Fat percentage cutoff values to define obesity and prevalence of sarcopenic obesity in community-dwelling older adults in Turkey

Abstract

Purpose

Sarcopenic-obesity (SO) is associated with low-functional-status and mortality. Few studies evaluated the definition and prevalence of SO. We aimed to investigate the fat-percentage cut-off values for obesity and prevalences of obesity, SO in community-dwelling older adults in Turkey.

Methods

Body-composition was measured using bioimpedance-analysis. Sarcopenia was defined by European-Working-Group-on-Sarcopenia-in-Older-People criteria. Obesity was defined by two different methods, a fat-percentile above 60th percentile (Zoico-method) or a BMI of ≥30 kg/m² (WHO-definition).

Results

We enrolled 992 subjects (308 men, 684 women). Body fat-percentage thresholds for obesity were 27.3% for men and 40.7% for women according to Zoico-method. The rates of obesity were about 40% in both genders by Zoico-method; 29.2% versus 53.7% for men and women by WHO definition. Prevalences-of-sarcopenia was 3.1% versus 0.4%; SO was 0.3% versus 0.1% when obesity was assessed with Zoico-method in men and women, respectively. No case of SO was defined when obesity was assessed using WHO-definition.

Conclusion

The threshold for obesity definition according to Zoico-method was similar to other European-populations. While obesity-prevalences were considerably high, SO prevalences were low but comparable to other populations. This low-prevalence seems to be due to underestimation of sarcopenia in obese subjects when skeletal-muscle-mass was adjusted by height² to recognize low-muscle-mass.

Keywords:

• Sarcopenia
• obesity
• sarcopenic obesity
• fat
• threshold
• cut-off

Introduction

Obesity is highly prevalent in older adults and a rapid increase in its prevalence is expected. Some studies (observational and interventional, i.e. weight loss studies) suggest negative impact of obesity on functional status in old age [1–6]. Sarcopenia is another prevalent concept in the older adult related to body composition and is known to be associated with functional impairment. However, there is not a global consensus on the definitions of obesity or sarcopenia.

In the literature, obesity is defined in several ways. The most commonly used is the World Health Organization (WHO) definition. WHO defines obesity as a body mass index greater than 30 kg/m² [7]. Recently, Zoico et al. defined obesity independent from the body mass index but according to the body fat percentage [8]. They considered obesity as a fat percentile above 60^th percentile of the study population [8]. Sarcopenia is defined as the presence of low muscle mass (LMM) and function (i.e. low muscle strength and/or gait speed) [9]. Low muscle mass is considered frequently as a skeletal muscle mass index [skeletal muscle mass (kg)/height² (m²)] lower than two standard deviations of young healthy adults and the normative population specific young reference population data are suggested to evaluate its cutoff values [9]. Low muscle mass lower than one standard deviation of young healthy adults is regarded as Class I LMM, and lower than two standard deviations of young adults is regarded as Class II LMM [8]. To assess sarcopenic obesity (SO), Zoico et al. were the first that put forward obesity definition by the fat percentile above 60th percentile of the study population and in their study, they regarded sarcopenia as the presence of class I or class II LMM (kg/m²) [8]. Consequently, Rolland et al. also considered obesity similarly and evaluated SO with the Zoico’s obesity definition and sarcopenia component as the presence of class II LMM (kg/m²) [10]. There is very limited data that includes sarcopenia component of SO definition with the co-presence of LMM and low muscle function [11] albeit this approach is the current globally recommended definition [9,12–14].

The combination of sarcopenia and obesity is expected to cause more problem, i.e. the functional impairment than that is caused only by one [15]. Muscle and fat mass are strongly interconnected from a pathogenetic point of view [16,17] which is also valid for bone and related osteoporosis [18–21]. Sarcopenic obesity has drawn attention in recent years with accumulating data on its relation with poor functional status and survival [22–24]. Despite the growing importance of SO, only few studies evaluated the definition and prevalence of SO. In this study, we aimed to determine the obesity threshold by the Zoico method and examine the prevalences of obesity and SO using different suggested definitions of obesity i.e with the Zoico definition and the WHO definition among the Turkish community-dwelling older adults aged ≥60 years old admitted to geriatric outpatient clinic of our university hospital.

