The review aims to address the following research question: “What is the efficacy and safety of herbal interventions (cinnamon, curcumin, fenugreek, ginger, moringa, saffron) in improving glycaemic control, lipid profiles, and other metabolic parameters in patients with T2DM?”

The PICO questions were assessed in this systematic review:

Population: Adult patients (≥18 years) diagnosed with T2DM.

Intervention: Any herbal intervention (e.g., cinnamon, curcumin, fenugreek, ginger, moringa, saffron) delivered in any form (e.g., capsules, powders, extracts).

Control: Placebo, routine care, or other active therapies.

Outcomes: Primary outcomes were glycaemic control (e.g., fasting blood glucose, glycated haemoglobin (HbA1c), insulin resistance) and lipid profiles (e.g., total cholesterol, LDL, HDL, triglycerides). Secondary outcomes included safety profiles (e.g., adverse events) and various metabolic markers (e.g., body weight, oxidative stress).

Eligible studies were randomized controlled trials (RCTs), including double-blind, triple-blind, and placebo-controlled designs, published in English. Studies involving participants with type 1 diabetes, gestational diabetes, or non-diabetic populations were excluded. Non-randomized studies, observational designs, case reports, review articles, animal studies, and studies with insufficient methodological clarity or incomplete data were also excluded. Additionally, trials employing combination therapies in which the independent effect of the herbal intervention could not be isolated were excluded.

The strategy of this research was undertaken in multiple search engines and databases (the Cochrane Library, Scopus, PubMed, and Web of Science). The search method included a combination of MeSH terms and phrases linked to T2DM, herbal therapies, and outcomes of interest. Examples of search keywords include: “type 2 diabetes mellitus”, “herbal medicine”, “cinnamon”, “curcumin”, “fenugreek”, “ginger”, “moringa”, “saffron”, “glycaemic control”, and “lipid profile”.

Two independent reviewers examined the titles and abstracts of all identified papers to determine eligibility based on the inclusion and exclusion criteria.

Full-Text Review: The complete texts of potentially eligible papers were retrieved and read separately by two reviewers. Disagreements were addressed by discussion or contact with a third reviewer.

The PRISMA flow diagram was used to describe the study selection process, with the number of studies identified and excluded for any reasons at each stage.

Data Extraction Form: A standardised data extraction form was designed and pilot tested. Data were extracted separately by two reviewers, and differences were handled by consensus.

The risk of bias was assessed using the Cochrane Risk of Bias Tool (RoB-2) for RCTs. The assessed domains were bias originating from the randomisation method, bias owing to variations from desired interventions, bias owing to missing outcome data, bias in measuring the outcome, and bias in the selection of the reported outcome. Judgment: Each domain was rated as “low risk,” “some concerns,” or “high risk” of bias. The total RoB for each study and in summary was presented.

The Critical Appraisal Skills Programme (CASP) tool was used to assess the quality of the 34 included studies. An 11-item checklist was used to assess the quality of these included studies.

The table highlights a range of studied herbal therapies, including cinnamon, curcumin, fenugreek, ginger, moringa, and saffron. Each plant was investigated in several trials, with variations in dosage and duration.

Cinnamon was the most frequently investigated herb, included in studies such as Anderson et al., 2016 (500 mg/day), Lira Neto et al., 2022 (3 g/day), and Zare et al., 2018 (500 mg capsules, twice a day) [20, 21, 25]. The dosages vary from 500 mg to 6 g per day, suggesting a wide range of studied quantities.

Curcumin (the active component of turmeric) was also frequently investigated, with doses ranging from 80 mg/day (e.g., Asadi et al., 2019) to 1500 mg/day (e.g., Hodaei et al., 2019, and Yaikwawong et al., 2024). Some research, such as Panahi et al. (2018), coupled curcumin with piperine (5 mg/day) to increase bioavailability [27-29, 31].

Fenugreek was explored in studies like Geberemeskel et al. (2019) (25 g twice a day) and Najdi et al. (2019) (2 g/day), with varied doses and durations [33, 35]. Ginger was examined in dosages ranging from 1200 mg/day (e.g., Carvalho et al., 2020) to 3000 mg/day (e.g., Shidfar et al., 2015), with most research utilising around 2000 mg/day [39, 42]. Moringa was investigated in several forms, including capsules (e.g., Taweerutchana et al., 2017) and leaf powder (e.g., Sissoko et al., 2020), with doses ranging from 1 g to 8 g/day [44, 45]. Saffron was evaluated in dosages of 15 mg/day (e.g., Moravej Aleali et al., 2019) to 1 g/day (e.g., Azimi et al., 2015), with most research utilising lower quantities [48, 49].

