Inositol supplementation efficacy in improving key cardiometabolic and anthropometric indices: a GRADE-assessed systematic review and meta-analysis of randomized controlled trials

Abstract

Background

Inositol and its derivatives may help mitigate risks associated with cardiovascular diseases; however, existing evidence remains inconsistent.

Objectives

The primary aim of this systematic review and meta-analysis of RCTs was to quantify the effects of inositol and its stereoisomers on anthropometric, and cardiometabolic measures.

Methods

A systematic review and meta-analysis of RCTs was conducted to clarify this. Searches in PubMed and Scopus, along with hand-searching references, identified RCTs on inositol supplementation lasting ≥ 4-week. Using random-effects models, the analysis determined mean effect sizes as weighted mean differences (WMD) with 95% CIs. Heterogeneity was assessed via the Cochrane Chi-squared test and Galbraith plots, while the ROBI tool evaluated bias risk. The strength of the evidence was assessed using the GRADE framework.

Results

Eighteen RCTs (n = 898) were totally included in this meta-analysis. Significant reductions in BMI (WMD (95%CIs):-0.57 kg/m²(-1.10, -0.03), I²=88.6), waist-to-hip ratio(WMD (95%CI):-0.02(-0.04, -0.001), I²=84.1), and waist circumference (WMD (95%CI):-2.36 cm(-4.39, -0.33) I²=55.0) were noted with high heterogeneity and low to very low certainty evidence. Inositol significantly decreased glucose levels (WMD (95%CI):-7.25 mg/dL(-10.98, -3.52), I²=90.7), insulin (WMD (95%CI):-4.74µU/mL(-6.16, -3.32), I²=90.6) and HOMA-IR (WMD (95%CI):-1.21(-1.58, -0.85), I²=85.0), both with moderate evidence certainty and high heterogeneity. Notable reductions in triglycerides (WMD (95%CI):-29.80 mg/dL(-48.16, -11.44), I²=96.0) and total-cholesterol (WMD (95%CI):-18.26 mg/dL(-30.75, -5.77), I² = 95.8) were observed, with high and low evidence certainty, respectively, and high heterogeneity. LDL-C and HDL-C improved with moderate certainty (WMDs (95%CIs):-5.15 mg/dL(-8.89, -1.42), I²= 0.0; and 2.76 mg/dL(1.16, 4.36), I²=52.9). Additionally, inositol significantly lowered systolic (WMD (95%CI):-5.34mmHg(-6.91, -3.78), I²=38.0) and diastolic blood pressure (WMD (95%CI):-6.12mmHg(-8.44, -3.80), I²=69.7) with low, and very low evidence certainty.

Conclusion

Overall, inositol may offer modest cardiometabolic benefits, with moderate-to-high certainty for improvements in insulin resistance and lipid profiles. However, existing studies show a high risk of bias and low certainty of evidence, particularly for anthropometric outcomes, creating cautious interpretation. Future research should involve large-scale, rigorous trials with standardized protocols, longer follow-up, and diverse populations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13098-025-01980-6.

Introduction

Individuals with cardiovascular diseases (CVDs) commonly experience comorbid metabolic disturbances. These include overweight or obesity as indicated by higher waist circumference (WC) and body mass index (BMI). Additional conditions include impaired glycemic control, such as hyperglycemia and insulin resistance. Among those with CVDs, hypertension (high blood pressure) and dyslipidemia characterized by elevated levels of total cholesterol (TC), triglycerides (TG), and low-density lipoprotein cholesterol (LDL-C), along with low high-density lipoprotein cholesterol (HDL-C) are frequently observed [1–6]. Notably, these risk factors have been recognized as indirect effects of diet on cardiometabolic-related morbidity and mortality [5, 7–10]. It is noteworthy that globally, in 2019 approximately 17.6 million deaths were associated with different types of CVDs, with around 52% (about 9.1 million) of these deaths being linked to dietary risk factors. Furthermore, nearly 45% of deaths related to cardiometabolic conditions are linked to poor dietary patterns [7, 11–13]. Given the significant economic and health burdens that CVDs impose on individuals and populations, prioritizing prevention, early detection, and risk factor management is essential to slowing their progression [8, 14].

In this context, non-pharmacological approaches have garnered considerable interest, as they tend to have lower costs, produce fewer side effects, and may promote better adherence than pharmacological or more invasive treatments. Dietary supplements with potential cardiometabolic benefits have been extensively studied, particularly compounds with anti-hyperglycemic, anti-hyperlipidemic, antioxidant, and anti-inflammatory properties [5, 7–10, 15–18]. Inositol and its derivatives have been recognized for its contribution to a range of physiological processes, such as cell growth, hormonal and metabolic balance, stress responses, calcium metabolism, and neurotransmitter-induced signal transduction. There has also been a surge in interest in examining the use of inositol for metabolic disorders. As a member of the glucose family, inositol is considered to be a hexahydroxycyclohexane alcohol [19]. There is substantial evidence from experimental studies demonstrating the beneficial effects of inositol on metabolic disturbances [19–24]. However, the current evidence regarding the effects of inositol for mitigating cardiometabolic conditions remains inconclusive.

While previous meta-analyses have highlighted inositol’s benefits for conditions like polycystic ovary syndrome (PCOS) [25] and metabolic disorders [26], these studies mainly focused on specific outcomes—such as glycemic control and lipid levels—or on homogeneous populations. This approach left gaps in understanding its broader cardiometabolic effects across diverse risk profiles. Additionally, current evidence lacks rigorous assessment of clinical relevance through minimal clinically important differences (MCID) and standardized dosing protocols [25–29]. The present systematic review and meta-analysis addresses these gaps by evaluating inositol’s efficacy across populations both with and without diagnosed metabolic disorders, including those with heterogeneous cardiometabolic conditions like obesity, non-alcoholic fatty liver disease (NAFLD), type 2 diabetes (T2DM), and polycystic ovary syndrome (PCOS). We incorporate MCID thresholds to assess clinical significance. The outcomes examined encompass various measures of glycemic control (plasma glucose, HbA1c, insulin, HOMA-IR), lipid profiles (TC, TG, LDL-C, HDL-C), anthropometric parameters (BMI, waist circumference, waist-to-hip ratio), and blood pressure (systolic and diastolic).

Methods

Search strategy

A comprehensive literature review was conducted using Medline (PubMed), Scopus, and additional grey literature sources from their inception through July 2024. Gray literature searches included theses, dissertations, and clinical trial registries (e.g., ClinicalTrials.gov). Using a combination of medical subject headings (MeSH) and free-text queries, our aim was to identify studies pertinent to our research question. The scope was confined to studies of human participants available in the English language. Additionally, a search of cross-references along with hand searches of the references of included articles was performed to identify further related publications that may have been excluded in the initial database searches. Detailed information regarding the search terms used—encompassing titles, abstracts, and keywords—are provided in Supplementary File 1; Table 1.

Table 1.

PICOS criteria for inclusion of studies

Parameter	Criterion
Population	- Adults ≥ 18 years - Human participants diagnosed with metabolic disorders
intervention	- Inositol - Myo-Inositol-1-Phosphate Synthase - Myo-Inositol - Chiro-Inositol - D-Chiro-Inositol - Pinitol
Comparison	- Placebo - No intervention - Standard diet
Outcome	Anthropometric and cardiometabolic risk factors, including: - Anthropometric measures (Body Mass Index [BMI], Waist Circumference [WC], and Waist-Hip Ratio [WHR] - Blood glucose - Hemoglobin A1c (HbA1c%) - Insulin resistance (HOMA-IR, Insulin) - Lipid profiles (TG, TC, LDL, HDL) - Blood Pressure (SBP, DBP)
Study design	Cross-over or Parallel randomized, controlled trials

In this systematic review and meta-analysis, we followed the 2020 Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [30]. A detailed protocol for this systematic review was registered in the International prospective register of systematic reviews (PROSPERO) database, which can be identified by the registration number CRD42024532838 (Supplementary File 2, Table 2).

Table 2.

