Body composition patterns among normal glycemic, pre-diabetic, diabetic health Chinese adults in community: NAHSIT 2013-2016.

BACKGROUND: Central obesity is known to be associated with diabetes. Increasing lower extremity circumference was hypothesized in association with lower risk of diabetes.
OBJECTIVE: This study determined which anthropometric patterns correlates the best with pre-diabetic and diabetic status among healthy adults.
DESIGN: Cross-sectional study with nationwide population sampling of participants was designed. PARTICIPANTS: In total, 1,358 ethnic Chinese adult participants were recruited from the Nutrition and Health Survey in Taiwan 2013-2016; the whole-body composition was measured through dual-energy X-ray absorptiometry. MAIN OUTCOME MEASURES: Fat and lean mass in whole and specific parts of body among heathy Asian adults with normal glycemic, pre-diabetic, and diabetic states were measured, separately. STATISTICAL ANALYSES PERFORMED: The generalized linear model was used to investigate the association between body composition (lean and fat mass) and hyperglycemic status. The reduced rank regression (RRR) was used to confirm the correlation between glycemic status and predicting factors (body composition parameters).
RESULTS: Trunk fat positively correlated with the fasting glucose level (r = 0.327, P < 0.001) and HbA1c (r = 0.329, P < 0.001), whereas limb fat negatively correlated with the fasting glucose level (r = -0.325, P < 0.001) and HbA1c (ρ = -0.342, P < 0.001), respectively. In RRR analyses, fasting glucose and HbA1c exhibited a high positive association on fat amount per lean mass of the trunk (factor loading = 0.5319 and 0.5599, respectively) and of android area (0.6422 and 0.6104) and a high negative association fat amount per lean mass of the legs (-0.3863 and -0.3083) and gynoid area (-0.3414 and -0.3725).
CONCLUSIONS: For healthy community participants, increasing trunk fat had a greater risk of hyperglycemic status. Increasing lower extremity mass may confer lower risk of diabetes.

Introduction

Central obesity is markedly associated with numerous hazardous health effects, such as cardiovascular diseases [1, 2], increased insulin resistance [3, 4], and diabetes mellitus [3, 5]. The prevalence of obesity has increased rapidly in Taiwan [6, 7] and other Asia-Pacific regions [8]. Studies [9, 10] have indicated that changes in dietary habits and lifestyles among residents of Taiwan, such as increasing consumption of sweetened beverages, decreasing physical activity, and increasing sedentary lifestyle, are contributing to the increasing prevalence of metabolic syndrome and central obesity. The Nutrition and Health Survey in Taiwan (NAHSIT), a nationwide research program, assesses the nutritional status and the association between dietary patterns and health among dwellers in Taiwan [6, 11]. Additionally, three waves of the NAHSIT from 1993–1996, 2005–2008, to 2013–2016 have indicated a 2-fold increase in the prevalence of obesity (BMI ≥ 27 kg/m2) from 11.8% to 22.0% [6] and diabetes mellitus from 5.3% to 9.1% or above [12] in Taiwan.

Numerous single measurement of the adiposity indices, such as BMI, waist circumference, and waist–hip ratio, are simple surrogate markers for central obesity, but they are limited by overlooking the lean and fat mass distribution in adults [13-15]. Markers of measurement of visceral fat only also limited by neglecting the protective effects of increasing lean mass in hips and lower limbs [16, 17]. Dual-energy X-ray absorptiometry (DXA) is accurate in measuring fat and lean mass in whole and each specific parts of body [18]. We sought to determine the anthropometric patterns which correlates the best for pre-diabetic and diabetic healthy adults.

Materials and methods

Subjects

In the NAHSIT 2013–2016 study, a team of well-trained interviewers and technicians conducted a nationwide cross-sectional assessment of the nutritional status and the association between dietary patterns and health among the general population of Taiwan (ethnic Chinese) during 2013–2016. The nutritional evaluation was continued and extending from previous NAHSIT 2005–2008 study [19], including both the behavior and health outcome indicators, such as questionnaires of dietary recall, of food frequency and habits, of dietary and nutritional knowledge, and serum measurement for clinical biochemistry analytes. The detailed items were described elsewhere [19]. By using a three-stage probability sampling covering 359 townships or city districts, the representative participants for the general Taiwanese population were selected [20]. Additionally, this team obtained health information from the participants by door-to-door visits, and using a standardized nutritional questionnaire. Individual socioeconomic status (iSES) was assessed by a sum of z scores of personal income and education years. For iSES, the lowest, middle, and highest classes corresponded to tertiles of total z scores of personal income and education years. The participants’ serum specimens were tested by a centralized laboratory. Finally, from the NAHSIT 2013–2016 study, 1,358 participants who completed the body composition examination and laboratory tests of fasting glucose (FG) and HbA1c were included in the present study. This study was approved by the Institutional Review Board on Biomedical Science Research, Academia Sinica, Taiwan (AS-IRB02-103003) and Research Ethics Committee, National Health Research Institutes, Taiwan (EC1020110). Data cannot be shared publicly and uploaded because of legal restrictions imposed by the government of Taiwan on the distribution of the personal health data in relation to the “Personal Information Protection Act.” Data are only available from the formal proposal to the Health and Welfare Data Science Center (HWDC), Ministry of Health and Welfare, Taiwan for researchers who meet the criteria for access to confidential data.

Definition of prediabetes and diabetes

Blood samples were collected for measuring FG and HbA1c. Definition of prediabetes and diabetes is according to the latest standards of American Diabetes Association [21]. For FG, levels of <100 mg/dl, 100–125 mg/dl, and >125 mg/dl were categorized into normal, prediabetic, and diabetic groups, respectively [21]. HbA1c levels of <5.7%, 5.7–6.4%, and >6.4% were classified into normal, prediabetic, and diabetic groups, respectively [21]. Diagnosis of diabetes mellitus was confirmed on the basis of self-reported diabetes with treatment, or laboratory tests with FG ≥126 mg/dL, or HbA1c ≥ 6.5%.

Data collection of body composition

Demographic information, and results of anthropometric assessment of weight, waist circumference, and BMI were recorded. Body composition was measured using the mobile DXA (Prodigy, GE Lunar Health Care, Wisconsin, USA) when conducting the nutritional status survey. Furthermore, the body composition parameters were defined as follows:

Fat body weight (%): (Total fat mass/1000)/ Body weight × 100

Limb fat body weight (%): ((Arms fat mass + Legs fat mass)/1000)/ Body weight × 100

Trunk fat body weight (%): (Trunk fat mass)/1000)/ Body weight × 100

Lean body weight (%): (Total lean mass/1000)/ Body weight × 100

Limb lean body weight (%): ((Arms lean mass + Legs lean mass)/1000)/ Body weight × 100

Trunk lean body weight (%): (Trunk lean mass)/1000)/ Body weight × 100

Limb in fat (%): (Arms fat mass + Legs fat mass)/Total fat mass × 100

Arms in fat (%): (Arms fat mass/Total fat mass) × 100

Legs in fat (%): (Legs fat mass/Total fat mass) × 100

Android in fat (%): (Android fat mass/Total fat mass) × 100

Gynoid in fat (%): (Gynoid fat mass/Total fat mass) × 100

Total region fat (%): (Fat mass in limbs and trunk/ Body weight) × 100

Total tissue fat (%): (Total fat mass/ Body weight) × 100

Statistical analysis

Continuous variables were expressed as mean and standard deviation and analyzed using the Kruskal–Wallis H test. The Pearson correlation statistics was used to examine the association between the blood glucose or HbA1c level and body composition markers. The generalized linear model (GLM) was used to estimate body composition parameters according to the blood glucose and HbA1c levels after adjustment for covariates of age, sex, blood pressure, and high density lipoprotein cholesterol (HDL-C). Reduced rank regression (RRR) is a modern method in epidemiology assessing the role of the response variables and latent variables [22, 23]. By using RRR, body composition patterns associated with blood glucose and HbA1c were derived. Sensitivity analysis was performed by excluding participants with diabetes mellitus who received treatment. All analyses were performed using SAS software 9.4 (Cary, NC). An alpha level of <0.05 was defined as statistically significant.

