The meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (18) and registered on PROSPERO (Registration Number: CRD42024606929).

A comprehensive electronic database search strategy was constructed to identify the effects of the YR on human health. The systematic search was performed on PubMed, Cochrane Library, Springer link, and Web of Science using “yacon” (MeSH) or “Smallanthus sonchifolius” (MeSH) terms from 2000 to November 2025, and major Chinese databases, including CNKI, Wanfang, and VIP, using “xuelianguo,” “yagong,” “XLG,” “yacon,” and “Smallanthus sonchifolius” terms from inception to November 2025. No search filters or limits were applied.

The PICOS (population, intervention, control, and outcomes) model, which outlines the inclusion and exclusion criteria, was used to select the eligible studies for the meta-analysis (Table 1).

Inclusion criteria were as follows: (1) availability of full-text articles; (2) clinical trials; (3) healthy people or people with metabolic disorders aged over 18 years; (4) single YR either in pure or standardized extract forms as a standalone intervention (5) reporting the mean and standard deviation (or standard error) or providing necessary data for these values at baseline and postintervention in both intervention and control groups; (6) reporting the sample size.

Exclusion criteria were as follows: (1) had participants involved adolescents; (2) involved the combined effects of other interventions; (3) presented insufficient data on the primary outcomes in both intervention and control groups; (5) reported in languages other than English; (6) showed only sensory outcome measures and (7) repeated publications of the same population in which the data were included for once unless there were different outcome measures.

Two investigators independently extracted data from the included studies after reviewing the titles, abstracts, and full texts. Data were extracted from the graphs using GetData Graph Digitizer 2.25 (19). A third investigator independently assessed and verified all data extraction. The researchers discussed any inconsistencies and disagreements, and a consensus was reached regarding the opinion of the researcher (Dr. Zhu), if necessary.

General information (authors, journal, publication year, and study design), participant information (age, gender ratio, sample size, participant descriptions, and baseline characteristics), intervention and control details (type, dosage of YR, and study duration), and outcome measures (weight, waist circumference, BMI, blood glucose indices, blood lipid indices, bowel movement status, and Nutrient intake) were extracted from the eligible literatures.

The Cochrane Handbook was utilized for the evaluation of risk bias. Two investigators evaluated the methodological quality from selection bias, performance bias, detection bias, attrition bias, reporting bias, and other biases. The methodological assessment for each trial was categorized using nominal scales: “yes” indicated a low risk of bias, “unclear” indicated an uncertain risk, and “no” indicated a high risk of bias.

Forest plots were generated only when two or more studies were available, in accordance with the Cochrane Handbook. If only one study was available, the findings were summarized narratively.

All the meta-analyses were performed using Stata software (version 17.0, Stata Corporation, College Station, TX) and Review Manager (RevMan) 5.4 software (Cochrane Collaboration, Oxford, UK). Mean difference (MD) was calculated with 95% confidence interval (CI), and I² test was used to evaluate the heterogeneity of the data. Considering the significant heterogeneity between different studies, a random-effects meta-analysis was performed to analyze the data by pooling the outcomes of the included studies. Subgroup analysis was conducted to investigate the effects of various variables on clinical outcomes.

The screening process for eligible studies, as illustrated in the PRISMA flowchart, is presented in Figure 1. A total of 3,599 articles were identified through searching databases, resulting in 1,461 records after excluding 2,138 duplicates. Subsequently, 1,449 articles were excluded due to non-compliance with the inclusion criteria, and ultimately 12 eligible articles were selected for eligibility.

