Basic characteristics of the study population
Descriptive differences of basic characteristics of the participants were presented in Table 1. There were 4,934 males (45.3%) and 5,964 females (54.7%) among the 10,898 participants. The average BMI and WC was 24.4 kg/m2 and 82.6 cm in males, 24.0 kg/m2 and 81.0 cm in females, respectively. The intake of dietary energy, fats, proteins, and carbohydrates was 2273.9 kcal/d, 91.5 g/d, 74.0 g/d, and 282.4 g/d in males, and 1919.4 kcal/d, 77.7 g/d, 62.3 g/d, and 242.0 g/d in females, respectively. Some variables were observed to have statistically significant differences between the sexes, such as age, income, smoking, drinking, energy and macronutrients intake (p < 0.001). Males consumed more calories, dietary fats, proteins, carbohydrates than females (p < 0.001), and had higher BMI and WC than females (p < 0.001).
<Table 1 > Demographic characteristics of the sample
Table 1. Baseline characteristics of the study population in CHNS(2015) by genders1,2
|
General Characteristic
|
|
Female(N=5,964)
|
|
Men(N=4,934)
|
p-Value
|
Age,n (%)
|
|
|
|
|
< 0.001
|
18-44 years
|
|
2318(38.9)
|
|
1753(35.5)
|
|
45-64 years
|
|
3646(61.1)
|
|
3181(64.5)
|
|
Education ,n (%)
|
|
|
|
|
< 0.001
|
Primary/illiterate
|
|
1752(29.4)
|
|
933(18.9)
|
|
Middle school
|
|
2122(35.6)
|
|
1929(39.1)
|
|
High/above
|
|
2090(35.0)
|
|
2072(42.0)
|
|
Income , n (%)
|
|
|
|
|
0.005
|
Low
|
|
1883(31.6)
|
|
1429(29.0)
|
|
Medium
|
|
1963(32.9)
|
|
1630(33.0)
|
|
High
|
|
2118(35.5)
|
|
1875(38.0)
|
|
Geogerphical region , n (%)
|
|
|
0.437
|
Rural
|
|
3727(62.5)
|
|
3119(63.2)
|
|
Urban
|
|
2237(37.5)
|
|
1815(36.8)
|
|
PA level, n (%)
|
|
|
|
|
< 0.001
|
Low
|
|
1923(32.2)
|
|
1708(34.6)
|
|
Medium
|
|
2108(35.3)
|
|
1536(31.1)
|
|
High
|
|
1933(32.4)
|
|
1690(34.3)
|
|
Smoking , n (%)
|
|
|
|
|
< 0.001
|
Ever/Never
|
|
5870(98.4)
|
|
2155(43.7%)
|
|
Current
|
|
94(1.6)
|
|
2779(56.3%)
|
|
Drinking , n (%)
|
|
|
|
|
< 0.001
|
Ever/Never
|
|
5563(93.3)
|
|
2153(43.6)
|
|
Current
|
|
401(6.7)
|
|
2781(56.4)
|
|
BMI category, n (%)
|
|
|
|
|
< 0.001
|
Thin
|
|
291(4.9)
|
|
183(3.7)
|
|
Normal
|
|
2931(49.1)
|
|
2220(45.0)
|
|
Overweight
|
|
1933(32.4)
|
|
1820(36.9)
|
|
Obesity
|
|
809(13.6)
|
|
711(14.4)
|
|
BMI (kg/m2)
|
|
24.0±0.1
|
|
24.4±0.1
|
< 0.001
|
WC(cm)
|
|
81.0±0.2
|
|
86.2±0.2
|
< 0.001
|
Dietary intake
|
|
|
|
|
|
Daily energy (k cal)
|
|
1919.4±8.9
|
|
2273.9±11.3
|
< 0.001
|
Fat (g)
|
|
77.7±0.6
|
|
91.5±0.7
|
< 0.001
|
Protein (g)
|
|
62.3±0.3
|
|
74.0±0.4
|
< 0.001
|
Carbohydrate (g)
|
|
242.0±1.4
|
|
282.4±1.8
|
< 0.001
|
Fat (% E3)
|
|
35.9±0.2
|
|
36.0±0.2
|
0.819
|
Protein (% E 4)
|
|
13.2±0.1
|
|
13.2±0.1
|
0.197
|
Carbohydrate (% E5)
|
|
50.7±0.2
|
|
49.8±0.2
|
< 0.001
|
Note: 1Data for categorical variable expressed as number (%); 2Values are mean± s.e. for continuous variables. % E3,4,5 means the percentage of energy intake from fat, protein and carbohydrate, respectively.
