SYSTEMATIC REVIEW article

Front. Pharmacol., 18 June 2026

Sec. Drugs Outcomes Research and Policies

Volume 17 - 2026 | https://doi.org/10.3389/fphar.2026.1846714

Efficacy and safety of incretin-based therapies in patients with type 2 diabetes mellitus: a network meta-analysis based on clinical trials

  • 1. Shandong Medicine and Health Key Laboratory of Clinical Pharmacy, Shandong Engineering and Technology Research Center for Pediatric Drug Development, The First Affiliated Hospital of Shandong First Medical University / Shandong Qianfoshan Hospital, Jinan, China

  • 2. School of Pharmacy, Shandong University of Traditional Chinese Medicine, Jinan, Shandong, China

Abstract

Objectives:

This study systematically assessed the efficacy and safety of 15 incretin-based therapies (IBTs) for type 2 diabetes mellitus (T2DM), including mono- (GLP-1RA), dual- (GIP/GLP-1RA or GLP-1/GCGR), and triple- (GIP/GLP-1/GCGR) receptor agonists, and explored how dosage and treatment duration affect clinical outcomes to support individualized treatment decisions.

Methods:

A systematic review and Bayesian network meta-analysis were performed following PRISMA 2020 and PRISMA-NMA guidelines (PROSPERO: CRD420251152746). We searched MEDLINE, Cochrane Library, and Embase up to 1 August 2025, for randomized controlled trials (RCTs). Study selection, data extraction, risk-of-bias assessment (Cochrane RoB 2.0) and CINeMA framework were conducted. Analyses included SUCRA ranking, sensitivity, subgroup, and meta-regression in R Studio 4.4.3.

Results:

102 RCTs (98,693 T2DM patients) with low overall bias were included. IBTs were superior to placebo in glycemic control (tirzepatide, orforglipron, semaglutide best), weight loss (retatrutide best), lipid, blood pressure, cardiorenal, and insulin function improvement. The most common adverse events were mild, transient gastrointestinal reactions, and the risk of hypoglycemia was low. Higher doses improved efficacy but increased adverse events; 45 mg orforglipron balanced benefits and tolerability. Optimal duration and combination efficacy varied by agent.

Conclusion:

IBTs provide comprehensive benefits in T2DM. Efficacy and safety differ across agents, doses, durations, and combinations, supporting individualized therapy. Long-term high-quality RCTs are needed. Cardiovascular protection may stem from synergistic improvements in lipids, blood pressure, renal function, and insulin sensitivity.

Systematic Review Registration:

https://www.crd.york.ac.uk/prospero/, identifier CRD420251152746.

1 Introduction

Type 2 diabetes mellitus (T2DM) is a major global public health challenge, with its increasing prevalence placing substantial strain on healthcare systems and markedly impairing patients’ quality of life (). Beyond glycemic dysregulation, T2DM is frequently associated with comorbidities such as obesity, dyslipidemia, hypertension (), as well as elevated risks of target organ damage—including atherosclerotic cardiovascular disease (ASCVD) (), chronic kidney disease (CKD) (; ), and metabolic dysfunction-associated steatotic liver disease (MASLD) ()--all of which complicate clinical management. Incretin-Based Therapies (IBTs), including mono- (glucagon-like peptide-1 receptor agonist, GLP-1RA), dual (glucose-dependent insulinotropic polypeptide/GLP-1 receptor agonist, GIP/GLP-1RA; or GLP-1/glucagon receptor agonist, GLP-1/GCGR), and triple (GIP/GLP-1/GCGR) receptor agonists, have emerged as cornerstone therapies in T2DM due to their pleiotropic benefits, including glycemic control, weight reduction, lipid regulation, and cardiorenal protection.

Recent clinical guidelines have updated T2DM treatment strategies to reflect the central role of IBTs. The Guidelines for the Prevention and Treatment of Diabetes Mellitus in China (2024 Edition) no longer list metformin as an absolute first-line therapy. Instead, they propose a risk-based hierarchical approach, recommending GLP-1RAs as first-line options for T2DM patients with established ASCVD (or high risk), CKD, overweight, obesity or MASLD—especially those with a history of ischemic stroke (). Similarly, the 2025 American Diabetes Association (ADA) Standards of Medical Care in Diabetes emphasize patient-centered, individualized treatment strategies, prioritizing GLP-1RAs or GIP/ GLP-1 dual agonists for T2DM patients with MASLD, overweight, obesity or metabolic dysfunction-associated steatohepatitis (MASH). Furthermore, GLP-1RAs are strongly recommended for individuals with CKD and eGFR < 30 mL/min/1.73 m2 due to their favorable safety profile and proven cardiovascular benefits ().