Materials and methods

We enrolled the study participants at the geriatric outpatient clinic of a university hospital between November 2012 and March 2016 retrospectively. The outpatient clinic accepts all older adults that are ≥60 years old. Among 2550 community-dwelling geriatric outpatients medically assessed, 1297 patients of them were not offered participation because of staffing issues, 150 patients of them had poor medical condition for acute problems, and 111 patients of them did not give informed consent. The acute problems for which the 150 patients were excluded from body composition analyses were acute infections, e.g. pneumonia, urinary system infections; acute execerbations of chronic diseases, e.g. chronic obstructive pulmonary disease, renal failure, heart failure, etc. As a result, the final study population was composed of 992 older adults.

Body mass index was calculated as body weight divided by height squared (weight/height²). Weight was measured by kilogram (kg) and height was measured by meter (m), and the unit for BMI was kg/m². Body composition was estimated using bioimpedance analysis (BIA) (Tanita-BC532-Japan). Fat free mass (FFM) was measured by BIA. Skeletal muscle mass (SMM) was calculated using the formula: SMM (kg) = 0.566 * FFM (kg) which had been validated against the SMM data calculated by use of 24-h creatinine excretion [25]. Skeletal muscle mass index was calculated by SMM/height² [26]. Low muscle mass (defined as Class II low muscle mass) was defined by our national data, i.e. ≤9.2 kg/m² versus 7.4 kg/m² [27]. To be able to make comparisons with the previously published studies, the Class I low muscle mass threshold was defined as mean-1 standard deviation of the young reference population and calculated from our previous publication as 10.1 kg/m² in males and 8.2 kg/m² in females [27].

Muscle strength was evaluated by handgrip strength (HGS) using a Jamar hydraulic hand-dynamometer. The participants were asked to exert maximum grip strength with bilateral hands with resting intervals of at least 30 s. The maximum grip strength was noted as the test outcome. Cut-off points were regarded as ≤22 kg in females and ≤32 kg in males according to our national data [27]. Physical performance was evaluated using 4-meter usual gait speed (UGS) at usual pace. UGS was noted impaired if it was below 0.8 m/s.

Sarcopenia was defined according to European Working Group on Sarcopenia in Older People (EWGSOP) criteria as presence of low muscle mass (LMM) and function (usual gait speed and/or muscle strength) [9]. Obesity was defined by two different methods, a fat percentile above 60th percentile (the Zoico method) or a BMI of ≥30 kg/m² (the WHO definition). Sarcopenic obesity (SO) was defined by co-presence of sarcopenia [9] and obesity either by WHO definition [7] or Zoico definition [8]. We performed the study in accordance with the guidelines of Declaration of Helsinki. Informed consents were received by all participants. This study was approved by the Istanbul University Istanbul Medical Faculty Ethics Committee.

Statistical analysis

This study is a descriptive study. The continuous variables were evaluated for normal distribution. Numerical variables were given as mean ± standard deviation or median as appropriate. The 60th percentile value is assessed for body fat percentage. Categorical variables were represented as frequencies. SPSS (statistical package for social sciences) for Windows 21.0 (SPSS Inc., Chicago, IL) was used for statistical analysis.

Results

We enrolled 992 subjects (308 men and 684 women) with a mean age of 75.2 ± 7.2 years. The demographic, body composition and muscle strength data of the subjects are listed in Table 1 according to gender. We calculated the 60th percentile of the fat ratio in different genders to determine the obesity threshold using Zoico method. Body fat% thresholds for obesity were 27.3% for men and 40.7% for women.

Table 1. Age, muscle strength, and body composition data of the subjects by gender.

