Clinical and lipid metabolic responses to diacylglycerol oil administration in Chinese adults with overweight/obesity or central obesity: a randomized, double-blind, placebo-controlled trial

Abstract

Abnormalities in triacylglycerol metabolism can lead to excessive visceral fat accumulation. Though diacylglycerol (DAG) administration could reduce serum total triacylglycerol (TG), its impact on visceral fat deposition and potential mechanisms remains unclear. This trial aimed to evaluate the impact of substituting regular rapeseed cooking oil (TAG) with DAG oil on primary outcomes such as anthropometric measurements and lipid profiles, as well as secondary outcomes including visceral fat and serum lipidomics in Chinese adults with overweight/obesity or central obesity. Ninety-five participants (BMI: 25.93 ± 2.92 kg m⁻²) were assigned to the DAG or TAG group through random allocation. Over 8 weeks, participants were provided similar diets cooked with DAG or TAG oil, respectively. By week 8, the serum TG (P = 0.026) and small dense low-density lipoprotein cholesterol (sdLDL-C) (P = 0.024) levels in the DAG group were significantly lower than those in the TAG group. The change in sdLDL-C was notably greater in the DAG group than in the TAG group (−0.10 ± 0.12 vs. −0.03 ± 0.16), and a significant decrease in the levels of waist circumference, hip circumference, total cholesterol, and sdLDL-C was only observed in the DAG group compared with the baseline (all P < 0.05). Imaging analyses revealed that attenuation of hepatic steatosis was observed in the DAG group compared with the TAG group (P = 0.035), and a decrease in visceral fat area was found only in the DAG group compared with the baseline (P < 0.001). Lipidomic profiling demonstrated DAG induced enrichment of serum triacylglycerol and phosphatidylethanolamine species containing mono/polyunsaturated fatty acids, which were associated with the enhanced adipocyte lipolysis and thermogenesis. These findings suggested that DAG-mediated lipid remodeling might be related to preventing lipid metabolic disorders through visceral fat regulation.

---

1. Introduction

Due to the dysregulation between lipogenesis and lipolysis, aberrant fat accumulation not only induces obesity but also confers a higher risk of hypertriglyceridemia, metabolic dysfunction-associated fatty liver disease (MAFLD), and cardiovascular diseases (CVDs), with central obesity exhibiting stronger epidemiological associations than generalized obesity.¹ Dysregulated triglyceride metabolism is implicated not only in elevating blood triglyceride levels, but also in driving fat deposition in visceral organs, particularly the liver and heart.² The estimated global prevalence of hypertriglyceridemia was 28.8% in adults,³ and hypertriglyceridemia is a common lipid abnormality in persons with visceral obesity and metabolic syndrome.⁴ Lifestyle modification is a foundational component of hypertriglyceridemia management. Dietary intervention, in particular, has demonstrated significant efficacy in reducing the body weight (BW), waist circumference,⁵ and triglyceride levels, thereby contributing to the mitigation of MAFLD and CVDs risk. A meta-analysis of randomized controlled trials demonstrated that consuming canola oil, compared to olive oil, significantly reduced the levels of total triglyceride and low-density lipoprotein cholesterol, suggesting that the type of edible oil plays a critical role in managing dyslipidemia.⁶

Diacylglycerol (DAG), a natural component of edible oils typically present at concentrations below 10%, has garnered considerable scientific interest in weight management and dietary interventions due to its distinct metabolic properties compared to triacylglycerol (TAG).⁷ Clinical evidence consistently indicates that DAG-enriched oil consumption promotes fat oxidation while maintaining the energy expenditure, potentially modulating appetite regulation and energy homeostasis.⁸ A short-term intake of DAG can effectively mitigate the postprandial elevation of blood lipids, particularly total triacylglycerol (TG) in serum,⁹ and long-term adherence to a diet high in DAG can reduce total fat and BW.¹⁰ Furthermore, specific DAG varieties, particularly those rich in alpha-linolenic acid, may enhance fat oxidation in individuals with insulin resistance and facilitate body fat reduction.¹¹ Although the effects of DAG on BW, lipid metabolism, and energy metabolism have been widely studied in humans, whether it can affect the accumulation of visceral fat is still not known, especially in the improvement of liver fat. Furthermore, the processes by which DAG influences lipid accumulation in humans remain poorly understood, despite some research studies that have attempted to investigate the regulation of lipid metabolism by dietary intervention via lipidomics. Nevertheless, the genetic background and dietary habits of Chinese residents are significantly different from those in the Western countries and Japan, and evidence on the effects of DAG on lipid metabolism and lipid accumulation is limited in the Chinese population.

This study investigated the potential effects of DAG administration on lipid metabolism and lipid accumulation in Chinese participants with overweight/obesity or central obesity through a randomized, double-blind, placebo-controlled trial. The participants consumed the standard meals cooked with DAG or TAG oil, and a nutritional survey was conducted to ensure a balanced intake of nutrients between the intervention group and the control group before and after the intervention. Primary outcomes included various physical measures (e.g., body weight, waist circumference) and serum parameters of lipid metabolism, such as lipid and lipoprotein levels. Secondary outcomes comprised visceral fat area, measured via dual-energy X-ray absorptiometry, and hepatic fat accumulation, assessed using conventional ultrasound, to evaluate the effect of DAG substitution on lipid metabolism and accumulation. Furthermore, serum lipidomics was used to conduct an in-depth investigation of the regulatory mechanisms governing lipid metabolism and fat deposition.

2. Materials and methods

2.1. Study design

This was a randomized, placebo-controlled, double-blind study conducted from March to May 2024 at China Agricultural University in Beijing. Ethical approval was obtained from the Research Ethics Committee of China Agricultural University (CAUHR-20240104) on 26 January 2024, and the study was registered in the Chinese Clinical Registry (ChiCTR-2400081188) on 26 February 2024. Participants met the inclusion criteria and provided written informed consent before enrollment.

The inclusion criteria were as follows: (1) aged 18–50 years, (2) body mass index (BMI) ≥22 kg m⁻² or waist circumference ≥85 cm for men and ≥80 cm for women,¹² (3) no food allergy history, and (4) willingness to complete the questionnaires for this trial. Individuals with the following conditions were excluded: (1) weight loss of more than 4.50 kg during the last 2 months, (2) impaired liver and kidney function, (3) medicinally controlled endocrine disease (such as diabetes mellitus, thyroid disease, etc.), (4) established hyperlipidaemia, (5) established cancer, (6) significant alcohol consumption during the previous year, (7) currently smoking, and (8) breastfeeding or pregnant.

Based on previous research performed on DAG oil, the sample size was calculated based on the primary outcome (TG). 44 patients per group would be necessary to observe ≥15% reduction of TG in the DAG group with 80% power and 5% significance level (one-sided). Considering a dropout rate of 15%, 102 randomized patients (51 per group) were needed.¹³ The eligible participants (n = 103, 61 men) were randomized in a 1 : 1 ratio to either the intervention (DAG group) or the control group (TAG group). Computer-generated random permuted blocks of sizes four were used, stratified by baseline BW (<70/70–80/≥80 kg) and BMI (<23/23–27.5/≥27.5 kg m⁻²). After adjustment for baseline equivalence, the two randomly assigned groups were designated as Group A and Group B. Neither the researchers nor the participants were aware of the treatment allocation. To ensure blinding, DAG and TAG oils were coded as A or B by the manufacturer's scientific expert. The principal investigator administered the treatments to adults in a double-blind manner based on these codes. The oil codes were documented on individual information forms, and the researcher responsible for data collection remained unaware of the treatment types. The participants underwent anthropometry, blood testing, and clinical analysis at weeks 0, 4, and 8 during the treatment period. They were instructed to maintain their living habits such as eating and physical activities. Physical exercise was evaluated using an international physical activity questionnaire at weeks 0, 4, and 8. During the intervention, 8 participants did not complete the follow-up intervention: 5 failed to adhere to the experimental diet, 1 dropped out because of hyperthyroidism, and 2 dropped out for personal reasons. Ultimately, 95 participants completed the study (Fig. 1), and most (85.30%) were overweight or obese according to the World Health Organization classification for Asian populations (BMI ≥23.0 kg m⁻¹ and ≥27.5 kg m⁻², respectively).¹⁴ The flowchart in Fig. 1 shows the number of participants who were recruited, randomly assigned, dropped out, and analysed during the study.

