Effect of tocotrienol-rich fraction (TRF) on lipid profile in hyperlipidemic experimental animal model: a systematic review and meta-analysis

Abstract

Hyperlipidemia, characterized by elevated lipid levels, is a major risk factor for cardiovascular diseases (CVD), including atherosclerosis. Tocotrienol-rich fraction (TRF), a potent form of vitamin E, has demonstrated promising lipid-lowering effects due to its antioxidant and anti-inflammatory properties. This systematic review and meta-analysis aimed to evaluate the effects of TRF on lipid profiles in hyperlipidemic experimental animal models and to comparatively assessed their suitability for preclinical research. A total of eight eligible studies published between 2002 and 2024 were included. TRF significantly reduced total cholesterol (effect size = − 4.675, p < 0.0001), LDL (− 4.847, p < 0.0001), and triglycerides (− 4.736, p < 0.0001), while significantly increasing HDL (4.001, p < 0.0001), particularly in rats and mice. Meanwhile, comparative analysis showed that New Zealand white rabbits, hamsters, and genetically modified mice exhibit lipid metabolism profiles closer to humans and respond to lower TRF doses over shorter durations. Conversely, rats, though less prone to diet-induced atherosclerosis, responded well to higher doses over longer periods. However, the predominance of rat studies, absence of standard drug comparators, and variability in dosing and duration limit translational interpretation. Future studies should utilize physiologically relevant models and adopt standardized experimental protocols to better assess TRF’s therapeutic potential in CVD management.

Introduction

Hyperlipidemia is defined by increased lipid levels in the body, which can result from various genetic or acquired conditions. In adults, hyperlipidemia is recognized as a significant risk factor for the development of cardiovascular disease (CVD). Several factors contribute to the growing concern of hyperlipidemia, including longer life expectancy and a decline in communicable diseases. Lifestyle-related factors such as smoking, high blood pressure, high cholesterol, and diabetes also play a significant role¹. Hyperlipidemia itself is often a silent condition, with no obvious symptoms. In most cases, it is only discovered during routine check-ups or when it reaches a critical stage, leading to adverse events like strokes or heart attacks. The main goal of hyperlipidemia treatment is to reduce the long-term risk of CVD caused by atherosclerosis². Typically, treatment begins with a combination of statin therapy and lifestyle changes, including a healthier diet and more physical activity³. Interestingly, research suggests that tocotrienols, a form of vitamin E, may also offer promising health benefits. Studies indicate they could help lower cholesterol, reduce inflammation, and support overall cardiovascular health⁴.

Tocotrienols, commonly referred to as T3 or toco-E, are one of the two main groups of Vitamin E compounds, alongside tocopherols (TP). Both belong to the family of lipid-soluble vitamins, but they differ in structure and function. Each group consists of four isomers, alpha (α), beta (β), gamma (γ), and delta (δ), classified based on the number and arrangement of methyl groups on their chromanol ring. While all Vitamin E compounds share a chromanol ring, their side chains set them apart. TP have a long, saturated side chain, whereas T3 have unsaturated double bonds, giving them distinct biological properties⁵.

Recent studies suggest that T3 may be more therapeutically potent than TP, offering benefits beyond their role as antioxidants. Research highlights their potential as antihypertensive, anti-cancer, anti-inflammatory, lipid-lowering, and neuroprotective agents, among other health benefits. These findings are supported by a substantial body of preclinical and clinical studies⁶.

In CVD, T3 exhibits strong antioxidant properties and help regulate blood lipid levels. They enhance the activity of key enzymes such as superoxide dismutase (SOD), NADPH: quinone oxidoreductase (NQO1), and glutathione peroxidase (GPx), which neutralize harmful free radicals. Additionally, they have been shown to lower cholesterol levels in both animal and human studies. This effect is primarily due to their ability to suppress 3-Hydroxy-3-Methylglutaryl-Coenzyme A (HMG-CoA) reductase, the enzyme responsible for cholesterol synthesis through a post-translational mechanism⁴.

Emerging study also points to the potential of γ- and δ-T3 in cancer prevention and treatment. This potential stems from their powerful anti-proliferative properties and direct pro-apoptotic effects on mitochondria in cancer cells. These compounds specifically inhibit pathways involving Nuclear Factor kappa-light-chain-enhancer of activated B cells (NF-κB) and B-cell lymphoma 2 (Bcl-2) while simultaneously stimulating the c-Jun N-terminal kinase /Bcl-2-associated X protein (JNK/BAX) pathways. This process leads to mitochondrial dysfunction and the release of cytochrome c into the cytosol, which can be effective against various solid and haematological tumours⁷.

Tocotrienol-rich fraction (TRF) refers to a concentrated extract of T3, commonly found in natural sources such as palm oil, rice bran oil, and oats. Recent systematic review and meta-analysis has investigated the impact of T3 supplementation on hyperlipidemia in adults. While human clinical trials showed that T3 significantly elevate high-density lipoprotein (HDL) cholesterol levels; however, the evidence for substantial reductions in low-density lipoprotein (LDL) cholesterol, total cholesterol (TC), or triglycerides (TG) remains inconclusive. Additionally, meta-analysis indicated that dosages of 200 mg or higher, taken for up to 8 weeks, were most effective in improving lipid profiles⁸.

