J Community Health DOI 10.1007/s10900-014-9881-3
ORIGINAL PAPER
The Performance of Obesity Screening Tools Among Young Thai Adults Panita Limpawattana • Thepkhachi Kengkijkosol Prasert Assantachai • Orapitchaya Krairit • Jiraporn Pimporm
•
Ó Springer Science+Business Media New York 2014
Abstract Obesity is a worldwide medical condition that leads to physical and psychological impairment. Specific ethnicity, gender and age group are related to different performances of anthropometric indices to predict obesity. The objectives of this study were to estimate the performance of the anthropometric indices for detecting obesity based on percentage of body fat (PBF), to study the correlation among those indices, and to determine the optimal cut-off point of the indices among young Thai adults. This is a cross-sectional study of healthy urban subjects in Khon Kaen, Thailand who were aged 20–39 years. Baseline characteristics and anthropometric measures were collected. PBF was determined using bioelectrical impedance analysis. Demographic data and anthropometric variables were analyzed using descriptive statistics. Receiver-
operating characteristic (ROC) curves were used to compare the performance of anthropometric measures as predictors of obesity. One-hundred men and 100 women were recruited for this study. Body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR) and waistto-stature ratio (WSR) were significantly correlated to PBF. BMI demonstrated the best performance according to the area under the ROC curves in both sexes at cut-off points of 22.5 in women or 25 kg/m2 in men. WC and WSR showed better performance than WHR to detect obesity. In conclusion, anthropometric indices in young Thai adults were correlated well with PBF to predict obesity as shown in prior reports. Different cut-off points of these indices to define obesity in young Thai adults are recommended. The global cut-off points of WSR in women regardless of ethnicity are supported.
A1 A2 A3 A4 A5
P. Limpawattana (&) Division of Geriatric Medicine, Department of Internal Medicine, Faculty of Medicine, Khon Kaen University, Khon Kaen 40002, Thailand e-mail: [email protected]
Keywords Body mass index Waist circumference Waist-to-hip ratio Waist-to-statue ratio Percentage body fat
A6 A7 A8
T. Kengkijkosol Department of Internal Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
Introduction
A9 A10 A11 A12
P. Assantachai Department of Preventive and Social Medicine, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand
A13 A14 A15 A16
O. Krairit Divison of Geriatric Medicine, Department of Internal Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
A17 A18 A19
J. Pimporm Outpatient Clinic, Srinagarind Hospital, Faculty of Medicine, Khon Kaen University, Khon Kaen, Thailand
Obesity is a medical condition in which excessive adipose tissue accumulates in the body leading to an impairment of both physical and psychological health and well-being [1]. It has been proven as a strong risk factor for cardiovascular disease [1, 2]. Obesity is a global public health problem. An estimated 315 million people worldwide are obese and at least 2.8 million die annually as a result of being overweight or obesity [1, 3]. In Western and westernized countries, the prevalence of obesity is about 20 % in men and 25 % in women [1]. For Asian countries, its prevalence is lower but it has tended to increase considerably. Women
123
J Community Health
are more likely to be obese than men and in the WHO regions for Africa, Eastern Mediterranean and South East Asia, women had approximately twice the obesity prevalence of men [3]. A measurement of percentage body fat (PBF) is considered the reference method for defining obesity in which PBF [25 % for men or 30 % and for women, were cut-off points to diagnose obesity since these points were associated with health-related adverse outcomes of obesity [2, 4]. Dual energy X-ray absorptiometry (DXA) is a frequently used tool to measure PBF. The limitations of its use are the unavailability in general settings and being an expensive tool particularly in developing countries [2, 4]. Bioimpedance analysis (BIA) is another technique to measure body composition including PBF. It was developed to estimate the volume of body fat and lean body mass. It is portable, easy to use and has a relatively lower cost than DXA [2, 4]. The accuracy of BIA is interfered with by some factors such as dehydration, strenuous exercise, pregnancy and having a heavy meal; controlling these factors is generally regarded as being essential to making this method an accurate instrument to measure body composition [4]. Anthropometric indices including body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR) and waist-to-height ratio (WHR) have been studied widely as the commonly used tools for assessing obesity due to the simplicity, low cost, short time consuming and their strong correlations with PBF, adverse metabolic profiles such as diabetes and metabolic syndrome, and cardiovascular disease [2, 5–11]. There is strong evidence that Asians should have lower cut-off values than Europeans for BMI, WC and WHR [6]. For WSR, a systematic review revealed that a cut-off point at 0.5 was a universal value regardless of ethnicity [12]. The cut-off points of BMI, WC, WHR and WSR in Asian studies; however, remained diverse across countries [4, 8]. Additionally, they were sex and age group specific in some reports [2, 13]. For example, obesity was defined as BMI [27 in men or [25 kg/m2 in women for Thais aged 20–84 years and [25 kg/m2 in both sexes for Singaporean persons [4]. For Malaysian adults, WC cut-offs of 81 cm. in men and 80 cm. in women were associated with abdominal obesity [13] while in China, the optimal values of WC were 78.9 cm in men and 65.8 cm in women for adults aged 20–30 years, and 82.4 cm in men and 71.4 cm in women for adults aged 31–45 years [2]. The studies in the Thai population for specific age groups are lacking. Therefore, the primary objective of this study was to estimate the performance of the anthropometric indices for detecting obesity. The second objective was to study the correlation among those indices and to determine the optimal cut-off points of the indices among young Thai adults.