Dietary intake by social demographic characteristics
The distribution of dietary intake by social demographic characteristics in different sexes was presented in Table 2. There were significant differences in the energy intake between different income and PA groups in females (p < 0.001) and between different regions and PA groups in males (p < 0.001). Dietary fat intake significantly increased as income levels increased (p < 0.001; 94.9 g in males and 80.5 g in females). Dietary protein intake in females aged 18-44 years (62.5 g) was significantly higher than that in females aged 45-64 years (61.8 g) (p < 0.001). Additionally, there were significant differences in dietary protein intake among different education levels, income levels, and regions in both sexes (p < 0.001). Males belonging to urban areas had the highest protein intake of 77.4 g. Dietary carbohydrate intake varied significantly across social demographic characteristics, and the highest levels of PA showed the most intake (296.3 g in males and 256.6 g in females; p < 0.001).
<Table 2 > Distribution of energy intake and macronutrient composition
There were significant differences in distribution of energy from macronutrients as per BMI categories by sex differentiation (Table 3). Males showed greater absolute intakes of energy and macronutrients as compared to females in the BMI categories comparison (p < 0.001). In the normal BMI category, the percentage of energy intake from carbohydrate in females (50.7%) was significantly higher than that in males (49.9%) (p < 0.001). In the overweight BMI category, the percentage of energy intake from carbohydrates in females (50.6%) was significantly higher than that in males (49.5%), while the percentage of energy intake from proteins in females (13.1%) was significantly lower than that in males (13.4%)(p < 0.001).
<Table 3 >Energy and macronutrients consumption in different weight outcomes by sex
Table 2 The distribution of daily energy and macronutrient intake by social-demographic characteristics in 20151,2
|
General Characteristic
|
Women (N=5,964)
|
|
Men(N=4,934)
|
Energy (k cal)
|
Fat(g)
|
Protein(g)
|
Carbohydrate(g)
|
|
Energy (k cal)
|
Fat(g)
|
Protein(g)
|
Carbohydrate (g)
|
Age
|
|
|
|
|
|
|
|
|
|
18–44 years
|
1931.4±14.5
|
77.8±0.9
|
62.5±0.5 **
|
244.5±2.2*
|
|
2300.6±19.5
|
91.8±1.2
|
74.7±0.7
|
290.8±3.2 ***
|
45–64 years
|
1911.7±11.3
|
77.7±0.7
|
61.8±0.4
|
240.4±1.7
|
|
2259.2±13.9
|
91.3±0.9
|
73.6±0.5
|
277.7±2.1
|
Education level
|
|
|
|
|
|
|
|
|
|
Primary/illiterate
|
1915.3±16.4
|
76.5±1.0
|
59.0±0.6***
|
246.9±2.6***
|
|
2258.7±26.2
|
90.8±1.6
|
69.2±0.9***
|
283.5±4.3 ***
|
Middle school
|
1938.8±15.6
|
77.3±1.0
|
61.5±0.6
|
248.4±2.5
|
|
2297.1±19.1
|
91.1±1.1
|
73.6±0.7
|
288.5±3.0
|