In contrast, animal studies have consistently demonstrated the lipid-lowering effects of T3, showing significant reductions in TC and TG in hyperlipidemic models. Thus, our study aims to review the effects of TRF on lipid profile in hyperlipidemic experimental animal models focusing on the question: “Does tocotrienol-rich fraction (TRF) demonstrate superior hypolipidemic effects compared to control and untreated high fat diet group in hyperlipidemic animal models?”.

In addition, this review seeks to compare the suitability and characteristics of different animal models used in CVD and atherosclerotic research, offering insights into how model selection may influence the observed efficacy of TRF.

Materials and methods

This systematic review has been registered with PROSPERO International Prospective Register of Systematic Reviews (Identifier CRD420251003070).

Inclusion and exclusion criteria

Articles were included if the following criteria were fulfilled: (1) the study assessed the effects of TRF in animal models, with no restrictions on age, sex, or body weight of the animals; (2) the study exclusively utilized TRF, with no restrictions on the source, method of preparation, dosage, route of administration, or duration of treatment; (3) the article was an original full research article; (4) the study was published in English due to limitations in translation resources and (5) the study was published from January 2002 to December 2024.

Articles were excluded from the review if: (1) the article was a review article, case report, or letter to the editor; (2) the study involved clinical trials conducted on human subjects; (3) the study was presented only as an abstract; (4) the study assessed the effects of TRF on unrelated blood or clinical parameters, specifically studies that do not report on any lipid profile outcomes; (5) the study investigated the effects of TRF in combination with other plants, supplements, or exercise; and (6) duplicated studies.

Searching strategy

A systematic review of the literature was conducted to investigate relevant studies on the effects of TRF on lipid profiles in animal studies focusing on rodents’ and rabbits’ model. The research question was formulated based on the Population, Intervention, Comparison, and Outcome (PICO) strategy. The research question for this study was “Does tocotrienol-rich fraction (TRF) demonstrate superior hypolipidemic effects compared to control and untreated high fat diet group in hyperlipidemic animal models?”.

We conducted a comprehensive search of online databases, including Web of Science, PubMed, ScienceDirect, and Scopus. The search was limited to articles published from 2002 to 2024 and restricted to studies published in English. Only original research articles were considered. The following key search terms were used in the search strategy: (“tocotrienol*” OR “tocotrienol-rich fraction*”) AND (“lipid profile*” OR “cholesterol*” OR “hyperlipidemia*”) AND (“animal model*” OR “rat*” OR “rodent*” OR “mice*” OR “rabbit*” OR “hamster*”).

Study selection

In conducting the systematic review, we utilized EndNote 21 to streamline the study selection process. The references from the databases were exported in .RIS format and subsequently imported into EndNote 21. To ensure a manageable dataset, we created a dedicated group within EndNote 21 specifically for this systematic review, allowing for easy access to the imported references. For example, the groups were categorized according to the online databases name i.e. “Web of Science,” “PubMed,” “Science Direct,” and “Scopus.” Following the import, we utilized the “Find Duplicates” feature to identify and remove duplicate entries, ensuring that each reference was unique. We then conducted an initial screening of the references based on our predefined inclusion and exclusion criteria. Tags such as “KIV” and “Final” were created to facilitate a more organized review process.

The risk of bias in the included studies was assessed using the SYRCLE Risk of Bias tool for animal studies. Each study was evaluated across multiple domains, including sequence generation, allocation concealment, blinding, and incomplete outcome data utilizing Microsoft Excel and subjective judgement from reviewers.

Data extraction

Following the initial screening, we excluded studies that did not meet the inclusion criteria based solely on their titles, abstracts, and keywords. Subsequently, we performed an in-depth evaluation of the full-text articles to further exclude any studies that did not align with our predefined criteria. In instances of duplicate publications, only the most recent version was retained for inclusion in the review.

Data of interest were systematically extracted from the selected studies, and the following information was documented: (1) Sample Population; (2) Intervention; (3) Controls/Comparison; and (4) Outcomes on Lipid Profile. To assess the potential hypolipidemic activity of TRF in in vivo studies, we evaluated lipid profile parameters, including TG, TC, LDL, and HDL.

Studies that met the inclusion criteria were organized into a final group designated for included studies tagged as “Final.”

Review method

All potential articles were identified and screened by two independent reviewers. The study selection and data extraction processes were conducted following detailed discussions between the reviewers. Any disagreements or uncertainties regarding the inclusion of study was resolved through consensus.

Data analysis

The meta-analysis of continuous outcomes was conducted using Comprehensive Meta Analysis (CMA) software to compute the standard mean difference (SMD) for effect size evaluation. We refer to meta-analysis articles which utilize summarized data instead of raw data to calculate Cohen’s d, providing valuable insights into the therapeutic effects of curcumin on glycaemic and lipid profiles in type 2 diabetes, as well as the safety and efficacy of proprotein convertase subtilisin-kexin type 9 inhibitors following acute coronary syndrome^9,10. The parameters included TC, LDL, HDL and TG levels. Following the calculation, forest plot was generated to visually present the findings along with funnel plot to assess publication bias.

Additionally, to support the meta-analysis result, key model characteristics used in CVD or atherosclerotic research were extracted, tabulated, and reviewed for interspecies comparisons based on previous studies with a focus on identifying how species and model selection may influence the interpretation of TRF’s therapeutic effects.