The doses of herbal therapies used varied widely between different studies, reflecting the variation in traditional and experimental applications of these plants. For example, cinnamon dosages ranged from 500 mg/day to 6 g/day, with doses between 500 mg and 3 g/day the most common. Curcumin doses ranged from 80 mg/day to 1500 mg/day, with some studies mixing it with piperine to increase absorption. Fenugreek dosages ranged from 2 g/day to 25 g twice daily.

There was also a large variation between the duration of treatments, from single doses (e.g., Rachid et al., 2022) to 12 months (e.g., Yaikwawong et al., 2024) [23, 29]. Most interventions lasted between 8 weeks and 3 months, suitable for testing short- to medium-term effects on glycaemic management and other metabolic parameters.

The majority of research used placebo-controlled designs, which are critical for assessing the efficacy of the herbal therapies. For instance, placebo groups were included in the designs of Anderson et al. (2016) and Lira Neto et al. (2022) to assess the effects of cinnamon extract and Adab et al. (2019) and Asadi et al. (2019) to examine the effects of curcumin and nano-curcumin, respectively [20,21.26.27]. Some research, such as Geberemeskel et al. (2019), instead used active controls (e.g., metformin) to test the efficacy of herbal remedies against conventional diabetic drugs [33]. A small number of studies, such as Rachid et al. (2022), employed non-placebo controls (control groups without any intervention), which may restrict the capacity to make significant conclusions about the efficacy of the herbal intervention [23].

The overall sample sizes (Table 1) and those for control groups (Table 2) are presented to allow for an assessment of the balance between groups. For instance, Anderson et al. (2016) reported on 64 participants in the cinnamon group and 73 in the placebo group [20]. Lira Neto et al. (2022) had 71 participants in the cinnamon group, and in the placebo group, 69 participants [21]. Most investigations have somewhat balanced group sizes, which is critical for assuring the validity of the statistical comparisons. The characteristics of the study participants, such as their age, gender distribution, baseline blood glucose levels, and final blood glucose levels following intervention, are reported in Table 3.

Different studies included a range of ages, from younger (Ahmad et al. 2017; mean age: 24.1 years) to older individuals (Anderson et al. 2016; mean age: 61.3 years) [20, 43]. The majority of studies used middle-aged to older participants, in line with the demographic typically affected by T2DM.

Sex distribution also varied between studies; some used a more balanced sample (e.g., Anderson et al. (2016), with 65.7% male participants) while others studied a predominantly female population (e.g., Yaikwawong et al. (2024), with 85% female participants) [20, 29]. In Rahimi et al. (2016), the intervention group (nano-curcumin) had a balanced gender distribution of 51.5% females, 48.5% males [32].

Most studies provide baseline glucose readings from participants (Table 3), indicating the severity of diabetes in the research populations. The large range of baseline glucose levels is a reflection of disparities in the severity of diabetes between patient populations and differences in inclusion criteria between studies. Baseline glucose levels ranged from 8.85 mmol/L (Anderson et al., 2016) to 123.65 mg/dL (Yaikwawong et al., 2024) and 180 mg/dL (Mirfeizi et al., 2016) [20, 22, 29]. It is difficult to assess the severity of diabetes across all included studies, as some, such as Geberemeskel et al. (2019), describe aberrant baseline glucose levels without using precise values [33].

Most studies also provide the end glucose levels following the intervention (Table 3), allowing evaluation of the effects of herbal treatments on lowering blood glucose levels.

For example, Anderson et al. (2016) reported a small decrease in final glucose levels of 8.19 mmol/L [20]. Yaikwawong et al. (2024) recorded an average final glucose level of 115.49 mg/dL, significantly lower than the baseline [29]. Rahimi et al. (2016) reported a final mean glucose level of 118.38 mg/dL, which was lower than that of the placebo group [32]. Certain studies, such as Zare et al. (2018), suggest a drop in glucose levels based on the data but do not provide the final values [25].

Additional baseline variables, such as body mass index (BMI), lipid profiles, and insulin levels, are provided by certain studies and are crucial for comprehending the overall metabolic health of study participants. For instance, baseline insulin, cholesterol, and glucose values are reported by Panahi et al. (2018), while Rahimi et al. (2016) offer glucose levels with comprehensive baseline lipid profiles (TC, TG, LDL-C, and HDL-C) [31, 32].

The main results and conclusions of the included studies are summarised in Supplementary 1, which focuses on how different herbal therapies affect lipid profiles, glycaemic management, and other metabolic parameters in individuals with T2DM.

According to several included studies, postprandial glucose (PPG), HbA1c, and fasting blood glucose (FBG) levels significantly improved after herbal therapies. Anderson et al. (2016) reported that cinnamon extract enhanced insulin sensitivity (HOMA-IR) and decreased FBG by 0.66 mmol/L [20]. Yaikwawong et al. (2024) observed that curcumin supplements dramatically decreased HbA1c (6.12% vs. 6.47% in the placebo group) and FBG (115.49 mg/dL vs. 130.71 mg/dL in the placebo group) [29]. Meanwhile, Rahimi et al. (2016) showed that nano-curcumin decreased HbA1c by 0.9% and FBG by 17.12 mg/dL when compared to placebo [32]. However, other investigations, such as Talaei et al. (2017), found no effect of cinnamon supplements on glycaemic control (FBG, HbA1c, HOMA-IR) [24].