Sensitivity analyses

Two sensitivity analyses were performed. First, we analyzed the association between body composition markers and different blood glucose and HbA1c levels by excluding patients of diabetes mellitus with medical treatment. Second, we investigated the relationship between body composition markers and insulin resistance (IR) by triglyceride glucose-waist circumference index (TyG-WC index) [24, 25]. The TyG and TyG-related markers have been used as the surrogate indicators for IR [25, 26]. Of these markers, the TyG-WC index was more accurate in estimating IR [25]. The formula of the TyG-WC index is Ln (triglyceride [mg/dL] × fasting glucose [mg/dL]/2) × waist circumference [24, 25]. A cutoff value of TyG-WC index ≥ 850 indicates IR [25, 26].

Results

Participant characteristics

Among the enrolled 1,358 participants, 214 (15.8%) had a confirmed diagnosis of diabetes mellitus (Table 1). The average FG and HbA1c levels in the diabetic and nondiabetic groups were 144.9 mg/dL and 97.0 mg/dL, and 7.3% and 5.6%, respectively. The proportions of male participants, levels of blood pressure and triglycerides, and the ever and current smokers were higher in the diabetic group. The average HDL-C level and z score of iSES were lower in the diabetic group. The nondiabetic group had higher SES when compared to the diabetic group. Between the diabetic and nondiabetic groups, alcohol consumption and physical activity was no significantly difference. Besides, 50.9% (109/214) of the diabetic participants took anti-diabetic medications.

Table 1

Markers that related obesity to diabetes mellitus (n = 1358).

Markers, units	Total		Diabetes mellitus				P valuea
	n = 1358		Yes (n = 214)		No (n = 1144)
	Mean	SD	mean	SD	mean	SD
Age, years	52.2	16.9	62.4	11.8	50.4	17.0	<0.001*
Male, n (ratio)	n = 688	(50.7%)	n = 128	(59.8%)	n = 560	(49.0%)	0.004*
HbA1c, %	5.9	0.9	7.3	1.4	5.6	0.4	<0.001*
Glucose, mg/dL	104.6	26.9	144.9	47.5	97.0	8.8	<0.001*
SBPb, mmHg	123.3	18.2	133.1	16.8	121.4	17.8	<0.001*
DBPb, mmHg	75.2	10.6	77.1	11.1	74.9	10.5	0.008*
TG, mg/dL	117.3	79.4	163.3	106.2	108.7	70.1	<0.001*
HDL, mg/dL	54.3	15.6	47.2	13.2	55.7	15.7	<0.001*
Smokingc
Non-smokers, n (ratio)	n = 905	(69.7%)	n = 133	(63.0%)	n = 772	(71.0%)	0.072
Ever smokers, n (ratio)	n = 222	(17.1%)	n = 44	(20.9%)	n = 178	(16.4%)
smokers, n (ratio)	n = 172	(13.2%)	n = 34	(16.1%)	n = 138	(12.7%)
Alcohol consumptiond (g/d)	21.2	100.6	32.1	135.7	19.0	91.8	0.838
Non-drinkers, n (ratio)	n = 582	(56.0%)	n = 103	(58.2%)	n = 479	(55.6%)	0.123
< 16 g/d, n (ratio)	n = 273	(26.3%)	n = 36	(20.3%)	n = 237	(27.5%)
16–35 g/d, n (ratio)	n = 72	(6.9%)	n = 11	(6.2%)	n = 61	(7.1%)
36–64 g/d, n (ratio)	n = 35	(3.4%)	n = 9	(5.1%)	n = 26	(3.0%)
> 64 g/d, n (ratio)	n = 77	(7.4%)	n = 18	(10.2%)	n = 59	(6.8%)
iSES level (z score)	0.0	1.6	-0.53	1.6	0.10	1.6	<0.001*
Lowest SES, n (ratio)	n = 446	(32.8%)	n = 97	(45.3%)	n = 349	(30.5%)	<0.001*
Middle SES, n (ratio)	n = 459	(33.8%)	n = 69	(32.2%)	n = 390	(34.1%)
Highest SES, n (ratio)	n = 453	(33.4%)	n = 48	(22.4%)	n = 405	(35.4%)
Physical activitye, MET hours/week	5.3	24.2	5.3	22.8	5.3	24.5	0.570

Data are expressed as mean and standard deviation (SD) for continuous variables and number (%) for discrete variables. DBP, diastolic blood pressure; HDL, high-density lipoprotein; iSES, individual socioeconomic status; MET, metabolic equivalent of task; n, number; SBP, systolic blood pressure; TG, triglycerides.

*Statistical significance as P < 0.05.

aTested by the Mann-Whitney U test and Chi-square test.

bn = 1277.

cn = 1299.

dn = 1039.

en = 1045.

Body composition markers on fasting glucose level

Participants were categorized into three groups based on the FG levels of ≥126 mg/dL, 100–125 mg/dL, and <100 mg/dL (Table 2). Group with FG < 100 mg/dL exhibited significantly lower weight, BMI, waist circumference, but heavier limb weight in both fat and lean mass than the groups with higher FG levels. Three groups showed no significant difference for lean body weight in total body and trunk area. The Pearson correlation statistics revealed modest positive correlation between FG and tissue percentage of trunk in fat mass (r = 0.327, P <0.001), negative correlation between FG and limb in fat mass percentage (r = -0.325, P <0.001) (Table 3). With increasing of FG levels, GLM regression analyses revealed a consistent increase of trunk weight and decrease of limb weight in both fat and lean mass.

Table 2

Body composition markers that related obesity to fasting glucose levels (n = 1358).