Results

Search results or study selection

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart, which details the study selection process, is provided in Fig. 1.

Fig. 1.

The preferred reporting items for systematic reviews and meta-analyses (PRISMA) flowchart.

Meanwhile, a complete breakdown of the risk of bias assessment is provided in Table 1 which revealed that most studies had a high risk of bias in random outcome assessment, blinding of caregivers and assessors. In contrast, sequence generation, baseline characteristics and allocation concealment were often rated as unclear while random housing was inconsistently reported across studies. Conversely, incomplete outcome data and selective outcome reporting were generally at low risk across studies, implying that data handling and reporting were relatively robust. Although the SYRCLE risk of bias assessment was not used as an exclusion criterion, the domain-specific risk profiles of individual studies were carefully considered during data synthesis. Specifically, findings from studies with high risk in key domains were interpreted with appropriate caution, particularly in terms of their influence on overall conclusions of the meta-analysis. No subgroup or sensitivity analyses were conducted based on SYRCLE risk of bias scores, due to the limited number of included studies and the overall similarity in risk of bias profiles across the dataset. In addition, no statistical weighting adjustment was applied based on risk of bias. This approach allowed for a comprehensive inclusion of available data while acknowledging potential limitations in methodological quality.

Table 1.

SYRCLE Risk of bias tool assessment.

Study characteristics

Among the eight included studies, three studies used male Wistar rats, two studies examined male Sprague Dawley rats, one study each focused on strain B6; 129S7-Ldlrtm1Her/J mice, New Zealand white (NZW) rabbits and Golden Syrian hamsters. The studies included in this systematic review utilized a variety of gender, sex, and weight characteristics in their sample populations. The summary of the characteristics of the studies is depicted in Table 2.

Table 2.

Summary of the effect of TRF on lipid profile in hyperlipidemic experimental animal model: species, strain, age, dosage, route of administration, and treatment duration.

Study	Species/strain	N	Sex	Age (weeks)/Weight (g)	TRF type and composition	Dose (mg/kg/day)	Route of TRF administration	Duration (weeks)	Control	Outcomes	References
1	Sprague Dawley rats	23	Male	NR/260–290	Palm TRFa	200	Oral	8	NDM, Diabetic untreated	TC↓, LDL↓, HDL↑, TG↓	11
2		21	Male	3 /45–70	Tocovid SupraBio™b	60	Oral	4	Control diet, HFD	TC↓19%, non-HDL↓29%, TG↑	12
3	Wistar rats	30	Male	NR/175–200	Palm TRFc & Rice bran TRFd	200 (PO-TRF), 400 (RBO-TRF)	Oral	16	HFD/STZ, Normal saline	TC↓, LDL↓, VLDL↓, HDL↑, TG↓	13
4		30	Male	8–12 / 100–120	Palm TRFc	100	Oral	6	Normal, Disease control	TC↓32%, LDL↓39%, VLDL↓48%, HDL↑73%, TG↓64%	14
5		30	Male	12/350–400	Palm TRFe	60, 100	Oral	12	Baseline, Normal, HCHF	TC↓, LDL↓, HDL↑ (100 mg), TG↓	15
6	NZW rabbits	30	NR	NR/2000–2500	Annatto T3f.	4, 15	Oral	8	T3-negative	LDL↓, TC↓ (4 mg), No HDL/TG change	16
7	Golden Syrian hamsters	70	Male	6 / 90–110	Mixed T3g and purified γ-T3h	23, 39, 58, 263	Oral	4	SIM, No treatment	TC↓7–23%, LDL↓6–37%, HDL ↔ , TG ↔	17
8	LDLr -/- mice (B6;129S7)	63	NR	4 / 20–22	γT3, δT3, γδT3 ± αTP/STi	50	Oral	4	DMSO, SIM	TC↓25%, LDL↓51%, TG↓19%, HDL ↔	18

TC, Total cholesterol; LDL, Low-density lipoprotein; VLDL, Very low-density lipoprotein; HDL, High-density lipoprotein; TG, Triglycerides; NDM, Normal non-diabetic; NZW, New Zealand white; NR, Not reported; BW, Body weight; HFD, High-fat diet; STZ, Streptocotozin; HCHF, High cholesterol, high fat; SIM, Simvastatin; DMSO, Dimethyl sulfoxide.

^aComposition not fully specified.

^bTocovid SupraBio™: 23.5% α-, 43.2% γ-, 9.8% δ-tocotrienols (T3), 23.5% α-tocopherol (TP).

^cPalm Oil (PO-TRF): 23 mg% α-TP, 29 mg% α-, 35 mg% γ-, 13 mg% δ-T3.

^dRice Bran Oil (RBO-TRF): 14.5 mg% α-, 2.3 mg% β-, 6.2 mg% γ-, 6.2 mg% δ-T3, 39.2% tocopherols, 31.6 mg% unidentified matters.

^ePalm TRF: 21.9% α-TP, 24.7% α-, 4.5% β-, 36.9% γ-, 12.0% δ-T3.

^fAnnatto T3: 90% δ-, 10% γ-T3.

^gMixed-T3: 29.5% α-, 3.3% ß, 41.4% γ, 0.1% δ-T3, and 25.1% others (mainly 9.9% γ-TP).

