We searched all publications from database inception to 15 January 2025 using three electronic databases: EMBASE, MEDLINE, and The Cochrane Central Register of Controlled Trials. No language restrictions were applied. The search strategy included the following terms: (exp ilex paraguariensis/or ((“mate” or “chimarrao” or “ilex paraguariensis” or “terere” or “yerba-mate”).ab.)) AND (randomized controlled trial.pt. or controlled clinical trial.pt. or randomized.ab. or placebo.ab. or drug therapy.fs. or randmo*.ab. or trial.ab. or groups.ab. not exp animals/not humans.sh.). To optimize the search outcomes, suitable term combinations and truncations were carefully chosen and customized for each database. The detailed search strategies are comprehensively outlined in Supplementary Table 1 . To ensure the identification of all potentially relevant studies, reference lists of included publications were manually scrutinized. Other study sources are mainly supplemented through reference list searching.

The present review included RCTs that assessed the effects of Yerba Maté intake on metabolic health indicators. The inclusion criteria for the studies are systematically presented in Table 1 according to the PICOS principle. No restrictions were placed on language or time. Studies were excluded based on the following criteria: (1) Non-original research publications including reviews, letters, posters, conference abstracts, case reports and opinion pieces; (2) Studies where Yerba Maté was combined with other plants or supplements; (3) Studies without available information of interest.

Two investigators (Li Huang and Yadong Liu) independently screened titles and abstracts, followed by full-text reviews of eligible studies. Data were extracted according to predefined criteria. Disagreements were resolved through consensus discussions involving a third independent reviewer (Jirong Yue). The extracted data included: first author’s last name, publication year, country, study design, sample size (intervention and control groups), mean age, gender, dosage of Yerba Maté, study duration, treatment duration, and baseline and outcome data for triglycerides, total cholesterol, HDL-C, LDL-C, fasting glucose, fasting insulin, postprandial glucose, HbA1c, waist circumference, body mass index (BMI), homeostasis model assessment (HOMA) index, and adverse events (AEs).

Two investigators (Daiping Li and Liantian Yue) independently assessed the risk of bias in the included RCTs using the RoB 2 tool. The risk of bias in parallel RCTs was assessed using the RoB 2.0 tool for RCTs (Version dated 22 August 2019). For crossover trials, the RoB 2.0 tool specific to crossover studies was utilized (Version dated 18 March 2021). Both tools evaluate the risk of bias across five domains: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of reported results. Disagreements were resolved through discussion with a third author (Jirong Yue). In cases where pertinent information was lacking, we proactively contacted the respective authors for clarification.

The overall quality of the evidence was evaluated using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach (21–23). The GRADE framework assesses the quality of evidence across five domains: risk of bias, directness of evidence, heterogeneity, precision of effect estimates, and risk of publication bias. The quality of evidence was independently evaluated by two reviewers (Daiping Li and Liantian Yue). In cases of discrepancies, an additional experienced rater was consulted (Jirong Yue).

The primary outcomes assessed in this study included triglycerides, total cholesterol, HDL-C, LDL-C, fasting glucose, fasting insulin, HbA1c, postprandial glucose, waist circumference, BMI, and AEs. All outcomes were reported as the change from baseline values. The incidence of adverse events was also analyzed.

Meta-analyses were performed using Review Manager (RevMan 5.4) when two or more studies from the same comparator provided sufficient data (i.e., mean difference [MD], standard deviation [SD], and number of participants in each intervention group). The I² statistic was calculated to quantify between-trial heterogeneity and inconsistency. When summary measures were not reported as MD and SD, previously published conversion tools were utilized (24). The randomized crossover trial used data from both intervention periods for analysis in the meta-analysis.

We quantified heterogeneity with I², interpreting 0-40% as low, 40-60% as moderate, and 60-90% as high (25). All meta-analyses were conducted using a random-effects model irrespective of I² value, in accordance with Cochrane Handbook recommendations. The chi-square P value was calculated to assess the statistical significance of heterogeneity, with P ≤ 0.05 indicating significant heterogeneity among the studies included in the meta-analysis. Sensitivity analyses were performed to evaluate the robustness of the results by: (1) excluding low-quality studies and (2) excluding studies with fewer than 10 participants.

