Psyllium supplementation and lipid profiles: systematic review and dose-response meta-analysis of randomized controlled trials

Abstract

Background

Some studies indicate that psyllium supplementation may change lipid profile levels. This study assessed the impact of psyllium consumption on lipid profile (Low-Density Lipoprotein Cholesterol, Triglyceride, High-Density Lipoprotein Cholesterol, and cholesterol).

Main text

We started searching articles using Scopus, Institute for Scientific Information Web of Science, and PubMed to identify eligible publications from March 15, 2022 to August 2, 2025. The effect of psyllium on lipid profiles in adults was evaluated through Randomized Controlled Trials. We calculated Weighted Mean Differences with 95% Confidence Intervals using a random effects model. In this study, 41 Randomized Controlled Trials articles and 2049 participants were included. Psyllium showed a significant decrease in Low-Density Lipoprotein Cholesterol and total Cholesterol, a nonsignificant reduction in Triglyceride, and an insignificantly increase in High-Density Lipoprotein Cholesterol; cholesterol: (Weighted Mean Differences: -9.05; 95% Confidence Intervals: -13.71, -4.40; p-value < 0.05), High-Density Lipoprotein Cholesterol: (Weighted Mean Differences: 0.57; 95% Confidence Intervals: -0.88, 2.04; p-value > 0.05), Triglyceride: (Weighted Mean Differences: -5.29; 95% Confidence Intervals: -12.14, 1.54; p-value > 0.05), and Low-Density Lipoprotein Cholesterol: (Weighted Mean Differences: -8.55; 95% Confidence Intervals: -12.92, -4.19; p < 0.001). Considerable heterogeneity was found for cholesterol: (I-squared index = 89.30%, p-value < 0.001), High-Density Lipoprotein Cholesterol: ( I-squared index = 77.96%, p-value < 0.001), Low-Density Lipoprotein Cholesterol: (I-squared index = 88.46%, p-value < 0.001); and Triglyceride: (I-squared index = 83.25%, p-value < 0.001) and duration and dosage of psyllium had a nonsignificant linear influence on lipid profiles.

Conclusion

Results showed that psyllium can significantly decrease Low-Density Lipoprotein Cholesterol and total cholesterol, can insignificantly decrease triglyceride, and can insignificantly increase High-Density Lipoprotein Cholesterol following psyllium consumption.

Registration

The study was approved by the Medical Ethics Committee of Isfahan University of Medical Sciences (IR.MUI.RESEARCH.REC.1402.017, Grant number: 140206). This research was registered in the PROSPERO system (CRD42023402987).

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12263-025-00786-5.

Introduction

Traditionally, lipid profiles were determined by analyzing blood samples after fasting for 8 to 12 h. However, the practice of fasting lipid measurements is gradually being replaced by random and non-fasting measurements in many regions. Several societies, guidelines, and statements now support the use of non-fasting measurements [1]. Hypertension and hyperlipidemia are key contributors and potential risk factors for cardiovascular disease [2]. A positive correlation between total cholesterol and cardiovascular mortality has been demonstrated, and there is also an association between hyperlipidemia and the development of heart failure [3]. Hyperlipidemia is characterized by abnormally high levels of lipoproteins or lipids due to metabolic dysfunction. This condition can adversely affect the structure and function of the heart and blood vessels, leading to cardiovascular diseases [4]. Elevated concentrations of total cholesterol, LDL-C, and TG, along with low HDL-C levels, are associated with increased CHD risk in women [5, 6]. Dietary fibers can serve as supplements with potential lipid-lowering effects [7]. They significantly affect the metabolism and absorption of fats and carbohydrates. A higher intake of dietary fiber can reduce total and LDL-C levels, resulting in a more favorable lipid profile [8]. Studies have shown that dietary fiber intake is negatively correlated with plasma total cholesterol and LDL-C, while it is positively correlated with HDL cholesterol. Thus, nutrition education aimed at increasing dietary fiber intake from sources such as legumes, grains, fruits, and vegetables can lead to lipid-reducing effects and optimal plasma lipid levels [9, 10]. Evidence suggests that both dietary soluble fiber and purified viscous soluble fiber intake are associated with lower serum cholesterol levels [11]. In 2002, the Institute of Medicine differentiated between dietary fiber (lignin and nondigestible carbohydrates) and functional fiber (nondigestible carbohydrates and isolates) [12]. Psyllium is an important source of highly soluble dietary fiber [13], derived from the seed husk of *Plantago ovata*, which has gel-forming properties [14–16]. The seed coat of psyllium contains a mixture of polysaccharides, including hexose sugars, pentose sugars, and uronic acids. It is viscous and forms a gel when mixed with water [17]. Psyllium is native to Iran, India, and other countries in the Middle East [18]. The consumption of dietary fiber can decrease total cholesterol, LDL-C, and the LDL-C: HDL-C ratio while increasing HDL-C levels [9–11]. Studies have shown that psyllium consumption offers various nutritional benefits, including management of diarrhea, constipation, colon cancer, ulcerative colitis, irritable bowel syndrome, diabetes, and hypercholesterolemia [20]. Psyllium can effectively lower LDL-C and TG when combined with diet and medications. It has been shown to decrease cholesterol and increase HDL-C in patients with chronic constipation and type II diabetes [1–5]. Additionally, psyllium husk is known to neutralize issues related to diabetes, high cholesterol, hypertriglyceridemia, and high blood pressure [23]. When included in a low-fat diet, psyllium can reduce serum cholesterol levels [24]. Moreover, psyllium husk can lower blood lipids when combined with food [25]. Psyllium husk consumption deceases sd-LDL, but it didn’t make a difference in HDL subclasses [6]. Considering conflicting evidence and limited research in this area, we conducted a systematic review and meta-analysis to better understand the effects of psyllium husk consumption on lipid profiles in the adult population.