Fig. 1 Study design and consort flowchart. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol.

2.2. Test diet

The DAG oil and regular rapeseed cooking oil (TAG oil) were obtained from Fastco Biotech Co., Ltd (Hangzhou, China). The DAG oil was prepared from natural rapeseed oil and glycerol using Novozyme 435 (Candida antarctica lipase B).¹⁵ After molecular distillation to remove monoglycerides and free fatty acids from the crude product, it was deodorized by stripping to obtain DAG oil, which was more than 80% pure by weight, at a ratio of 66.80 : 33.20 for 1,3-DAG to 1,2-DAG. The prepared TAG oil was used as the raw oil to produce DAG oil, which had a DAG concentration of approximately 2.10%. The DAG and TAG oils have similar contents of monoacylglycerol and free fatty acid and a similar composition of fatty acids but differ only in the contents of DAG and TAG, which are shown in Table 1. Given the minimal usage of cooking oil in Chinese breakfasts, only lunch and dinner were provided to participants daily during the intervention period (limited to 30.00 g of cooking oil per person per day). The participants in the DAG group were provided with food cooked with DAG-rich cooking oil, whereas those in the TAG group were provided with regular rapeseed cooking oil. The average daily intake of oil per person in the DAG group was 22.39 g, whereas that in the TAG group was 21.67 g. The food was uniformly prepared in the school canteen, with the same recipes and meal processing, except for the use of the different cooking oils. Diet consumption was not restricted during the intervention period. After the meals were prepared, the researchers packaged them in portions using four dishes per meal for every group. The dishes for the DAG and TAG groups were prepared and packaged separately with labels to distinguish them.

Table 1 The fatty acid composition of DAG and TAG oils

Constituents	DAG oil	TAG oil
Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol.
Acylglycerols, g per 100 g
DAG	81.34	2.10
Monoacylglycerol	1.33	0.20
TAG	17.25	96.80
Free fatty acid	0.08	0.05
Fatty acids, g per 100 g
C14:0	0.090	0.060
C16:0	6.380	4.070
C16:1	0.290	0.235
C17:0	0.080	0.050
C18:0	1.980	1.710
C18:1	64.320	62.920
C18:2	17.190	18.990
C18:3	6.465	8.280
C20:0	0.860	0.570
C20:1	1.130	1.140
C20:2	0.050	0.080
C22:0	0.235	0.350
C22:1	0.110	0.040
C24:0	0.060	0.220
C24:1	0.060	0.150

---

The raw materials (including cooking oil and seasonings) and the weight of each meal were objectively documented prior to cooking for 3 consecutive days (2 weekdays and 1 weekend) pre-intervention (week 0) and post-intervention (week 8). All ingredients were prepared separately for the intervention, and not for other purposes. The participants were supplied with the pre-weighed dishes and staple meals in the school canteen, and at the end of the meal, the residual dishes and staple foods were weighed and recorded using an electronic scale (accurate to 0.005 kg). The nutritional consumption of the participants was evaluated over 3 days by integrating their self-reported intake of additional foods or beverages with their consumption of the supplied meals.

2.3. Anthropometry

Height and BW were measured using a height and weight measuring instrument (Meifu Electronics Co., Ltd, Shenzhen, China). BMI was calculated as BW (kg) divided by height (m) squared. Waist and hip circumferences were measured using non-stretch anthropometry tape. Waist circumference was measured 1 cm above the umbilicus and hip circumference was measured at the greatest lateral lumbar level. The waist circumference/hip circumference ratio (W/H ratio) was calculated as waist circumference (cm) divided by hip circumference (cm). Triceps skinfold thickness was measured using skinfold callipers. All these indicators were measured after the participants removed clothing including shoes, hats, and coats. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured on the left arm using an automatic blood pressure monitor (YE680CR, Yuyue Medical Equipment & Supply Co., Ltd, Jiangsu, China)

2.4. Clinical and biomedical measurements

Blood samples were collected from fasted participants at weeks 0, 4, and 8. All analyses were performed using serum samples. The detection indicators were as follows: (1) fat metabolism-related indicators. Serum total cholesterol (TC) and TG concentrations were measured enzymatically. High-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) were measured directly. Serum concentrations of apolipoprotein AI (ApoAI) and apolipoprotein B (ApoB) were measured using a turbidimetric immunoassay. Small dense low-density lipoprotein cholesterol (sdLDL-C) levels were measured using peroxide enzymatic methods. Serum lipoprotein (α) (Lp (α)) was measured by turbidimetric immune assays. Lipoprotein-associated phospholipase A2 (Lp-PLA2), very-low-density lipoprotein (VLDL), and oxidized low-density lipoprotein (Ox-LDL) levels were measured using enzyme-linked immunosorbent assay (ELISA) kits.¹⁶ (2) Liver and kidney function-related indicators. The concentrations of alanine aminotransferase (ALT) were measured using the alanine substrate method, and the concentrations of aspartate transaminase (AST) were measured using the rate method. Creatinine (Cr) was measured using enzymatic kinetics. (3) Glucose metabolism-related indicators. The glucose panel (GLU) was determined using the glucose oxidase method. Fasting insulin (FINS) was measured using chemiluminescence. Insulin resistance was analysed as a model assessment for insulin resistance (HOMA-IR), which is calculated as (FINS × GLU)/22.5. All ELISA kits were purchased from Meibiao Biotechnology Co., Ltd (Jiangsu, China) and Autobio Diagnostics Co., Ltd (Zhengzhou, China).

2.5. Clinical analysis

The morphology of the liver was examined using conventional ultrasound at 2–5.5 MHz (R10, SAMSUNG Electronics Co., Ltd, Republic of Korea). Hepatic steatosis severity was qualitatively scored by quantified physicians based on the echogenicity of the liver parenchyma, diaphragm condition, and blood vessel walls within the liver, indicating none, mild, moderate, and severe fatty livers, respectively.¹⁷ Tissue attenuation imaging (TAI) and tissue scatter distribution imaging (TSI) were used to quantitatively examine and diagnose hepatic steatosis. Visceral fat area (VFA) and total body fat (TBF) were examined using dual-energy X-rays (HORIZON-A-CN, HOLOGIC Inc., Marlborough, MA, USA). The above indicators were measured in a fasting state, with the participants removing their outer clothing and metal items. All the measurements were performed by licensed physicians.

2.6. Lipidomics

Lipidomic analyses were performed using serum samples from liquid chromatography–mass spectrometry (LC-MS).¹⁸ Initially, the samples were pre-treated to eliminate proteins and contaminants, followed by the extraction of lipid metabolites. Subsequently, data were acquired by detecting metabolites in both the positive and negative ionization modes from the LC-MS, yielding mass spectrometry and tandem mass spectrometry data. These data were processed using LipidSearch software (Thermo Fisher Scientific, Waltham, MA, USA), which facilitated reverse peak extraction, alignment, and identification, ultimately resulting in a comprehensive data matrix that included retention time, peak area, mass-to-charge ratio, and identification details. This process produced a list of metabolites along with a data matrix, which were imported into the Majorbio cloud platform for preprocessing. Differential metabolites were identified using orthogonal partial least squares discriminant analysis (OPLS-DA) based on the Variable Importance in Projection value (VIP > 1) and P value (<0.05). Additional personalized analyses such as a chemical bond, saturation, and Kyoto Encyclopaedia of Genes and Genomes analysis were performed to extract biological information from the differential lipids.

2.7. Statistical analyses

The data were tested for normality using the Kolmogorov–Smirnov test, presented in means ± standard deviation and analysed using the Statistical Package for the Social Sciences version 27.0 (IBM Corp., Armonk, NY, USA). Significant differences between the two groups were assessed using the analysis of covariance (ANCOVA), t-test or chi-square test, with a P value < 0.05 indicating a significant difference. An analysis of variance (ANOVA) was employed to analyse the significant differences between each time point and the baseline data within each group. The graphs were created using GraphPad Prism (version 9.5, GraphPad Software., Boston, MA, USA), Origin (version 2021, OriginLab Corp., Northampton, MA, USA), and R (version 4.2.2) programming language.