123
Materials and Methods Study Subjects and Setting This was a cross-sectional study. The setting was the urban area of Khon Kaen province, Thailand which is the secondlargest Northeastern province. It is located 445 km away from Bangkok with a population of 1.8 million [14]. The potential subjects were sent a letter of invitation asking if they would like to participate in the study. Subjects were a sample of young healthy Thai adults who aged 20–39 years. The exclusion criteria were the persons who could not stand upright, had a pacemaker, used drugs, herbs or any hormones that affect muscle mass and strength such as contraceptive drugs, steroid hormones and thyroid hormone, drank alcohol 12 h prior to the test, had vigorous exercise within 12 h before analysis, were dehydrated, and women during the menstrual period or pregnancy. Data Collection Baseline data were collected and anthropometric measures were taken while subjects were lightly clothed and wore no shoes. WCs were taken midway between the inferior margin of the last rib and the iliac crest at the end of expiration and hip circumference (HC) was measured at the widest part of hip. Measurements of standing height were performed without shoes with a stadiometer. Body weight and percent body fat (PBF) estimation were determined using the Tanita bioelectrical impedance analysis (BIA) system (model BC-418). BMI was calculated as weight (kg) divided by height squared (m2). The waist-to-hip ratio (WHR) and waist-to-stature ratio (WSR) were then calculated. Obesity was defined by PBF [30 % in women and [25 % in men [2]. Procedure Eligible subjects had no food or drink 4 h prior the test and written informed consent was obtained from all subjects. Baseline data and anthropometric indices were collected by trained nurses and then body composition was measured by the body composition monitor with a scale. Sample Size Calculation Sample size calculations were based on the areas under the receiver operating characteristic curves (ROC) and areas under the curves (AUC) according to the methodology of Hanley and McNeil (1983) [15]. ROC curves were used to summarize the accuracy of diagnostic tests. This method varies the sample size until a sufficiently small standard error (SE) of the area under the ROC curve was achieved.
J Community Health Table 1 Baseline characteristics of subjects by sex Variables
Women (N = 100)
Men (N = 100)
Age (years); med (IQR1, IQR3)
25 (29,34.5)
29 (24,34)
Body weight (kg); med (IQR1, IQR3)
55.05 (48.55,63.95)
65.6 (59,73.25)
Height (m); mean (SD) BMI; med (IQR1, IQR3)
157.49 (5.91) 22.05 (19.75,26)
169.77 (7.09) 23.05 (20.30,25.45)
WC (cm); med (IQR1, IQR3)
78 (70.5,86)
81 (75.5,88)
HC (cm); med (IQR1, IQR3)
93 (89,102)
93 (88,99)
WHR; med (IQR1, IQR3)
0.81 (0.78,0.88)
0.88 (0.82,0.93)
WSR; med (IQR1, IQR3)
0.50 (0.44,0.55)
0.48 (0.44,0.52)
PBF (%); med (IQR1, IQR3)
32 (26.9,38.05)
19.8 (14.75,24.25)
Med median, IQR inter-quartile range, bmi body mass index (kg/m2), WC waist circumference (cm), HC hip circumference (cm), WHR waist-to-hip ratio, WSR; waist-to-stature ratio, PBF percentage body fat (%)
Because of the complexity of the formula, a web-based calculator (www.anaesthetist.com/mnm/stats/roc/#stderr) was used to determine the standard error. Finally, a sample size of 200 participants of men and women was found to be adequate and feasible to conduct the trial in clinical practice at the AUC of 0.9 and S.E. of 0.04.
Statistical Analyses Demographic data and anthropometric variables were analyzed using descriptive statistics, presentations in percentage, mean and standard deviations. If the distribution of these data was not a normal distribution, then medians, and inter-quartile ranges were used instead. Spearman’s rank correlation coefficients were calculated to quantify associations between anthropometric measures (BMI, WC, WHR and WSR) and PBF to predict obesity. The ROC curve was used to summarize the overall accuracy of the anthropometric indices for obesity detection. Then optimal cut-off points were determined. The performances of the tests were summarized by the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and likelihood ratios. All of the data analyses were performed by using STATA version 10.0 (StataCorp, College Station, TX, USA). Ethics approval was provided by the ethics committee of Medicine Faculty, Khon Kaen University according to the Helsinki Declaration.
Fig. 1 Spearman rank correlation (rho) between percent body fat and other parameters by sex BMI body mass index (kg/m2), WC waist circumference (cm), WHR waist-to-hip ratio, WSR waist-to-stature ratio, PBF percentage of fat (%)
Results There were 100 men and 100 women recruited in this study. Basic anthropometric indices are shown in Table 1. It was found that the median/average values of age, body weight, height, BMI, WC, HC and WHR in men were greater than those found in women while WSR and PBF were lower in men than in women. Simple correlations between PBF and anthropometric indices in both sexes are presented in Fig. 1. All the correlations were statistically significant (p \ 0.05). In women, all indices except WHR (q \ 0.5) had good correlations whereas in men, a slightly lower but still significant correlation was detected. WHR remained the lowest correlation (q 0.37) in men as in women. Obesity, when defined based on the PBF as [30 % in women or [25 % in men [7]; the majority of women (60 %) and minority of men (21 %) were obese. The performances of the anthropometric indices including the cutoff points for all variables corresponding to the criterion values with the best tradeoffs are shown in Table 2. The ROC curves for obesity diagnosis among anthropometric indices in women and men are demonstrated in Figs. 2 and 3. According to the curves, BMI was the most predictable for obesity in both sexes and WHR was the worst one. In particular, BMI in women and WSR in men showed obvious good performances for obesity detection, given very high likelihood ratios.
Discussion The anthropometric indices of young Thai adults were quite similar to the data of the young Chinese population. In addition, the PBF in men was much lower than the one for women [2]. Although the Chinese study used DXA to measure body composition, the BIA technique used here,
123
J Community Health Table 2 Performance of obesity parameters by sex Indices
BMI
WC
WHR
WSR
Woman
Man
Woman
Man
Woman
Man
Woman
Man
AUC 95 % (CI)
0.96 (0.94, 0.99)
0.93 (0.87, 0.90)
0.91 (0.85, 0.97)
0.91 (0.81, 1.00)
0.72 (0.62, 0.83)
0.75 (0.63, 0.86)
0.90 (0.84, 0.96)
0.91 (0.81, 1.00)
Cut-off point
22.5
25
75
89
0.776
0.75
0.503
0.533
Sensitivity Specificity
80 % 97.5 %
90.48 % 89.97 %
85.00 % 77.5 %
76.19 % 92.41 %
90 % 97.5 %
76.1 % 68.35 %
71.67 % 85 %
71.43 % 98.73 %
PPV
98 %
70.37 %
85 %
76.19 %
76.47 %
57.14 %
87.76 %
93.75 %
NPV
76.47 %
97.26 %
77.5 %
93.6 %
86.67 %
93.01 %
66.67 %
92.86 %
LR?