Method

This research was registered in the PROSPERO system (CRD42023402987). The research followed the guidelines established by the PRISMA guidelines [27]. We conducted a systematic review to evaluate the effects of psyllium supplementation on lipid profile levels. We started searching from March 15, 2022 to August 2, 2025 and was restricted to English-language articles (PubMed, Institute for Scientific Information WOS, and Scopus).

Data were extracted meticulously using specific keywords: (“lunelax” OR “mucilage polysaccharides” OR “mucilage” OR “Plantago ovata” OR “plant mucilage” OR “psyillium” OR “ispaghul” OR “Metamucil” OR “isogel” OR “plantago” OR “psyllium-husk” OR “ispaghula” OR “Plantago psyllium” OR “Psyllium fiber”) AND (“Pilot study” OR “trial” OR “Controlled Clinical Trials as Topic” OR “randomly” OR “double-blind randomized controlled trial” OR “cluster randomized controlled trials” OR “Clinical Trial” OR “Randomized Controlled Trial” OR “RCTs” OR “Controlled Clinical Trial” OR “RCT” OR “cRCTs” OR “Clinical Trials as Topic” OR “controlled trial*” OR “clinical trial*” OR “Randomized” OR “intervention*” OR “single-blind” OR “double-blind” OR “placebo” OR “single-blind randomized controlled trial” OR “Random Allocation” OR “Single-Blind Method” OR “Cross-Over Studies” OR “Comparative Study” OR “Follow-Up Studies” OR “Double-Blind Method” OR “assignment” OR “parallel” OR “cross-over”). These terms could be used in combination with “OR” and/or “AND” or alone to identify additional studies.

To ensure accuracy, our search was limited to studies involving human subjects only. To prevent duplication, two independent researchers (Zeinab.Gholami and Zeinab.Paknahad) evaluated both the main titles and abstracts. Additionally, we manually searched for other gray literature articles and we set up alerts in the PubMed and Scopus databases, and since 2023 no new studies on this topic have been published. When we encountered articles that we could not access, we reached out to the authors via email. It is needed to notify that it was going to invite the third author if we didn’t agree with overall the findings, but fortunately we both agreed in all steps.

Eligibility criteria

This meta-analysis study utilized the Intervention, Population, Comparison, and Outcomes (PICOS) criteria. The outcomes assessed included changes in cholesterol levels, HDL-C, LDL-C, and TG. The comparisons were made against a control group, the intervention involved psyllium, and the focus was on adults over the age of 18. The study design included randomized controlled trials. both placebo-controlled and active comparator trials were considered. Furthermore, we have performed subgroup analyses separating psyllium vs. control trials from psyllium vs. active comparator trials. This distinction helps clarify that the observed heterogeneity is partly attributable to differences in comparators and not only to dosage variation. The inclusion criteria were as follows: (a) participants must be adults over 18 years old; (b) the study must evaluate the effects of psyllium husk on changes in cholesterol, LDL-C, TG, and HDL-C with a control group; (c) RCTs must have either a crossover or parallel design. Exclusion criteria included: (a) non-RCT studies; (b) reviews, letters, conference summaries, workshops, or studies containing inaccurate data; (c) individuals under 18 years of age; (d) in vitro studies or research involving animals or cell cultures; (e) studies without a control group; (f) studies lacking standard deviation data; (g) defective data sets; (h) articles without referenced means and SDs; (i) non-English articles; (j) studies with a research period of fewer than two weeks.

Data extraction

The data extraction form was completed by two experienced investigators, Zeinab.Gholami. and Zamzam.Paknahad., using a combination of spreadsheet software and a word processor. They conducted a thorough review of all selected papers. To access previously inaccessible articles, they established communication with the corresponding authors via email. After a comprehensive examination of the full texts, the following data were retrieved: publication year, authors’ names, study design (crossover or parallel), location of the study, average age of participants, characteristics of the study population, sample size, health status of participants, gender, duration of intervention, dosage of psyllium, and mean ± SD of lipid profile levels before and after the intervention. Additional manual data extraction was conducted and documented in the specified format. Data from published figures were extracted using Graph Digitizer Get Data software.

Quality assessment

To evaluate the potential for bias, two researchers (Zamzam.Paknahad. and Zeinab.Gholami.) used the Cochrane Collaboration’s risk of bias assessment tool (Cochrane RoB 2.0 tool). Seven criteria were evaluated for each study: (a) Randomization process, (b) Deviations from intended intervention, (c) Measurement of the outcome, (d) Missing outcome data, (e) Selection of the reported result, (f) Overall risk of bias. The overall risk of bias judgment in ROB 2 is determined as low risk if all domains are low risk, some concerns if at least one domain has some concerns but none are high risk, and high risk if any domain is rated high risk [7, 8]. This classification is summarized in Table 2.

Table 2.