IBTs vary in molecular structure, medication regimens, and administration cycles. With the continuous development of these drugs, there remains a lack of comprehensive and unified evaluations regarding their comparative efficacy, safety profiles, and applicable populations across different clinical scenarios. Most previous studies have only analyzed a single class of drugs, routine indicators, or individual outcome measures. For instance, a network meta-analysis conducted by Wu et al., which included 360 randomized controlled trials (RCTs) involving a total of 157,696 patients, confirmed that incretin drugs can significantly improve HOMA-βand HOMA-IR (). Stefanou et al. performed a systematic review and meta-analysis of 16 RCTs with 28,168 patients, verifying that GLP-1 RAs and dual GIP/GLP-1 RAs can reduce the incidence of MACE and all-cause mortality (). A meta-analysis by Mendonça et al., based on 8 RCTs enrolling 68,572 patients, found that GLP-1 RAs could lower the risk of renal function deterioration by 16% (). Furthermore, in clinical practice and previous trials, incretin drugs are frequently combined with other oral antidiabetic drugs (OADs) or insulin. However, prior meta-analyses failed to thoroughly explore the impacts of background medications and active control groups. Waldrop et al. integrated trial data from 2008 onwards, stratified the results according to control regimens such as placebo and monotherapy, and comprehensively assessed the hypoglycemic and weight-loss effects of the drugs. They also analyzed the influences of baseline combined use of OADs and the intragroup efficacy differences within the same drug classes on glycemic control (). In contrast, the present study included 15 types of incretin drugs. It comprehensively analyzed pancreatic function indicators (HOMA-β, HOMA-IR, and fasting C-peptide) as well as cardiac and renal function indicators (MACE, eGFR, and UACR). The study also updated data after 2016 and conducted stratified analyses on treatment regimens and control regimens in the subgroup analysis, addressing the lack of comprehensive multi-dimensional and multi-target evaluation in existing research.

As a robust tool for synthesizing evidence from multiple trials, network meta-analysis enables both direct and indirect comparisons of interventions, facilitating comprehensive efficacy-safety ranking. In accordance with the PRISMA 2020 and PRISMA-NMA guidelines, this study constructed a systematic literature search and rigorous screening of RCTs evaluating 15 IBTs for T2DM. It comprehensively evaluated their performance in key domains including glycemic control, weight management, lipid and blood pressure regulation, cardiorenal protection, and insulin function improvement, while also exploring the impacts of dose, treatment duration and combination therapy on outcomes. Based on receptor targeting profiles, the 15 agents were categorized into three main groups: GLP-1RAs (semaglutide, liraglutide, dulaglutide, exenatide, lixisenatide, orforglipron, polyethylene glycol loxenatide, efpeglenatide, ITCA 650); dual receptor agonists, including the GIP/GLP-1R agonist (tirzepatide) and the GLP-1R/GCGR agonist (mazdutide, survodutide); and the triple GIP/GLP-1R/GCGR agonist (retatrutide). The primary objective is to provide high-quality, actionable evidence to inform guideline-concordant clinical decision-making and optimize the benefit-risk balance in the management of T2DM.

2 Methods

The protocol of this systematic review and network meta-analysis was registered with PROSPERO (registration number: CRD420251152746). This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 edition () and the PRISMA Extension Statement for Network Meta-Analyses (PRISMA-NMA) ().

2.1 Search strategy

A comprehensive literature search was conducted in Medline, the Cochrane Library, Embase, and selected Chinese databases to identify RCTs evaluating IBTs in T2DM patients, from database inception to 1 August 2025. An additional supplementary search focusing on RCTs involving cardiorenal outcomes and insulin function parameters was performed on 11 September 2025, to minimize the risk of missing relevant studies. Three reviewers independently conducted the search and screening processes, with discrepancies resolved through consultation with a fourth reviewer. Furthermore, manual searches were carried out on the official websites of the European Association for the Study of Diabetes (EASD) and the ADA scientific meetings, as well as on ClinicalTrials.gov. The reference list of included studies and previously published systematic reviews were also screened to identify potentially eligible trials.

The search terms of this study are as follows. (“tirzepatide” OR “orforglipron” OR “semaglutide” OR “retatrutide” OR “liraglutide” OR “dulaglutide” OR “exenatide” OR “lixisenatide” OR “mazdutide” OR “survodutide” OR “visepegenatide” OR “polyethylene glycol loxenatide” OR “efpeglenatide” OR “ITCA 650″) AND (“type 2 diabetes” OR “type 2 diabetes mellitus” OR “T2DM”) AND (“randomized controlled trial” OR “RCT” OR “randomized” OR “placebo-controlled”). Complete search strategies are provided in Supplementary Appendix 1.

