviations, and for categorical variables frequencies and percentages were used. The Kolmogorov–Smirnov test was performed to evaluate the normal distribution of the data. To analyze the differences between participants with and without risk of falling (ABC), Student's t test (continuous independent variables), and the Chi-squared test (sex) were used. In order to analyze the independent associations, a multivariate logistic regression was performed. Those variables with significant individual associations (*p* < 0.05) were selected for the stepwise logistic regression model. The odds ratio (OR) can be considered as significant when the 95% confidence interval (CI) does not include 1.00. The Chi-squared and Hosmer–Lemeshow tests were conducted to assess the overall goodness-of-fit for each of the steps of the model, as well as for the final model. To explore the possible individual associations of muscle strength and functional balance with PSQI, HADS, FSS, MNA, and IPAQ-SF scores, as well as with age (independent variables), Pearson's correlation was used. As for the analysis of the independent associations, the same procedure was applied, but using a stepwise multivariate linear regression. Functional balance and muscle strength were individually introduced as dependent variables in separate models. We first looked into the bivariate correlation coefficients, and any independent variables with significant associations (*p* < 0.05) were included in the multivariate linear regression. Adjusted R<sup>2</sup> was used to calculate the effect size coefficient of multiple determination in the linear models. R<sup>2</sup> can be considered insignificant when <0.02, small if between 0.02 and 0.15, medium if between 0.15 and 0.35, and large if >0.35 [34]. A 95% confidence level was used (*p* < 0.05). Data management and analysis were performed using the SPSS statistical package for the social sciences for Windows (SPSS Inc., Chicago, IL, USA).

#### **3. Results**

A total of 304 participants (72.04 ± 7.88 years) took part in this study. When studying the ABC score (23.42 ± 7.25), 24.01% of participants were at risk of falling. The analysis revealed (Table 1) that participants with an ABC score < 67 were individually associated with the largest values of anxiety (*p* = 0.002), depression (*p* = 0.001), fatigue (*p* = < 0.001), increased age (*p* < 0.001), and worse nutritional status (*p* = 0.002).


**Table 1.** Individual differences according to the risk of falling.


**Table 1.** *Cont*.

ABC: Activities-Specific Balance Confidence Scale. MET: Metabolic Equivalent of Task. PSQI: Pittsburgh Sleep Quality Index. SD: Standard Deviation.

The multivariate logistic regression that looked into the risk of falls as assessed with the ABC score revealed several significant results. Higher values of anxiety (OR = 1.10, 95% CI = 1.02–1.18), fatigue (OR = 1.04, 95% CI = 1.02–1.06), and older age (OR = 1.08, 95% CI = 1.04–1.12) were independently associated with ABC scores < 67%. The Hosmer–Lemeshow test showed a good fit of the model (Chi-squared = 2.403, *p* = 0.966), which was able to classify correctly 78.29% of all participants at high risk of suffering a future fall, according to the ABC score (Table 2).

**Table 2.** Multivariate logistic regression analyses for factors associated with the risk of falling (determined through the ABC score).


ABC: Activities-Specific Balance Confidence Scale. CI: Confidence Interval. OR: Odds Ratio.

As for functional balance (9.86 ± 2.91 s) and muscle strength (19.43 ± 6.42 kg), the individual associations are shown in Table 3. Muscle strength was only associated with anxiety (*p* = 0.001), fatigue (*p* = 0.020), and nutritional status (*p* = 0.038), whereas poor functional balance was related to greater age (*p* < 0.001) and physical activity level (*p* = 0.035), as well as with the use-of-sleeping-medication domain in PSQI (*p* = 0.028). Regarding sex differences, men displayed greater muscle strength (both *p* < 0.001), but worse functional balance (*p* = 0.005).

**Table 3.** Pearson's correlations of functional balance and muscle strength, with PSQI scores and possible confounders.



**Table 3.** *Cont*.

MET: Metabolic Equivalent of Task. PSQI: Pittsburgh Sleep Quality Index. r: Pearson's Correlation Coefficient.

Lastly, the linear regression analysis (Table 4) revealed that greater muscle strength was independently associated with the male sex (*p* < 0.001) and a better nutritional status (*p* = 0.001), and the effect size was large (adjusted R<sup>2</sup> = 0.392). On the other hand, being older (*p* < 0.001) and the use of sleeping medication (*p* = 0.033) were linked to poorer functional balance, although the effect size was small (adjusted R<sup>2</sup> = 0.115).

**Table 4.** Multivariate linear regression analyses for functional balance and muscle strength.


B: Unstandardized Coefficient. β: Standardized Coefficient. CI: Confidence Interval. MET: Metabolic Equivalent of Task. PSQI: Pittsburgh Sleep Quality Index.

#### **4. Discussion**

The objective of this study was to evaluate the associations of psychological distress and sleep quality with balance confidence, functional balance, and muscle strength among community-dwelling middle-aged and older individuals. In our study, anxiety, fatigue, older age, and the use of sleeping medication were shown to be associated with the risk of falling and poorer functional balance. Muscle strength was associated with being male and nutritional status.

In general, balance confidence scores are able to predict perceived physical function and even mobility in older adults [35]. Similar to our own study, a previously published paper also employed regression models to find a significant association of anxiety with confidence in balance, while depression was shown to be associated with avoidance of activity [36]. A systematic review with meta-analysis found an association between balance confidence and anxiety [37], and a similar link was established between depression and level of physical activity [38]. Regarding the association of age with balance confidence, Medley et al. [39] reported that participants with low balance confidence were older than those who reported high balance confidence. In our study, only anxiety, age, and fatigue were independently associated with the balance confidence. To our knowledge, this is the first study to observe an association between confidence in balance and fatigue in healthy middle-aged and older people, although there are studies that demonstrate this association, but in people with some pathology [40,41].

Muscle strength plays an important role in the execution of many activities of daily living and is considered an indicator of functional decline among community-dwelling older adults [42]. Low grip strength is predictive of poor outcomes and indicative of prolonged hospital stays, increased functional limitations, poor quality of life, and death [43]. For example, it has been shown that people who have a high level of grip strength have a significantly lower fear of falling than those who show lower levels [44,45]. In addition, it has been observed that the strength of the abductor muscles can identify older adults at risk of falling [46]. A recent study looking into the association between falls and lower-limb strength failed to find any link at a one-year follow-up [47]. Our study found no associations whatsoever between muscle strength and sleep quality, and increased muscle strength was independently associated only with being male (as in previous studies by Buchman et al. [48]) and with improved nutritional status. Other authors have agreed before that a poor diet is significantly associated with lower muscle strength, but they also linked it to lower physical function, longer TUG test time, depression, and risk of falling [49], although the results of the present study should be interpreted with caution since they are correlations and a cause-effect relationship cannot be established. Some recent studies even recommend the intake of supplementary proteins given their significant effect in increasing muscle mass and strength among elderly people with sarcopenia [50]. We must consider, however, that disparities in the literature may be due to a variety of population ages, measurement methods, and educational and cultural levels, which may have a confounding effect.

Balance confidence contributes to functional mobility performance [39], and there seems to be a strong link between balance self-efficacy and function capabilities [51]. A study by Brandão et al. [52] identified an association between excessive daytime sleepiness and quality of life, and also characterized the profile of older adults with poor sleep quality. Sleep duration is associated with inflammation markers (serum interleukin-6, tumor necrosis factor α, and C-reactive protein) in older adults, and in turn with mortality [53]. Loss of functional balance, as measured by the TUG test, is known to be one of the first signs of aging and is considered a marker for general health that is strongly associated with the risk of mortality [54]. In our results, and as far as individual associations are concerned, higher age, poorer sleep quality (use of sleep medication), and decreased levels of physical activity were linked to lower TUG scores. However, in the multivariate analysis model, such associations only held for the first two variables (age and poor sleep quality). The results of a study conducted among women indicate that a shorter sleep duration increased wakefulness after sleep onset, and decreased sleep efficiency are risk factors for functional or physical impairment in older women [55].

There are some limitations to our study that must be acknowledged. Firstly, its cross-sectional design did not allow for the evaluation of causal relationships. Secondly, sleep quality was assessed using self-report methods, and therefore the influence of recall bias must be considered. Thirdly, our study was conducted among people from a specific geographic area, and any generalization of its results should be limited to individuals with characteristics similar to those of our population sample. Future studies should consider exploring prospective designs, employing objective sleep quality assessment methods (polysomnography or actigraphy), and applying them to a general population of older adults.

#### **5. Conclusions**

Among middle-aged and older Spanish people, greater levels of anxiety and fatigue, as well as older age were associated with an increased risk of falling (assessed with the Activities-specific Balance Confidence scale). No associations were found with sleep quality and depression. Greater muscle strength was associated with being male and having a better nutritional status. Finally, increased age and the use of sleeping medication were linked to poorer functional balance.

**Author Contributions:** Conceptualization: R.F.-C. and F.H.-C.; methodology: R.F.-C. and A.A.-A.; formal analysis: F.H.-C. and R.F.-C.; supervision: A.M.-A. and A.A.-A.; writing—original draft preparation: R.F.-C. and F.H.-C.; writing—review and editing: A.M.-A. and A.A.-A.; funding acquisition: A.M.-A. and F.H.-C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the project 1260735, integrated into the 2014–2020 Operational Programme FEDER in Andalusia.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Brief Report* **Dynapenia and Low Cognition: A Cross-Sectional Association in Postmenopausal Women**

**Julie A. Pasco 1,2,3,4,\* , Amanda L. Stuart <sup>1</sup> , Sophia X. Sui <sup>1</sup> , Kara L. Holloway-Kew <sup>1</sup> , Natalie K. Hyde <sup>1</sup> , Monica C. Tembo <sup>1</sup> , Pamela Rufus-Membere <sup>1</sup> , Mark A. Kotowicz 1,2,3 and Lana J. Williams <sup>1</sup>**


**Abstract:** Dynapenia is a key contributor to physical frailty. Cognitive impairment and dementia accompany frailty, yet links between skeletal muscle and neurocognition are poorly understood. We examined the cross-sectional relationship between lower limb muscle strength and global cognitive function. Participants were 127 women aged 51–87 years, from the Geelong Osteoporosis Study. Peak eccentric strength of the hip-flexors and hip abductors was determined using a hand-held dynamometer, and dynapenia identified as muscle strength *t*-scores < −1. Cognition was assessed using the Mini-Mental State Examination (MMSE), and MMSE scores below the median were rated as low. Associations between dynapenia and low cognition were examined using logistic regression models. Hip-flexor dynapenia was detected in 38 (71.7%) women with low cognition and 36 (48.7%) with good cognition (*p* = 0.009); for hip abductor dynapenia, the pattern was similar (21 (39.6%) vs. 9 (12.2%); *p* < 0.001). While the observed difference for hip-flexor strength was attenuated after adjusting for age and height (adjusted Odds Ratio (OR) 1.95, 95%CI 0.86–4.41), low cognition was nearly 4-fold more likely in association with hip abductor dynapenia (adjusted OR 3.76, 95%CI 1.44–9.83). No other confounders were identified. Our data suggest that low strength of the hip abductors and low cognition are associated and this could be a consequence of poor muscle function contributing to cognitive decline or vice versa. As muscle weakness is responsive to physical interventions, this warrants further investigation.

**Keywords:** cognition; brain-body cross-talk; muscle strength; older persons; sarcopenia

#### **1. Introduction**

Dynapenia refers to age-associated loss of skeletal muscle strength [1]. From about age 50 years, muscle strength declines by 10–15% per decade up to age 70 years, reaching losses of 25–40% per decade thereafter [2,3]. The rate of decline in muscle strength surpasses age-related loss of skeletal muscle mass and is a key contributor to physical frailty; physical frailty is known to accompany cognitive impairment and dementia [4].

Low muscle strength is also a key characteristic of sarcopenia. This is evident in the revised operational definition from the European Working Group on Sarcopenia in Older People (EWGSOP2), which focuses on low muscle strength as the primary parameter of sarcopenia; low muscle mass (or quality) confirms the diagnosis, and poor physical performance identifies severe sarcopenia [5]. Sarcopenia has been associated with cognitive impairment and Alzheimer's disease [6].

Muscle deterioration during ageing is a consequence of decreases in the number and cross-sectional area of muscle fibres [7] and reductions in the number of motoneurons [8].

**Citation:** Pasco, J.A.; Stuart, A.L.; Sui, S.X.; Holloway-Kew, K.L.; Hyde, N.K.; Tembo, M.C.; Rufus-Membere, P.; Kotowicz, M.A.; Williams, L.J. Dynapenia and Low Cognition: A Cross-Sectional Association in Postmenopausal Women. *J. Clin. Med.* **2021**, *10*, 173. https://doi.org/10.3390/jcm10020173

Received: 8 December 2020 Accepted: 4 January 2021 Published: 6 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Thus, loss of muscle strength in older people is attributable, at least in part, to neurologic mechanisms that alter the number of functioning motor units [9]. However, links between dynapenia and cognitive function are poorly understood. As handgrip strength is easily measured, predicts adverse health outcomes, and is considered to indicate global skeletal muscle strength [10], EWGSOP2 recommends handgrip strength for assessing muscle strength in the diagnosis for sarcopenia. However, not all studies support good agreement between handgrip and lower limb muscle strength [11]. Several reviews have described associations between upper body strength measures and cognition by assessing handgrip strength [12–14], but links between lower body strength measures (e.g., hip flexor or hip abductor strength) have not been well investigated. The main goal of this study was to examine the relationship between lower limb skeletal muscle strength and global cognitive function in older women (>50 years), as older ages are at highest risk for age-related physical and mental decline. We hypothesise that dynapenia will be associated with low cognition.

#### **2. Materials and Methods**

An age-stratified, population-based cohort of 1494 women (age 20–94 years) was recruited for the Geelong Osteoporosis Study between 1994 and 1997, with 77.1% response, using a random-selection process from electoral rolls [15]; the cohort was predominantly Caucasian (~98%). A listing on the electoral roll for the Barwon Statistical Division fulfilled inclusion criteria; residence in the region for less than 6 months and inability to provide informed consent necessitated exclusion. Six years later (2000–2003), 638 of 1048 women who were re-assessed at follow-up were aged over 50 years. Among these, 127 (ages 51–87 years) provided measures of cognitive function in combination with muscle strength, weight, height, and information about their lifestyle behaviours, meeting criteria for inclusion in this analysis. There were no further exclusion criteria for the analysis. Written informed consent was obtained from all participants. The Barwon Health Human Research Ethics Committee approved the study (project 92/01).

Global cognitive function was assessed using the Mini-Mental State Examination (MMSE) [16]; scores below the median were rated as low cognition and scores above the median were rated as good cognition. Peak eccentric muscle strength of the hip flexors and abductors were measured using a hand-held dynamometer (HHD; Nicholas Manual Muscle Tester, Lafayette Instrument Company, Lafayette, IN, USA) [17]. HHD provides good to excellent reliability and validity when compared with fixed dynamometry for most measures of isometric lower limb strength [18]. To measure hip flexion strength, the seated participant raised the test thigh 10 cm above the bench; with the HHD positioned 5 cm proximal to the patella, the examiner applied a downward force while the participant resisted, until resistance could no longer be sustained. For hip abduction strength, the sidelying participant raised the outstretched test leg 20 cm above the bench; the HHD was positioned 10 cm proximal to the lateral malleolus. Measurements were repeated bilaterally, and the maximum of triplicate measures for each muscle group was used in analyses. Multiplying the maximal registered force (kg) by 9.81 converted the force to Newtons (N). Dynapenia refers to muscle strength *t*-scores < 1 for hip flexors and hip abductors [19]. All MMSE tests were conducted by one of the authors (A.L.S.), and HHD assessments were performed by other trained research personnel.

Data on current smoking, alcohol use, and mobility were collected by self-report. The usual consumption of different alcoholic beverages was recorded as glasses per week, and the average daily total was categorised as <1 or ≥1 glass/day. Participants with mobility described as someone who 'moves, walks, and works energetically and participates in vigorous exercise' were classified as being physically active.

Descriptive characteristics of participants were presented as mean (±SD), median (IQR), or *n* (%). Intergroup differences were identified by Student's *t*-test for parametric data, Mann–Whitney test for non-parametric data, and chi-square test for categorical data. Associations between dynapenia (exposure) and low cognitive performance (outcome)

were examined using binary logistic regression models before and after adjusting for potential confounders and effect modifiers. Statistical analyses were performed using Minitab (version 16, Minitab, State College, PA, USA).

#### **3. Results**

Participant characteristics are listed in Table 1. The median MMSE score was 29 (range 22–30). Mean muscle strength measures for hip flexors and abductors were lower for women with low cognition in comparison with those with good cognition. The group with low cognition was older, had shorter stature, and was less likely to consume, on average, one or more alcoholic drinks each day; otherwise no other differences were detected.


**Table 1.** Participant characteristics for the whole group and according to categories of cognition. Data are shown as mean (±SD), median (interquartile range, IQR), or number (%).

\* MMSE Mini-Mental State Examination. \*\* BMI: Body Mass Index.

While hip flexor dynapenia was detected in 38 (71.7%) women with low cognition and 36 (48.7%) with good cognition (*p* = 0.009), this difference was attenuated after adjustments were made for age and height (adjusted Odds Ratio (OR) 1.95, 95%CI 0.86*–*4.41, *p* = 0.110).

For hip abductors, dynapenia was detected for 21 (39.6%) women with low cognition and 9 (12.2%) of those with good cognition (*p* < 0.001). This association was sustained after adjustment for age and height; those with hip abductor dynapenia were nearly four-fold more likely to have low cognition (adjusted OR 3.76, 95%CI 1.44*–*9.83, *p* = 0.007). Body Mass Index (BMI), smoking, alcohol consumption, and mobility did not contribute to the model. No effect modifiers were identified.

#### **4. Discussion**

Here we present data that describe a cross-sectional association between lower limb muscle strength and global cognitive function in postmenopausal women. These findings are concordant with cross-sectional data from the I-Lan Longitudinal Aging Study (ILAS) involving 731 elderly men and women (mean age 73.4 years) for whom MMSE-derived global cognitive impairment was associated with low handgrip strength (adjusted OR 2.23, 95%CI 1.29–3.86) [20]. Similarly, cross-sectional data for 292 men (ages 60–96 years) from the Geelong Osteoporosis Study revealed that handgrip strength was associated with psychomotor function and overall cognitive performance assessed by the CogState Brief Battery [21], and another cross-sectional study of 39 men (ages 61–79 years) reported that knee extensor strength was positively associated with global cognitive function also

assessed by MMSE [22]. Moreover, there are longitudinal data that describe increases in muscle strength [23] and torque [24] in association with improvements in cognitive performance following loading of skeletal muscle through progressive resistance training regimens. Conclusions from a recent systematic review were that resistance training elicited functional changes in the brain, including improvements in executive function [25].

Previous work in animal models suggest that restriction of skeletal muscle activity in the hind legs of mice affect neurogenic areas of the brain [26] and produce changes in memory and spatial learning [27]. These findings align with observations in the literature that physical inactivity is a common risk factor for Alzheimer's disease and, conversely, that voluntary exercise improves cognitive function [28,29].

Cognitive changes following skeletal muscle loading may operate through the release from contracting muscle of chemical messengers, such as brain-derived neurotrophic factor (BDNF), which trigger neurobiological changes [29,30]. Age-related declines in skeletal muscle and cognitive capabilities have common pathophysiological pathways, including chronic inflammation, oxidative stress, and endocrine imbalances; they also share risk factors associated with adverse lifestyle behaviours, such as physical inactivity, smoking, and excessive use of alcohol, that might contribute to brain–body cross-talk by mediating these pathophysiological pathways [31–34]. While we accounted for differences in body habitus and lifestyle behaviours, biomarker data that indicate biological imbalances were not available.