High/above
|
1903.0±14.5
|
79.2±0.6
|
65.4±0.6
|
231.3±2.0
|
|
2259.1±16.5
|
92.1±1.4
|
76.6±0.6
|
276.1±2.5
|
Income level
|
|
|
|
|
|
|
|
|
|
Low
|
1964.4±17.4***
|
78.8±1.1***
|
60.8±0.6***
|
251.9±2.7***
|
|
2300.9±22.4
|
90.9±1.3***
|
71.4±0.7***
|
293.4±3.7 ***
|
Medium
|
1894.7±15.0
|
73.7±0.9
|
60.4±0.5
|
246.6±2.5
|
|
2260.3±19.7
|
88.1±1.2
|
72.4±0.7
|
287.7±3.2
|
High
|
1903.1±14.1
|
80.5±1.0
|
64.8±0.5
|
228.9±2.0
|
|
2265.1±17.5
|
94.9±1.1
|
77.4±0.7
|
269.3±2.5
|
Geogerphical region
|
|
|
|
|
|
|
|
|
Rural
|
1914.9±11.3
|
74.9±0.7***
|
59.9±0.4***
|
249.4±1.8 ***
|
|
2295.2±14.5**
|
90.4±0.9***
|
72.4±0.5***
|
290.9±2.3 ***
|
Urban
|
1927.0±14.6
|
82.4±1.0
|
65.8±0.6
|
229.6±2.0
|
|
2237.2±18.0
|
93.3±1.1
|
76.8±0.7
|
267.6±2.7
|
PA level
|
|
|
|
|
|
|
|
|
|
Low
|
1892.7±16.0***
|
77.8±1.1
|
61.9±0.6
|
235.4±2.4***
|
|
2258.8±19.8***
|
91.7±1.2
|
73.7±0.7
|
279.0±3.0***
|
Medium
|
1890.9±14.4
|
77.7±0.9
|
62.3±0.5
|
234.6±2.2
|
|
2221.8±18.8
|
90.8±1.1
|
74.3±0.7
|
270.7±3.0
|
High
|
1976.9±16.0
|
77.5±1.0
|
62.1±0.6
|
256.6±2.6
|
|
2336.4±20.0
|
91.9±1.2
|
74.0±0.7
|
296.3±3.2
|
Note: 1 Values are mean± s.e. for continuous variables .2 ***, ** and * indicate statistical significance at the 1%, 5%and 10% level, respectively.
Table 3 Dietary energy and macronutrient intake among different BMI categories by gender in 20151,2
|
Dietary intake
|
|
The thin BMI category
|
|
The normal BMI category
|
|
The overweight BMI category
|
|
The obesity BMI category
|
|
Women
(N=291)
|
Men
(N=183)
|
|
Women
(N=2931)
|
Men
(N=2220)
|
|
Women
(N=1933)
|
Men
(N=1820)
|
|
Women
(N=809)
|
Men
(N=711)
|
Energy
(k cal)
|
|
1887.7±40.0
|
2296.8±65.7***
|
|
1908.0±12.7
|
2234.7±16.8***
|
|
1948.2±16.1
|
2290.3±18.4***
|
|
1903.2±23.1
|
2348.2±30.2***
|
Fat
(g)
|
|
76.8±2.7
|
92.4±4.1***
|
|
77.0±0.8
|
90.0±1.1***
|
|
79.0±1.0
|
92.4±1.1***
|
|
77.4±1.6
|
93.6±1.8***
|
Protein
(g)
|
|
60.6±1.4
|
71.4±2.1***
|
|
62.1±0.5
|
72.2±0.6***
|
|
62.9±0.6
|
75.5±0.7***
|
|
60.7±0.8
|
76.6±1.1***
|
Carbohydrate
(g)
|
|
237.9±6.0
|
290.0±10.3***
|
|
240.5±1.9
|
277.7±2.6***
|
|
245.0±2.5
|
282.6±2.9***
|
|
240.1±3.4
|
294.1±4.7***
|
Fat
(% E3)
|
|
35.9±0.7
|
36.1±1.0
|
|
35.8±0.2
|
35.9±0.3
|
|
36.1±0.3
|
36.1±0.3
|
|
35.9±0.4
|
35.6±0.4
|
Protein
(% E4)
|
|
13.0±0.2
|
12.8±0.2
|
|
13.3±0.1
|
13.2±0.1
|
|
13.1±0.1
|
13.4±0.1**
|
|
13.2±0.1
|
13.0±0.1
|
Carbohydrate
(% E5)
|
|
50.9±0.7
|
50.4±1.0
|
|
50.7±0.2
|
49.9±0.3 **
|
|
50.6±0.3
|
49.5±0.3**
|
|
51.0±0.4
|
50.3±0.5
|
Note: 1 Values are mean (s.e.) for continuous variables and n (%) for categorical variable.2 ***, ** and * indicate statistical significance at the 1%, 5%and 10% level, respectively. % E3,4,5 means the percentage of energy intake from fat, protein and carbohydrate, respectively.