32
8.93
3.78
10.02
1.68
2.26
4.78
56.43
LR-
0.31
0.11
0.19
2.26
0.25
0.42
0.33
0.29
2
Bmi body mass index (kg/m ), WC waist circumference (cm), WHR waist-to-hip ratio, WSR waist-to-stature ratio, AUC area under receiving operating characteristic curve, CI confidence interval, PPV positive predictive values, NPV negative predictive values, LR? likelihood ratio positive, LR- likelihood ratio negative
Fig. 2 ROC curves of anthropometric indices to predict obesity using percent body fat as a gold standard in women
after controlling interfering factors, the results showed similar patterns supporting the prior studies [2, 16–18]. Comparing the PBF in the same age group of the current study to one report in Europe where the average PBF ranged from 24.9 to 26.1 % in women and 21.7 to 24.5 % in men, a higher PBF in Thai women and lower PBF in Thai men, confirmed the differences between ethnicities, body size, life styles, and cultural backgrounds that might influence the diversity of body composition [4, 18]. Therefore, different cut-off points for anthropometric indices should be identified. All indices correlated significantly with the PBF with a greater degree (0.67–0.94) than the preceding study of Chinese populations (0.57–0.75) for both sexes except in the area of WHR that showed lower correlations (0.37–0.42 vs. 0.47–0.65) [2]. As high PBF is correlated with adverse health outcomes [7]. and existing data showed these
123
Fig. 3 ROC curves of anthropometric indices to predict obesity using percent body fat as a gold standard in men
anthropometric indexes all correlated to cardiovascular risk factors, the results from this study were comparable to these studies [5, 6, 9, 11, 19, 20]. The current results determined the different cut-off points of BMI, WC, WHR and WSR for detection of obesity among young adults. The overall performances of screening tools according to the AUC curves were good and correlated with existing data [8–11, 13, 20]. The optimal cut-off points of BMI, WC and WHR were lower than European countries, consistent with many Asian studies though different cut-off points were reported [8–11, 21]. Even for prior Thai studies e.g. the National Thai population that recommends a BMI [23 kg/m2 of both sexes, or[25 kg/m2 in women and[27 kg/m2 in men aged 20–84 years as obesity [4], the differences can be explained by the study populations. In this study that included only young Thai adults the outcomes using the
J Community Health
National study would indicate obesity. One study in Malaysia recommended the cut-off points of WC in women \80 cm and men \81 cm which were lower than most Asian studies for cut-off points in men which the authors explained these differences from the higher prevalence of cardiovascular disease in Malaysian populations than other Asian countries [13]. BMI showed the best performance in this study based on the AUC of the ROC curves. This result implies that total body fat (BMI) is likely to play a more major role than abdominal fat in predicting obesity in young Thai adults since BMI is considered an indicator of body fatness [1]. The optimal cut-off points of the WSR in this study showed similar results for women and a slightly higher result for men than existing reports [5, 12], supporting the evidence that the cut-off points of 0.5 could be a suitable global boundary values independent of ethnicities [12]. The performance of WSR in this study was comparable to WC and slightly lower than BMI to predict obesity based on the AUC of the ROC curves though some studies showed that WSR was a better screening tool than WC and BMI [9, 22]. Interpretation must be careful because those studies’ outcomes were for cardiovascular disease risk. The usefulness of WHR to detect obesity in this study showed the lowest performance. As compared to the previous Thai study in adults of an average age of 41.9 ± 10.6 years, the performances were similar and are in contrast to the report of Taiwanese that showed WHR had the higher adjusted odds ratios than BMI to predict the risk of type 2 diabetes [11, 20]. This can be explained by the limitation of WHR that reflect differences of body proportion in different age groups and ethnicities. For HC, there are numerous reports regarding the inverse effects on metabolic profiles including reduced cardiovascular risks and mortality after adjustment for WC, possibly due to measurement of HC alone represents only subcutaneous fat, not including visceral fat since visceral fat is correlated to increased risks of cardiovascular disease [23, 24]. Therefore, analyzing the performance of HC alone in these cross-sectional data was inappropriate. Given the dramatically increased obesity prevalence in adolescents and young adults [25] in addition to the adverse health outcomes of obesity, early screening with different cut-off points in different age groups and sexspecific criteria are recommended. One study in China also demonstrated various cut-off points for each age group provided higher performances than single age criteria [2]. Although body fat is closely related to obesity and other unfavorable outcomes, measurement by DXA or even BIA in the general population especially in the Southeast Asia region is impractical. Early detection with simple, available and reliable anthropometric tools would be more realistic. Additionally, early recognition could lead to early weight management strategy and there is no harm done. The cost
effectiveness would be superior than the cost of treatments when cardiovascular disease exists. Therefore, the results of the study could be generalized to the young Thai adult population. The limitations of this study were, firstly, the cut-off points of anthropometric indices defining obesity may be population-specific. Secondly, waist and hip measurements may differ from other studies; however, this issue may not be significant as previous studies showed that different techniques were indifferent to cardiovascular outcomes [26]. Lastly, selection and information bias is considered unlikely by virtue of the prospective design. In conclusion, anthropometric indices are highly correlated to the PBF in young Thai adults. BMI is superior to WC, WSR and WHR in predicting obesity. Optimal cut-off points for sex and age are specific and recommended for use. The global cut-off points of WSR in women regardless of ethnicity are supported. These findings increase public health awareness to stimulate early diagnosis of obesity with simple screening tools in young adults. Acknowledgments We wish to acknowledge Professor James A. Will, University of Wisconsin-Madison, for editing the manuscript via the Faculty of Medicine Publication Clinic, Khon Kaen University, Thailand. This manuscript was funded by the Neuroscience Research and Development Group, Khon Kaen University, Thailand. Conflict of interest The author (s) declared no potential conflicts of interest with respect to the authorship and/or publication of this article.