The characteristics of the included studies are summarized in Table 2. Among the included 12 studies (12–16, 20–26), four employed a self-controlled design (n = 123), and eight used placebo-controlled design (n = 302). Each study included between 15 and 55 subjects. The ages of participants ranged from 24.87 ± 2.75 to 67.11 ± 6.12 years. The studies originated from various geographical locations, including Asia: Japan (n = 1), South America: Peru (n = 1), Argentina (n = 1), and Brazil (n = 9). The studied population consisted of healthy subjects, excess-weight adult volunteers, patients with type 2 diabetes, and elderly subjects. The intervention duration varied from 1 day to 5 months. The dosage was calculated based on the FOS in YR, ranging from 6.4 g to 14 g FOS/day. Participants received interventions through yacon syrup, freeze-dried powdered yacon (FDY), a yacon-based product (YBP), yacon flour, and yacon shake. The primary outcome measures included BMI, body weight, waist circumference, stool frequency, stool consistency, fecal pH and short-chain fatty acids, as well as glucose, insulin and HOMA-IR levels, lipid and lipoprotein levels and nutrient intake. 3.3 Risk of bias of the included trials.

The risk of bias of the 12 studies was assessed using the Cochrane Collaboration tool. In selection bias, 10 trials employed random allocation schemes, classifying low risk. Additionally, 10 trials reported adequate allocation concealment, also deemed low risk. Nine studies implemented double-blind protocols, which were considered low risk for both performance and detection biases. All studies presented complete data, leading to a low risk assessment for attrition bias. In summary, out of the 12 studies, 9 trials exhibited an overall low risk of bias, with no areas of high or unclear risk identified (Figure 2).

Three studies were pooled to assess the effect of YR on the stool frequency in subjects. Compared with the control group, the stool frequency in the YR group was significantly increased (SMD = 4.03, 95% CI (0.54, 7.52), p = 0.02, I² = 97.0%) (Table 3; Supplementary Figure S4A).

Three studies assessed stool consistency using the Bristol Stool Form Scale, and the meta-analysis results showed that YR consumption significantly softened the stool consistency of participants (SMD = 0.94, 95%CI (0.08, 1.80), p = 0.032, I² = 76.5%) (Table 3; Supplementary Figure S4B).

Two studies evaluated the effect of YR on fecal pH. The meta-analysis revealed a significant decrease in fecal pH in the YR interventional group (SMD = −1.12, 95%CI = −1.61, −0.62, p = 0.000) with no heterogeneity (I² = 0.0%) (Table 3; Supplementary Figure S4C).

Two studies analyzed the effect of YR on the contents of short-chain fatty acids (SCFAs) including acetate, propionate, and butyrate. The meta-analysis revealed that no significant differences in SCFAs levels before and after YR intervention but with tendencies: acetate: SMD = −0.04, 95%CI (−0.49, 0.42), p = 0.88, I² = 0.0%; propionate: SMD = −0.08, 95% CI (−0.53, 0.38), p = 0.74, I² = 0.0%; butyrate: SMD = 1.07, 95%CI (−1.12, 3.25), p = 0.34, I² = 94.2% (Table 3; Supplementary Figure S5).

Five studies concerned the changes in fasting blood glucose following YR consumption. Mean values and standard deviations of changes before and after intervention were extracted for meta-analysis. Results indicated YR showed no effects on fasting blood glucose (SMD = −0.55 95%CI (−1.81, 0.72), p = 0.4) with high heterogeneity (I² = 92.1%) (Table 3; Supplementary Figure S6A). Four studies reported fasting insulin statistics and HOMA-IR, respectively. The meta-analysis revealed no statistical significance in fasting insulin (SMD = −0.55, 95%CI (−1.81, 0.72), p = 0.4, I² = 92.1%) and HOMA-IR (SMD = −0.69, 95%CI (−2.21, 0.82), p = 0.371, I² = 95.0%) between pre- and post-intervention (Table 3; Supplementary Figures S6B,C).

Two studies (12, 16) containing three trials investigated the effect of YR on postprandial blood glucose. Meta-analysis results indicated no significant difference between experimental and control groups (SMD = −0.32, 95%CI (−0.69, 0.06), p = 0.10, I² = 0.0%) but with decrease tendency (Table 3; Supplementary Figure S7A).