However, an important forte of our study design is that assessment of muscle strength and cognitive function were obtained independently, thereby minimising potential differential measurement bias. We also recognise as weaknesses that the data are cross-sectional, and there is the possibility that low cognitive function might have impacted on the ability to perform the muscle strength tests. Although several variables, including lifestyle behaviours, were considered as confounders, residual confounding is likely. It is noted that our results should be interpreted with caution because of the small numbers of participants in some sub-groups (such as nine women with hip abductor dynapenia and good cognition), and we recognise that the findings may not be generalisable beyond our sample of white postmenopausal women. Moreover, while we have investigated only lower limb muscle strength in association with global cognition, there is scope to explore this further using other indices of muscle performance and cognitive function in specific domains.

#### **5. Conclusions**

Our cross-sectional analyses suggest that low muscle strength and low cognition are associated, and this could be related to muscle function deficits contributing to cognitive decline or vice versa. Longitudinal studies are needed to address temporal changes in skeletal muscle performance and cognitive function. We propose that future cross-sectional studies use different neuroimaging methods (e.g., functional near-infrared spectroscopy, electroencephalography, functional magnetic resonance imaging) and/or assess neurochemical substances (e.g., neurotransmitters, neurotrophic factors) in order to further elucidate the underlying neurobiological mechanisms of the relationship between skeletal muscle strength and cognitive performance. Emerging pharmacological therapies for preventing muscle loss, such as antibodies against myostatin and activin receptor types IIA and IIB [35,36], might provide potential novel agents for the management of cognitive decline. Similarly, pharmacological therapies for preventing cognitive decline might provide potential novel agents for the management of muscle decay. There is evidence to support life course approaches for preventing and managing physical and cognitive performance [29,37–40]. Thus, there may be potential for future clinical trials involving such novel therapies, together with interventions focused on specific modes of exercise, diet, and health behaviours that could be championed as public health strategies for both physical and mental health benefits.

**Author Contributions:** Conceptualization: J.A.P.; methodology: J.A.P. and L.J.W.; formal analysis: J.A.P.; investigation: A.L.S., K.L.H.-K., N.K.H., M.C.T., and P.R.-M.; resources: J.A.P., M.A.K., and L.J.W.; data curation: J.A.P.; writing—original draft preparation: J.A.P.; writing—review and editing, J.A.P., A.L.S., S.X.S., K.L.H.-K., N.K.H., M.C.T., P.R.-M., M.A.K., and L.J.W.; project administration: J.A.P. and M.A.K.; funding acquisition: J.A.P., M.A.K., and L.J.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** The Geelong Osteoporosis Study was funded by the National Health and Medical Research Council (NHMRC), Australia (projects 251638, 628582). S.X.S., M.C.T., and P.R.M. were supported by Deakin Postgraduate Scholarships; K.L.H.-K. was supported by an Alfred Deakin Postdoctoral Research Fellowship; N.K.H. was supported by a Deakin Postdoctoral Research Fellowship; and L.J.W. was supported by a NHMRC Career Development Fellowship (1064272) and a NHMRC Investigator grant (1174060). However, these funding organisations played no role in the design and conduct of the study, in the collection, management, analysis and interpretation of the data, or in the preparation, review, and approval of the manuscript.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Human Research Ethics Committee at Barwon Health (project 92/01).

**Informed Consent Statement:** Informed consent was obtained from all participants involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical restrictions.

**Acknowledgments:** The authors acknowledge the study participants.

**Conflicts of Interest:** The authors declare no conflict of interest. J.A.P. has received grants from the National Health and Medical Research Council (NHMRC) Australia, the BUPA Foundation, Amgen-GSK OA-ANZBMS, Amgen Australia, Deakin University, the Western Alliance, and the Beischer Foundation; K.L.H.-K. has received grants from Deakin University, Amgen-GSK OA-ANZBMS and Amgen Australia; N.K.H. has received grants from Deakin University and the Beischer Foundation; M.A.K. has received grants from the NHMRC, Amgen-GSK OA-ANZBMS, Amgen Australia, and the Western Alliance; L.J.W. has received grants from Deakin University and the NHMRC.

#### **References**


### **Oxidative Stress, Telomere Shortening, and Apoptosis Associated to Sarcopenia and Frailty in Patients with Multimorbidity**

**Máximo Bernabeu-Wittel 1,\* , Raquel Gómez-Díaz <sup>2</sup> , Álvaro González-Molina <sup>1</sup> , Sofía Vidal-Serrano <sup>3</sup> , Jesús Díez-Manglano <sup>4</sup> , Fernando Salgado <sup>5</sup> , María Soto-Martín 6 , Manuel Ollero-Baturone <sup>1</sup> and on behalf of the PROTEO RESEARCHERS** †


Received: 6 August 2020; Accepted: 14 August 2020; Published: 18 August 2020

**Abstract:** Background: The presence of oxidative stress, telomere shortening, and apoptosis in polypathological patients (PP) with sarcopenia and frailty remains unknown. Methods: Multicentric prospective observational study in order to assess oxidative stress markers (catalase, glutathione reductase (GR), total antioxidant capacity to reactive oxygen species (TAC-ROS), and superoxide dismutase (SOD)), absolute telomere length (aTL), and apoptosis (DNA fragmentation) in peripheral blood samples of a hospital-based population of PP. Associations of these biomarkers to sarcopenia, frailty, functional status, and 12-month mortality were analyzed. Results: Of the 444 recruited patients, 97 (21.8%), 278 (62.6%), and 80 (18%) were sarcopenic, frail, or both, respectively. Oxidative stress markers (lower TAC-ROS and higher SOD) were significantly enhanced and aTL significantly shortened in patients with sarcopenia, frailty or both syndromes. No evidence of apoptosis was detected in blood leukocytes of any of the patients. Both oxidative stress markers (GR, *p* = 0.04) and telomere shortening (*p* = 0.001) were associated to death risk and to less survival days. Conclusions: Oxidative stress markers and telomere length were enhanced and shortened, respectively, in blood samples of polypathological patients with sarcopenia and/or frailty. Both were associated to decreased survival. They could be useful in the clinical practice to assess vulnerable populations with multimorbidity and of potential interest as therapeutic targets.

**Keywords:** multimorbidity; polypathological patients; sarcopenia; frailty; oxidative stress; telomere length; apoptosis

#### **1. Introduction**

As a result of populations' aging throughout the world, the prevalence of chronic conditions has drastically increased; these coexist frequently in the same patient, conditioning deleterious relationships, faster clinical and functional deterioration, poorer quality of life, and higher mortality. Taking multimorbidity seriously is of nuclear importance for the sustainability of all healthcare

systems [1–3]. Multimorbidity is narrowly correlated to aging. As a matter of fact, there is a direct and strong correlation between the development of different chronic conditions and longevity. The easy explanation of this correlation is the longer exposure to different risk factors (environmental agents, unhealthy lifestyles, inherited risk factors, overuse deterioration), which impact in the development of diseases and failures of multiple organs and systems [3,4].

Sarcopenia and frailty are two major geriatric syndromes closely related to the aging process [5,6]. The development of one or both of them is linked to progressive functional disability, loss of quality of life and death. Their prevalence in elderly populations approximates 10% and 15%, respectively; however, in the presence of chronic conditions and multimorbidity, these prevalences can raise to 20% and 60%, respectively [7]. Both syndromes are narrowly interrelated; as a matter of fact, they have currently an identical therapeutic approach based on physical activity and optimal nutrition. In a recent study, both sarcopenia and frailty were present in the same patient in 18% of the studied cases; that is to say that most sarcopenic patients were frail, and about one third of frail patients were sarcopenic [7]. Nevertheless, these percentages are different in other studies, probably due to sample selection criteria [8].

Both syndromes have commonalities sharing a nuclear issue, which is the physical function impairment, usually assessed by different tools like walking speed and hand grip strength. Such impairment may be responsible for the concurrent existence of a disability in both phenotypes, but they also express differences; as a matter of fact, sarcopenia rather tends to assume the lineaments of cachexia and "muscle wasting", whereas frailty status is largely dominated by a low physical performance, homeostasis disruption to stressors, and disabling condition.

The deep and intimal relation between sarcopenia and frailty probably reflects that they share similar or identical pathophysiological routes and molecular mechanisms. In this field, many metabolic imbalances and other molecular factors have been studied and correlated in some ways to both geriatric syndromes. Sarcopenia has been associated to genetic expression of apoptosis and muscular autophagy, muscle androgenic and vitamin D receptors, chronic inflammation, oxidative stress, and telomere shortening [9–11]. On the other hand, frailty has been associated to inflammation pathways (demonstrated in the case of C-reactive protein, interleukin-1β, the IL-1 receptor antagonist, IL-18, and tumor necrosis factor alpha), unspecific immunological alterations linked to immunosenescence (mainly thymus involution and the corresponding decrease of T and B lymphocyte precursors and the reduction in the proliferative capacity of the T and B lymphocytes), and oxidative stress [12–14].

From these data, the narrow relation between frailty and sarcopenia can be extracted. This is more so in patients with multimorbidity, in which aging and chronic conditions may trigger more oxidative stress, telomere shortening, and apoptosis. In these patients, sarcopenia and frailty could be the results of a multisite "rusting" produced by chronic inflammation processes and their consequent imbalance between the production of reactive oxygen species (ROS) and cellular antioxidant defenses, present in chronic neurological, pulmonary, and cardiovascular diseases, along with atherosclerosis, diabetes, obesity, and arthritis. Nevertheless, the role and weight of any of these molecular alterations in sarcopenic and/or frail populations with multimorbidity remain unknown.

For all these reasons, we have explored the main oxidative stress markers, telomere length, and apoptosis parameters in a hospital-based multicenter cohort of multimorbidity patients. We hypothesized that all these biological markers have a deep impact and association to sarcopenia and frailty.

#### **2. Patients and Methods**

#### *2.1. Development of the Study*

This was a prospective observational, multi-institutional (6 centers) study carried out by researchers from the Polypathological Patient and Advanced Age Study Group of the Spanish Society of Internal Medicine (all participant centers are listed on the PROTEO Researchers list). The study was approved by the ethics committee of all participant centers. The study inclusion period ranged from January 2012 to March 2016.

All patients treated in the Internal Medicine and Geriatric areas who accomplished inclusion criteria (≥18 years old and fulfilling criteria of polypathological patient (PP)) were included, after providing their written informed consent. The patient's sample was collected by performing prevalence surveys every 14 days during the study period. A total of 155 surveys were performed (29 ± 19 surveys per hospital).

After receiving informed consent, a complete set of demographical, socio-familial, clinical, functional, biological, and pharmacological data were collected from all included patients.

Sarcopenia was defined following EWGSOP criteria [15]. This was established by the presence of a gait speed <sup>≤</sup> 0.8 m/seg, plus a skeletal muscle mass <sup>&</sup>lt;6.76 Kg/m<sup>2</sup> in women, and <10.76 Kg/m<sup>2</sup> in men (for those patients able to walk) or a hand grip strength lower than 50 percentile of his/her age group and gender, and a skeletal muscle mass <6.76 Kg/m<sup>2</sup> in women and <10.76 Kg/m<sup>2</sup> in men (for those patents unable to walk). Frailty was defined when fulfilling 3 or more of Fried's criteria (slowness, weakness, weight loss, exhaustion, and low physical activity) [16].

All patients were clinically followed during a 12-month period in order to assess mortality, as previously described [7]. Time survival was assessed, and in case of death, chronology of the demise was incorporated. Therefore, we looked at mortality as a time-dependent outcome. For the dichotomous outcome, subjects were categorized depending on whether or not they survived 12 months from their initial interview date. For the continuous outcome, survival time was defined as the number of days between the baseline interview and the date of death. All these data were collected by clinicians in charge who were active members of the investigation team.

Ethics Committee Approval: The present study has been approved by the ethics committee of all participant centers (ethical approval code: CEI2012/PI242). Ethical Guidelines for Authorship and Publishing: The authors certify that they comply with the ethical guidelines for publishing in the Journal Clinical Medicine.

#### *2.2. Biological Parameters Determination*

We determined blood or plasma biological parameters of all included patients, including oxidative stress markers, apoptosis expression, and telomere length.

Oxidative stress markers: We determined activity/levels of catalase, glutathione reductase (GR), total antioxidant capacity to reactive oxygen species (TAC-ROS), and superoxide dismutase (SOD). Colorimetric studies were performed using a monochromator-based UV–VIS spectrophotometer (Multiskan® GO; Thermo Fisher Scientific Corporation, Carlsbad, CA, USA).

Catalase activity (nmol/min/mL) was measured in patients' plasma using the colorimetrical procedure provided by Cayman's Catalase Assay Kit, Item No. 707002 (Cayman Chemical, Ann Arbor, MI, USA). The method is based on the reaction of the enzyme wit methanol in the presence of an optimal concentration of H2O2. The formaldehyde produced is measured colorimetrically with Purpald as the chromogen. Purpald specifically forms a bicyclic heterocycle with aldehydes, which upon oxidation changes from colorless to purple color [17,18].

Glutathione reductase activity (U/mL; 1 Unit = the amount of enzyme that will cause the oxidation of 1.0 nmol of NADP to NADP+ per minute at 25 ◦C) was analyzed in patients' plasma by measuring the rate of NADPH oxidation, using for this purpose the Cayman's Glutathione reductase Assay Kit, Item No. 703202 (Cayman Chemical, Ann Arbor, MI, USA). The oxidation of NADPH is accompanied by a decrease in absorbance at 340 nm and is directly proportional to the GR activity in the sample [18,19].

Total antioxidant capacity to reactive oxygen species (mM Trolox equivalents) was analyzed measuring the ability of patients' plasma antioxidants to inhibit the oxidation of ABTS® (2,2′ -Azino-di-3-ethylbenzthiazoline sulphonatel) to ABTS®+ by metmyoglobin. For this purpose, the Cayman's Antioxidant Assay Kit, Item No. 709001 (Cayman Chemical, Ann Arbor, MI, USA) was used. The antioxidants cause suppression of the absorbance at 750 nm or 405 nm to a degree that is

proportional to their concentration. This capacity of the antioxidants is compared to that of Trolox, a water-soluble tocopherol analogue, and is quantified as millimolar Trolox equivalents [20,21].

Superoxide dismutase activity (U/mL) was measured in patients' plasma using the colorimetrical absorbance procedure provided by Cayman's Superoxide Dismutase Assay Kit, Item No. 706002 (Cayman Chemical, Ann Arbor, MI, USA). The method utilizes a tetrazolium salt for detection of superoxide radicals generated by xanthine oxidase and hypoxanthine. One unit of SOD is defined as the amount of enzyme needed to exhibit 50% dismutation of the superoxide radical. This assay measures all types of SOD (Cu/Zn-SOD, Mn-SOD, and Fe-SOD) [18].

Telomere length: We assessed telomere length following the procedure described by O'Callaghan and Fenech, in which the absolute telomere length (aTL) was measured [22]. For this purpose, we used Telomere standard Human/rodent (teloF and teloR) as primers for telomere length (TL) analysis and 36B4 standard human primers for single copy gene (SCG) determinations. All these were supplied by TaqMan™ Array Human Telomere Extension by Telomerase (Thermofisher Scientific, Waltham, MA, USA).

First standard curves were constructed for both experiments (TL and SCG). Then, all patients' samples were analyzed, and aTL was calculated dividing the absolute result of TL by the result of SCG. This result was again divided by 92 (each somatic human cell has 46 chromosomes, and each chromosome has 2 telomeres) in order to obtain the mean aTL per single telomere [22].

Apoptosis: In order to detect apoptosis, we evaluated possible DNA fragmentation in patients' leucocytes. For this purpose, we performed a DNA conventional constant field gel electrophoresis loading in a 0.8% agarose gel panel a total or 300 ng from a normalized purified DNA mixture with a DNA concentration of 30 ng/uL. DNA was purified by means of standard techniques already established [22]. When apoptosis is present, the result is fragmentation of DNA into multiples of 180 base-pair lengths; a characteristic "ladder" effect is obtained when these fragments are resolved in the agarose gel electrophoresis [23].

Statistical analysis: The dichotomous variables were described as whole numbers and percentages, and the continuous variables as mean and standard deviation (or median and interquartile rank (IQR) in those with no criteria of normal distribution). The distribution of all variables was analyzed with the Kolmorogov–Smirnov test. Possible biological parameters associated to the presence of sarcopenia and death were investigated performing the Student's *t* for normally distributed quantitative variables, and Mann–Whitney *U* test.

Finally, we also evaluated the association of these biological parameters with functional status (by means of basal Barthel index), death risk (by means of PROFUND index), and survival (considering death as a time-dependent variable), using linear regression models. Statistical significance was considered when obtained *p* values were ≤0.05. Statistics were performed using the SPSS 22.0 software (IBM, Armonk, NY, USA).

#### **3. Results**

We included 444 patients with a mean age of 77.3 ± 8.4 years. Fifty-five percent were male. The main clinical features and biological parameters of the recruited patients are detailed in Table 1. Sarcopenia was present in 97 (21.8%), frailty in 278 (62.6%), and the remaining 69 (15.6%) were robust. Eighty patients (18% of the whole cohort) out of those with sarcopenia or frailty had simultaneously both phenotypes.

And combined sarcopenia and frailty were present in 80 (18%) patients. Mortality in the 12-months follow-up period was 40% (*N* = 178). A detailed clinical description of the included patients has already been published [7]; briefly, sarcopenia was more frequent in men, and associated to chronic lung diseases, cancer, lower BMI, and previous hospital admissions, whereas frailty was more frequent in women and associated to a higher number of polypathology categories, chronic pain, anxiety, and pressure ulcers; both phenotypes shared association with age, asthenia, and lower BI scores.


**Table 1.** Main clinical and biological features of a multicenter sample of 444 polypathological patients recruited for sarcopenia and frailty assessment.

SD: standard deviation; IQR: interquartile range; NYHA: New York Heart Association; mMRC: Medical Research Council. BMI: body mass index.

#### *3.1. Oxidative Stress Markers*

Median catalase and GR activity were 53 nmol/min/mL (IQR = 20–83), and 9.8 U/mL (IQR = 6.6–13.2), respectively. Total antioxidant activity against ROS was 2.4 mM Trolox equivalents (IQR = 1.8–3). Finally, median SOD activity was 4.6 U/mL (IQR = 2.8–6.6).

Differences of oxidative stress markers in patients with sarcopenia, frailty, or those with both conditions with respect to those without sarcopenia, robust, or those without both conditions are detailed in Table 2.


**Table 2.** Differences of oxidative stress markers, telomere length, and apoptosis markers in patients with multimorbidity according to their sarcopenia and frailty assessment.

CAT: catalase (nmol/min/mL); GR: Glutathione reductase (U/mL); TAC-ROS: total antioxidant activity against reactive oxygen species (mM Trolox equivalents); SOD: superoxide dismutase (U/mL); absolute telomere length (kbases/telomere); \* Interquartile range; WBC: white blood cells; DNA: deoxyribonucleic acid.

#### *3.2. Absolute Telomere Length Analysis*

Mean aTL was 2 kbases per telomere (IQR = 0.1–55). Differences of aTL in patients with sarcopenia, frailty, or those with both conditions with respect to those without sarcopenia, robust, or those without both conditions are also detailed in Table 2.

#### *3.3. Apoptosis*

Apoptosis by means of DNA fragmentation analysis was not present in any of the patients included in the study.

#### *3.4. Functional Parameters, Death Risk by PROFUND Index and Survival according to Di*ff*erent Molecular Parameters*

A worse functional status by means of lower Barthel's index score was associated to shorter telomere length (Beta = 1.25 (1.07–1.34)); *p* = 0.001), but not with any of the oxidative stress markers.

