In this study, with mean carbohydrate intake relatively low accompanied by a higher fat intake (50% and 36% of total energy, respectively), the usual LCD score was positively associated with the risk of MetS in men. A significant trend was found between plant-based LCD score and the MetS (Pfor trend<0.05) in men. However, no association between any kind of LCD score and the MetS was found in women. In addition, a direct positively relationship was identified between the mean values of the three LCD scores and the number of MetS components in men, but not in women. It suggests that the usual LCD score may be an independent risk factor for the MetS in men but not in women. Furthermore, the LCD scores were associated with individual components of the MetS in men, especially for central obesity, hyperglycemia and hypertriglyceridemia.
The underlying mechanism of this gender difference was unclear, and possibly it is related to the differences in lifestyle and dietary characteristics between men and women [27–29]. Besides, different food choices in men and women may exist. In Chinese men, there was the existence of an “alcohol” or “animal and fried food” dietary pattern, characterized by animal foods (eg, meat, poultry, fish), fried dough and alcohol[28, 29], which was the main source of total fat, especially saturated fat[30]. Our study showed that intake of alcohol consumption and red meat was higher in men than that in women, despite fat intake was similar. For example, in Q1 of the usual LCD, the drinking rate of men and women was 54% and 7.3% respectively, and the average intake of red meat was 46.6 g and 38.6 g respectively. For Q4, the drinking rate of men and women was 53.1% and 10.3%, and average red meat intake was 104.1 g and 90.4 g, respectively. High alcohol consumption has been confirmed to be positively associated with the risk of MetS and its components[31, 32]. And red meat consumption was related to insulin resistance, lipid peroxidation and metabolic syndrome reported in Brazilian middle-aged men[33].
Epidemiologic researches linking LCD score and the risk for MetS were limited and controversial[19–22]. A cohort study in Tehran showed LCD may be associated with a decreased risk of MetS and its components in adults[19], but not in children and adolescents[20]. And other studies also displayed null association between LCD score and MetS[20, 22]. Our study showed that the high usual LCD score was significantly positively associated with the risk of MetS and its components in men, such as hyperglycaemia, hypertriglyceridemia, central obesity. One of the possible explanations for this controversy in the present study might be related to the composition of LCD, such as the quantity and quality of macronutrients. Low-carbohydrate diets usually have a relatively lower percentage of carbohydrate intake from total energy and a higher percentage of fat intake from total energy. In this study, the average carbohydrate intake and fat intake of the highest quartile groups of the LCD score were 39–40% and 43–46% of total energy in men, respectively, which were different from that in Tehranian adults (53.5% for carbohydrate and 35.2% for fat, respectively) and the Korean adults (46–48% for carbohydrate and 32–33% for fat, respectively). Our study showed that high fat intake of total energy was slightly higher than the recommended levels of China and WHO.
High fat intake has been shown to be positively associated with metabolic syndrome risk[34–36],although the 2015–2020 Dietary Guidelines for Americans emphasize the types of fat rather than total fat intake[37]. In our investigation༌fat intake of total energy was over 40% in men in the highest quartile groups of the LCD score༌accompanied with higher consumptions of different types of fats (SFA, MUFA and PUFA). The main types of fat in Chinese diet are MUFA and PUFA, because vegetable cooking oil consumption is the main source of fat for people. However, high LCD score including a high plant-based LCD score were associated with the high incidence of type 2 diabetes in Chinese adults[11]. In the present study similar results were found that a high plant-based LCD score increased the risk of metabolic diseases, such as hyperglycaemia, central obesity and MetS in men, which was inconsistent with the results of western countries[13, 17, 18]. In our study, the vegetable cooking oil is the main source of fat in a high plant-based LCD score, and the vegetable cooking oil intake was up to 63g in men with the highest quartile of the plant-based LCD score, which was two folds as much the recommended maximum intake level (recommended intake is 25-30g). Previous reports showed individuals with a higher total fat intake(> 37E%) did not have beneficial impact from MUFA[38], and only in subjects with a lower total fat intake (< 35E%) was found that the association of changes in estimated desaturase activities with changes in insulin sensitivity[39]. The present study showed that excessive fat intake, even MUFA and PUFA, may be dangerous, and the overall dietary feature should be considered whether it is an animal-based diet or a plant-based diet.
To our knowledge, this is the first epidemiology study examining the association between LCD score and the MetS in Chinese adults who undergoing nutrition transition. As a cross-sectional study of the relatively large sample, this study used the 3 consecutive days of dietary recollection combined with condiments weighing method to evaluate LCD score according to food source and explore the relationship between LCD score and the MetS, adjusting for various potential confounders such as age, sex, smoking, drinking, BMI, physical activity, drug use and total energy intake in the multiple regression models. The overall results are relatively reliable.
The study has several limitations. First, causal relationship can not be identified due to the cross-sectional nature of our data. Second, the observation from our study may not be generalizable to all Chinese populations because of the survey location, which is only one Chinese megacity. Other limitations may include measurement bias and residual confounding, as all other observational studies.