3. Results

After the 8-week intervention, the participants had good compliance with the research protocol, and no adverse reactions were reported. The distribution of two groups of participants is illustrated in Fig. 1. A total of 95 participants (50 in the DAG group and 45 in the TAG group) successfully completed the trial. The baseline characteristics of participants are presented in Table 2. The DAG and TAG groups did not differ significantly in demographic and baseline variables. Although significant baseline differences in protein and cholesterol intakes were observed, the 8-week intervention revealed no between-group differences in energy and nutrient intakes, and physical activity levels or their changes throughout the intervention period (Table 3).

Table 2 Participant characteristics (n = 95)

Variables	DAG group, n = 50	TAG group, n = 45	P
Means ± standard deviations for both the DAG and TAG groups. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; BW, body weight; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; W/H ratio, waist circumference/hip circumference; LDL-C, low-density lipoprotein cholesterol; HDL-C, high density lipoprotein cholesterol; TG, total triacylglycerol; TC, total cholesterol; ApoAI, apolipoprotein AI; ApoB, apolipoprotein B; sdLDL-C, small dense low-density lipoprotein cholesterol; Lp (α), lipoprotein (α); Lp-PLA2, lipoprotein-associated phospholipase A2; VLDL, very low-density lipoprotein; Ox-LDL, oxidized low-density lipoprotein; Cr, creatinine; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GLU, glucose; FINS, fasting serum insulin; HOMA-IR, homeostasis model assessment of insulin resistance.
Age	24.86 ± 5.62	23.00 ± 3.97	0.068
Sex			0.843
Female	21 (42.00%)	18 (40.00%)
Male	29 (58.00%)	27 (60.00%)
BW (kg)	75.83 ± 9.86	76.56 ± 13.87	0.765
BMI (kg m^−2)	25.77 ± 2.83	26.11 ± 3.04	0.572
Waist circumference (cm)	87.82 ± 7.64	86.44 ± 8.13	0.395
Hip circumference (cm)	102.16 ± 4.60	101.69 ± 6.57	0.688
Skinfold thickness (cm)	19.25 ± 6.75	18.80 ± 7.12	0.752
W/H ratio	0.86 ± 0.05	0.85 ± 0.05	0.332
SBP (mmHg)	123.50 ± 14.90	123.04 ± 12.82	0.874
DBP (mmHg)	79.78 ± 8.90	80.42 ± 9.94	0.740
LDL-C (mmol L^−1)	2.57 ± 0.58	2.50 ± 0.60	0.567
HDL-C (mmol L^−1)	1.32 ± 0.29	1.23 ± 1.23	0.109
TG (mmol L^−1)	1.24 ± 0.62	1.19 ± 0.75	0.731
TC (mmol L^−1)	4.59 ± 0.73	4.41 ± 0.73	0.215
ApoAI (g L^−1)	1.41 ± 0.20	1.35 ± 0.17	0.122
ApoB (g L^−1)	0.77 ± 0.21	0.75 ± 0.75	0.601
sdLDL-C (mmol L^−1)	0.74 ± 0.24	0.73 ± 0.19	0.718
Lp (α) (mg L^−1)	220.22 ± 255.80	177.56 ± 160.20	0.339
Lp-PLA2 (ng mL^−1)	15.00 ± 4.83	15.38 ± 4.46	0.695
VLDL (mmol L^−1)	11.71 ± 3.39	11.87 ± 2.97	0.808
Ox-LDL (ng mL^−1)	583.86 ± 163.27	598.26 ± 153.20	0.660
Cr (μmol L^−1)	72.80 ± 15.57	71.16 ± 15.00	0.602
UREA (mmol L^−1)	4.57 ± 1.21	4.47 ± 0.78	0.623
ALT (U L^−1)	22.58 ± 16.29	21.20 ± 21.20	0.638
AST (U L^−1)	19.20 ± 7.76	18.51 ± 4.12	0.596
GLU (mmol L^−1)	4.38 ± 0.48	4.43 ± 0.41	0.582
FINS (μU mL^−1)	14.17 ± 6.77	13.48 ± 7.21	0.631
HOMA-IR	2.79 ± 1.44	2.64 ± 2.64	0.608

---

Table 3 Dietary intake and MET in the DAG and TAG groups

		DAG, n = 50	TAG, n = 45	P
Means ± standard deviations for both the DAG and TAG groups. Δ, the difference at week 8 from the baseline. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; SFASs, saturated fatty acids; MUFAs, monounsaturated fatty acids; PUFAs, polyunsaturated fatty acids; MET, metabolic equivalent of task.
Week 0 (baseline)	Energy (kcal)	3247.58 ± 587.61	3392.37 ± 636.76	0.252
	Protein (g)	116.56 ± 21.56	126.68 ± 21.21	0.024
	Fat (g)	89.04 ± 18.22	93.42 ± 18.80	0.252
	Dietary fiber (g)	12.00 ± 2.67	12.50 ± 2.01	0.307
	Carbohydrate (g)	223.48 ± 55.75	226.52 ± 54.90	0.790
	Vitamin E (mg)	11.12 ± 2.64	11.88 ± 1.81	0.108
	Vitamin C (mg)	139.91 ± 34.57	152.02 ± 28.74	0.068
	Cholesterol (mg)	1004.67 ± 304.50	1127.32 ± 286.19	0.047
	SFAs (g)	15.26 ± 7.80	18.60 ± 9.81	0.068
	MUFAs (g)	23.21 ± 7.46	25.94 ± 9.62	0.124
	PUFAs (g)	22.68 ± 3.88	22.71 ± 4.55	0.972
	MET (min per week)	1484.03 ± 1500.24	1685.14 ± 1445.26	0.918
Week 8	Energy (kcal)	3089.35 ± 600.39	3152.59 ± 776.51	0.656
	Protein (g)	104.39 ± 25.49	106.72 ± 29.24	0.680
	Fat (g)	72.37 ± 14.58	73.31 ± 18.36	0.782
	Dietary fiber (g)	18.79 ± 5.08	19.33 ± 6.82	0.661
	Carbohydrate (g)	222.51 ± 43.14	228.27 ± 57.48	0.580
	Vitamin E (mg)	16.08 ± 4.25	16.75 ± 4.66	0.467
	Vitamin C (mg)	113.41 ± 34.21	120.42 ± 62.87	0.496
	Cholesterol (mg)	914.04 ± 533.46	840.92 ± 380.27	0.448
	SFAs (g)	8.08 ± 2.19	7.27 ± 2.63	0.107
	MUFAs (g)	13.63 ± 3.66	12.18 ± 4.48	0.086
	PUFAs (g)	14.30 ± 3.71	13.02 ± 4.73	0.142
	MET (min per week)	1749.95 ± 1808.96	1912.68 ± 1708.99	0.508
Δ	Energy (kcal)	−158.24 ± 877.21	−239.78 ± 1075.07	0.685
	Protein (g)	−12.17 ± 36.23	−19.96 ± 38.99	0.315
	Fat (g)	−16.67 ± 26.40	−20.11 ± 29.00	0.547
	Dietary fiber (g)	6.79 ± 5.66	6.83 ± 7.30	0.978
	Carbohydrate (g)	−0.97 ± 71.03	1.75 ± 77.63	0.859
	Vitamin E (mg)	4.96 ± 4.86	4.87 ± 5.13	0.928
	Vitamin C (mg)	−26.50 ± 43.66	−31.61 ± 68.64	0.663
	Cholesterol (mg)	−90.63 ± 630.65	−286.39 ± 423.81	0.082
	SFAs (g)	−7.18 ± 9.07	−11.33 ± 11.30	0.051
	MUFAs (g)	−9.58 ± 9.64	−13.76 ± 12.18	0.066
	PUFAs (g)	−8.38 ± 5.02	−9.69 ± 6.75	0.281
	MET (min per week)	265.92 ± 1667.94	227.53 ± 1945.13	0.654

---

3.1. Anthropometry

After adjusting for the baseline, the ANCOVA revealed significant differences in waist circumference and W/H ratio between the two groups at week 4 (P = 0.041, P = 0.017, respectively), although no inter-group differences were found in other anthropometry indicators (Fig. 2A–E). Both the DAG and TAG groups had reductions in average BW, BMI, waist circumference, and hip circumference during the intervention period. Compared with the baseline, the waist circumference and hip circumference were significantly reduced in the DAG group at week 8 (P = 0.037 and P = 0.002, respectively) (Fig. 3C and D). However, no significant differences were observed in the TAG group when compared with the baseline (Fig. 3A–E). It is noteworthy that the changes in waist circumference and W/H ratio in the DAG group were significantly higher than those in the TAG group (P = 0.032, P = 0.018, respectively) (Fig. 4A–E).