Participant and study characteristics

Ref. No.	Geographical location (First author, year)	Study design	Health conditions	Biological Sex	Comparison	Type of intervention/ ingredients	Inositol Dosage	Intervention duration	Mean age, in Intervention/ Control years	N intervention/control	Mean BMI, in intervention/control
[50]	Caracas, Venezuela (Nestler,1999)	Double-blind, RCT	Obese, PCOS	Female	Placebo	D-Chiro-Inositol	1200 mg	6–8 weeks	Intervention: 29 ± 6, Control: 26 ± 5	Intervention:19, Control: 6	Intervention: 31.3 ± 2.4 Control: 31.0 ± 2.2
[48]	San Francisco, California, USA (Davis, 2000)	Double-blind, RCT	Obese, T2DM, glucose intolerant	Male and female	Placebo	Pinitol	20 mg/kg/day	4 weeks	51 ± 12	Intervention: 12, Control: 10	36.6 ± 4.9 kg/m2
[51]	Venezuela (Iuorno, 2002)	Double-blind, RCT	PCOS	Female	Placebo	D-Chiro-Inositol	600 mg	6–8 weeks	Intervention: 28.2 ± 1.5, Control: 26.5 ± 1.4	Intervention:10, Control:10	Intervention 22.4 ± 0.3, Control: 22.1 ± 0.3
[47]	Korea (Kim, 2005)	double-blinded, RCT	T2DM	Male, female	Placebo	Pinitol	600 mg	13 weeks	Intervention: 59.97 + 3.1, control: 61.77 + 2.0	Intervention:15, control: 15	Intervention: 24.4 ± 0.7, Control: 23.8 ± 0.5
[37]	Italy (Genazzani, 2008)	Controlled clinical study	Overweight, PCOS	Female	Folic acid	Myo-inositol + folic acid	2 g of myo-inositol + 200 mcg of folic acid	12 weeks	Reproductive age (Mean not reported)	Intervention:10, control:10	Intervention: 29 + 1.6, Control: 27.8 + 2.1
[38]	Vicenza, Italy (Costantino, 2009)	Double-blind, RCT	PCOS	Female	Folic acid	Myo-inositol + folic acid	4 g Myo-inositol + 400 mcg folic acid	12–16 weeks	interventions: 28.8 ± 1.5, Control: 27.1 ± 1.4	Intervention: 23, Control: 19	Intervention: 22.8 ± 0.3, Control: 22.5 ± 0.3
[39]	Padua, Italy (Dona, 2012)	Randomized clinical study	PCOS	Female	Placebo	Myo-inositol	1200 mg	12 weeks	Intervention: 23.5 ± 2.1, Control: 23.6 ± 1.4	Intervention: 18, Control: 8	Intervention: 21.6 + 1.9, Control 21.9 + 0.6
[46]	Korea (Kim, 2012)	double-blind RCT	T2DM (poorly controlled)	Male, Female	Placebo	3-O-methyl-chiro-inositol	1200 mg	12 weeks	Intervention: 56.3 ± 9.8, Control: 52.9 ± 10.5	Intervention: 33, Control: 33	Intervention: 25.30 ± 3.25, Control: 26.13 ± 3.03
[40]	Messina, Italy (Santamaria, 2012)	RCT	Postmenopausal women with metabolic syndrome	Female	Placebo	Myo-inositol	2 g twice daily	24 weeks	Intervention: 55.6 ± 3.2, Control: 55.0 ± 3.2	Intervention: 40, Control: 40	Intervention: 31.5 ± 2.4, Control: 30.7 ± 2.5
[41]	Modena, Italy (Artini 2013)	RCT	PCOS	Female	Folic acid	Myo-Inositol + folic acid	2 g myo-inositol + 200 mcg folic acid	12 weeks	Intervention: 34.9 ± 2.1, Control: 36.2 ± 2.3	Intervention: 25, Control: 25	Intervention: 26.5 ± 6.1, Control: 26.3 ± 6.8
[42]	Milan, Italy (D’Anna, 2013)	Double-blind RCT	Women with BMI > 30 kg/m², metabolic syndrome	Female	Folic acid	Myo-Inositol + folic acid	2 g myo-inositol + 200 mcg folic acid	12–13 weeks	Intervention: 30.9 (18–44), Control: 31.7 (19–43)	Intervention: 80, Control: 80	Intervention: 33.8 (30.0–46.9), Control: 33.8 (30.0–46.0)
[52]	Spain (Banuls, 2015)	Double-blind RCT	Healthy subjects	Male, Female	Sucrose-Sweetened	Inositol-enriched beverage	2.23 g per dose	12 weeks	Intervention: 34.0 ± 10.9, Control: 32.5 ± 10.2	Intervention: 20, Control: 20	Intervention: 23.8 ± 3.2, Control: 24.2 ± 3.5
[43]	Italy (Benelli, 2016)	RCT	Obese PCOS	Female	Folic acid	Myo-Inositol + D-Chiro-Inositol + folic acid	550 mg Myo-Inositol, 13.8 mg D-Chiro-Inositol, 200 mcg folic acid	24 weeks	Intervention: 23 ± 6.8, Control: 25 ± 7.3	Intervention: 21, Control: 25	Intervention: 32 ± 4.8, Control: 31 ± 4.6
[31]	Korea (Lee, 2019)	Double-blind RCT	NAFLD	Male, Female	dextrin and magnesium stearate	Pinitol	Low: 300 mg/day, High: 500 mg/day	12 weeks	Low: 46.1 ± 1.7, High: 45.6 ± 2.1, Control: 47.7 ± 2.1	Low: 30, High: 30, Placebo: 30	Low: 27.1 ± 0.6, High: 28.6 ± 0.5, Control: 27.1 ± 0.5
[53]	Iran (Tutunchi, 2023)	RCT	Obese, NAFLD	Male, Female	Dietary recommendation	Myo-Inositol sachet containing 4 g powder + dietary recommendation	4 g powder	8 weeks	Intervention: 38.52, Control: 36.33	Intervention: 23, Control: 24	Inositol Intervention: Control: 33.75
[49]	East Carolina, Greenville (Campbell, 2004)	RCT	Older adults, varying insulin sensitivity (Healthy subjects)	Male, Female	Placebo	Pinitol	1000 mg twice daily	7 weeks	Intervention:66 + 11, Control: 65 + 6	Intervention: 7, Control: 8	27.9 ± 3.3 kg/m²
[45]	Italy (Capasso, 2013)	RCT	Postmenopausal women with metabolic syndrome at risk of breast cancer	Female	Placebo	Inositol + Alpha Lipoic Acid	Not specified	24 weeks	Intervention: 57.71 ± 7.9, Control: 58.2 ± 5.6	Intervention: 77, Control: 78	Intervention: 30.35 ± 5.3, Control: 29.21 ± 4.82
[44]	Catania, Italy (Cianci, 2015)	Randomized clinical study	PCOS	Female	Control group	D-chiro-inositol (DCI) and alpha lipoic acid	D-chiro-inositol: 1000 mg per day - Alpha lipoic acid: 600 mg per day	24 weeks	23.8 ± 2.5	Intervention: 26, Control: 20	Intervention: 28.7 ± 2, Control: 28.6 ± 2

RCT: randomized clinical (controlled) trial; NAFLD: non-alcoholic fatty live disease; MetS: metabolic syndrome; T2DM: type 2 diabetes mellitus; PCOS: Polycystic ovary syndrome

Study selection

The criteria for defining the research question were established using the PICOS (Participants, Intervention, Comparison, Outcomes, Study design) framework, which is illustrated in Table 1. Although the grey literature was searched, only published studies including cross-over or parallel randomized controlled trials and prospective designs, ensuring a comprehensive evaluation of the effectiveness of the interventions, data completeness, methodological rigor, and reproducibility.

The studies needed to investigate the effects of inositol supplementation for a minimum duration of four weeks on at least one primary outcome of interest. The outcomes of interest included anthropometric data, glycemic control factors, lipid profiles, as well as blood pressure. The population targeted consisted of adults aged 18 years and older, and the interventions encompassed various forms of inositol, including Myo-Inositol-1-Phosphate Synthase, myo-inositol, chiro-inositol, di-chiro-inositol, and pinitol. Comparisons included placebo, no intervention, folic acid, or standard diet. Eligible studies were required to provide both baseline and end-of-trial data for both intervention and control groups or to report changes over the trial period.

To ensure the relevance of the selected articles to the research question, the following exclusion criteria were established. First, studies were disqualified if they involved the administration of any other concurrent intervention alongside inositol, or included a customized weight loss program provided only to the intervention group without a parallel treatment for the control group. Second, articles focusing on pregnant or lactating participants, participants under 18 years of age, individuals with weakened immune systems, or those with conditions such as respiratory, hormonal, infectious diseases, or psychiatric disorders—including, but not limited to, type 1 diabetes mellitus, gestational diabetes, rheumatoid arthritis, multiple sclerosis, and severe renal issues or acute medical events like heart failure and myocardial infarction—were excluded. Third, any study where inositol supplementation lasted less than four weeks was disqualified. Fourth, studies that did not include control condition were ruled out. Additionally, any research that did not meet the minimum data requirements for this review was also excluded. Additional exclusions included case studies, book chapters, observational studies, animal or in vitro research, and purely review articles.

The selection of eligible publications was conducted through a two-step approach. In the first step, PM, FA, and MAD examined the titles, abstracts, and keywords of the records obtained from the searches to pinpoint relevant studies. In the subsequent step, the selected articles were independently reviewed in their entirety by three researchers (PM, FA, and ZGh) to evaluate their eligibility for inclusion in the final analysis. Disagreements concerning eligibility were mediated by a third researcher (MAD) to reach consensus.

Data extraction

Upon identifying eligible trials for the current meta-analysis, data extraction was carried out by PM, with subsequent verification by MAD and ZGh. The extracted data encompassed essential publication details including the name of the first author, year of publication, and geographic location; and study characteristics including design, participant health conditions, biological sex distribution, inositol dosage, duration of intervention, mean participant ages, and sample sizes for both intervention and control groups. Additionally, both baseline and post-study outcome levels of BMI, WC, WHR, blood glucose, HbA1c%, fasting insulin, HOMA-IR, TG, TC, LDL-C, and HDL-C, systolic and diastolic blood pressure (SBP and DBP), were extracted, as well as any changes reported in these parameters. Studies that did not specify the measurement methods were still included, provided they reported baseline and post-intervention values with sufficient statistical detail. For RCTs featuring multiple time points, only the final time point data were analyzed, omitting intermediate values. In the case of the study conducted by Lee et al. [31], which involved two intervention arms comparing varying dosages of inositol against a placebo, each intervention was treated as a distinct study, while the effective sample size for the control group was halved to maintain accurate participant representation [32].

Risk of bias assessment

To assess the risk of bias in the studies included in this meta-analysis, we utilized the Cochrane Risk of Bias Version 1 (ROB 1) tool, which is specifically designed for RCTs [33]. Six key domains that affect the quality of RCTs, including [1], random sequence generation (selection bias) [2], allocation concealment (selection bias) [3], blinding of participants and personnel (performance bias) [4], blinding of outcome assessment (detection bias) [5], incomplete outcome data (attrition bias), and [6] selective outcome reporting (reporting bias) were examined accordingly. Two independent reviewers (PM and ZGh) assessed each study, assigning each domain a risk of bias rating as “Low risk” (Adequate methods reported with minimal concerns for bias), “Unclear risk” (Insufficient information provided to determine the level of bias), and “High risk” (Significant methodological concerns that may impact study validity). Overall study classification was also determined correspondingly. Disagreements between reviewers were resolved through discussion with a third reviewer (MAD). Ultimately, each study was classified as having a high risk of bias (poorly conducted), moderate risk of bias (fairly conducted), or low risk of bias (well-conducted, or good quality), which provided critical insight into the reliability of the trial findings and their suitability for inclusion in the meta-analysis.

Statistical analysis

Meta-analysis was conducted using STATA 16 software (StataCorp LC, Texas, USA). After the extraction of pertinent data, all outcome measures were standardized to uniform units, which included BMI in kg/m², WHR in centimeters, WC also in centimeters, as well as biochemical markers such as insulin (µU/mL), HOMA-IR, glucose levels, and lipid parameters comprising TC, TG, LDL-C, and HDL-C, expressed in milligrams per deciliter (mg/dL). Blood pressure measurements, specifically SBP and DBP, were expressed in millimeters of mercury (mmHg). In instances where adequate data regarding the mean differences between pre-intervention and post-intervention levels were not available, the corresponding standard deviations (SDs) for each of the group means were computed using the specified formula, as follows:

SD _change = square root (SD _baseline²+SD _final²-(2× r × SD _baseline × SD _final)), assuming a correlation coefficient (r) = 0.8; [32].

Mean differences between groups were derived by taking the mean change in the targeted variable noted in the intervention group and subtracting it from the mean change reported in the control group. This difference was standardized by dividing it by the average of the SD.