^hγ-T3: 0.6% α-, 7.0% ß-, 86.1% γ-, 0.1% δ-T3, and 6.2% others (mainly 1.2% γ-TP).

ⁱVarious combinations: γδT3 alone or with α-tocopherol (αTP) or Simvastatin (ST).

The sources of TRF used across the studies included Palm Oil (PO), Rice Bran Oil (RBO), and Annatto Oil, with variations in the specific formulations and compositions. Meanwhile, the range of TRF dosages administered across the studies was from 4 to 400 mg/kg body weight with duration of treatment ranging from 4 to 16 weeks.

Review outcomes

Data analysis outcomes

The standardized mean difference (SMD) was calculated based on lipid profile outcomes depicted in Table 3, along with 95% confidence intervals (CI) to evaluate the effectiveness of tocotrienol across various species. The degree of heterogeneity among the studies was assessed using the Q statistic with a corresponding p-value indicating whether the observed differences could be attributed to random effects. Additionally, publication bias was assessed using Egger’s test, and in the event of funnel plot asymmetry, Duval and Tweedie’s trim and fill method was performed to adjust for potential bias by estimating and imputing missing studies.

Table 3.

Summarized data of lipid profile outcomes in mmol/L.

Study	Group	Lipid profiles (mmol/L)
		TC	LDL	HDL	TG
1	NDM Group	1.18 ± 0.06	0.49 ± 0.03	0.54 ± 0.04	0.42 ± 0.01
	Non-TRF Group	1.86 ± 0.09*	1.01 ± 0.23*	0.11 ± 0.01*	1.69 ± 0.13*
	TRF Group	1.38 ± 0.08#	0.38 ± 0.03*#	0.46 ± 0.03*#	1.19 ± 0.05*#
2	Control diet	1.76 ± 0.07	NR	1.20 ± 0.05	0.94 ± 0.23
	HFD	2.16 ± 0.08*	NR	1.39 ± 0.06	2.22 ± 0.18*
	TRF	1.76 ± 0.12#	NR	1.22 ± 0.08	2.34 ± 0.28*#
3	Control	2.81 ± 0.20	1.19 ± 0.29	0.98 ± 0.13	1.18 ± 0.22
	HFD/STZ	5.46 ± 0.51*	3.52 ± 0.28*	0.65 ± 0.09*	2.31 ± 0.13*
	R-400	3.94 ± 0.52*#	2.30 ± 0.26*#	1.08 ± 0.21#	1.42 ± 0.31#
	P-200	2.94 ± 0.26#	1.32 ± 0.43#	1.11 ± 0.17#	1.13 ± 0.39#
4	Normal	1.18 ± 0.21	0.36 ± 0.06	0.08 ± 0.02	0.85 ± 0.13
	Atherogenic diet	3.68 ± 0.38*	2.89 ± 0.24*	0.05 ± 0.01*	2.28 ± 0.17*
	Untreated atherogenic rats	3.58 ± 0.32	2.73 ± 0.22	0.05 ± 0.01	2.15 ± 0.14
	TRF-treated atherogenic rats	1.76 ± 0.17#	1.29 ± 0.20#	0.19 ± 0.03#	0.81 ± 0.11#
5	Baseline	2.13 ± 0.05	0.56 ± 0.06	1.30 ± 0.06	0.33 ± 0.04
	Normal	2.35 ± 0.13	0.69 ± 0.04	1.21 ± 0.06	0.35 ± 0.03
	HCHF	3.67 ± 0.19*	1.57 ± 0.20*	0.54 ± 0.18*	2.44 ± 0.15*
	HCHF + 60 pT3	2.64 ± 0.33#	0.89 ± 0.08#	1.11 ± 0.23	1.27 ± 0.24*#
	HCHF + 100 pT3	2.53 ± 0.20#	0.85 ± 0.15#	1.35 ± 0.07#	1.05 ± 0.16#
6	B1: T3-negative	0.60 ± 0.10	0.10 ± 0.00	0.30 ± 0.10	0.60 ± 0.00
	B1: T3-4 mg/kg	0.8 ± 0.00	0.20 ± 0.00	0.60 ± 0.00	0.70 ± 0.10
	B1: T3-15 mg/kg	0.68 ± 0.10	0.20 ± 0.00	0.40 ± 0.10	0.60 ± 0.10
	B2: T3-negative	7.10 ± 0.70	7.40 ± 0.30	1.00 ± 0.00	0.70 ± 0.20
	B2: T3-4 mg/kg	12.50 ± 2.90	13.30 ± 4.10	2.00 ± 0.40	0.70 ± 0.10
	B2: T3-15 mg/kg	7.80 ± 0.80	7.40 ± 0.80	0.90 ± 0.10	0.60 ± 0.10
7	Control	5.48 ± 0.62	1.58 ± 0.26	3.43 ± 0.47	3.29 ± 0.96
	Mixed T3 (39 mg/kg)	5.11 ± 0.74	1.35 ± 0.29	3.35 ± 0.69	3.00 ± 0.41
	Mixed T3 (263 mg/kg)	4.93 ± 0.39#	1.22 ± 0.23#	3.34 ± 0.31	3.04 ± 0.86
	γ-T3 (23 mg/kg)	5.10 ± 0.68	1.38 ± 0.19	3.34 ± 0.58	3.01 ± 0.38
	γ-T3 (58 mg/kg)	4.64 ± 0.60#	1.10 ± 0.23#	3.20 ± 0.63	2.90 ± 0.50
	γ-T3 (263 mg/kg)	4.72 ± 0.72#	1.06 ± 0.35#	3.22 ± 0.47	3.77 ± 0.53
8	Control	7.50 ± 0.50	2.70 ± 0.20	4.50 ± 0.40	1.80 ± 0.20
	Simvastatin	7.60 ± 0.60	2.80 ± 0.20	4.70 ± 0.40	1.70 ± 0.20
	γδT3	5.40 ± 0.40#	1.30 ± 0.10#	4.60 ± 0.30	1.50 ± 0.10#
	Simvastatin + γδT3	5.60 ± 0.30#	1.40 ± 0.10#	4.00 ± 0.30	1.40 ± 0.10#
	γδT3 + αTP	5.20 ± 0.30#	1.40 ± 0.10#	4.00 ± 0.30	1.40 ± 0.10#