Fig. 2 Comparison of anthropometry and blood indicators between the DAG group and the TAG group (means ± standard deviations). *Significant difference between the groups based on the results of ANCOVA adjusted for the baseline. (*P < 0.05). (A) BW; (B) BMI; (C) Waist circumference; (D) Hip circumference; (E) W/H Ratio; (F) LDL-C; (G) HDL-C; (H) TG; (I) TC; (J) ApoAI; (K) sdLDL-C; (L) Lp (α); (M) Lp-PLA2; (N) VLDL; (O) Ox-LDL; (P) GLU; (Q) FINS; (R) HOMA-IR. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; BW, body weight; BMI, body mass index; W/H ratio, waist circumference/hip circumference; LDL-C, low-density lipoprotein cholesterol; HDL-C, high density lipoprotein cholesterol; TG, total triacylglycerol; TC, total cholesterol; ApoAI, apolipoprotein AI; sdLDL-C, small dense low-density lipoprotein cholesterol; Lp (α), lipoprotein (α); Lp-PLA2, lipoprotein-associated phospholipase A2; VLDL, very low-density lipoprotein; Ox-LDL, oxidized low-density lipoprotein; GLU, glucose; FINS, fasting serum insulin; HOMA-IR, homeostasis model assessment of insulin resistance.

Fig. 3 Comparison of anthropometry and blood indicators between the baseline and different intervention times in both the DAG and TAG groups (means ± standard deviations). *Significant difference from the baseline based on the results of ANOVA. (*P < 0.05, **P < 0.01, ***P < 0.001). (A) BW; (B) BMI; (C) Waist circumference; (D) Hip circumference; (E) W/H Ratio; (F) LDL-C; (G) HDL-C; (H) TG; (I) TC; (J) ApoAI; (K) sdLDL-C; (L) Lp (α); (M) Lp-PLA2; (N) VLDL; (O) Ox-LDL; (P) GLU; (Q) FINS; (R) HOMA-IR. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; BW, body weight; BMI, body mass index; W/H ratio, waist circumference/hip circumference; LDL-C, low-density lipoprotein cholesterol; HDL-C, high density lipoprotein cholesterol; TG, total triacylglycerol; TC, total cholesterol; ApoAI, apolipoprotein AI; sdLDL-C, small dense low-density lipoprotein cholesterol; Lp (α), lipoprotein (α); Lp-PLA2, lipoprotein-associated phospholipase A2; VLDL, very low-density lipoprotein; Ox-LDL, oxidized low-density lipoprotein; GLU, glucose; FINS, fasting serum insulin; HOMA-IR, homeostasis model assessment of insulin resistance.

Fig. 4 Comparison of changes in anthropometry and blood indicators between the DAG group and the TAG group (means ± standard deviations). *Significant difference between the groups. (*P < 0.05). Δ, the difference at each stage of intervention from the baseline. (A) BW; (B) BMI; (C) Waist circumference; (D) Hip circumference; (E) W/H Ratio; (F) LDL-C; (G) HDL-C; (H) TG; (I) TC; (J) ApoAI; (K) sdLDL-C; (L) Lp (α); (M) Lp-PLA2; (N) VLDL; (O) Ox-LDL; (P) GLU; (Q) FINS; (R) HOMA-IR. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; BW, body weight; BMI, body mass index; W/H, waist circumference/hip circumference; LDL-C, low-density lipoprotein cholesterol; HDL-C, high density lipoprotein cholesterol; TG, total triacylglycerol; TC, total cholesterol; ApoAI, apolipoprotein AI; sdLDL-C, small dense low-density lipoprotein cholesterol; Lp (α), lipoprotein (α); Lp-PLA2, lipoprotein-associated phospholipase A2; VLDL, very low-density lipoprotein; Ox-LDL, oxidized low-density lipoprotein; GLU, glucose; FINS, fasting serum insulin; HOMA-IR, homeostasis model assessment of insulin resistance.

3.2. Blood indicators

The results of blood indicators are displayed in Fig. 2F–J, 3F–J, and 4F–J. Based on the results of ANCOVA, there were significant differences in LDL-C and sdLDL-C between the groups after 4 weeks of administration (P = 0.019, P = 0.034, respectively) (Fig. 2F and K). By the 8-week follow-up, significant differences were found in TG and sdLDL-C between the two groups (P = 0.026, P = 0.024, respectively) (Fig. 2H and K). Additionally, both groups exhibited significant reductions in serum LDL-C and HDL-C levels compared with baseline measurements (Fig. 3F–G). Notably, the serum TG levels, the primary outcomes, were significantly lower at week 4 than at the baseline (P = 0.001) in the DAG group (Fig. 3H). Similarly, the serum TC levels in the DAG group were significantly lower at week 8 than that at the baseline (P = 0.048) (Fig. 3I). Further analysis revealed that there were no changes in LDL-C, HDL-C, TG, TC, or ApoAI concentrations between the groups after 4 weeks of administration. However, the variation in ApoAI after 8 weeks of administration was significantly lower in the DAG group (P = 0.025) (Fig. 4J). The variation trend for serum TG of the primary outcomes in the DAG group was superior to that for the TAG group (P = 0.025) after 8 weeks of administration (Fig. 4H). The results of lipid-related indicators are shown in Fig. 2K–O, 3K–O, and 4K–O. Although there were no significant changes in Lp (a), Lp-PLA2, VLDL, and Ox-LDL concentrations between the groups, a significant decrease in sdLDL-C was observed in the DAG group compared to that in the TAG group after 4 weeks and 8 weeks of administration (P = 0.021, Fig. 2K). Correspondingly, the sdLDL-C concentration was consistently reduced in the DAG group compared to the baseline (Fig. 3K).

The results of the glucose-related indicators are shown in Fig. 2P–R, 3P–R, and 4P–R. Although significant differences in GLU, FINS, and HOMA-IR concentrations between the groups were not observed, significant decreases in GLU at week 4 were found in both groups compared to the baseline (Fig. 3P), which might be related to the healthy diet administration.

3.3. Clinical analysis

The analysis of fatty liver diagnoses indicated a reduction in the number of patients exhibiting severe fatty liver in the DAG group following the intervention (P = 0.035) (Fig. 5A and B). Conversely, no significant changes were observed in the TAG group in patients with severe fatty liver disease. However, there were no significant differences in TSI and TAI based on ultrasonic examination (Fig. 5C and D). Notably, TBF and VFA in the DAG group showed more significant decreases after the 8-week intervention when compared to the baseline (P = 0.016 and P = 0.001, respectively) (Fig. 5E–G).

Fig. 5 Effect of the DAG intervention on liver fat and body fat. *Significant difference from the baseline in the DAG group based on the results of ANOVA. (*P < 0.05, **P < 0.01). A: Results of the semi-quantitative analysis of fatty liver. B: Ultrasound diagnostic detection morphology of the liver at the baseline (above) and week 8 (below) of a participant in the DAG intervention group. C: Change in the TSI of the liver relative to the baseline. D: Change in the TAI of the liver relative to the baseline. E: Change in VFA relative to the baseline. F: Change in TBF relative to the baseline. G: Dual-energy X-ray detection images at the baseline and week 8 in the TAG and DAG groups. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; Severe, severe fatty liver; Moderate, moderate fatty liver; Mild, mild fatty liver; None, non-fatty liver; TSI, tissue attenuation imaging; TAI, tissue scatter distribution imaging; VFA, visceral fat area; TBF, total body fat.