A total of 1451 citations were identified through the initial database search. After eliminating duplicates and screening titles, 1294 citations were considered potentially relevant. During the title/abstract screening phase, 1265 studies were excluded due to failure to meet the study objectives. Full-text screening further excluded 16 studies (12, 26–40). The reasons for excluding these studies are detailed in Supplementary Table 2 . Ultimately, 13 RCTs (17, 21, 41–51) fulfilled the inclusion criteria for this systematic review and meta-analysis. Figure 1 is a flow diagram of the study selection process.

A detailed overview of the study characteristics is presented in Table 2 . The 13 studies enrolled participants with ages ranging from 28 to 57 years. Among these studies, 9 were parallel RCTs and 4 were crossover RCTs. The studies were conducted across various countries: five in Brazil (17, 43, 46, 50, 51), three in Korea (21, 47, 48), two in Italy (44, 45), one in Germany (49), one in Norway (41), and one in Argentina (42). The study sample comprised healthy adults (41, 49), overweight or obese individuals (21, 42, 47, 48), participants with metabolic abnormalities (17, 43–46), and people with HIV (50, 51). Interventions were delivered as Yerba Maté capsules (21, 41, 44–51) or as Yerba Maté tea (42, 43, 49). Daily capsule doses ranged from 0.5 g (44), 1.0 g (45), 2.25 g (46), 3.0 g (21, 47, 48, 50, 51) to 5 g (41); tea doses were 20 g (17, 43) or 100 g (42). The extract used by Opala 2006 (49) (containing asparagus, green tea, black tea, guarana, maté, and kidney beans) did not specify individual amounts. Treatment durations spanned 5 days (41), 15 days (50, 51), 4 weeks (46), 6 weeks (47, 48), 60 days (17), 3 months (44, 45), 12 weeks (21, 42, 49) and 90 days (43).

Among the nine parallel RCTs included, potential bias was identified across several domains in Supplementary Figure 1 . Specifically, issues with the Randomization process (D1) were noted in three studies (17, 42, 43) and concerns regarding Deviations from the intended interventions (D2) were observed in two study (17, 43). One study (17) was rated as some concerns in the Missing outcome data (D3) domain. All studies, however, demonstrated a low risk of bias in the Measurement of the outcome (D4) domain. Furthermore, three studies (17, 42, 47) exhibited some concerns in the Selection of the reported result (D5) domain. Five studies (21, 44, 45, 48, 49) exhibited low risk of bias across all domains, increasing confidence in their findings.

Among the four crossover RCTs included, potential bias was identified across several domains in Supplementary Figure 2 . Of the four crossover RCTs assessed, three studies raised some concerns about the Randomization process (D1) (41, 46, 50), and one study showed bias issues caused by menstruation and carryover effects (DS) (41). In terms of Deviations from the intended interventions (D2), one study presented a high risk of bias (50), while two studies indicated some concerns (41, 51). For Missing outcome data (D3), two studies were rated as high risk (50, 51). The Measurement of the outcome (D4) domain generally showed a low risk across studies, suggesting a reliable assessment of outcomes. As for the Selection of the reported result (D5) domain, three studies were identified with some concerns (41, 50, 51). Three studies (41, 50, 51) were rated as having some concerns in the domain of Selection of the reported result (D5). Two studies (50, 51) were classified as high risk and other two studies (41, 46) were identified with some concerns across all domains.

Applying the GRADE framework ( Table 3 ), we rated the certainty of evidence as high, moderate, or low for all outcomes. We downgraded waist circumference evidence by one level because of marked inconsistency (I²>50%). The evidence for fasting insulin and HOMA index was downgraded by one level due to detected publication bias (more than half of the studies come from the same team). Total cholesterol, triglycerides, LDL-C, and fasting glucose were downgraded by two level for very serious limitations, combining serious inconsistency and serious imprecision (wide 95% CI), resulting in low certainty. Postprandial glucose and HbA1c were downgraded by two levels due to very serious limitations, combining serious inconsistency and detected publication bias, resulting in low certainty.

