Higher childhood weight gain, lower skeletal muscle mass, and higher cereal consumption in normal-weight Japanese women with high-percentage trunk fat: a subanalysis study

Abstract

Normal-weight but high-percentage trunk fat phenotype was characterized in a setting where adiposity is not associated with educational and socioeconomic status. Body size trajectory since birth, current body composition measured using whole-body dual-energy X-ray absorptiometry, cardiometabolic traits, serum adipokines, and dietary intake were measured cross-sectionally in 251 normal weight Japanese female university students whose fasting triglyceride and homeostasis model assessment-insulin resistance (HOMA-IR) averaged 56 mg/dL and 1.2, respectively. They were grouped according to tertile of percentage trunk fat. Although HOMA-IR did not differ among three groups, high-percentage trunk fat was associated with higher triglyceride and apolipoprotein B, and lower HDL cholesterol and apolipoprotein A1. In multivariate logistic regression analyses, weight-adjusted skeletal muscle mass (OR: 0.13, 95% CI: 0.04–0.38, p < 0.001), weight gain from birth to age 12 years (OR: 1.214、95% CI: 1.008–1.463、p = 0.04), and cereal consumption (OR:1.008, 95% CI: 1.000–1.016, p = 0.04) were associated with high-percentage trunk fat independent of birthweight, HOMA-IR, adipose tissue-insulin resistance index (the product of fasting insulin and free fatty acid), triglyceride, HDL cholesterol, apolipoprotein A1 and B, leptin, adiponectin, blood pressure, and high-sensitivity C-reactive protein. Early childhood growth, lower skeletal muscle mass, and higher cereal consumption may be associated with normal-weight but high-percentage trunk fat phenotype in Japanese female university students in this subanalysis study. Atherogenic profile of lipids and apolipoproteins may be directly related to abdominal fat accumulation.

Introduction

In epidemiological and clinical settings, general obesity is usually defined by body mass index (BMI) and central or abdominal obesity by waist circumference, waist-to-hip ratio, or waist-to-height ratio [1]. Normal weight (BMI: 18.5–24.9 kg/m²) is associated with a healthy cardiometabolic risk profile and a low risk of type 2 diabetes and cardiovascular disease [2]. However, normal-weight individuals who have several cardiometabolic risk factors, typically referred to as metabolically unhealthy, normal weight (MUNW) [3]. In addition, some normal weight people with central obesity (NWCO) show a cardiometabolic risk profile similar to obese people [4–6]. NWCO has been shown to be a higher risk of cardiovascular events and all-cause mortality [1, 7–11]. Therefore, it is important to understand the underlying risk factors and mechanisms.

Studies have shown that nutrition and environmental stressors during perinatal period can induce long-term adaptations that increase risk of diabetes and cardiovascular disease [12]. Li et al. reported that birthweight and genetics may play important roles in predicting the MUNW phenotype in children [13]. We [14] have reported in young normal-weight Japanese women that birthweight was associated not only with lower appendicular muscle mass but also with lower gluteofemoral fat mass, both of which appear to be characteristic body composition of a MUNW phenotype [3]. Viitasalo et al. have shown that an increase in BMI from childhood to adulthood may be associated with MUNW in adults aged 24–39 years [15, 16].

Rice consumption may be associated with lower waist circumference in a country where eating rice is not common [17]. In contrast, in a country where eating rice is common, rice consumption may be associated with higher waist circumference or metabolic syndrome in women [18–20]. Studies on dietary intake in NWCO are lacking.

There are limited studies which employed accurate technology such as abdominal computed tomography scanning or dual-energy X-ray absorptiometry (DXA) to assess visceral or truncal fat in normal weight individuals. Although we found three studies in middle-aged people employing computed tomography scanning [21–23] and a single study in postmenopausal women using DXA [9] as discussed later, we did not find studies in young individuals which employed accurate technology. Therefore, we explore the associations between early-life factors, dietary factors, body composition using DXA, cardiometabolic risk factors, and the normal-weight but high-abdominal fat accumulation phenotype in Japanese female university students, a population in which adiposity is not associated with educational and socioeconomic status, that is reported to be associated with adiposity [24].

