-
The study was conducted in five provinces of China, including the northern regions of China, such as Mudanjiang in the Heilongjiang province, Shijiazhuang in the Hebei province, and Beijing, and the southern regions of China, such as Chengdu in the Sichuan province and Shenzhen. Rural regions were selected in Mudanjiang, Shijiazhuang, and Chengdu, and urban regions were chosen in Beijing, Shenzhen and Chengdu.
We first used the statistical formula [n = (Uα × σ/δ)2] to calculate the lowest sample size required for this study. In the formula, α was significant level and the value was 0.05, and U value was 1.96; σ represented the standard deviation of basal metabolism values for adult males and females, for which we referred to values reported by previous studies[20,21]; δ was acceptable error. By the calculation, the lowest number was 166 for males and 133 for females. Each region was assigned the same number of participants. To ensure an adequate sample size, double the minimum subjects were enrolled. The subjects were enrolled in different regions and underwent medical examination. Individuals with thyroid diseases, hepatic diseases, renal diseases, insulin-dependent diabetes mellitus, or any other metabolic disorder were excluded. For female subjects, their menstruation was required to be regular with no menstruation during the experiment. Pregnant and lactating women were not selected during the experiment. In total, 470 healthy adults of normal body weight (BMI between 18.5 and 23.9 kg/m2) aged from 20 to 45 years old were selected finally.
All procedures involving human subjects were approved by the National Institute for Nutrition and Health Chinese Center for Disease Control and Prevention Ethical Review Committee (Ethical approval no.: 2009−0212). Written informed consent was obtained from all subjects.
-
The height of subjects was determined with the Fix Feet Tall (Lameris, Utrecht, Netherlands) with an accuracy of 0.01 meter. Participants stood straight and barefoot on the baseboard of the Fix Feet Tall to measure their height, and their weight was determined with the Digital Weight Scale (HW100KGL, Japan) with an accuracy of 0.01 kg. After a night of fasting, participants in their underwear stood on the baseboard of the Digital Weight Scale to measure their body weight (Body weight = measured weight − weight of their underwear). The body surface area (BSA) was calculated using the following formula proposed by Zhao et al.[22,23]; for men, BSA (m2) = 0.00607 height (cm) + 0.0127 weight (kg) − 0.0698; for women, BSA (m2) = 0.00586 height (cm) + 0.0126 weight (kg) − 0.0461.
-
The BEE was measured with a portable indirect device called the cardiopulmonary function tester (Cosmed, K4B2, Italy). On the day before the experiment, each subject stayed in a single room with a stable temperature of 20–25 ℃ and humidity of 40%–60%. The subjects were asked to get accustomed to the apparatus, face mask, and the surrounding environment. BEE was measured in the morning when subjects with 12 h fasting were awakened gently from sleeping and asked to lie down quietly. During the procedure, the subjects could not move or speak. Once the consumption of oxygen and production of carbon dioxide were stable, the measurement lasted for 4 minutes. At the same time, the temperature, heart rate, and respiratory rate of subjects, and the temperature and humidity of the room were measured and recorded. Values for energy expenditure (EE) were calculated from VO2 and VCO2 using Weir’s equation[24].
The K4b2 is a portable piece of equipment that can monitor the real-time exhaled gas of a subject with remote sensing technology. It determinates the amount of gas exhaled by a subject and then calculates the amount of oxygen and carbon dioxide expenditure, from which the respiratory quotient is acquired. Before each test, the Cosmed K4b2 system was warmed up for at least 45 minutes, and the O2 and CO2 analyzers were calibrated using ambient air and reference gas with 16% O2 and 5% CO2. The flow meter was calibrated using a 3 liter syringe (Quinton Instruments, Seattle, WA, USA).
-
The predictive equations of Henry et al.[25], Schofield et al.[11], H-B et al.[13], and Liu et al.[14] were used to calculate the BEE. These equations were chosen either because they had previously been widely used in healthy Chinese population studies (Schofield equation, H-B equation)[11,13], derived based on a Chinese database (Liu equation)[14], or reported to be better suitable for a Chinese population (Henry equation) [25]. The predictive equations chosen for the estimation of BEE are presented in Table 1.