Markers, unit	Fasting Glucose (mg/dL)						P valuea
	≥126 (n = 139)		100–125 (n = 456)		<100 (n = 763)
	Mean	SD	Mean	SD	Mean	SD
Weight, kg	68.7	13.4	67.9	12.4	62.0	12.6	<0.001*
BMI, kg/m2	26.3	4.2	25.5	3.6	23.5	3.8	<0.001*
Waist, cm	91.6	10.4	88.2	10.4	80.8	10.3	<0.001*
Total fat mass, g	22070.3	7830.1	21249.5	7023.3	18651.4	7009.0	<0.001*
Total lean mass, g	43235.6	8362.4	43121	8812.0	40009.3	8652.7	<0.001*
Total region fat, %	32.1	7.8	31.6	7.9	30.3	8.6	0.026*
Total tissue fat, %	33.3	8.0	32.8	8.1	31.5	8.8	0.035*
Fat body weight, %	31.7	7.7	31.2	7.9	29.8	8.6	0.024*
Limb fat body weight, %	11.6	3.7	12.1	3.9	12.9	4.3	<0.001*
Trunk fat body weight, %	19.0	4.7	18.0	4.5	15.8	4.9	<0.001*
Lean body weight, %	63.3	7.3	63.7	7.6	64.8	8.4	0.078
Limb lean body weight, %	27.3	3.5	28.2	4.0	28.5	4.2	0.018*
Trunk lean body weight, %	31.1	4.0	30.4	3.8	31.0	4.2	0.067
Limb in fat, %	36.2	5.3	38.4	5.6	43.0	6.0	<0.001*
Trunk in fat, %	60.0	5.4	57.9	5.7	52.8	6.4	<0.001*
Limb in lean, %	43.2	2.6	44.3	2.6	44.0	2.5	<0.001*
Trunk in lean, %	49.1	2.3	47.8	2.1	47.8	2.0	<0.001*

BMI, body mass index.

*Statistical significance as P < 0.05.

aKruskal–Wallis H test

Table 3

Body composition markers that related obesity to blood glucose levels (n = 1358).

Markers, unit	Pearson correlation (glucose)	P value	Regression coefficient (glucose)a		P value
			Beta	95% CI
Weight, kg	0.143	<0.001	0.03496	(0.0109 to 0.05902)	0.004*
BMI, kg/m2	0.184	<0.001	0.01184	(0.00366 to 0.02001)	0.005*
Waist, cm	0.268	<0.001	0.04082	(0.0197 to 0.06193)	<0.001*
Total fat mass, g	0.109	<0.001	15.5139	(-2.3243 to 33.3521)	0.088
Total lean mass, g	0.121	<0.001	21.27296	(10.17281 to 32.3731)	<0.001*
Total region fat, %	0.035	0.204	0.00194	(-0.01467 to 0.01855)	0.819
Total tissue fat, %	0.032	0.241	0.00197	(-0.01511 to 0.01905)	0.821
Fat body weight, %	0.036	0.188	0.00002	(-0.00015 to 0.00019)	0.802
Limb fat body weight, %	-0.119	<0.001	-0.00007	(-0.00013 to -0.00001)	0.019*
Trunk fat body weight, %	0.164	<0.001	0.00012	(0.00001 to 0.00024)	0.040*
Lean body weight, %	-0.022	0.416	0.00002	(-0.00012 to 0.00015)	0.798
Limb lean body weight, %	-0.059	0.030	-0.00004	(-0.00009 to 0.00002)	0.183
Trunk lean body weight, %	0.037	0.177	0.00007	(-0.00001 to 0.00014)	0.090
Limb in fat, %	-0.325	<0.001	-0.00026	(-0.00034 to -0.00018)	<0.001*
Trunk in fat, %	0.327	<0.001	0.00035	(0.00023 to 0.00046)	<0.001*
Limb in lean, %	-0.097	<0.001	-0.00007	(-0.00012 to -0.00003)	0.002*
Trunk in lean, %	0.173	<0.001	0.00009	(0.00005 to 0.00013)	<0.001*

aTested by GLM and adjusted age, sex, systolic blood pressure, diastolic blood pressure, triglycerides, HDL, and iSES level. (n = 1277).

Body composition markers on HbA1c level

Based on the HbA1c levels, participants were categorized into three groups of ≥6.5%, 5.7–6.4%, and <5.7% (Table 4). Similar to the results for FG levels, groups with lower HbA1c levels exhibited greater decrease in weight, BMI, waist circumference, total fat mass, total lean mass, total region fat, and tissue fat. Limb in fat showed moderate negative (r = -0.342, P <0.001) and trunk in fat revealed positive correlations with HbA1c (r = 0.329, P <0.001), respectively (Table 5). With increasing levels of HbA1c, our regression analyses consistently revealed decrease of limb weight and increase of trunk weight for both lean and fat mass.

Table 4

Body composition markers that related obesity to HbA1c levels (n = 1358).

Markers, units	HbA1c (%)						P valuea
	≥6.5 (n = 177)		5.7–6.4 (n = 530)		<5.7 (n = 651)
	Mean	SD	mean	SD	Mean	SD
Weight, kg	68.8	13.6	65.2	12.6	63.15	12.79	<0.001*
BMI, kg/m2	26.3	4.1	24.9	3.8	23.63	3.82	<0.001*
Waist, cm	91.5	10.3	86.0	10.8	81.14	10.46	<0.001*
Total fat mass, g	22055.8	7556.8	20274.6	7194.3	18954.11	7027.83	<0.001*
Total lean mass, g	43231.7	8498.8	41433.3	8912.2	40842.31	8753.02	0.003*
Total region fat, %	32.2	7.6	31.4	8.5	30.20	8.32	0.003*
Total tissue fat, %	33.4	7.8	32.6	8.7	31.42	8.56	0.006*
Fat body weight, %	31.7	7.6	30.9	8.4	29.84	8.36	0.01*
Limb fat body weight, %	11.6	3.7	12.2	4.1	12.91	4.22	<0.001*
Trunk fat body weight, %	18.9	4.4	17.5	4.9	15.79	4.79	<0.001*
Lean body weight, %	63.2	7.2	63.8	8.1	64.96	8.13	0.004*
Limb lean body weight, %	27.4	3.4	28.1	4.0	28.75	4.18	<0.001*
Trunk lean body weight, %	30.8	4.0	30.6	4.1	30.92	4.05	0.177*
Limb in fat, %	36.3	5.2	39.4	5.7	43.13	6.20	<0.001*
Trunk in fat, %	59.9	5.3	56.6	6.0	52.82	6.58	<0.001*
Limb in lean, %	43.4	2.7	44.0	2.5	44.22	2.62	0.008*
Trunk in lean, %	48.7	2.4	48.0	2.0	47.61	2.10	<0.001*

BMI, body mass index.

*Statistical significance as P < 0.05.

Kruskal–Wallis H test

Table 5

Body composition markers that related obesity to HbA1c levels (n = 1358).