2.2 Eligibility criteria

Studies that enrolled adult participants with T2DM and its associated complications were considered eligible if they involved treatment with IBTs and included a minimum follow-up duration of 12 weeks. Interventions consisted of either IBTs monotherapy or combination therapy with non-randomized background antihyperglycemic agents. Control groups received placebo, a different IBTs, or other antidiabetic drugs. Only peer-reviewed, published RCTs were included; studies presented solely as conference abstracts, crossover designs, and trials involving discontinued or withdrawn drugs were excluded.

2.3 Screening process

The retrieved database records were imported into EndNote X9 to remove duplicates and were then merged with results from other sources for comprehensive comparison. The screening process consisted of three stages. Firstly, three reviewers independently screened articles based on titles, retaining any studies with uncertain eligibility for further assessment. Secondly, all articles retained in the initial stage underwent abstract review, during which discrepancies among reviewers were resolved through consensus discussions, with arbitration by a fourth reviewer when necessary. Thirdly, full-text version of articles that met the preliminary inclusion criteria based on titles and abstracts were assessed against predefined eligibility criteria to determine final inclusion.

The criteria for full-text review were as follows: the outcome indicators of the article met the study’s requirements, there was no substantial missing data, double-blind design RCTs were preferred, and open-label trials were eligible for inclusion if their outcome indicators fulfilled the study’s objectives.

2.4 Data extraction

For each eligible study, data were independently extracted by two reviewers using a standardized, pre-designed form. The extracted information included study characteristics (e.g., publication year, treatment duration), study population demographics (e.g., age, gender, sample size, diabetes duration, body mass index), intervention details (e.g., agent name and dosage), and outcome indicators. The following efficacy outcomes were collected: mean change from baseline in glycated hemoglobin (HbA1c), change in fasting plasma glucose (FPG) concentration from baseline, proportion of patients achieving glycemic targets of HbA1c <7.0% and <6.5%, change in body weight from baseline, as well as changes in blood pressure systolic blood pressure (SBP) and diastolic blood pressure (DBP)) and serum lipid profiles (such as high-density lipoprotein (HDL), low-density lipoprotein (LDL), total cholesterol (TC), and triglyceride (TG) levels). In addition, key clinical endpoints were extracted to assess cardiorenal effects and insulin function, including all-cause mortality (ACM), non-fatal stroke (NFS), non-fatal myocardial infarction (NFM), cardiovascular mortality (CVM), major adverse cardiovascular events (MACE), estimated glomerular filtration rate (eGFR), urinary albumin/creatinine ratio (UACR), fasting insulin level (FIL), homeostasis model assessment of β-cell function (HOMA-β), homeostasis model assessment of insulin resistance (HOMA-IR), and C-peptide levels during treatment. Safety data were comprehensively evaluated by extracting all reported adverse events without predefined restrictions. For studies presenting data graphically, numerical values were estimated using OriginPro software. All extractions were performed by two independent reviewers and subsequently verified, with discrepancies resolved through discussion or adjudication by a third reviewer. Baseline characteristics of the included RCTs are summarized in Supplementary Appendix 3.

2.5 Quality assessment of evidence

The Cochrane Risk of Bias Tool for Randomized Trials (RoB 2.0) was used to assess the risk of bias in all included trials across the following domains: random sequence generation, allocation concealment, blinding, incomplete outcome data, and selective reporting of results (). If the risk of bias was judged to be low in all domains, the trial was assigned an overall low risk of bias (score 1). If the risk of bias was high in at least one domain, the trial was classified as having a high overall risk of bias (score 3). In all other cases--such as when some concerns were identified in one or more domains--the overall risk of bias was rated as having some concerns (score 2). Risk of bias assessments were conducted independently by two reviewers, with disagreements resolved by consensus discussion. To quantitatively assess funnel plot symmetry, Egger’s regression test and Begg’s rank correlation test were performed (). Funnel plots were generated using estimates from direct evidence to visually inspect small-study effects, with each comparison evaluated separately. We further applied the Confidence in Network Meta-Analysis (CINeMA) framework to comprehensively assess the certainty of evidence generated by network meta-analysis, covering six core key domains: within-study bias, reporting bias, indirectness, imprecision, heterogeneity, and inconsistency. To evaluate intransitivity (a key component of indirectness assessment in the CINeMA framework), we systematically compared potential effect modifiers, including baseline age, baseline HbA1c concentration, diabetes duration, and body mass index (BMI), between studies that provided direct and indirect evidence for each pair of treatment comparisons. In accordance with standardized CINeMA assessment criteria, each domain of the CINeMA assessment was rated on a hierarchical scale (very low, low, moderate, high), and the overall certainty of evidence for each primary and secondary outcome measure was synthesized based on the ratings of the individual domains. The CINeMA assessment was conducted by two independent reviewers, with any discrepancies resolved through discussion and, if necessary, adjudicated by a third senior reviewer.