A higher death risk by means of PROFUND index was associated to shorter telomere length (Beta = 0.5 (0.14–0.65); *p* = 0.001) and to a higher GR activity (Beta = 1.7 (1.2–2); *p* = 0.04). On the other hand, a lower number of survival days was associated to shorter telomere length (Beta = 1.2 (1.01–1.32); *p* = 0.003) and to a higher GR activity (Beta = 0.3 (0.1–0.24)); *p* = 0.02).

#### **4. Discussion**

In the present study, we have detected enhanced oxidative stress and significant telomere shortening in PP with sarcopenia, frailty, or both syndromes combined. On the contrary, no evidence of apoptosis was detected.

Sarcopenia was prevalent in our cohort of polypathological patients and was associated to a significant higher SOD activity; other oxidative stress markers activity was also elevated, and the TAC-ROS decreased, but differences in these last were not significant. In the same way, we observed a significant telomere length shortening in these patients compared to other PP without sarcopenia. These results are highly concordant with the pathogenesis of sarcopenia in the elderly as already demonstrated by other authors [24–27]. Many authors have compared these markers among elderly and young people [28]; in the present study we have also detected important differences among elderly patients with chronic conditions with or without sarcopenia. These findings could have two major clinical applications: first, to use them as biological markers of sarcopenia in the elderly compared to persons of the same age; and second, to guide future treatments towards these targets in order to avoid or delay the development of sarcopenia. With respect to oxidative stress, SOD was the marker with the largest differences among PP with or without sarcopenia. As a matter of fact, SOD has been already strongly linked to muscular weakness, muscular wasting, and sarcopenia in clinical and experimental scenarios [29–31]; in this sense, among others, probably SOD could be the optimal oxidative stress marker in the evaluation of sarcopenia.

Frailty was also highly prevalent in the studied PP cohort and was associated to a significant increase in SOD activity and a decreased plasma TAC-ROS. It was also associated to a significant telomere length shortening compared to other PP without frailty. The deep relation between sarcopenia and frailty is already known; they share molecular and physiological pathways, symptoms, signs, and clinical phenotypes [32,33], so the presence of these molecular alterations in both of them is biologically coherent. In this case, we also observed a decreased antioxidant fitness of the plasma in frail PP. As main differences, frailty is more age related, whereas sarcopenia is also related to disease, starvation, and disuse [34]; additionally, despite criteria defining the two conditions overlap, frailty requires weight loss, whereas sarcopenia requires muscle loss [34,35].

In PP with sarcopenia and/or frailty we have observed the coexistence of telomere shortening and enhanced oxidative stress. There is accumulating evidence of the role of oxidative stress in DNA damage and telomere shortening with aging and chronic diseases [36]. These changes have been observed in humans, as well as in mouse models and cell cultures [36]. There are probably mixed mechanisms in this narrow relation of oxidative stress and telomere length. In aging and in many chronic conditions, processes associated to chronic inflammation play a nuclear role. Chronic inflammation is characterized by higher oxidative stress in affected tissues and circulating plasma. This may lead to direct cell DNA damage, including telomere regions. Additionally, inflammatory states are associated to enhanced necrosis and cellular regeneration cycles, and this increased cell turnover directly affects telomere length [37–42]. Many authors already point out that targeting oxidative stress could be of notable benefit in telomere length maintenance, especially in populations with chronic conditions like patients in the present study [39–42].

We did not detect any DNA fragmentation in our patients' leucocytes, so no apoptosis evidence could be detected in PP's blood samples by this technique. Apoptosis pathways have been classically associated to sarcopenia and frailty and nowadays are considered one of the main causes of these two syndromes [43,44]. As a matter of fact, there is multiple evidence of apoptosis presence in muscle tissue of experimental animal models, as well as in humans with sarcopenia [45–49]. Nevertheless, no information is available about the presence of apoptosis evidence in blood leukocytes of patients with sarcopenia and/or frailty. Some authors have described indirect apoptosis pathways data in blood leucocytes in elderly and in patients with dementia (like less resistance to experimental apoptosis inducers; senescence of CD8+ T-cells; and increased expression of HLA-DR, CD95, and Bcl-2 in CD3+ lymphocytes) [50]. Recently, increased ROS production and DNA fragmentation has been observed

in blood monocytes of atherosclerotic mice, uprising again the interrelations of oxidative stress and apoptosis signaling [51]. Apoptosis will for sure be present in muscle tissues of patients with sarcopenia and frailty, like enhanced oxidative stress and telomere shortening. Nevertheless, according to our data, an easy detection of its presence in blood samples from these patients is probably not useful in the clinical setting, and demonstrating it in tissue specimens is not clinically justified.

A poorer functional status, higher mortality risk, and less survival days in the 12-month follow-up were associated to shorter telomere length; besides, mortality risk and survival days were also associated to enhanced GR activity. These data are in concordance with previous studies in which telomere shortening has been associated to poorer survival in cancer, diabetes, cardiovascular diseases, and even to higher all-cause mortality [52–56]; additionally, oxidative stress has also been related to poor health outcomes in many clinical scenarios, and to all-cause mortality [57–60]. Our data confirm this deleterious relationships with sarcopenia and frailty in patients with multimorbidity, as well as the association to poorer functional status. Some authors have already claimed the clinical usefulness of biomarkers' panels including aTL, if we want to accurately assess and predict outcomes in vulnerable aged populations [61]. We suggest including also oxidative stress markers in these panels, mainly GR, TAC-ROS, and SOD.

This study has some limitations that should be noted. The results could be limited by the number of patients, but on the other hand, the cohort was recruited in various centers, was homogeneous, and probably represents adequately hospital-based populations with moderate–severe multimorbidity. Additionally, the studied biomarkers are also associated to some of the chronic conditions of the included patients and could raise the question of their real correlation to sarcopenia and frailty; this issue always underlies the frailty and sarcopenia phenotypes, since they have multiple concurrent causes, with a prominent role of debilitating chronic diseases; in our opinion, they behave as parts of the same clinical-molecular syndrome; as a matter of fact, the term "inflamm-aging" is already established, and probably in the future, it will be necessary to add chronic conditions and call it "inflamm-chronic-aging".

In conclusion, oxidative stress and telomere shortening, but not apoptosis markers, were enhanced in blood samples of polypathological patients with sarcopenia and/or frailty with respect to those patients without these two geriatric syndromes. Telomere shortening was associated to functional decline, and both, oxidative stress markers and telomere shortening, were associated to higher mortality risks and decreased survival. Both of these biomarkers could be useful in the clinical evaluation of vulnerable patients prone to sarcopenia and frailty and of potential interest as therapeutic targets.

**Author Contributions:** Conceptualization, M.B.-W. and M.O.-B.; Formal analysis, M.B.-W.; Investigation, M.B.-W., R.G.-D., Á.G.-M., S.V.-S., J.D.-M., F.S. and M.S.-M.; Methodology, M.B.-W., R.G.-D. and J.D.-M.; Supervision, M.O.-B.; Visualization, M.O.-B.; Writing–original draft, M.B.-W.; Writing–review & editing, Á.G.-M. and M.O.-B. All authors have contributed substantially to the work. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** Special thanks to all researchers from the proteo project listed below. Máximo Bernabeu-Wittel (1), Álvaro González-Molina (1), Rocío Fernández-Ojeda (2), Jesús Díez-Manglano (3), Fernando Salgado-Ordóñez (4), María Soto-Martín (5), Marta Muniesa (6), Manuel Ollero-Baturone (1), Juan Gómez-Salgado (7), Sofía Vidal-Serrano (2), Adriana Rivera-Sequeiros (2), Antonio Fernández-Moyano (2), Lourdes Moreno-Gaviño (1), Dolores Nieto-Martín (1), Nieves Ramírez-Duque (1), Esther del Corral-Beamonte (3), Pablo Martínez-Rodés (3), María Sevil-Puras (3), Rosa Bernal-López (4), Ricardo Gómez-Huelgas (4), Bosco Barón-Franco (5). Hospitals: (1) Hospital Universitario Virgen del Rocío, Sevilla, Spain; (2) Hospital San Juan de Dios del Aljarafe, Sevilla, Spain; (3) Hospital Royo Villanova, Zaragoza, Spain; (4) Hospital Regional, Málaga, Spain; (5) Hospital Juan Ramón Jiménez, Huelva, Spain; (6) Hospital San Juan de Dios de Pamplona, Pamplona, Spain; (7) School of Nursery, University of Huelva, Spain.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Measures Derived from Panoramic Ultrasonography and Animal-Based Protein Intake Are Related to Muscular Performance in Middle-Aged Adults**

**Nathaniel R. Johnson <sup>1</sup> , Christopher J. Kotarsky <sup>2</sup> , Kyle J. Hackney <sup>1</sup> , Kara A. Trautman <sup>3</sup> , Nathan D. Dicks <sup>4</sup> , Wonwoo Byun <sup>5</sup> , Jill F. Keith <sup>6</sup> , Shannon L. David <sup>1</sup> and Sherri N. Stastny 1,\***


**Abstract:** Ultrasonography advantageously measures skeletal muscle size and quality, but some muscles may be too large to capture with standardized brightness mode (B-mode) imaging. Panoramic ultrasonography can capture more complete images and may more accurately measure muscle size. We investigated measurements made using panoramic compared to B-mode ultrasonography images of the rectus femoris with muscular performance. Concurrently, protein intake plays an important role in preventing sarcopenia; therefore, we also sought to investigate the association between animal-based protein intake (ABPI) and muscular performance. Ninety-one middle-aged adults were recruited. Muscle cross-sectional area (CSA) and thickness were obtained using B-mode and panoramic ultrasound and analyzed with Image J software. Muscular performance was assessed using isokinetic dynamometry, a 30-s chair test, and handgrip strength. Three-day food diaries estimated dietary intakes. Linear regression models determined relationships between measures from ultrasonography and muscular performance. Mixed linear models were used to evaluate the association between ABPI and muscular performance. Muscle CSA from panoramic ultrasonography and ABPI were positively associated with lower-body strength (β ± S.E.; CSA, 42.622 ± 20.024, *p* = 0.005; ABPI, 65.874 ± 19.855, *p* = 0.001), lower-body endurance (β ± S.E.; CSA, 595 ± 200.221, *p* = 0.001; ABPI, 549.944 ± 232.478, *p* = 0.020), and handgrip strength (β ± S.E.; CSA, 6.966 ± 3.328, *p* = 0.004; ABPI, 0.349 ± 0.171, *p* = 0.045). Panoramic ultrasound shows promise as a method for assessing sarcopenia. ABPI is related to better muscular performance.

**Keywords:** panoramic ultrasound; echogenicity; specific force; isokinetic dynamometry; protein intake; muscle quality; strength; endurance

#### **1. Introduction**

Earlier and more frequent assessments of muscle strength, mass, size, and quality and physical performance could help prevent sarcopenia by indicating a need for treatment or other intervention. According to the European Working Group on Sarcopenia in Older People 2, low muscle strength is the first criteria of sarcopenia, and low muscle mass or quality is the second; both must be assessed to determine sarcopenia [1]. Low physical

**Citation:** Johnson, N.R.; Kotarsky, C.J.; Hackney, K.J.; Trautman, K.A.; Dicks, N.D.; Byun, W.; Keith, J.F.; David, S.L.; Stastny, S.N. Measures Derived from Panoramic Ultrasonography and Animal-Based Protein Intake Are Related to Muscular Performance in Middle-Aged Adults. *J. Clin. Med.* **2021**, *10*, 988. https://doi.org/ 10.3390/jcm10050988

Academic Editor: Gianluca Testa

Received: 30 January 2021 Accepted: 22 February 2021 Published: 2 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

performance in addition to low muscle strength and quantity is considered severe sarcopenia [1]. Measures of muscle strength, such as handgrip strength and physical performance (e.g., 30-s chair stand), however, can be performed with minimal equipment and are used across various settings [1]. Although several methods can be used to accurately assess muscle quantity and quality such as computed tomography (CT), magnetic resonance imaging (MRI), and dual x-ray absorptiometry, these techniques require expensive equipment and are not portable, limiting their utility. Ultrasonography is a portable and relatively low-cost method of assessing muscle size [2], making it a potentially useful tool for evaluating sarcopenia for clinical or research purposes [3,4]. Beyond this, ultrasonography records a measure of muscle quality in the form of echogenicity or echo intensity [5–7], making ultrasound a potentially more powerful tool than bioelectrical impedance for assessing sarcopenia or signs of pre-sarcopenia in middle age.

Others have used ultrasonography to successfully diagnose sarcopenia [7–9]. However, two of these studies were performed with either frail elderly patients or older adults diagnosed with chronic kidney disease [8,9]. Not only are the causes of sarcopenia thought to start earlier in life [1], making middle-aged adults a population of interest, but also, older adults often have smaller muscles that can be captured using a traditional ultrasound image at 50% of leg length. Although Ismail and colleagues [7] were able to discriminate between those with sarcopenia and those without in a younger cohort, they did this by using longitudinal and not transverse images of the rectus femoris. The crux of the issue is that in populations that have greater muscle mass at the midpoint of the thigh, such as younger populations, the entire transverse rectus femoris may be too large to capture in one image [10]. Assuming the goal is to image the entire transverse rectus femoris, then there are two workarounds: one is to use a feature, like the panoramic feature, to record the entire rectus femoris at the midpoint of the thigh, and the other is to move the imaging site distally down the leg where the rectus femoris has smaller transverse sections. Other researchers have validated panoramic ultrasound of the quadriceps with MRI [11], but to our knowledge, the relationship between ultrasonographic measures of the transverse rectus femoris captured using the panoramic feature and muscular performance, in particular that of the knee extensors, has not been investigated. Because muscle strength is more closely related to sarcopenia than muscle mass [1,12], the association warranted investigation.

Beyond this, specific force, the amount of force produced per unit of muscle, like echogenicity [6], is considered a measure of muscle quality [12]. Although echogenicity of the rectus femoris is related to muscle quality assessed using CT [8], and to a lesser extent knee extensor strength [13], the echogenicity of the rectus femoris has not been directly related to the specific force of the muscle. However, Ismail and colleagues [7] reported a significant relationship between echogenicity of rectus femoris and handgrip strength relative to bodyweight, a crude measure of specific force. If echogenicity and specific force reflect the muscle quality of the rectus femoris, then they should be closely related. We also sought to determine this relationship.

Outside of assessing the condition, nutrition is another important consideration for preventing and treating sarcopenia. Although there are many nutritional factors that can impact sarcopenia [14], dietary protein is perhaps of greatest interest because of its ability to stimulate muscle protein synthesis [15]. Recently though, the role of protein intake in performance has come into question, with one group finding no relationship between protein intake and measures of muscular performance, such as handgrip strength, knee extensor strength, and 30-s chair stand test performance [16]. Foods from animal and plant sources, of course, differ in their digestibility and amino acid content [17], and therefore in their ability to stimulate muscle protein synthesis [18]. Due to the differential impact that animal-based protein has on muscle protein synthesis, we secondarily sought to determine the relationship between animal-based protein intake (ABPI) and lower-body strength and endurance, handgrip strength, and 30-s chair stand performance, measures of muscular performance.

#### **2. Materials and Methods**

This was a cross-sectional study conducted in the North Dakota State University Healthy Aging Lab from October 2016 to December 2018. A total of 50 women and 41 men from the local community were recruited using e-mail, flyers, and word-of-mouth to visit the research lab for two sessions. During the first session, anthropometric, ultrasonographic, and performance variables were measured, and accelerometers and three-day food diaries were provided. Within seven to 14 days, participants returned their accelerometers and their completed food diaries to the lab. Participants were between 40 and 67 years of age, not currently using any nicotine product, free of any untreated or nonresponsive diseases or conditions including neuromuscular disease or conditions that might undermine muscle health, such as diabetes, ambulatory without any assistance, and had to include both animal-based and plant-based foods in their diets. Participants were screened using the 2011 Physical Activity Readiness Questionnaire [19], a more detailed health history questionnaire, and an orthostatic hypotension test. Participants were also instructed to refrain from exercise and strenuous physical activity at least 48 h prior to the first session. The study was approved by the North Dakota State University Institutional Review Board (#HE26929 & 26153) and complied with the Helsinki Declaration of 1983. Written informed consent was obtained from all participants in this study.

#### *2.1. Participant Heath Screening and Anthropometric Measures*

To screen participants for orthostatic hypotension, related to regulatory and safety concerns, resting blood pressure and standing blood pressure were measured manually with a stethoscope and Diagnostix 703 sphygmomanometer (American Diagnostic Corporation, Hauppauge, NY, USA). Those whose blood pressure dropped by more than 10 mm Hg, either systolic or diastolic, from resting to standing during the orthostatic hypotension test were excluded (*n* = 0). Following the orthostatic hypotension test, anthropometric variables were measured. Age (years) was self-reported. Height (cm) was measured using a stadiometer (Seca 213, Chino, CA, USA) and body mass (kg) was recorded using a digital balance (Denver Instrument DA-150, Arvada, CO, USA).

#### *2.2. Ultrasonography*

Images of the right rectus femoris muscle were captured using a Philips ultrasound system (model HD11 XE; Bothell, WA, USA) with a L12-5 50 mm linear array probe by three trained research assistants. Images were taken while participants were standing at marked sites 50% and 75% of the measured distance from the superior iliac spine of the hip to the lateral condyle of the knee. Participants were instructed to use their left leg as a base of support, while relaxing their right, resulting in a slight bend in the right knee. Previous works have shown high test–retest reliability of ultrasound measures of muscle thickness of healthy adults taken in the standing position [20,21]. A more recent study found the intraclass correlation coefficient (ICC) for standing measures of the anterior thigh muscles was 0.89, while the ICC for the same measures taken while participants were recumbent was 0.90 [22]. Following generous application of ultrasonic gel, the probe was placed on the skin perpendicular to the leg, and light, consistent pressure was applied to avoid excessive depression of the dermal surface until a full, clear image was obtained. The probe was removed from participants' skin between each image acquisition, and markings were used to ensure the same area was measured. Because our participants were younger and likely have greater muscle size, the panoramic feature was used at the 50% site to record the entire transverse rectus femoris [10]. For panoramic ultrasonography, the lateral side of the right rectus femoris was identified, and the probe was moved medially until the entire transverse rectus femoris was recorded. B-mode image captures were taken at the 75% site where transverse sections of the rectus femoris are smaller. Three images were captured at each site using a frequency of 37 Hz with a standardized depth of 7 cm and gain of 100%.

After each image was captured, a 1 cm line was added to each image to act as a known distance during analysis. Images were transferred to personal computers, calibrated, and

analyzed. ImageJ software (National, Institutes of Health, Bethesda, MD, USA, version 1.42) was used to analyze echogenicity, cross-sectional area (CSA), and muscle thickness [23]. Echogenicity was defined as the mean pixel intensity of the rectus femoris measured in arbitrary units (A.U.) ranging between 0 (i.e., black) and 255 (i.e., white). Anatomical muscle CSA was determined by tracing the inside of the epimysium of the rectus femoris using the polygon tool. Rectus femoris thickness was assessed with a single measurement using the straight-line tool; using ImageJ, a line was made through the largest, middle portion of the muscle perpendicular to the skin. Intraclass correlation coefficients (ICC) were used to examine the reliability of these analyses. All three research assistants completed reliability training prior to being allowed to be an operator for the testing in the study. The test–retest reliability of three images obtained by the research assistants using ICCs and 95% confidence intervals were as follows: panoramic muscle thickness = 0.98 (0.90, 0.95), B-mode muscle thickness = 0.98 (0.97, 0.99), panoramic muscle area = 0.95 (0.93, 0.96), B-mode muscle area = 0.97 (0.97, 0.98), panoramic muscle echogenicity = 0.98 (0.97, 0.98), and B-mode echogenicity = 0.81 (0.75, 0.87). For consistency, these measurements were all analyzed by the same member of the research team. The mean of each participant's values across the three images at each site (i.e., 50% and 75%) was used in our analyses. Figure 1 displays an example of muscle thickness and CSA captured and analyzed at each site.