Markers, units	Pearson correlation (HbA1c)	P value	Regression coefficient (HbA1c)a		P value
			Beta	95% CI
Weight, kg	0.096	<0.001	0.84539	(0.12748 to 1.5633)	0.021*
BMI, kg/m2	0.165	<0.001	0.33533	(0.09016 to 0.58051)	0.007*
Waist, cm	0.242	<0.001	0.96860	(0.33674 to 1.60046)	0.003*
Total fat mass, g	0.082	0.003	196.49335	(-315.76508 to 708.75177)	0.452
Total lean mass, g	0.073	0.007	615.02581	(281.70332 to 948.3483)	<0.001*
Total region fat, %	0.036	0.180	-0.20021	(-0.69218 to 0.29177)	0.425
Total tissue fat, %	0.033	0.225	-0.21164	(-0.71762 to 0.29434)	0.412
Fat body weight, %	0.033	0.220	-0.00230	(-0.00728 to 0.00268)	0.366
Limb fat body Weight, %	-0.128	<0.001	-0.00363	(-0.00548 to -0.00178)	<0.001*
Trunk fat body weight, %	0.164	<0.001	0.00208	(-0.00129 to 0.00546)	0.226
Lean body weight, %	-0.029	0.279	0.00166	(-0.00244 to 0.00576)	0.427
Limb lean body weight, %	-0.084	0.002	-0.00047	(-0.00223 to 0.00129)	0.599
Trunk lean body weight, %	0.043	0.113	0.00239	(0.00007 to 0.00471)	0.043*
Limb in fat, %	-0.342	<0.001	-0.00902	(-0.01155 to -0.00648)	<0.001*
Trunk in fat, %	0.329	<0.001	0.01088	(0.00745 to 0.0143)	<0.001*
Limb in lean, %	-0.137	<0.001	-0.00189	(-0.0033 to -0.00048)	0.009*
Trunk in lean, %	0.206	<0.001	0.00250	(0.00123 to 0.00378)	<0.001*

Tested by GLM and adjusted age, sex, systolic blood pressure, diastolic blood pressure, triglycerides, HDL, and iSES level. (n = 1277)

First, on excluding patients of diabetes mellitus with medical treatment, the effects of body composition markers on different blood glucose and HbA1c levels were found coherent to our primary analyses, as mentioned previously (S1–S3 Tables). Second, 236 and 1122 participants were respectively categorized into with IR and non-IR groups (S4 Table). Consistent to the participants with higher FG or HbA1c, the IR group showed increased values of weight, BMI, waist, total fat mass, and markers of limb in fat (%), and trunk in fat (%). With increasing values of TyG-WC index, the participants showed decrease of lean mass weight and increase of fat mass weight (S5 Table).

Body composition parameter patterns explained by fasting glucose and HbA1c

RRR analyses revealed that FG and HbA1c shared similar factor loadings (Table 6). Both FG and HbA1c exhibited higher positive factor loadings (explaining the positive correlation between the observed variables of body composition and latent variable of FG or HbA1c) for the trunk in lean percentage (0.2983 and 0.3093, respectively) and the trunk in fat percentage (0.4282 and 0.3909, respectively) and higher negative loadings for the legs in lean percentage (−0.2595 and −0.2864, respectively) and the legs in fat percentage (−0.4514 and −0.4545, respectively). The total lean mass and total fat mass in grams also showed a similar pattern.

Table 6

RRR loading of blood glucose and HbA1c levels.

Body Area Loading	Unit: percentage (%)		Unit: gram
	Glucose (mg/dL)	HbA1c (%)	Glucose (mg/dL)	HbA1c (%)
Total lean and fat mass	(a) Total body composition (%)		(b) Total body composition
Arms in lean	0.0583	0.2030*	0.1375	0.0564
Legs in lean	−0.2595*	0.0915	−0.0266	−0.2864*
Trunk in lean	0.2983*	0.4056*	0.3274*	0.3093*
Android in lean	0.3479*	0.4742*†	0.4363*†	0.3705*
Gynoid in lean	−0.1803	0.1374	0.0416	−0.1715
Arms in fat	0.0130	0.2089*	0.3165*	0.1135
Legs in fat	−0.4514*†	−0.1476	−0.2006*	−0.4545*†
Trunk in fat	0.4282*†	0.4478*†	0.4740*†	0.3909*
Android in fat	0.3946*	0.5168*†	0.5362*†	0.3461*
Gynoid in fat	−0.3791*	−0.0952	−0.1686	−0.4004*†
Stratification by lean mass	(c) Total lean mass (%)		(d) Total lean mass (g)
Arms	0.1041	0.3001*	0.2435*	0.0957
Legs	−0.4636*†	0.1352	−0.0470	−0.4858*†
Trunk	0.5330*†	0.5994*†	0.5797*†	0.5247*†
Android	0.6216*†	0.7008*†	0.7727*†	0.6285*†
Gynoid	−0.4575*†	0.2031*	0.0737	−0.2908*
Stratification by fat mass	(e) Total fat mass (%)		(f) Total fat mass (g)
Arms	0.0157	0.2837*	0.3835*	0.1406
Legs	−0.5447*†	−0.2004*	−0.2431*	−0.5627*†
Trunk	0.5167*†	0.6082*†	0.5744*†	0.4840*†
Android	0.4762*†	0.7019*†	0.6497*†	0.4285*†
Gynoid	−0.4575*†	−0.1293	−0.2043*	−0.4958*†
(g) Fat amount per lean (Ratio of Fat mass/Lean mass)
Arms	0.1290	0.3198*
Legs	−0.3863*	−0.3083*
Trunk	0.5319*†	0.5599*†
Android	0.6422*†	0.6104*†
Gynoid	−0.3725*	−0.3414*

*Factor loading > 0.2.

†Factor loading > 0.4.

On stratification by lean mass, FG and HbA1c still exhibited the highest positive loadings on the android area [(0.7008 and 0.7727 in grams, respectively) and (0.6216 and 0.6285 in percentage, respectively)] and high negative loadings on the legs (−0.4636 and −0.4858 in percentage, respectively) and the gynoid area (−0.4575 and −0.2908 in percentage, respectively). On stratification by fat tissue, FG and HbA1c still exhibited the highest positive loading on the android area [(0.4762 and 0.4285 in percentage, respectively) and (0.7019 and 0.6497 in grams, respectively)] and a high negative loading on the legs (−0.5447 and −0.5627 in percentage, respectively) and the gynoid area (−0.4575 and −0.4958 in percentage, respectively). For the index of fat amount per lean mass ratio (fat mass/lean mass), FG and HbA1c shared high positive loadings on the trunk (0.5319 and 0.5599, respectively) and the android area (0.6422 and 0.6104, respectively) and negative loadings on the legs (−0.3863 and −0.3083, respectively) and the gynoid area (−0.3414 and −0.3725, respectively).

Discussion

This study offered comprehensive exploration for the association between body composition parameters, measured through DXA, and normal glycemic, prediabetic, and diabetic status, defined by FG and HbA1c levels, with a nationwide sampling ethnic Chinese population. On analyzing the body composition, we found that trunk fat positively correlated with increasing FG and HbA1c levels. On the contrary, lower extremity weight (in both fat and lean mass) showed negatively correlation with FG and HbA1c levels.