References 1. James, P. T. (2004). Obesity: The worldwide epidemic. Clinics in Dermatology, 22(4), 276–280. 2. Yang, F., Lv, J. H., Lei, S. F., Chen, X. D., Liu, M. Y., Jian, W. X., et al. (2006). Receiver-operating characteristic analyses of body mass index, waist circumference and waist-to-hip ratio for obesity: Screening in young adults in central south of China. Clinical Nutrition, 25(6), 1030–1039. 3. World Health Organization. (2014). Global Health Observatory (GHO). http://www.who.int/gho/ncd/risk_factors/obesity_text/en/ . Accessed 14 March 2014. 4. Tawonrungrojn, S. (2011). Indicators and their range of values to indicate obesity. Songkla Medicine Journal, 29(2), 89–96. 5. Cai, L., Liu, A., Zhang, Y., & Wang, P. (2013). Waist-to-height ratio and cardiovascular risk factors among Chinese adults in Beijing. PLoS One, 8(7), e69298. 6. Cheong, K. C., Yusoff, A. F., Ghazali, S. M., Lim, K. H., Selvarajah, S., Haniff, J., et al. (2013). Optimal BMI cut-off values for predicting diabetes, hypertension and hypercholesterolaemia in a multi-ethnic population. Public Health Nutrition, 16(3), 453–459. 7. Okorodudu, D. O., Jumean, M. F., Montori, V. M., RomeroCorral, A., Somers, V. K., Erwin, P. J., et al. (2010). Diagnostic performance of body mass index to identify obesity as defined by body adiposity: A systematic review and meta-analysis. International Journal of Obesity (London), 34(5), 791–799. 8. Paniagua, L., Lohsoonthorn, V., Lertmaharit, S., Jiamjarasrangsi, W., & Williams, M. A. (2008). Comparison of waist circumference,
123
J Community Health
9.
10.
11.
12.
13.
14.
15.
16.
17.
body mass index, percent body fat and other measure of adiposity in identifying cardiovascular disease risks among Thai adults. Obesity Research & Clinical Practice, 2(3), I–II. Park, S. H., Choi, S. J., Lee, K. S., & Park, H. Y. (2009). Waist circumference and waist-to-height ratio as predictors of cardiovascular disease risk in Korean adults. Circulation Journal, 73(9), 1643–1650. Wang, F., Wu, S., Song, Y., Tang, X., Marshall, R., Liang, M., et al. (2009). Waist circumference, body mass index and waist to hip ratio for prediction of the metabolic syndrome in Chinese. Nutrition Metabolism and Cardiovascular Diseases, 19(8), 542–547. Worachartcheewan, A., Dansethakul, P., Nantasenamat, C., Pidetcha, P., & Prachayasittikul, V. (2012). Determining the optimal cutoff points for waist circumference and body mass index for identification of metabolic abnormalities and metabolic syndrome in urban Thai population. Diabetes Research and Clinical Practice, 98(2), e16–e21. Browning, L. M., Hsieh, S. D., & Ashwell, M. (2010). A systematic review of waist-to-height ratio as a screening tool for the prediction of cardiovascular disease and diabetes: 0.5 could be a suitable global boundary value. Nutrition Research Reviews, 23(2), 247–269. Cheong, K. C., Ghazali, S. M., Hock, L. K., Yusoff, A. F., Selvarajah, S., Haniff, J., et al. (2013). Optimal waist circumference cut-off values for predicting cardiovascular risk factors in a multi-ethnic Malaysian population. Obesity Research & Clinical Practice. doi: 10.1016/j.orcp.2013.03.004. Pongchaiyakul, C., Limpawattana, P., Kotruchin, P., & Rajatanavin, R. (2013). Prevalence of sarcopenia and associated factors among Thai population. Journal of Bone and Mineral Metabolism, 31(3), 346–350. Hanley, J. A., & McNeil, B. J. (1983). A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology, 148(3), 839–843. Kyle, U. G., Gremion, G., Genton, L., Slosman, D. O., Golay, A., & Pichard, C. (2001). Physical activity and fat-free and fat mass by bioelectrical impedance in 3853 adults. Medicine and Science in Sports and Exercise, 33(4), 576–584. Mijnarends, D. M., Meijers, J. M., Halfens, R. J., ter Borg, S., Luiking, Y. C., Verlaan, S., et al. (2013). Validity and reliability of tools to measure muscle mass, strength, and physical performance in community-dwelling older people: A systematic
123
18.
19.
20.
21.
22.
23.
24.
25.
26.
review. Journal of the American Medical Directors Association, 14(3), 170–178. Chen, L. K., Liu, L. K., Woo, J., Assantachai, P., Auyeung, T. W., Bahyah, K. S., et al. (2013). Sarcopenia in Asia: Consensus report of the asian working group for sarcopenia. Journal of the American Medical Directors Association, 15(2), 95–101. Janghorbani, M., & Amini, M. (2010). Comparison of body mass index with abdominal obesity indicators and waist-to-stature ratio for prediction of type 2 diabetes: The Isfahan diabetes prevention study. Obesity Research and Clinical Practice, 4(1), e1–e82. Cheng, C. H., Ho, C. C., Yang, C. F., Huang, Y. C., Lai, C. H., & Liaw, Y. P. (2010). Waist-to-hip ratio is a better anthropometric index than body mass index for predicting the risk of type 2 diabetes in Taiwanese population. Nutrition Research, 30(9), 585–593. Samsen, M., Hanchaiphiboolkul, S., Puthkhao, P., Tantirittisak, T., & Towanabut, S. (2012). Appropriate body mass index and waist circumference cutoffs for middle and older age group in Thailand: Data of 19,621 participants from Thai Epidemiologic Stroke (TES) study. Journal of the Medical Association of Thailand, 95(9), 1156–1166. Ashwell, M., Gunn, P., & Gibson, S. (2011). Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: Systematic review and metaanalysis. Obesity Reviews, 13(3), 275–286. Cameron, A. J., Magliano, D. J., Shaw, J. E., Zimmet, P. Z., Carstensen, B., Alberti, K. G., et al. (2012). The influence of hip circumference on the relationship between abdominal obesity and mortality. International Journal of Epidemiology, 41(2), 484–494. Oka, R., Kobayashi, J., Yagi, K., Tanii, H., Miyamoto, S., Asano, A., et al. (2008). Reassessment of the cutoff values of waist circumference and visceral fat area for identifying Japanese subjects at risk for the metabolic syndrome. Diabetes Research and Clinical Practice, 79(3), 474–481. Ekelund, U., Neovius, M., Linne, Y., Brage, S., Wareham, N. J., & Rossner, S. (2005). Associations between physical activity and fat mass in adolescents: The Stockholm Weight Development study. American Journal of Clinical Nutrition, 81(2), 355–360. Wang, Z., Ma, J., & Si, D. (2010). Optimal cut-off values and population means of waist circumference in different populations. Nutrition Research Reviews, 23(2), 191–199.