Heterogeneity

To assess heterogeneity in our analysis, we employed the Cochrane Chi-squared test in conjunction with Galbraith plots, interpreting the I² statistic as follows: values between 0% and 40% indicate low or negligible heterogeneity, 30% to 60% suggest moderate heterogeneity, 50% to 90% reflect substantial heterogeneity, and 75% to 100% signify considerable heterogeneity. We considered an I² statistic of 50% or higher, along with a P value of less than 0.10, as indicative of significant heterogeneity among the studies. In light of the observed heterogeneity and the diverse characteristics of the included studies—such as variations in age, biological sex, and health conditions—we utilized a random-effects model for the meta-analysis. This approach enabled us to calculate mean effect sizes, reported as weighted mean differences (WMD) with their corresponding 95% confidence intervals (95% CIs) [34].

Subgroup analyses

This study considered four main subgroup analyses: (1) participant health conditions (with obesity or overweight, metabolic syndrome (MetS), type 2 diabetes (T2DM), non-alcoholic fatty liver disease (NAFLD), PCOS, and healthy participants (those without a diagnosed metabolic disorder); (2) intervention durations (≤ 8 weeks vs. >8 weeks); (3) inositol daily dosage (low doses : <4 g daily vs. high doses: ≥ 4 g daily); and (4) the risk of bias according to the ROB1 scoring (classified as either low/moderate or high risk of bias).

Meta-regression

To explore the relationships between age as a continuous variable and its potential moderating effects on various measured outcomes, random-effects meta-regressions were executed. These analyses assessed estimated net changes in parameters such as BMI, and metabolic markers including glucose, insulin, HOMA-IR, TC, TG, LDL-C, HDL-C, SBP and DBP levels according to age. The unrestricted maximum likelihood method was employed for these calculations.

Publication bias

The potential for publication bias was evaluated through the visual examination of funnel plots for asymmetry, along with conducting Egger’s weighted regression tests. A p-value of less than 0.05 was set as the cut-off for statistical significance for the Egger’s test, indicating the presence of publication bias. Additionally, we visually examined funnel plots for asymmetry, which can also suggest potential bias if smaller studies with negative outcomes are missing from the analysis. Upon confirming publication bias via Egger’s tests, we employed a non-parametric random-effects trim-and-fill approach, specifically the Duval and Tweedie “trim and fill” method, to adjust for the identified biases in the dataset [35] (Supplementary file 3; Supplementary Figs. 1 to 7).

Sensitivity/influence analysis

To ascertain the effect of each study on the overall effect size, we conducted a leave-one-out sensitivity analysis. Moreover, we performed a separate sensitivity analysis to investigate the effects of risk of bias on the results by excluding those studies categorized with a high risk of bias (available in Supplementary file 4; Supplementary Figs. 8 to 20).

Certainty of evidence

To enhance transparency, the strength of the evidence was thoroughly assessed using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) framework [36], as outlined in Supplementary file 7, Table 3. This tool systematically assesses risk of bias, inconsistency, indirectness, imprecision, and publication bias. RCTs are primarily considered to offer a high level of evidence; however, this classification may be downgraded in light of several significant considerations. These include the risk of bias, inconsistency—marked by notable heterogeneity that lacks reasonable justifications (indicated by I² values greater than 50% alongside p-values below 0.1)—indirectness, which pertains to factors that might limit the applicability of conclusions to wider populations, and imprecision represented by broad 95% confidence intervals for effect sizes or studies with fewer than 400 participants. The potential for publication bias is also accounted for. The GRADE methodology ultimately allows for the overall certainty of the evidence to be rated as high, moderate, low, or very low.

Table 3.

Quality assessment of the studies using the Cochrane risk of bias tool for randomized controlled trials

	Quality	Random sequence generation	Allocation concealment	Selective reporting bias	Blinding of participants	Incomplete outcome data	Outcome assessor blinding	Other bias
Nestler, et al., 1999	Moderate risk of bias	U	L	U	L	L	L	L
Davis, et al., 2000	Moderate risk of bias	U	L	U	L	L	L	L
Iuorno, et al., 2002	Moderate risk of bias	U	L	U	L	L	L	L
Campbell, et al., 2004	High risk of bias	U	U	U	L	L	L	H
Kim, et al., 2005	High risk of bias	U	U	U	L	L	L	L
Genazzani, et al., 2008	High risk of bias	U	U	U	U	L	U	L
Costantino, et al., 2009	High risk of bias	U	U	U	L	L	L	L
Dona, et al., 2012	High risk of bias	L	U	U	U	L	U	L
Kim, et al., 2012	High risk of bias	U	U	U	L	L	L	H
Santamaria, et al., 2012	High risk of bias	L	U	U	U	L	U	L
Artini, et al., 2013	high risk of bias	L	L	U	U	L	U	H
Capasso, et al., 2013	High risk of bias	U	U	U	U	L	U	H
D’Anna, et al., 2014	High risk of bias	L	U	U	H	L	U	H
Banuls, et al., 2015	High risk of bias	U	U	U	L	L	L	L
Cianci, et al., 2015	High risk of bias	L	U	U	H	L	U	L
Benelli, et al., 2016	High risk of bias	U	U	U	U	L	U	L
Lee, et al., 2019	Low risk of bias	L	L	L	L	L	L	L
Tutunchi, et al., 2023	Low risk of bias	L	L	L	L	L	L	L

Results

After the preliminary evaluation of articles by screening titles and/or abstracts, a total of 41 articles were identified as potentially relevant RCTs from approximately 3,046 records obtained in the initial literature search (Fig. 1). Following a comprehensive review of these selected articles, 18 studies were determined to meet the eligibility criteria for inclusion in the final meta-analysis. The search process for this meta-analysis is illustrated in Fig. 1.

Fig. 1.

Flow diagram of the meta-analysis study selection procedure

Characteristics of included studies

Table 2 presents the participant and study characteristics of the studies included in this systematic review and meta-analysis. This systematic review and meta-analysis included 18 RCTs, involving a total of 889 participants, with 428 in intervention groups and 421 in control groups. The studies, published between 1999 and 2023, predominantly used a double-blind RCT design. All included RCTs were parallel-arm studies. Of these, nine studies were conducted in Italy ([37–45], three in Korea [31, 46, 47], and two in the United States [48, 49]. The remaining studies were from various countries, with two trials in Venezuela [50, 51], Spain [52], and Iran [53]. Participants in the intervention groups had an age range of 23 to 66 years, with a pooled mean age of 43 years, while control group participants were similarly aged (23.5–65 years, mean 43 years). The interventions primarily consisted of inositol supplements, administered in various forms such as capsules, powders, or drinks, with dosages ranging from 600 mg/day to 4 g/day. The most common formulations included myo-inositol (2–4 g/day), D-chiro-inositol (600–1200 mg/day), and pinitol (300–1200 mg/day), sometimes combined with folic acid (200–400 mcg/day) or alpha-lipoic acid. The study by Lee et al. [31], included two intervention groups—one receiving a high dose and the other a low dose inositol supplement—that were compared to a placebo; each intervention group was treated as an individual study in the meta-analysis. Control participants received a placebo, dietary counseling, or no intervention, depending on the study design. In studies where dietary counseling was provided, it was either given to both the intervention and control groups or only to the control group as a comparison condition. The duration of supplementation varied from 4 weeks to 24 weeks, with the majority of trials (11 studies) lasting 12 weeks. Importantly, no significant adverse effects were reported in association with inositol supplementation.

The included RCTs examined a range of health conditions, with eight studies on PCOS, four on overweight/obesity or metabolic syndrome, two on T2DM, two on NAFLD, and two studies involving those without diagnosed metabolic disorders (healthy individuals).

Quality assessment

Among the 18 RCTs, only two had a low overall risk of bias according to ROB 1 scoring. Three trials were categorized as moderate risk of bias, while the remaining 13 were classified as having high risk of bias. Most of the studies did not provide sufficient details of the techniques employed for random sequence generation and allocation concealment. The specifics of quality evaluation for each, based on the Cochrane Risk of Bias Tool for RCTs, are shown in Table 3.

Meta-analyses

Body mass index

Analysis using random-effects models of 10 RCTs revealed a significantly larger reductions in BMI among intervention groups inositol supplements as compared with control groups (WMD (95% CI) = −0.57 (−1.10, −0.03) kg/m²; P = 0.037). There was a high level of heterogeneity (I² = 88.6; p < 0.001) and a low evidence certainty according to GRADE scoring (Table 4; Fig. 2) (Supplementary file 7; Table 3). Subgroup analyses indicated that the quality of the RCTs and the duration of interventions were among the identified contributors to this heterogeneity. Notably, the efficacy of inositol supplements varied significantly based on intervention duration, with the most substantial reduction in BMI observed when the supplementation lasted more than eight weeks, albeit still with high heterogeneity (WMD (95% CI) = −0.90 (−1.61, −0.20); P = 0.011; I²=92.2; P heterogeneity < 0.001)) as compared with intervention duration of less than eight weeks. Subgroup analysis indicated no significant differences by health status (Obesity, MetS, NAFLD, vs. PCOS), or dosage of inositol administered (< 4 g/d vs. ≥ 4 g/d) (Table 4 and Supplementary File 5, Supplementary Figs. 21.a to 21.c). Subgroup analysis according to risk of bias showed that the studies that were rated as high risk of bias showed that inositol effectively reduced BMI as compared with controls (WMD (95% CI) = −0.90 (−1.60, −0.20) kg/m²; P-value = 0.011; I² = 92.2; P for heterogeneity < 0.001), whereas, low risk of bias studies did not indicate that inositol effectively reduced BMI when compared to control groups (Table 4 and Supplementary File 5, Supplementary Fig. 21.d). The meta-regression analysis focusing on age as a potential moderator revealed a coefficient of −0.03 (95% CI: −0.09 to 0.02), which did not achieve significance, indicating that age did not moderate the relationship between inositol supplementation and BMI changes (Table 5, Supplementary File 6, and Supplementary Fig. 34).

Table 4.