Three studies assessed fasting total cholesterol, and five studies examined fasting HDL-c, LDL-c and triglycerides before and after intervention. The meta-analysis showed that YR did not reduce fasting total cholesterol, HDL-c, LDL-c and triglycerides but with regulatory tendencies (total cholesterol: SMD = 0.02, 95% CI (−0.29, 0.33), p = 0.90, I² = 0.0%; LDL-c: SMD = 0.12, 95%CI (−0.76, 1.01), p = 0.76, I² = 89.5%; HDL-c: SMD = −0.04, 95%CI (−0.31, 0.22), p = 0.76, I² = 0.0%; triglycerides: SMD = 0.54, 95% CI (−0.39, 1.46), p = 0.25, I² = 90.1%) (Table 3; Supplementary Figure S8).

One study with two trials evaluated the effect of YR on postprandial triglyceride levels in both healthy and overweight individuals. The meta-analysis showed that YR decreased postprandial triglyceride levels in a certain degree (SMD = −0.13, 95%CI (−0.57, 0.31), p = 0.55, I² = 0.0%) (Table 3; Supplementary Figure S7B).

Two studies assessed fiber intake before and after intervention. The meta-analysis revealed fiber intake was significantly increased in interventional group (SMD = 0.58, 95%CI (0.18, 0.99), p = 0.01, I² = 0.0%). Two studies analyzed energy intake before and after intervention. The meta-analysis indicated energy intake was not significantly changed, but with a decrease tendency (SMD = −0.16, 95%CI (−0.56, 0.24), p = 0.42, I² = 0.0%). Three studies reported carbohydrate, fat, and protein intake before and after intervention. The meta-analysis showed the intake of carbohydrate, fat and protein in interventional group showed a decreasing trend (Carbohydrates: SMD = −0.08, 95%CI (−0.42, 0.27), p = 0.67, I² = 0.0%; Fat: SMD = −0.15, 95%CI (−0.50, 0.20), p = 0.39, I² = 0.0%; Protein: SMD = −0.12, 95%CI (−0.47, 0.23), p = 0.49, I² = 0.0%) (Table 3; Supplementary Figure S9).

Inulin-type fructans, including fructooligosaccharides (FOS, DP < 10) and inulin (DP 2–60), are low-calorie dietary fibers. They can replace high-calorie carbohydrates, reduce energy intake, slow gastric emptying, suppress hunger, and decrease the small intestine’s absorption of sugar and fat. In addition, inulin-type fructans resistant to digestive enzymes could reach colon, where optimize the intestinal microenvironment via promoting the generation of beneficial bacteria and production of postbiotics, and reducing the generation of pathogenic bacteria and production of intestinal endotoxin (27, 28). The intestinal microenvironment optimization improves dysbiosis-related conditions, including but not limited to obesity, constipation and metabolic syndrome. For example, beneficial bacteria enhance intestinal barrier function, decrease permeability, and limit the entry of pro-inflammatory substances like lipopolysaccharide (LPS) into the bloodstream, reducing chronic low-grade inflammation associated with obesity (29–31). Meanwhile, SCFAs (mainly acetic, propionic, and butyric acids), one of a type of postbiotics yielded by FOS, regulate various physiological processes in the human body. SCFAs lower colonic pH (32, 33), maintain intestinal moisture and osmotic pressure, increase stool viscosity and moisture, promote defecation, and alleviate constipation (34). They also reduce appetite and food intake by modulating appetite-related hormones, increasing peptide YY levels, and inhibiting growth hormone-releasing peptide and leptin secretion (35–37).