ORIGINAL PAPER
The Performance of Obesity Screening Tools Among Young Thai Adults Panita Limpawattana • Thepkhachi Kengkijkosol Prasert Assantachai • Orapitchaya Krairit • Jiraporn Pimporm
•
Ó Springer Science+Business Media New York 2014
Abstract Obesity is a worldwide medical condition that leads to physical and psychological impairment. Specific ethnicity, gender and age group are related to different performances of anthropometric indices to predict obesity. The objectives of this study were to estimate the performance of the anthropometric indices for detecting obesity based on percentage of body fat (PBF), to study the correlation among those indices, and to determine the optimal cut-off point of the indices among young Thai adults. This is a cross-sectional study of healthy urban subjects in Khon Kaen, Thailand who were aged 20–39 years. Baseline characteristics and anthropometric measures were collected. PBF was determined using bioelectrical impedance analysis. Demographic data and anthropometric variables were analyzed using descriptive statistics. Receiver-
operating characteristic (ROC) curves were used to compare the performance of anthropometric measures as predictors of obesity. One-hundred men and 100 women were recruited for this study. Body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR) and waistto-stature ratio (WSR) were significantly correlated to PBF. BMI demonstrated the best performance according to the area under the ROC curves in both sexes at cut-off points of 22.5 in women or 25 kg/m2 in men. WC and WSR showed better performance than WHR to detect obesity. In conclusion, anthropometric indices in young Thai adults were correlated well with PBF to predict obesity as shown in prior reports. Different cut-off points of these indices to define obesity in young Thai adults are recommended. The global cut-off points of WSR in women regardless of ethnicity are supported.
A1 A2 A3 A4 A5
P. Limpawattana (&) Division of Geriatric Medicine, Department of Internal Medicine, Faculty of Medicine, Khon Kaen University, Khon Kaen 40002, Thailand e-mail: [email protected]
Keywords Body mass index Waist circumference Waist-to-hip ratio Waist-to-statue ratio Percentage body fat
A6 A7 A8
T. Kengkijkosol Department of Internal Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
Introduction
A9 A10 A11 A12
P. Assantachai Department of Preventive and Social Medicine, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand
A13 A14 A15 A16
O. Krairit Divison of Geriatric Medicine, Department of Internal Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
A17 A18 A19
J. Pimporm Outpatient Clinic, Srinagarind Hospital, Faculty of Medicine, Khon Kaen University, Khon Kaen, Thailand
Obesity is a medical condition in which excessive adipose tissue accumulates in the body leading to an impairment of both physical and psychological health and well-being [1]. It has been proven as a strong risk factor for cardiovascular disease [1, 2]. Obesity is a global public health problem. An estimated 315 million people worldwide are obese and at least 2.8 million die annually as a result of being overweight or obesity [1, 3]. In Western and westernized countries, the prevalence of obesity is about 20 % in men and 25 % in women [1]. For Asian countries, its prevalence is lower but it has tended to increase considerably. Women
123
J Community Health
are more likely to be obese than men and in the WHO regions for Africa, Eastern Mediterranean and South East Asia, women had approximately twice the obesity prevalence of men [3]. A measurement of percentage body fat (PBF) is considered the reference method for defining obesity in which PBF [25 % for men or 30 % and for women, were cut-off points to diagnose obesity since these points were associated with health-related adverse outcomes of obesity [2, 4]. Dual energy X-ray absorptiometry (DXA) is a frequently used tool to measure PBF. The limitations of its use are the unavailability in general settings and being an expensive tool particularly in developing countries [2, 4]. Bioimpedance analysis (BIA) is another technique to measure body composition including PBF. It was developed to estimate the volume of body fat and lean body mass. It is portable, easy to use and has a relatively lower cost than DXA [2, 4]. The accuracy of BIA is interfered with by some factors such as dehydration, strenuous exercise, pregnancy and having a heavy meal; controlling these factors is generally regarded as being essential to making this method an accurate instrument to measure body composition [4]. Anthropometric indices including body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR) and waist-to-height ratio (WHR) have been studied widely as the commonly used tools for assessing obesity due to the simplicity, low cost, short time consuming and their strong correlations with PBF, adverse metabolic profiles such as diabetes and metabolic syndrome, and cardiovascular disease [2, 5–11]. There is strong evidence that Asians should have lower cut-off values than Europeans for BMI, WC and WHR [6]. For WSR, a systematic review revealed that a cut-off point at 0.5 was a universal value regardless of ethnicity [12]. The cut-off points of BMI, WC, WHR and WSR in Asian studies; however, remained diverse across countries [4, 8]. Additionally, they were sex and age group specific in some reports [2, 13]. For example, obesity was defined as BMI [27 in men or [25 kg/m2 in women for Thais aged 20–84 years and [25 kg/m2 in both sexes for Singaporean persons [4]. For Malaysian adults, WC cut-offs of 81 cm. in men and 80 cm. in women were associated with abdominal obesity [13] while in China, the optimal values of WC were 78.9 cm in men and 65.8 cm in women for adults aged 20–30 years, and 82.4 cm in men and 71.4 cm in women for adults aged 31–45 years [2]. The studies in the Thai population for specific age groups are lacking. Therefore, the primary objective of this study was to estimate the performance of the anthropometric indices for detecting obesity. The second objective was to study the correlation among those indices and to determine the optimal cut-off points of the indices among young Thai adults.