Materials and methods

This subanalysis study investigated 251 normal-weight female Japanese students among 307 female Japanese students of Department of Food Sciences and Nutrition, Mukogawa Women’s University, who were recruited as volunteers whose details were reported previously [25]. Among 251 women, 181 and 166 women provided data on weight trajectory and dietary intake, respectively. There was no significant difference in current body composition and cardiometabolic risk factors between women with and without these data (data not shown). We excluded women with clinically diagnosed acute or chronic diseases, those on hormonal contraception, and those on a diet to lose weight from the study.

Biological mothers of participants were asked to report weight at birth, and height and weight at age 12 and 15 of their daughters, which were depicted in either maternal health check notes or child health notebook records (issued by each municipal office). Among 251 participants, 181 reported birthweight data and 167 reported height and weight at age 12 and 15.

Participants underwent blood sampling, measurement of anthropometric indices, blood pressure, and body composition after 12-h overnight fasting as previously described [26–28]. Brachial blood pressure was measured using an automated sphygmomanometer (BP-203RV II, Colin, Tokyo, Japan) after participants were seated at least for 5 min. A standard 75 g oral glucose tolerance test (OGTT) was done with multiple postload glucose and insulin measurements over a 30–120-min period in 99 women. The area under the response curve of glucose (AUCg) and serum insulin (AUCi) was calculated by the trapezoidal method. Plasma glucose, serum insulin, triglyceride (TG), cholesterol, high-density lipoprotein (HDL) cholesterol, apolipoprotein A1 and B (ApoA1 and ApoB, respectively), free fatty acid, leptin, adiponectin, alanine-aminotransferase (ALT), and high-sensitivity C-reactive protein (hsCRP) were measured as previously reported [26, 27]. Low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald’s formula. Adipose tissue-insulin resistance index (AT-IR, the product of fasting insulin and free fatty acid) and homeostasis model assessment-insulin resistance (HOMA-IR) were calculated as previously reported [27, 28]. The ratio of leptin to adiponectin was calculated as a marker of compromised adipose tissue function [29].

Body composition was measured using whole-body dual-energy X-ray absorptiometry (DXA) (Hologic QDR-2000 software version 7.20D, Waltham, MA) [26]. General adiposity was assessed using BMI, percentage body fat (%BF), and fat mass index (FMI), the last of which was calculated as body fat mass in kg divided by height in meter squared. Waist circumference and percentage trunk fat (% trunk fat) were considered as markers of abdominal fat accumulation. Muscle mass were evaluated by relative appendicular skeletal muscle mass (ASM) as percentage of body mass (%ASM) and absolute ASM index (ASM/height² in kg/m²). %ASM has reported to be a better predictor of insulin resistance than ASM or ASM index [30].

One hundred and sixty-six women reported dietary intake using the self-administered diet history questionnaire [31], which is a validated 16-page structured questionnaire, which assesses dietary habits in the preceding one-month period. The questionnaire consists of major cooking methods, amount of six alcoholic beverages, semiquantitative portion size of 121 selected food and nonalcoholic beverage items, dietary supplements, consumption frequency and amount of cereals (rice, bread, noodles and other wheat foods). This has been widely used in Japan and its validity with respect to commonly studied nutrition factors has been confirmed.