3.4. Lipidomics analysis

Lipidomics was performed for serum samples from all participants at weeks 0, 4, and 8. After data preprocessing, 1011 lipid metabolites were detected in the positive and negative ion detection modes, which were mainly classified into five groups using LIPID MAPS (https://www.lipidmaps.org) that included fatty acids, glycerolipids, glycerophospholipids, sphingolipids, and sterol lipids. A principal component analysis of the lipid metabolite species in both anionic and cationic modes was performed separately for the two groups at weeks 0, 4, and 8 (Fig. 6A). A representative point of the individuals in the same group was projected to clearly observe the changes after the intervention. Notably, there was no significant difference between the lipid species of the two groups before the intervention, whereas the distance between the two groups gradually increased as the intervention continued (especially in the anionic mode). We counted the differences in the relative expression of metabolites between two groups in each period (Fig. 6B) and the number of unique/shared metabolites between two groups in each period (Fig. 6C). The results showed that a considerable number of metabolites had intergroup and intragroup differences before and after the intervention. Notably, DAG administration increased metabolites at weeks 4 (n = 194) and 8 (n = 163) compared to the TAG group, suggesting that the DAG intervention may have promoted activation of certain metabolic pathways, resulting in an overall change in metabolite expression patterns. We summarized all lipid subclasses that had significantly different changes during the intervention and found that intergroup differences in metabolites were very limited before the intervention. However, lipid subclasses such as sphingomyelin (phytosphingosine) (phSM), lyso-phosphatidylethanol, gangliosides, and monogalactosyldiacylglycerol had significant changes (among them, TG, diacylglycerol (DG), phosphatidylethanolamine (PE), and cardiolipin (CL) had the most significant changes) after 8-weeks of intervention, with up-regulated expressions of TG, PE, and DG, and down-regulated expression of phSM and CL, in the DAG group compared to the TAG group (Fig. 6D). To further explore the relationship between lipid metabolites and VFA, we focused on observing the significant changes in the TG and PE characteristics, which had the most significant changes (Fig. 6D and 7A, B). The analysis revealed that the TGs and PEs containing saturated fatty acids (SFAs) exhibited no significant differences between the groups at week 0 but had a notable decrease at week 8 in both groups. The DAG group had significantly higher levels of TGs containing monounsaturated fatty acids (MUFAs) and polyunsaturated fatty acids (PUFAs) than those in the TAG group; a similar trend was observed for PEs. Although TGs and PEs containing SFAs exhibit high levels in the DAG group, TGs and PEs containing MUFAs and PUFAs were also significantly increased compared to the TAG group. Specifically, in the TGs, the number of double bonds (N = 2, N = 3, N = 7, N = 8, and N = 12; all P < 0.001) increased in the DAG intervention group after the 8-week intervention when compared to the TAG group. For the PEs, the number of double bonds—N = 2 (P < 0.001), N = 4 (P = 0.017), and N = 6 (P < 0.001)—also increased in the DAG group after the 8-week intervention compared to those in the TAG group.

Fig. 6 Effect of the DAG intervention on serum lipidomics. A: A principal component analysis for the metabolites’ changes at weeks 0, 4, and 8; the upper image shows that the detection was in cationic mode, and the lower one shows that the detection was in anion mode. B: The number of compounds with significant difference intragroup and intergroup. C: The Venn diagram of the number of shared and unique metabolites. D: Differential abundance score between the two groups at weeks 0 and 8. Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; POS, positive ion detection; NEG, negative ion detection CerG2GNAc1, Simple Glc series; LPMe, lyso-phosphatidylethanol; PG, phosphatidylglycerol; GM3, gangliosides; MGDG, monogalactosyldiacylglycerol; PA, phosphatidic acid; LPE, lyso-phosphatidylethanolamine; FA, fatty acid; PS, phosphatidylserine; OAHFA, (O-acyl)-1-hydroxy fatty acid; LPI, lyso-phosphatidylinositol; dMePE, dimethylphosphatidyl-ethanolamine; PI, phosphatidylinositol; Hex2Cer, sphingosine ceramides; MLCL, monolysocardiolipin; LdMePE, lysodimethylphosphatidylethanolamine; MG, monoglyceride; CerP, ceramides phosphate; ZyE, zymosteryl; Cer, ceramides; HeX1Cer, sphingosine ceramides; CL, cardiolipin; AcCa, acyl carnitine; PE, phosphatidylethanolamine; DG, total diacylglycerol; MePC, methylatedphosphatidylcholine; TG, total triacylglycerol; PC, phosphatidylcholine; SM, sphingomyelin; phSM, sphingomyelin (phytosphingosine); PEt, phosphatidylethanol; PIP2, phosphatidylinositol; CmE, clathrin-mediated endocytosis; SPH, sphingosine; WE, wax esters; ChE, cholesteryl ester.

Fig. 7 Effect of the DAG intervention on metabolic pathways of serum lipidomics. A: Comparison of the TG double bond numbers for the two groups at weeks 0 and 8. B: Comparison of the PE double bond numbers for the two groups at weeks 0 and 8 (the blank indicates that it was not detected). C: KEGG classification between the two groups at week 8. D: Differential abundance scores between the two groups at week 8. *Significant difference between the groups. (*P < 0.05, **P < 0.01, ***P < 0.001). Abbreviations: DAG, diacylglycerol; TAG, triacylglycerol; TG, total triacylglycerol; PE, phosphatidylethanolamine; KEGG, Kyoto Encyclopedia of Genes and Genomes.

To clarify the metabolic mechanisms underlying hepatic lipid accumulation and serum TG changes, we performed pathway enrichment analyses using metabolites and lipid species that were significantly altered at different times. Fig. 7C shows the metabolic pathway enrichment analysis which revealed that the lipid metabolic pathway was the predominant metabolic pathway in the DAG group after the intervention. The analyses also revealed that the expression of pathways involved in the regulation of adipocyte lipolysis, thermogenesis, fat digestion and absorption tended to be upregulated in the DAG group after the intervention (Fig. 7D), suggesting that these pathways may be involved in VFA decline.

4. Discussion

The objective of this clinical trial was to evaluate the effect of substituting conventional cooking oil with DAG oil as the main source of dietary fat on lipid metabolism and accumulation in the human body. This study implemented an 8-week dietary intervention involving an overweight/obese or central obese population with an average BMI of 25.93 ± 2.92 kg m⁻². Significant decreases in waist circumference, TG and sdLDL-C levels, and significant improvements in TBF and VFA were observed in the DAG intervention group. Notably, patients in the DAG intervention group with severe fatty liver exhibited symptom improvement after the intervention, whereas the number of patients with severe fatty liver in the TAG group remained unchanged. The serum lipidomic analysis revealed that the DAG intervention caused considerable amounts of changes in metabolites, particularly a significant increase in TGs and PEs containing MUFAs and PUFAs. The expression of pathways such as the regulation of lipolysis in adipocytes, thermogenesis, and fat digestion and absorption in the DAG group exhibited elevated activity in response to the intervention, which might be involved in the declines of VFA and liver fat.

Following an 8-week intervention, both the DAG and TAG groups exhibited reductions in BW and BMI, with the DAG intervention group having a more substantial decrease than the TAG group. However, no significant differences were observed in either the intragroup or intergroup comparisons. A clinical trial indicated that after a 16-week intervention, both DAG and TAG intakes were associated with BW reduction.¹⁹ Notably, the DAG group experienced a significant decrease in BW and BMI compared to their baseline values, although no significant differences were found between the two groups. In contrast, a 1-year dietary intervention revealed that the BW of the participants in the DAG group was significantly lower than those at the baseline, whereas no changes were observed in the TAG group.²⁰ Furthermore, the research conducted by Yamamoto et al.²¹ demonstrated that after a 6-month clinical intervention trial, the group that consumed DAG exhibited BW reduction, whereas the group that consumed TAG experienced BW gain. The discrepancies observed in our studies may be attributed to the relatively short intervention durations and participant variations in BW loss responses, which led to non-significant changes in metrics such as BW and BMI.