Numerous studies document changes in lipid profiles, including increases in HDL-C and decreases in TG, TC, and LDL-C. According to Geberemeskel et al. (2019), fenugreek seed powder increased HDL-C (by 21.7%) while considerably lowering TC (by 13.6%), TG (by 23.53%), and LDL-C (by 23.4%) [33]. According to Neta et al. (2021), taking supplements of Curcuma longa L. increased HDL-C and decreased TG levels by 21% [30]. According to Shidfar et al. (2015), ginger supplements dramatically raised HDL-C levels and decreased LDL-C levels [42]. Nevertheless, certain studies, such as Kandhare et al. (2018), indicated that herbal therapies did not significantly alter lipid profiles [34].

Improvements in pancreatic β-cell function (HOMA-β) and HOMA-IR were reported in certain investigations. Yaikwawong et al. (2024) observed decreased HOMA-IR and enhanced HOMA-β after their curcumin intervention; however, this was contradicted by another study, which found no effect (Hodaei et al., 2019) [28, 29]. According to Panahi et al. (2018), curcuminoids and piperine decreased FBG levels and enhanced insulin sensitivity [31].

Several studies showed decreases in oxidative stress and inflammatory markers after herbal therapy. According to Shidfar et al. (2015), ginger supplements raised antioxidant capacity (TAC) and decreased oxidative stress indicators (such as Malondialdehyde (MDA)) [42]. Moravej Aleali et al. (2019) reported that saffron enhanced antioxidant status and decreased inflammatory indicators (hs-CRP) [48]. However, Shahbazian et al. (2019) found no effect of saffron supplementation on inflammatory biomarkers like TNF-α or IL-6 [52].

Following herbal therapies, some studies showed changes in waist circumference, body weight, and BMI. Adab et al (2019) found that turmeric supplements significantly lowered body weight and BMI [26]. Hodaei et al. (2019) observed decreased hip circumference and weight after curcumin treatment compared to placebo [28]. Nevertheless, other studies, including Talaei et al. (2017), indicated no discernible effects on body weight or BMI after cinnamon supplementation [24].

The conclusions varied widely between studies. Numerous authors concluded that lipid profiles, glycaemic management, and other metabolic parameters were improved by the specific herbal therapy given. Anderson et al. (2016) stated that in individuals with high blood glucose, cinnamon extract decreased FBG, insulin, and cholesterol levels [20]. Yaikwawong et al. (2024) reported that curcumin decreased FBG, decreased insulin resistance, and enhanced β-cell activity [29]. However, in other reports, such as Rachid et al. (2022), the herbal intervention had no discernible impact on glycaemic control (e.g., postprandial glucose response) or other measures [23].

The risk of bias assessment identified five forms of bias: randomisation, variations from intended intervention, missing outcome data, assessment of outcome, and selection of reported results. Most studies exhibited minor risks of bias (Figs. 2 and 3). Green circles indicate a low risk in all domains. Some studies (Ahmad et al., 2017; Sissoko et al., 2020; Azimi et al., 2015; Rahimi et al., 2016; Verma et al., 2016; and Suchitra et al., 2015) exhibit moderate concerns, with yellow circles in one or more areas [32, 36, 37, 43-45, 49]. Domains D1 (randomisation procedure) and D4 (measuring of outcome) typically caused concern. A high risk of bias was detected for certain studies, with red markers suggesting a high risk in at least one category. These studies may have problems with randomisation or group allocation techniques (Taweerutchana et al., 2017; Hameed et al., 2023; Leone et al., 2018) [44, 46, 47]. Overall, most studies were deemed low risk, fulfilling essential methodological quality requirements. A few studies exhibited some concerns, and a few presented substantial risks in certain domains.

Upon the eleven items included for the quality of this study, the quality assessment showed that all studies addressed a clearly formulated research question, and their participants were randomised to the interventions. Suchitra (2015) did not report some of participant basic characteristics and baseline glucose levels. Besides, Azimi (2015), Ebrahimi (2019), and Milajerdi (2018) did not mention the gender distribution in the study, neglecting the gender effect on the study outcomes. Regarding the study duration, Yaikwawong's 2024 study continued the study for 12 months, including the probability of long-lasting effect of curcumin. While Rahimi H.R. 2016 took only 3 months of therapy, this may not give the long-lasting effect of the same drug. Because these studies tested a safe herbal product, the benefits of the experimental intervention outweigh the harms, neglecting the costs' effect. All studies with blind controlled trials affirmed the blindness, except for Rania A Najdi, 2019 [35], who reported an open-label study.