Overall certainty of evidence (GRADE ratings), and subgroup analysis according to risk of bias, health condition, intervention duration, inositol daily dosage

	N		WMD (95%CI) *	P value	I2 (%)	P heterogeneity	P between	Certainty of evidence (GRADE ratings)¥¥
Body Mass Index (BMI) (kg/m2)								Low
All included trials		10	−0.57 (−1.10, −0.03)	0.037	88.6	< 0.001	-
Health conditions
Obesity, MetS, NAFLD		6	−0.88 (−1.51, −0.25)	0.006	15.5	0.306	0.425
PCOS		3	−0.51 (−1.17, 0.16)	0.135	92.3	< 0.001	-
Intervention duration
≤ 8 weeks		3	0.33 (−0.13, 0.79)	0.162	0.00	0.925	0.004
> 8 weeks		7	−0.90 (−1.61, −0.20)	0.011	92.2	< 0.001
Inositol daily dosage
< 4 g/d		6	−0.72 (−1.41, −0.04)	0.038	83.6	< 0.001	0.447
≥ 4 g/d		4	−0.31 (−1.14, 0.52)	0.468	79.3	0.002	-
Risk of bias
Good/Fair		3	0.33 (−0.13, 0.79)	0.162	0.00	0.925	0.004
Poor		7	−0.90 (−1.60, −0.20)	0.011	92.2	< 0.001	-
Waist-to-Hip Ratio (WHR)								Low
All included trials		5	−0.02 (−0.04, −0.001)	0.035	84.1	< 0.001	-
Health conditions
Obesity, MetS, NAFLD		2	0.001 (−0.01, 0.01)	0.778	0.00	0.453	< 0.001
PCOS		3	−0.03 (−0.04, −0.02)	< 0.001	0.00	1.000	-
Intervention duration
≤ 8 weeks		3	−0.02 (−0.04, −0.01)	0.007	0.00	0.473	0.700
> 8 weeks		2	−0.01 (−0.04, 0.01)	0.314	95.7	< 0.001	-
Inositol daily dosage
< 4 g/d		3	−0.02 (−0.04, 0.01)	0.175	70.3	0.034	0.645
≥ 4 g/d		2	−0.02 (−0.04, −0.01)	0.012	56.9	0.128	-
Risk of bias
Good/Fair		3	−0.02 (−0.04, −0.01)	0.007	0.00	0.473	0.700
Poor		2	−0.01 (−0.04, 0.01)	0.314	95.7	< 0.001	-
Waist circumference (WC) (cm)								Very low
All included trials		5	−2.36 (−4.39, −0.33)	0.023	55.00	0.064
Inositol daily dosage
< 4 g/d		3	−1.64 (−3.84, 0.56)	0.143	52.3	0.148	0.488
≥ 4 g/d		2	−3.22 (−7.11, 0.67)	0.104	64.4	0.060
Glucose levels (mg/dL)								Very low
All included trials		14	−7.25 (−10.98, −3.52)	< 0.001	90.7	< 0.001	-
Health conditions
Obesity, MetS, T2DM, NAFLD		7	−12.66 (−18.55, −6.77)	< 0.001	88.6	<0.001	0.006
PCOS		5	−3.08 (−8.78, 2.62)	0.289	92.0	< 0.001	-
Healthy *		2	−1.99 (−4.96, 0.989)	0.191	0.00	0.982	-
Intervention duration
≤ 8 weeks		5	−3.62 (−7.68, 0.45)	0.081	40.7	0.150	0.086
> 8 weeks		9	−9.15 (−14.01, −4.30)	< 0.001	93.7	0.982	-
Inositol daily dosage
< 4 g/d		10	−8.41 (−14.48, −2.34)	0.007	92.4	< 0.001	0.556
≥ 4 g/d		4	−6.25 (−10.13, −2.37)	0.002	84.8	< 0.001	-
Risk of bias
Good/Fair		4	−4.04 (−9.16, 1.08)	0.122	54.4	0.087	0.205
Poor		10	−8.49 (−13.07, −3.90)	< 0.001	93.0	< 0.001	-
Insulin levels (µU/mL)								Moderates
All included trials		16	−4.74 (−6.16, −3.32)	< 0.001	90.6	< 0.001	-
Health conditions
Obesity, MetS, NAFLD, T2DM		2	−3.70 (−6.18, −1.22)	< 0.001	88.7	<0.001	< 0.001
PCOS		3	−6.85 (−8.74, −4.97)	0.003	86.9	< 0.001	-
Healthy *		2	−0.86 (−2.77, 1.04)	0.374	46.2	0.173	-
Intervention duration
≤ 8 weeks		5	−1.78 (−4.62, 1.07)	0.221	65.2	0.022	0.022
> 8 weeks		11	−5.55 (−7.08, −4.02)	< 0.001	91.6	< 0.001	-
Inositol daily dosage
< 4 g/d		12	−3.96 (−5.35, −2.58)	< 0.001	86.3	< 0.001	0.296
≥ 4 g/d		4	−7.26 (−13.28, −1.24)	0.018	95.9	< 0.001	-
Risk of bias
Good/Fair		4	−3.32 (−7.47, 0.83)	0.117	71.3	0.015	0.433
Poor		12	−5.09 (−6.59, −3.59)	< 0.001	91.4	< 0.001	-
The Homeostatic Model Assessment for Insulin Resistance (HOMA-IR)								Moderates
All included trials		11	−1.21 (−1.58, −0.85)	< 0.001	85.0	< 0.001	-
Health conditions ¥
Obesity, MetS, NAFLD, T2DM		5	−1.66 (−2.32, −0.99)	< 0.001	66.00	0.019	0.157
PCOS		5	−1.13 (−1.42, −0.83)	< 0.001	58.9	0.045
Inositol daily dosage
< 4 g/d		10	−1.33 (−1.63, −1.02)	< 0.001	65.8	0.003	< 0.001
≥ 4 g/d		2	−0.46 (−0.81, −0.11)	0.009	24.7	0.249	-
Total Cholesterol (TC) levels (mg/dL)								Low
All included trials*		11 (12 effect sizes)	−18.26 (−30.75, −5.77)	0.004	96.0	< 0.001	-
Health conditions ¥
Obesity, MetS, T2DM		4	−19.82 (−32.20, −7.45)	0.002	90.5	< 0.001	0.145
PCOS		4	−28.25 (−58.27, 1.77)	0.065	97.9	< 0.001
NAFLD		3	−6.31 (−16.35, 3.72)	0.217	0.00	0.896
Intervention duration
≤ 8 weeks		4	−26.15 (−41.59, −10.7)	0.001	83.3	< 0.001	0.310
> 8 weeks		8	−14.17 (−31.39, 3.05)	0.107	97.1	< 0.001	-
Inositol daily dosage
< 4 g/d		8	−15.61 (−26.83, −4.38)	0.006	88.8	< 0.001	0.542
≥ 4 g/d		4	−23.261 (−45.19, −1.33)	0.038	96.3	< 0.001	-
Risk of bias
Poor		6	−17.04 (−36.60, 2.52)	0.088	97.9	< 0.001	0.796
Good/Fair		6	−20.19 (−33.83, −6.56)	0.004	80.00	< 0.001	-
Triglyceride (TG) levels (mg/dL)								High
All included trials*		12 (13 effect sizes)	−29.80 (−48.16, −11.44)	0.001	96.0	< 0.001
Health conditions ¥
Obesity, MetS, T2DM		5	−28.12 (−48.19, −8.06)	0.006	95.3	< 0.001	0.013
PCOS		4	−58.63 (−105.94, −11.33)	0.015	97.3	< 0.001	-
NAFLD		3	6.09 (−15.28, 27.47)	0.576	0.00	0.945	-
Intervention duration
≤ 8 weeks		4	−44.37 (−75.61, −13.13)	0.005	84.8	< 0.001	0.288
> 8 weeks		9	−23.382 (−46.29, −0.47)	0.045	97.1	< 0.001	-
Inositol daily dosage
< 4 g/d		9	−23.37 (−35.66, −11.09)	< 0.001	82.6	< 0.001	0.307
≥ 4 g/d		4	−41.01 (−72.54, −9.49)	0.011	96.00	< 0.001	-
Risk of bias
Poor		7	−29.45 (−54.42, −4.48)	0.021	97.8	< 0.001	0.956
Good/Fair		6	−30.50 (−57.94, −3.06)	0.029	81.3	< 0.001
LDL-C levels (mg/dL)								Moderates
All included trials*		6 (7 effect sizes)	−5.15 (−8.89, −1.42)	0.007	0.00	0.630	-
Health conditions
Obesity, MetS, T2DM, PCOS		3	−3.90 (−9.51, 1.71)	0.173	29.0	0.245	0.275
NAFLD		3	−9.74 (−18.58, −0.89)	0.031	0.00	0.907	-
Intervention duration
≤ 8 weeks		3	−3.55 (−10.142, 3.033)	0.290	26.3	0.257	0.351
> 8 weeks		4	−7.64 (−13.16, −2.13)	0.007	0.00	0.979	-
Inositol daily dosage
< 4 g/d		5	−4.18 (−8.42, 0.06)	0.035	0.00	0.559	0.344
≥ 4 g/d		2	−8.49 (−16.35, −0.63)	0.034	0.00	0.497	-
Risk of bias
Poor		2	−7.73 (−13.67, −1.79)	0.011	0.00	0.678	0.275
Good/Fair		5	−3.48 (−8.27, 1.32)	0.155	0.00	0.560	-
HDL-C levels (mg/dL)								Moderates
All included trials*		10 (11 effect sizes)	2.76 (1.16, 4.36)	0.001	52.9	0.020	-
Health conditions ¥
Obesity, MetS, T2DM		5	4.08 (2.33, 5.82)	< 0.001	47.6	0.106	0.206
PCOS		2	2.36 (−3.46, 8.19)	0.426	59.5	0.116	-
NAFLD		3	1.09 (−1.77, 3.94)	0.456	0.00	0.388	-
Intervention duration
≤ 8 weeks		3	0.96 (−1.212, 3.128)	0.387	0.00	0.540	0.064
> 8 weeks		8	3.58 (1.84, 5.32)	< 0.001	45.8	0.074	-
Inositol daily dosage
< 4 g/d		8	2.38 (0.47, 4.29)	0.014	39.2	0.118	0.581
≥ 4 g/d		3	3.38 (0.40, 6.36)	0.026	67.2	0.047	-
Risk of bias
Poor		6	4.32 (2.87, 5.76)	< 0.001	28.1	0.224	0.002
Good/Fair		5	0.52 (−1.46, 2.50)	0.606	0.00	0.702	-
SBP levels								Low
All included trials		7	−5.34 (−6.91, −3.78)	< 0.001	38.0	0.139	-
Health conditions ¥
Obesity, MetS, T2DM		4	−7.64 (−11.38, −3.91)	< 0.001	4.4	0.306	0.240
PCOS		2	−5.146 (−6.98, −3.31)	< 0.001	46.9	0.130	-
Intervention duration
≤ 8 weeks		2	−3.75 (−8.63, 1.12)	0.131	73.9	0.050	0.391
> 8 weeks		5	−5.94 (−6.99, −4.88)	< 0.001	7.2	0.366	-
Inositol daily dosage
< 4 g/d		4	−4.73 (−8.13, −1.341)	0.006	49.9	0.112	0.543
≥ 4 g/d		3	−5.87 (−7.29, −4.46)	< 0.001	20.00	0.287	-
Risk of bias €
DBP levels								Low
All included trials		7	−6.12 (−8.44, −3.80)	< 0.001	69.7	0.003	-
Health conditions ¥
Obesity, MetS, T2DM		2	−8.18 (−13.98, −2.38)	0.006	64.2	0.095	0.596
PCOS		4	−6.31 (−10.06, −2.56)	0.001	76.8	0.005	-
Intervention duration
≤ 8 weeks		2	−6.10 (−8.21, −4.00)	< 0.001	0.00	0.552	0.996
> 8 weeks		5	−6.11 (−9.32, −2.91)	< 0.001	79.4	0.001	-
Inositol daily dosage
< 4 g/d		4	−6.07 (−9.98, −2.15)	0.002	65.1	0.035	0.920
≥ 4 g/d		3	−6.34 (−9.92, −2.76)	0.001	81.1	0.005	-

Risk of bias ^€

NAFLD: Non-alcoholic fatty live disease; MetS: Metabolic syndrome; T2DM: Type 2 diabetes mellitus; PCOS: Polycystic ovary syndrome. SBP: Systolic blood pressure; DBP: Diastolic blood pressure. WMD: Weighted mean differences; CI; Confidence Interval

*Healthy populations are generally defined as individuals without diagnosed metabolic disorders

^¥Since only one trial on healthy subjects was available for this subgroup analysis, it was excluded

^€Not applicable as none of the included RCTs were ranked as low risk of bias

**Some subgroup analyses were not applicable for some outcomes due to a limited number of available studies

^¥¥Described in details in Supplementary File 7 Table 3

Fig. 2.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on Body Mass index (BMI)

Table 5.