Data are represented as mean ± SD.

TC, Total cholesterol; LDL, Low-density lipoprotein; HDL, High-density lipoprotein; TG, Triglycerides; NDM, normal non-diabetic group; HFD, high fat diet; NR, not reported; STZ, streptozotocin; R-400, rice bran oil TRF 400 mg/kg; P-200, palm oil TRF 200 mg/kg; HCHF, high cholesterol high fat diet; pT3, palm tocotrienol; T3, tocotrienol.

*p < 0.05 compared with normal diet control group.

^#p < 0.05 compared with untreated HFD control group.

A random-effects model was utilized due to expected differences in experimental condition and animal models. The findings are reported as pooled effect sizes for TC, LDL, HDL, and TG, with subgroup analyses performed for comparisons specific to different species.

Effects of TRF on total cholesterol

Figure 2 presents the meta-analysis on the efficacy of TRF in reducing TC levels across different species which yielded varying effect sizes. In rats (n = 5), the pooled SMD was − 4.675 (95% CI − 5.477 to − 3.872, p < 0.0001), indicating a significant decrease in TC levels with TRF treatment.

Fig. 2.

Meta-analysis of the efficacy of TRF on TC levels in animal models: forest plot and summary results. Test for overall effect: SMD = -2.925, p = 0.008; Test for heterogeneity: Q = 80.430, df = 3, p < 0.0001.

For rabbit (n = 1) study, the effect size was − 1.295 (95% CI − 2.233 to − 0.356, p = 0.007) suggesting a reducing effect. In hamster (n = 1) study, TRF significantly lowered TC levels with an SMD of − 0.927 (95% CI − 1.482 to − 0.372, p = 0.001). Similarly, the mice (n = 1) study demonstrated a strong cholesterol-lowering effect with an SMD of − 4.893 (95% CI − 6.062 to − 3.723, p < 0.0001).

Overall, TRF reduced TC by − 2.925 (p = 0.008) while the Q-test for heterogeneity (Q = 80.430, df = 3, p < 0.0001) suggests substantial variability between studies and p-value indicates that the heterogeneity is statistically significant.

Effects of TRF on LDL

Figure 3 summarizes the overall effect of TRF on LDL levels compared to the control group. In rats (n = 4), four out of five studies with one study reported on VLDL level instead of LDL level, reveal a highly significant reduction in LDL (SMD = − 4.847, p < 0.0001) and CI for this effect ranges from − 5.693 to − 4.002. In hamsters (n = 1) and mice (n = 1), TRF led to significant reductions in LDL, with effect sizes of − 1.327 (p < 0.0001) and − 8.415 (p < 0.0001), respectively. In contrast, rabbit study (n = 1) found no significant effect (p = 0.120).

Fig. 3.

Meta-analysis of the efficacy of TRF on LDL levels in animal models: forest plot and summary results. Test for overall effect: SMD = -3.775, p = 0.006; Test for heterogeneity: Q = 104.097, df = 3, p < 0.0001.

To conclude, the findings from this meta-analysis indicate that TRF significantly reduces LDL cholesterol levels, as evidenced by a large negative effect size (SMD = − 3.775, p = 0.006). However, the substantial heterogeneity among studies (Q = 104.097, p < 0.0001) suggests that the effects of TRF on LDL vary considerably depending on study characteristics.

Effects of TRF on HDL

Figure 4 depicts the meta-analysis of the efficacy of TRF on HDL levels compared to the control group. In rats (n = 5), TRF showed a significant elevation in HDL levels, with an effect size of 4.001 (p < 0.0001) and CI ranged from 3.065 to 4.938.

Fig. 4.

Meta-analysis of the efficacy of TRF on HDL levels in animal models: forest plot and summary results. Test for overall effect: SMD = 1.605, p = 0.115; Test for heterogeneity: Q = 35.872, df = 3, p < 0.0001.

Meanwhile, the studies on rabbits (n = 1), hamsters (n = 1), and mice (n = 1) do not show significant effects with SMD of 1.000 (p = 0.357), 0.278 (p = 0.519), and 1.023 (p = 0.074) respectively.

To conclude, while the positive SMD (1.605) suggests a potential increase in HDL cholesterol, the lack of statistical significance (p = 0.115) means this effect is not conclusive. Additionally, the high heterogeneity (Q = 35.872, p < 0.0001) suggests that the effect varies across studies.