Quality assessment

Article	Randomization process	Deviations from intended intervention	Measurement of the outcome	Missing outcome data	Selection of the reported result	Overall risk of bias
Sonia Vega-López et al. Study arm 1	H	H	H	L	L	H
Sonia Vega-López et al. Study arm 2	H	H	H	L	L	H
Sonia Vega-López et al. Study arm 3	H	H	H	L	L	H
Ong Pui Wen et al.	H	H	H	L	L	H
Noureddin Soltanian,	H	H	H	L	L	H
Gary w.neal	H	H	H	L	L	H
James W. Anderson,	H	L	H	L	L	H
James W. Anderson	H	L	H	L	L	H
James W Anderson,	H	L	H	L	L	H
ARRIGO F.G. CICERO	H	L	H	L	L	H
Arrigo F.G. Cicero	H	L	H	L	L	H
Fatemeh Pourbehi1	H	L	L	L	L	H
Vijay Ganji Study arm 1	H	H	H	L	L	H
Vijay Ganji Study arm 2	H	H	H	L	L	H
Gregory T. Everson	H	L	L	L	L	H
JAVED ASGHAR1	H	L	L	L	L	H
David Ja Jenkins Study arm 1	H	H	H	L	L	H
David ja Jenkins Study arm 2	H	H	H	L	L	H
Mark M.M Study arm 1	H	L	L	L	H	H
Mark M.M Study arm 1	H	L	L	L	H	H
Martha R.M	H	H	H	L	L	H
Noureddin Soltaniana	H	H	H	H	L	H
David c k r	H	L	H	L	L	H
Rosa Sola	H	H	H	L	L	H
Rosa Solàa	H	L	H	L	L	H
Gordon Schectman,	H	H	H	H	L	H
Seyedeh Ferdows Jazayeri	L	L	H	L	L	H
SHAH MURAD	H	H	H	L	L	H
Amane Sheikh1	H	H	H	L	H	H
Sudeep Shrestha Study arm 1	H	L	H	H	L	H
Sudeep Shrestha Study arm 2	H	L	H	H	L	H
Sudeep Shrestha Study arm 3	H	L	H	H	L	H
J. David Spence,	L	L	H	H	L	H
Dennis L. Sprecher	H	L	H	L	L	H
Dennis L. Sprecher	H	L	H	L	L	H
C. D. Summerbell,	L	L	H	L	L	H
Thomas MS W	H	H	H	L	H	H
Seyed Ali Ziai	H	L	H	L	L	H
James J. Maciejko	H	L	H	L	L	H
Larry P. Bell,	H	L	H	L	L	H
Ricklefs Ka	H	H	H	L	L	H

H: high; L: low

Data synthesis and statistical analysis

Our goal was to evaluate changes in LDL, HDL-C, TG, and cholesterol levels using a random-effects model to calculate the mean changes and standard deviation [29]. We employed CMA software alongside the random-effects model to assess study heterogeneity and provide accurate pooled prevalence estimates with 95% CI. In this analysis, an I-squared index (I²) value over 50% indicates significant heterogeneity, suggesting the use of a random-effects model to account for data variability. To further address heterogeneity, we employed meta-regression and subgroup analyses. In this study, meta-regression was conducted to explore the relationship between study duration and psyllium dosage. A significant level of P < 0.05 was deemed appropriate for all statistical analyses, which were performed using version 3 of the CMA software. If the SD of the mean difference was not reported in the articles, we calculated it using the following formula:

Equation 1: SD change = √([SD baseline]² + [SD final]² - [2R × SD baseline × SD final]) [9]. For cases where the SE was provided but not the SD, we used Eq. 2: SD = SE×√n. In instances where data was presented only in graphical form without accompanying measures of central tendency and variability, we utilized the Get Data Graph Digitizer software to extract the necessary quantitative data. The I² statistic was used to assess heterogeneity: (I² < 25%) indicates low heterogeneity, (I² = 25–50%) indicates moderate heterogeneity, (I² = 50–75%) indicates severe heterogeneity, and (I² >75%) indicates high heterogeneity [10].Prespecified subgroup analyses were conducted based on baseline levels of LDL, HDL-C, TG, and cholesterol, mean age of participants, health status, sample size, publication bias, study duration (in weeks), sensitivity analysis, and psyllium dosage (in mg/day).

Results

Search results

This research was registered in the PROSPERO system (CRD42023402987). The screening and selection process of the study is illustrated in Fig. 1 using a flowchart. To stay updated on relevant publications, alert notifications have been configured in both Scopus and PubMed. A preliminary search in PubMed returned 998 results, which increased to 1029 after activating the alert feature. The initial search in Scopus yielded 2,320 results, which grew to 2,327 once the alerts were enabled. Additionally, we found 1,048 records in the Web of Science. We eliminated 2,070 duplicate articles and investigated 2,131 abstracts and titles. However, there were 14 studies without full text available. Ultimately, 189 full-text articles were retained for further review. During the screening process, 146 articles were excluded for the following reasons: 101 studies were not relevant, 2 focused on animals, 5 lacked a control group, 19 had an intervention duration of less than 2 weeks, 5 involved non-adult populations, 2 did not report baseline means and standard deviations, 4 did not provide standard deviations, 2 were not RCTs, and 6 were published in languages other than English. As a result, 41 RCTs were included in the final meta-analysis (see Fig. 1).

Fig. 1.