A total of 11 RCTs were included to assess the impact of Yerba Maté on lipid levels, with follow-up durations ranging from 15 to 90 days ( Figure 2 ). Comparative analysis of Yerba Maté versus control groups demonstrated non-significant differences in total cholesterol ( Figure 2A ), triglycerides ( Figure 2B ), HDL-C ( Figure 2C ) or LDL-C ( Figure 2D ).

Seven RCTs were included to evaluate the effects of Yerba Maté on glycemic status, with follow-up durations ranging from 5 days to 3 months. The quantitative analysis results are summarized in Figure 3 . Meta-analysis revealed a significant reduction in postprandial glucose levels (MD -12.76, 95% CI -16.78 to -8.74, I² = 91%, p=0.001, N = 2) ( Figure 3C ), HbA1c levels (MD -0.37, 95% CI -0.56 to -0.18, I² = 64%, p=0.1; N = 2) ( Figure 3D ) and the HOMA index (MD -0.24, 95% CI -0.37 to -0.11, I² = 0%, p=0.82; N = 2) ( Figure 3E ). No significant differences were observed in fasting glucose levels ( Figure 3A ) or fasting insulin concentrations ( Figure 3B ) between the Yerba Maté group and the control group.

Eight RCTs were included to assess the effects of Yerba Maté on waist circumference and BMI, with follow-up durations ranging from 4 weeks to 3 months ( Figure 4 ). Comparative meta-analysis showed no significant differences in waist circumference ( Figure 4A ) or BMI ( Figure 4B ) between Yerba Maté and control groups.

We performed subgroup analyses of intervention population, intervention methods, and intervention duration. Subgroup analyses of intervention doses were not possible because most intervention doses were only available in 1–2 studies. We conducted a subgroup analysis of the intervention group for Yerba Maté ( Figures 5 , 6 ), including: dyslipidemic volunteers, overweight and obese and non-dyslipidemic, normal-weight volunteers. Subgroup analyses found a significant difference in triglycerides (MD -15.34, 95%CI -20.37, -10.31, I² =0%, p<0.001, n=5) ( Figure 5B ) and fasting glucose (MD -9.08, 95%CI -15.29, -2.88, I² = 95%, p=0.004, n=3) ( Figure 6A ) among dyslipidemic volunteers.

We conducted a subgroup analysis of our intervention methods for Yerba Maté ( Figure 7 ), including Yerba Maté extract and Yerba Maté. In this subgroup analysis, the intervention with Yerba Maté extract significantly reduced triglycerides(MD -9.75, 95%CI -17.60, -1.91, I² =52%, p=0.01, n=7) ( Figure 7B ). Most intervention methods included in the study were Yerba Maté extracts, with no more than one study involving Yerba Maté tea. Therefore, no subgroup analysis of intervention methods based on fasting glucose, waist circumference, and BMI was conducted.

We conducted a subgroup analysis of the intervention duration for Yerba Maté, including periods of less than 12 weeks and 12 weeks or longer. However, no significant differences were found ( Figures 8 , 9 ).

The safety profile of Yerba Maté was reported in four studies (12, 18, 21, 49), with adverse effects summarized in Table 4 . Yerba Maté tea consumption has been associated with adverse effects in certain individuals, including oral or gastric mucosal irritation, insomnia, nausea, acid reflux, tachycardia, angina pectoris, headache, and abdominal discomfort.

To evaluate the reliability of the results regarding Yerba Maté’s effects on lipid profiles and fasting blood glucose levels, sensitivity analyses were performed ( Figure 10 ). Excluding low-quality studies and those with fewer than 10 participants did not alter the overall findings, confirming the robustness of the meta-analysis results. No sensitivity analysis was performed on fasting insulin, postprandial glucose, HbA1c, HOMA index, waist circumference and BMI, since the studies did not include low-quality research or studies with fewer than 10 participants.