Figure 1a shows a significant difference in the percentage of energy intake from fat between males and females within the DRIs recommended range (20-30%). Females in the underweight, overweight obesity BMI categories, and abdominal obesity groups had significantly higher percentage of energy intake from fat than males, in line with the DRIs (20-30%) (p < 0.001). Moreover, with weight gain in females, the percentage of energy intake from fat was significantly lower than the DRIs recommended range (20-30%) (p < 0.001). The proportion of percentage of energy intake from fat over the DRIs standard (>30%) was significantly different between males and females (Figure 1b). Specifically, in the underweight and the normal BMI category, females who exceeded the DRI standard (>30%) was significantly more than that of males, as opposed to the overweight/obese BMI categories and abdominal obesity groups (p < 0.001). There was a significant difference in males and females with respect to the percentage of energy intake from carbohydrates in line with the DRIs standard (50-65%) (Figure 1c). Females in the normal/overweight/obesity BMI categories, and abdominal obesity groups had significantly higher percentage of energy intake from carbohydrates than males, in line with the DRIs (50-65%), as opposed to the underweight group. Figure 1c suggested that the proportion of the percentage of energy intake from carbohydrates below the recommended DRIs (<50%) increased significantly with weight gain (p < 0.001) Furthermore, males below the recommended DIRs (<50%) (Figure 1d) were significantly more than that of females in the normal/overweight/obese BMI categories, and abdominal obesity groups, as opposed to the underweight group (p < 0.001).
Figure 1. Energy intake from fat and carbohydrate compared with the DRIs standard in subgroups with different weight outcomes
< Figure 1a > the proportion of dietary energy from fat within the recommended values (20-30%E) among body weight outcomes by sex
< Figure 1b > the proportion of dietary energy from fat beyond the recommended values (>30%E) among body weight outcomes by sex
< Figure 1c > the proportion of dietary energy from carbohydrate within the recommended values (50-65% E) among body weight outcomes by sex
< Figure 1d > the proportion of dietary energy from carbohydrate below the recommended values (<50%E) among body weight outcomes by sex
As shown in Table 4, the associations between energy and macronutrient consumption and BMI were estimated using quantile regression. From the adjusted model for males, significant coefficients for BMI were observed at the 25 th, 50 th, 75 th, and 95 th dietary energy quantiles (p < 0.05), at the 75th and 95 th dietary fat quantiles (p < 0.05), at the 5 th, 25 th, 50 th, and 75 th dietary protein quantiles (p < 0.05), and at the 75th and 95th dietary carbohydrate quantiles (p < 0.05). Furthermore, the increase in dietary intake was higher at the upper end of the distribution, suggesting that males with higher BMI had an additional dietary intake than individuals with lower BMI. Moreover, from our findings, 10% increase in BMI would result in an additional intake of 0.002-0.004 kcal/d of dietary energy, 0.032-0.057 g/d of dietary fat, 0.039-0·084 g/d of dietary protein, and 0.018-0.028 g/d of dietary carbohydrate across all the quantiles in males (p < 0.05). In females, the quantile regression coefficient of dietary intake in all BMI quantile percentages was not statistically significant.
<Table 4 > Association between energy macronutrient intakes and BMI
With respect to our findings stated in Table 5, significant coefficients for WC in males were observed at the 75th and 95th dietary energy quantiles (p < 0.05) and at the 75 th and 95 th dietary protein quantiles (p < 0.05); whereas, in females, the statistically significant coefficients were observed at the 25 th and 50 th quantiles dietary carbohydrate quantiles (p < 0.05). A 10% increase in WC would result in an additional intake of 0.004-0.008 kcal/d of dietary energy and 0.051-0.052 g/d of dietary carbohydrates across all quantiles for males (p < 0.05), and additional intake of 0.060-0.150 kcal/d of dietary fat in females (p < 0.05). On the other hand, it would result in reduced intake of dietary carbohydrate of 0.051-0.052 g/d across all quantiles for females (p < 0.05).