Flow diagram of study selection

Study characteristics

A total of 41 studies were deemed eligible for inclusion in this review. These studies were published between 1988 and 2022, with durations ranging from 14 to 182 days. The total number of participants across these studies was 2,049 (925 controls and 1,124 psyllium intervention). Table 1 presents the characteristics of the articles, which were conducted in various geographical locations, including Iran, Malaysia, Italy, Spain, the USA, Pakistan, Mexico, the UK, Australia, Germany, Canada, the Palestine, and Poland. The average age of participants ranged from 24 to 66.2 years, with more articles published that included both men and women. The psyllium dosage varied from 0.002 to 25 g per day. The apparent “duplicates studies in Table 1” are not repeated entries; they represent different intervention arms reported within the same trial. To avoid confusion, we indicated “Study arm 1,” “Study arm 2,” and etc. or conducted in different years.

Table 1.

General specifications of the recognized studies put in meta-analysis

studies	year	Country	Participant/health status	Study design	outcome	Sex	Sample size		In_Mean Age	Duration (day)	Type of psyllium	Type of control	Intervention
							IG	CG					Dose/day (g / day)
Sonia V. L et al. [11] Study arm 1	2001	USA	Healthy adult/H	crossover	LDL, HDL, TG, cholesterol	M			43.7	30/30	psyllium supplement cookies	control cookies	15
Sonia V. L et al. [11] Study arm 2	2001	USA	Healthy adult/H	crossover	LDL, HDL, TG, cholesterol	F pre m			39.3	30	psyllium supplement cookies	control cookies	15
Sonia V. L et al. [11] Study arm 3	2001	USA	Healthy adult/H	crossover	LDL, HDL, TG, cholesterol	F post m			54.6	30	psyllium supplement cookies	control cookies	15
Ong P.W et al. [12]	2022	Malaysia.	Healthy male adult/H	P, r,pc, un	LDL, HDL, TG, cholesterol	M	14	15	26	30	Psyllium husk	mixed herbs	25
Noureddin S et al. [13]	2018	Iran	patients with T2D, constipation/UH	P, r,pc, si	LDL, HDL, TG, cholesterol	M/F	24	27	58	84	Psyllium cookie	Flaxseed and place cookie	20
Gary W.N et al. [14]	1988	USA	Patient with cholesterol/UH	P, r,oc, open	LDL, HDL, TG, cholesterol	M/F	27	27		49	Psyllium in diet		20.4
James W.A et al. [15]	1988	USA	M hypercholesterolemia. / UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M	14	14	47.6	70	Psyllium Hydrophilic Mucilloid	cellulose placebo	10.2
James W.A et al. [16]	1992	USA	Hypercholesterolemia/ UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	21	23	57	70	Psyllium flake	Wheat bran	10
James W.A et al. [17]	2000	USA	primary hypercholesterolemia. / UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	197	51	51	182	psyllium husk fiber	cellulose placebo	10.2
ARRIGO F.G. C et al. [18]	2007	Italy	severe hyperlipoproteinemia; uncontrolled diabetes/ UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	48	48	48	180	soluble psyllium husk powder	hydrolyzed guar gum	7
Arrigo F.G.C et al. [19]	2010	Italy	Caucasian patients/ UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	48	48	48	180	soluble psyllium husk powder	hydrolyzed guar gum	7
Fatemeh P et al. [20]	2020	Iran	non-diabetic women with PCOS/ H	P, r,pc, d	LDL, HDL, TG, cholesterol	F	24	24	24	56	psyllium	microcrystalline cellulose	10
Vijay G et al. [21] Study arm 1	1996	USA	normolipidemic humans. /H	cr, r,pc, d	LDL, HDL, TG, cholesterol	M/F			31	28	psyllium husk fiber supp	soybean AND PSY	20
Vijay G et al. [21] Study arm 2	1996	USA	normolipidemic humans. / H	cr, r,pc, d	LDL, HDL, TG, cholesterol	M/F			31	28	psyllium husk fiber supp	coconut oil AND PSY	20
Gregory T.E et al. [22]	1992	USA,	moderate hypercholesterolemia/ UH	cr, r,pc, d	LDL, HDL, TG, cholesterol	M			44	40	psyllium	cellulose placebo	15