Our systematic review and meta-analysis demonstrate that Yerba Maté consumption may improve glycemic outcomes, including decreased postprandial glucose levels, HbA1c levels, and the HOMA index. However, no significant effects were observed on lipid profiles (triglycerides, total cholesterol, HDL-C, LDL-C), waist circumference, or BMI. These findings suggest that Yerba Maté may exert glycemic control benefits independent of lipid metabolism modulation.

It is important to highlight that the two studies indicating Yerba Maté’s beneficial effects on blood glucose were conducted by the same research team (44, 45), which may introduce reporting bias and warrants cautious interpretation. In these two studies (44, 45), the subjects were individuals with impaired fasting glucose and impaired glucose tolerance. While other studies (41, 46–49) included diverse populations such as healthy volunteers, athletes, and overweight individuals with normal or well-regulated blood glucose, who may respond less noticeably to interventions, leading to non-significant results. In addition, these two studies employed a longer intervention period of three months, providing a more extended window to observe potential changes in blood glucose indicators, whereas some shorter-term studies may not have been as effective in capturing significant improvements (41, 46–48). However, it should be noted that the two studies were conducted by the same research team and used a standardized formulation produced by the same company. This might lead to reporting bias. Therefore, their positive results should be interpreted with caution.

To contextualize our findings, we compared the effect sizes for fasting glucose and triglycerides with those reported for cinnamon and fenugreek (52, 53). Yerba Maté and cinnamon achieve identical fasting-glucose reductions (MD -9.08), surpassing fenugreek (MD -0.84); for triglycerides cinnamon (54) retains the largest effect (WMD -29.59), while yerba maté provides a moderate but clinically relevant benefit (MD -20.37).

Some systematic reviews have yielded inconsistent results concerning the metabolic impacts of Yerba Maté supplementation. For example, Masson et al. (10) showed a significant reduction in LDL-C in non-diabetic populations, Clemente et al. (15) found HDL-C improvements in obese populations using 3g/d standardized extracts, but José et al. (55) observed no effects on glycemic parameters in type 2 diabetes patients given <1g/d extracts. Four factors might account for the disparity between our findings and prior meta-analyses. First, there are population-specific differences in metabolic pathways. José et al. (55) focused on chronic disease populations but excluded the acute metabolic responses in our athlete subgroup. While Clemente et al. (15) reported HDL-C improvement in homogeneous obese cohorts, our study included HIV patients and athletes whose lipid metabolism may be affected by antiretroviral therapy (50) or exercise-induced adaptations (41). Second, dose-response heterogeneity is crucial. Unlike Luís et al. (21), who focused on 3g/d standardized extracts, our analysis included extreme dosing like 50g/d tea in Avena 2019 (42). Third, there were marked differences in outcome measurement timelines. The HbA1c reduction in our study mainly came from ≥84-day interventions (44, 45), whereas the negative waist circumference findings included short-term trials like the 5-day cycling protocol in Areta 2018 (41), which contrasts with Luís et al.’s (21) 90-day focused analysis. Fourth, heterogeneity in study design caused bias. Our review strictly included only parallel RCTs and crossover RCTs to maximize internal validity. In contrast, Masson et al. (14) combined RCTs and observational studies (e.g., cohort and case-control studies), which might have allowed lifestyle confounders such as diet and exercise to interfere with the independent link between Yerba Maté intervention and metabolic outcomes.

Subgroup analysis of the intervention population revealed significant differences in triglycerides and fasting glucose levels among dyslipidemic volunteers. Subgroup analysis of the intervention methods revealed Yerba Maté extracts were significantly differences in triglycerides. Dyslipidemic volunteers and Yerba Maté extracts as possible sources of heterogeneity. Future studies may explore larger-scale interventions with Yerba Maté extracts among dyslipidemic volunteers. However, subgroup analysis conducted in subgroups according to different intervention duration showed no significant differences in total cholesterol, triglycerides, HDL-C, LDL-C, fasting glucose, waist circumference and BMI between Yerba Maté and control groups.