Data were presented as mean ± SD. Due to deviation from normal distribution, fasting insulin, HOMA-IR and hsCRP were logarithmically transformed for analyses. Because there is no clearly established cut point of % trunk fat for central obesity and because there was no woman whose waist circumstance exceeded 90 cm, a cutoff point for abdominal obesity in women according to Japan Society for the Study of Obesity [32], participants were grouped into three groups according to tertile of % trunk fat: low, median, and high % trunk fat. Differences among three groups were analyzed by analysis of variance and then Bonferroni's multiple comparison procedure. Dietary data were analyzed by Kruskal–Wallis test. Stepwise multivariate logistic regression analyses were used to identify most important determinants of high % trunk fat. Independent variables included were variables which showed significant difference among three groups: weight gain until 12 years, %ASM, TG, HDL cholesterol, ApoA1, ApoB, leptin, adiponectin, hsCRP and consumption of cereals (model A). As both weight transition and dietary intake were available in 92 women, the final model included 92 participants. There was no difference between 92 and remaining 159 women in the percentage of the low (30.4 vs. 34.6 %), median (33.7 vs. 32.1 %), and high (35.9 vs. 33.3 %) tertile of % trunk fat (p=0.79). A two-tailed p<0.05 was considered statistically significant. All calculations were performed with SPSS system 23.0 (SPSS Inc, Chicago, IL).

Results

As previously reported [25], current BMI averaged 20.5 kg/m², waist 71.5 cm, and % trunk fat 28.8 % in 251 normal-weight Japanese women. Their fasting TG, HOMA-IR, and ALT averaged 56 mg/dL, 1.2, and 12.9 U/L, respectively.

There was no difference in birthweight and height at 12, 15 and 20 years (Table 1). However, weight at 12 and 15 years and, hence, weight gain until 12 years and BMI at 12 and 15 years were higher in women with high % trunk fat compared with other two groups of women (Fig. 1). BMI, waist, percentage leg and body fat, and FMI increased in a stepwise fashion from the low to high %TF. %ASM decreased in a stepwise fashion from the low to the high %TF although there was no difference in ASM index.

Table 1.

Weight trajectories since birth and current body composition in young women grouped according to tertile of trunk fat percentage

	Low tertile	Median tertile	High tertile	p values	#
Range: trunk fat (%)	17.5–26.3	26.4–31.3	31.4–43.4
Age (years)	20.6 ± 1.2	20.6 ± 1.2	20.5 ± 1.3	0.830
Birthweight (g)	3241 ± 421	3180 ± 378	3174 ± 373	0.590
Weight (kg)
at 12 years	42.2 ± 5.8	42.3 ± 5.2	45.6 ± 7.3	0.004	b, c
at 15 years	47.9 ± 4.8	48.6 ± 5.1	51.7 ± 6.1	< 0.001	b, c
at 20 yeas	49.6 ± 4.0	51.4 ± 4.4	54.7 ± 4.7	< 0.001	a, b, c
Δ Body weight0–12 (kg)	39.0 ± 5.7	39.1 ± 5.2	42.4 ± 7.3	0.003	b, c
Height (cm)
at 12 years	149.6 ± 6.0	151.7 ± 6.2	151.5 ± 6.5	0.155
at 15 years	156.4 ± 5.3	157.0 ± 5.3	157.7 ± 4.3	0.380
at 20 years	159.0 ± 5.4	158.6 ± 5.1	159.1 ± 4.4	0.805
BMI (kg/m2)
at 12 years	18.7 ± 2.0	18.4 ± 1.6	19.8 ± 2.3	0.001	b, c
at 15 years	19.6 ± 1.4	19.7 ± 1.5	20.8 ± 2.1	< 0.001	b, c
at 20 years	19.6 ± 0.9	20.4 ± 1.2	21.6 ± 1.4	< 0.001	a, b, c
Waist (cm)	68.9 ± 4.6	71.6 ± 4.1	74.1 ± 4.9	< 0.001	a, b, c
Leg fat (%)	27.6 ± 3.3	30.7 ± 3.2	34.5 ± 4.3	< 0.001	a, b, c
Trunk fat (%)	23.4 ± 2.2	28.6 ± 1.5	35.5 ± 2.8	< 0.001	a,b,c
Body fat (%)	23.6 ± 2.3	27.9 ± 1.9	33.2 ± 2.8	< 0.001	a, b, c
ASM (%)	30.3 ± 1.5	29.1 ± 1.4	26.8 ± 1.6	< 0.001	a, b, c
ASMI (kg/m2)	5.9 ± 0.4	5.9 ± 0.5	5.8 ± 0.5	0.077
FMI (kg/m2)	4.6 ± 0.6	5.6 ± 0.5	7.1 ± 0.9	< 0.001	a, b, c