	Male (n = 303)	Female (n = 667)	Total (n = 970)
Age	76.3 ± 6.9	74.3 ± 7.2	75.2 ± 7.2
BMI (kg/m^2)	27.7 ± 4.3	30.7 ± 5.6	29.7 ± 5.4
Fat (%)	26.4 ± 6.3	38.2 ± 7.1	34.5 ± 8.8
SMMI (kg/m^2)	10.9 ± 1.1	10.1 ± 1.2	10.3 ± 1.2

BMI: body mass index; SMMI: skeletal muscle mass index; HGS: hand grip strength.

The rates of obesity were similar and about 40% in both genders (i.e. 40.3% vs 40.5%) by Zoico definition but much more prevalent in females, i.e. 29.2% versus 53.7% for men and women by WHO definition, respectively (Table 2). The comparison of the body fat% thresholds for obesity according to Zoico method (fat percentile above 60th percentile) and obesity prevalences in our study versus different populations published [8,28] so far are given in Table 2.

Table 2. The body fat% thresholds for obesity according to Zoico method (fat percentile above 60th percentile) and obesity prevalences in our study and comparison with the different populations.

	Turkey^a	Italy		Korea (> =60)		France
	(BIA)	(DXA)^b		(DXA)^c		(DXA)^d
	Male	Female	Female	Male	Female	Female
	n = 308	N = 684	n = 167	n = 198	n = 328	n = 1308
Obesity (fat%) threshold (Zoico definition)	27.3%	40.7%	42.9%	20.2%	31.7%	40%
Obesity prevalence (Zoico)	40.3% (124)	40.5% (277)	39.6%	20.3%	16.5%	36%
Obesity prevalence (WHO)	29.2% (92)	53.7% (380)	15.8%

^aBody composition was assessed by BIA.

^b Body composition was assessed by DXA. The study included only the female participants [8].

^c Body composition was assessed by DXA [28].

^d Body composition was assessed by DXA [10].

^e The prevalence of obesity by Zoico method was analyzed by the figures given in the corresponding articles.

The measurements of muscle strength and usual gait speed were performed in 294 males and 668 females. The prevalence of sarcopenia noted by the EWGSOP definition was 3.1% in men and 0.4% in women and 1.2% in total study population. The rate of SO was low as 0.3% and 0.1% in men and women, respectively when obesity was assessed with Zoico method while no case of SO was defined in the whole population when obesity was assessed using WHO definition. On the other hand, in our study, the presence of Class I low muscle mass and obesity by Zoico definition was not very low, i.e. 5.5% in the male participants (Table 3). Comparison of the prevalence of low muscle mass, EWGSOP defined sarcopenia and sarcopenic obesity in our study from Turkey and the studies from Italy [8] and France [10] is shown in Table 3.

Table 3. The prevalence of low muscle mass, EWGSOP defined sarcopenia and sarcopenic obesity in our study and comparison with the different populations.

	Turkey^a (BIA)		Italy^b (DXA)	France (DXA)^c
	Male	Female	Female	Female
	n = 308	n = 684	n = 167	n = 1308
Class I low muscle mass (n = 308 vs n = 684)	22.4% (69)	3.1% (21)	27.2%
Class II low muscle mass (n = 308 vs n = 684)	5.2% (16)	0.6% (4)		9.6% (126)
Sarcopenia (EWGSOP) (n = 294 vs n = 668)	3.1% (9)	0.4% (3)
SO (Zoico)^d (Class I LMM + Obesity)	5.5% (17)	0.3% (2)	12.4%
SO (Zoico)^e (Class II LMM + Obesity)	1% (3)	0.1% (1)		2.8%
SO (Zoico)^f (EWGSOP Sarcopenia + Obesity) (n = 294 vs n = 668)	0.3% (1)	0.1% (1)
SO (WHO)^g (n = 294 vs n = 668)	0	0

The data are given as percent (numbers).

SO: sarcopenic obesity; LMM: low muscle mass; WHO: World Health Organization.

Class I and Class II low muscle mass were defined as skeletal muscle mass index lower than 1 and 2 standard deviations of the young reference population.