<Table 5 > Association between energy macronutrient intakes and WC
Table4 Quantile regression estimates for the association between energy intake and macronutrient composition and BMI (kg/m2) in CHNS 20151,2,3
|
Dietary intake
|
|
MEN
|
|
WOMEN
|
5th
|
25th
|
50th
|
75th
|
95th
|
|
5th
|
25th
|
50th
|
75th
|
95th
|
Energy (kcal)
|
Model 1
|
|
0.0001
|
0.0001*
|
0.0025**
|
0.0026***
|
0.0046***
|
|
0.0001
|
0.0001
|
0.0001
|
-0.0001
|
0.0001
|
|
Model 2
|
|
0.0001
|
0.0002**
|
0.0003**
|
0.0004***
|
0.0004**
|
|
0.0001
|
0.0001
|
0.0001
|
0.0001
|
-0.0001
|
Fat (g)
|
Model 1
|
|
0.0023
|
0.0016
|
0.0031
|
0.0025
|
0.0049***
|
|
0.0009
|
0.0014
|
0.0014
|
-0.0017
|
-0.0007
|
|
Model 2
|
|
0.0030
|
0.0020
|
0.0022
|
0.0032***
|
0.0057***
|
|
0.0021
|
0.0010
|
0.0010
|
0.0006
|
-0.0035
|
Protein (g)
|
Model 1
|
|
0.0056*
|
0.0100***
|
0.0107***
|
0.0085***
|
0.0088
|
|
0.0033
|
0.0022
|
0.0016
|
-0.0044
|
-0.0056
|
|
Model 2
|
|
0.0039**
|
0.0059**
|
0.0086***
|
0.0084***
|
0.0081
|
|
0.0031
|
0.0027
|
0.0025
|
-0.0001
|
-0.0001
|
Carbohydrate (g)
|
Model 1
|
|
-0.0003
|
0.0004
|
0.0009***
|
0.0013***
|
0.0025**
|
|
0.0002
|
0.0005
|
0.0007
|
-0.0005
|
-0.0003
|
|
Model 2
|
|
0.0001
|
0.0009
|
0.0016*
|
0.0018**
|
0.0028***
|
|
0.0003
|
0.0007
|
0.0001
|
0.0004
|
-0.0001
|
Fat (% E4)
|
Model 1
|
|
0.0009
|
0.0005
|
0.0013
|
0.0020
|
-0.0099
|
|
0.0001
|
0.0001
|
0.0030
|
-0.0020
|
-0.0021
|
|
Model 2
|
|
0.0015
|
0.0006
|
0.0039
|
0.0022
|
0.0133
|
|
0.0071
|
-0.0003
|
0.0044
|
-0.0015
|
-0.0150
|
Protein (% E5)
|
Model 1
|
|
0.0489**
|
0.0565***
|
0.0427**
|
0.0202*
|
-0.0593**
|
|
0.0155
|
-0.0045
|
-0.0242
|
-0.0410
|
-0.0520
|
|
Model 2
|
|
0.0256
|
0.0180
|
0.0069
|
0.0172
|
-0.0623
|
|
0.020
|
0.0156
|
0.0060
|
-0.0160
|
-0.0373
|
Carbohydrate(% E6)
|
Model 1
|
|
-0.0152
|
-0.0043
|
-0.0024
|
0.0014
|
0.0028
|
|
-0.0040
|
-0.0010
|
-0.0005
|
0.0043
|
0.0042
|
|
Model 2
|
|
-0.0019
|
0.0007
|
0.0040
|
0.0024
|
0.0173
|
|
-0.0066
|
0.0003
|
-0.0050
|
0.0036
|
0.0164
|
Note: 1 Model 1 included energy intake and macronutrient composition in the quantile regression model to investigate the association with BMI/waist circumference.
2 Model 2 adjusted for age (18-44y and 45-64y), education (primary/illiterate, junior and high school/above), income (low, medium and high), region (urban/rural) and physical activity (low, medium and high), current smoker (yes/no), current drinker (yes/no).
3 ***, ** and * indicate statistical significance at the 1%, 5%and 10% level, respectively.
% E4, 5, 6 means the percentage of energy intake from fat, protein and carbohydrate, respectively.