Author (age) Male Female Henry (18–30) (kJ/d) 51W + 3,500 47W + 2,880 Henry (30–60) (kJ/d) 53W + 3,070 39W + 3,070 Schofield (18–30) (kJ/d) 63W + 2,896 62W + 2,036 Schofield (30–60) (kJ/d) 48W + 3,653 34W + 3,538 HB (≥ 18)(×4.184 kJ/d) 66.473 + 5.003H + 13.752W – 6.775A 655.096 + 1.850H + 9.563W – 4.676A Liu (≥ 18) (×4.184 kJ/d) 13.88W + 4.16H to 3.43A – 112.40S + 54.34 (male = 0 and female = 1) Note. HB: Harris-Benedict, W: Weight (kg), H: Height (cm), A: Age (years), S: Sex. Table 1. Predictive equations chosen for the estimation of basal energy expenditure
-
The data were analyzed with the SPSS software (version 19.0; SPSS, Inc., Chicago, IL, USA). Descriptive data are presented as mean ± standard error of mean (SEM). Independent or paired t-test was used to compare mean differences (Kcal/day) between the measured and predicted values among subgroups. Multiple linear regressions were applied to derive new predictive equations to estimate the BEE for males and females. The bias, accuracy rate, concordance correlation coefficient (CCC), and root mean square error (RMSE) were used to evaluate the accuracy of the predictive equations. Accuracy was calculated as the percentage of subjects with pBEE values within 10 percent of mBEE[26]. A prediction < 90% of the measured mBEE was classified as an underestimation, whereas a prediction > 110% of the measured mBEE was classified as an overestimation. Chi square analysis was used to determine whether the differences in categorical variables such as sex, region, and accuracy rate were significant. Significance for all analyses was set at P < 0.05.
-
The distribution of participants is summarized in Table 2. There were 470 subjects in total, including 232 males and 238 females. The distribution of participants was similar between northern and southern areas, as well as between urban and rural places.
Region Total Male (n, %) Female (n, %) Total 470 232 (49.4) 238 (50.6) North/South North 250 126 (50.4) 124 (49.6) South 220 106 (48.2) 114 (51.8) Urban/Rural Urban 286 144 (50.3) 142 (49.7) Rural 184 88 (47.8) 96 (52.2) Table 2. Distribution of all the participants
The characteristics of all the subjects are summarized in Table 3. Females and males were of similar ages. Significant differences were observed in weight, height, BMI, and body surface between females and males.
Variables Total Female Male P-value Age (year) 27.76 ± 0.36 28.02 ± 0.53 27.49 ± 0.49 0.471 Weight (kg) 57.35 ± 0.36 52.92 ± 0.36 61.89 ± 0.46 < 0.001 Height (cm) 164.03 ± 0.37 158.76 ± 0.37 169.44 ± 0.41 < 0.001 BMI (kg/cm2) 21.24 ± 0.08 20.99 ± 0.11 21.51 ± 0.11 0.001 Body surface (m2) 1.62 ± 0.01 1.53 ± 0.01 1.72 ± 0.01 < 0.001 Table 3. Characteristics of all participants
-
As presented in Table 4, males expended significantly higher energy (5,954 kJ/d) than females (5,089 kJ/d, P < 0.001). Females who lived in northern areas expended similar energy as those in southern areas. Females in rural area expended significantly higher energy than those in urban areas (P < 0.001). Males in southern areas had significantly higher energy than those in northern areas (P < 0.001), while males expended similar energy in both northern and southern areas.
Variables Total (kJ/d) P-value Female (kJ/d) P-value Male (kJ/d) P-value Total 5,516 ± 70 5,089 ± 97 5,954 ± 93 North/South North 5,495 ± 111 0.744 5,265 ± 172 0.057 5,721 ± 138 0.006 South 5,540 ± 81 4,897 ± 73 6,232 ± 115 Urban/Rural Urban 5,279 ± 84 < 0.001 4,662 ± 89 < 0.001 5,887 ± 123 0.354 Rural 5,885 ± 117 5,720 ± 183 6,065 ± 139 Table 4. The measured basal energy expenditure for all participants
-
Considering the statistical difference of the BEE values between males and females, we derived two different equations according to variables such as weight, height, BSA, and regions. The predictive equation was: BEE (kJ/d) = 2625.201 + 60.003 × weight (kg) −707.702 × region (North = 1, South = 0) for males (r2 = 0.117, n = 232), and BEE (kJ/d) = −7141.710 + 82.444 × height (cm) − 1437.918 × region (city = 1, rural = 0) for females (r2 = 0.203, n = 238).