YR contains abundant phenolic compounds, including chlorogenic acid, caffeic acid, ferulic acid, quinic acid, and quercetin, which are bioactive substances crucial for managing chronic conditions like obesity, diabetes, and metabolic disorders. Phenolic compounds mainly exert functional effects through the following manners. (1) Phenolic compounds regulate lipid metabolism. For example, chlorogenic acid inhibits porcine pancreatic lipase (EC 3.1.1.3), reduces lipid absorption, and upregulates peroxisome proliferator-activated receptor (PPAR) mRNA, promoting fat metabolism and reducing weight gain and visceral fat accumulation in the liver (38). Kaempferol inhibits lipogenesis by suppressing Akt and mTORC1 activation, blocking downstream SREBP1C signaling, and directly activating AMPK to further inhibit SREBP1C-mediated adipogenesis (39). Quercetin enhances lipid metabolism and promotes lipolysis through SIRT1 and Akt pathway activation (40). (2) Phenolic substances regulate blood glucose metabolism. For instance, chlorogenic and caffeic acids inhibit α-amylase, lowering blood glucose and increasing insulin levels, thereby improving pancreatic β-cell function (41). Quercetin improves insulin signaling and lowers blood glucose by promoting Akt and glycogen synthase kinase-3 (GSK-3) phosphorylation while its antioxidant properties reduce oxidative stress and repair damaged pancreatic β-cells (40). Ferulic acid neutralizes streptozotocin toxicity via antioxidant activity, protecting β-cells, promoting their proliferation and insulin secretion, and enhancing hepatic glucose utilization (42). Quinic acid stimulates insulin secretion and regulates blood glucose by mobilizing intracellular calcium ions and increasing the NADPH/NADP⁺ ratio (43).

In summary, we speculate that the function of YR may be the result of the combined action of inulin-type oligosaccharides, phenolic substances, etc. in YR (Figure 3). However, the precise mechanisms underlying YR’s clinical effects require further investigation.

This study has several strengths. First, we conducted a systematic and comprehensive search across multiple major English-language databases and applied strict, explicit and reproducible inclusion and exclusion criteria. This approach improved the completeness of the literature sources and enhanced the reliability of the findings. Second, only clinical trials using YR as the sole intervention were included, which minimized confounding from other dietary components or combined supplements and allowed a more clinically targeted and accurate attribution of effects. Third, this study integrated multiple outcome dimensions, including anthropometric indicators, bowel function, glucose and lipid metabolism, and dietary intake, providing a systematic evaluation of the potential clinical benefits of YR in weight management, metabolic health, and gut function. Previous studies have suggested that YR may be valuable for individuals with obesity, metabolic disorders, and functional constipation, possibly through its prebiotic effects, appetite regulation, and gut microbiota–mediated metabolic improvements.

This study has several limitations. First, for only English-language publications were primarily included, there is a possibility of language bias and retrieval bias due to limited database coverage. We conducted additional searches in CNKI, Wanfang, and VIP using the keywords “xuelianguo,” “yagongguo,” and “XLG,” but the relevant Chinese literature were animal studies, and no randomized controlled clinical trials meeting the inclusion criteria were identified. In addition, nine of the included studies were conducted in Brazil, which may introduce regional bias and affect the generalizability of the findings. Second, the included studies generally had small sample sizes, and the participants varied across healthy subjects, overweight or obese individuals, patients with type 2 diabetes, and elderly populations. Baseline differences among these groups may lead to variations in effect size. Although subgroup analyses suggested that YR may have more pronounced effects in overweight individuals, this observation may be influenced by confounding factors and requires further validation. Finally, some outcomes showed high statistical heterogeneity, indicating substantial differences across studies in terms of intervention form, dosage, duration, and lifestyle background. These factors may affect the stability of the effect estimates, and therefore the results should be interpreted with caution.

The findings of this study indicate that YR, as a safe and natural source of dietary fiber or prebiotics, may be applied in nutritional interventions for individuals with mild obesity, metabolic disturbances, and functional constipation. It may also serve as a dietary option for populations with inadequate fiber intake. In addition, these results provide scientific support for the development of YR-based functional foods. The potential value of YR in personalized and precision nutrition strategies also warrants further exploration.

Given the limited scale and high heterogeneity of the current evidence, future studies should include large-sample, multicenter, and long-term randomized controlled trials, with a particular focus on populations at metabolic risk. These studies should systematically evaluate the dose–response relationship of different formulations (e.g., syrup, powder), dosages, and intervention durations. Furthermore, mechanistic indicators such as gut microbiota and metabolites should be incorporated to identify the most suitable target populations and to elucidate the underlying mechanisms of action.