**Figure 1.** Examples of rectus femoris muscle thickness and CSA captured via ultrasonography for one participant. (**a**) Rectus femoris muscle thickness at 50% of leg length captured using the panoramic feature. (**b**) Same as A but showing muscle CSA. (**c**) Rectus femoris muscle thickness at 75% of leg length captured using a standardized B-mode image. (**d**) Same as C but showing muscle CSA.

#### *2.3. Performance Measures*

Participants performed a self-paced, low to moderate intensity warm-up for five minutes using a cycle ergometer. Muscle strength and endurance of the lower body were tested using isokinetic dynamometry on a Biodex Pro IV System (Biodex Medical Systems, Shirley, NY, USA). Lower body muscular strength was assessed using peak torque performed during a three-repetition test at 60 ◦ per second for knee extension– flexion and a three-repetition test at 30 ◦ per second for plantar-dorsiflexion. Lower body muscular endurance was evaluated using the total amount of work performed during a 21-repetition test at 180 ◦ per second for knee extension–flexion and 60 ◦ per second for plantar-dorsiflexion [24]. Muscular strength and then endurance were first assessed in upper leg (i.e., knee extension–flexion) and then in the lower leg (i.e., plantar-dorsiflexion). A warm-up set was completed before each lower-body strength test (i.e., knee extension– flexion and plantar-dorsiflexion); participants were instructed to perform three repetitions at ≤75% of their perceived maximal effort. Thirty seconds of rest was given between all extension-flexion tests. One minute of rest was provided between plantar-dorsiflexion tests. To optimize performance, participants were encouraged to employ "all-out effort" by research staff during all muscle function tests. To better capture muscular performance of the entire right leg, peak torques from the isokinetic strength test and total work from the isokinetic endurance test were added together to create summed peak torque and summed total work (i.e., knee extension + knee flexion + plantarflexion + dorsiflexion).

Maximal handgrip strength (kg) was assessed using an analog Jamar Handheld Dynamometer (Bolingbrook, IL, USA). Participants were instructed to grasp the dynamometer in their dominant hand and to keep their elbow at their side with a 90◦ bend between the upper arm and forearm, while standing. Participants were told to squeeze the dynamometer as hard as possible for two to three seconds. Each participant performed three maximal attempts; the highest grip strength was used.

Participants then performed a 30-s chair stand test on a chair with a 43 cm floor-to-seat height. All trials were performed with participants' arms crossed and feet at a comfortable distance apart (i.e., about hip to shoulder width). With a straight back, participants were instructed to fully sit down and stand-up for each repetition, and practice repetitions were performed to ensure adequate performance during the test. The total number of repetitions completed in 30-s period was recorded, and the 30-s period began when participants started to rise.

#### *2.4. Physical Activity Assessment*

Following performance testing, participants were given accelerometers and three-day food diaries. Physical activity was recorded using Actigraph (ActiGraph Corp, Pensacola, FL, USA) GT9X accelerometers. Participants were instructed to wear accelerometers on their right hip during all waking hours, excluding activities where the device may get wet (e.g., bathing or swimming), for a period of one week and to keep a sleep log to record the time that the accelerometer was removed at night and put back on in the morning. The raw acceleration data were collected at 80 Hz and processed in R software (http://cran.r-project.org, accessed on 6 September 2016) using the GGIR package (version 1.10-10) [25]. Non-wear time was defined as intervals of at least 90 min of zero counts with allowance of a two-minute interval of non-zero counts within a 30-min window [26]; thus, only valid time during waking hours of each day was included for statistical analyses. Although accelerometry captures many aspects of physical activity (e.g., sedentary time, light physical activity, etc.), we decided to use moderate-to-vigorous physical activity (MVPA) in our analyses because of its relationship with performance variables [27,28].

#### *2.5. Nutrition Analysis*

After performance testing, participants were also given three-day food diaries, received training on how to record dietary intakes by a member of the research team, and were required to watch a prerecorded training video. Dietary intakes from three-day food diaries, including nutritional supplements, were entered into Food Processor Nutrition Analysis Software (ESHA Research, Salem, OR, USA), which uses FoodData Central (USDA National Nutrient Database) by trained research assistants. Data entry was then line-by-line verified by a registered dietitian. Animal- and plant-based protein intakes were estimated using a line-by-line examination of dietary intake by a registered dietitian. Food items that contained less than 1 g of total protein were excluded from these calculations. Foods containing both animal- and plant-based protein were split according to their ingredients to distinguish protein sources. Animal-based protein sources included meat, fish and seafood, dairy, eggs, poultry, and wild game.

#### *2.6. Statistical Analyses*

Alpha was set at 0.05, and all statistics were performed in SPSS version 27 (IBM, Armonk, NY, USA). All data are available as a supplemental file.

2.6.1. Primary Analyses: Measures from Ultrasonography and Their Relationships with Muscular Performance and the Association between Rectus Femoris Echogenicity and Specific Force

Three male participants could not be included in analyses of ultrasonography because our ultrasound machine suffered a catastrophic failure near the very end of the data collection window, precluding ultrasonography for these male participants. Thus, all analyses related to ultrasonography have 88 as opposed to 91 participants.

We used multiple-linear regression models to determine the relationships between variables derived from ultrasonography (i.e., rectus femoris muscle thickness, echogenicity, and CSA) using the two different methodologies (i.e., panoramic versus B-mode images) and sites (i.e., 50% and 75% of right leg length) with measures of muscular performance. Each of these variables from ultrasonography were assessed in separate multiple-linear regression models. Although we consider summed peak torque and summed total work to be more representative of lower-body performance, we specifically included knee extensor peak torque and total work in these analyses, because ultrasonography was used to measure the rectus femoris, one of the knee extensors. Separate multiple-linear regression models were also used to evaluate the relationship between echogenicity and specific force of the rectus femoris, two measures of muscle quality. All aforementioned regression models were adjusted for gender (i.e., 0 = women, 1 = men), age, and body mass in kilograms divided by the square of height in meters (BMI), because these variables are routinely collected in both clinical and research settings.

#### 2.6.2. Secondary Analyses: Animal-Based Protein Intake and Muscular Performance

All participants completed a three-day food diary, completed all performance measures (i.e., isokinetic dynamometry, handgrip strength, and 30-s chair stand test), and wore an accelerometer. For our analyses investigating nutritional variables, we first used simple linear regression models to verify that our estimates of animal-based and plant-based protein intakes together agreed with total protein intake. Animal-based and plant-based protein intakes, determined by line-by-line analysis of three-day food diaries by a registered dietitian and expressed either as relative intakes or percentages of energy intakes, were entered as predictor variables, and total protein, without partitioning into animal- or plant-based protein intakes, was the outcome variable.

Analyses of nutritional data are complicated by the shared variance of many variables. Energy intake and macronutrient intakes, which we examined in this work, are directly related, that is, a person's macronutrient intake, withstanding alcohol, determines their energy intake (i.e., protein + carbohydrates + fat = energy). Therefore, when analyzing dietary variables, relative energy (kcals/kg/day) and the relative intakes of all the macronutrients (g/kg/day) cannot be entered simultaneously. We used Pearson Product– Moment Coefficients to examine the collinearity of both relative macronutrient intakes and macronutrient intakes as percentages of energy intake with one another and with relative energy intake. Although there are other methodologies, we chose to include relative energy intake (kcal/kg/day) in our analyses and to express the intake of the macronutrients as percentages of energy intake. This method allowed us to control for both relative energy intake and macronutrient intakes in our statistical models.

Mixed linear models were used to evaluate the impact of ABPI on muscular performance. The 41 men and 50 women were first blocked according to self-reported gender (0 = women, 1 = men). Then, each gender was split at their median of energy intake from animal-based protein. More specifically, gender and ABPI (below median = 0, above median = 1) were entered as fixed factors. Age, BMI, MVPA, relative energy intake, and percent energy from protein, fat, and carbohydrates were entered as continuous covariates.

Models were evaluated for equality of error of variance using Levene's Test of Equality of Variance and for heteroscedasticity using White's Test of Heteroscedasticity; mixed models that were significantly unequal in their variances or heteroscedastic were transformed using the square root function. Out of an abundance of caution, we chose to use the HC3 method to calculate the standard errors of our variables, as it is more robust to unequal variances, heteroscedasticity, and multicollinearity than the ordinary least squares method [29]. We did not hypothesize that there would be interaction between gender and ABPI, so only main effects were examined in these mixed models. For those models in which ABPI is significant, we evaluated effect size using partial eat squared. We also sought to verify that ABPI and not total protein intake is important to performance. We verified our results by performing the same aforementioned methods, but we split each gender at median of total protein intake as a percentage of energy intake and included ABPI as a percentage of energy intake as a continuous covariate.

Estimates of physical activity from accelerometry are considered valid when the devices are worn for 10 h per day for at least four days [28], and three participants failed to meet these criteria despite our instruction to wear the devices during all waking hours for one week. Nonetheless, all other participants achieved at least four or more days including one weekend day with an average of 10 or more hours of time wearing the device. These three participants who failed to wear accelerometers as directed represent a small portion of our sample (3.3%), and physical activity was included in our mixed models as a covariate; physical activity is not the focus of this work, but we feel it is essential to control for in our mixed models evaluating ABPI. For these reasons and due to small sample size, particularly when split into groups, we decided to include these three participants, using their limited physical activity data in our analyses.

#### 2.6.3. Descriptive Statistics

For our descriptive statistics, we described the four groups from the secondary analyses in our all of our tables, even though we chose not to investigate the association between ABPI and measures from ultrasonography, because the three men who were precluded from ultrasonography were, coincidently, above the median for animal-based protein intake as a percentage of energy. Within these tables, we chose to use the Brown–Forsythe method for comparisons, because we did not assume equal variances. We compared those above the median of ABPI as a percentage of energy to those below the median within each gender, so we did not adjust for multiple comparisons.

#### **3. Results**

Table 1 describes participants self-reported age, measured height, weight, and calculated BMI. There were no statistically significant differences between those below or above the median of ABPI as a percentage of total energy within each gender.


**Table 1.** Self-reported age and anthropometrics for 41 men and 50 women.

All values are medians. Comparisons within gender and between those below and above the median for animal-based protein intake (ABPI) as a percentage of energy intake were made using the Brown–Forsythe method.

> Table 2 describes right rectus femoris muscle thickness, echogenicity, and CSA measured using the panoramic ultrasonography at 50% and B-mode images at 75% of the

distance of the right leg. Within each gender, there were no statistically significant differences in these measures between those above the median of ABPI and those below.

**Table 2.** Rectus femoris muscle thickness, echogenicity, and cross-sectional area assessed via ultrasonography captured using the panoramic feature at 50% and with regular B-mode images at 75% of the right leg in 88 middle-aged men and women.


All values are medians. CSA = Muscle Cross-Sectional Area. A.U. = Arbitrary Units. Comparisons within gender and between those below and above the median for animal-based protein intake (ABPI) as a percentage of energy intake were made using the Brown–Forsythe method.

> Table 3 presents the results of the sperate multiple-linear regression models investigating the relationship between different measures derived from ultrasonography and muscular performance. Measures of rectus femoris size assessed using panoramic ultrasonography were less related to knee extensor performance but more strongly related to overall muscular performance. More specifically, both muscle thickness (*p* = 0.302) and CSA (*p* = 0.056) assessed using the panoramic feature of the right leg were unrelated to knee extensor peak torque, whereas the same measures assessed using a B-mode image at of the right leg were related to knee extensor peak torque. Similarly, muscle thickness assessed using the panoramic feature was unrelated to knee extensor total work (*p* = 0.197). Although muscle CSA captured with the panoramic feature was related to knee extensor total work (*p* = 0.049), it was less closely related than muscle CSA (*p* = 0.013) or thickness (*p* = 0.036) assessed with a B-mode image at 75% of leg length. Conversely, measures of muscle thickness (*p* = 0.001) and CSA (*p* = 0.004) derived from panoramic ultrasound were significantly related to handgrip strength performance, whereas the same measures collected using B-mode were not. Muscle CSA from panoramic ultrasound was also most closely related to summed peaked torque (*p* = 0.005), a relationship that was only close to significance (*p* = 0.051) with a B-mode image. Both methodologies (i.e., panoramic and B-mode) produced measures of muscle thickness and CSA that were associated with summed total work.

> Echogenicity of rectus femoris was unrelated to both knee extensor and summed peak torque but was significantly associated with knee extensor total work when captured using either panoramic (*p* = 0.001) or B-mode images (*p* = 0.004). Echogenicity of the rectus femoris from both panoramic (*p* = 0.008) and B-mode (*p* = 0.007) images was also associated with handgrip strength. Interestingly, although echogenicity was related to knee extensor total work, it was not related to summed total work when using either methodology. No ultrasonographic measure was associated with 30-s chair stand performance.

> Table 4 describes our evaluation of echogenicity with specific force, two measures of muscle quality. Echogenicity was not related to specific force in any regression model nor was any model significant. We found measures from the 50% site, taken using the panoramic feature, created better fitting models. In fact, echogenicity assessed at 50% trended toward significance (*p* = 0.077).

 *Clin. Med.* **2021**, *10*, 988

*J.*

**Table 3.** The associations between different ultrasonographic measures of the right rectus femoris using the panoramic feature (50% of leg upper length) and a B-mode image (75% of upper leg length) in a sample of 88 middle-aged men and women when controlling for age, gender, and BMI.



**Table 4.** Association of echogenicity assessed via ultrasonography captured using the panoramic feature and B-mode images of the right leg with various assessments of knee extensor specific force in 88 middle-aged men and women.

> A.U. = Arbitrary Units. S.E. = Standard Error. Age: years. Gender: Women = 0; Men = 1. BMI: kg/m<sup>2</sup> .

> > Table 5 describes the nutritional variables assessed from three-day food diaries for study participants. There were significant differences in macronutrient intake between those above the median for ABPI as a percentage of energy intake and those below within each gender; relative carbohydrate intake, carbohydrate intake as percentage of energy, protein intake as percentage of energy, relative ABPI, ABPI as a percentage of energy, and relative plant-based protein intake were all significantly different in both men and women. Those above the median consumed less carbohydrates, more protein, and more animal based protein than those below. In women, there were also significant differences in relative fat and calcium intake with those above the median consuming less fat and more calcium. In men, on the other hand, there was a significant difference in relative energy intake with those below the median of ABPI consuming more energy.

> > Table 6 lists physical activity variables recorded using accelerometry. Excluding wear days, which were greater in men below the median compared to men above the median, there were no significant differences between those above the median of animal-based protein as percentage of energy intake and those below.

> > Regression models examining estimates of animal-based and plant-based protein intakes with total protein intake showed good agreement between our estimates and total protein. Estimates of relative animal-based and relative plant-based protein intakes explained 98.4% of the variance in relative protein intake (F2,88 = 2788.702, *p* < 0.001), and estimates of animal- and plant-based protein intakes as percentages of energy explained 94.0% of the variance in protein as a percentage of energy (F2,88 = 683.550, *p* < 0.001).

> > Table 7 shows Pearson Product–Moment Correlations between relative macronutrient intakes, macronutrient intakes as percentages of energy intake, and relative energy intake. Relative macronutrient intakes showed stronger relationships with relative energy intake than macronutrient intakes expressed as a percentage of energy intake. Withstanding the association between percent of energy from fats and carbohydrates, macronutrient intakes expressed as percentages of energy were less strongly correlated amongst one another than relative macronutrient intakes. These results suggest macronutrient intakes should be expressed as percentages of energy intake in statistical models including relative energy intake to limit collinearity.


**Table 5.** Dietary intakes accessed from three-day food diaries in 41 middle-aged men and 50 middle-aged women.

All values are medians. Comparisons between those below and above the median for animal-based protein intake (ABPI) as a percentage of energy intake within gender were made using the Brown–Forsythe method. \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001.


**Table 6.** Physical activity variables assessed using accelerometry in 41 middle-aged men and 50 middle-aged women.

All values are medians. Comparisons between those below and above the median for animal-based protein intake (ABPI) as a percentage of energy intake within gender were made using the Brown–Forsythe method. \* *p* < 0.05.

**Table 7.** Pearson Product–Moment Correlations of macronutrient intakes, including animal-based protein, and relative energy intake in 41 middle-aged men and 50 middle-aged women.


Table 8 and Figure 2 present the results of our investigation of the relationship between ABPI with performance measures. To create homoscedastic models with equal variances, data from the handgrip strength test (kg) and the 30-s chair stand test (repetitions) were transformed using the square root function. Using these transformed variables, all of these mixed models had equal variances according to Levene's Test and were homoscedastic according to White's test (i.e., *p* > 0.05).

**Table 8.** Animal-based protein intake and muscular performance in middle-aged men and women.


S.E. = standard error. Age: years. Gender: Women = 0, Men = 1. BMI: kg/m<sup>2</sup> . Relative energy intake: kcal/kg/day. Animal-based protein intake was split at the median of percent energy from animal-based protein within both men and women; below median = 0, above median = 1. Nutritional variables were assessed using three-day food diaries. Summed isokinetic peak torque was calculated by adding the peak torques recorded during the isokinetic strength test, 60◦ per second for knee extension–flexion and 30◦ per second for plantar-dorsiflexion. Summed isokinetic endurance was calculated by adding total work performed during a 21-repetition test at 180◦ per second for the knee extension–flexion and 60◦ per second for plantar-dorsiflexion. Total repetitions performed during the 30-s chair stand test and handgrip strength were transformed using the square root function. The height of the chair for the 30-s chair stand test was 43 cm.

> Our mixed models explained 78.6% of the variance of summed peak torque performed during the isokinetic strength test, 75.7% of the variance of summed work performed during the isokinetic endurance test, and 83.3% of the variance in handgrip strength transformed using the square root function, indicating good model fit for these performance variables. However, our mixed model investigating the results of the 30-s chair stand test only explained 19.1% of the variance in this measure, indicating relatively poor model fit. Nonetheless, all models were significant.

> Animal-based protein intake was significant to mixed models evaluating lower-body muscular strength, lower-body muscular endurance, and handgrip strength. Those consuming above the median of animal-based protein as percentage of energy intake performed better on these tests of muscular strength and endurance than those below the median. The effect sizes assessed using partial eta squared of the ABPI median split were 0.120, 0.065, and 0.049 for summed lower-body peak torque, summed lower-body total work, and handgrip strength, respectively. Animal-based protein intake was not related to performance in the 30-s chair stand test.

**Figure 2.** Animal-based protein intake and muscular performance. Animal-based protein intake was split at the median of percent energy from animal-based protein within both men and women; below median = 0, above median = 1. Covariates included age, gender, BMI, MVPA, relative energy intake, and percentages of energy intake from fat, carbohydrate, and protein. All bars are means, and error bars represent 95% confidence intervals. (**a**) Summed isokinetic peak torque by gender and animal-based protein intake. Summed isokinetic peak torque was calculated by adding the peak torques recorded during the isokinetic strength test, 60 ◦ per second for knee extension–flexion and 30 ◦ per second for plantar-dorsiflexion. (**b**) Summed isokinetic endurance by gender and animalbased protein intake. Summed isokinetic endurance was calculated by adding total work performed during a 21-repetition test at 180 ◦ per second for the knee extension–flexion and 60 ◦ per second for plantar-dorsiflexion. (**c**) Square root transformed 30-s chair stand test repetitions by gender and animal-based protein intake. The height of the chair for the 30-s chair stand test was 43 cm. (**d**) Square root transformed handgrip strength by gender and animal-based protein intake.