2.6 Data synthesis and analysis

We employed an evidence-based evaluation framework integrating Bayesian network meta-analysis and frequentist methods, using R software (version 4.4.3) as the analytical platform. A comprehensive analytical system was constructed using the GeMTC, R JAGS, and netmeta packages. Within the Bayesian framework, non-informative priors, a consistent random-effects model, and a normal likelihood-identity link function were set (). Parameter estimation was performed via the Markov Chain Monte Carlo (MCMC) algorithm, and model convergence was assessed using the Potential Scale Reduction Factor (PSRF), along with trace plots and density plots (). Mean differences (MD) were used for continuous outcomes description; for binary outcomes, odds ratios (OR) were used to quantify effect sizes. Relative effects of interventions were presented through league tables and forest plots, while efficacy ranking was conducted using rank probability analysis and Surface Under the Cumulative Ranking (SUCRA) values. Network consistency was evaluated using the node-splitting method (). Heterogeneity was assessed using the Bayesian I2 statistic--the proportion of between-study variance to total variance (). Missing data were supplemented by searching clinical trial registries or referring to comparable studies. Additionally, network meta-regression analyses were conducted to incorporate baseline characteristics (e.g., disease duration, age, glycosylated hemoglobin, etc.) and intervention-related factors (e.g., dosage, follow-up duration, concomitant treatment regimens) to explore potential effect modifiers, and subgroup analyses were performed according to the above intervention factors. Under the frequentist framework, the netmeta package was utilized to assess publication bias through Egger’s test and corrected funnel plots.

3 Results

3.1 Literature selection and study characteristics

A total of 21,041 citations were identified through the initial literature search, of which 462 full-text articles were assessed for eligibility. The literature search was updated to ensure inclusion of the most current evidence. Based on the predefined inclusion criteria, 102 RCTs, involving 98,693 adult participants, were included in the analyses (Figure 1). The sample size of individual studies ranged from 43 to 14,752 participants. Intervention duration varied from 12 to 280 weeks. The mean age across studies was 57.13 years (standard deviation [SD] 9.86), with a mean proportion of male participants at 45.29%. The mean duration of diabetes was 8.57 years (SD 6.52). At baseline, the mean body mass index was 31.30 kg/m2 (SD 6.47), and mean HbA1c level was 8.70% (SD 0.98) (Supplementary Appendix 2). No evidence of asymmetry in the funnel plot was observed, suggesting low risk of publication bias (Supplementary Appendix 9).

FIGURE 1

3.2 Risk of bias, certainty of evidence, and consistency

The risk of bias for each included trial is presented in Supplementary Appendix 4. A key limitation identified was incomplete reporting of blinding procedures for participants, researchers, and outcome assessors. Among the 102 trials, all were assessed as having a low risk of bias in random sequence generation, 101 (99.01%) in deviation from intended interventions, 101 (99.01%) in missing outcome data, and 92 (90.20%) in outcome measurement. All 102 studies (100%) exhibited a low risk of selective reporting bias. Overall, only one study (0.98%) was judged to have a high overall risk of bias.

Although some comparisons showed potential inconsistency, for most outcomes there was no statistically significant evidence of network inconsistency. Given that this review strictly included RCTs according to predetermined eligibility criteria, inter−study heterogeneity was generally low. Heterogeneity test using the I2 statistic revealed no substantial heterogeneity across the network, with most comparisons showing low or low−to−moderate levels of heterogeneity (Supplementary Appendix 5).

The results of the evidence certainty assessment based on the CINeMA framework show that after assessing the level of evidence using CINeMA, most of the results of the pairwise comparisons were of moderate or high confidence (Supplementary Appendix 10). All networks met the principle of transitivity, endowing the validity of indirect comparisons (Supplementary Appendix 10, Supplementary Table S10.1).

3.3 Sensitivity analyses and meta−regressions

To further evaluate the robustness of the results, sensitivity analyses were conducted by sequentially excluding all previously included trials comparing IBTs with other classes of antihyperglycemic−agents −− retaining only RCTs with placebo as control −− along with additional assessments stratified by dosage, exclusion of open−label trials, and removal of RCTs employing combination therapy designs. As illustrated in Supplementary Appendix 10, the results of sensitivity analysis were consistent with the primary outcomes, thereby supporting the stability and reliability of the study findings.