-
As shown in Table 5, the BEE values predicted by the equation of Henry, Schofield, H-B, and Liu were significantly different from those of the mBEE for males (all P < 0.05), while the values predicted by the equation developed in the present study were similar to the mBEE. The CCC of the new equation was the highest among all the predictive equations; however, the CCC of all predictive equations was lower than 0.8. The new equation had the smallest RMSE and the maximum positive error (MPE), while the maximum negative error (MNE) was similar among all the predictive equations. The accuracy of all the predictive equations was lower than 50%, and the Chi square analysis showed that there was no significant difference among them (all P > 0.05).
Predictive equations Average bias (kJ/d) CCC (95% CI) RMSE (kJ/d) MNE (%) MPE (%) Accuracy rate (%) Under/over-estimation (%) Henry −670 ± 91* 0.093 (0.034–0.151) 1,517 −42.3 170.0 34.5 65.5 Schofield −797 ± 63* 0.100 (0.034–0.151) 1,594 −38.9 171.6 32.8 67.2 H-B −652 ± 93* 0.129 (0.053–0.204) 1,552 −43.5 163.8 34.1 65.9 Liu −422 ± 91* 0.142 (0.065–0.216) 1,450 −45.0 152.9 35.3 64.7 New equation in present study 0.03 ± 87 0.210 (0.136–0.282) 1,328 −50.4 122.2 37.9 62.1 Note. H-B: Harris-Benedict, CCC: Coherent correlation coefficient, RMSE: Root mean square error, MPE: Maximum positive error, MNE: Maximum negative error, *Compared to mBEE, P < 0.05. Table 5. Evaluation of the predictive equations for healthy Chinese males
As presented in Table 6, the BEE values predicted by the equation of Henry, Schofield, and H-B were significantly different from those of the mBEE for females (all P < 0.05). No significant difference was observed between the mBEE and the predicted values from the Liu equation and the equation derived in the present study. The CCC of the new equation was the highest among all the predictive equations. However, the CCC of all predictive equations was lower than 0.8. The new equation had the smallest RMSE, maximum positive error (MPE), and the maximum negative error (MNE). The accuracy of all the predictive equations was lower than 50%, and the Chi square analysis showed that there was no significant difference among them (all P > 0.05).
Predictive equations Average bias (kJ/d) CCC (95% CI) RMSE (kJ/d) MNE (%) MPE (%) Accuracy rate (%) Under/over- estimation (%) Henry −199 ± 98* 0.006 (−0.038−0.050) 1,520 −52.6 158.0 34.5 65.5 Schofield −235 ± 96* 0.051 (0.004–0.098) 1,495 50.4 158.0 32.8 67.2 H-B −450 ± 99* −0.011 (−0.057–0.035) 1,593 −50.7 170.0 29.0 71.0 Liu −103 ± 99 0.033 (−0.032–0.098) 1,523 −54.7 155.9 37.8 62.2 New equation in present study 0.02 ± 86 0.338 (0.255–0.416) 1,328 −45.2 115.4 35.3 64.7 Note. H-B: Harris-Benedict, CCC: Coherent correlation coefficient, RMSE: Root mean square error, MPE: Maximum positive error, MNE: Maximum negative error, *Compared with mBEE, P < 0.05. Table 6. Evaluation of the predictive equations for healthy Chinese females
Basal Energy Expenditure of Chinese Healthy Adults: Comparison of Measured and Predicted Values
doi: 10.3967/bes2020.075
- Received Date: 2019-10-22
- Accepted Date: 2020-03-11
-
Key words:
- Basal energy expenditure /
- Chinese healthy adults /
- Indirect calorimetry /
- Predictive equations
Abstract:
Citation: | MAO De Qian, WU Jing Huan, HUANG Cheng Yu, LI Ke Ji, LIU Xiao Li, ZHANG Shi Lian, WANG Yan Ling, CHEN Wei, LI Ming, YANG Xiao Guang, PIAO Jian Hua. Basal Energy Expenditure of Chinese Healthy Adults: Comparison of Measured and Predicted Values[J]. Biomedical and Environmental Sciences, 2020, 33(8): 566-572. doi: 10.3967/bes2020.075 |