JAVED A et al. [23]	2011	Pakistan	hyperlipidemic patients/ UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	18	20		90	psyllium husk		10
David J.J et al. [24] Study arm 1	1997	Canada	Healthy subject/ H	cr, r,pc, d	LDL, HDL, TG, cholesterol	M/F			57.5	28	Psyllium + 6% mufa	Wheat bran	11.4
David J.J et al. [24] Study arm 2	1997	Canada	Healthy subject/ H	cr, r,pc, d	LDL, HDL, TG, cholesterol	M/F			58	28	Psyllium + 12% mufa	Wheat bran	12.4
Mark M.M (25) Study arm 1	1998	UK	primary hypercholesterolemia. / UH	P, r,pc, d	LDL, cholesterol	M/F	44	38	51.2	84	ispaghula husk	sucrose	7
Mark M.M (25) Study arm 2	1998	UK	primary hypercholesterolemia. / UH	P, r,pc, d	LDL, cholesterol	M/F	46	35	51.2	84	ispaghula husk	sucrose	10.5
Martha R.M et al. [26]	1998	Mexico	Dm2/UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	63	60	57	42	Plantago Psyllium	microcrystalline cellulose	15
Noureddin S et al. [5]	2018	Iran	patients with type 2 diabetes and chronic constipation/UH	P, r,pc, s	LDL, HDL, TG, cholesterol	M/F	24	27	56.8	84	Psyllium cookie	Placebo cookie	20
David c k r et al. [27]	1994	Australia.	Hyperlipidemia/UH	cr, r,pc, d	LDL, HDL, TG, cholesterol	M			49.1	84	psyllium	Wheat bran	12
Rosa S et al. [28]	2007	Spain	CVD/UH	cr, r,pc, s	LDL, HDL, TG, cholesterol	M	15	13	61.4	56	ovata husk ovata seeds	Insoluble fiber	10.5
Rosa S et al. [29]	2010	Spain	CVD/UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	94	93	54.24	56	Po-husk	microcrystalline-cellulose	14
Gordon S et al. [30]	1993	USA	Type Ifa hyperlipidemia/UH	P, r,pc,	LDL, HDL, TG, cholesterol		98	79	66.2	60	Psyllium	Niacin	10.4
Seyedeh Ferdows Jazayeri [31]	2021	Iran	NAFLD/UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	31	32	43.3	84	Plantago major	toasted flour powder	4
SHAH M et al. [32]	2010	Pakistan	primary hyperlipidemic/UH	P, r,pc, s	LDL, HDL, TG, cholesterol	M/F	18	20	51.2	90	psyllium husk,	partly grinded wheat	10
Amane Sh et al. [33]	2017	Iran	type 2 diabetes/UH	P, r,pc, s	TG	M/F				42	psyllium powder	pears	10
Sudeep Sh et al. [34] Study arm 1	2006	UK	healthy adults/H	cr, r,pc, d	LDL, HDL, TG, cholesterol	M				30/28	psyllium	Plant sterol	7.28
Sudeep Sh et al. [34] Study arm 2	2006	UK	healthy adults/H	cr, r,pc, d	LDL, HDL, TG, cholesterol	F PRE				30/28	psyllium	Plant sterol	7.28
Sudeep Sh et al. [34] Study arm 3	2006	UK	healthy adults/H	cr, r,pc, d	LDL, HDL, TG, cholesterol	F POST				30/28	psyllium	Plant sterol	7.28
J. David S et al. [35]	1995	UK	moderate primary hypercholesterolemia/UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	24	29	55	70	psyllium	Placebo cellulose Combination Therapy Colestipol Alone	15
Dennis L.S et al. [36] Study arm 1	1993	USA	Healthy/H	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	18	19	50.9	112	psyllium High-fat diet	placebo	10.2
Dennis L.S et al. [36] Study arm 2	1993	USA	Healthy/H	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	41	40	60	112	psyllium low-fat diet	placebo	10.2
C. D. Summerbell et al. [37]	1994	UK	moderate hypercholesterolemia/UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	19	18	47	63	psyllium		14.2
Thomas MS W et al. [38]	1994	Canada	raised serum cholesterol/UH	cr, r,pc, d	LDL, HDL, TG, cholesterol	M/F				14	Psyllium fiber	wheat bran	7.3
Seyed A.Z et al. [39]	2005	Iran	diabetic patients/UH	P, r,pc, d	LDL, HDL, TG, cholesterol		21	15	51.9	56	Psyllium	microcrystalline cellulose	10.2
James J. M et al. [40]	1994	USA	primary hypercholesterolemia/UH	cr, r,pc, d	LDL, HDL, TG, cholesterol	M/F	18			42	psyllium hydrophilic mucilloid and cholestyramine	cholestyramine and placebo
Larry P. B et al. [41]	1989	USA	moderate hypercholesterolemia/UH	P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	40	35	46.2	56	psyllium		10.2
Ricklefs K et al. [42]	2017	USA		P, r,pc, d	LDL, HDL, TG, cholesterol	M/F	8	9	58.5	56	ground psyllium	ground flaxseeds	9