Mean ± SD

n = 83 or 84 in each group, except for birth weight and height and weight at 12 and 15 years, where number ranges 51–68

Δ Body weight0–12: weight gain from birth to 12 years

BMI body mass index, ASM appendicular skeletal muscle mass, ASMI ASM index, FMI fat mass index

#p < 0.05 or less by Bonferroni's multiple comparison procedure

a low vs. median, b low vs. high, c median vs. high

Fig. 1.

Birthweight, weight at age 12 years, weight gain from birth to 12 years of age (ΔBody weight_0–12), and daily intake of cereals in women with the low, median, and high tertile of percentage trunk fat (blue, yellow, and red squares, respectively). Mean ± SD of n = 83 or 84, except for cereal intake (see Table 3). Means not sharing the same alphabet are significantly different with each other at p < 0.05 or less by Bonferroni’s multiple comparison procedure

Despite a remarkable increase in % trunk fat, fasting glucose, insulin, ALT, HOMA-IR, and AT-IR did not differ among three groups (Table 2). Glucose and insulin concentrations after oral glucose loading did not differ as well (Table 3). However, serum and LDL cholesterol, TG and ApoB were higher and HDL cholesterol and ApoA1 were lower in women with high % trunk fat compared with the other two groups (Fig. 2). Serum leptin showed a stepwise increase (Fig. 3). Adiponectin was lower and log hsCRP was higher in high % trunk fat than in the other two groups. There was a stepwise increase in leptin/adiponectin ratio (Fig. 3). Blood pressure did not differ (Table 2).

Table 2.

Cardiometabolic risk factors in young women grouped according to tertile of percentage trunk fat

	Tertile of percentage trunk fat			p values	#
	Low	Median	High
Fasting glucose (mg/dL)	82 ± 7	84 ± 7	83 ± 6	0.119
Fasting insulin (μU/mL)	5.7 ± 2.9	5.9 ± 3.2	6.5 ± 3.8	0.353
HOMA-IR	1.1 ± 0.6	1.2 ± 0.8	/	0.469
AT-IR	3.4 ± 3.1	3.4 ± 3.1	3.5 ± 2.4	0.980
ALT (U/L)	12.1 ± 4.8	12.6 ± 4.5	13.8 ± 10.1	0.259
Triglyceride (mg/dL)	52 ± 21	54 ± 20	62 ± 32	0.021	b
Cholesterol (mg/dL)	179 ± 25	184 ± 25	185 ± 29	0.225
HDL cholesterol (mg/dL)	77 ± 12	77 ± 13	71 ± 14	0.011	b, c
LDL cholesterol (mg/dL)	92 ± 20	96 ± 21	101 ± 26	0.019	b
Apolipoprotein A1 (mg/dL)	166 ± 18	166 ± 20	159 ± 21	0.023	b
Apolipoprotein B (mg/dL)	66 ± 12	70 ± 13	74 ± 17	0.002	b, c
Leptin (ng/mL)	6.3 ± 1.9	8.7 ± 3.0	10.8 ± 4.1	< 0.001	a, b, c
Adiponectin (µg/mL)	12.0 ± 4.4	12.0 ± 4.1	10.3 ± 3.8	0.008	b, c
hsCRP (μg/dL)	19 ± 58	26 ± 63	41 ± 89	0.139
log hsCRP	0.92 ± 0.42	1.02 ± 0.48	1.20 ± 0.52	0.001	b, c
Systolic BP (mmHg)	105 ± 8	106 ± 11	107 ± 10	0.496
Diastolic BP (mmHg)	60 ± 7	62 ± 7	61 ± 7	0.325

Mean ± SD

n = 83 or 84 in each group

AT-IR adipose tissue-insulin resistance, HOMA-IR homeostasis model assessment-insulin resistance, ALT alanine-aminotransferase, hsCRP high-sensitivity C-reactive protein, BP blood pressure

^#The same as in Table 1

Table 3.