Our study consistently showed two major patterns between anthropometric parameters and glycemic status: (a) increasing levels of FG and HbA1c increased trunk fat mass and trunk lean mass, and (b) increasing lower extremity weight (either in fat or in lean mass) conferred decreased levels of FG and HbA1c. For the first pattern, this was consistent to the finding that central obesity was more common in type 2 diabetes [3, 5, 8]. Compared to the second pattern, previous studies showed inconsistent results. A study through analyzing the NHANES 1999–2004 found no significant association in limb lean mass and glycemic status on adjusting age, body mass index, ethnicity, smoking, alcohol, and physical activity [27]. In contrast, some studies found that lean mass is significantly negatively associated with insulin resistance among prediabetic patients [28, 29]. We considered that this difference reflect the different severity of insulin resistance among individuals from the sampling population. Lean mass in limbs may be an alternative to insulin resistance as a body composition marker. Lastly, our analyses should also support the evidence supporting the hypothesis that increasing limb mass decreased risk of diabetes. To validate our findings, we performed a confirmatory RRR analysis.

The RRR was similar to both exploratory principal component analysis and factor analysis, it is superior to these traditional dimension reduction methods because it involves determination of linear functions of the predicting factors (body composition parameters) by maximizing the explained variation in the response variables (blood glucose and HbA1c) [23]. Because RRR employs both information of the response variables (FG and HbA1c) and predicting factors (body composition parameters), it represents a posteriori method and thus validates the correlation between the response variables and predicting factors.

Additionally, the RRR analysis showed that individuals with higher FG and HbA1c levels accumulated more fat mass and increased lean mass in the android area, but less fat and lean mass in their lower extremities. We considered these findings were compatible to previous studies indicating functional differences of fat tissue between upper body (android area) and lower body [30-32]. Previous studies found the opposite association of upper and lower body fat with 2-hour post-load glucose level of oral glucose tolerance test [32] and diabetic risk [33, 34]. Recent studies indicated this difference was medicated by different adipokines [31] and site-specific sets of genes between upper and low body fat tissues [30, 35], respectively. On the other hand, the RRR also showed android lean mass was positively associated with FG and HbA1c. We found some studies showed the similar findings in their original data: diabetic participants with significantly higher trunk lean mass in comparison to non-diabetes [34, 36]. To our recognition, this should be cautious interpreted since DXA may severely overestimate trunk and android lean mass [37, 38]. When compared to magnetic resonance imaging (MRI) modality, a study found DXA overestimate 7% and 48% of trunk and android lean mass, respectively [38]. The association of trunk and android lean mass with FG and HbA1c should be reserved to answer with MRI.

The distinctive strengths of this study are as follows: (1) a nationwide population-based study investigating the association between body composition, measured through DXA, and differed glycemic status among healthy individuals in general communities, (2) a comprehensive exploration of the association between lean mass, fat mass, and blood glucose in different body regions, and (3) the validation of the findings through RRR. Besides, the sensitivity analysis helps confirm the major results and explain the mechanism between body composition markers and glycemic change was by insulin resistance.

The major study limitation, which was not addressed, was that a causal relationship between these body composition indexes and blood glucose levels could not be drawn by using the cross-sectional study design. However, the strong association and our sensitivity analysis results, based on the diagnosis of diabetes, were consistent. Thus, these findings suggest a direct correlation between body composition and blood glucose levels. Second, serum insulin was unmeasured in the NAHSIT 2013–2016. Our investigation of association between body composition and IR was replaced by the TyG-WC index. In fact, the recent studies suggested the TyG-WC index is a reliable alternative marker to homeostasis model assessment of insulin resistance (HOMA-IR) [24, 25]. Third, dynapenia or decreased muscle strength may influence glycemia status [39]. However, the strength was not measured and investigated in this study. Lastly, unlike the hospital-based study, the NAHSIT aimed to enroll healthy participants in the community. The unfavorable health outcome indicators, such as diabetic complications, were not explored in this survey.

Besides, numerous regions or countries in east Asia, like Japan, Korea, and China, was mono-ethnicity. In fact, more than 95% of residents in Taiwan were Han Chinese. A subethnic study in Taiwan also found no significant difference in the prevalence of diabetes and impaired glucose tolerance (IGT) between Taiwanese Aborigines and Han Chinese [40]. Therefore, no ethnic difference for on body composition, phenotype of type 2 diabetes, and glycemic status were explored in this study.

Conclusions

In conclusion, abdominal or trunk fat is found strongly associated with hyperglycemia for healthy adults in the general community. High fat and low lean mass percentage in abdomen increases the risk of hyperglycemia. Increasing lower extremity mass may confer lower risk of hyperglycemia. Furthermore, studies with physical activity and lifestyle intervention for increasing lower extremity mass could be implemented among the general population in the future.

The body composition markers that related obesity according to diabetes mellitus diagnosis.

(n = 1358)–sensitivity analysis.

(DOCX)

Click here for additional data file.

(n = 1358)–sensitivity analysis.

(DOCX)

Click here for additional data file.

The body composition markers that related obesity according to DM diagnosis.

(n = 1358)–sensitivity analysis.

(DOCX)

Click here for additional data file.

Body composition markers that related obesity to Triglyceride Glucose-Waist Circumference (TyG-WC) index.

(DOCX)

Click here for additional data file.

(DOCX)

Click here for additional data file.

6 Aug 2020

PONE-D-20-13486

Body Composition Patterns among Normal Glycemic, Pre-diabetic, Diabetic Health Chinese Adults in Community: NAHSIT 2013–2016

PLOS ONE

Dear Dr. Bai,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Sep 20 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Cheng Hu

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please refrain from stating p values as 0.000, either report the exact value or employ the format p<0.001.

3.Thank you for stating the following financial disclosure:

[The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.].

At this time, please address the following queries:

Please clarify the sources of funding (financial or material support) for your study. List the grants or organizations that supported your study, including funding received from your institution.

State what role the funders took in the study. If the funders had no role in your study, please state: “The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.”

If any authors received a salary from any of your funders, please state which authors and which funders.

If you did not receive any funding for this study, please state: “The authors received no specific funding for this work.”

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

4.We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

5. Please amend your list of authors on the manuscript to ensure that each author is linked to an affiliation. Authors’ affiliations should reflect the institution where the work was done (if authors moved subsequently, you can also list the new affiliation stating “current affiliation:….” as necessary).

6. Please upload a copy of Figure 1 and 2, to which you refer in your text on page 4. If the figure is no longer to be included as part of the submission please remove all reference to it within the text.

7. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript entitled "Body Composition Patterns among Normal Glycemic, Pre-diabetic, Diabetic Health Chinese Adults in Community: NAHSIT 2013–2016" reports the association between body composition parameters and normal glycemic, prediabetic, and diabetic status in Chinese adult population sampled for the nationwide survey. Although the association between body composition parameters and each glycemic status was clearly described, the authors should discuss some additional points to emphasize a rationale of this study.

Specific comments:

1. The authors should discuss the possible confounding factors of body composition and glycemic status. For example, smoking, alcohol and socioeconomic status could be confounding factors, especially in older adults (Osteoporosis and sarcopenia 2018;4:109-113, Lancet Diabetes Endocrinol 2015;3:958-967, Int J epidemiol 2011;40:804-818; J Diabetes Investig. 2020 Mar 29. doi: 10.1111/jdi.13260. etc).

2. The authors should consider the diabetic complications and anti-diabetes drugs to appropriately assess the relationship between body composition and glycemic status in patients with diabetes. Diabetic neuropathy and use of SU and glinides may affect skeletal muscle mass (J Am Med Dir Assoc 2016;17:896-901, J Diabetes Complications 2017;31:1465-1473, etc).