N: number; In N: intervention number; Pl N; placebo number; IN: intervention; M: male; F: female; IG: intervention group; CG: control group; H: Healthy; UH: Unhealthy; F pre m: Female pre-menopausal; F post m: Female post-menopausal; Pc, placebo control; cr.r: randomized crossover design; P.r, randomized parallel design; d: Double-Blind Study; s: Single-Blind Study; LDL-C: Low-Density Lipoprotein Cholesterol; TG: Triglyceride; HDL_C: High-Density Lipoprotein Cholesterol

Meta-analysis results

A total of 41 articles involving 2,049 (925 controls and 1,124 psyllium intervention) assessed the influence of psyllium on lipid profiles. A random-effects model was employed to demonstrate a significant reduction in LDL-C and total cholesterol levels, a nonsignificant decline in TG, and a nonsignificant increase in HDL-C (see Fig. 2). The following results were observed: - TG: Weighted Mean Difference (Weighted Mean Differences (WMD)): -5.29; 95% CI: -12.14 to 1.54; p > 0.05 - HDL-C: WMD: 0.57; 95% CI: -0.88 to 2.04; p > 0.05 - LDL-C: WMD: -8.55; 95% CI: -12.92 to -4.19; p < 0.0001 - Cholesterol: WMD: -9.05; 95% CI: -13.71 to -4.40; p < 0.05 Significant heterogeneity was found for the following: - LDL-C: I² = 88.46%, p < 0.0001 - TG: I² = 83.25%, p < 0.001 - HDL-C: I² = 77.96%, p < 0.001 - Cholesterol: I² = 89.30%, p < 0.0001 A sensitivity analysis was conducted to assess the robustness of these findings. The association remained statistically significant when each study was excluded sequentially from the analysis. The results were robust for HDL-C (WMD changed from 0.17 to 1.05), cholesterol (WMD changed from − 10.14 to -7.65), and LDL-C (WMD changed from − 9.44 to -7.28). However, the sensitivity varied between the two studies on TG (WMD changed from − 9.31 to -3.66). The apparent “duplicates studies in Table 1” are not repeated entries; they represent different intervention arms reported within the same trial. To avoid confusion, we indicated “Study arm 1,” “Study arm 2,” and etc. or conducted in different years. (see Table 2). To assess the robustness of the results and identify potential sources of heterogeneity, a sensitivity analysis was conducted using the leave-one-out method. In this approach, one study was removed at a time from the meta-analysis, and the impact on the overall effect size and heterogeneity was evaluated. Overall, the sensitivity analysis indicated that the results were partially dependent on specific studies, and their removal could influence both the overall effect estimate and the degree of heterogeneity. This highlights the importance of considering the methodological quality of included studies and conducting additional analyses to assess the robustness of the findings. In this meta-analysis, a very high level of heterogeneity was observed (I² >90%), indicating substantial variation across the included studies. Several factors may have contributed to this heterogeneity. First, differences in psyllium dosage across studies (ranging from low doses to more than 10 g/day) could have significantly influenced the outcomes, as the metabolic effects of psyllium are likely dose-dependent. Second, the duration of intervention varied widely among studies, with some lasting only a few weeks while others extended for several months. Short-term interventions may not have been sufficient to produce meaningful changes in lipid profiles, thereby increasing variability. Additionally, variation in study quality, including aspects like randomization methods, blinding, and control for confounders, may have also played a role. This high degree of heterogeneity limits the reliability and generalizability of the pooled estimates and highlights the importance of sensitivity and subgroup analyses to identify potential sources of variability. Therefore, findings should be interpreted with caution, and future research should prioritize rigorous study design and standardized.

Fig. 2.

(a-d) Forest plot illustrating weighted mean difference and 95% confidence intervals for the impact of psyllium consumption on LDL-C (a), HDL_C (b), TG (c), and cholesterol (d). LDL-C: Low Density Lipoprotein Cholesterol; HDL-C: High Density Lipoprotein Cholesterol; TG: Triglyceride

Subgroup analysis

Subgroup analysis revealed variability in the effects of psyllium on LDL-C and cholesterol levels. The significance of cholesterol and LDL-C reduction varied with dose (greater than 10 g per day) and duration, with notable effects observed for intervention periods both shorter and longer than 50 days. For TG, a significant reduction was observed with psyllium consumption at a dosage below 10 g per day and for durations shorter than 50 days. However, HDL-C levels did not show a significant dependence on either dosage or duration of psyllium husk treatment. Overall, meta-regression indicated that dosage and duration had no statistically significant linear impact on the lipid profile. Table 3 provides a summary of the outcomes from the subgroup analyses.

Table 3.

Subgroup analysis

Variable	Dose		Duration
	< 10 g /day	> 10 g/day	< 50 day	> 50 day
LDL-C
Number of comparisons	5	21	3	23
WMD (95% CI)	1.99 (-1.29, 5.28)	-11.06 (-15.52, -6.60)	-11.80 (-15.24, -8.37)	-8.29 (-13.40, -7.86)
P value	0.23	0.001	0.00	0.00
I squared	0.001	86.81	0.00	90.82
p- heterogeneity	0.60	0.00	0.93	0.00
HDL-C
Number of comparisons	4	20	3	21
WMD (95% CI)	1.39 (-0.45, 3.24)	0.47 (-1.23, 2.17)	2.16 (-4.72, 9.07)	0.55 (-0.92, 2.02)
P-value	0.13	0.58	0.53	0.46
I squared	52.73	79.56	90.78	74.19
p- heterogeneity	0.09	0.001	0.001	0.001
TG
Number of comparisons	4	19	3	20
WMD (95% CI)	-11.81 (-18.22, -5.40)	-3.34 (-11.92, 5.23)	-12.93 (-20.30, -5.56)	-3.40 (-11.56, 4.74)
P-value	0.001	0.44	0.001	0.41
I squared	0.001	78.70	0.001	77.53
p- heterogeneity	0.67	0.00	0.42	0.001
Cho
Number of comparisons	6	21	3	24
WMD (95% CI)	-0.88 (-4.01,2.24)	-11.58 (-16.92, -6.24)	-11.98 (-20.88, -3.08)	-8.72 (-14.00, -3.45)
P value	0.58	0.001	0.001	0.001
I squared	0.00	91.38	53.66	91.91
p- heterogeneity	0.96	0.00	0.11	0.00

LDL-C: Low Density Lipoprotein Cholesterol; HDL_C: High Density Lipoprotein Cholesterol; TG: triglyceride; Cho: cholesterol; CI: confidence interval; WMD: weighted mean difference

Dose-response analyses

Meta-regression analyses were conducted to explore potential sources of heterogeneity and to examine dose-response relationships. Subgroup analyses helped identify sources of heterogeneity, while meta-regression indicated linear trends between intervention characteristics and outcomes. A steeper slope in the regression line indicates a stronger association. Psyllium dosage had an insignificant negative effect on LDL-C (slop: -0.34; p value > 0.05), HDL-C (slop: -0.01; p value > 0.05), TG (slop: -0.56; p value > 0.05), and cholesterol (slop: -0.50; p value > 0.05). The duration did not significantly affect any of these variables: LDL-C (slop: 0.02; p value > 0.05), HDL-C (slop: 0.00; p value > 0.05), TG (slop: -0.04; p value > 0.05), and cholesterol (slop: -0.06; p value > 0.05) (Fig. 3). Table 4 presents the results of the meta-regression.