123
Materials and Methods Study Subjects and Setting This was a cross-sectional study. The setting was the urban area of Khon Kaen province, Thailand which is the secondlargest Northeastern province. It is located 445 km away from Bangkok with a population of 1.8 million [14]. The potential subjects were sent a letter of invitation asking if they would like to participate in the study. Subjects were a sample of young healthy Thai adults who aged 20–39 years. The exclusion criteria were the persons who could not stand upright, had a pacemaker, used drugs, herbs or any hormones that affect muscle mass and strength such as contraceptive drugs, steroid hormones and thyroid hormone, drank alcohol 12 h prior to the test, had vigorous exercise within 12 h before analysis, were dehydrated, and women during the menstrual period or pregnancy. Data Collection Baseline data were collected and anthropometric measures were taken while subjects were lightly clothed and wore no shoes. WCs were taken midway between the inferior margin of the last rib and the iliac crest at the end of expiration and hip circumference (HC) was measured at the widest part of hip. Measurements of standing height were performed without shoes with a stadiometer. Body weight and percent body fat (PBF) estimation were determined using the Tanita bioelectrical impedance analysis (BIA) system (model BC-418). BMI was calculated as weight (kg) divided by height squared (m2). The waist-to-hip ratio (WHR) and waist-to-stature ratio (WSR) were then calculated. Obesity was defined by PBF [30 % in women and [25 % in men [2]. Procedure Eligible subjects had no food or drink 4 h prior the test and written informed consent was obtained from all subjects. Baseline data and anthropometric indices were collected by trained nurses and then body composition was measured by the body composition monitor with a scale. Sample Size Calculation Sample size calculations were based on the areas under the receiver operating characteristic curves (ROC) and areas under the curves (AUC) according to the methodology of Hanley and McNeil (1983) [15]. ROC curves were used to summarize the accuracy of diagnostic tests. This method varies the sample size until a sufficiently small standard error (SE) of the area under the ROC curve was achieved.
J Community Health Table 1 Baseline characteristics of subjects by sex Variables
Women (N = 100)
Men (N = 100)
Age (years); med (IQR1, IQR3)
25 (29,34.5)
29 (24,34)
Body weight (kg); med (IQR1, IQR3)
55.05 (48.55,63.95)
65.6 (59,73.25)
Height (m); mean (SD) BMI; med (IQR1, IQR3)
157.49 (5.91) 22.05 (19.75,26)
169.77 (7.09) 23.05 (20.30,25.45)
WC (cm); med (IQR1, IQR3)
78 (70.5,86)
81 (75.5,88)
HC (cm); med (IQR1, IQR3)
93 (89,102)
93 (88,99)
WHR; med (IQR1, IQR3)
0.81 (0.78,0.88)
0.88 (0.82,0.93)
WSR; med (IQR1, IQR3)
0.50 (0.44,0.55)
0.48 (0.44,0.52)
PBF (%); med (IQR1, IQR3)
32 (26.9,38.05)
19.8 (14.75,24.25)
Med median, IQR inter-quartile range, bmi body mass index (kg/m2), WC waist circumference (cm), HC hip circumference (cm), WHR waist-to-hip ratio, WSR; waist-to-stature ratio, PBF percentage body fat (%)
Because of the complexity of the formula, a web-based calculator (www.anaesthetist.com/mnm/stats/roc/#stderr) was used to determine the standard error. Finally, a sample size of 200 participants of men and women was found to be adequate and feasible to conduct the trial in clinical practice at the AUC of 0.9 and S.E. of 0.04.
Statistical Analyses Demographic data and anthropometric variables were analyzed using descriptive statistics, presentations in percentage, mean and standard deviations. If the distribution of these data was not a normal distribution, then medians, and inter-quartile ranges were used instead. Spearman’s rank correlation coefficients were calculated to quantify associations between anthropometric measures (BMI, WC, WHR and WSR) and PBF to predict obesity. The ROC curve was used to summarize the overall accuracy of the anthropometric indices for obesity detection. Then optimal cut-off points were determined. The performances of the tests were summarized by the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and likelihood ratios. All of the data analyses were performed by using STATA version 10.0 (StataCorp, College Station, TX, USA). Ethics approval was provided by the ethics committee of Medicine Faculty, Khon Kaen University according to the Helsinki Declaration.
Fig. 1 Spearman rank correlation (rho) between percent body fat and other parameters by sex BMI body mass index (kg/m2), WC waist circumference (cm), WHR waist-to-hip ratio, WSR waist-to-stature ratio, PBF percentage of fat (%)
Results There were 100 men and 100 women recruited in this study. Basic anthropometric indices are shown in Table 1. It was found that the median/average values of age, body weight, height, BMI, WC, HC and WHR in men were greater than those found in women while WSR and PBF were lower in men than in women. Simple correlations between PBF and anthropometric indices in both sexes are presented in Fig. 1. All the correlations were statistically significant (p \ 0.05). In women, all indices except WHR (q \ 0.5) had good correlations whereas in men, a slightly lower but still significant correlation was detected. WHR remained the lowest correlation (q 0.37) in men as in women. Obesity, when defined based on the PBF as [30 % in women or [25 % in men [7]; the majority of women (60 %) and minority of men (21 %) were obese. The performances of the anthropometric indices including the cutoff points for all variables corresponding to the criterion values with the best tradeoffs are shown in Table 2. The ROC curves for obesity diagnosis among anthropometric indices in women and men are demonstrated in Figs. 2 and 3. According to the curves, BMI was the most predictable for obesity in both sexes and WHR was the worst one. In particular, BMI in women and WSR in men showed obvious good performances for obesity detection, given very high likelihood ratios.
Discussion The anthropometric indices of young Thai adults were quite similar to the data of the young Chinese population. In addition, the PBF in men was much lower than the one for women [2]. Although the Chinese study used DXA to measure body composition, the BIA technique used here,
123
J Community Health Table 2 Performance of obesity parameters by sex Indices
BMI
WC
WHR
WSR
Woman
Man
Woman
Man
Woman
Man
Woman
Man
AUC 95 % (CI)
0.96 (0.94, 0.99)
0.93 (0.87, 0.90)
0.91 (0.85, 0.97)
0.91 (0.81, 1.00)
0.72 (0.62, 0.83)
0.75 (0.63, 0.86)
0.90 (0.84, 0.96)
0.91 (0.81, 1.00)
Cut-off point
22.5
25
75
89
0.776
0.75
0.503
0.533
Sensitivity Specificity
80 % 97.5 %
90.48 % 89.97 %
85.00 % 77.5 %
76.19 % 92.41 %
90 % 97.5 %
76.1 % 68.35 %
71.67 % 85 %
71.43 % 98.73 %
PPV
98 %
70.37 %
85 %
76.19 %
76.47 %
57.14 %
87.76 %
93.75 %
NPV
76.47 %
97.26 %
77.5 %
93.6 %
86.67 %
93.01 %
66.67 %
92.86 %
LR?