3. The authors should discuss not only muscle mass but also muscle strength as well as exercise capacity. Dynapenia could also influence glycemic status (J epidemiol 2015 ; 25 : 656-662); and the authors should discuss the possible relationship in the manuscript.

4. The authors should discuss possible effects of ethnic differences on body composition and glycemic status. In the previous study, it was suggested that ethnic differences may affect body composition, diet and physical activity (Womens Health(Lond) 2015;11:913-27).

5. Since the phenotype of type 2 diabetes differ from those of other ethnic groups (Lancet Diabetes Endocrinol. 2016 Jan;4(1):2-3; Curr Diab Rep. 2015 Jun;15(6):602; J Diabetes Investig. 2016 Apr;7 Suppl 1(Suppl 1):102-9 etc). The authors should discuss possible etnic differences in their observed results.

Reviewer #2: Major

1. In line 214-215, you are writing “lean mass…”. Therefore, not only comparison between glucose profile and body composition but also the comparison with an insulin level and HOMA-IR and the body composition is necessary.

2. In line 243-244, you are writing “low lean mass percentage in abdomen…”. Results of RRR show that trunk lean and fat mass have a positive association with a glucose profile. Trunk in fat and lean has a positive association with glucose profile, and limb in fat and lean seems to have a negative association with glucose profile. Therefore, the discussion about this difference seems to be necessary.

Minor

1. In line 85, you are writing “…assessment of the nutritional status”. What kind of item did you use for nutritional evaluation, blood nutritional items (albumin, cholinesterase etc.) or nutrition score (MNA, CONUT etc.)?

2. In the section of Data Collection of Body Composition (117-125), the calculating formula is (body composition/1000)/weight *100, isn't it?

3. In line 150, is the classification of the FG levels ≥126mg/dL, 100-126 mg/dL and <100mg/dL? It is necessary to conform them to data of Table2.

4. In line 161, did you mistake “≥6.5%” for “>6.5%”?

5. In line 205, you wrote "…increase trunk fat mass". But not only "trunk fat mass" but also "trunk lean mass" increased.

6. Explanation of total region fat, total tissue fat, android (in fat) and gynoid (in fat) is necessary in the method.

7. In Table 1, I think the data of Male is "number and ratio" not "mean and SD".

8. The unit of weight is “kg” not “cm” (Table2-5).

9. Please revise the data of SD from 10.41 to 10.4 in the item of the waist in Table 2.

10. Please revise “Statistical significance as P<0.5” to “Statistical significance as P<0.05” in Table 2.

11. I think that HbA1c is unnecessary for a title of Table3.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

7 Sep 2020

We thank the reviewers for their constructive comments. We have made revisions to the manuscript to address all the questions and comments raised by the three reviewers. We highlights changes made to the original version by setting the text color to red. Our specific responses to each comment are as follows:

Responses to reviewers #1:

� The manuscript entitled "Body Composition Patterns among Normal Glycemic, Pre-diabetic, Diabetic Health Chinese Adults in Community: NAHSIT 2013–2016" reports the association between body composition parameters and normal glycemic, prediabetic, and diabetic status in Chinese adult population sampled for the nationwide survey. Although the association between body composition parameters and each glycemic status was clearly described, the authors should discuss some additional points to emphasize a rationale of this study.

� We are grateful for all of your constructive comments. In the revised manuscript, we addressed the all of the additional points in the following sections.

� 1. The authors should discuss the possible confounding factors of body composition and glycemic status. For example, smoking, alcohol and socioeconomic status could be confounding factors, especially in older adults (Osteoporosis and sarcopenia 2018;4:109-113, Lancet Diabetes Endocrinol 2015;3:958-967, Int J epidemiol 2011;40:804-818; J Diabetes Investig. 2020 Mar 29. doi: 10.1111/jdi.13260. etc).

� Thank you for the comment. We agreed that smoking, alcohol, and socioeconomic status were possible confounding effects between body composition and glycemic status. We revised our manuscript by adding analyses for smoking, alcohol, and socioeconomic status.

� For smoking, we categorized participants into current, ever, and no smokers. The ever and current smoker were more common in the diabetic groups but without statistical significance. (Please see Table 1, line 165-171, results section)

� For alcohol, we obtained the average alcohol consumption (gram per day) and classified participants into non-drinker, <16, 16-35, 36-64, >64 g/day. Alcohol consumption between the diabetic and nondiabetic groups showed no significant difference. (Please see Table 1, line 165-171, results section)

� For socioeconomic status, we acquired z scores of personal income and education years. Individual socioeconomic status (iSES) was obtained by a sum of z scores of personal income and education years. Thereafter, the iSES was classified into the lowest, middle, and highest classes according to the tertiles of the z score. When compared to the diabetic group, the nondiabetic group had higher z score of iSES. (Please see and line 92-95, method section, and Table 1, line 165-171, results section)

� Accordingly, the regression analyses for body composition markers that related to blood glucose (Table 3) and HbA1c (Table 5) were adjusted for age, sex, systolic and diastolic blood pressure, triglycerides, HDL, and iSES level. (Please see line 199, Table 3, and line 217, Table 5, results section)

� 2. The authors should consider the diabetic complications and anti-diabetes drugs to appropriately assess the relationship between body composition and glycemic status in patients with diabetes. Diabetic neuropathy and use of SU and glinides may affect skeletal muscle mass (J Am Med Dir Assoc 2016;17:896-901, J Diabetes Complications 2017;31:1465-1473, etc).

� Thank you for the comment. We added the proportion of using anti-diabetes drugs in the diabetic group. For them, 50.9% (109/214) of the diabetic participants took anti-diabetic drugs. (Please see line 169-170, results section)

� On the other hand, the enrolled participants in the NAHSIT study were healthy subjects, or people who had chronic diseases, such as diabetes, but generally these participants were in good medication control and favorable functional status. Compared to hospital-based subjects, our participants with diabetes were in the early disease status. Most of these healthy participants has not developed the diabetic complication yet. Diabetic complications were categorized into marco-vascular and micro-vascular complications. Among them, diabetic neuropathy was one of the micro-vascular complication. In most condition, type 1 and type 2 diabetes need > 20 or > 7 years of clinical course to develop the microvascular complications.[1] Since the NAHSIT was a community-based study and focused on healthy subjects, data of diabetic complication such diabetic neuropathy was not collected in this survey. We addressed this limitation in the discussion section. (Please see line 324-326, discussion section)

� 3. The authors should discuss not only muscle mass but also muscle strength as well as exercise capacity. Dynapenia could also influence glycemic status (J epidemiol 2015 ; 25 : 656-662); and the authors should discuss the possible relationship in the manuscript.

� Thank you for the comment. Since muscle strength was not measured and the data was unavailable in the NAHSIT 2013-2016, we addressed this limitation in the discussion section. (Please see discussion section, line 322-323)

� 4. The authors should discuss possible effects of ethnic differences on body composition and glycemic status. In the previous study, it was suggested that ethnic differences may affect body composition, diet and physical activity (Womens Health(Lond) 2015;11:913-27).