Glucose and insulin concentrations during a 75 g oral glucose tolerance test in young women grouped according to tertile of percentage trunk fat

	Tertile of percentage trunk fat			p values
	Low (n = 29)	Median (n = 34)	High (n = 36)
Fasting glucose (mg/dL)	82 ± 7	84 ± 7	83 ± 6	0.119
30-min glucose (mg/dL)	121 ± 20	118 ± 26	120 ± 21	0.882
60-min glucose (mg/dL)	108 ± 32	99 ± 34	107 ± 36	0.537
120-min glucose (mg/dL)	89 ± 21	90 ± 25	98 ± 24	0.247
Fasting insulin (μU/mL)	5.7 ± 2.9	5.9 ± 3.2	6.5 ± 3.8	0.236
30-min insulin (μU/mL)	47 ± 34	60 ± 41	50 ± 27	0.275
60-min insulin (μU/mL)	44 ± 29	44 ± 23	40 ± 24	0.682
120-min insulin (μU/mL)	34 ± 16	40 ± 22	44 ± 33	0.268
AUCg (mg/dL/2 h)	206 ± 36	199 ± 46	209 ± 46	0.606
AUCi (μU/mL/2 h)	75 ± 40	85 ± 36	78 ± 39	0.567

Mean ± SD Fasting glucose and insulin were the same in Table 2

AUCg and AUCi the area under the response curve of glucose and insulin, respectively

Fig. 2.

HDL cholesterol, triglyceride, apolipoprotein A1 and B in women with the low, median, and high tertile of percentage trunk fat (blue, yellow, and red squares, respectively). Mean ± SD of n = 83 or 84. See Fig. 1 legend for statistical significance

Fig. 3.

Serum leptin, adiponectin, leptin/adiponectin ratio, and logarithmically transformed high-sensitivity C-reactive protein (log hsCRP) in women with the low, median, and high tertile of percentage trunk fat (blue, yellow, and red squares, respectively). Mean ± SD of n = 83 or 84. See legend for Fig. 1 for statistical significance

Daily intake of carbohydrate was tended to be higher, and intake of cereals was significantly higher in women with high % trunk fat than in the low tertile (Table 4, Fig. 1). However, there was no difference in daily intake of energy, protein, fat and fibers among three groups of women.

Table 4.

Daily dietary intake in young women grouped according to tertile of percentage trunk fat

	Tertile of percentage trunk fat			p values	#
	Low	Median	High
	(n = 57)	(n = 56)	(n = 53)
Energy (kcal)	1789 ± 442	2002 ± 531	2092 ± 1139	0.119
Energy (kcal/kg/body weight)	35.9 ± 8.4	38.9 ± 10.3	38.5 ± 21.0	0.274
Carbohydrate (g)	238 ± 61	272 ± 83	288 ± 173	0.048
Protein (g)	60 ± 15	66 ± 18	68 ± 30	0.257
Fat (g)	63 ± 19	69 ± 22	70 ± 37	0.566
Carbohydrate (%)	53.3 ± 5.2	54.0 ± 5.9	54.9 ± 6.2	0.544
Protein (%)	13.5 ± 1.6	13.4 ± 2.0	13.2 ± 1.6	0.812
Fat (%)	31.5 ± 4.5	30.9 ± 4.9	30.1 ± 5.6	0.582
SFA (g)	17.8 ± 5.6	19.7 ± 6.9	20.5 ± 12.7	0.405
MUFA (g)	21.5 ± 6.9	23.6 ± 7.9	23.9 ± 11.9	0.378
PUFA (g)	13.6 ± 4.3	14.6 ± 4.8	14.3 ± 6.3	0.548
Fibers (g)	11.8 ± 4.2	13.8 ± 5.2	13.4 ± 7.4	0.09
Cereals (g)	337 ± 103	399 ± 140	433 ± 223	0.001	a, b