Effects of TRF on triglycerides

Figure 5 illustrates the overall effect of TRF on TG levels compared to the control group. In rats (n = 5), TRF led to a significant decrease in TG, with an effect size of − 4.736 (p < 0.0001) and CI ranged from − 5.944 to − 3.528. In mice (n = 1), TRF caused a significant decrease in TG, with an effect size of − 2.313 (p = 0.004) and CI ranged from − 3.872 to − 0.754.

Fig. 5.

Meta-analysis of the efficacy of TRF on TG levels in animal models: forest plot and summary results. test for overall effect: SMD =  − 1.985, p = 0.090; Test for heterogeneity: Q = 29.622, df = 3, p < 0.0001.

In contrast, the results in rabbits (n = 1) and hamsters (n = 1) were not significant. In the rabbit study, the effect size was − 0.315 (p = 0.749), and the CI ranged from − 2.243 to 1.613, indicating no effect on TG. Similarly, the hamster study showed a minimal effect, with SMD of − 0.436 (p  = 0.459).

Overall, although TRF shows a moderate reduction in TG (SMD = − 1.985), the effect is not statistically significant (p = 0.090). Additionally, the substantial heterogeneity (Q = 29.622, p < 0.0001) indicates variability in the effects across studies.

Publication bias

Figures 6, 7, 8 and 9 illustrate the funnel plots used to assess publication bias in the meta-analysis of TRF’s effects on TC, LDL, HDL, and TG levels, respectively. The Duval and Tweedie’s trim and fill analysis for TC, LDL, HDL, and TG revealed varying degrees of publication bias and heterogeneity. For TC and LDL, five studies were trimmed, indicating potential publication bias, but the adjusted values still supported a significant reduction. HDL showed no trimmed studies, suggesting no publication bias, and the effect estimates remained unchanged. TG had three studies trimmed, indicating some bias, and the adjusted effect size became slightly less negative, suggesting an overestimated initial effect. Despite some publication bias, the findings remain robust, supporting the efficacy of TRF in improving lipid profiles.

Fig. 6.

Funnel plot for assessing publication bias in the meta-analysis of TRF’s effect on TC levels. The publication bias is adjusted by imputing the missing studies based on the asymmetry of the funnel plot. (●) Plot Imputed and (○) Plot observed studies. Egger’s linear regression test (Intercept -6.5891, t = 11.7214, p < 0.0001).

Fig. 7.

Funnel plot for assessing publication bias in the meta-analysis of TRF’s effect on LDL levels. The publication bias is adjusted by imputing the missing studies based on the asymmetry of the funnel plot. (●) Plot Imputed and (○) Plot observed studies. Egger’s linear regression test (Intercept − 6.4991, t = 10.3743, p < 0.0001).

Fig. 8.

Funnel plot for assessing publication bias in the meta-analysis of TRF’s effect on HDL levels. (○) Plot observed studies. Egger’s linear regression test (Intercept 6.6351, t = 14.6710, p < 0.0001).

Fig. 9.

Funnel plot for assessing publication bias in the meta-analysis of TRF’s effect on TG levels. The publication bias is adjusted by imputing the missing studies based on the asymmetry of the funnel plot. (●) Plot Imputed and (○) Plot observed studies. Egger’s linear regression test (Intercept − 6.3269, t = 9.3093, p < 0.0001).

Comparison among atherosclerotic animal models

Table 4 displays the susceptibility and relevance of various animal models commonly used in TRF studies related to CVD, particularly atherosclerosis, summarized based on previous studies. Sprague Dawley and Wistar rats are relatively resistant to atherosclerosis and typically require HCD or genetic modifications for disease induction, limiting their relevance to human pathology, however, they remain valuable for specific CVD studies. Conversely, NZW rabbits are highly susceptible to diet-induced atherosclerosis and develop plaques that closely mimic human atherosclerotic lesions, including lipid accumulation, inflammation, and potential plaque rupture. This makes rabbits a robust and translational model for studying the pathogenesis and treatment of atherosclerosis. This is also supported by previous findings showing that after 8 weeks of a high-cholesterol diet (HCD), rabbits exhibited increased body weight, hypercholesterolemia, atherosclerotic plaque development, and elevated expression of proteins linked to endothelial activation and pro-inflammatory biomarkers.

Table 4.

Susceptibility and relevance of animal models in TRF studies.

Model	Susceptibility	Induction methods	Relevance to human disease	References
SD Rat	Resistant	HCD, genetic modifications	Limited, useful for specific CVD studies	19–21
Wistar Rat	Resistant	HCD, genetic modifications	Limited, useful for specific CVD studies	19–21
NZW Rabbit	Highly susceptible	HCD, mechanical injury	High, closely resembles human atherosclerosis	22–24
Mice (genetically modified)	Moderately susceptible	Genetic modifications, HCD	Moderate, useful for genetic and mechanistic studies	19,25
Hamster	Highly susceptible	HCD, genetic modifications	High, similar lipoprotein metabolism to humans	20

SD, Sprague Dawley; NZW, New Zealand white; HCD, High cholesterol diet; CVD, Cardiovascular disease.