Meta−regression analysis was performed to assess the potential impact of baseline factors on the main outcomes, including age, proportion of male participants, duration of diabetes, BMI, baseline HbA1c, and combination therapy. The analysis revealed that combination therapy was significantly associated with greater treatment effects on the primary outcome measure (P < 0.05; see Supplementary Appendix 11).

3.4 Glycemic control

The network meta−analysis on HbA1c reduction included 102 trials involving 98,693 participants. Compared with placebo, all 15 IBTs demonstrated statistically significant efficacy in reducing HbA1c levels in adult T2DM patients (Figure 2). Tirzepatide exhibited the greatest effect size, with a MD of −2.30% ([95% CrI −2.70 to −1.80], SUCRA 99.72%), followed by orforglipron (MD −1.80% [95% CrI −2.70 to −0.96], SUCRA 91.71%). Full ranking results based on SUCRA values are shown in Supplementary Appendix 7, Supplementary Table S7.1.

FIGURE 2

As assessed by FPG, the network meta−analysis included 87 RCTs involving 45,510 participants, confirming the placebo−controlled efficacy of all 12 IBTs (Figure 3). Tirzepatide showed the greatest reduction, with a MD of −3.30 mmol/L ([95% CrI −4.00 to −2.60], SUCRA 98.73%), followed by orforglipron (MD −2.60 mmol/L [95% CrI −4.20 to −1.10], SUCRA 87.54%). Complete ranking results based on SUCRA are presented in Supplementary Appendix 1, Supplementary Table S7.2.

FIGURE 3

For the proportion of T2DM patients achieving the HbA1c target of 7.0%, the network meta−analysis included 73 RCTs involving 41,660 participants, demonstrating the statistically significant efficacy of all 12 IBTs compared with placebo (Supplementary Appendix 6, Supplementary Table 6.1). Orforglipron showed the greatest effect, with a MD of 59.00% ([95% CrI 49.00 to 69.00], SUCRA 96.03%), followed by tirzepatide (MD 58.00% [95% CrI 52.00 to 65.00], SUCRA 95.64%). (The SUCRA ranking is shown in the Supplementary Appendix 7, Supplementary Table 7.4).

Regarding the proportion of T2DM patients achieving the more stringent HbA1c target of 6.5%, the analysis included 67 RCTs involving 37,559 participants, confirming the placebo−controlled efficacy of all 12 IBTs (Supplementary Appendix 6, Supplementary Table 6.2). Tirzepatide ranked first (MD 60.00% [95% CrI 53.00 to 66.00], SUCRA 96.41%), followed by orforglipron (MD 59.00% [95% CrI 49.00 to 70.00], SUCRA 95.80%). Full SUCRA−based rankings are presented in Supplementary Appendix 7, Supplementary Table 7.5.

3.5 Weight loss

This analysis included 98 trials involving 54,156 participants for body weight change. Retatrutide demonstrated the greatest efficacy in weight reduction, with a MD of −17.00 kg ([95% CrI −20.00 to −14.00], SUCRA 99.81%), and was ranked as the most effective IBTs (Figure 4). Compared with placebo, tirzepatide, orforglipron, mazdutide, semaglutide, and survodutide also showed significant reductions in body weight. Complete ranking results based on SUCRA values are provided in Supplementary Appendix 7, Supplementary Table S7.3.

FIGURE 4

3.6 Lipid profiles

For HDL elevation, liraglutide demonstrated the greatest effect, with a MD of 0.03 mmol/L ([95% CrI −0.003 to 0.06], SUCRA 90.96%). For LDL−C reduction, visepegenatide showed the largest decrease (MD −0.31 mmol/L [95% CrI −0.78 to 0.16], SUCRA 85.93%). For TC reduction, exenatide was the most effective (MD −0.24 mmol/L [95% CrI −0.33 to −0.16], SUCRA 93.33%). For TG reduction, visepegenatide again ranked highest (MD −0.31 mmol/L [95% CrI −0.78 to 0.16], SUCRA 84.07%. Detailed results are summarized in Supplementary Appendix 6, Supplementary Tables S6.2–S6.6. Complete SUCRA−based rankings are presented in Supplementary Appendix 7, Supplementary Tables S7.6–S7.9.

3.7 Blood pressure

For SBP reduction, eight IBTs demonstrated statistically significant efficacy. Orforglipron showed the greatest effect, with a MD of −8.50 mmHg ([95% CrI −10.00 to −6.90], SUCRA 99.99%), followed by liraglutide (MD −3.50 mmHg [95% CrI −4.50 to −2.40], SUCRA 77.66%). For DBP reduction, visepegenatide was the most effective (MD −2.10 mmHg [95% CrI −5.20 to 0.98], SUCRA 79.08%). Detailed results are provided in Supplementary Appendix 6, Supplementary Tables S6.7, S6.8. Complete SUCRA−based rankings are shown in Supplementary Appendix 7, Supplementary Tables S7.10, S7.11.