32
8.93
3.78
10.02
1.68
2.26
4.78
56.43
LR-
0.31
0.11
0.19
2.26
0.25
0.42
0.33
0.29
2
Bmi body mass index (kg/m ), WC waist circumference (cm), WHR waist-to-hip ratio, WSR waist-to-stature ratio, AUC area under receiving operating characteristic curve, CI confidence interval, PPV positive predictive values, NPV negative predictive values, LR? likelihood ratio positive, LR- likelihood ratio negative
Fig. 2 ROC curves of anthropometric indices to predict obesity using percent body fat as a gold standard in women
after controlling interfering factors, the results showed similar patterns supporting the prior studies [2, 16–18]. Comparing the PBF in the same age group of the current study to one report in Europe where the average PBF ranged from 24.9 to 26.1 % in women and 21.7 to 24.5 % in men, a higher PBF in Thai women and lower PBF in Thai men, confirmed the differences between ethnicities, body size, life styles, and cultural backgrounds that might influence the diversity of body composition [4, 18]. Therefore, different cut-off points for anthropometric indices should be identified. All indices correlated significantly with the PBF with a greater degree (0.67–0.94) than the preceding study of Chinese populations (0.57–0.75) for both sexes except in the area of WHR that showed lower correlations (0.37–0.42 vs. 0.47–0.65) [2]. As high PBF is correlated with adverse health outcomes [7]. and existing data showed these
123
Fig. 3 ROC curves of anthropometric indices to predict obesity using percent body fat as a gold standard in men
anthropometric indexes all correlated to cardiovascular risk factors, the results from this study were comparable to these studies [5, 6, 9, 11, 19, 20]. The current results determined the different cut-off points of BMI, WC, WHR and WSR for detection of obesity among young adults. The overall performances of screening tools according to the AUC curves were good and correlated with existing data [8–11, 13, 20]. The optimal cut-off points of BMI, WC and WHR were lower than European countries, consistent with many Asian studies though different cut-off points were reported [8–11, 21]. Even for prior Thai studies e.g. the National Thai population that recommends a BMI [23 kg/m2 of both sexes, or[25 kg/m2 in women and[27 kg/m2 in men aged 20–84 years as obesity [4], the differences can be explained by the study populations. In this study that included only young Thai adults the outcomes using the
J Community Health
National study would indicate obesity. One study in Malaysia recommended the cut-off points of WC in women \80 cm and men \81 cm which were lower than most Asian studies for cut-off points in men which the authors explained these differences from the higher prevalence of cardiovascular disease in Malaysian populations than other Asian countries [13]. BMI showed the best performance in this study based on the AUC of the ROC curves. This result implies that total body fat (BMI) is likely to play a more major role than abdominal fat in predicting obesity in young Thai adults since BMI is considered an indicator of body fatness [1]. The optimal cut-off points of the WSR in this study showed similar results for women and a slightly higher result for men than existing reports [5, 12], supporting the evidence that the cut-off points of 0.5 could be a suitable global boundary values independent of ethnicities [12]. The performance of WSR in this study was comparable to WC and slightly lower than BMI to predict obesity based on the AUC of the ROC curves though some studies showed that WSR was a better screening tool than WC and BMI [9, 22]. Interpretation must be careful because those studies’ outcomes were for cardiovascular disease risk. The usefulness of WHR to detect obesity in this study showed the lowest performance. As compared to the previous Thai study in adults of an average age of 41.9 ± 10.6 years, the performances were similar and are in contrast to the report of Taiwanese that showed WHR had the higher adjusted odds ratios than BMI to predict the risk of type 2 diabetes [11, 20]. This can be explained by the limitation of WHR that reflect differences of body proportion in different age groups and ethnicities. For HC, there are numerous reports regarding the inverse effects on metabolic profiles including reduced cardiovascular risks and mortality after adjustment for WC, possibly due to measurement of HC alone represents only subcutaneous fat, not including visceral fat since visceral fat is correlated to increased risks of cardiovascular disease [23, 24]. Therefore, analyzing the performance of HC alone in these cross-sectional data was inappropriate. Given the dramatically increased obesity prevalence in adolescents and young adults [25] in addition to the adverse health outcomes of obesity, early screening with different cut-off points in different age groups and sexspecific criteria are recommended. One study in China also demonstrated various cut-off points for each age group provided higher performances than single age criteria [2]. Although body fat is closely related to obesity and other unfavorable outcomes, measurement by DXA or even BIA in the general population especially in the Southeast Asia region is impractical. Early detection with simple, available and reliable anthropometric tools would be more realistic. Additionally, early recognition could lead to early weight management strategy and there is no harm done. The cost
effectiveness would be superior than the cost of treatments when cardiovascular disease exists. Therefore, the results of the study could be generalized to the young Thai adult population. The limitations of this study were, firstly, the cut-off points of anthropometric indices defining obesity may be population-specific. Secondly, waist and hip measurements may differ from other studies; however, this issue may not be significant as previous studies showed that different techniques were indifferent to cardiovascular outcomes [26]. Lastly, selection and information bias is considered unlikely by virtue of the prospective design. In conclusion, anthropometric indices are highly correlated to the PBF in young Thai adults. BMI is superior to WC, WSR and WHR in predicting obesity. Optimal cut-off points for sex and age are specific and recommended for use. The global cut-off points of WSR in women regardless of ethnicity are supported. These findings increase public health awareness to stimulate early diagnosis of obesity with simple screening tools in young adults. Acknowledgments We wish to acknowledge Professor James A. Will, University of Wisconsin-Madison, for editing the manuscript via the Faculty of Medicine Publication Clinic, Khon Kaen University, Thailand. This manuscript was funded by the Neuroscience Research and Development Group, Khon Kaen University, Thailand. Conflict of interest The author (s) declared no potential conflicts of interest with respect to the authorship and/or publication of this article.