^aLMM was noted by SMMI adjusted for height2. [SMMI= skeletal muscle mass (kg)/height (m2’)] [26]. Sarcopenia was defined by the EWGSOP criterion. Body composition was assessed by BIA estimating the total skeletal muscle mass.

^bThe authors regarded sarcopenia as the presence of LMM. Body composition was assessed by DXA. They used both appendicular SMMI adjusted for height2 and total SMMI adjusted for total body weight [total skeletal muscle mass*100 (kg) /body weight (kg)]. They defined presence of sarcopenia either as the presence of LMM when SMMI was adjusted by height2 or SMMI adjusted by total body weight and gave a composite figure. They only noted the SO data when sarcopenia was defined as the presence of Class I LMM. It included only the female subjects [8].

^cThe authors regarded sarcopenia as the presence of LMM. Body composition was assessed by DXA. They used appendicular SMMI adjusted for height2. They only noted the SO data when sarcopenia was defined as the presence of Class II LMM. It included only the female subjects [10].

^dSO was defined as presence of Class I LMM plus obesity (Zoico definition).

^eSO was defined as presence of Class II LMM plus obesity (Zoico definition).

^fSO was defined as sarcopenia (EWGSOP definition) plus obesity (Zoico definition).

^gSO was defined as sarcopenia (EWGSOP) plus obesity (WHO definition).

Discussion

In this study, we defined obesity cutoff threshold according to Zoico definition as 27.3% in males and 40.7% in females. Physiologically, females have higher percentage of fat in their body [29]. Hence, their fat percentage cutoff threshold to define obesity is logical to be higher than that of the males. Our fat percentage cutoff thresholds were in accordance with the published cutoff studies from Europe, i.e. Italy [8] and France [10] but higher than that of an Asian country, Korea [28]. These results are in line with the suggestions that obesity cutoff thresholds are lower in the Asian population than the European counterparts [30–32].

To date, the most widely accepted and adapted methodology to define obesity is the WHO definition [7]. It is a simple measure that can be calculated in anywhere with measures of weight and height and can be replicated very easily. While these are the advantages of WHO method, it does not take gender differences to define obesity into account. WHO definition is a rough measure to estimate excess body fat, whereas the body composition allows us to detect the real fat percentage. Our results indicate that the obesity prevalence is comparable between the genders by Zoico definition but much more common in the females by the WHO definition. Many researches to date have been performed due to early release and ease of application of WHO definition. While many studies in the younger age group noted associations of WHO-defined obesity with worse functionality, in the older adults it gave conflicting results [6]. While some studies [1–6] suggested the high BMI as a risk factor for low functional status, some studies [1,2,5,6,33] suggested conversely the low BMI as a risk factor for functional status. Of course, the more the extreme of BMI (may be higher or lower), the risk for functional impairment is the higher [1]. These results are conflicting to the hypothesis that the excess fat would compromise the functional status. We suggest that the reported association of better functional status with obesity maybe due to these limitations of WHO definition of obesity. The future researches are needed to comment whether Zoico definition of obesity is associated with poor functional status or would give conflicting results.

Our results are important to make a comparison between countries regarding the prevalence of obesity. The prevalence of obesity by the Zoico method were comparable between Turkey and Italy (about 40%) [8], which were higher than in Korea (15–20%) [28]. Regarding the WHO definition, in this study, among older adults ≥60 years old, our obesity prevalence was 47.5%, being more common in females. In the US, prevalence of obesity is reported about 35% by WHO definition in older adults aged 65 and over with a similar prevalence in both genders [34]. In this study, the prevalence of obesity by educational attainment was also examined. No linear trends in obesity prevalence by educational attainment were observed in men or women aged 75 and over. In the age group between 65–74 years, among males, the prevalence of obesity was higher among those with some college compared with those with a college degree and among females, and there was a decrease in obesity prevalence with increasing education [34]. Very recently, the prevalence of obesity from Europe by WHO definition including data from 10 European countries among individuals aged ≽50 years was reported 15–20% [35]. The prevalence of obesity was about 15% in older males and 20% in older females in the European study [35]. This figure is similar to the Italian study reported by Zoico that reported 15.8% obesity prevalence by WHO definition among 67–78 years healthy older females [8]. So, this study provides insight for very different prevalences of obesity in between the countries which may direct researchers to find the underlying reasons and consequently possible solutions with providing comparison between the countries.