3.8 MACE

Tirzepatide demonstrated the most favorable overall profile, showing the greatest efficacy across four outcomes: MACE (OR 0.57 [95% CrI 0.39 to 0.79], SUCRA 96.21%), non−fatal myocardial infarction (OR 0.62 [95% CrI 0.33 to 1.1], SUCRA 88.39%), cardiovascular death (OR 0.64 [95% CrI 0.10 to 3.7], SUCRA 67.16%), and all−cause mortality (OR 0.58 [95% CrI 0.08 to 3.8], SUCRA 73.65%). It also ranked highly for non−fatal stroke (OR 0.69 [95% CrI 0.28 to 1.60], SUCRA 69.62%). Subgroup analyses indicated that ITCA650 did not demonstrate statistically significant benefits for MACE, cardiovascular death, and all−cause mortality. Similarly, lixisenatide did not show significant efficacy in reducing non−fatal stroke and non−fatal myocardial infarction. Detailed results are presented in Supplementary Appendix 6, Supplementary Tables S6.9–S6.13. Complete SUCRA−based rankings are provided in Supplementary Appendix 7, Supplementary Tables 7.12–S7.16.

3.9 Renal function

Regarding the changes in eGFR, polyethylene glycol loxenatide showed the greatest improvement compared with the placebo (MD 2.30 mL/min/1.73 m2 [95% CrI −4.50 to 9.20]). For UACR reduction, exenatide demonstrated the most favorable effect (MD −17.00 mg/g [95% CrI −43.00 to −8.20]). Detailed results are provided in Supplementary Appendix 6, Supplementary Tables S6.14, S6.15.

3.10 Function of insulin

For the outcome of reduced FIL levels, tirzepatide showed the greatest efficacy (MD −23.00 pmol/L [95% CrI −32.00 to −14.00], SUCRA 98.55%). For reduced C−peptide levels, tirzepatide again the most effective (MD −0.20 ng/mL [95% CrI −0.36 to −0.05], SUCRA 97.33%). For improvement in insulin resistance as measured by HOMA−IR, liraglutide ranked highest (MD −0.58 [95% CrI −0.90 to −0.25], SUCRA 82.94%). For enhancement of beta−cell function assessed by HOMA−β, tirzepatide demonstrated optimal performance (MD 31.00 [95% CrI 21.00 to 41.00], SUCRA 86.86%). Detailed results are summarized in Supplementary Appendix 6, Supplementary Tables S6.16–S6.19. Complete SUCRA−based rankings are provided in Supplementary Appendix 7, Supplementary Table S7.1.

3.11 Adverse events

All evaluated IBTs were associated with a statistically significant increase in the risk of gastrointestinal adverse events compared with placebo. For nausea incidence, ITCA650 had the highest risk (OR 6.70 [95% CrI 2.50 to 18.00]), followed by visepegenatide (OR 5.30 [95% CrI 2.30 to 12.00]), polyethylene glycol loxenatide (OR 4.90 [95% CrI 1.60 to 17.00]) and liraglutide (OR 3.60 [95% CrI 2.80 to 4.80]). In terms of the incidence of vomiting, ITCA650 had the highest risk (OR 12.00 [95% CrI 5.50 to 27.00]), followed by visepegenatide (OR 4.20 [95% CrI 2.00 to 9.60]), polyethylene glycol loxenatide (OR 3.60 [95% CrI 1.30 to 11.00]). In terms of the incidence of diarrhea, mazdutide had the highest risk (OR 4.70 [95% CrI 2.90 to 7.80]), and the risks of tirzepatide (OR 2.30 [95% CrI 1.80 to 2.90]) and retatrutide (OR 2.40 [95% CrI 1.30 to 4.20]) were similar. In terms of the incidence of constipation, retatrutide had the highest risk (OR 4.40 [95% CrI 1.90 to 9.90]), followed by tirzepatide (OR 2.60 [95% CrI 1.50 to 4.50]) (Supplementary Appendix 6, Supplementary Tables S6.20–S6.23).