References 1. James, P. T. (2004). Obesity: The worldwide epidemic. Clinics in Dermatology, 22(4), 276–280. 2. Yang, F., Lv, J. H., Lei, S. F., Chen, X. D., Liu, M. Y., Jian, W. X., et al. (2006). Receiver-operating characteristic analyses of body mass index, waist circumference and waist-to-hip ratio for obesity: Screening in young adults in central south of China. Clinical Nutrition, 25(6), 1030–1039. 3. World Health Organization. (2014). Global Health Observatory (GHO). http://www.who.int/gho/ncd/risk_factors/obesity_text/en/ . Accessed 14 March 2014. 4. Tawonrungrojn, S. (2011). Indicators and their range of values to indicate obesity. Songkla Medicine Journal, 29(2), 89–96. 5. Cai, L., Liu, A., Zhang, Y., & Wang, P. (2013). Waist-to-height ratio and cardiovascular risk factors among Chinese adults in Beijing. PLoS One, 8(7), e69298. 6. Cheong, K. C., Yusoff, A. F., Ghazali, S. M., Lim, K. H., Selvarajah, S., Haniff, J., et al. (2013). Optimal BMI cut-off values for predicting diabetes, hypertension and hypercholesterolaemia in a multi-ethnic population. Public Health Nutrition, 16(3), 453–459. 7. Okorodudu, D. O., Jumean, M. F., Montori, V. M., RomeroCorral, A., Somers, V. K., Erwin, P. J., et al. (2010). Diagnostic performance of body mass index to identify obesity as defined by body adiposity: A systematic review and meta-analysis. International Journal of Obesity (London), 34(5), 791–799. 8. Paniagua, L., Lohsoonthorn, V., Lertmaharit, S., Jiamjarasrangsi, W., & Williams, M. A. (2008). Comparison of waist circumference,
123
J Community Health
9.
10.
11.
12.
13.
14.
15.
16.
17.
body mass index, percent body fat and other measure of adiposity in identifying cardiovascular disease risks among Thai adults. Obesity Research & Clinical Practice, 2(3), I–II. Park, S. H., Choi, S. J., Lee, K. S., & Park, H. Y. (2009). Waist circumference and waist-to-height ratio as predictors of cardiovascular disease risk in Korean adults. Circulation Journal, 73(9), 1643–1650. Wang, F., Wu, S., Song, Y., Tang, X., Marshall, R., Liang, M., et al. (2009). Waist circumference, body mass index and waist to hip ratio for prediction of the metabolic syndrome in Chinese. Nutrition Metabolism and Cardiovascular Diseases, 19(8), 542–547. Worachartcheewan, A., Dansethakul, P., Nantasenamat, C., Pidetcha, P., & Prachayasittikul, V. (2012). Determining the optimal cutoff points for waist circumference and body mass index for identification of metabolic abnormalities and metabolic syndrome in urban Thai population. Diabetes Research and Clinical Practice, 98(2), e16–e21. Browning, L. M., Hsieh, S. D., & Ashwell, M. (2010). A systematic review of waist-to-height ratio as a screening tool for the prediction of cardiovascular disease and diabetes: 0.5 could be a suitable global boundary value. Nutrition Research Reviews, 23(2), 247–269. Cheong, K. C., Ghazali, S. M., Hock, L. K., Yusoff, A. F., Selvarajah, S., Haniff, J., et al. (2013). Optimal waist circumference cut-off values for predicting cardiovascular risk factors in a multi-ethnic Malaysian population. Obesity Research & Clinical Practice. doi: 10.1016/j.orcp.2013.03.004. Pongchaiyakul, C., Limpawattana, P., Kotruchin, P., & Rajatanavin, R. (2013). Prevalence of sarcopenia and associated factors among Thai population. Journal of Bone and Mineral Metabolism, 31(3), 346–350. Hanley, J. A., & McNeil, B. J. (1983). A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology, 148(3), 839–843. Kyle, U. G., Gremion, G., Genton, L., Slosman, D. O., Golay, A., & Pichard, C. (2001). Physical activity and fat-free and fat mass by bioelectrical impedance in 3853 adults. Medicine and Science in Sports and Exercise, 33(4), 576–584. Mijnarends, D. M., Meijers, J. M., Halfens, R. J., ter Borg, S., Luiking, Y. C., Verlaan, S., et al. (2013). Validity and reliability of tools to measure muscle mass, strength, and physical performance in community-dwelling older people: A systematic
123
18.
19.
20.
21.
22.
23.
24.
25.
26.
review. Journal of the American Medical Directors Association, 14(3), 170–178. Chen, L. K., Liu, L. K., Woo, J., Assantachai, P., Auyeung, T. W., Bahyah, K. S., et al. (2013). Sarcopenia in Asia: Consensus report of the asian working group for sarcopenia. Journal of the American Medical Directors Association, 15(2), 95–101. Janghorbani, M., & Amini, M. (2010). Comparison of body mass index with abdominal obesity indicators and waist-to-stature ratio for prediction of type 2 diabetes: The Isfahan diabetes prevention study. Obesity Research and Clinical Practice, 4(1), e1–e82. Cheng, C. H., Ho, C. C., Yang, C. F., Huang, Y. C., Lai, C. H., & Liaw, Y. P. (2010). Waist-to-hip ratio is a better anthropometric index than body mass index for predicting the risk of type 2 diabetes in Taiwanese population. Nutrition Research, 30(9), 585–593. Samsen, M., Hanchaiphiboolkul, S., Puthkhao, P., Tantirittisak, T., & Towanabut, S. (2012). Appropriate body mass index and waist circumference cutoffs for middle and older age group in Thailand: Data of 19,621 participants from Thai Epidemiologic Stroke (TES) study. Journal of the Medical Association of Thailand, 95(9), 1156–1166. Ashwell, M., Gunn, P., & Gibson, S. (2011). Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: Systematic review and metaanalysis. Obesity Reviews, 13(3), 275–286. Cameron, A. J., Magliano, D. J., Shaw, J. E., Zimmet, P. Z., Carstensen, B., Alberti, K. G., et al. (2012). The influence of hip circumference on the relationship between abdominal obesity and mortality. International Journal of Epidemiology, 41(2), 484–494. Oka, R., Kobayashi, J., Yagi, K., Tanii, H., Miyamoto, S., Asano, A., et al. (2008). Reassessment of the cutoff values of waist circumference and visceral fat area for identifying Japanese subjects at risk for the metabolic syndrome. Diabetes Research and Clinical Practice, 79(3), 474–481. Ekelund, U., Neovius, M., Linne, Y., Brage, S., Wareham, N. J., & Rossner, S. (2005). Associations between physical activity and fat mass in adolescents: The Stockholm Weight Development study. American Journal of Clinical Nutrition, 81(2), 355–360. Wang, Z., Ma, J., & Si, D. (2010). Optimal cut-off values and population means of waist circumference in different populations. Nutrition Research Reviews, 23(2), 191–199.