We reported the very low prevalence of SO when sarcopenia was defined by the EWGSOP methodology. The prevalence of SO was about 0.2% with similar figures in between genders when the obesity was defined by Zoico methodology. Strikingly, there was no case of SO when obesity was defined with the WHO methodology. To our knowledge, there is only one study reported so far that denotes the SO prevalence by the sarcopenia component evaluated by the EWGSOP criterion [11]. They noted a 0.1% of SO by sarcopenia (EWGSOP definition) and obesity (WHO definition) among a total of 965 older (≥70 years) males from Germany [11]. Consequently, they suggested that with respect to obesity, it was evident that BMI-based approaches were absurd when addressing sarcopenic obesity. In this study, they also evaluated obesity by different thresholds of body fat percentages and declared the prevalence of SO as 2.9% with sarcopenia (EWGSOP definition) plus obesity (fat percentage >30%). While SO figures are also low in Germany, it was somewhat higher than our study population. Of note, it is reported that the prevalence of obesity among older adults increases most at Germany [35] among the European countries.

Despite the considerably high prevalence of obesity in this study, the very low prevalence of SO seems to be due to the low prevalence of sarcopenia. The prevalence of sarcopenia was 1.2% in our study population (3.1% in males vs 0.4% in females). Furthermore, in this study, skeletal muscle mass was adjusted by height². While it is the method most widely adapted in current definitions of sarcopenia [9,12–14], this method of adjustment is well-known to fail for identifying most obese people with sarcopenia. Newman et al. denoted that the skeletal muscle mass/height² index is highly correlated with body mass index (BMI) [36]. Thus, this index primarily identifies thin people as sarcopenic; hence, it could have limited applications as it would underestimate sarcopenia in overweight or obese people [37]. Recognizing these considerations, Janssen et al. [38] proposed a definition of skeletal muscle mass index i.e the skeletal muscle mass adjusted for weight (skeletal muscle mass (kg) * 100/weight (kg)). Also, recently, FNIH definition of sarcopenia has been put forward [39] that included skeletal muscle mass adjustment by body mass index [40]. These approaches would apparently identify more obese subjects as having low muscle mass and sarcopenia. Future studies on the prevalence of SO with the other adjustment indices of skeletal muscle mass and their relative associations with functional status are needed and awaited.

Some limitations of our study should be considered. We have analyzed data from the older adults admitting to a geriatric outpatient clinic of a university hospital. Therefore, these figures do not apply as the obesity and SO prevalences of the community-dwelling older adults in this region. However, our outpatient clinic accepts all older adults ≥60 years from the general population. They are not necessarily only the fragile population but include fit older adults that apply for preventive services and routine control of stable diseases as hypertension and diabetes. It would be better to take representative samples from the general community-dwelling older adults. However, our study provides figures for the outpatient clinic patients that whom we serve in medical practice routinely. Second, we assessed body composition by BIA. Although the assessment by BIA is suggested reliable in both clinical practice and research settings [9], it is a result of an estimation, not an exact measurement. Moreover, due to easy application with lower cost efficiency and lack of X-ray exposure, BIA technique may be a better choice to determine body composition in large cohorts.

In conclusion, we defined obesity cut off thresholds of fat percentage according to Zoico definition of obesity and prevalences of SO with the most adapted sarcopenia definition (EWGSOP criterion) in community-dwelling older adults aged 60 years and older that admitted to our geriatric outpatient clinic. We have used and reported SO with two different obesity definitions. Our results indicated that while obesity prevalences were considerably high, SO prevalences were low with either methodology. This low prevalence seems to be due to underestimation of sarcopenia in obese subjects when skeletal muscle mass was adjusted by height² to recognize low muscle mass.

Disclosure statement

No potential conflict of interest was reported by the authors.