Mean ± SD

SFA, MUFA and PUFA saturated, monounsaturated and polyunsaturated fatty acids, respectively

^#The same as in Table 1

We had multivariate logistic regression analyses for high % trunk fat as a dependent variable (Table 5), which included as independent variables, anthropometric, cardiometabolic, dietary and other variables, that showed significant differences among three groups of women: weight gain until 12 years, %ASM, TG, HDL cholesterol, ApoA1, ApoB, leptin, adiponectin, hsCRP, and consumption of cereals (model A). High % trunk fat was independently associated with %ASM, weight gain until 12 years and consumption of cereals. Results did not change when leptin/adiponectin ratio was further included. In model B where ASMI was included instead of %ASM, ASMI, ApoB, and leptin were associated with high % trunk fat tertile in addition to weight gain until 12 years and consumption of cereals.

Table 5.

Multivariate logistic regression analysis for the high tertile of percentage trunk fat

	Odds ratio	95% CI		p values
		Lower	Upper
Model A
%ASM	0.131	0.045	0.380	0.000
Δ Body weight0–12	1.214	1.008	1.463	0.041
Cereal consumption	1.008	1.000	1.016	0.044
Model B
Apolipoprotein B	1.057	1.010	1.107	0.018
Leptin	1.268	1.022	1.572	0.031
ASMI	0.206	0.054	0.783	0.02
Δ Body weight0-12	1.108	1.007	1.220	0.036
Cereal consumption	1.005	1.000	1.010	0.044

Other independent variables included: Model A: triglyceride, HDL cholesterol, apolipoprotein A1, leptin, adiponectin and hsCRP. Model B: ASMI was included instead of %ASM

Abbreviations are the same as in Tables 1 and 2

Discussion

The current study has demonstrated that high % trunk fat tertile was associated with early childhood growth and lower skeletal muscle mass in Japanese female university students. In addition, even in young women whose fasting triglyceride and HOMA-IR averaged 56 mg/dL and 1.2, respectively, normal weight but high % trunk fat phenotype was associated not only with higher TG and lower HDL cholesterol as previously reported [5, 6], but also with higher ApoB and lower ApoA1. Although markers of insulin resistance (fasting insulin, HOMA-IR, and AT-IR) did not differ among three groups of women despite a stepwise increase in % trunk fat, a reliable and accurate estimate of abdominal fat accumulation, leptin to adiponectin ratio, a marker of compromised adipose tissue function [29], was higher in normal weight women with high % trunk fat. We confirmed previously reported positive association of trunk fat with hsCRP, a marker of systemic low-grade inflammation [33], in the present study. Finally, normal weight but high % trunk fat phenotype was associated with higher consumption of cereals. Among these variables, early childhood growth, lower skeletal muscle mass, and higher consumption of cereals were independently associated with normal weight but high % trunk fat phenotype.

There is an increasing number of studies on NWCO [1, 4–11, 34–37]. However, studies are limited which employed accurate technology to assess visceral or truncal fat in normal weight individuals. We found a single study using DXA [9], which found that higher percent trunk fat was associated with increased risk of cardiovascular disease in 2683 postmenopausal women with normal BMI. To the best of our knowledge, the present study may be the first to investigate normal weight, high abdominal fat accumulation phenotype in a young female population using DXA, a sophisticated measures of body composition. There are three studies which defined central obesity as visceral fat areas ≥100 cm² measured by computed tomography scans in normal weight middle-aged individuals [21–23]. Consistent with the present study, NWCO was associated with higher triglyceride [21–23] and lower adiponectin [22, 23].