Mice, particularly genetically modified strains like ApoE-/- and Ldlr-/-, are widely used due to their ease of handling and genetic tractability. Although mouse lipoprotein metabolism and plaque features differ from humans, they are ideal for mechanistic and genetic studies. Hamsters, on the other hand, offer a more comparable lipoprotein metabolism to humans among rodent models. When fed high-cholesterol diets, especially in genetically altered forms, hamsters develop atherosclerotic lesions that closely resemble human plaques in lipid composition and distribution.

Discussion

The results of the meta-analysis across four species namely rats, rabbits, hamsters, and mice revealed varying effects of TRF on lipid profiles which may be attributed to interspecies differences or the influence of TRF dosage and treatment duration. While a few included studies evaluated TRF in comparison with standard pharmacological agents such as statins, their limited number and methodological heterogeneity prevented a valid pooled or structured comparative analysis. Therefore, this meta-analysis focused on synthesizing the standalone lipid-lowering effects of TRF in hyperlipidemic animal models.

Based on the result obtained, TRF significantly reduced TC, LDL, and TG levels in rats and mice, with large effect sizes in rats (TC: − 4.675, LDL: − 4.847, TG: − 4.736; p < 0.0001) and strong significance in mice (p < 0.0001), while no significant effects were observed in rabbits and hamsters. Meanwhile for HDL, TRF significantly increased levels in rats (effect size: 4.001; p < 0.0001), with inconclusive results in mice and no significant changes in rabbits and hamsters.

In the context of species-specific reasoning, mice were found to have the highest levels of polyunsaturated fatty acid (PUFA)-containing cholesteryl esters (ChEs) and significant differences in phosphatidylcholines (PCs) and ether phospholipids (ePLs). This distinctive lipid composition may affect the way these animals respond to TRF treatment. Since PUFA-containing ChEs play a key role in cellular signalling and membrane structure, the way these lipids interact with TRF could enhance the compound’s lipid-lowering effects in mice²⁶.

In contrast, rats displayed lower levels of ether-type phospholipids (PLs) and phosphatidylethanolamines (PEs) compared to humans, along with distinct differences in ChEs. Since ether-type PLs are essential for maintaining membrane integrity and function, their lower abundance in rats may influence how these animals metabolize TRF. The variations in ChEs could also play a role in how TRF affects lipid metabolism in rats, potentially explaining the strong response seen in this species²⁶.

Rabbits, on the other hand, had significantly elevated levels of sulfatides, a type of sulfonated glycosphingolipids compared to both mice and rats which suggests a different mode of lipid metabolism in rabbits. Sulfatides are involved in cell signalling and formation of myelin sheath, and their elevated levels in rabbits may indicate a unique lipid processing pathway that could influence the absorption, distribution, and action of TRF²⁶.

However, the varying effects of TRF across studies may also be influenced by differences in animal models, correlating the dosage and treatment duration used. Among the models in this present study, Sprague Dawley and Wistar rats were most commonly used, despite their natural resistance to diet-induced atherosclerosis. Interestingly, significant lipid-lowering effects were still observed. These effects may be attributed to the administration of higher doses over extended treatment durations.

For instance, in Sprague Dawley rats, Study 1 used a relatively high dose of 200 mg/kg/day over 8 weeks, which produced notable improvements, whereas Study 2 used a lower dose of 60 mg/kg for only 4 weeks, potentially explaining the more modest effect. Similarly, Wistar rat studies that utilized doses of 200–400 mg/kg over 12 to 16 weeks in Study 3 and Study 5 tended to show stronger therapeutic responses compared to shorter, lower-dose studies such as Study 4 with 100 mg/kg for 6 weeks. These findings suggest that both higher dosages and extended treatment durations enhance tocotrienol’s lipid-modulating effects, especially in rodent models with lower baseline susceptibility^11–15.

In contrast, NZW rabbits, which are highly susceptible to atherosclerosis, responded to much lower doses (4–15 mg/kg) over 8 weeks¹⁶. Despite the lower dosing, significant anti-atherogenic effects were reported, reflecting the species’ sensitivity to both dietary cholesterol and therapeutic interventions. Considering the original dosage of 15 mg/kg administered to rabbits in Study 6, the equivalent dose for rats was calculated using standard interspecies dose conversion based on body surface area normalization, applying species-specific Km factors. According to this method, the conversion factor from rabbit (Km = 12) to rat (Km = 6) results in a two-fold increase, yielding an equivalent rat dosage of 30 mg/kg²⁷. Despite the adjustment, this translated dose remains the lowest among the studies evaluated (Study 1–5), where TRF doses ranged significantly higher. This suggests that the rabbit study employed a relatively conservative dosage, which may partially explain differences in treatment efficacy outcomes observed in the meta-analysis. Future studies employing NZW rabbits as atherosclerotic models should consider administering TRF at doses exceeding 15 mg/kg to potentially elicit more pronounced lipid-lowering effects, provided that such doses are supported by toxicity and safety evaluations.

Similarly, Golden Syrian hamsters, responded positively to a broad dose range (23–263 mg/kg) over just 4 weeks, with higher doses showing stronger lipid-lowering effects¹⁷. This enhanced response is attributed to their human-like lipoprotein metabolism, making them a valuable model for studying lipid-related interventions. In genetically modified mice (LDLr-/-), Study 8 used a 50 mg/kg/day dose for 4 weeks and observed improvements in cholesterol parameters, despite the shorter duration and lower dose relative to rat studies¹⁸. This further emphasizes that genetic susceptibility significantly influences the therapeutic threshold, allowing for efficacy at lower dosages or shorter treatment periods.