Because ABPI was significant to lower-body muscular strength, lower-body muscular endurance, and handgrip strength, we wanted to verify that these findings were due to ABPI and not to greater total protein intake. Although we did control for total protein intake as percentage of energy in our mixed models where participants were split at the median of ABPI, Table 9 shows our analyses where participants were split at the median of total protein intake as percentage of energy intake and ABPI as a percent of energy intake was entered as a continuous covariate. With the exception of square root transformed 30-s chair stand repetitions, all of these mixed models had equal variances according to Levene's Test and were homoscedastic according to White's test (i.e., *p* > 0.05). Square root transformed 30-s chair stand performance was homoscedastic but showed unequal variances between groups (*p* = 0.024) according to Levene's test. Because our earlier analysis of square root transformed 30-s chair stand performance (i.e., Table 8) showed equal variances between groups, was homoscedastic, and produced nonsignificant results regarding protein intake and ABPI, we did not transform 30-s chair stand performance using a different methodology (e.g., Log). In other words, square root transformed 30-s chair stand performance was included in Table 9 despite showing unequal variances between groups, although the HC3 method is considered to be more robust to violations of unequal variance [28]. Total protein intake split at the median of energy intake was not significant to any performance variable, whereas APBI split at the median was significant to lower-body muscular strength, lower-body muscular endurance, and handgrip strength, indicating that APBI is more closely related to muscular performance than total protein intake.


**Table 9.** Total protein intake and muscular performance in middle-aged men and women.

S.E. = standard error. ABPI = animal-based protein intake. Age: years. Gender: Women = 0, Men = 1. BMI: kg/m<sup>2</sup> . Relative energy intake: kcal/kg/day. Total protein intake was split at the median of percent energy from protein within both men and women; below median = 0, above median = 1. Nutritional variables were assessed using three-day food diaries. Summed isokinetic peak torque was calculated by adding the peak torques recorded during the isokinetic strength test, 60◦ per second for knee extension–flexion and 30◦ per second for plantar-dorsiflexion. Summed isokinetic endurance was calculated by adding total work performed during a 21-repetition test at 180◦ per second for the knee extension–flexion and 60◦ per second for plantar-dorsiflexion. Total repetitions performed during the 30-s chair stand test and handgrip strength were transformed using the square root function. The height of the chair for the 30-s chair stand test was 43 cm.

#### **4. Discussion**

We found that measures of muscle size from standardized B-mode ultrasound images better captured the performance of the knee extensors, whereas measures of muscle size assessed from panoramic images were more closely related to overall muscular performance, producing significant associations between muscle size with summed peak torque and handgrip strength. However, our methodology differed from that of others who have utilized panoramic ultrasound. We took panoramic images of the rectus femoris at one location (i.e., 50% of leg length) as opposed to using a template to image the entire length of the quadriceps, although one research group advocated for an investigation of a single site at the mid-quadriceps [11].

Nonetheless, the lack of a significant relationship between muscle thickness and CSA measured using the panoramic feature and knee extensor strength is surprising, considering these measures of muscle size were more closely related both to lower-body strength (i.e., summed peak torque) and upper-body strength. Low muscle strength is the first criterion of sarcopenia according to the European Working Group on Sarcopenia in Older People 2 and should be, albeit not necessarily linearly, related to muscle mass [1]. In other words, changes in muscle mass or size are not as meaningful as changes in muscle strength. Measures of muscle size or mass that are unrelated to muscle strength then may have limited utility in assessing or screening for sarcopenia. Despite the fact measures from panoramic ultrasonography lacked face validity in the form of a significant relationship with knee extensor peak torque, our findings suggest that the panoramic feature is a suitable method for assessing sarcopenia in those with greater muscle at the midpoint of thigh, as it is related to both lower-body and upper-body strength.

We also report that in our sample echogenicity was unrelated to both knee extensor, strength, overall lower-body strength, and rectus femoris specific force, another measure of

muscle quality. Although Strasser and colleagues [13] reported a significant correlation between echogenicity and knee extensor strength, the relationship was only found in younger and not older adults. In contrast, Akima and colleagues [30] found a significant relationship between echogenicity and sit-to-stand performance in older Japanese men and women. However, in a subsequent work, the same research group reported no relationship between echogenicity and knee extensor strength [6]. We also did not find a significant relationship between echogenicity and knee extensor strength, and we were the first, at least to our knowledge, to directly compare the echogenicity of the rectus femoris to the muscle's specific force. None of the relationships were significant. However, we did find an association between echogenicity with handgrip strength and knee extensor muscular endurance. Echogenicity has been related to both intramuscular fat [31] and fibrous tissue [32] content of muscle. In a large study of older Italian men and women, De Stefano and colleagues [33] reported a negative association between intramuscular fat and physical performance but found that those who were overweight or "Class I" obese had greater knee extensor strength than those with a normal BMI, suggesting that intramuscular fat plays a greater role in physical performance than in maximal strength. Our findings regarding echogenicity support that view. Echogenicity, then, is not closely related to specific force as it is with other muscular qualities such as endurance, because specific force is dependent on maximal muscle strength.

Our secondary findings regarding dietary intake indicate a positive relationship between ABPI and muscle strength when controlling for gender, age, BMI, relative energy intake, and macronutrient composition. More specifically, those above the median of ABPI as percentage of energy intake showed greater lower-body strength and endurance and greater handgrip strength than those below. Although greater protein intake is thought to be protective from developing sarcopenia [34–36], a recent cross-sectional study of older Danish adults utilizing methods similar to ours (e.g., three-day food diary and physical activity assessment) reported that protein intake was not related to knee extensor strength, handgrip strength, and 30-s chair stand test performance [16]. In contrast to their methodology where participants were divided into groups based on relative protein intake, we split ours according to ABPI as a percentage of energy intake. Although recommendations for protein intake are made on a g/kg basis [36], an advantage of expressing intakes as percentages of energy intake is that one can control for relative energy intakes and for macronutrient composition in the same statistical model. There is a high degree of collinearity between relative intakes of macronutrients and relative energy intake. In fact, one of the main findings from Højfeldt and colleagues' study of older Danish adults was that relative protein intakes and relative energy intakes are related [16]. Collinearity can bias estimates of betas in multivariate analyses [37]. Although there is still a degree of collinearity between macronutrient intakes as percentages of energy and relative energy intakes, we addressed this issue by using the HC3 method of calculating standard errors, which is more robust to collinearity and heteroscedasticity [29]. Outside of expressing intakes as percentages of energy, our methodology also differed because we evaluated ABPI. Plant-based proteins generally contain amino acids that are oxidized to be used as energy to a greater extent than higher quality animal-based proteins [18]. Thus, total protein intake is likely less strongly related to muscle mass and strength than protein intake from higher quality sources, and our findings particularly support this notion. When split at its median, total protein intake as a percentage of energy intake was not related to lower-body strength, lower-body endurance, and handgrip strength, whereas ABPI split at the median was positively associated with all these measures.

There are some limitations to our investigations. We cannot determine from our primary results if the panoramic feature inaccurately quantified muscle size, because our study lacked a measure of criterion validity in the form rectus femoris muscle thickness and cross-sectional area assessed using MRI or CT. Another caveat to our findings regarding ultrasonography is the skill of our sonographers. Although our sonographers were trained and showed good reliability, ICCs were greater than 0.95 for all measures other than B-mode

echo intensity, which was equal to 0.81; they were and are not professional sonographers. Panoramic ultrasound is a more difficult method to perform, as the probe must be moved while keeping light, consistent pressure during imaging. Our results regarding panoramic ultrasonography and knee extensor performance may indicate, then, that the method should only be performed by those with highest levels of skill. Nonetheless, measures from panoramic ultrasonography were related to summed peak torque and handgrip strength, indicating these measures were related to overall performance. Another potential limitation was the assessment of anatomical as opposed to physiological CSA, as physiological CSA of pennate muscles, such as the rectus femoris, is thought to be more closely related to strength [10].

Regarding the limitations of our secondary analysis, this was a cross-sectional study incapable of establishing causality, the self-reported nature of our food-diary recording limits its accuracy, and we included three participants' physical activity data despite the fact these participants did not have enough valid wear days. Our secondary investigation did have some strengths. We objectively measured and controlled for physical activity. We verified our partitioning of protein intake into animal- and plant-based sources using regression models. We included relative energy and macronutrient intakes in our mixed models to control for differences in participants' diets outside of ABPI. Lastly, we confirmed the importance of ABPI to muscular performance by performing another set on analyses where participants were spilt at the median of percent energy from total protein.

#### **5. Conclusions**

We report that measures of muscle thickness and CSA derived from panoramic ultrasonography are more closely related to overall strength than the same measures derived from B-mode ultrasound images. Thus, panoramic images may be a suitable method to measure muscle size and estimate overall muscle mass when the entire transverse area of a muscle cannot be measured with a standardized B-mode image. However, measures of muscle size from B-mode images were more closely related to the performance of knee extensors alone, suggesting that B-mode images may be better measures of individual muscles or muscle groups. Echogenicity of the rectus femoris was unrelated to its specific force and to overall lower-body strength. Instead, echogenicity was related to handgrip strength and knee extensor endurance. Finally, we found a positive relationship between ABPI and lower-body strength, lower-body endurance, and handgrip strength when controlling for physical activity and diet.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2077-038 3/10/5/988/s1: Supplementary Data File.

**Author Contributions:** Conceptualization: W.B., S.L.D., K.J.H., J.F.K., and S.N.S.; methodology: W.B., K.J.H., and S.N.S.; validation: W.B., K.J.H., and S.N.S.; formal analysis: W.B., C.J.K., and S.N.S.; investigation: N.D.D., N.R.J., C.J.K., and K.A.T.; resources: W.B., S.L.D., K.J.H., J.F.K., and S.N.S.; data curation: N.R.J.; writing—original draft preparation: N.R.J.; writing—review and editing: W.B., N.D.D., S.L.D., K.J.H., J.F.K., C.J.K., S.N.S., and K.A.T.; visualization: C.J.K. and N.R.J.; supervision: W.B., S.L.D., K.J.H., J.F.K., and S.N.S; project administration: K.J.H. and S.N.S.; funding acquisition: W.B., S.L.D., K.J.H., J.K., and S.N.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This project was funded by the Beef Checkoff through the National Cattlemen's Beef Association and the Minnesota Beef Council FAR26929.

**Institutional Review Board Statement:** The study was approved by the North Dakota State University Institutional Review Board (#HE26929 & 26153) and complied with the Helsinki Declaration of 1983.

**Informed Consent Statement:** Written informed consent was obtained from all participants in this study.

**Data Availability Statement:** Data is available as a Supplemental Data File.

**Acknowledgments:** We are grateful for the undergraduate and graduate students who helped complete the project.

**Conflicts of Interest:** This project was funded by the National Cattlemen's Beef Association. The sponsor had no role in the design, execution, interpretation, or writing of the study.

#### **References**


*Article*

## **Correlations between the Quality of Life Domains and Clinical Variables in Sarcopenic Osteoporotic Postmenopausal Women**

#### **Mariana Cevei <sup>1</sup> , Roxana Ramona Onofrei 2,\* , Felicia Cioara <sup>1</sup> and Dorina Stoicanescu <sup>3</sup>**


Received: 30 December 2019; Accepted: 4 February 2020; Published: 6 February 2020

**Abstract:** (1) Background: both sarcopenia and osteoporosis are major health problems in postmenopausal women. The aim of the study was to evaluate the quality of life (QoL) and the associated factors for sarcopenia in osteoporotic postmenopausal women, diagnosed according to EWGSOP2 criteria. (2) Methods: the study sample comprised 122 osteoporotic postmenopausal women with low hand grip strength and was divided into two groups: group 1 (probable sarcopenia) and group 2 (sarcopenia). QoL was assessed using the validated Romanian version of SarQol questionnaire. (3) Results: the D1, D4, D5, D7 and total SarQoL scores were significantly lower in women from group 2 compared to group 1. In group 2, women older than 70 years had significant lower values for D1, D3, D4, D6 and total SarQoL scores. Age, history of falls and the presence of confirmed and severe sarcopenia were predictors for overall QoL. (4) Conclusions: the frequency of sarcopenia was relatively high in our sample, with body mass index and history of falls as predictors for sarcopenia. Older osteoporotic postmenopausal women, with previous falls and an established sarcopenia diagnosis (low muscle strength and low muscle mass), were more likely to have a decreased quality of life.

**Keywords:** sarcopenia; quality of life; osteoporosis; postmenopausal women

#### **1. Introduction**

Sarcopenia is characterized by decreased muscle strength, loss of muscle mass and poor physical performance [1]. The condition is associated with aging. Aging is a complex process, involving many variables that interact with each other and include, besides genetic factors, lifestyle and chronic diseases. Even if sarcopenia is more common among older individuals, it can also occur earlier in life. It typically begins in the fourth decade of life, but the decline is accelerated after the sixth decade [2,3].

The decrease in muscle strength and muscle mass contributes to the loss of the ability to live independently and thus becomes an important public health problem. Sarcopenia is associated with physical disability, poor physical performance, functional decline, falls, and hospitalization [4]. Multimorbidity is frequent in older individuals and some diseases, such as heart failure or chronic obstructive pulmonary disease, accelerate the loss of muscle strength and mass, creating a vicious cycle [5]. All these have a major impact on the patient′ s quality of life [6]. Sarcopenia also increases the risk of falls. There is a high risk for hip fractures, as loss of muscle mass is frequently associated with loss of bone [5]. The high risk of falls in sarcopenic patients was found to be regardless of age,

gender and other confounding factors [7]. In turn, falls are associated with functional deterioration, physical disability, impairment in activities of daily living, increased morbidity and mortality [8]. In a meta-analysis that included 17 studies, a significant association between sarcopenia and fractures was found, independent of study design, study population, gender, sarcopenia definition, geographical area or study quality [9].

Considering the specificities of older individuals, a sarcopenia-specific quality of life questionnaire (SarQoL) has been developed [10,11] and validated [12]. The Romanian version of the SarQoL® was validated in 2017 [13]. Previous studies have reported the associated factors and the effects on the quality of life in adults with sarcopenia, using different diagnostic criteria. Only a few studies used the revised criteria of the European Working Group on Sarcopenia in Older People (EWGSOP2) [14–16]. The purpose of the present study was to evaluate the quality of life and the associated factors for sarcopenia in Romanian osteoporotic postmenopausal women, using the EWGSOP2 diagnostic criteria.

#### **2. Materials and Methods**

#### *2.1. Study Design and Participants*

Participants for this observational study were recruited from the postmenopausal women admitted to Medical Rehabilitation Clinical Hospital Băile Felix, România. To be selected, participants had to be previously diagnosed with primary osteoporosis (T-score ≤ −2.5, evaluated by DXA) and to have low hand grip strength. Low hand grip strength was defined according to the EWGSOP2 recommended cut-off of < 16 kg for women and was used to quantify the loss of muscle strength [1]. Criteria for exclusion were: (1) severe mobility disorders of the weight-bearing joints and cases with neurological conditions that affect balance and gait; (2) inability to walk for at least 10 min without a walking aid; (3) history of hip or knee arthroplasty; (4) inflammatory musculoskeletal conditions; (5) malignancies; (6) infectious diseases, (7) diabetic neuropathy; (8) cognitive impairments.

All participants provided written informed consent. The study complied with the Declaration of Helsinki and was approved by the Local Ethics Commission for Scientific Research of Medical Rehabilitation Clinical Hospital Băile Felix, România (4016/30.04.2018).

#### *2.2. Assessments*

Socio-demographic and clinical data (age, weight, height, body mass index, marital status, occupational status, years of menopause, history of and tendency towards falls, history of osteoporotic fractures, clinical conditions) were collected by interview and from medical documents. From the medical documents, the appendicular lean muscle mass determined by dual-energy X-ray absorptiometry was recorded for each participant in the study. Based on these results, and according to the recommended EWGSOP2 cut-off points for skeletal muscle mass index (appendicular lean mass/height<sup>2</sup> ), the participants were categorized as having low muscle mass (<5.5 kg/m<sup>2</sup> ) and normal muscle mass [1].

#### 2.2.1. Physical Performance

Physical performance was examined by the Timed Up&Go test, with the G-Walk system (BTS Bioengineering, Milan, Italy). It uses a validated wireless inertial sensor, made up of four inertial platforms, each composed of a tri-axial accelerometer, a tri-axial gyroscope and a magnetometer [17]. The G-sensor was attached to the participants fifth lumbar vertebra. The subjects were asked to stand up from a chair, to walk along a 3 m pathway at a self-selected speed, turn around and walk back to the chair and sit down. The recorded data were transmitted to the PC through a Bluetooth connection and processed by the BTS G-studio software (BTS Bioengineering, Milan, Italy). Women who scored ≥ 20 s were considered to have low physical performance.

Cases with low muscle strength were classified as having probable sarcopenia. We considered all participants that met the two EWGSOP2 diagnostic criteria-low muscle strength and low muscle mass—to have confirmed sarcopenia. Women with confirmed sarcopenia and low physical performance were categorized as having severe sarcopenia, according to the EWGSOP2 revised criteria [1]. We divided the study sample in two groups: group 1 comprised participants with probable sarcopenia (*n* = 58) and group 2, those with an established sarcopenia diagnosis, which included participants with confirmed and severe sarcopenia, according to EWGSOP2 (*n* = 64).

#### 2.2.2. Quality of Life

The quality of life was assessed using the validated Romanian version of SarQol questionnaire (Sarcopenia Quality of Life). This is a multidimensional questionnaire, evaluating seven domains of health-related quality of life—physical and mental health (D1), locomotion (D2), body composition (D3), functionality (D4), activities of daily living (D5), leisure activities (D6) and fears (D7) [10]. The 22 questions are rated on a 4-point Likert scale. Each domain is scored from 0 to 100 and an overall score is calculated. A higher score reflects a higher quality of life [12]. The SarQol questionnaire has good internal consistency and construct validity, good discriminative power and good responsiveness [12–14,18].

#### *2.3. Statistical Analysis*

The statistical analysis was performed using the Medcalc Statistical Software version 19.1 (MedCalc Software bv, Ostend, Belgium). All data were tested for normality with the Shapiro–Wilk's test. Descriptive statistics were calculated for all socio-demographics' characteristics (frequencies, means and standard deviation), SarQoL scores and TUG (median and interquartile range (IQR)). Between-groups differences were assessed using the independent*t*-test and Mann–Whitney test, respectively. Categorical data were compared using Chi-squared test. Logistic regression analysis was used to identify the factors associated with sarcopenia. Odds ratios (OR), 95% confidence intervals (CI) and *p* values were reported. Spearman rank correlation coefficient was used to assess the relationship between socio-demographic and clinical factors and the SarQoL scores. Variables that demonstrated significance were then entered into a stepwise multiple linear regression analysis to assess the predictors of quality of life, with SarQoL domains and total scores as a dependent variable. The significance level was set at *p* < 0.05 for all tests.

#### **3. Results**

The study sample comprised 122 women (mean age 67.02 ± 8.3 years) (ranging between 48 and 83 years) that met the inclusion criteria and agreed to participate in the study. More than half of the participants (52.46%) were diagnosed with confirmed and severe sarcopenia.

Table 1 summarizes the characteristics of the participants. There were no significant differences between the two groups in participants' characteristics, except for weight, BMI and fall history. There was a higher percent of overweight and obese women in group 1 compared to group 2 (*p* < 0.0001). The proportion of overweight or obese women in our sample was 69.67%. A total of 93.10% of participants with probable sarcopenia and 48.43% of those with sarcopenia were overweight or obese. Women with sarcopenia had a higher frequency of history of falls than those with probable sarcopenia (*p* = 0.03).