Meta-regression analysis for age as a potential moderator^*

	Coefficient (95%CI)	P value	Residual heterogeneity: I2 (%)	R-squared (%)
Inositol and BMI
Age	−0.031 (−0.09, 0.02)	0.206	88.84	15.17
Inositol and Glucose
Age	−0.30 (−0.68, 0.08)	0.106	88.90	28.24
Inositol and Insulin
Age	0.15 (−0.04, 0.33)	0.115	87.25	12.25
Inositol and HOMA-IR
Age	−0.02 (−0.05, 0.002)	0.071	51.9	73.82
Inositol and TC
Age	0.76 (−0.17, 1.69)	0.096	84.63	30.44
Inositol and TG
Age	1.59 (−0.20, 3.38)	0.075	92.09	30.86
Inositol and LDL
Age	−0.11 (−0.78, 0.56)	0.641	0.00	-
Inositol and HDL
Age	0.16 (0.001, 0.31)	0.048	18.95	77.67
Inositol and SBP
Age	−0.09 (−0.32, 0.13)	0.298	37.72	-
Inositol and DBP
Age	−0.08 (−0.39, 0.22)	0.481	74.07	−29.69

Waist-to-hip ratio (WHR)

Random-effects meta-analysis involving five RCTs with 352 participants indicated that inositol supplementation reduced WHR significantly more than control groups (WMD (% CI) = −0.02 (−0.04, −0.001); P = 0.035). There was a high level of heterogeneity (I² = 84.1; P < 0.001) and a low evidence certainty according to GRADE scoring (Table 4; Fig. 3) (Supplementary file 7; Table 3).

Fig. 3.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on waist-to-hip ratio (WHR)

Subgroup analysis according to health condition showed that only patients with PCOS had significant decreases in WHR following inositol supplementation as compared with control groups demonstrating no significant heterogeneity (WMD (95% CI) = −0.03 (−0.04, −0.02) cm; P < 0.001; I² = 0.00; P for heterogeneity = 1.000). The intervention groups with obesity, MetS, or NAFLD, inositol was not significantly different from controls (Table 4 and Supplementary File 5, Supplementary Figs. 22.a to 22.c). Subgroup analysis according to risk of bias demonstrated that the reduction in WHR following inositol supplementation remained significant and exhibited no significant heterogeneity after the exclusion of studies deemed to have a high risk of bias (WMD (95% CI) = −0.02 (−0.04, −0.01) cm; P = 0.007; I² = 0.00; P for heterogeneity = 0.473) (Table 4 and Supplementary File 5, Supplementary Fig. 22.d).

Waist circumference (WC)

Five RCTs comprising 352 participants, showed that inositol supplementation resulted in a significantly larger decrease in WC when compared with control groups (WMD (95% CI) = −2.36 (−4.39, −0.33) cm; P-value = 0.023). There was high heterogeneity and very low certainty of evidence (I² = 55.0; P = 0.064) (Table 4; Fig. 4) (Supplementary file 7; Table 3). Due to the low number of available trials, only the subgroup analysis based on inositol daily dosage was performed, which indicated no significant differences when dosages were separated (< 4 g/d vs. ≥ 4 g/d). The quality of the studies for WC were all rated as high risk of bias based on the risk of bias assessments, therefore a subgroup analysis based on study quality was not possible (Table 4 and Supplementary File 5, Supplementary Figs. 23.a).

Fig. 4.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on waist circumference (WC)

Blood glucose

Fourteen RCTs with a total of 683 participants assessed the effects of inositol on glucose levels. Meta-analysis showed significantly larger decrease in glucose levels among inositol groups when compared with control groups (WMD (95% CI) = −7.25 (−10.98, −3.52) mg/dL; P-value < 0.001). There was high heterogeneity among studies and very low certainty of evidence; (I² = 90.7; P < 0.001) (Table 4; Fig. 5) (Supplementary file 7; Table 3). Subgroup analysis indicated that participants’ health conditions may contribute to the high heterogeneity. Specifically, studies involving individuals with obesity, MetS, NAFLD, or T2DM demonstrated significant reductions in glucose levels following inositol supplementation compared to control groups (WMD (95% CI) = −12.66 (−18.55, −6.77) mg/dL; P-value < 0.001). There was high heterogeneity among studies (I² = 88.6; P < 0.001). Subgroup analysis showed no significant differences in glucose reductions for trials of individuals with PCOS or healthy controls (those without a diagnosed metabolic disorder). In addition, subgroup analysis based on risk of bias showed that the significantly larger reductions in glucose persisted only when combining results from trials identified as high risk of bias (WMD (95% CI) = −8.49 (−13.07, −3.90); P < 0.001; I² = 93.0; P for heterogeneity < 0.001). Trials with low risk of bias, when considered alone, did not reduce glucose levels significantly more than controls (Table 4 and Supplementary File 5, Supplementary Figs. 24.a to 24.d). Meta-regression analyzing the moderating effects of age on the association between inositol supplementation and glucose levels, with a coefficient of −0.30 (95% CI: (−0.68, 0.08)), was not statistically significant (Table 5) (Supplementary File 6; Supplementary Fig. 35).

Fig. 5.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on Glucose levels

HbA1c%

There were three studies with a total of 118 participants that included HbA1c levels. There were no statistically significant differences between inositol supplementation and controls on HbA1c levels (WMD (95%CI) = −0.48 (−1.38, 0.42) P-value = 0.292). There was high heterogeneity and very low certainty of evidence (I² = 94.2; P < 0.001) (Fig. 6).

Fig. 6.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on HbA1c levels

Fasting insulin

A total of 16 trials comprising 749 participants showed that inositol supplementation decreased serum insulin concentrations significantly more than controls (WMD (95% CI) = −4.74 µU/mL (−6.16, −3.32); P < 0.001). There was high heterogeneity among included studies, and a moderate certainty of evidence (I² = 90.6; P < 0.001) (Table 4; Fig. 7) (Supplementary file 7; Table 3).

Fig. 7.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on Insulin levels

Subgroup analyses based on health conditions and follow-up durations, both contributed to the high heterogeneity observed among the studies. Those diagnosed with PCOS exhibited the most marked decrease in insulin levels (WMD (95% CI) = −6.85 µU/mL (−8.74, −4.97); P = 0.003; I² = 86.9; P for heterogeneity < 0.001) when compared with controls, followed by individuals with obesity, MetS, T2DM, or NAFLD (WMD (95% CI) = −3.70 µU/mL (−6.18, −1.22); P < 0.001; I² = 88.7; P for heterogeneity < 0.001). However, trials involving healthy participants (those without a diagnosed metabolic disorder) did not show greater effects for inositol supplementation versus controls for insulin. Inositol supplementation lasting more than 8 weeks reduced insulin significantly more than controls (WMD (95% CI) = −5.55 µU/mL (−7.08, −4.02); P < 0.001; I² = 91.6; P for heterogeneity < 0.001); whereas. studies with intervention durations of 8 weeks or less did not show significantly larger effects for inositol supplementation (Table 4 and Supplementary file 5, Supplementary Figs. 25.a to 25.d). For the subgroup analysis of low-risk vs. high-risk studies, low-risk studies did not indicate significantly larger effects for inositol supplementation as compared with controls (WMD (95% CI) = −3.32 µU/mL (−7.47, 0.83); P = 0.117; I² = 71.3; P for heterogeneity < 0.015). A meta-regression examining the effects of average age of participants on insulin changes did not reveal significant results, indicating that age was not a moderator of the association between inositol supplementation and insulin changes. The analysis showed a coefficient of 0.15 (95% CI: −0.04, 0.33) (Table 5) (Supplementary File 6; Supplementary Fig. 36).

Homeostatic model assessment for insulin resistance (HOMA-IR)

Twelve RCTs with 621 participants demonstrated a significantly larger reduction in HOMA-IR following inositol supplementation as compared with controls (WMD (95% CI) = −1.21 (−1.58, −0.85); P < 0.001). There was substantial heterogeneity among studies and moderate certainty of evidence (I² = 85.0; P < 0.001) (Table 4; Fig. 8) (Supplementary file 7; Table 3).

Fig. 8.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on HOMA-IR levels

Subgroup analysis for the dosages of inositol appeared to be the main contributor to the observed heterogeneity. RCTs administering more than 4 g of inositol showed significant reductions in HOMA-IR as compared with controls (WMD (95% CI) = −0.46 (−0.81, −0.11); P = 0.009; I² = 24.7; P for heterogeneity = 0.249), with the largest effects shown in studies providing less than 4 g daily (WMD (95% CI) = −1.33 (−1.63, −1.02); P < 0.001; I² (%) = 65.8; P for heterogeneity < 0.003) (Table 4 and Supplementary File 5, Supplementary Figs. 26.a to 26.b). Nevertheless, the health conditions of the participants (i.e. Obesity, MetS, NAFLD, T2DM, vs. PCOS) have not been found to play a significant role regarding heterogeneity. The subgroup analysis based on risk of bias was not conducted, as nearly all studies, except for one trial including HOMA-IR, were rated as moderate or high risk of bias. Meta-regression analyzing age in relation to changes in HOMA-IR, with a coefficient of −0.02 (95% CI: −0.05, 0.002), did not yield significant results (Table 5) (Supplementary File 6; Supplementary Fig. 37).