� Thank you for the comment. We reviewed this literature and agreed the possible effects of ethnic difference (such as White, Black, and Chinese) on glycemic status and body composition. In fact, numerous countries or regions in east Asia were of a single ethnic group, such as China (91.6% of Han Chinese), South Korea (96% of Koreans), Japan (98% of Japanese), and Taiwan (95.4% of Han Chinese).

� In fact, more than 95% of residents were in Taiwan were Han Chinese, and only 2.3% were Taiwanese aborigines. Therefore, the participants characteristic of our NAHSIT study was mono-ethnicity.[2] In our revised manuscript, we added the description of mono-ethnic characteristic in Taiwan in our discussion section. (Please see discussion section, line 328-333)

� 5. Since the phenotype of type 2 diabetes differ from those of other ethnic groups (Lancet Diabetes Endocrinol. 2016 Jan;4(1):2-3; Curr Diab Rep. 2015 Jun;15(6):602; J Diabetes Investig. 2016 Apr;7 Suppl 1(Suppl 1):102-9 etc). The authors should discuss possible etnic differences in their observed results.

� Thank you for the comment. After reviewing these literatures, we agreed that the phenotype of type 2 diabetes or β-cell function and insulin sensitivity may vary by different ethnic groups, such as ethnic Caucasians, Japanese, Koreans, and Chinese. In Taiwan, 95.4% of residents were in fact Han Chinese, and 2.3% were Taiwanese aborigines (such as Ami, Atayal).[3] (Please see discussion section, line 300-305)

� A subethnic study in Taiwan for β-cell function and insulin sensitivity also found no significant difference in the prevalence of diabetes and impaired glucose tolerance (IGT) between Aborigines and Chinese.[4] (Please see discussion section, line 328-333)

Responses to reviewers #2:

Major:

� 1. In line 214-215, you are writing “lean mass…”. Therefore, not only comparison between glucose profile and body composition but also the comparison with an insulin level and HOMA-IR and the body composition is necessary.

� We are grateful for all of your constructive comments. Since insulin level was not available in the NAHSIT 2013-2016 survey, we addressed this limitation. (Please see line 305-307, in the discussion section)

� On the other hand, we used the triglyceride glucose-waist circumference (TyG-WC) index as a surrogate marker for insulin resistance[5, 6]. According to a recent large scale nutritional survey for Asian population, the TyG index and TyG-related markers were highly correlated with homeostasis model assessment of insulin resistance (HOMA-IR).[6]

� Of these TyG-related markers, we noticed that TyG-WC index had the highest diagnostic consistency to HOMA-IR for both male and female Asian participants. Therefore, we decided to conduct a sensitivity analyses for the association between body composition and insulin resistance with TyG-WC. A cut-off value of for TyG-WC ≥ 850 was defined as insulin resistance.

� The sensitivity analysis of the association between body composition markers and TyG-WC was shown in Table S4 and S5. With increasing values of TyG-WC index, the participants showed decrease of lean mass weight and increase of fat mass.

� (Please see line 150-159, methods section; line 226-231 results section; line 314-315, discussion section; line 321=326, discussion section; and Table S4-S5)

� 2. In line 243-244, you are writing “low lean mass percentage in abdomen…”. Results of RRR show that trunk lean and fat mass have a positive association with a glucose profile. Trunk in fat and lean has a positive association with glucose profile, and limb in fat and lean seems to have a negative association with glucose profile. Therefore, the discussion about this difference seems to be necessary.

� Thank you for the great comment. We addressed (1) the different functional characteristics fat tissue between upper body (android) and lower body, and (2) the paradoxically positive association of trunk lean mass with FG and HbA1c levels in the discussion section.

� First, we considered these findings were compatible to previous studies indicating functional differences of fat tissue between upper body (android area) and lower body.[7-9] Previous studies found the opposite association of upper and lower body fat with 2-hour post-load glucose level of oral glucose tolerance test[9] and diabetic risk[10, 11]. Recent studies indicated this difference was medicated by different adipokines[8] and site-specific sets of genes between upper and low body fat tissues[7, 12], respectively

� Second, the RRR also showed android lean mass was positively associated with FG and HbA1c. We found some studies showed the similar findings in their original data: diabetic participants with significantly higher trunk lean mass in comparison to non-diabetes.[11, 13] This should be cautious interpreted since DXA may severely overestimate trunk and android lean mass.[14, 15] When compared to magnetic resonance imaging (MRI) modality, a study found DXA overestimate 7% and 48% of trunk and android lean mass, respectively.[15] The association of trunk and android lean mass with FG and HbA1c should be reserved to answer with MRI. (Please see line 292-307, discussion section)

Minor:

� 1. In line 85, you are writing “…assessment of the nutritional status”. What kind of item did you use for nutritional evaluation, blood nutritional items (albumin, cholinesterase etc.) or nutrition score (MNA, CONUT etc.)?

� Thank you for the comment. Continued from our previous 2005-2008 Nutrition and Health Survey in Taiwan (NAHSIT)[16], the 2013-2016 NAHSIT assessed both the (1) behavior and (2) health outcome indicators.

� For behavior indicators, we collected information of socioeconomic items, diet behavior and belief (including dietary recall, food frequency and habits, dietary and nutritional knowledge).

� For health outcome indicators, we measured the body composition by dual-energy X-ray absorptiometry and clinical biochemistry, including items of serum cholesterol (including total cholesterol, LDL-C and HDL-C), triglycerides, blood glucose, uric acid, CRP, creatinine, liver function tests, amylase, complete blood count, vitamins, minerals, serum pH, pCO2 , pO2, iron and ferritin, TIBC, phospholipid, homocysteine, BUN, alkaline phosphatase, PTH, and DNA from white blood cells for genetic analysis. (Please see line 85-90 in the revised manuscript)

� 2. In the section of Data Collection of Body Composition (117-125), the calculating formula is (body composition/1000)/weight *100, isn't it?

� Thank you for the comment. We had revised the calculating formula. (Please see line 122-136)

� 3. In line 150, is the classification of the FG levels ≥126mg/dL, 100-126 mg/dL and <100mg/dL? It is necessary to conform them to data of Table2.

� Thank you for the comment. The classification of FG levels were of ≥126 mg/dL, 100–125 mg/dL, and <100 mg/dL, which was conform to data of Table 2 in the revised manuscript. (Please see line186, and line 196, Table 2, results section)

� 4. In line 161, did you mistake “≥6.5%” for “>6.5%”?

� Thank you for the comment. Th participants were categorized into three groups of ≥6.5%, 5.7–6.4%, and <5.7%. (Please see line 207, results section)

� 5. In line 205, you wrote "…increase trunk fat mass". But not only "trunk fat mass" but also "trunk lean mass" increased.

� Thank you for the comment. We revised the text with adding “trunk lean mass.” (Please see line 270-271, discussion section)

� 6. Explanation of total region fat, total tissue fat, android (in fat) and gynoid (in fat) is necessary in the method.

� Thank you for the comment. These terms were defined as the following.