Numerous studies have demonstrated a significant reduction in serum TG levels in both DAG and TAG groups following the intervention.^22,23 Consumption of oils high in DAG markedly decreases serum TG levels and other related indicators in C57BL/6J mice.²⁴ In this study, assessments of serum lipid metabolism-related indicators revealed that serum TG levels in both groups initially declined before subsequently increasing. At the end of the intervention, there was a significant difference in serum TG between the groups, and the TAG group exhibited elevated serum TG levels relative to the baseline values. Notably, the DAG group experienced a significant reduction in serum TG levels compared to the baseline at the mid of the intervention. In a study conducted by Meguro et al.,²⁵ the DAG group had a slight decrease in serum TG levels post-intervention, whereas the TAG group experienced a slight increase; however, this change was not significant. The results of our study align closely with those of the aforementioned study. Nevertheless, there remains a lack of a plausible explanation for the pronounced decrease in serum TG levels observed in both groups at the mid of the intervention, warranting further investigation in conjunction with lipidomic data. Additionally, this study found that post-intervention TC and sdLDL-C levels in the DAG intervention group significantly decreased compared to the baseline. sdLDL-C is positively associated with the risk of MAFLD in populations without obesity and indicates hepatic-centered metabolic abnormalities.^26,27 Moreover, previous research has demonstrated a strong association between sdLDL-C and atherosclerosis,²⁸ suggesting that reducing sdLDL-C levels may improve blood lipid profiles and provide new perspectives for atherosclerosis prevention.²⁹

Extensive research has determined that DAG consumption may improve substantial insulin resistance which is the common pathophysiologic basis of multiple metabolic diseases (such as obesity, type 2 diabetes mellitus and hypertension),³⁰ which are associated with the development of MAFLD. A previous study demonstrated that significant improvements of glucose metabolism were observed in diabetic participants with normal BW after DAG oil intervention;³¹ however, the positive effects of DAG oil on glucose metabolism in overweight/obese individuals were not found.³² Similarly, our present study showed that there was no significant change in HOMA-IR in both groups, indicating that insulin resistance in overweight/obese participants was not improved by short-term DAG intervention.

Reductions in BW, TBF, and localized fat accumulation, especially in visceral adipose fat, were observed.¹⁰ Mice showed a marked decrease in adipose fat tissue in a 12-week study;³³ the DAG group showed a significant reduction in TBF compared to the baseline levels, whereas the TAG group did not. Although there were no significant effects on metrics such as BW and BMI, these results align with the results of the study by Ando et al.¹¹ Notably, all patients diagnosed with severe fatty liver in the DAG group exhibited improvements, transitioning to a mild or moderate fatty liver status post-intervention; however, no improvements were observed in the TAG group. Research has shown that inhibiting fat synthesis and reducing fatty acid uptake have beneficial therapeutic effects on MAFLD.³⁴ Therefore, it can be postulated that the distinct metabolic characteristics of DAG in the small intestine confer similar benefits. The ingestion of DAG is associated with reductions in VFA and TBF, ameliorating the pattern of adipose tissue deposition in the human body and potentially aiding in the prevention of obesity and MAFLD.

Although TAI and TSI offer reproducible and quantitative correlates of hepatic steatosis, their region-of-interest-based approach samples only a limited area of the liver.³⁵ This localized assessment may inadequately reflect heterogeneous fat distribution, especially following short-term nutritional interventions. The present study used a lipidomic analysis to explore the regulatory mechanisms of DAG intervention in lipid metabolism. Our results showed that the TBF and VFA (especially liver fat) were significantly improved in the DAG group, which is similar to the findings in model animals.³⁶ Furthermore, in our study, increased DG levels were observed in the DAG intervention group and led to changes in lipid profiles which presented as the upregulation of TGs and PEs containing MUFAs and PUFAs. Previous studies have shown that the therapeutic effects of lipids are mainly attributed to the type of TG and PE.^37,38 Increase in TGs containing PUFAs can alter lipid metabolism and phospholipids containing PUFAs are negatively correlated with liver fat.^39,40 DAG oil enhances fatty acid oxidation through the modulation of AMPK and other enzymes,⁴¹ which subsequently leads to alterations in the fatty acid composition of hepatic phospholipids, characterized notably by an increased proportion of PUFAs.⁴² A high intake of MUFAs and PUFAs can contribute to a reduction in steatosis in alleviating MAFLD;⁴³ notably, patients with MAFLD had a significant reduction in liver fat and beneficial changes in serum lipid profiles after long-term PUFA treatment.⁴⁴ Furthermore, PUFAs have been shown to oxidize more rapidly,⁴⁵ which positively affects lipid accumulation. This evidence suggests that PUFAs enriched in TGs contribute to the mobilization of adipose tissue, thereby reducing the accumulation of visceral fat in the body.

To further elucidate the effect of DAG-induced lipid metabolism on liver health, we analyzed the signaling and metabolic pathways associated with lipid metabolism using Kyoto Encyclopedia of Genes and Genomes analysis. The present study showed that the DAG intervention group had an overall tendency to upregulate the expression of pathways such as the regulation of lipolysis in adipocytes, thermogenesis, and fat digestion and absorption, which might contribute to the increasing energy expenditure and reducing lipid accumulation in adipocytes. One study revealed that stimulating pathways such as thermogenesis, fat digestion, and absorption may contribute to mitigating MAFLD.⁴⁶ Therefore, the improvement in DAG-related lipid metabolic pathways might help in reducing fat accumulation in visceral organs, particularly in the liver.

This study provided all participants with the same standardized meals without restricting food intake or energy, and the results are close to the intervention effects of diacylglycerol in real-world applications. Furthermore, this study employed dual-energy X-ray detection, which is recognized as the “gold standard” for assessing visceral fat distribution. Additionally, non-targeted lipid metabolomics was used to explore the potential mechanisms in which triacylglycerol influences lipid accumulation and metabolism. However, identifying specific lipid metabolites associated with DAG oil administration is challenging because of the short-term intervention and the variability among individuals; thus, the health effects of long-term DAG intervention are necessary as well as their persistence effect and rebound. In the preparation process of glycerol diester oil, enzymatic hydrolysis and molecular distillation are usually used, which inevitably leads to the loss of volatile components, plant sterols, and thermosensitive vitamins in the raw oil, which may affect the biological effects of DAG oil. The detection of key components would contribute to understanding the impact of DAG intake on body lipids. Furthermore, the study participants tend to be younger and overweight or mildly obese, which may limit the generalizability of the findings when applied to older adults and more obese populations. Although we provided standard meals throughout the study, individualized dietary needs and the difficulty in accurately assessing dietary intake might lead to differences in nutrient intake between the two groups at the baseline, thereby introducing some confusion regarding the intervention effects. The depth of lipidomics analysis is limited by relatively limited application in diacylglycerol oil research and the scarcity of reference data available for comprehensive mechanistic interpretation.

5. Conclusions

This randomized, double-blind, placebo-controlled trial with overweight/obese or central obese adults showed that consuming DAG oil instead of TAG oil reduced the levels of sdLDL-C, TBF, VFA, and liver fat. Increased levels of serum TGs and PEs containing MUFAs and PUFAs induced by DAG intervention might help reduce fat accumulation in visceral tissues, especially in the liver. These findings suggested that DAG-mediated lipid remodelling might be related to preventing lipid metabolic disorders through visceral fat regulation. Further studies are required to clarify the long-term effects of dietary DAG on lipid metabolism, fat accumulation, and cardiovascular diseases in humans.

Author contributions

Conceptualization: Y. X.; methodology: L. Q.; validation: L. Q., and Y. X.; formal analysis: L. Q.; investigation: L. Q., W. Z., Z. L., X. B., R. L. and Q. H.; resources: Z. Y.; data curation: L. Q.; writing—original draft preparation: L. Q.; writing—review and editing: L. Q., Z. Y. and Y. X.; visualization: L. Q.; supervision: W. Z., Z. L., X. B., R. L., Q. H., Q. S., H. J., Y. X. and Y. Z.; project administration: Y. X.; funding acquisition: Y. X. and Y. Z. All authors have read and agreed to the published version of the manuscript.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The funder of the study had no role in designing the study, patient recruitment, data collection, analysis and interpretation, writing of manuscripts, the decision to submit for publication, or other aspects pertinent to the study.