Total cholesterol (TC) levels

Eleven RCTs (with 12 effect sizes) with a total of 616 participants showed statistically significant differences between inositol supplementation and control groups for TC levels (WMD (95% CI) = −18.26 (−30.75, −5.77) mg/dL; P = 0.004). There was high heterogeneity among included studies (I² = 95.8; P for heterogeneity < 0.001) (Table 4; Fig. 9). The level of certainty regarding this trend was low according to the GRADE scoring (Supplementary file 7; Table 3).

Fig. 9.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on Total Cholesterol (TC) levels

Subgroup analysis based on factors such as the participants’ health conditions, (Obesity, MetS, T2DM, vs. PCOS, vs. NAFLD), intervention duration (≤ 8 weeks vs. >8 weeks), and inositol daily dosage (< 4 g/d vs. ≥ 4 g/d) were not significant contributors to heterogeneity, and inositol efficacy was comparable across the subgroups (Table 4 and Supplementary File 5, Supplementary Figs. 27.a to 27.d). Subgroup analysis based on risk of bias indicated that these significant reductions in TC levels persisted only when combining results from trials identified as being at a low risk of bias (WMD (95% CI) = −20.19 (−33.83, −6.56) mg/dL; P-value = 0.004; I² = 80.0; P for heterogeneity < 0.001). Meta-regression analysis focusing on age, with a coefficient of 0.76 (95% CI: −0.17, 1.69), did not indicate significant moderating effects for age (Table 5) (Supplementary File 6; Supplementary Fig. 38).

Triglyceride (TG) levels

Twelve RCTs (with 13 effect sizes) involving 622 participants, revealed a significant reduction in serum/plasma TG levels after inositol supplementation (WMD (95% CI) = −29.80 (−48.16, −11.44) mg/dL; P = 0.001). There was high heterogeneity among included studies (I² (%) =96.0; P for heterogeneity < 0.001) (Table 4; Fig. 10). The level of certainty regarding this trend was high according to the GRADE scoring (Supplementary file 7; Table 3). According to the subgroup analysis, the health condition of the studied participants was a significant factor contributing to the heterogeneity. The RCTs of patients with PCOS exhibited the largest effects on TG levels with high heterogeneity (WMD (95% CI) = −58.63 (−105.94, −11.33) mg/dL; P = 0.015; I² = 97.3; P heterogeneity < 0.001). Other subgroups, including those with obesity, MetS, and T2DM (WMD (95% CI) = −28.12 (−48.19, −8.06) mg/dL; P = 0.006; I² = 95.3; P heterogeneity < 0.001). No significant differences between inositol supplementation and controls were noted in studies involving patients with NAFLD (Table 4 and Supplementary File 5, Supplementary Figs. 28.a to 28.d). In addition, significantly larger decreases in TG levels were also evident in the lower risk of bias RCTs (WMD (95% CI) = −30.50 (−57.94, −3.06) mg/dL; P = 0.029). There was high heterogeneity among studies (I² = 81.3; P < 0.001). Additionally, the meta-regression analysis regarding age did not yield significant results. The analysis reports a coefficient of 1.59 (95% CI: −0.20, 3.38) (Table 5) (Supplementary File 6; Supplementary Fig. 39).

Fig. 10.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on total Triglyceride (TG) levels

Low-density lipoprotein cholesterol (LDL-C) levels

Six RCTs (with 7 effect sizes), involving 273 participants, showed that inositol supplementation reduced LDL-C significantly more than controls (WMD (95% CI) = −5.15 (−8.89, −1.42) mg/dL; P = 0.007). There was very low heterogeneity among studies and a moderate certainty of evidence (I² = 0.0; P = 0.630) (Table 4; Fig. 11) (Supplementary file 7; Table 3). The effects of inositol supplementation were consistent across various subgroups, including the health conditions of the studied participants (Obesity, MetS, T2DM, PCOS vs. NAFLD), intervention duration (≤ 8 weeks vs. >8 weeks), inositol daily dosage (< 4 g/d vs. ≥4 g/d), as well as risk of bias assessment. The significant reduction in LDL-C levels following inositol supplementation remained significant when only taking into account the studies rated as high risk of bias (WMD (95% CI) = −7.73 (−13.67, −1.79) mg/dL; P = 0.011). There was very low heterogeneity among studies (I² = 0.0; P heterogeneity = 0.678). However, this effect was not evidence after excluding studies with high risk of bias (Table 4 and Supplementary File 5, Supplementary Figs. 29.a to 28.d). The meta-regression analysis with respect to age did not show significant moderating effects for age, with a coefficient of −0.11 (95% CI: −0.78, 0.56) (Supplementary File 6; Supplementary Fig. 40).

Fig. 11.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on LDL-C levels

High-density lipoprotein cholesterol (HDL-C) levels

Ten RCTs (with 11 effect sizes) and a total of 604 participants showed that inositol supplementation increased HDL-C levels significantly more than controls (WMD (95% CI) = 2.76 (1.16, 4.36) mg/dL; P = 0.001). There was high heterogeneity among studies and moderate certainty of evidence; (I² = 52.9; P = 0.020) (Table 4; Fig. 12) (Supplementary file 7; Table 3). According to the subgroup analysis, the effects of inositol supplementation were consistent across various subgroups, including the health conditions of the studied participants (Obesity, MetS, T2DM, vs. PCOS vs. NAFLD), intervention durations (≤ 8 weeks vs. >8 weeks), and inositol daily dosages (< 4 g/d vs. ≥4 g/d). The subgroup analysis according to studies risk of bias assessment demonstrated that the increases in HDL-C levels remained significantly larger as compared with controls, in the subgroup of trials rated as high risk of bias, with low heterogeneity (WMD (95% CI) = 4.32 (2.87, 5.76) mg/dL; P < 0.001; I² = 28.1; P heterogeneity = 0.224). When focusing solely on studies with low risk of bias (rated as good quality), there were no statistically significant differences between inositol supplementation and controls (Table 4 and Supplementary File 5, Supplementary Figs. 30.a to 30.d). Additionally, the meta-regression analysis showed a weak but statistically significant positive correlation between changes in HDL-C and the average age of participants (Coefficient (95% CI) = 0.16 (0.001, 0.31) (Supplementary File 6; Supplementary Fig. 41). With increasing age, the beneficial effects of inositol supplementation on HDL-C levels were larger. However, it should be noted that only 8 out of 11 trials provided information on the mean ages of the studied participants.

Fig. 12.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on HDL-C levels

Systolic and diastolic blood pressure (SBP and DBP)

Seven RCTs with a combined total of 282 participants showed that inositol supplementation led to significantly larger reductions in SBP as compared with controls (WMD (95% CI) = −5.34 (−6.91, −3.78) mmHg; P < 0.001). There was low heterogeneity among studies (I² = 38.0; P = 0.139), with low certainty of evidence (Table 4; Fig. 13) (Supplementary file 7; Table 3). Subgroup analyses by health conditions of the studied participants (Obesity, MetS, T2DM, vs. PCOS), intervention duration (≤ 8 weeks vs. >8 weeks), and inositol daily dosage (< 4 g/d vs. ≥4 g/d) showed consistent results on SBP for inositol supplementation as compared with controls (Table 4 and Supplementary File 5, Supplementary Figs. 31.a to 31.c).

Fig. 13.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on SBP levels

In addition, inositol lowered DBP significantly more than control groups (WMD (95% CI) = −6.12 (−8.44, −3.80) mmHg; P < 0.001). There was high heterogeneity among studies and very low certainty of evidence (I² = 69.7; P = 0.003) (Table 4; Fig. 14) (Supplementary file 7; Table 3). Subgroup analyses indicated no significant between group differences for health conditions of the studied participants (Obesity, MetS, T2DM, vs. PCOS), intervention duration (≤ 8 weeks vs. >8 weeks), and inositol daily dosage (< 4 g/d vs. ≥4 g/d), and therefore did not explain the heterogeneity for DBP (Table 4 and Supplementary File 5, Supplementary Figs. 32.a to 32.c). Furthermore, the meta-regression examining age did not show significant moderation by age for SBP (coefficient = −0.09 (95% CI: −0.32, 0.13)), or DBP (coefficient = −0.08 (95% CI: −0.39, 0.22)) (Supplementary File 6; Supplementary Fig. 42–43).

Fig. 14.

Forest plot depicting weighted mean differences (WMD) and 95% confidence intervals (CI) for the effects of Inositol supplementation on DBP levels

Influence/sensitivity analysis

The overall WMDs for all outcomes (including BMI, WHR, WC, glucose, HbA1c%, insulin, HOMA-IR, TC, TG, LDL-C, HDL-C, SBP, and DBP) did not change in significance or direction when each individual study was systematically excluded from the analysis (Supplementary File 4, Supplementary Figs. 8–20).

Publication bias

The results from Egger’s linear regression tests revealed no evidence of publication bias for inositol supplementation outcomes, including BMI, glucose, insulin, TG, TC, and HOMA-IR, while evidence of likely publication bias was indicated for HDL-C (P = 0.028). Furthermore, an assessment of the funnel plots—which illustrate study precision (represented as the inverse of the standard error of the mean, SEM) against effect size (average changes)—showed probable asymmetry for three outcomes: BMI, HOMA-IR, and TC (Supplementary File 3, Supplementary Figs. 1–7). A nonparametric trim-and-fill analysis for HDL-C involved 14 studies, with 11 observed and 3 imputed, resulting in an adjusted pooled effect size of 3.489 and a 95% CI of (1.97, 5.00). For BMI, no additional studies were imputed, confirming the absence of bias. In the case of HOMA-IR, the imputation of two studies yielded a WMD of −1.10, with a 95% CI of (−1.51, −0.69). Similar findings were observed for TC, where the imputation of two studies resulted in a WMD of −1.10, with a 95% CI of (−21.77, −12.07), indicating a significant negative effect.