Overall, the relationship between dose, duration, and species susceptibility is evident in the meta-analysis results. Rodent models such as rats may require higher and prolonged dosing due to lower baseline sensitivity, while more susceptible models like rabbits, hamsters, and genetically altered mice achieve comparable or superior effects with lower doses and shorter durations.

Limitations of study

Although this meta-analysis incorporates various animal models including Sprague Dawley, Wistar rats, NZW rabbits, mice, and hamsters, the predominance of rat studies limits robust interspecies comparisons. While the lipid-lowering effects of TRF were consistently observed, the limited number of studies per species and insufficient statistical power hinder definitive subgroup analyses. Additionally, heterogeneity in animal models, TRF composition, dosing regimens, and treatment durations contributes to variability that may affect the pooled outcomes. Moreover, the influence of methodological quality on the pooled results could not be formally explored due to the small number and similar risk profiles of included studies.

A notable limitation across the majority of studies included in this meta-analysis is the absence of standard drug comparators. Specifically, six out of the eight studies (Studies 1–6) compared tocotrienol treatment solely to untreated control or disease model groups, without benchmarking against established lipid-lowering agents such as statins. Only two studies (Study 7 and 8) included a standard comparator, simvastatin, enabling a direct assessment of tocotrienol’s efficacy relative to a clinically validated therapeutic agent.

The lack of standard comparators poses significant challenges in interpreting the clinical relevance of tocotrienol. While reductions in lipid parameters may be statistically significant compared to untreated controls, it remains unclear whether tocotrienol performs comparably, better, or worse than standard pharmacological options. This omission limits translational insight and may hinder the potential positioning of tocotrienol as a complementary or alternative treatment for hyperlipidemia and atherosclerosis.

In term of publication bias, although funnel plots and Duval and Tweedie’s trim and fill analyses were conducted, the results suggest a degree of bias may still be present. While the adjusted effect sizes largely remained consistent with the original findings, the presence of bias, particularly for TC and LDL, indicates that the true effect of TRF might be slightly overestimated.

In addition, several limitations affect the translational relevance to humans. Firstly, physiological and metabolic differences can influence treatment response²⁸. Secondly, variations in dosage, study duration, and disease induction methods among animal models complicate cross-species comparisons²⁹. Moreover, the controlled conditions of animal studies do not fully capture the complexity and heterogeneity of human populations³⁰. These factors highlight the challenges in extrapolating preclinical findings to clinical outcomes and underscore the need for well-designed human trials to validate TRF’s therapeutic potential.

Lastly, this meta-analysis also faces broader limitations. The literature search was confined to studies published in English between 2002 and 2024, which may have excluded relevant findings from earlier publications or non-English sources.

Future direction

Tocotrienols are well known for their antioxidant and anti-inflammatory properties, which may play a role in their ability to lower lipid levels. However, the exact molecular mechanisms behind these effects are still not fully understood. Future research should aim to uncover these pathways while also exploring the clinical potential of TRF in different formulations as well as including standard drug comparators in experimental designs. Recent study indicates that improving the permeability of γ-tocotrienol could enhance its bioavailability through advanced formulation techniques or structural modifications to γ-tocotrienol itself. The potential of nanoformulations to enhance TRF’s bioavailability and therapeutic efficacy presents an exciting avenue for future exploration, whereas delivery systems in the form of nanovesicles or nanoparticles strongly modified the absorption and congener selectivity of tocotrienols after oral administration. Additionally, recent expansive development of diverse nanoemulsion (NE) vehicles also emphasized their vast potential to improve the effective dosing of different clinical and experimental drugs of lipophilic nature including TRF^31–33.

Conclusions

This meta-analysis demonstrates that TRF exerts significant lipid-lowering effects, particularly reductions in total cholesterol, LDL, and triglycerides, and increased HDL, in hyperlipidemic animal models, most notably in rats and mice. However, interspecies variations in lipid metabolism, susceptibility to atherosclerosis, and physiological differences significantly influence treatment outcomes. While rodent models required higher doses and longer treatment durations to elicit effects, more susceptible models such as rabbits, hamsters, and genetically modified mice responded to lower doses, underscoring the importance of species selection in preclinical research. Future research should prioritize the inclusion of underrepresented but physiologically relevant species, standardized dosing regimens, and direct comparisons with established lipid-lowering agents.

Author contributions

Abdah H. W. contributed to the conceptualization, methodology, investigation, data curation, and writing of the original draft. Hanafi N. I. was responsible for formal analysis, validation, visualization, and writing during the review and editing process. Abdul Muid S. provided resources, managed the project, and supervised the research. Ibrahim N. contributed to statistical analysis, software use, data curation, and manuscript review and editing. Mohd Kasim N. A. was involved in conceptualization, supervision, funding acquisition, review and editing of the manuscript, and served as the corresponding author. All authors reviewed and approved the final version of the manuscript.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request. Due to ethical and confidentiality considerations, certain restrictions may apply to the availability of the data.

Declarations

Competing interests

The authors declare no competing interests.