Our systematic review and meta-analysis suggest that Yerba Maté consumption may be associated with improvements in glycemic control, as indicated by reductions in postprandial glucose, HbA1c levels, and HOMA index. However, no significant effects were observed on fasting glucose or insulin levels. These findings hint at potential mechanisms that might underlie the observed effects, such as enhanced insulin sensitivity and postprandial glucose regulation, though the exact mechanisms remain to be fully elucidated. Existing literature proposes that Yerba Maté polyphenols could potentially modulate hepatic insulin signaling by suppressing TNF-α (56), and may interact with the PI3K-AKT pathway, involving elements like Akt2 and Irs1, possibly contributing to improved peripheral insulin resistance (57). Regarding postprandial glucose control, it has been hypothesized that bioactive compounds in Yerba Maté might inhibit intestinal SGLT1-mediated glucose absorption (58). Additionally, its anti-glycation properties, which seem to surpass those of green tea (59), could potentially protect β-cells by inhibiting AGE formation during the secondary glycation phase. While these mechanistic insights offer plausible explanations for the observed effects, they are largely derived from preclinical studies and in vitro evidence. As such, they should be interpreted with caution and require further validation through well-designed clinical trials before Yerba Maté can be positioned as an adjuvant therapy for diabetes management.

The present study found no significant effects of Yerba Maté consumption on waist circumference or body mass index (BMI), suggesting limited efficacy in directly reducing abdominal adiposity or overall body weight. However, preclinical evidence indicates its potential to modulate energy metabolism through three interconnected mechanisms: (1) Caffeine-mediated activation of the sympathetic nervous system enhances lipolytic enzyme activity, though clinical relevance requires exceeding dose-dependent thresholds (>3 g/day of standardized extracts or >20 g/day of traditional infusion) to achieve thermogenic effects (60–62); (2) While acute interventions (e.g., 5-day trials) demonstrate efficacy in delaying gastric emptying and enhancing satiety via gastrointestinal modulation, these effects fail to translate into sustained body composition improvements in longitudinal studies, likely due to compensatory metabolic adaptations (28, 63); (3) Although Yerba Maté supplementation reduces chronic inflammation markers such as C-reactive protein (CRP) – a known correlate of visceral adiposity – its anti-inflammatory properties exhibit paradoxical dissociation from abdominal fat redistribution, necessitating extended clinical trials to validate causal relationships (55, 62).

Although Yerba Maté (Ilex paraguariensis) is recognized for its numerous health benefits, its potential adverse effects and carcinogenic risks warrant attention. Studies have indicated that long-term, high-volume consumption of Yerba Maté may increase the risk of certain cancers, including oral, esophageal, laryngeal, and lung cancers (64–68). Notably, the risk of carcinogenesis is significantly elevated when Yerba Maté is consumed at temperatures exceeding 65 °C, which may be attributed to thermal injury to esophageal cells caused by hot beverages (65–68). Additionally, Yerba Maté contains various chemical constituents, such as tannins, nitrosamines, and polycyclic aromatic hydrocarbons, which may possess carcinogenic potential (68–72). However, other research has suggested that the polyphenolic compounds in Yerba Maté may exert protective effects against certain cancers (70–72). Despite these findings, the current evidence is inconsistent, with some studies indicating that the association between hot Yerba Maté consumption and increased cancer risk may be influenced by other factors, such as smoking and alcohol consumption (64, 73). The randomized controlled trials included in this systematic review all had intervention periods of no more than three months. Future research necessitates long-term follow-up studies to further investigate the safety profile of Yerba Maté, particularly its health impacts under different consumption temperatures and dosages.

The methodological quality assessment revealed notable variations across study designs. Among the nine parallel RCTs, five studies (21, 44, 45, 48, 49) demonstrated low risk of bias across all domains, while the remaining four exhibited some concerns or high risk in specific domains (particularly randomization process and deviations from intended interventions). The crossover RCTs showed more substantial methodological limitations, with two studies (50, 51) classified as high risk and two (41, 46) having some concerns across multiple domains.