**Table 1.** Socio-Demographic and Clinical Characteristics.

The associations between the socio-demographic and clinical factors and the presence of sarcopenia were analysed by logistic regression. The factors significantly associated with sarcopenia were BMI (OR 0.79, 95%CI 0.71–0.88, *p* < 0.0001) and the history of falls (OR 2.84, 95%CI 1.13–7.09, *p* = 0.003). After adjusting for covariates (age, marital status, number of comorbidities and years since menopause), multiple logistic regression showed that BMI (OR 0.77, 95%CI 0.69–0.87, *p* < 0.0001) and the history of falls (OR 3.95, 95%CI 1.38–11.29, *p* = 0.01) together can predict the sarcopenic status. A lower BMI associated with at least one fall in the past would predispose osteoporotic postmenopausal women to sarcopenia.

Table 2 presents the total scores, as well as each domain scores of the SarQoL questionnaire. The D1, D4, D5, D7 and total SarQoL scores were significantly lower in women from group 2 compared to group 1.

In the whole study sample, significant lower scores were observed for D1, D4, D5, D7 and total SarQoL scores in the > 70 years group compared to the other two age groups, and for D2 and D3 in the > 70 years compared to the < 60 years group. In the probable sarcopenia group, no significant differences in all the SarQoL scores were observed between age groups. In group 2, women older than 70 years had significantly lower values for D1, D3, D4, D6 and total SarQoL scores than those from the other two age groups. For the D5 and D7 domains, women from group 2, older than 70 years, had significantly lower scores than those younger than 60 years (*p* < 0.05). When comparing the SarQoL scores between the two groups based on age, significantly lower scores were recorded only in the >70 years old group for D3, D4, D5 and total scores.


**Table 2.** Results of the SarQol Questionnaire in the Three Age Groups.

Data are presented as median and (IQR); *p* a relates to group 1–group 2 comparison (*p* < 0.05); <sup>b</sup> relates to the > 70 years and 60–69 years comparison (*p* < 0.05); <sup>c</sup> relates to the >70 years and <60 years (*p* < 0.05).

Physical performance did not differ significantly between the two groups. Low physical performance assessed with TUG (TUG > 20 s) was observed in 28 women from group 1 (48.27%) and in 34 women from group 2 (53.12%). According to the EWGSOP2 criteria, 53.12% women from group 2 were classified as having severe sarcopenia when using TUG performance. No age differences were observed between those with confirmed sarcopenia and those with severe sarcopenia.

In the whole study sample and in the probable sarcopenia group, significant greater TUG scores were observed in women older than 70 years compared to those younger than 60 years (21.8(17.9–30.6) vs. 16.3(13.3–21.38) s, *p* = 0.001 for the whole sample; 25.10(19.06–33.3) vs. 13.90(9.54–17.18) s, *p* < 0.001 for the probable sarcopenia group).

Significant negative correlations were found between SarQoL domains and total scores and some of the socio-anthropometric data for the whole study sample (Table 3). The history of falls and the number of comorbidities were negatively correlated with all SarQoL scores, except the D6 domains.


**Table 3.** Correlations between Sarqol Domaines and Clinical Variables.

Data represents the Spearman correlation coefficient; \* *p* < 0.05

The stepwise multiple linear regression analysis with SarQoL total score as a dependent variable revealed a negative association with age, history of falls and being sarcopenic (adjusted *R* <sup>2</sup> = 0.238; F3,118 = 13.59, *p* < 0.0001). Being older, sarcopenic with at least one fall in the past would negatively affect the quality of life of osteoporotic postmenopausal women. Table 4 shows the results of the regression analysis for all the SarQoL scores. In all regression models, history of falls was negatively correlated with all quality of life questionnaire domains, indicating that osteoporotic postmenopausal women with low muscle strength and falls in the past will have a poorer quality of life.


**Table 4.** Multiple Linear Regression Analysis with the Total and the Seven Domain Scores of SarQoL as Dependent Variables.

#### **4. Discussion**

The main aim of this study was to assess the relationship between sarcopenia and the quality of life in osteoporotic postmenopausal women. Both sarcopenia and osteoporosis are major health problems in postmenopausal women, negatively affecting the quality of life [19–21], the incidence of

falls, and mortality [22–25]. To the best of our knowledge, there are no studies investigating the quality of life in Romanian postmenopausal osteoporotic women diagnosed with sarcopenia according to the updated EWGSOP diagnostic criteria.

There are several definitions, diagnostic criteria and cut-offs used for the diagnosis of sarcopenia [1,26–31]. In our study, we used the revised EWGSOP2 criteria. The percentage of confirmed and severe sarcopenia in osteoporotic postmenopausal women aged between 48 and 83 years at the time of assessment was 52.46%. Similar results were also found in other studies, showing the association of sarcopenia and osteoporosis [32–37]. Walsh et al., reported a similar prevalence of sarcopenia of 50% in osteoporotic postmenopausal women, using the loss of muscle mass for the sarcopenia diagnosis [38]. Hamad et al. found that sarcopenia was present in 74.6% of postmenopausal women with osteoporosis, supporting the results of Yoshimura that osteoporosis increases the risk of sarcopenia [39,40]. Studies indicate that the prevalence of sarcopenia increases with age [41,42]. In our study, the percentage of osteoporotic postmenopausal women diagnosed with sarcopenia increased with age, with 43.75% of women being older than 70 years. The true prevalence of sarcopenia cannot be correctly estimated, since various definitions, cut-offs or populations were used across studies.

History of falls and BMI were significantly associated with the presence of sarcopenia in osteoporotic postmenopausal women. Our results showed that osteoporotic postmenopausal women with at least one fall in the past had a significantly higher risk of developing sarcopenia. Similar results were presented by Clynes et al., who reported an association of falls in the last year and sarcopenia, diagnosed using the IWGS (International Working Group of Sarcopenia) definition, but not the EWGSOP one [43]. In their meta-analysis, Yeung et al. also reported a positive association between sarcopenia and falls [9]. Sepulveda-Loyola et al. found a strong association between osteosarcopenia (defined as the concomitant presence of osteoporosis/osteopenia with sarcopenia [44]) and falls and fractures history in community-dwelling older adults [45]. Other prospective studies have reported the association between sarcopenia and the incidence and risk of falls [7,46–48].

We found that BMI was lower in sarcopenic women than in those with probable sarcopenia from group 1. There was a higher percent of overweight and obese women with probable sarcopenia compared to those with an established sarcopenia diagnosis. In the sarcopenic group, we identified 14.06% cases of sarcopenic obesity. In recent years, the prevalence of obesity combined with sarcopenia had increased, resulting in a high-risk geriatric syndrome. Affected individuals are at risk of synergistic complications from both sarcopenia and obesity [49].

The logistic regression results in our study showed that osteoporotic postmenopausal women with a higher body mass index had a significantly reduced risk of developing sarcopenia. Similar results were found in previous studies [50–52], although in these studies the comparisons were made with non-sarcopenic subjects. Other studies also reported the protective effect of high body mass against sarcopenia in Asian population [53–55]. Moreno-Aguilar et al. found that a higher BMI represents a protective factor against the presence of osteosarcopenia [56]. Despite these findings, in a recent meta-analysis Shen et al. suggested, as well as Gonzales et al. in 2017, that BMI should not be used for making clinically important decisions at the individual patient level, since it could not differentiate between body weight components (body fat and lean mass) [57,58].

The SarQoL questionnaire is a specific health-related quality of life questionnaire for sarcopenia and muscle impairments [59]. Previous studies have demonstrated the ability of SarQoL to discriminate sarcopenic individuals with regard to their quality of life, as long as for the diagnosis of sarcopenia both muscle mass and muscle strength criteria were used [12,13,59–61]. The present study showed that osteoporotic postmenopausal women with probable and established sarcopenia had a reduced quality of life, as assessed with the SarQoL questionnaire. Our results were slightly lower than those obtained in previous studies by the sarcopenic participants [12,14,59–62]. We have to mention that, in previous studies, the EWGSOP criteria were used for establishing the diagnosis of sarcopenia, with a few exceptions where the revised EWGSOP2 criteria were used [14,62,63]. We found that the domains of physical and mental health (D1), functionality (D4), activities of daily living (D5), fears

(D7) and total SarQoL scores were significantly lower in women with sarcopenia than those with probable sarcopenia. For locomotion (D2), body composition (D3) and leisure activities (D6) domains we have not found significant differences between sarcopenic groups. Similar results were found for the D6 domain in the study of Gasparik et al., with no significant differences between sarcopenic and non-sarcopenic participants when using the Romanian version of the SarQoL, as well as in the study of Konstantynowicz et al., who used the Polish version of the SarQoL [13,60]. The reason could be due to the fact that Romanian and Polish older people are not involved in many leisure activities [60].

In our study, osteoporotic postmenopausal women with sarcopenia who were older than 70 years had significantly lower values for physical and mental health (D1), body composition (D3), functionality (D4), leisure (D6) and total SarQoL scores than the younger ones. In the probable sarcopenia cases, the SarQoL scores were not influenced by age.

The negative impact of sarcopenia on quality of life has been largely investigated, although different criteria and questionnaires were used. The physical function domain of the quality of life has been proved to be impaired in sarcopenic patients, as assessed by the SF-36 questionnaire [52,64–67].

The multiple regression analysis in the present study showed a significant impact of age, history of falls and the presence of sarcopenia on the overall quality of life of postmenopausal osteoporotic women, as assessed with the SarQoL questionnaire. Older osteoporotic postmenopausal women with previous falls were more likely to have lower scores on physical and mental health (D1), functionality (D4) and activities of daily living (D5) domains. In association with the history of falls, the number of comorbidities was found to be a predictor only in the physical and mental health domain (D1) and body composition domain scores, respectively. Years since menopause, along with the history of falls, negatively influenced the fear domain (D7) score.

Several limitations of this study should be addressed. The study sample comprised only osteoporotic postmenopausal women with low grip strength, and no control group (premenopausal, non-sarcopenic) was included. Another issue that has to be mentioned is that the number of comorbidities was quite high and could influence the quality of life. The sample could have also been biased compared to the normal population, since the subjects were recruited from a rehabilitation clinic.

#### **5. Conclusions**

In summary, in our sample of osteoporotic postmenopausal women, the frequency of sarcopenia, as defined with the EWGSOP2 criteria, was relatively high. The body mass index and the history of falls could predict, together, sarcopenia in osteoporotic postmenopausal women. Our results showed that osteoporotic postmenopausal women with at least one fall in the past and a lower body mass index had a significantly higher risk of developing sarcopenia. History of falls and the number of comorbidities were negatively correlated with all quality of life questionnaire domains, indicating that postmenopausal women with low muscle strength and falls in the past will have a poorer quality of life. Older osteoporotic postmenopausal women, with previous falls and a confirmed sarcopenia diagnosis (low muscle strength and low muscle mass) were more likely to have a decreased quality of life. Future studies are required to identify women at risk, in order to reduce the prevalence of sarcopenia and its negative effects.

**Author Contributions:** Conceptualization, M.C., D.S. and R.R.O.; methodology, M.C., D.S., F.C. and R.R.O.; formal analysis, M.C., D.S., F.C. and R.R.O.; investigation, M.C. and F.C.; data curation, D.S. and R.R.O.; writing—original draft preparation, M.C., D.S., F.C. and R.R.O.; writing—review and editing, M.C., D.S. and R.R.O.; visualization, M.C., D.S., F.C. and R.R.O.; supervision, M.C. All authors have read and agreed to the published version of the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Using the Updated EWGSOP2 Definition in Diagnosing Sarcopenia in Spanish Older Adults: Clinical Approach**

**Anna Arnal-Gómez 1,2 , Maria A. Cebrià i Iranzo 1,3,4,\* , Jose M. Tomas 5,6 , Maria A. Tortosa-Chuliá 7,8 , Mercè Balasch-Bernat 1,4 , Trinidad Sentandreu-Mañó 1,6 , Silvia Forcano <sup>3</sup> and Natalia Cezón-Serrano 1,4**


**Abstract:** Recently the European Working Group on Sarcopenia in Older People (EWGSOP2) has updated diagnostic criteria for sarcopenia, which consist of one or more measures of muscle strength, muscle mass, and physical performance, plus an initial screening test called SARC-F. The main objective was to compare the number of cases of sarcopenia, using the different measurements and screening options. A cross-sectional study was conducted on Spanish older adults (*n* = 272, 72% women). Combining the different measures proposed by the steps described in the EWGSOP2 algorithm, 12 options were obtained (A–L). These options were studied in each of the three models: (1) using SARC-F as initial screening; (2) not using SARC-F; and (3) using SARC-CalF instead of SARC-F. A χ 2 independence test was statistically significant (χ 2 (6) = 88.41, *p* < 0.001), and the association between the algorithm used and the classification of sarcopenia was moderate (Cramer's V = 0.226). We conclude that the different EWGSOP2 measurement options imply case-finding differences in the studied population. Moreover, when applying the SARC-F, the number of people classified as sarcopenic decreases. Finally, when SARC-CalF is used as screening, case finding of sarcopenic people decreases. Thus, clinical settings should consider these outcomes, since these steps can make preventive and therapeutic interventions on sarcopenia vary widely.

**Keywords:** sarcopenia; older adults; diagnostic criteria; clinical

#### **1. Introduction**

The prevalence and impact of sarcopenia increase with age, and consequently, global aging of the population has turned sarcopenia into a public health concern of great priority both for clinicians and researchers [1]. Thus, the concept of sarcopenia has evolved in recent years at the same time that the number of scientific publications has increased in order to identify its possible causes and consequences [2–4].

Although there are different international teams which have published their guidelines or consensus for sarcopenia [5], the European Working Group on Sarcopenia in Older People of 2010 (EWGSOP) guideline has been one of the most widely used and has catalyzed research activity of sarcopenia worldwide [6–8]. In 2018, the Working Group updated the original definition (EWGSOP2), which since then considers low muscle strength as an essential characteristic of sarcopenia, uses detection of low muscle quantity or quality to

**Citation:** Arnal-Gómez, A.; Cebrià i Iranzo, M.A.; Tomas, J.M.; Tortosa-Chuliá, M.A.; Balasch-Bernat, M.; Sentandreu-Mañó, T.; Forcano, S.; Cezón-Serrano, N. Using the Updated EWGSOP2 Definition in Diagnosing Sarcopenia in Spanish Older Adults: Clinical Approach. *J. Clin. Med.* **2021**, *10*, 1018. https://doi.org/10.3390/ jcm10051018

Academic Editor: Gianluca Testa

Received: 30 December 2020 Accepted: 18 February 2021 Published: 2 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

confirm its diagnosis, and regards poor physical performance as confirmation of severe sarcopenia [7].

Therefore, in this recent definition (EWGSOP2), muscle strength is brought to the forefront of the diagnostic algorithm [9]. To measure muscle strength, handgrip strength or chair stand are recommended; for measuring muscle mass, two different options are given in order to adjust Appendicular Skeletal Muscle Mass (ASM) either by height squared, weight, or body mass index (BMI); finally, for physical performance, four assessment options are given: gait speed, the Short Physical Performance Battery (SPPB), the Timed-Up and Go test (TUG), and the 400 m walk [7]. Therefore, one or more measures of muscle strength, muscle mass, and/or physical performance together with gender-specific cut-off points for some of these measurements are needed for diagnosing sarcopenia [10,11]. From the clinical perspective, it has to be taken into account that these different options imply that the correct implementation of sarcopenia diagnosis in daily clinical practices requires many factors such as acquisition and financial costs of diagnostic measurement equipment, evaluator training and knowledge, and time constraints of diagnostic measures, among other factors [12]. The assessment in sarcopenia has become a challenge for healthcare professionals in order to identify those who may benefit from intervention [13], leading to a small percentage using diagnostic measures in clinical practice [12]. Therefore, while all different options of the definition are convenient and reliable [12,14], the impact of the different measurements on case finding of sarcopenia is to be elucidated and could help transfer the diagnosis of sarcopenia from research to the clinical context [12,14].

In addition, to facilitate the detection of sarcopenia, a screening test called SARC-F has been proposed to be carried out, before performing the measurements of strength and muscle mass, as indicative of the risk of sarcopenia [15]. SARC-F consists of five questions answered by the patients themselves, so it is a simple, practical, and easily applied screening tool for older adults and for the applicant. However, the use of SARC-F is not mandatory for healthcare professionals, except with screening purposes in high-risk patients [16]. Thus, EWGSOP2 recommends the SARC-F questionnaire as a way to obtain self-reports from patients with signs of sarcopenia and as a formal approach [16].

Moreover, although in previous studies conducted in community-dwelling older adults, SARC-F has shown very good specificity to diagnose sarcopenia, its sensitivity is low, which may be not desirable for a questionnaire aimed at screening purposes [17–20]. With the intention to solve this, SARC-CalF, which adds calf circumference (CC) to SARC-F, has been suggested as an option that may significantly increase the sensitivity of SARC-F [21]. If not only community-dwelling people are studied, but also institutionalized older adults are included, a broader population is characterized and therefore clinicians have more information about the use of these tools. Therefore, they should be validated in different populations and living settings [21], plus the new EWGSOP2 definition has to be taken into account.

It was hypothesized that although there are different measurement options for each step of the algorithm of the EWGSOP2 definition, no difference in case finding will be found in older adults, allowing healthcare professionals to use the most feasible in their daily clinical practice. We also hypothesized that by not using the SARC-F, case finding of sarcopenia could be increased. Moreover, it may be increased when using the SARC-CalF instead of the SARC-F in these populations.

Therefore, the aim of this study was to compare the number of cases of sarcopenia in older adults using the different measurement options of each step of the algorithm of the European Working Group on Sarcopenia in Older People 2018 (EWGSOP2). We also aimed to evaluate the impact of using SARC-F, SARC-CalF, or no screening on the case finding of sarcopenia in Spanish older adults living in the province of Valencia.

#### **2. Experimental Section**

#### *2.1. Study Design*

A multicenter cross-sectional study was carried out between January 2019 and February 2020 in institutionalized and community-dwelling older adults, living in the province of Valencia (Spain). This study was approved by the Ethics Committee for Human Research of the University of Valencia (H1542733812827) and was conducted in accordance with the Declaration of Helsinki. This research was registered in the ClinicalTrials.gov database (ID: NCT03832608). Before entering in the study, participants signed a written consent, briefed beforehand.

#### *2.2. Participants*

The sample included 272 adults aged 65 or older, living in the community (*n* = 139) or institutionalized in residential facilities (*n* = 133). Candidates were not included if they: (1) had edema which could interfere with the bioimpedance analysis (BIA); (2) had a cognitive impairment measured with the Mini-Mental State Examination (MMSE) < 18 points [22]; (3) were suffering from any acute or unstable chronic disease, or had a hospital admission in the last month.

#### *2.3. Sarcopenia Definition*

The algorithm of the EWGSOP2 was followed for case finding and diagnosing sarcopenia and determining its severity [7]. It included the SARC-F and the measurements of muscle strength, muscle quantity, and physical performance.

The SARC-F questionnaire is composed of five items questioning strength, assistance in walking, rise from a chair, stair climbing, and falls. It is scored between 0 and 2, and it allows identifying cases with a score of ≥4 points from a total of 12 points [15].