3.12 Subgroup analyses of each GLP−1RA by doses and administration regimen

All IBTs at varying dosages significantly reduced HbA1c, FBG, and body weight: retatrutide 12 mg achieves the greatest FBG reduction (MD −3.20 mmol/L [95% CrI −4.30 to −2.10]) and weight loss (MD −15.00 kg [95% CrI −17.00 to −13.00]), while tirzepatide 15 mg induces the most notable HbA1c decrease (MD −2.00% [95% CrI to 2.20 to −1.80]). Most IBTs are linked to increased gastrointestinal adverse events: efpeglenatide 12 mg (MD 44.00% [95% CrI 28.00 to 61.00]) and orforglipron 24 mg (MD 28.00% [95% CrI: 18.00 to 39.00]) carry the highest nausea risk; efpeglenatide 14 mg (MD 28.00% [95% CrI 12.00 to 45.00]) and survodutide 1.8 mg (MD 19.00% [95% CrI 8.60 to 29.00]) present notable vomiting risk; mazdutide 4.5 mg (MD 32.00% [95% CrI 24.00 to 40.00]) and 3 mg (MD 28.00% [95% CrI 20.00 to 36.00]) pose higher diarrhea risk; orforglipron 24 mg (MD 13.00% [95% CrI 7.20 to 20.00]) and 12 mg (MD 11.00% [95% CrI 4.50 to 17.00]) have elevated constipation risk (see Supplementary Appendix 12).

3.13 Subgroup analyses of individual GLP−1RA by treatment duration

All IBTs achieved significant HbA1c reduction across all durations, with tirzepatide exhibiting optimal long−term efficacy (MD −1.80% [95% CrI −2.10 to −1.50]). Most IBTs also yielded significant FPG reduction, with dulaglutide demonstrating the best mid−term efficacy (MD −4.40 mmol/L [95% CrI −7.70 to −1.10]) most IBTs induced significant weight loss, with tirzepatide exerting the strongest long−term effect (MD −9.90 kg [95% CrI −12.00 to −7.90]); regarding safety, most IBTs elevated the risks of nausea, vomiting, diarrhea and constipation across durations, with semaglutide having the highest long−term nausea risk (OR 11.00 [95% CrI 6.20 to 22.00]) and mid−term vomiting risk (OR 8.30 [95% CrI 5.60 to 12.00]), and liraglutide showing the greatest long−term risks of diarrhea (OR 26.00 [95% CrI 8.60 to 85.00]) and constipation (OR 22.00 [95% CrI 12.00 to 51.00]) (see Supplementary Appendix 13).

3.14 Subgroup analysis of combination therapy regimens

All IBTs combination regimens achieved significant reductions in HbA1c, with the optimal hypoglycemic effect observed in the regimens of tirzepatide 15 mg (MD −2.80% [95% CrI −3.40 to −2.20]) and 10 mg (MD −2.80% [95% CrI −3.30 to −2.20]) in combination with an oral hypoglycemic agent at a stable dose; regimens involving “combination with insulin” or “high−dose monotherapy” generally exerted more pronounced HbA1c−lowering effects (e.g., tirzepatide 15 mg plus insulin: MD −2.40%; semaglutide 1 mg monotherapy: MD −1.90%); FPG reduction, most regimens had significant effects, with the liraglutide 1.8 mg plus multiple antidiabetic drugs regimen being the most effective (MD −2.90 mmol/L [95% CrI −5.30 to −0.46]), high−dose tirzepatide (15 mg) across all regimens consistently demonstrating substantial FPG−lowering potential (monotherapy: MD −2.50 mmol/L; plus insulin: MD −2.40 mmol/L); in terms of weight reduction, most regimens produced significant effects, with the tirzepatide 15 mg plus stable oral antidiabetic drugs regimen showing the strongest efficacy (MD −10.00 kg [95% CrI −13.00 to −7.80]), high−dose tirzepatide (15 mg) across all regimens exhibiting marked weight−lowering effects (monotherapy: MD −9.40 kg; plus insulin: MD −8.10 kg) and the semaglutide 1 mg plus insulin regimen also having robust efficacy (MD −5.00 kg [95% CrI −7.00 to −3.00]), while the liraglutide 0.6 mg monotherapy regimen even showed a trend toward weight gain (MD 0.37 kg [95% CrI −2.10 to 2.90]) (see Supplementary Appendix 14).