Studies have shown that a MUNW phenotype was associated with birthweight and a rapid increase in BMI from childhood to adulthood [13, 15, 16]. The Bogalusa Heart Study has shown that rapid increase in BMI during childhood (aged 4–19 years) were associated with BMI and waist in young adults [38]. Sutharsan et al. reported that rapid weight gain in the first 5 years of life was associated with BMI and waist in 21-year-old adults [39]. The latter two studies may be in line with the present observation that normal weight but high % trunk fat phenotype was associated with weight gain from birth to 12 years in young women although it was not associated with birthweight in the present study.

The liver is well known to produce and release glucose and TG in very-low density lipoprotein particles. In contrast, skeletal muscle is a site of clearance of circulating TG and production of HDL particles in the fasting state [40] Further, skeletal muscle is a major site of insulin-mediated glucose disposal in the postprandial state [41]. As a consequence, in addition to waist circumference, low skeletal muscle mass assessed by %ASM has been shown to be associated with insulin resistance, type 2 diabetes, and metabolic syndrome [42–44]. Therefore, it seems reasonable to assume that a combination of higher trunk fat and lower skeletal muscle mass may contribute to atherogenic profile of lipids and apolipoproteins, and systemic low-grade inflammation in young women with high % trunk fat. Studies have shown that low weight-adjusted muscle mass was observed in people with MUNW, whose waist was higher than metabolically healthy normal weight individuals [45–47].

In women with high % trunk fat, waist circumference averaged 74.1 cm and ALT 13 U/L, suggesting a minimum TG accumulation, if any, in the trunk and the liver, respectively. In addition, their serum TG and HOMA-IR averaged 62 mg/dL and 1.3, respectively, and there was no difference in markers of insulin resistance among three groups of women. However, high % trunk fat was associated with higher leptin to adiponectin ratio, a marker of compromised adipose tissue function [29]. These observations may suggest that atherogenic profile of lipids, lipoproteins and apolipoproteins may be directly related to abdominal fat accumulation although these differences are very small from the clinical point of view.

Azadbakht et al. [18] reported a significant positive association between the prevalence of central obesity and rice intake among Iranian female adolescents. Ahn et al [19]. examined the rice-eating patterns in relation to the risk for central obesity and found positive association between the two in premenopausal and postmenopausal women. In the present study, cereals’ consumption was associated with high % trunk fat in young Japanese women. Approximately 70 % of cereals consumed by young Japanese women were rice [48] although a decreasing trend of rice intake in Japan in recent decades. A systematic review and meta-analysis have shown that white rice consumption was associated with a higher prevalence and/or incidence of metabolic syndrome, particularly in Eastern Asian and Iranian populations, where rice is the staple [49]. In contrast, a study from USA reported that rice consumers had lower waist circumference and were significantly more likely to have a normal BMI [17]. In Western and Mediterranean countries, it has been proposed that central fat distribution may be related to diets with a high ratio of saturated to unsaturated fatty acids [50].

The accurate and reliable measures of general and central fat accumulation by DXA are the strength of the present study. There are several limitations such as a single measurement of biochemical variables, the cross-sectional design and relatively small sample size. In addition, multivariate logistic regression analyses were done in a smaller sample size than the total population. Statistical power and sample size were not calculated. As participants were young Japanese women, the results may not be generalized to other gender, age populations, races or ethnicities. However, this might be considered as a strength because participants were female university students, a homogeneous population including educational and socioeconomic status as previously reported in detail previously [26].

In conclusion, normal-weight but high-percentage trunk fat phenotype may be associated with early childhood growth, lower skeletal muscle mass, and higher cereal consumption in young Japanese women. Atherogenic profile of lipids and apolipoproteins may be directly related to abdominal fat accumulation.

Data availability

The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

The authors declare that there is no duality of interest associated with this manuscript.

Ethical Approval

The study was approved by the Ethics Committees of the Mukogawa Women’s University (No. 07-28 on 19/02/2008) and followed the tenets of the Declaration of Helsinki.

Informed consent

All participants gave written informed consent after the experimental procedure had been explained.