The current results manifested that higher fat or protein intakes were associated with decreased risks of liver dysfunction in male, low energy intakes, no drinking and normal BMI tuberculosis patients. Additionally, RCS showed J-shaped associations between the protein intakes, fat intakes, the fat to energy percentages or the carbohydrate to energy percentages and the risk of liver dysfunction for TB patients.
Our finding of the negative association between protein intakes and liver dysfunction is consistent with previous studies, such as a large-scale, community-based prospective cohort of Korean adults[24], randomized controlled trials among patients with long-chain fatty acid oxidation disorders (LC-FAODs) [25], and among participants with morbid obesity [26]. Additionally, the results from an animal experiment on rats revealed that higher protein intake in diet exhibited lower liver weights and less fat deposition in the liver [27]. The reason might be amino acids, such as methionine, N-acetylcysteine, and glycine, could mitigate or prevent oxidative stress and damage in the liver to humans and other animals [28]. Our RCS model predicts that the risk of liver dysfunction will decline apparently by consumption of dietary protein of 70.69 g per day, which was consistent with a study aimed at investigating the quantitative relationship between dietary protein intake and liver disease [29]. Previous studies have indicated that inadequate intakes of protein were prevalent in TB patients [9, 30]. For example, protein intakes in TB patients (44.6 g/day in male and 35.9 g/day in female) were much lower than Dietary Reference Intake in the normal population (65 g/day in male and 55 g/day in female) [9], which were also lower than the cut-off in our RCS models. Therefore, it makes sense to recommend an increased protein intake for TB patients.
The negative association between the prevalence of liver dysfunction and fat intakes was found in our study, which was consistent with the previous cohort studies [31, 32]. However, there were conflicting results from other studies. After an average follow-up time of 26.6 years, there was a null association between total fat intake and hepatocellular carcinoma cancer (HCC) in 160 cases [33]. Some studies suggested that high fat intake could worsen NAFLD, possibly because excess saturated fatty acids and trans fatty acids might have an impact on hepatocyte steatosis through chylomicron uptake and fatty decomposition of adipose tissue and de novo hepatic adipogenesis (DNL) [34–36]. Because Chinese participants, living in rural, consumed fewer animal fats, such as unprocessed red meat and processed meat; more vegetable oil dietary and fat from the plant, such as whole grains, nuts and legumes [37, 38]. Moreover, dietary total fat intake up to 90g/ day has a significant protective effect on HCC; much higher than the protective effect of 47.50 g/ day dietary fat intake on liver function suggested by our RCS model [39]. It may be caused by Chinese and American difference in eating habits. Research showed that dietary fat provided 35% of total dietary energy for American, and only 22% for Chinese [40–42]. In addition, our result of 47.50g/d dietary fat intake is much lower than average 76.9g in the daily fat intake of the Chinese rural population [43], which may be the fact that the population in our data was TB population.
We determined that the energy ratio of protein had no significant effect on liver dysfunction in the RCS model. Our novel finding revealed that fat or carbohydrate to energy percentages had a non-linear relationship with liver function, indicating that energy percentages of macronutrients might play essential roles in liver dysfunction in TB patients. As for their ratio, RCS revealed that when less than the cut-off values (the fat (22%) OR carbohydrate (69%) to energy percentages), the OR value of liver dysfunction gradually decreased with the increase of the energy supply ratio; otherwise, the opposite trends appeared. In agreement with previous findings [15, 44–46], over fat or carbohydrate energy supply diet was significantly higher on liver dysfunction as compared with a normal calories diet, which was in part explained by the dose-response relationship between dietary fat or carbohydrates and liver dysfunction. The possible underlying mechanisms are as follows. First, studies have shown that intrahepatic triglyceride (IHTG) synthesized via the DNL and chylomicron particles may trigger relevant potential pathways pathway leading to non-normal liver function [47]. DNL is an intrinsic metabolic process that converts excessive carbohydrates to fatty acids esterified to become TGs for storage [48]. Then, for intravascular hydrolysis mediated by lipoprotein lipase, chylomicron particles transport dietary lipids in the systemic circulation to peripheral tissues including adipose tissue and skeletal muscle. Fatty acids that are not absorbed by these tissues spill over into the bloodstream and can be absorbed by the liver and utilized to make triglycerides [48]. Second, hepatic lipid accumulation is the rate of DNL relative to fatty acid oxidation. Acetyl-CoA carboxylase (ACC) is a key enzyme, regulating the production of malonyl-CoA, a substrate for DNL and an inhibitor of carnitine palmitoyl-transferase 1 (CPT1) and thereby mitochondrial fat acid entry [48, 49]. Third, energy percentages control could decrease the ratio of glutathione: oxidized glutathione (GSH: GSSG) in the mitochondria of liver tissue by lowering GSSG levels, and reduce liver damage [50, 51]. Because studies have demonstrated that Sirt3 could modulate mitochondrial fatty-acid oxidation in mammals, mediate oxidative damage and enhance the glutathione antioxidant defense system under energy restriction condition [52, 53]. Besides, isocitrate dehydrogenase 2 (Idh2), an enzyme that converts nicotinamide adenine dinucleotide phosphate (NADP) to NADPH in mitochondria, could regulates the redox state of mitochondria. Sirt3 stimulates Idh2 activity in mitochondria, leading to increased levels of NADPH and GSH: GSSG, the major redox couple in the cell. Therefore, over-consumption should also be avoided.
In addition, there is a consensus that gradual weight loss through caloric reduction improves serum liver enzymes, liver fat, liver inflammation, and degree of fibrosis [54]. Previous studies have suggested that intrahepatic fat contents loss and improved liver profile were achieved in adults with NAFLD by reducing calorie intake and improving diet quality [55, 56]. And, there was consistent with our results suggested an inverse association between dietary protein or fat intake and odds of liver dysfunction in patients with energy intakes below 1646.80 kcal per day. According to a single-center, parallel-group, double-blind RCT for 12 weeks, it suggested that compared to a standard energy-restricted diet, the combination of lower energy intake and higher dietary protein intake could lead to more significant improvements in oxidative stress and inflammation and more adiponectin in overweight/ obese NAFLD adults [57].
The strengths of this study included the large-scale epidemiological study, which investigates the relationship between macronutrients intakes and liver dysfunction during tuberculosis treatment and considers the effect of macronutrients to energy percentages on liver dysfunction among tuberculosis patients. Far more convincing, we also conducted the RCS model among participants who were tested for cut-off values of macronutrients intakes or the energy percentages, an attempt less made earlier. However, several limitations common to observational studies should be mentioned. Due to the nature of cross-sectional studies, the temporal sequences may not be clear. But providing clues to the cause for the benefits of increasing nutritional education and supplementing appropriate protein and fat on the treatments and prognosis of tuberculosis patients with liver dysfunction, which was consistent with previous studies [58, 59]. In addition, observational studies cannot be free from some residual confounding factors, although we adjusted for many potential risk factors of liver dysfunction, including age, gender, education, alcohol intake, physical activity, marital status, smoking status, TG, TCHO and HDL. Then, we just focused on the pure nutrients and no data on dietary types were taken into consideration. Finally, all participants in this study are Chinese, which limits its generalizability.