Effects of Macronutrient Distribution on Weight and Related Cardiometabolic Profile in Healthy Non-Obese Chinese: A 6-month, Randomized Controlled-Feeding Trial

Abstract

Background: It has been suggested that the increase in carbohydrate at the expense of fat has contributed to the obesity epidemic in North America and some European countries. However, obesity rates in China have increased rapidly in parallel with a transition from the traditional low fat, high carbohydrate diet to a diet relatively high in fat and reduced in carbohydrate. Therefore, the current study aimed to determine whether the traditional Chinese diet was likely to be more effective than a diet with higher fat and lower carbohydrate — which is consumed in most Western societies, at weight control among a non-obese healthy population in China.

Methods: The 6-month, two-center, three-arm, randomized, parallel-group, controlled-feeding trial was conducted at People's Liberation Army General Hospital in north China and Zhejiang University in south China. We recruited healthy young adults (aged 18–35 years, body mass index < 28) who lived in the university campus or the hospital dormitory during the whole study intervention period. They were required to eat only the foods provided, and to avoid excessive or unusual strenuous exercise during the trial. Participants were simultaneously enrolled and randomized using a computer-generated number (stratified by clinic center, age, sex, and body mass index) by data manager to one of the three isocaloric diets (1:1:1): a lower fat, higher carbohydrate diet (fat 20%, carbohydrate 66% energy); a moderate fat, moderate carbohydrate diet (fat 30%, carbohydrate 56% energy); a higher fat, lower carbohydrate diet (fat 40%, carbohydrate 46% energy). Protein provided 14% energy in all diets. We provided all food and beverages throughout the 6-month intervention. Laboratory personnel were masked to treatment allocation. Body weight was the primary outcome and measured each month. Data were primarily analyzed according to an intention-to-treat approach, supplemented with per-protocol analysis. The study was approved by the Ethics Committee at Zhejiang University. Each participant provided written informed consent. The study was registered at Clinicaltrials.gov, number NCT02355795">NCT02355795.

Findings: Between April 30, 2016, and October 30, 2016, 307 participants were randomly assigned to the lower fat diet (n = 101), the moderate fat diet (n = 105) and the higher fat diet (n = 101), and 245 (79.8%) participants completed the study. Reduction in body weight was significantly greater in the lower fat, higher carbohydrate group throughout the intervention (P < 0.001 for the interaction between diet group and time) than in the two other groups. Weight change at 6 months was − 1.6 kg (95% CI − 1.8 to − 1.4) in the lower fat, higher carbohydrate group; − 1.1 kg (95% CI − 1.3 to − 0.9) in the moderate fat, moderate carbohydrate group, and − 0.9 kg (95% CI − 1.1 to − 0.6) in the higher fat, lower carbohydrate group. Reduction in waist circumference, total cholesterol, low-density lipoprotein cholesterol and non-high-density lipoprotein cholesterol on the lower fat, higher carbohydrate group were greater than those observed on the other two diet groups.

Interpretation: A lower fat, relatively higher carbohydrate diet, similar in macronutrient composition to that traditionally eaten in China appears to be less likely to promote excessive weight gain and be associated with a lower cardiometabolic risk profile than a diet more typical of that eaten in Western countries in healthy non-obese Chinese. Findings from studies in European and North American populations suggesting possible benefits of carbohydrate restriction may not apply to people of other ethnicities.

Fig. 1

Trial profile. LF-HC = lower fat, higher carbohydrate. MF-MC = moderate fat, moderate carbohydrate. HF-LC = higher fat, lower carbohydrate.

Fig. 2

Mean changes and diet contrast in body weight and waist circumferences. Panel A shows mean changes in body weight and waist circumference. Error bars indicate 95% confidence intervals. Data are based on mixed-model analysis of variance. The P value at the lower left indicates the test of whether the change between baseline and intervention period (mean of each month) differed significantly between participants assigned to three diet groups. The P values for the comparison between the LF-HC group and the MF-MC group are < 0.001 for body weight and < 0.001 for waist circumference. The P values for the comparison between the HF-LC group and the MF-MC group are 0.44 for body weight and 0.21 for waist circumference. The P values for the comparison between the LF-HC group and the HF-LC group are < 0.001 for body weight and < 0.001 for waist circumference. Panel B shows the between-diet differences in body weight and waist circumference at 1 to 6 month. Error bars indicate 95% confidence intervals. LF-HC = lower fat, higher carbohydrate. MF-MC = moderate fat, moderate carbohydrate. HF-LC = higher fat, lower carbohydrate.

Fig. 3

Mean changes in lipid profile throughout the 6-month intervention. Panel A shows the results for total cholesterol, Panel B for HDL cholesterol, Panel C for LDL cholesterol, Panel D for non-HDL cholesterol, Panel E for total/HDL cholesterol, Panel F for apolipoprotein A1, Panel G for apolipoprotein B, Panel H for triglyceride. Data are based on mixed-model analysis of variance. The P value at the lower left indicates the test of whether the change between baseline and intervention period (mean of each month) differed significantly between participants assigned to three diet groups. The P values for the comparison between the LF-HC group and the MF-MC group are 0.004 for total cholesterol, 0.03 for HDL cholesterol, 0.05 for LDL cholesterol, and 0.03 for non-HDL cholesterol. The P values for the comparison between the HF-LC group and the MF-MC group are 0.19 for the total cholesterol, 0.77 for HDL cholesterol, 0.48 for LDL cholesterol, and 0.27 for non-HDL cholesterol. The P values for the comparison between the LF-HC group and the HF-HC group are < 0.001 for total cholesterol, 0.004 for HDL cholesterol, 0.001 for LDL cholesterol, and 0.003 for non-HDL cholesterol. LF-HC = lower fat, higher carbohydrate. MF-MC = moderate fat, moderate carbohydrate. HF-LC = higher fat, lower carbohydrate.