Abbreviations

ANCOVA	Analysis of covariance
ANOVA	Analysis of variance
ApoAI	Apolipoprotein AI
ApoB	Apolipoprotein B
BMI	Body mass index
BW	Body weight
CL	Cardiolipin
Cr	Creatinine
DAG	Diacylglycerol
DG	Diacylglycerol
ELISA	Enzyme-link immunosorbent assay
FINS	Fasting insulin
GLU	Glucose
HDL-C	High density lipoprotein cholesterol
HOMA-IR	Homeostasis model assessment of insulin resistance
LC-MS	Liquid chromatography-mass spectrometry
LDL-C	Low-density lipoprotein cholesterol
Lp (α)	Lipoprotein α
Lp-PLA2	Lipoprotein-associated phospholipase A2
MAFLD	Metabolic dysfunction-associated fatty liver disease
MUFA	Monounsaturated fatty acids
Ox-LDL	Oxidized low-density lipoprotein
PE	Phosphatidylethanolamine
phSM	Sphingomyelin (phytosphingosine)
PUFA	Polyunsaturated fatty acid
SBP	Systolic blood pressure
sdLDL-C	Small dense low-density lipoprotein cholesterol
SFA	Saturated fatty acid
TAG	Triacylglycerol
TAI	Tissue attenuation imaging
TBF	Total body fat
TC	Total cholesterol
TG	Total triacylglycerol
TSI	Tissue scatter distribution imaging
VFA	Visceral fat area
VLDL	Very low-density lipoprotein
W/H ratio	Waist circumference/hip circumference ratio

Data availability

The raw data supporting the conclusions of this article will be made available upon request.

Acknowledgements

The trial was funded by the National Key Research and Development Program of China (2021YFD2100200/2021YFD2100201). The authors would like to thank the fellows in the team and nurses who offered assistance for the progress of this trial.