Muscle strength was measured by:


Muscle quantity as Appendicular Skeletal Muscle Mass (ASM) was measured with BIA using the Bodystat® 1500MDD (Bodystat Ltd., Douglas, UK). This device was calibrated previous to the measurements. Prior to the assessment, the following criteria were checked [26,27]: participants could not have done previous physical exercise; 2–3 h of fasting was needed, including alcohol or a large amount of water, and emptying their bladder; every metal piece was taken off; and the test was not implemented if they were wearing a pacemaker and/or had edema (diagnosed by the physician). When applying the BIA test (alternating sinusoidal electric current of 200 µA at 50 kHz), the patient was asked to lie in supine position, on a nonconductive surface, with no contact between the limbs. Electrodes were applied with an ipsilateral tetrapolar method, on previously cleaned skin. The electrodes of the upper limb were placed at the knuckles and wrist, and those of the lower limb were placed at the metatarsal head bones line and the anterior side of the ankle. ASM was calculated following Sergi's BIA equation: ASM (kg) = −3.964 + (0.227 × RI) + (0.095 × weight) + (1.384 × sex) + (0.064 × Xc) [28]. The proposed ASM cut-offs were:


Physical performance of participants was measured by:


Following these assessments, participants were classified according to the EWGSOP2 algorithm [7,8]: (1) they had probable sarcopenia with a score of ≥4 points SARC-F and low muscle strength (grip strength < 27 kg for men and <16 kg for women; or chair stand > 15 s); (2) they had confirmed sarcopenia when low quantity muscle was also detected (ASM < 20 kg for men and <15 kg for women; or ASMI <7.0 kg/m<sup>2</sup> for men and <5.5 kg/m<sup>2</sup> for women); and (3) they had severe sarcopenia, when low physical performance was added (gait speed < 0.8 m/s; SPPB ≤ 8 points; or TUG ≥ 20 s).

#### *2.4. Additional Measurements*

Anthropometric variables: Age and gender were registered; body weight (kg) was measured using a Tanita BC 601 (TANITA Ltd., Amsterdam, The Netherlands); height (cm) was assessed with a stadiometer SECA 213 (Seca Ltd., Hamburg, Germany); and finally, BMI (kg/m<sup>2</sup> ) was calculated.

SARC-CALF consists of the same five items as SARC-F which are scored the same [36] and adds the CC that was measured as the widest circumference of calf. The CC item is scored as 0 points when the participant had more than 31 cm circumference and as 10 points if it was less than or equal to 31 cm. A SARC-CalF ≥ 11 indicates positive screening for sarcopenia [37–39].

All the assessments were done on the same day for each participant, and different physiotherapists took these measurements for all the samples. Intraclass Correlation Coefficients (ICCs) were calculated to know the interrater reliability, and they ranged from 0.802 to 0.985, which may be considered very good reliability (values between 0.75 and 0.90 indicate good reliability; values over 0.90 show excellent reliability) [40].

#### *2.5. Applied Models*

Three models were applied: Model 1: using SARC-F as initial screening; Model 2: not using any initial screening; and Model 3: using SARC-CalF as initial screening instead of SARC-F. By combining the different measures proposed by the steps described in the EWGSOP2 algorithm (Find–Assess–Confirm–Severity), 12 options were obtained (A to L), and to each one of the three models, each of their twelve options was tested (Table 1).



\* Model 1: using SARC-F as initial screening; Model 2: not using any initial screening; and Model 3: using SARC-CalF instead of SARC-F.

#### *2.6. Statistical Analyses*

With descriptive purposes, means, standard deviations, and 95% confidence intervals (CI) for all variables were calculated. All statistical analyses were performed with R [41], also employing the packages vcd [42] and DescTools [43]. Descriptive statistics (proportions) of multinomial variables were performed [44] including 95% CI for the proportions of each category by the method of Glaz and Sison [45,46]. Chi-square tests of goodness-of-fit and independence were also performed together with their association measures (Pearson residuals and Cramer's V). The CI for V coefficient was bias-corrected [47]. Whenever multiple statistical tests were made, the Sidak correction was employed.

#### **3. Results**

#### *3.1. Sample Characteristics*

A total of 272 participants were included in this study. The age range for all the participants was 65–97 years, the mean age was 77.0 (8.7) years old, and according to setting, the mean was 72.3 and 81.9 years old for community dwelling and institutionalized participants, respectively. Seventy-two percent of participants (*n* = 197) were women (Table 2).


**Table 2.** Characteristics of the participants (*n* = 272) according to setting and gender: mean (standard deviation) and [95% confidence interval].

Abbreviations: BMI = Body Mass Index; ASM = Appendicular Skeletal Muscle Mass; SPPB = Short Physical Performance Battery; TUG = Timed-Up and Go test. *p*-value unpaired Student's t-test. \* *p* < 0.05; † *p* < 0.001.

#### *3.2. Analysis of the Models*

For each of the three models, and each of their 12 options (A to L), 95% CI and each category of classification (no sarcopenia, probable sarcopenia, confirmed sarcopenia, and severe sarcopenia) were calculated. These CIs are presented in Figures 1–3.

**Figure 1.** Multinomial 95% confidence intervals for the proportion of each category in the 12 options of Model 1.

**Figure 2.** Multinomial 95% confidence intervals for the proportion of each category in the 12 options of Model 2.

**Figure 3.** Multinomial 95% confidence intervals for the proportion of each category in the 12 options of Model 3.

Once these CIs were calculated, they were averaged for each model, and each of the 12 options (A to L) in each model was compared with a goodness-of-fit chi-square test with expected probabilities the average probabilities of each model. Therefore, the 12 tests within each model (algorithm) tested whether the classification of the different steps was statistically equal or different. Table 3 offers the results of all these chi-square tests. Regarding Model 1, all but two tests showed statistical significance, indicating that the different steps of the algorithm significantly affect the classification. Model 2 tests showed significant results in all cases, and therefore this supports that the different steps of the algorithm lead to different classifications. However, in Model 3, the results showed no statistical significance, and therefore for this algorithm, the different steps do not lead to significantly different classifications.

**Table 3.** Goodness-of-fit chi-square tests, probability level corrected with Sidak method.


Notes: Model 1: using SARC-F; Model 2: not using any initial screening; Model 3: using SARC-CalF; *p*-values corrected with Sidak's correction.

> A chi-square independence test was performed to compare the classifications into the different groups the three algorithms made. This chi-square was statistically signifi

cant (χ 2 (6) = 88.41, *p* < 0.001), and the association between the algorithm used and the classification of sarcopenia was moderate (Cramer's V = 0.226, 95% CI [0.177, 0.276]).

In addition, we have analyzed how each model on the whole is associated with severity levels. Figure 4 graphically presents the association based on the Pearson's residuals. It can be seen that Model 1 is not significantly associated with any classification as represented by the grey color. However, Model 2 is associated with the classification into the different groups of sarcopenia with a positive association (blue color) with the levels of severity, being higher with probable sarcopenia, and with negative association (red color) with nonsarcopenic older adults. On the contrary, Model 3 is associated positively (blue color) with no sarcopenia.

#### **4. Discussion**

The present study showed that using the different measurement options for each step of the EWGSOP2 implied differences in case finding in the studied population. These differences have been analyzed in relation to each of the steps, describing which measurements detect more or less cases of sarcopenia. Moreover, our results indicate that when applying the SARC-F, case finding of sarcopenia decreases, thus by not applying it, more cases are found, especially among those with probable sarcopenia. Finally, when SARC-CalF is used as screening, the number of people classified as sarcopenic decreases.

To the best of our knowledge, this is the first study to analyze the different measurement options of the EWGSOP2 in Spanish older adults, and the fact that there are differences in case finding has important clinical consequences. Taking into account that sarcopenia is frequently not noticeable in earlier stages [48], detecting probable sarcopenia is of paramount importance in order to be able to start intervention. In Model 1, using SARC-F, and Model 2, using no screening, there are significant differences in case finding among most of the options. When analyzing these differences, it can globally be seen that those which use the chair stand for measuring muscle strength (G to L) are the ones that find more probable sarcopenic participants. This is interesting considering that previous research has highlighted that handgrip strength seems to be used widely for the measurement of muscle strength [13]. However, it requires the use of a calibrated handheld dynamometer under well-defined test conditions [23]. Therefore, commercial dynamometers are usually limited in clinical settings by the need to purchase specialized equipment, the relative expense, and the lack of trained staff [13]. In addition, sometimes measurement of grip is not possible due to hand disability, such as with patients who are suffering from advanced

arthritis or stroke [7]. On the whole, these facts could explain why only a small percentage of healthcare professionals use diagnostic measures in clinical practice as stated before [12]. On the other hand, in previous research, the chair stand has been shown to be able to provide a valid tool for assessing lower body strength [49]. This is in line with our results, which seem to show chair stand can be a reliable method for case finding of probable sarcopenia in the studied population. From the clinical approach, detection of cases as early as possible is important considering that it is better to prevent the skeletal muscle mass depletion and loss of strength and function rather than trying to restore them when they have progressed [50]. Therefore, for clinical settings where a handgrip dynamometer is not always available, the chair stand could be used as an alternative assessment of muscle strength [13]. This way, preventive strategies together with treatment interventions could be implemented before the muscle deterioration occurs [50].

After detecting probable sarcopenia cases, the second step of the EWGSOP2 algorithm evaluates muscle quantity. The EWGSOP2 consensus presents cut-off points for both ASMI (kg/height squared) [51] and ASM (kg) [29] for use when calculating muscle mass. In relation to Model 1, when analyzing the different options, there are more cases of severe sarcopenia in those options that previously have confirmed it by using the ASM (kg) cut-off for muscle quantity (B1, C1, E1, H1, I1, K1). Considering that low muscle mass is highly related to disability and frailty in older adults [52], measuring muscle mass in a precise way is crucial for confirming sarcopenia in this population. There is an ongoing debate about the preferred adjustment for muscle mass indices and whether the same method can be used for all populations [7]. For our population, the results show the ASMI is a more restrictive cutoff, whereas with the ASM, more sarcopenic participants are detected and, consequently, more are classified as suffering severe sarcopenia. This could also explain why G1, J1, and L1 show more cases of probable sarcopenia, since they are using the ASMI and therefore more participants are not being confirmed with sarcopenia and stay as probable. Therefore, some participants could present low strength, however, their amount of muscle mass would still be within the EWGSOP2 criteria, preventing categorization in more advanced stages of the pathology, as similarly stated in previous research [53]. Although the most accurate way to define muscle mass remains uncertain [54], our results show ASMI is more restrictive for our population, classifying them mostly as probable, whereas they could have been classified as severe if the ASM had been used. Thus, methods used to define low lean mass can make preventive and therapeutic interventions on sarcopenia vary widely.

In relation to Model 2, with no initial screening, there are significant differences in case finding among all of the options. When analyzing them individually, again the same trend can be found in relation to muscle strength and muscle mass, that is, chair stand detects more probable sarcopenia (options G to L) and ASMI is more restrictive (A, D, F, G, J, L). In relation to physical performance, the options which confirm sarcopenia with the ASM and then classify its severity with the SPPB or gait speed (B, C, and G to J) are the ones which detect more cases of severe sarcopenia. Detection of low physical performance predicts adverse outcomes [7], so it becomes of paramount importance for the clinical approach. However, in older populations, physical performance is frequently difficult to measure due to acute illness or because of dementia, gait disorder, or a balance disorder [55–57], thus finding a safe and valid assessment becomes necessary. Gait speed is considered a quick and reliable test for sarcopenia, which is why it is widely used in practice [58]. Although the SPPB also predicts outcomes [34], it is a more time-consuming test to apply and therefore, it is more used in research than in clinical practice. Therefore, and considering our results, clinicians can rely on SPPB and gait speed to detect the severe cases, although the latter can be considered as a more approachable measurement in the clinical context.

Regarding the use or not of SARC-F, it was not implemented in Model 2, and more cases were found as probable, confirmed, or severe sarcopenia. Therefore, when using SARC-F for screening in our population, it is at the expense of missing cases who would have been at least in the category of probable sarcopenia, since more cases were detected in Model 2. Although the SARC-F has shown excellent specificity [15,17,18,59–61], it has

shown some problems in relation to its low sensitivity [17,59], which means that there is a high risk of missed diagnosis of individuals who have sarcopenia. Moreover, as noted in the EWGSOP2 definition, in clinical settings, case finding should start when a patient has symptoms or signs of sarcopenia, and in these situations, further testing is recommended, the use of any screening tool not being mandatory [16]. This is in line with our results, which suggest SARC-F does not always detect possible cases of sarcopenia in our sample.

In relation to Model 3, which screens using the SARC-CalF, our results show case finding of sarcopenic people is not increased. Moreover, it is the model with which a lower amount of sarcopenic people are found. Although previous research has shown promising results regarding SARC-CalF, with a better sensitivity than SARC-F [21,39], this does not concur with our results. This may be explained by the cut-offs that have been used, since again different options are found in this regard [37], and should be addressed in future research. Moreover, the few participants that were detected as suffering from sarcopenia with any of the options of Model 3 were classified as severe sarcopenia, thus indicating they were highly impaired in their physical performance, which allows less options of recovery. From the clinical approach, if a sarcopenia screening test is used, it is expected to dismiss from further testing as many healthy individuals as possible but should also guarantee diagnosis of those who do have sarcopenia [39] in order to start the appropriate intervention, thus this may not be possible using the SARC-CalF in a population like ours.

Considering that from the three models, Model 2 has shown the highest positive associative probability in case finding of participants with sarcopenia, especially in probable and confirmed, this finding would allow clinicians to detect sarcopenia in earlier stages. This model has different options which have shown statistical differences and using one or other to detect the presence of sarcopenia can be time consuming and expensive and might require highly specialized equipment [50]. Moreover, selecting a way of diagnosing sarcopenia requires balancing the possible benefit of including functional and ASM measurements against the difficulties related to their inclusion [62]. Therefore, on the whole, those options of Model 2 which include the chair stand and use the ASM may be finding more cases of sarcopenia in its different classifications.

#### *Limitations and Strengths*

The main limitation of our study, common to other studies, is related to sample size. A larger sample size would be advisable, as well as studying case finding and implementing this model analysis in other populations besides the Spanish one to confirm our promising results. Another limitation is that the sample had a higher percentage of women, and although this is characteristic related to aged population in Spain, greater gender equality would be important in future research. Another interesting line to be implemented in the future could be analyzing how those older adults found to be sarcopenic in one model behave in the other models. However, this study offers the novelty of analyzing the different options of the EWGSOP2 to show the one that can find sarcopenia cases in an accurate way, which would promote an adequate and early intervention.

#### **5. Conclusions**

There are differences in case finding of sarcopenia in the studied Spanish older adults when the different measurement options for each step of the EWGSOP2 definition are applied. For muscle strength, the chair stand seems to be detecting more cases of probable sarcopenia, for muscle mass, ASM detects more confirmed and severe, and for physical performance, SPPB and gait speed seem to be reliable options. In addition, more sarcopenia cases are identified when no initial screening is used, therefore, in clinical practice, when a patient shows symptoms or signs of sarcopenia, a screening questionnaire may be surpassed and further testing is recommended to confirm sarcopenia. Thus, clinical settings should take into consideration that the methods used to define these steps can make preventive and therapeutic interventions on sarcopenia vary widely.

**Author Contributions:** Conceptualization, M.A.C.i.I., A.A.-G. and N.C.-S.; methodology, M.A.C.i.I., A.A.-G., N.C.-S., M.A.T.-C., M.B.-B., S.F. and T.S.-M.; formal analysis, J.M.T., M.A.C.i.I., A.A.-G. and N.C.-S.; investigation, M.A.C.i.I., A.A.-G., N.C.-S., M.A.T.-C., M.B.-B., S.F. and T.S.-M.; resources, M.A.C.i.I., A.A.-G. and N.C.-S.; data curation, M.A.C.i.I., A.A.-G., N.C.-S. and M.B.-B.; writing original draft preparation, M.A.C.i.I., A.A.-G., N.C.-S. and J.M.T.; writing—review and editing, A.A.- G., M.A.C.i.I., N.C.-S., M.A.T.-C., M.B.-B., S.F., T.S.-M. and J.M.T.; project administration, M.A.C.i.I.; funding acquisition, M.A.C.i.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Generalitat Valenciana (GV/2019/131) and by FEDER/ Ministerio de Ciencia e Innovación—Agencia (RTI2018-093321-B-100).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee for Human Research of the University of Valencia (protocol code H1542733812827 approved 18 December 2018).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Acknowledgments:** We gratefully acknowledge the participation of all community-dwelling participants as well as residents and staff of the residential facilities of La Saleta Care, Parque Luz Xirivella, Parque Luz Catarroja, El Mas Torrent, and especially Mary Martínez Martínez, the Technical Manager of La Saleta Care.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


*Review*

### **Diagnosis, Treatment and Prevention of Sarcopenia in Hip Fractured Patients: Where We Are and Where We Are Going: A Systematic Review**

**Gianluca Testa 1,\* , Andrea Vescio <sup>1</sup> , Danilo Zuccalà 1 , Vincenzo Petrantoni <sup>1</sup> , Mirko Amico <sup>1</sup> , Giorgio Ivan Russo <sup>2</sup> , Giuseppe Sessa <sup>1</sup> and Vito Pavone <sup>1</sup>**


Received: 14 August 2020; Accepted: 15 September 2020; Published: 17 September 2020

**Abstract:** Background: Sarcopenia is defined as a progressive loss of muscle mass and muscle strength associated to increased adverse events, such as falls and hip fractures. The aim of this systematic review is to analyse diagnosis methods of sarcopenia in patients with hip fracture and evaluate prevention and treatment strategies described in literature. Methods: Three independent authors performed a systematic review of two electronic medical databases using the following inclusion criteria: Sarcopenia, hip fractures, diagnosis, treatment, and prevention with a minimum average of 6-months follow-up. Any evidence-level studies reporting clinical data and dealing with sarcopenia diagnosis, or the treatment and prevention in hip fracture-affected patients, were considered. Results: A total of 32 articles were found. After the first screening, we selected 19 articles eligible for full-text reading. Ultimately, following full-text reading, and checking of the reference list, seven articles were included. Conclusions: Sarcopenia diagnosis is challenging, as no standardized diagnostic and therapeutic protocols are present. The development of medical management programs is mandatory for good prevention. To ensure adequate resource provision, care models should be reviewed, and new welfare policies should be adopted in the future.

**Keywords:** sarcopenia; hip fracture; diagnosis; treatment; prevention; dual-energy X-ray absorptiometry; bisphosphonate; β-hydroxy-β-methylbutyrate; exercise intervention

#### **1. Introduction**

Sarcopenia-related falls and fractures play an important role in our society due to the increased average age of the population [1]. Hip fractures are becoming an evolving and more current problem, as well as one of the most serious medical and social concerns. Hip fractures result in enhanced mortality, perpetual physical morbidity and reduced activities of daily living (ADL) [2,3], with a decrease of the quality of life for caregivers and an increased economic impact on society and government spending [4–6]. Nowadays, the prevention of hip fractures is considered crucial for preserving an acceptable quality of life in older patients. For these reasons, the role of the muscles trophism and function is crucial to prevent traumas in older patients [1]. Ageing is inversely related to the mass and strength of skeletal muscles, and their loss accelerates after 65 years of age, leading to an increased risk of adverse outcomes [7]. For the last 30 years, a considerable effort has been made to understand the condition of sarcopenia, and several definitions have been proposed. Sarcopenia was first defined by

Rosenberg as an age-associated loss of skeletal muscle mass [8], but recently, it has been identified as a disease, and is included in the ICD-10 code (M62.84) [9]. Several disease descriptions were suggested during the last 20 years, but substantial operative variances are present concerning definitions, including nomenclature, the technique of assessment of lean mass, the technique of standardization of lean mass to body size, cut-points for weakness and cut-points for slowness [10]. One of the most accepted was described by the EWGSOP (European Working Group on Sarcopenia in Older Persons), updated in 2018 (EWGSOP2), considering sarcopenia as a "progressive loss of muscle mass and muscle strength, associated to an increased likelihood of adverse events, such as falls, fractures, physical disability and death" [7]. Several authors investigated the differences in sarcopenia cases, agreeing with EWGSOP1 and EWGSOP2 and noting substantial discordance and limited overlap of the definitions [11,12]. Nevertheless, the EWGSOP2 is crucial suggestion to evaluate a possible condition of sarcopenia by measuring the muscle strength, muscle mass and physical performance [13]. Aging is related to variations in body structure and uncontrolled weight loss. The progressive loss of skeletal muscle mass (SMM) and strength promotes functional and physical disability, leading to poor quality of life [7]. The body composition changes were reported in several studies [7,14,15]. Cruz-Jentoft et al. [7] showed a loss of muscle strength in older patients through measurement grip strength with a dynamometer. Hida et al. [14] demonstrated a greater sarcopenia prevalence and more diminished leg muscle mass in subjects following a hip fracture than uninjured subjects with the same age. The most efficient technique to date, dual energy X-ray absorptiometry (DXA), assesses lean mass [16]. Bioelectrical impedance analysis (BIA), CT, and MRI can be used in selected cases [16]. DXA has several advantages, including low cost, low irradiation exposure and easy availability and usability. However, the difficulty of performing this examination in patients with hip fracture or in subjects undergoing recent orthopaedic surgery, due to post-surgical pain and immobility, the use of machines with non-uniform calibrations between them and the lack of universally shared protocols, makes DXA not always reliable in the quantification of MM and in the instrumental diagnosis of sarcopenia [11,17]. No specific drugs have been approved for the treatment of sarcopenia and the literature lacks evidence [16]. Research activity is focused on developing new drugs for sarcopenia, although progress has not been straightforward. Initial interest in selective androgen receptor modulators is related to small phase I and II trials [18,19]. For these reasons, the interest in sarcopenia is rising in orthopaedic surgery, due to the high prevalence of older patients, especially those suffering for hip fractures [20], and sarcopenia could be considered as a hip fracture risk factor.