Significant heterogeneity persisted in this study’s meta-analysis, despite conducting subgroup and sensitivity analyses. Several factors likely contributed to this heterogeneity. First, the studies encompassed a wide range of participant characteristics, including age, gender, and health status, from healthy individuals to those with conditions like obesity, prediabetes, metabolic syndrome, and HIV infection. Such diversity in metabolic status and disease burden can lead to varied responses to Yerba Maté interventions. Second, differences in the form of intake of Yerba Maté and study design, such as differences in the mode of use (tea infusions, extract or capsules), whole and part amounts used, dosage, and duration of intervention, could act as confounders responsible for differences between effects in the short term and the long term. Third, studies were marked by strong differences in baseline characteristics. Some studies may have included subjects with high baseline triglyceride levels. In contrast, others have lower baseline levels that could have resulted in different effects of Yerba Maté being observed. Finally, heterogeneity may be attributed to study design and methodology differences, e.g., randomization, blinding and data collection. These issues can be addressed by future research that standardizes intervention and measurement protocols, broadens the study population, controls all confounding variables, and strengthens the study’s design. Since this approach reduces heterogeneity, it will aid in increasing the reliability and the generalizability of findings.

The present study has several strengths. Its exhaustive search strategy allows all relevant literature to be included. By minimizing the risk of publication bias, this approach guarantees that as many studies as possible are represented and thus makes the findings more robust. Moreover, the studies are assessed with rigorous evaluation methods, including the RoB 2 tool, for the risk of bias so that each study is thoroughly scrutinized. The GRADE approach adds to the analysis by practicing a systematic and transparent approach to evaluating the overall quality of evidence and, thus, the level of confidence in the results.

This review is subject to certain limitations. Given that research findings from South America may be published in languages such as Spanish, we ultimately decided to include studies without language restrictions. This is not consistent with our PROSPERO record. The major issue is a big heterogeneity in clinical and procedural aspects of the Yerba Maté consumption in the studies examined, including differences in the details of their participants, their Yerba Maté consumption regimens, and the different follow-up durations. Further, heterogeneity extends to assessing such outcomes as total cholesterol, LDL-C, fasting blood glucose, postprandial blood glucose, HOMA index and BMI, which may influence the comparison and the generalizability of the results. In addition, the study of outcomes, including postprandial blood glucose, HOMA index, and HbA1c, is limited to only two of the included studies, which may result in too few subjects to allow sound conclusions. Furthermore, three studies had a follow-up period under 20 days, which may raise a question of the length of follow-up was insufficient for the Yerba Maté consumption. Future research should strive to develop protocols and outcome measures that are standardized in order to facilitate comparison and reliability of results.

Although consumption of Yerba Maté has shown health benefits, future research should go further in studying the long-term effects and risks of this consumption. Future research investigations on Yerba Maté and glycemic control should focus on extending the intervention duration, employing Yerba Maté extracts, and enrolling subjects with metabolic disorders. As far as it is carcinogenic, this matter is still inconclusive because some studies on Yerba Maté show that carcinogenic substances (Polycyclic aromatic hydrocarbons, or PAHs, and tannins) come not from Yerba Maté itself but from the contamination during processing. Yet more studies have come to the conclusion that perhaps it is not the tea that poses a problem but the high temperature with which it is traditionally consumed. Moreover, smoking and alcohol drinking may contribute to the observed cancer risk. Given the present study’s limitations, large-scale prospective cohort studies are required to clarify whether Yerba Maté carries a carcinogenic risk and to differentiate temperature, PAH, and other variables.

Furthermore, the role of Yerba Maté in blood pressure and other metabolic syndrome indices (uric acid) has been widely ignored. The current research focus mainly involves short-term interventions with limited systematic evaluation of long-term effects. RCTs that are well designed are necessary to validate the effects of Yerba Maté on the metabolic indicators and elucidate the physiological mechanisms. Evaluation of the role of Yerba Maté in preventing and treating chronic diseases has not been possible because of the lack of long-term intervention studies. Studies of the incidence of severe cardiovascular events in prospective cohort studies with extended follow-up could supply important information about its long-term health effects.

Then, the possible effects of Yerba Maté on several different populations, e.g., the older adults, the ones who are athletes, and the diabetic, are variant due to the differences in each person. Its mechanisms of action and effects in specific populations merit future research, focusing on the feasibility of a personalized intervention. In doing this, we will augment our entity’s understanding of the health implications of Yerba Maté.