Table 5 Quantile regression estimates for the association between energy intake and macronutrient composition and WC(cm) in CHNS 20151,2,3
|
Dietary intake
|
|
MEN
|
|
WOMEN
|
5th
|
25th
|
50th
|
75th
|
95th
|
|
5th
|
25th
|
50th
|
75th
|
95th
|
Energy (kcal)
|
Model 1
|
|
-0.0008
|
-0.0001
|
0.0005
|
0.0007**
|
0.0011**
|
|
-0.0012
|
-0.0001
|
-0.0001
|
-0.0005
|
-0.0004
|
|
Model 2
|
|
-0.0006
|
0.0004*
|
0.0004*
|
0.0007**
|
0.0008**
|
|
-0.0014
|
-0.0001
|
-0.0002
|
-0.0003
|
-0.0001
|
Fat (g)
|
Model 1
|
|
-0.0103
|
0.0056*
|
0.0086**
|
0.0041
|
0.0062
|
|
-0.0190
|
0.0010
|
0.0001
|
0.0072
|
0.0140
|
|
Model 2
|
|
-0.0027
|
0.0082
|
0.0041
|
0.0050
|
0.0045
|
|
-0.0117
|
0.0001
|
0.0005
|
0.0060*
|
0.0150**
|
Protein (g)
|
Model 1
|
|
-0.0200
|
0.0139**
|
0.0194***
|
0.0170**
|
0.0238
|
|
-0.0500
|
-0.0166
|
-0.0061
|
-0.0189
|
-0.0217
|
|
Model 2
|
|
-0.0238
|
0.0107
|
0.0105
|
0.0131
|
0.0252
|
|
-0.0385
|
0.0001
|
-0.0084
|
-0.0108
|
-0.0098
|
Carbohydrate (g)
|
Model 1
|
|
-0.0044
|
-0.0001
|
-0.0001
|
0.0042*
|
0.0064*
|
|
-0.0061
|
-0.0007
|
-0.0001
|
-0.0006
|
0.0034
|
|
Model 2
|
|
-0.0049
|
0.0009
|
0.0023
|
0.0052**
|
0.0051**
|
|
-0.0085
|
-0.0001**
|
-0.0023**
|
0.0001
|
0.0045
|
Fat (% E4)
|
Model 1
|
|
0.0226
|
0.0001
|
0.0001
|
0.0236
|
-0.0308
|
|
0.0265
|
0.0050
|
0.0001
|
-0.0300
|
-0.0515
|
|
Model 2
|
|
0.0080
|
0.0069
|
0.0027
|
0.0374
|
-0.0323
|
|
0.0257
|
0.0001
|
0.0126
|
-0.0269
|
-0.0896
|
Protein (% E5)
|
Model 1
|
|
0.1543
|
0.0714
|
0.0960
|
0.0001
|
0.0810
|
|
-0.1222
|
-0.0974
|
-0.066
|
-0.103
|
-0.136
|
|
Model 2
|
|
0.0367
|
0.0307
|
-0.0335
|
-0.0778
|
0.0235
|
|
0.0001
|
0.0001
|
-0.0246
|
0.0001
|
-0.075
|
Carbohydrate(% E6)
|
Model 1
|
|
-0.0195
|
-0.0001
|
-0.0161
|
0.0211
|
0.0001
|
|
-0.0226
|
-0.0021
|
-0.0001
|
0.0414
|
0.0614
|
|
Model 2
|
|
-0.0172
|
-0.0020
|
0.0111
|
0.0364
|
0.0267
|
|
-0.0302
|
-0.0001
|
-0.0115
|
0.0302
|
0.0897
|
Note: 1 Model 1 included energy intake and macronutrient composition in the quantile regression model to investigate the association with BMI/waist circumference.
2 Model 2 adjusted for age (18-44y and 45-64y), education (primary/illiterate, junior and high school/above), income (low, medium and high), region (urban/rural) and physical activity (low, medium and high), current smoker (yes/no), current drinker (yes/no).
3 ***, ** and * indicate statistical significance at the 1%, 5%and 10% level, respectively.
% E4, 5, 6 means the percentage of energy intake from fat, protein and carbohydrate, respectively.