References

1. Q. Pang, J. Y. Zhang, S. D. Song, K. Qu, X. S. Xu, S. S. Liu and C. Liu, Central obesity and nonalcoholic fatty liver disease risk after adjusting for body mass index, World J. Gastroenterol., 2015, 21, 1650–1662.
2. D. Ferrara, F. Montecucco, F. Dallegri and F. Carbone, Impact of different ectopic fat depots on cardiovascular and metabolic diseases, J. Cell Physiol., 2019, 234, 21630–21641.
3. J. Ballena-Caicedo, F. E. Zuzunaga-Montoya, J. A. Loayza-Castro, L. E. M. Vasquez-Romero, R. Tapia-Limonchi, C. I. G. De Carrillo and V. J. Vera-Ponce, Global prevalence of dyslipidemias in the general adult population: a systematic review and meta-analysis, J. Health Popul. Nutr., 2025, 44, 308.
4. S. Subramanian and A. Chait, Hypertriglyceridemia secondary to obesity and diabetes, Biochim. Biophys. Acta, 2012, 1821, 819–825.
5. C. Zhou, M. Wang, J. Liang, G. He and N. Chen, Ketogenic Diet Benefits to Weight Loss, Glycemic Control, and Lipid Profiles in Overweight Patients with Type 2 Diabetes Mellitus: A Meta-Analysis of Randomized Controlled Trails, Int. J. Environ. Res. Public Health, 2022, 19, 10429.
6. B. Pourrajab, E. Sharifi-Zahabi, S. Soltani, H. Shahinfar and F. Shidfar, Comparison of canola oil and olive oil consumption on the serum lipid profile in adults: a systematic review and meta-analysis of randomized controlled trials, Crit. Rev. Food Sci. Nutr., 2023, 63, 12270–12284.
7. B. D. Flickinger and N. Matsuo, Nutritional Characteristics of DAG Oil, Lipids, 2003, 38, 129–132.
8. M. M. Kamphuis, D. J. Mela and M. S. Westerterp-Plantenga, Diacylglycerols affect substrate oxidation and appetite in humans, Am. J. Clin. Nutr., 2003, 77, 1133–1139.
9. H. Taguchi, H. Watanabe, K. Onizawa, T. Nagao, N. Gotoh, T. Yasukawa, R. Tsushima, H. Shimasaki and H. Itakura, Double-blind controlled study on the effects of dietary diacylglycerol on postprandial serum and chylomicron triacylglycerol responses in healthy humans, J. Am. Coll. Nutr., 2000, 19, 789–796.
10. T. Nagao, H. Watanabe, N. Goto, K. Onizawa, H. Taguchi, N. Matsuo, T. Yasukawa, R. Tsushima, H. Shimasaki and H. Itakura, Dietary Diacylglycerol Suppresses Accumulation of Body Fat Compared to Triacylglycerol in Men in a Double-Blind Controlled Trial, J. Nutr., 2000, 130, 792–797.
11. Y. Ando, S. Saito, S. Oishi, N. Yamanaka, M. Hibi, N. Osaki and Y. Katsuragi, Alpha Linolenic Acid-enriched Diacylglycerol Enhances Postprandial Fat Oxidation in Healthy Subjects: A Randomized Double-blind Controlled Trail, J. Oleo Sci., 2016, 65, 685–691.
12. K. R. Harshitha and M. Bhargava, Mid-upper arm circumference and neck circumference to screen for overweight-obesity in young adults in South India, Heliyon, 2022, 8, e12173.
13. R. Vazquez-Frias, A. Consuelo-Sanchez, C. P. Acosta-Rodriguez-Bueno, A. Blanco-Montero, D. C. Robles, V. Cohen, D. Marquez and M. Perez 3rd, Efficacy and Safety of the Adjuvant Use of Probiotic Bacillus clausii Strains in Pediatric Irritable Bowel Syndrome: A Randomized, Double-Blind, Placebo-Controlled Study, Paediatr. Drugs, 2023, 25, 115–126.
14. S. Yue, V. T. K. Thi, L. P. Dung, B. T. H. Nhu, E. Kestelyn, D. T. Thuan, L. Q. Thanh and J. E. Hirst, Clinical consequences of gestational diabetes mellitus and maternal obesity as defined by asian BMI thresholds in Viet Nam: a prospective, hospital-based, cohort study, BMC Pregnancy Childbirth, 2022, 22, 195.
15. Y. Mao, D. Zheng, L. He and J. Chen, The Lipid-Metabolism-Associated Anti-Obesity Properties of Rapeseed Diacylglycerol Oil, Nutrients, 2024, 16, 2003.
16. L. Wu, P. Shao, Z. Gao, S. Zhang, J. Ma, J. Bai and Y. Wei, Homocysteine and Lp-PLA2 levels: Diagnostic value in coronary heart disease, Medicine, 2023, 102, e35982.
17. G. Polti, F. Frigerio, G. Del Gaudio, P. Pacini, V. Dolcetti, M. Renda, S. Angeletti, M. Di Martino, G. Iannetti, F. M. Perla, E. Poggiogalle and V. Cantisani, Quantitative ultrasound fatty liver evaluation in a pediatric population: comparison with magnetic resonance imaging of liver proton density fat fraction, Pediatr. Radiol., 2023, 53, 2458–2465.
18. J. Wang, C. Wang and X. Han, Tutorial on lipidomics, Anal. Chim. Acta, 2019, 1061, 28–41.
19. K. Tomonobu, T. Hase and I. Tokimitsu, Dietary diacylglycerol in a typical meal suppresses postprandial increases in serum lipid levels compared with dietary triacylglycerol, Nutrition, 2006, 22, 128–135.
20. H. Kawashima, H. Takase, K. Yasunaga, Y. Wakaki, Y. Katsuragi, K. Mori, T. Yamaguchi, T. Hase, N. Matsuo, T. Yasukawa, I. Tokimitsu and W. Koyama, One-year ad libitum consumption of diacylglycerol oil as part of a regular diet results in modest weight loss in comparison with consumption of a triacylglycerol control oil in overweight Japanese subjects, J. Am. Diet. Assoc., 2008, 108, 57–66.
21. K. Yamamoto, K. Tomonobu, H. Asakawa, K. Tokunaga, T. Hase, I. Tokimitsu and N. Yagi, Diet therapy with diacylglycerol oil delays the progression of renal failure in type 2 diabetic patients with nephropathy, Diabetes Care, 2006, 29, 417–419.
22. K. Yamamoto, H. Asakawa, K. Tokunaga, H. Watanabe, N. Matsuo, I. Tokimitsu and N. Yagi, Long-Term Ingestion of Dietary Diacylglycerol Lowers Serum Triacylglycerol in Type II Diabetic Patients with Hypertriglyceridemia, J. Nutr., 2001, 131, 3204–3207.
23. K. Yamamoto, M. Takeshita, I. Tokimitsu, H. Watanabe, T. Mizuno, H. Asakawa, K. Tokunaga, T. Tatsumi, M. Okazaki and N. Yagi, Diacylglycerol oil ingestion in type 2 diabetic patients with hypertriglyceridemia, Nutrition, 2006, 22, 23–29.
24. H. Lu, T. Guo, Y. Fan, Z. Deng, T. Luo and H. Li, Effects of diacylglycerol and triacylglycerol from peanut oil and coconut oil on lipid metabolism in mice, J. Food Sci., 2020, 85, 1907–1914.
25. S. Meguro, K. Higashi, T. Hase, Y. Honda, A. Otsuka, I. Tokimitsu and H. Itakura, Solubilization of phytosterols in diacylglycerol versus triacylglycerol improves the serum cholesterol-lowering effect, Eur. J. Clin. Nutr., 2001, 55, 513–517.
26. H. Huang, J. Xie, L. Hou, M. Miao, L. Xu and C. Xu, Estimated small dense low–density lipoprotein cholesterol and nonalcoholic fatty liver disease in nonobese populations, J. Diabetes Invest., 2023, 15, 491–499.
27. T. Hirano, N. Satoh and Y. Ito, Specific Increase in Small Dense Low-Density Lipoprotein-Cholesterol Levels beyond Triglycerides in Patients with Diabetes: Implications for Cardiovascular Risk of MAFLD, J. Atheroscler. Thromb., 2024, 31, 36–47.
28. M. Balling, B. G. Nordestgaard, A. Langsted, A. Varbo, P. R. Kamstrup and S. Afzal, Small Dense Low-Density Lipoprotein Cholesterol Predicts Atherosclerotic Cardiovascular Disease in the Copenhagen General Population Study, J. Am. Coll. Cardiol., 2020, 75, 2873–2875.
29. X. Jin, S. Yang, J. Lu and M. Wu, Small, Dense Low-Density Lipoprotein-Cholesterol and Atherosclerosis: Relationship and Therapeutic Strategies, Front. Cardiovasc. Med., 2022, 8(804214).
30. E. Muzurovic, D. P. Mikhailidis and C. Mantzoros, Non-alcoholic fatty liver disease, insulin resistance, metabolic syndrome and their association with vascular risk, Metabolism, 2021, 119, 154770.
31. J. S. Zheng, L. Wang, M. Lin, H. Yang and D. Li, BMI status influences the response of insulin sensitivity to diacylglycerol oil in Chinese type 2 diabetic patients, Asia Pac. J. Clin. Nutr., 2015, 24, 65–72.
32. G. Reyes, K. Yasunaga, E. Rothenstein, W. Karmally, R. Ramakrishnan, S. Holleran and H. N. Ginsberg, Effects of a 1,3-diacylglycerol oil-enriched diet on postprandial lipemia in people with insulin resistance, J. Lipid Res., 2008, 49, 670–678.
33. M. Anikisetty, A. G. Gopala Krishna, V. Panneerselvam and A. N. Kamatham, Diacylglycerol (DAG) rich rice bran and sunflower oils modulate lipid profile and cardiovascular risk factors in Wistar rats, J. Funct. Foods, 2018, 40, 117–127.
34. Y.-M. Chen, C.-F. Lian, Q.-W. Sun, T.-T. Wang, Y.-Y. Liu, J. Ye, L.-L. Gao, Y.-F. Yang, S.-N. Liu, Z.-F. Shen and Y.-L. Liu, Ramulus Mori (Sangzhi) Alkaloids Alleviate High-Fat Diet-Induced Obesity and Nonalcoholic Fatty Liver Disease in Mice, Antioxidants, 2022, 11, 905.
35. X. Li, Z. Sun, W. Liu, L. Sun, J. Ren, Y. Xu, H. Yu and W. Bai, Methodology exploration and reproducibility evaluation of TAI and TSI for quantitative ultrasound assessment of hepatic steatosis, Heliyon, 2024, 10, e31904.
36. K. Feng, H. Fang, G. Liu, W. Dai, M. Song, J. Fu, L. Wen, Q. Kan, Y. Chen, Y. Li, Q. Huang and Y. Cao, Enzymatic Synthesis of Diacylglycerol-Enriched Oil by Two-Step Vacuum-Mediated Conversion of Fatty Acid Ethyl Ester and Fatty Acid From Soy Sauce By-Product Oil as Lipid-Lowering Functional Oil, Front. Nutr., 2022, 9, 884829.
37. D. L. Gorden, D. S. Myers, P. T. Ivanova, E. Fahy, M. R. Maurya, S. Gupta, J. Min, N. J. Spann, J. G. McDonald, S. L. Kelly, J. Duan, M. C. Sullards, T. J. Leiker, R. M. Barkley, O. Quehenberger, A. M. Armando, S. B. Milne, T. P. Mathews, M. D. Armstrong, C. Li, W. V. Melvin, R. H. Clements, M. K. Washington, A. M. Mendonsa, J. L. Witztum, Z. Guan, C. K. Glass, R. C. Murphy, E. A. Dennis, A. H. Merrill, D. W. Russell, S. Subramaniam and H. A. Brown, Biomarkers of NAFLD progression: a lipidomics approach to an epidemic, J. Lipid Res., 2015, 56, 722–736.
38. J. Xu, M. Shi, L. Chen, S. Chi, S. Zhang, J. Cao, B. Tan and S. Xie, Muscular lipidomics and transcriptomics reveal the effects of bile acids on lipid metabolism in high-fat diet-fed grouper, Fish Physiol. Biochem., 2023, 50, 127–143.
39. S. Blanc, I. Ottestad, S. Hassani, G. I. Borge, A. Kohler, G. Vogt, T. Hyötyläinen, M. Orešič, K. W. Brønner, K. B. Holven, S. M. Ulven and M. C. W. Myhrstad, Fish Oil Supplementation Alters the Plasma Lipidomic Profile and Increases Long-Chain PUFAs of Phospholipids and Triglycerides in Healthy Subjects, PLoS One, 2012, 7, e42550.
40. M. Orešič, T. Hyötyläinen, A. Kotronen, P. Gopalacharyulu, H. Nygren, J. Arola, S. Castillo, I. Mattila, A. Hakkarainen, R. J. H. Borra, M.-J. Honka, A. Verrijken, S. Francque, P. Iozzo, M. Leivonen, N. Jaser, A. Juuti, T. I. A. Sørensen, P. Nuutila, L. Van Gaal and H. Yki-Järvinen, Prediction of non-alcoholic fatty-liver disease and liver fat content by serum molecular lipids, Diabetologia, 2013, 56, 2266–2274.
41. J. Y. Xia, W. L. Holland, C. M. Kusminski, K. Sun, A. X. Sharma, M. J. Pearson, A. J. Sifuentes, J. G. McDonald, R. Gordillo and P. E. Scherer, Targeted Induction of Ceramide Degradation Leads to Improved Systemic Metabolism and Reduced Hepatic Steatosis, Cell Metab., 2015, 22, 266–278.
42. M. Murata, T. Ide and K. Hara, Reciprocal responses to dietary diacylglycerol of hepatic enzymes of fatty acid synthesis and oxidation in the rat, Br. J. Nutr., 1997, 77, 107–121.
43. M. Monserrat-Mesquida, M. M. Quetglas-Llabrés, C. Bouzas, O. Pastor, L. Ugarriza, I. Llompart, K. Cevallos-Ibarra, A. Sureda and J. A. Tur, Plasma Fatty Acid Composition, Oxidative and Inflammatory Status, and Adherence to the Mediterranean Diet of Patients with Non-Alcoholic Fatty Liver Disease, Antioxidants, 2023, 12, 1554.
44. V. Smid, K. Dvorak, P. Sedivy, V. Kosek, M. Lenicek, M. Dezortova, J. Hajslova, M. Hajek, L. Vitek, K. Bechynska and R. Bruha, Effect of Omega-3 Polyunsaturated Fatty Acids on Lipid Metabolism in Patients With Metabolic Syndrome and NAFLD, Hepatol. Commun., 2022, 6, 1336–1349.
45. S. A. Parry, F. Rosqvist, T. Cornfield, A. Barrett and L. Hodson, Oxidation of dietary linoleate occurs to a greater extent than dietary palmitate in vivo in humans, Clin. Nutr., 2021, 40, 1108–1114.
46. J. Liu, H. Wu, Y. Zhang, C. Hu, D. Zhen, P. Fu and Y. He, Phycobiliprotein Peptide Extracts from Arthrospira platensis Ameliorate Nonalcoholic Fatty Liver Disease by Modulating Hepatic Lipid Profile and Strengthening Fat Mobilization, Nutrients, 2023, 15, 4573.

---