The aim of this systematic review was to analyse diagnosis methods of sarcopenia in patients with hip fracture and evaluate prevention and treatment strategies described in literature.

#### **2. Experimental Section**

#### *2.1. Study Selection*

From their date of inception to 19th March 2020, two independent authors (AV and GT) systematically reviewed the main web-based databases, Science Direct and PubMed, agreeing to the Preferred Reporting Items for systematic Reviews and Meta-Analyses (PRISMA) recommendations [19]. The research string used was "sarcopenia AND (diagnosis OR treatment OR prevention) AND (femoral neck fracture OR hip fracture)". In order to extract the number of patients, mean age at treatment, sex, type of treatment, follow-up, and year of the study a standard data entry form was used for each included original manuscript. Three independent reviewers (MA, PV and DZ) performed the quality evaluation of the studies. Discussing conflicts about data were resolved by consultation with a senior surgeon (VP).

#### *2.2. Inclusion and Exclusion Criteria*

Eligible studies for the present systematic review included sarcopenia diagnosis, treatment and prevention in hip-fractured patients. The original titles and abstracts examination were selected using the following inclusion criteria: Sarcopenia, hip fractures, diagnosis and treatment and prevention with a minimum average of 6-months follow-up in last 20 years. The exclusion criteria were: Patients' cohort with no sarcopenia diseases, less than 6 months of symptoms and no human trials. Each residual duplicate, articles related on other issues or with inadequate technical methodology and available abstract were ruled out.

#### *2.3. Risk of Bias Assessment*

According to the ROBINS-I tool for nonrandomized studies [21], a three-stage assessment of the studies included risk of bias assessment was performed. The first step involved the design of the systematic review, the next phase was the assessment of the ordinary bias discovered in these manuscripts and the final was about the total risk of bias. "Low risk" and "Moderate risk" studies were considered acceptable for the review. The assessment was separately performed by three authors (MA, PV and DZ). Any discrepancy was discussed with the senior investigator (VP) for the final decision. All the authors agreed on the result of every stage of the assessment.

#### **3. Results**

#### *3.1. Included Studies*

Thirty-two manuscripts were recovered. Twenty-four articles were chosen, following the exclusion of duplicates. At the end of the first screening, according to the selection criteria previously described, nine articles were chosen as eligible for full-text reading. Metanalysis or systematic reviews were eliminated from the study. Finally, after reading the complete articles and examining the reference list, we chose seven manuscripts comprised of randomized controlled human trials (hRCT), prospective and retrospective cohort or series studies, according to previously described criteria. A selection and screening method PRISMA [22] flowchart is provided in Figure 1.

**Figure 1.** PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) flowchart.

#### *3.2. The Diagnosis of Sarcopenia in Patients A*ff*ected by Hip Fracture*

Kramer et al. [23] performed biopsies of vastus lateralis to assess the muscle changes. The sample was divided in to three groups: Healthy young women (HYW) (18–25 years), healthy older women (HEW) (>65 years) and older women (>65 years) affected by traumatic hip fracture (FEW). FEW Type 2 fibers (2.609 <sup>±</sup> <sup>185</sup> <sup>µ</sup>m<sup>2</sup> ) were noted significantly smaller compared to HEW (3.723 <sup>±</sup> <sup>322</sup> <sup>µ</sup>m<sup>2</sup> ; *p* = 0.03) and HYW (4.755 <sup>±</sup> <sup>335</sup> <sup>µ</sup>m<sup>2</sup> ; *p* < 0.001).

Hansen et al. [16] compared the Computed Tomography (CT) and dual-energy X-ray absorptiometry (DXA) efficiency in the assessment of midthigh muscle mass (SMM) and midthigh cross-sectional area (CSA) respectively, after a hip fracture with 12 months follow-up. The two measures were significantly linked to baseline (r = 0.86, *p* < 0.001). Ratios of midthigh fat to lean mass were comparably related (interclass correlation coefficient = 0.87, *p* < 0.001). Data of the change from baseline to follow-up showed a low correlation (interclass correlation coefficient = 0.87, *p* = 0.019). The assessment of muscle mass by DXA-derived midthigh slice has been shown to be reasonably accurate in comparison to a single-slice CT technique in this sample of frail older patients.

Villani et al. [24] evaluated the agreement degree between DXA and bioelectrical impedance spectroscopy (BIS) associated to corrected arm muscle area (CAMA). No significant changes (*p* = 0.78) were found when comparing fat-free mass (FFM) with BIS (FFMBIS) to FFM with DXA (FFMDXA) mean bias. Nevertheless, when included as an independent covariate, gender demonstrated an influence on variation in the mean bias over time (*p* = 0.007). The influence of BMI had no effect on change in the mean bias (*p* = 0.19). Similarly, no significant changes in the mean bias were observed between SMMDXA and SMMCAMA across each assessment time point (*p* = 0.18). At each assessment follow-up, both the techniques were observed overestimated SMM and FFM.

#### *3.3. Treatment of Sarcopenia in Patients A*ff*ected by Hip Fracture*

Flodin et al. [25] evaluated the efficacy of nutritional supplementation on body composition (BC), handgrip strength (HGS) and health-related quality of life (HRQoL) in 79 hip-fractured patients (mean age 79 ± 9 years). Patients were divided into a protein and bisphosphonate group (PB) group, bisphosphonate-only group (BO) and a control group (CG) with 12 months follow-up. All groups included the CG, received calcium and Vitamin D supplementation. No significant differences in changes of FFM Index, HGS and HRQoL were detected during the follow-up period between the groups.

Invernizzi et al. [26] assessed the essential amino acid supplementation (AAS) in hip-fractured patients. Thirty-two patients (sarcopenia-affected = 71.9%) underwent to a 2-month rehabilitative protocol combined with dietetic counselling. The AA group (16 subjects) had an AAS, while the NAA group did not receive AAS. According to Janssen criteria, both groups were divided in subgroups: Sarcopenic (Sac) and non-sarcopenic (No-Sac) patients. At 2 months follow-up, the Sac AA group (*n* = 10) obtained better significant results in the Iowa Level of Assistance scale (ILOA) and all the primary outcomes (*p* < 0.017) compared to Sac NAA cohort (*n* = 13). The No-Sac groups had similar results.

Malafarina et al. [27] investigated the effectiveness of β-hydroxy-β-methylbutyrate (HMB) oral NS on muscle mass and nutritional markers (BMI, proteins) in patients >65 years with hip fracture. Fifty-five patients (IG) received standard diet plus HMB NS and 52 patients (CG) received standard diet only. The authors used the EWGSOP criteria to diagnose sarcopenia and its prevalence among the entire population was 72%. The sarcopenia diagnostic markers were gait speed (GS), HGS and BC (assessed with BIA). Positive results were recorded in IG for grip work index (*p* = 0.188), muscle mass (MM) (*p* = 0.031) and appendicular lean mass (aLM) (*p* = 0.020). GS analysis did not show a significant difference (*p* = 0.367).

#### *3.4. Prevention of Sarcopenia in Patients A*ff*ected by Hip Fracture*

In a study by Ding-Cheng Chan et al. [28], 110 patients over 50 years of age with high-risk fracture underwent 3-month exercise interventions. According to different modalities of the exercise, the cohort were randomly divided into integrated care (IC) group and machine-based low extremities exercise (LEE) group. The authors observed a gain in limb mass in the entire cohort (1.13%, *p* < 0.05) with a significant change in the LEE group (1.13%, *p* < 0.01). Both groups obtained significant improvement in muscle strength measured with curl, press and leg extension, grip strength, gait speed, chair stand test and time up and go test. Improvements were seen in leg curl in the LEE group (29.78%, *p* = 0.001).

The most important results of the included articles were summarized (Table 1).


**Table 1.** Included studies summary. Dual-energy X-ray absorptiometry (DEXA); healthy young women (HYW); healthy elderly women (HEW); elderly women with a hip fracture (FEW); Dual-energy X-ray absorptiometry (DEXA); BIS (bioletrical impedance spectroscopy); corrected arm muscle area (CAMA); Fat-free mass (FFM) with BIS (FFMBIS); FFM with DXA (FFMDXA); handgrip strength (HGS) and health-related quality of life (HRQoL); Timed Up and Go test (TUG); Iowa Level of Assistance scale (ILOA); Mini Nutritional Assessment−Short Form (MNA-SF); Barthel index (BI); Functional Ambulation Categories (FAC).



**Table1.***Cont*.

*J. Clin. Med.* **2020**, *9*, 2997

#### **4. Discussion**

#### *4.1. General Considerations and Key Findings*

According to the review findings the diagnosis is still a challenge. The lack of an optimal instrumental tool for diagnosis in hip-fractured patients demonstrates the crucial role of physicians in these cases. The diagnosis is not instrumental data but the correct analysis of the clinical examination and patients' physical status evaluation in association with the results of the tool. At the same time, the nutritional supplementation and hip fracture prevention exercise program are mandatory to avoid the variances in body composition after midlife. Therefore, body composition evaluation is a crucial element for measuring health status in older adults.

The higher incidence of fractures, especially in the spinal column and femoral neck, is attributable to the condition of osteopenia or osteoporosis. Several authors have debated the correlation of bone mineral density (BMD) to muscle mass (MM). However, this association is controversial. Gillette-Guyonnet et al. [29] and Walsh et al. [30] claimed there was no muscle–bone relationship. On the other hand, Locquet et al. exhaustively explored the correlation between muscle and identified a subpopulation affected by the reduction in bone and muscle mass [31]. Moreover, Hirschfeld et al. suggested considering the two condition as a new pathologic disorder, where the subjects affected should be defined as "osteosarcopenic patients" [32]. The controversial findings should be explained by the different diagnosis protocols used. In fact, the sarcopenia diagnosis is often challenging, and there is not an instrumental method or standard algorithm commonly accepted for the evaluation. EWGSOP2 suggests combining clinical tests and instrumental investigations to evaluate the muscle strength, physical performance and muscle mass [11].

#### *4.2. Sarcopenia Diagnosis in Hip-Fractured Patients*

Determining grip strength is easy, inexpensive and routine in clinical practice. The evaluation requires calibrated handheld dynamometer use under well-definite exam circumstances with interpretive data from appropriate reference populations [11,33]. On the other hand, the technique measurements can be influenced by the examiner [33]. Similarly, the chair stand test (also called the chair rise test), aims to assess the quantity of time that the patient needs to rise five times from a seated position without using their arms [30]. The Gait Speed test is helpful in the evaluation of physical performance. The principles are the Short Physical Performance Battery (SPPB), and the Timed-Up and Go test (TUG), but the results can be influenced by patient compliance. The Gait Speed test is a rapid, secure and reliable test to assess sarcopenia by EWGSOP2 [11]. The patient walks for 4 m while the clinical staff records the walking speed using an electronic device or manually with a stopwatch. A Gait Speed of ≤0.8 m/s is a severe sarcopenia marker [34–36]. The SPPB is a complex test aimed to analyse gait speed using a balance test and a chair stand test. The highest score is 12 points, and a score of ≤8 points suggests inadequate physical performance. The TUG test assesses the taken time to rise from a standard chair, walk 3 m away, turn around, walk back and sit down again. A score of >20 s is indicative of poor physical performance [37].

Due to the reduced mobility in the hip-fractured patients, and consecutively, to the impossibility in performing the main tests used to assess the disease, the instrumental tools are important part of diagnosis, even if they can replace the clinical evaluation.

DXA is a more widely accessible tool to establish MM [38], and can be defined as total body SMM, as ASM or as muscle cross-sectional area of specific muscle groups [16]. New methods have been studied, including midthigh muscle measurement by CT or MRI, BIS, psoas muscle measurement with CT, the detection of specific biomarkers and other tests [16,24,25]. CT and MRI allow for a precise and detailed study of soft tissues and they offer reliable and universally shared data. On the other hand, these methods have a high cost and it is difficult to find institutes where it is possible to quickly perform them. Moreover, CT exposes patients to a high rate of irradiation [16,34]. Hansen et al. [16] compared SMM estimated by DXA to midthigh muscle CSA, determined by CT, in a group of older

patients with hip fracture, observing a positive correlation between CT-determined midthigh muscle CSA and DXA-derived midthigh SMM. The assessment of MM and body composition by DXA-derived midthigh slice has been shown to be reasonably accurate in comparison to a single-slice CT technique in this sample of frail older patients [16].

BIS is another technique used to estimate SMM. The measurement is not a direct evaluation of MM, but an estimation on the whole-body electrical conductivity, through conversion equations [37]. BIS needs highly trained personnel, and the institutes where it can be performed are very difficult to find. Villani et al. [24] compared BIS associated to CAMA and DXA, noting BIS were reliable, but the difficulties in carrying out the examination and in the use of conversion equations led to DXA as the preferred reference technique. Muscle mass evaluation is not the only parameter that can be associated to sarcopenia. A low muscle quality is considered as one of the diagnosis criteria by EWGSOP [11]. Muscle quality is one of the main determinants of muscle function, depending on different factors (fibre composition, architecture, metabolism, intermuscular adipose tissue, fibrosis, motor unit activation) [39]. In particular, the decline of type II muscle fibres (II-MF) is responsible for muscle mass reduction [40].

Kramer et al. [22] performed vastus lateralis biopsies in different groups, confirming a significant II-MF atrophy in older women with hip fractures when compared to healthy older or young women. Since muscle atrophy is associated to loss of function, the author suggested that II-MF atrophy could lead to a higher risk of falls and consequent fractures. This study has some limits. There was no measure of strength and the sample was exclusively female, but the findings could be relevant to treat sarcopenia and to understand the II-MF atrophy causes. The histological diagnosis of sarcopenia could be a valuable way to understand physiopathology of sarcopenia in patients with hip fractures, even if it is not obviously suitable for routine diagnosis.

#### *4.3. Sarcopenia Treatment in Hip-Fractured Patients*

The treatment of sarcopenia in patients affected by hip fractures is a multidisciplinary challenge and, according to our findings, great attention should be given to nutritional status. Malnutrition is a highly prevalent condition in the geriatric population affected by this fracture [27]. Therefore, oral nutritional supplementation (ONS), in addition to rehabilitation programs, has become the subject of debate between different authors. Flodin et al. [25] investigated the effects of protein-rich supplementation and bisphosphonate on body composition, handgrip strength and quality of life in patients with hip fracture at 12-months follow-up. In a group, the combination of bisphosphonates and protein supplementation had no significant effects on handgrip strength (HGS), body composition and health-related quality of life (HRQoL). In another group, a positive effect of protein-rich supplementation and bisphosphonates on HGS and HRQoL was demonstrated.

Malafarina et al. [27] showed good results using oral nutritional supplementation with β-hydroxy-β-methylbutyrate (HMB). This approach improves MM, function and general nutritional status in hip-fractured patients [27]. HMB, a metabolite of leucine, has beneficial effect on MM and function in older people [41], but considering the lack of evidence focused on hip-fractured people, more investigations are needed in the treatment of sarcopenia with HMB in these patients. On the other hand, the role of a nutritional intervention without exercise for the treatment of sarcopenia is debated [41]. Although many studies have described good results in increasing protein intake in the older population [42,43], the timing and distribution is unclear [44].

#### *4.4. Sarcopenia Prevention in Hip-Fractured Patients*

Despite the few studies focused on sarcopenia prevention in our study, it could be considered the major area of research for future clinical activity and observational epidemiological trials [39] in order to identify and modify the sarcopenia risk factors. A midlife lifestyle approach could be more proper to limit the sarcopenia incidence [45].

Physical activity programs have been suggested as a relevant technique in reducing the risk of hip fracture in older patients [46,47]. In the study by Piastra et al. [47], data showed a significant improvement in MM, muscle mass index, and handgrip strength in muscle reinforcement training group, demonstrating that a muscle reinforcement program moved participants from a condition of moderate sarcopenia at baseline to a condition of normality. Ding-Cheng Chan et al. [28] evaluated effects of programs in community-dwelling older adults with high risk of fractures (> or =3% for hip fracture). The exercise authors clarified the lack of differences in the types of exercise to improve sarcopenia when compared an integrated care model to a lower extremity exercise model. However, several authors promoted rehabilitation protocols for hip-fractured patients, consisting of oral nutritional supplementation with proteins and amino acids and exercise programs [46,47]. Singh et al. [47] proposed a new rehabilitation protocol in the older with hip fracture after orthopaedic surgery. The 12-month rehabilitation program was characterized by a high-intensity progressive resistance training and a targeted treatment of balance, osteoporosis, nutrition, vitamin D and calcium, depression, home safety and social support. The authors showed a statistically significant reduction in mortality, nursing home hospitalization and disability, especially in those subjects with a systematic good health status.

A life course approach to prevention is paramount and offers chance to intervention when lifestyle changes, inspiring the increase of physical activity with immediate to lifelong advantages for skeletal muscle health [16].

#### *4.5. Limits of the Study*

The limits of the study are represented by the heterogenicity of the definition of sarcopenia and by the tools considered to assess the patient functional outcome. We extensively searched and identified all relevant last 20 years sarcopenia diagnosis-, treatment- and prevention-related articles. Therefore, risk of bias assessment showed moderate overall risk, which could influence our analysis. Moreover, in the diagnosis section, only instrumental tool evaluations without clinical assessment were detected.

#### **5. Conclusions**

Sarcopenia is a physiological condition and contributes to the increased risk of falls and hip fractures in the older population. However, the diagnosis of sarcopenia is challenging, especially in hip-fractured patients, and there are currently no standardised diagnostic and therapeutic protocols. The development of medical management programs is mandatory for good prevention. To ensure adequate resource provision, care models should be reviewed, and new welfare policies should be adopted in the future.

**Author Contributions:** Conceptualization, G.T., A.V. and V.P. (Vito Pavone); methodology, A.V.; software, M.A.; validation, G.T., V.P. (Vincenzo Petrantoni) and G.I.R.; formal analysis, A.V.; investigation, M.A., V.P. (Vincenzo Petrantoni) and D.Z.; resources, A.V.; data curation, G.T.; writing—original draft preparation, D.Z.; writing—review and editing, G.T. and A.V.; visualization, G.I.R.; supervision, G.S.; project administration, V.P. (Vito Pavone); funding acquisition, V.P. (Vito Pavone). All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest. The author GT declares to be the Guest Editor of the Special Issue "Prevention and Treatment of Sarcopenia" of Journal of Clinical Medicine.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