Discussion

In the current systematic review and meta-analysis, we investigated the efficacy of inositol supplementation, including its various stereoisomers such as myo-inositol and D-chiro-inositol, on anthropometric measures and established cardiometabolic risk factors (blood glucose, insulin, HOMA-IR, lipid profile, and blood pressure). Results suggest that inositol supplements may be beneficial for improving various metabolic outcomes. Notably, supplementation with inositol was inversely affected BMI, WC, and WHR as compared with controls. However the GRADE analyses indicated low or very low certainty, and the effects were clinically small. Additionally, glucose, insulin, and HOMA-IR, decreased statistically significantly more following inositol consumption when compared with controls, with very low certainty for glucose and moderate certainty for both insulin and HOMA-IR. These effects indicate the effectiveness of inositol supplementation for improving insulin sensitivity. Furthermore, inositol consumption decreased TG and LDL-C levels, and increased HDL-C, statistically significantly more than controls, highlighting the effectiveness of inositol for ameliorating dyslipidemia. Grade scoring indicated this evidence had a moderate to high level of certainty. Although inositol supplementation also resulted in statistically significantly lower TC concentrations, the level of certainty for this effect was low. Blood pressure improvements were also observed with inositol supplementation as compared with controls, though this evidence had low or very low certainty. Given the high heterogeneity in these results, the effects of inositol appeared to differ by health condition, with the most pronounced metabolic benefits observed in individuals with underlying health conditions. Unhealthy groups, such as those with obesity, T2DM, MetS, or PCOS generally experienced more statistically significant improvements in glycemic control, insulin levels, and TG compared to healthy participants. However, these effects were not consistent across all conditions. In women with PCOS, inositol supplementation appeared particularly effective for improving insulin sensitivity and body fat distribution, as indicated by slightly higher reductions in fasting insulin and WHR levels compared to other groups. Notably, unlike other patient groups, while NAFLD patients showed no meaningful improvement in TG levels, LDL-C reductions were statistically significant in this subgroup. These variations suggest that inositol’s therapeutic potential may depend strongly on the specific metabolic context of each patient. Additionally, subgroup analyses indicated that longer intervention durations (more than 8 weeks versus shorter periods) were associated with greater effects on BMI and insulin levels. Inositol dosage also seemed to be the main contributor to the observed heterogeneity in insulin resistance status following inositol supplementation, as RCTs administering less than 4 g of the supplement showed statistically significantly larger reductions in HOMA-IR as compared with controls. Subgroup analyses also revealed a notable discrepancy in study quality. Trials with a high risk of bias showed exaggerated reductions in BMI and greater improvements in HDL-C levels compared to controls. In contrast, studies rated as low risk of bias found no statistically significant differences in effects. This inconsistency likely stems from methodological limitations common in high-risk studies, such as lack of blinding, small sample sizes, and inadequate randomization, which may have overestimated the intervention’s benefits. Meta-regression analyses examined age as a potential moderator but did not identify a statistically significant relationship with changes in BMI, glucose, insulin, HOMA-IR, or lipid levels. Despite these analyses, substantial heterogeneity remained unexplained for some outcomes, indicating the potential influence of unmeasured confounding factors such as dietary intakes, and physical activity levels, differences in baseline metabolic profiles, or variations in study methodologies. Specifically, the form of inositol used (myo-inositol vs. D-chiro-inositol vs. combined formulations) was not consistently reported across studies, which may contribute to differential effects. Besides, variations in control conditions (placebo, standard diet, or no intervention) could have influenced the observed effect sizes. These findings underscore the need for further research to refine our understanding of the variability in inositol’s effects.

The available evidence on inositol supplementation generally shows no publication bias for most outcomes, reinforcing the reliability of these findings. However, as potential bias may exist for HDL-C, HOMA-IR, and TC, there is a need for rigorous methodologies to confirm inositol’s effectiveness for these parameters. Future research should aim to include a broader range of studies to validate these results and draw more comprehensive conclusions regarding inositol supplementation. Besides, optimal dosing strategies, long-term effects, or specific populations, should additionally be explored to fully elucidate inositol’s role in cardiometabolic health. Additionally, it should be noted that the meta-analysis on HbA1c (%) levels appeared underpowered due to a very small number of included studies (n = 3). Consequently, results are highly susceptible to the influence of a single study and may not be generalizable. Additionally, the short duration of the included trials is unlikely to capture clinically meaningful changes in this long-term glycemic marker.

The current findings are supported by a number of available systematic reviews and meta-analyses [25–29]. These meta-analyses [25–29], have explored the effects of inositol supplementation on various cardiometabolic factors, in particular in specific conditions, such as PCOS [25] and metabolic disorders [26]. Consistent with the present findings, promising results have been shown for reducing BMI [25, 28], glucose [25, 27], insulin [25, 27], lipid profiles (TC, TG, and LDL-C) [26], SBP and DBP [29], particularly in people with cardiometabolic dysfunction. However, there are still some inconsistencies that lack sufficient evidence for factors like weight changes, WC, and HDL-C following inositol supplementation [26, 27]. A systematic review and meta-analysis conducted by Miñambres et al. evaluated the effects of inositol supplementation on glucose homeostasis across various clinical conditions. This analysis included 20 RCTs comprising 1,239 participants, mostly suffering from metabolic disorders, in addition to pregnant women. The study showed significant reductions in fasting insulin, HOMA-IR, fasting glucose, and 2-hour glucose following inositol supplementation, aligning with our findings. Specifically, inositol supplementation significantly reduced fasting glucose, 2-hour plasma glucose after an oral glucose tolerance test, fasting insulin, and HOMA-IR. However, the same meta-analysis found no significant differences in BMI, contrary to our findings, or HbA1c%, similar to our results, for inositol supplementation compared with controls [27]. Although mechanistic evidence supports inositol’s role in improving glycemic indices, the meta-analysis did not observe statistically significant effects on HbA1c. This is likely due to the limited number of studies measuring HbA1c—only three—thus reducing statistical power. Additionally, since HbA1c reflects long-term glycemic control over approximately three months, the relatively short durations of two of these studies (Davis et al., 2000, 4 weeks [48]; and Kim, 2012, 12 weeks [46]) may have been insufficient to capture meaningful changes in HbA1c levels.

On the other hand, in line with our results on BMI, a systematic review and meta-analysis by Zarezadeh et al. [28], investigated the effects of inositol supplementation on BMI. This study included 15 RCTs with a total of 891 participants. Of these, 10 studies focused on patients with PCOS, 2 on patients with MetS, and the remaining 3 examined pre-diabetic, diabetic, and healthy individuals. The analysis revealed that inositol supplementation significantly decreased BMI with high heterogeneity was observed among the studies. The findings indicated that inositol’s impact on BMI was particularly pronounced in individuals with PCOS and those classified as overweight or obese, as opposed to the current findings. Specifically, myo-inositol demonstrated a stronger effect on BMI reduction compared to other forms. The study concluded that oral inositol supplementation could be considered an effective adjunct treatment for improving BMI, although further research is warranted to explore its long-term effects and safety profile across diverse populations [28]. Another meta-analysis by Tabrizi, which included a search for relevant studies in various databases up to October 2017, encompassed a total of 14 RCTs conducted on patients with metabolic disorders. The pooled findings revealed that inositol supplementation, in partial agreement with our findings significantly and effectively reduced TG, TC, and LDL-C levels in these patients which aligns closely with our findings. Nonetheless, contrary to the current results, while the meta-analysis indicated potential benefits of inositol on HDL-C levels in patients with PCOS, there was no significant effect of inositol supplementation on HDL-C levels in non-PCOS patients [26]. Similar results have been noted in another systematic review and meta-analysis on patients with PCOS, conducted by Greff et al. [25]. Their results revealed a modest reduction in BMI with inositol supplementation, indicating a small to moderate effect on weight management in patients with PCOS, in contrast to our results. Additionally, inositol supplementation led to a significant decrease in fasting plasma glucose levels and a reduction in AUC-insulin levels, indicating a small to moderate positive effect on insulin sensitivity in patients with PCOS. This finding is consistent with our results regarding insulin. However, it should be noted that, in contrast to our inclusion criteria, this study included trials where the effects of inositol were compared with those of metformin [25].

Inositol can be found in nine stereoisomeric forms, with myo-inositol and D-chiro-inositol being the two most prominent types. Although similar in structure, D-chiro-inositol and myo-inositol differ in the configuration of one hydroxyl group. L-chiro-inositol, muco-inositol, scyllo-inositol, epi-inositol, neo-inositol, allo-inositol, and ci-inositol are among the other identified isomers of this agent [19]. Aging, insulin resistance, and glucose dysmetabolism are known to increase the need for myo-inositol in particular [54]. Additionally, emerging evidence suggests that current dietary practices, particularly high sugar intake, may influence inositol bioavailability and cardiometabolic effects. Given the shared transporter systems between glucose and myo-inositol, hyperglycemia can competitively inhibit cellular inositol uptake, potentially increasing inositol requirements. This interplay could contribute to the observed heterogeneity in our meta-analysis, as individuals with varying dietary glucose levels may exhibit differential responses to inositol supplementation [54, 55].

Several metabolic pathways, may explain how inositol exerts its beneficial effects on cardiometabolic health, including enhanced insulin sensitivity, reduced insulin resistance, improved lipid profiles, regulation of blood pressure, modulation of adipokines, and a decrease in oxidative stress. A key mechanism underlying these improvements appears to be the activation of the AMP-activated protein kinase (AMPK) pathway, a critical regulator of energy balance and glucose metabolism. Recent evidence supports this hypothesis; demonstrated that inositol supplementation activates the AMPK pathway, leading to enhanced insulin sensitivity and increased cellular glucose uptake in human endometrial cells exposed to a PCOS environment [56]. This finding bolsters our discussion of inositol’s role in insulin signaling and provides a mechanistic basis for the observed cardiometabolic benefits. Additionally, D-chiro-inositol has been suggested to enhance the activity of rate-limiting enzymes, particularly pyruvate dehydrogenase phosphatase and glycogen synthase, which are integral to glucose metabolism. This enhancement promotes cellular responsiveness to insulin and mediates its function in insulin-sensitive tissues [19]. Experimental studies further suggest that inositol stereoisomers, particularly myo-inositol, modulate pathways related to glycemic control by inhibiting intestinal glucose absorption and stimulating glucose uptake in skeletal muscle [20]. Besides, animal model research has also shown that the intracerebroventricular administration of a derivative of D-chiro-inositol in mice reduces body weight and decreases food intake, while also stimulating insulin secretion and enhancing peripheral sensitivity to insulin [21, 22]. In this respect, inositol is widely recognized for its role in facilitating insulin function by participating in the insulin signaling cascade as a secondary messenger. This involvement enhances the peripheral uptake of glucose by promoting the translocation of glucose transporter type 4 (GLUT-4) to the cell membrane. Furthermore, it has been proposed that the transport of inositol phosphoglycans within the cell is stimulated following the binding of insulin to its specific receptors. This mechanism further underscores the mediating effects of inositol on insulin action [57].