• Total region fat (%): (Fat mass in limbs and trunk/ Body weight) × 100

• Total tissue fat (%): (Total fat mass/ Body weight) × 100

• Android in fat (%): (Android fat mass/Total fat mass) × 100

• Gynoid in fat (%): (Gynoid fat mass/Total fat mass) × 100

(Please see our revised manuscript, line 133-136, method section)

� 7. In Table 1, I think the data of Male is "number and ratio" not "mean and SD".

� Thank you for the comment. The unit for data of male was revised. (Please see our revised manuscript, Table 1, results section)

� 8. The unit of weight is “kg” not “cm” (Table2-5).

� Thank you for the comment. The unit of weight was corrected now. (Please see the revised manuscript, Table 2 to 5, results section)

� 9. Please revise the data of SD from 10.41 to 10.4 in the item of the waist in Table 2.

� Thank you for the comment. We revised the data of SD from 10.41 to 10.4 in the item of the waist in Table 2. (Please see Table 2, results section)

� 10. Please revise “Statistical significance as P<0.5” to “Statistical significance as P<0.05” in Table 2.

� Thank you for the comment. We made a revision as P <0.05 in Table 2. (Please see Table 2, results section)

� 11. I think that HbA1c is unnecessary for a title of Table3.

� Thank you for the comment. We deleted HbA1c for a title of Table 3. (Please see Table 3, results section)

References

1. Association AD. 11. Microvascular Complications and Foot Care:. Diabetes Care. 2020;43(Suppl 1):S135-S51. doi: 10.2337/dc20-S011. PubMed PMID: 31862754.

2. https://en.wikipedia.org/wiki/Monoethnicity.

3. https://en.wikipedia.org/wiki/Demographics_of_Taiwan.

4. Chen HD, Shaw CK, Tseng WP, Chen HI, Lee ML. Prevalence of diabetes mellitus and impaired glucose tolerance in Aborigines and Chinese in eastern Taiwan. Diabetes Res Clin Pract. 1997;38(3):199-205. doi: 10.1016/s0168-8227(97)00104-6. PubMed PMID: 9483387.

5. Zheng S, Shi S, Ren X, Han T, Li Y, Chen Y, et al. Triglyceride glucose-waist circumference, a novel and effective predictor of diabetes in first-degree relatives of type 2 diabetes patients: cross-sectional and prospective cohort study. J Transl Med. 2016;14:260. Epub 2016/09/07. doi: 10.1186/s12967-016-1020-8. PubMed PMID: 27604550; PubMed Central PMCID: PMCPMC5015232.

6. Lim J, Kim J, Koo SH, Kwon GC. Comparison of triglyceride glucose index, and related parameters to predict insulin resistance in Korean adults: An analysis of the 2007-2010 Korean National Health and Nutrition Examination Survey. PLoS One. 2019;14(3):e0212963. Epub 2019/03/07. doi: 10.1371/journal.pone.0212963. PubMed PMID: 30845237; PubMed Central PMCID: PMCPMC6405083.

7. Karpe F, Pinnick KE. Biology of upper-body and lower-body adipose tissue--link to whole-body phenotypes. Nat Rev Endocrinol. 2015;11(2):90-100. Epub 2014/11/04. doi: 10.1038/nrendo.2014.185. PubMed PMID: 25365922.

8. Wu H, Qi Q, Yu Z, Sun Q, Wang J, Franco OH, et al. Independent and opposite associations of trunk and leg fat depots with adipokines, inflammatory markers, and metabolic syndrome in middle-aged and older Chinese men and women. J Clin Endocrinol Metab. 2010;95(9):4389-98. Epub 2010/06/02. doi: 10.1210/jc.2010-0181. PubMed PMID: 20519350.

9. Snijder MB, Dekker JM, Visser M, Bouter LM, Stehouwer CD, Yudkin JS, et al. Trunk fat and leg fat have independent and opposite associations with fasting and postload glucose levels: the Hoorn study. Diabetes Care. 2004;27(2):372-7. doi: 10.2337/diacare.27.2.372. PubMed PMID: 14747216.

10. Tatsukawa Y, Misumi M, Kim YM, Yamada M, Ohishi W, Fujiwara S, et al. Body composition and development of diabetes: a 15-year follow-up study in a Japanese population. Eur J Clin Nutr. 2018;72(3):374-80. Epub 2018/01/23. doi: 10.1038/s41430-017-0077-7. PubMed PMID: 29362458.

11. Lee JS, Auyeung TW, Leung J, Kwok T, Leung PC, Woo J. The effect of diabetes mellitus on age-associated lean mass loss in 3153 older adults. Diabet Med. 2010;27(12):1366-71. doi: 10.1111/j.1464-5491.2010.03118.x. PubMed PMID: 21059088; PubMed Central PMCID: PMCPMC3059762.

12. Pinnick KE, Nicholson G, Manolopoulos KN, McQuaid SE, Valet P, Frayn KN, et al. Distinct developmental profile of lower-body adipose tissue defines resistance against obesity-associated metabolic complications. Diabetes. 2014;63(11):3785-97. Epub 2014/06/19. doi: 10.2337/db14-0385. PubMed PMID: 24947352.

13. Kim KS, Park KS, Kim MJ, Kim SK, Cho YW, Park SW. Type 2 diabetes is associated with low muscle mass in older adults. Geriatr Gerontol Int. 2014;14 Suppl 1:115-21. doi: 10.1111/ggi.12189. PubMed PMID: 24450569.

14. Buckinx F, Landi F, Cesari M, Fielding RA, Visser M, Engelke K, et al. Pitfalls in the measurement of muscle mass: a need for a reference standard. J Cachexia Sarcopenia Muscle. 2018;9(2):269-78. Epub 2018/01/19. doi: 10.1002/jcsm.12268. PubMed PMID: 29349935; PubMed Central PMCID: PMCPMC5879987.

15. Rankin KC, O'Brien LC, Gorgey AS. Quantification of trunk and android lean mass using dual energy x-ray absorptiometry compared to magnetic resonance imaging after spinal cord injury. J Spinal Cord Med. 2019;42(4):508-16. Epub 2018/02/20. doi: 10.1080/10790268.2018.1438879. PubMed PMID: 29461936; PubMed Central PMCID: PMCPMC6718191.

16. Tu SH, Chen C, Hsieh YT, Chang HY, Yeh CJ, Lin YC, et al. Design and sample characteristics of the 2005-2008 Nutrition and Health Survey in Taiwan. Asia Pac J Clin Nutr. 2011;20(2):225-37. PubMed PMID: 21669592.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

9 Oct 2020

Body Composition Patterns among Normal Glycemic, Pre-diabetic, Diabetic Health Chinese Adults in Community: NAHSIT 2013–2016

PONE-D-20-13486R1

Dear Dr. Bai,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Cheng Hu

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: Thank you for revision and comment. I confirmed contents again.

There are no additional question and revision.

This paper is acceptable.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

22 Oct 2020

PONE-D-20-13486R1

Body composition patterns among normal glycemic, pre-diabetic, diabetic health Chinese adults in community: NAHSIT 2013–2016

Dear Dr. Bai:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Cheng Hu

Academic Editor

PLOS ONE