Fig. 3.

(a-d) Dose-response relations between psyllium dosage (mg/d) and mean difference in LDL-C (a), HDL_C (b), TG (c), cholesterol (d). LDL-C: Low Density Lipoprotein Cholesterol; HDL-C: High Density Lipoprotein Cholesterol; TG: Triglyceride

Table 4.

meta-regression

variable	Slope	95% CI	P-value
LDL-C
Dose	-0.34	-1.18, 0.49	0.41
duration	0.02	-0.07, 0.11	0.67
HDL_C
Dose	-0.01	-0.28, 0.25	0.91
duration	0.00	-0.01, 0.03	0.52
TG
Dose	-0.56	-1.91, 0.78	0.41
duration	-0.04	-0.19, 0.10	0.55
Cho
Dose	-0.50	-1.37, 0.36	0.25
duration	-0.06	-0.17, 0.04	0.26

LDL-C: Low Density Lipoprotein; HDL_C: High Density Lipoprotein Cholesterol; TG: triglyceride; Cho: cholesterol; CI: confidence interval

Publication bias

Estimates of publication bias are presented graphically in Supplementary Fig. 5. Egger’s test did not indicate any publication bias concerning the effects of psyllium husk on HDL-C (p = 0.26), TG (p = 0.06), or cholesterol (p = 0.19); however, a publication bias was observed in LDL-C (p = 0.04). Begg’s test showed no publication bias for TG (p = 0.77), HDL-C (p = 0.27), total cholesterol (p = 0.75), or LDL-C (p = 0.36). Trim-and-fill analyses were conducted, and reviews of 9, 8, 10, and 5 studies revealed that psyllium had a stronger effect on lowering HDL-C (WMD): -1.34, 95% CI: −2.81, 0.13), LDL-C (WMD: -14.58, 95% CI: −19.06, -10.10), cholesterol (WMD: -15.59, 95% CI: −20.36, -10.81), and TG (WMD: -13.20, 95% CI: −20.85, − 5.56). The results of the publication bias analysis are shown in Table 5.

Table 5.

Publication bias

variable	Corrected effect size			Begg				Egger							Fail safe n test
	Study trimmed	WMD	CI 95%	KENDALL TAU		Z value	P-value 2 tailed	intercept	CI 95%	T value		Df		P-value	n
LDL-C	8	-14.58	-19.06, -10.10	-0.12	0.90		0.36	1.47	0.01, 2.93	2.08	24.00		0.04		1182.00
HDL_C	9	-1.34	-2.81, 0.13	0.16	1.09		0.27	0.81	-0.67, 2.30	1.13	22.00		0.26		0.00
TG	5	-13.20	-20.85, -5.56	-0.04	0.29		0.77	0.93	-0.06, 1.93	1.93	21.00		0.06		148.00
Cho	10	-15.59	-20.36, -10.81	0.04	0.31		0.75	1.04	-0.58, 2.68	1.31	25.00		0.19		1206.00

LDL-C: Low Density Lipoprotein; HDL_C: High Density Lipoprotein Cholesterol; TG: triglyceride; Cho: cholesterol; CI: confidence interval; WMD: weighted mean difference; DF: degrees of freedom; N: number

Adverse events

The main side effects of psyllium husk reported in studies include abdominal cramps, respiratory tract infections, heartburn, constipation, gas, bloating, frequent nighttime urination, diarrhea, increased flatulence, fullness, more frequent defecation, mild gastrointestinal complications, and stomach pains.

Discussion

This study compared the effects of psyllium to a control group and found a significant decrease in LDL-C and total cholesterol levels, a nonsignificant decrease in TG, and a nonsignificant increase in HDL-C. Notable heterogeneity was observed in the lipid profiles. A sensitivity analysis assessed the robustness of these findings and indicated that the association remained statistically significant when each study was excluded from the analysis sequentially. The results were consistent for LDL-C and HDL-C, as well as total cholesterol, although the findings varied between the two TG studies. Subgroup analysis revealed variability in the effects of psyllium on LDL-C and cholesterol levels. The significance of cholesterol and LDL-C reduction varied with dose (greater than 10 g per day) and duration, with notable effects observed for intervention periods both shorter and longer than 50 days. For TG, a significant reduction was observed with psyllium consumption at a dosage below 10 g per day and for durations shorter than 50 days. However, HDL-C levels did not show a significant dependence on either dosage or duration of psyllium husk treatment. Overall, meta-regression indicated that dosage and duration had no statistically significant linear impact on the lipid profile, these effects were not uniformly significant. Additionally, Egger’s test indicated no evidence of publication bias in studies examining psyllium’s impact on lipid profiles except LDL-C. Begg’s test showed no publication bias for lipid profiles. Trim-and-fill analyses were conducted, and reviews of 9, 8, 10, and 5 studies revealed that psyllium had a stronger effect on lowering lipid profiles. no evidence of publication bias was observed for most lipid parameters, with the exception of LDL-C, for which a small-study effect was detected.