Inositol supplementation may also mitigate oxidative stress, a hallmark of cardiometabolic disorders, particularly in insulin-resistant states. By reducing oxidative stress, inositol can contribute to improved overall metabolic health and decreased inflammation related to insulin resistance and cardiometabolic conditions [24]. Recent findings from Rostami et al., reinforce this effect, showing that inositol supplementation significantly lowers oxidative stress markers in obese patients with NAFLD, contributing to its broader metabolic benefits [58]. However, in conditions like T2DM, renal reabsorption of myo-inositol may be impaired due to hyperglycemia and insulin resistance, leading to increased urinary excretion. This dysregulation could contribute to microvascular complications, highlighting a complex interplay in inositol metabolism [19].

Adipokines can also play a significant role in metabolic regulation, with adiponectin levels inversely related to visceral adiposity and insulin resistance. Myo-inositol supplementation may improve dyslipidemia by increasing adiponectin and reducing leptin levels [19, 23], further supporting its metabolic benefits.

As the current results suggested, myo-inositol supplementation might lead to statistically significant improvements in blood pressure. In alignment with our results, the systematic review and meta-analysis by Hashemi Tari et al. [29], investigated the effects of inositol supplementation on blood pressure by analyzing data from 7 RCTs published within 1999 to 2012, involving a total of 322 participants, mostly with PCOS or MetS. In partial agreement with the current results, the analysis revealed significant reductions in both SBP and DBP. Specifically, with some similarities with our findings, inositol supplementation resulted in a WMD of −5.69 mmHg for SBP (95%CI −7.35, −4.02), and − 7.12 mmHg for DBP (95%CI −10.18, −4.05), with subgroup analyses indicating more pronounced effects in individuals with MetS and those receiving higher doses (>4 gr) or longer treatment durations (>8 weeks). The study also assessed the risk of bias and heterogeneity among the included trials, finding low heterogeneity across studies for both SBP and DBP outcomes [29]. These benefits may be mediated by improved endothelial function and reduced vascular resistance, potentially linked to inositol’s enhancement of insulin sensitivity and adiponectin levels [19, 23, 24].

Clinical relevance of the current findings

Our meta-analysis demonstrates that inositol supplementation statistically significantly improves multiple anthropometric, and cardiometabolic parameters, though the clinical significance varies across outcomes. The observed reductions in anthropometric measures (−0.57 kg/m² BMI, −2.36 cm WC) demonstrate mixed clinical relevance. While the BMI reduction fell below its established threshold for clinical significance (MCID 0.95 kg/m²), the WC reduction showed potential clinical importance at the point estimate (−2.36 cm, exceeding the 2 cm MCID). However, the 95%CI for WC (−4.39 to −0.33 cm) includes values below the MCID threshold, indicating uncertainty about consistent clinically meaningful effects. These findings suggest that while inositol supplementation produces statistically significant anthropometric changes, its clinical relevance for meaningful of improvement in these measures in the studied populations remains uncertain [59, 60]. Glycemic parameters presented a similar pattern: while the − 7.25 mg/dL glucose reduction didn’t reach clinical significance (MCID ~ 28.8 mg/dL), the substantial decreases in fasting insulin (−4.74 µU/mL) and HOMA-IR (−1.21) surpassed their respective MCIDs (0.83 µU/mL and 0.05) [61]. Lipid profile improvements were clinically meaningful for TC (−18.26 mg/dL), LDL-C (−5.15 mg/dL), and TG (−29.80 mg/dL), all exceeding established MCID thresholds (TC:; LDL-C: 3.87 mg/dL; and TG: 7.97 mg/dL), though the HDL-C increase (2.76 mg/dL) was modest (MCID threshold: 3.87 mg/dL) [61]. Blood pressure reductions (−5.34/−6.12 mmHg for SBP/DBP) also exceeded clinical significance thresholds (2 mmHg) [61].

Regarding the optimal dosage and duration of inositol administration, subgroup analysis showed that while supplementation provided metabolic benefits across different dosages and durations, the most statistically significant improvements were observed with particular patterns. Longer intervention periods (> 8 weeks) were particularly effective for glycemic control and anthropometric outcomes, while higher doses (≥ 4 g/day) showed superior effects on insulin sensitivity and lipid profile modulation. The combination of adequate duration (≥ 8 weeks) with sufficient dosage (~ 4 g/day) emerged as the most effective approach for achieving comprehensive metabolic benefits. Importantly, the data indicate that while some benefits were observable at lower doses or shorter durations, the optimal parameters varied by specific metabolic outcome.

The current findings hold particular clinical significance for conditions marked by cardiometabolic dysfunction, including obesity, MetS, T2DM, and PCOS. These conditions share common pathological features such as hyperglycemia, insulin resistance, dyslipidemia, hypertension, and central obesity - all of which showed improvement with inositol supplementation to some extent in our analysis. Given that these metabolic disturbances are well-established risk factors for cardiovascular disease and stroke, our results suggest inositol may play an important role in mitigating cardiovascular risk through its beneficial effects on multiple metabolic pathways [62–64].

Overall, it is important to recognize that although inositol supplementation may offer modest clinical improvements in various cardiometabolic parameters, its clinical effectiveness is considerably less substantial than that of first-line pharmacological therapies such as metformin and statins. While the observed improvements in surrogate markers are promising, it remains unknown whether inositol supplementation translates into a reduced risk of major cardiovascular events. Furthermore, supplements should not be viewed as a substitute for foundational lifestyle modifications, which are paramount for managing cardiometabolic risk. Consequently, inositol should be regarded as a complementary approach rather than a replacement in managing cardiometabolic risk.

Strengths and limitations

This systematic review and meta-analysis has several important strengths that enhance the validity of its findings. We employed rigorous methodology following PRISMA guidelines, conducting a comprehensive search without restrictions on publication date or geographical location to minimize selection bias. Our analysis examined a broad spectrum of clinically relevant outcomes spanning anthropometric measures and cardiometabolic risk factors, addressing pressing global health priorities that are of particular concern to clinicians and patients alike. Further, the application of the GRADE framework represents another significant strength, marking to our knowledge the first systematic evaluation of evidence certainty regarding inositol’s effects across diverse health conditions. Besides, our exclusive focus on RCTs, considered the gold standard for intervention studies, strengthens the design. However, this strength is tempered by the methodological concerns that emerged regarding the included studies themselves.

Thereby, several limitations must be considered when interpreting these findings. Regarding the methodological concerns, it was noted that a significant proportion of the included RCTs were identified as having a high risk of bias due to inadequate randomization, blinding, and reporting practices. This limitation raises concerns about the reliability of the results and the validity of the conclusions drawn from these studies. While we employed random-effects models and conducted subgroup analyses where possible, a high heterogeneity persists as a notable limitation. The high heterogeneity across the included studies might be particularly related to participant characteristics (including age, baseline metabolic status, and specific health conditions) and intervention protocols (such as different inositol isoforms, dosages, and treatment durations). Beyond the factors analyzed, however, heterogeneity may also stem from unmeasured variables such as participants’ baseline dietary patterns, medications’, and physical activity levels, adherence to the supplementation protocol, ethnic and the specific formulation or isomer of inositol used (e.g., myo-inositol vs. D-chiro-inositol vs. combined formulations), which were not consistently reported across studies. In addition, our evaluation was constrained by missing information on the purity of the inositol supplements used, and the possible interactions with participants’ baseline inositol levels and genetic differences in inositol metabolism. Since inositol deficiency can result from multiple metabolic disturbances, the initial inositol status of participants could substantially affect the observed treatment effect, a variable that should be measured in future studies [55, 65].

Although we assessed publication bias using funnel plots and Egger’s tests, the potential for unpublished studies or selective outcome reporting remains a concern. The analysis was further constrained by the small number of available studies for certain outcomes, limited sample sizes in many trials, and the inclusion of studies using combination therapies that may obscure inositol’s specific effects. Most importantly, our GRADE evaluations revealed that the evidence supporting many outcomes was of low or very low certainty, primarily due to study limitations, imprecise effect estimates, and inconsistent findings across trials. These limitations warrant cautious interpretation of our findings and underscore the need for more rigorous, standardized research in this field.

To our knowledge, this synthesis represents the most comprehensive appraisal to date of inositol’s metabolic effects. Nevertheless, it highlights important gaps that future research should address. We recommend larger, multicenter RCTs with standardized protocols that explicitly compare isoform-specific effects (e.g., myo-inositol versus D-chiro-inositol). Such trials should use consistent dosing regimens and test for at least six months to evaluate long-term sustainability and safety. They should employ uniform control conditions and standardized measurement of anthropometric and outcome variables. Greater geographical representation, with increased inclusion of underrepresented regions (e.g., Africa, and Eastern Europe), would improve the global applicability of findings and enhance generalizability.

Conclusion

This comprehensive systematic review and meta-analysis suggests that inositol supplementation may offer modest benefits for several cardiometabolic parameters, though the strength of evidence varies considerably across different outcomes. We found moderate to high-certainty evidence supporting improvements in markers of insulin resistance (notably HOMA-IR and fasting insulin levels) and lipid profiles (including LDL-C, HDL-C, and TG). These findings point to insulin resistance and dyslipidemia as the most reliable therapeutic targets for inositol intervention. By contrast, the evidence for anthropometric measures (such as BMI, WC, and WHR), fasting glucose, TC, and blood pressure was of low or very low certainty. The most consistent benefits emerged from regimens providing approximately 4 g/day of inositol for at least 8 weeks, though optimal parameters may vary depending on specific treatment goals.

Thereby, inositol may be a beneficial adjunct therapy for cardiometabolic disorders, particularly in high-risk patients with insulin resistance. However, the high risk of bias and low certainty of evidence in existing studies, especially for anthropometric data, necessitate cautious interpretation. The current meta-analysis also reveals significant evidence gaps. To address these, future research should prioritize large-scale, methodologically rigorous trials with standardized protocols, longer follow-up periods (6–12 months or more), and more diverse population representation. In the meantime, clinicians should carefully weigh potential benefits against the limitations of the existing data when considering inositol for managing cardiometabolic conditions.

Author contributions

ZGh: Conceptualization, Project administration, Investigation, Methodology, Data curation, Formal analysis, Writing – original draft, Writing – review & editing. MAD: Conceptualization, Supervision, Project administration, Writing – review & editing. PM: Investigation, Data curation, Writing – review & editing. SR: Interpretation, Writing – review & editing. FA, NA: Investigation, Data curation.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All analyses were based on previously published studies; thus, no ethical approval is required.

Consent for publication

All analyses were based on previously published studies; thus, no consent for publication is required.

Competing interests

The authors declare no competing interests.

Human ethics and consent to participate

Not applicable.

Clinical trial number

Not applicable.

Systematic review and meta-analysis registration

A detailed protocol for this systematic review was registered in the International prospective register of systematic reviews (PROSPERO) database, which can be identified by the registration number CRD42024532838.