Dietary fiber is known to reduce total cholesterol, LDL-C, and the LDL-C to HDL-C ratio, while also increasing HDL-C [43, 44]. Psyllium husk can aid in conjunction with diet and medications aimed at improving low LDL-C and TG [45]. It has been shown to reduce total and LDL cholesterol levels [46], while also addressing conditions like diabetes, high cholesterol, hypertriglyceridemia, and hypertension [18]. In individuals with chronic constipation and type II diabetes, psyllium can decrease cholesterol levels and boost HDL-C [1–5]. Consumption of psyllium within a low-fat diet has been associated with lower serum cholesterol [47]. Furthermore, when psyllium husk is mixed with food, it can effectively lower blood lipids [38]. Although psyllium husk consumption reduced sd-LDL, it did not significantly affect HDL subclasses [6]. The fat-lowering effects of viscous fiber are thought to result from its ability to bind bile acids, thereby enhancing fecal excretion. Viscous fiber also promotes intestinal fermentation, leading to increased production of SCFA and modifications in cholesterol metabolism [46, 48]. Reduced postprandial insulin secretion may contribute to lower fatty acid synthesis and decreased circulating TG [46]. Dietary fiber can lower serum and liver lipids through several mechanisms, including binding with bile, increasing viscosity, and delaying food transit in the small intestine. Additionally, it inhibits lipid and glucose absorption, promotes the synthesis of short-chain fatty acids, and regulates lipid metabolism genes [49]. The primary mechanism of action for soluble fiber appears to relate to increased fecal bile acid excretion and a heightened rate of chenodeoxycholate production [43].

Strengths and limitations

Strengths of this study include: (a) Our meta-analysis included more articles than previous studies, allowing for a comprehensive evaluation of the effects of psyllium husk on lipid profile; (b) subgroup analyses were performed to identify sources of heterogeneity among the studies; (c) a sensitivity analysis was conducted to demonstrate the reliability of the findings, and (d) Some studies only conducted meta-analysis on specific diseases, but ours included both healthy and unhealthy individuals. However, there are several limitations to this study: (a) Several outcomes exhibited observed heterogeneity. Therefore, future research should aim to identify the sources of this heterogeneity and provide more optimized analyses through more rigorous study designs and better control of confounding variables; (b) several articles did not consider dietary intake, which may affect lipid profiles, we suggest examining dietary intakes in future studies; (c) The number of articles included in the analysis was limited. We suggest not to consider the exclusion Criteria in future studies and consider case-control, cross sectional, RCT, and etc. studies to increase data volume and improve the generalizability of the results; and (d) There was a variation in ages among participants. We suggest that future studies be conducted separately for children, young people, and the elderly, (e) restricting the search to English-language articles and only two databases may introduce a risk of selection bias.

Conclusion

This study investigates the effects of psyllium husk on cholesterol, TG, LDL_C, and HDL_C. We discovered that psyllium can decrease LDL_C and cholesterol levels, insignificantly decrease TG, and insignificantly increase HDL_C. Subgroup analysis revealed variability in the effects of psyllium on LDL-C and cholesterol levels. The significance of cholesterol and LDL-C reduction varied with dose (greater than 10 g per day) and duration, with notable effects observed for intervention periods both shorter and longer than 50 days. For TG, a significant reduction was observed with psyllium consumption at a dosage below 10 g per day and for durations shorter than 50 days. However, HDL-C levels did not show a significant dependence on either dosage or duration of psyllium husk treatment. Overall, meta-regression indicated that dosage and duration had no statistically significant linear impact on the lipid profile, these effects were not uniformly significant. Additionally. psyllium’s impact on lipid profile remains inconclusive and warrants further research. We emphasize the need for future trials with standardized dosages, longer durations, and control of confounding variables such as diet and physical activity.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ISI

Institute for Scientific Information

WOS

ISI Web of Science

RCTs

Randomized Controlled Trials

WMD

Weighted Mean Differences

CI

Confidence Intervals

LDL-C

Low-Density Lipoprotein Cholesterol

TG

Triglyceride

HDL_C

High-Density Lipoprotein Cholesterol

CHD

Coronary Heart Disease

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

CMA

Comprehensive Meta-Analysis

SD

Standard Deviation

SE

Standard Error

I²

I-squared index

sd-LDL

Small Dense-Low-Density Lipoprotein Cholesterol

SCFA

Short-Chain Fatty Acids

Author contributions

The manuscript was read and approved by all authors; Zeinab.Gholami: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Validation; Roles/Writing - original draft. Zamzam.Paknahad: Conceptualization; Supervision; Validation; Visualization; Writing - review & editing.

Funding

This study was supported by Isfahan University of Medical Sciences, (Code: IR.MUI.RESEARCH.REC.1402.017, Grant number: 140206).

Data availability

Data are not publicly available due to ethics committee restrictions. However, data are available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

The study was approved by the Medical Ethics Committee of Isfahan University of Medical Sciences (IR.MUI.RESEARCH.REC.1402.017, Grant number: 140206). This research was registered in the PROSPERO system (CRD42023402987). Zamzam Paknahad and all authors have read and approved the final version of the manuscript had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.

Consent for publication

Not applicable.

Approval of the submission

Zamzam.Paknahad. and all authors have read and approved the final version of the manuscript have full access to all of the data in this study, and take complete responsibility for the integrity of the data and the accuracy of the data analysis.

Transparency statement

The primary author, Zamzam.Paknahad. confirms the authenticity, accuracy, and transparency of the current study’s report. Moreover, no crucial elements of the research have been omitted, and any deviations from the original study design (if applicable, registered) have been clarified.

Full uncropped gels and blots image(s)

Not applicable.

Human ethics and consent to participate declarations

Not applicable.

Competing interests

The authors declare no competing interests.