4 Discussion

This network meta-analysis shows that IBTs deliver comprehensive benefits for T2DM. Tirzepatide, orforglipron and semaglutide achieve excellent glycemic control, while retatrutide is superior in weight reduction. Their effects on blood lipids and blood pressure vary by drug: liraglutide raises HDL, visepegenatide lowers LDL-C and TG, and exenatide reduces TC; orforglipron works best for SBP reduction, and visepegenatide is optimal for DBP management. In addition, tirzepatide exerts the strongest cardioprotective effects, and all IBTs have renal protective properties—polyethylene glycol loxenatide improves eGFR and exenatide decreases UACR. Tirzepatide and liraglutide also help optimize insulin function. Common adverse reactions included mild pulse rate elevation, nasopharyngitis, constipation, and injection site reactions (exenatide, lixisenatide, etc.) () Low-titer anti-drug antibodies were detected for some agents without impairing efficacy (). Hypoglycemia risk was low overall. Rare adverse events involved acute kidney injury (liraglutide, exenatide) (), cholelithiasis and malignant neoplasms (semaglutide, liraglutide, dulaglutide) (; ), and diabetic ketoacidosis (retatrutide) (). Despite minor agent-specific differences in adverse reaction incidence and manifestations, the overall safety profiles of IBTs remained consistent. In addition, safety shows clear dose-response relationships—higher doses enhance efficacy but increase adverse event risks, with orforglipron 45 mg achieving an optimal benefit-tolerability balance, while semaglutide, dulaglutide, and other IBTs also exhibit dose-dependent efficacy and formulation-related disparities (e.g., oral orforglipron 36 mg matches injectables in HbA1c or weight loss effects, and 45 mg reduces FBG close to injectables). Treatment duration impacts outcomes: most agents improve efficacy long-term, with divergent trends in gastrointestinal adverse events (e.g., exenatide lowers nausea or vomiting risks, tirzepatide reduces diarrhea risk) and varying short-/mid-/long-term efficacy across agents, while IBTs combination therapies yield drug-specific synergistic effects superior to monotherapy, further supporting individualized selection based on patient comorbidities (e.g., obesity, cardiorenal disease) to maximize therapeutic benefits while ensuring safety.

From the perspective of pathophysiological mechanisms, the differences in therapeutic efficacy among various incretin drugs can be partially attributed to the synergistic effects of receptor co-activation. Although the insulinotropic effect of GIP in patients with T2DM is significantly diminished due to impaired β-cell function (), resulting in limited hypoglycemic efficacy when used alone, complementary benefits are generated when it is co-activated with GLP-1RAs (). As a dual GIP/GLP-1RAs, tirzepatide acts on two incretin signaling pathways simultaneously. It not only promotes insulin secretion in a glucose-dependent manner but also synergistically enhances satiety, suppresses appetite, and improves lipid metabolism. Furthermore, the addition of glucagon receptor agonism forms a triple receptor co-activation strategy. Against the backdrop of hypoglycemic and weight-lowering effects mediated by GLP-1RAs, moderate activation of glucagon receptors yields additional metabolic benefits by boosting energy expenditure, enhancing lipolysis, and improving hepatic insulin sensitivity (). Retatrutide, as a GIP/GLP-1/GCGR triple receptor agonist, relies precisely on this mechanism to demonstrate superimposed advantages in weight loss and glycemic control (). While glucagon activation itself does not directly lower blood glucose, its indirect effects mediated by improving insulin resistance and promoting adipose tissue remodeling produce synergistic benefits with the GIP and GLP-1 pathways (). Therefore, the efficacy gradient among different drugs observed in this study essentially reflects the pathophysiological progressive logic from single-receptor activation to dual-receptor and even triple-receptor activation.

By comprehensively evaluating the benefit-risk profiles of various incretin-based drugs, this study found that although medications including tirzepatide, orforglipron and semaglutide statistically significant advantages in reducing HbA1c, their incidence of gastrointestinal adverse events (nausea, vomiting, diarrhea, constipation) increases correspondingly with a clear dose-response relationship. Specifically, as the administered dose escalates, the hypoglycemic efficacy is further improved, while the risk and severity of adverse events such as nausea and vomiting rise simultaneously, directly leading to treatment discontinuation in some patients. Subgroup analysis further confirmed that high-dose regimens (e.g., orforglipron 24 mg, efpeglenatide 12 mg) can achieve superior glycemic control but are associated with markedly reduced gastrointestinal tolerability and an elevated discontinuation rate. Notably, orforglipron 45 mg achieves a relatively optimal balance between efficacy and tolerability, whereas some drugs with high SUCRA rankings do not demonstrate an equivalent level of overall acceptability. This finding suggests that in real clinical practice, selecting the so-called “optimal” drugs solely based on the reduction in glycated hemoglobin or SUCRA ranking may overlook the critical impact of patient tolerability on treatment persistence. If patients discontinue medication early due to intolerance to gastrointestinal reactions, the net clinical benefit of drugs with potent hypoglycemic effects demonstrated in clinical trials will be greatly diminished in practical application. Therefore, clinical decision-making should comprehensively weigh hypoglycemic efficacy, dose-dependent side effects and anticipated discontinuation risks, and formulate individualized treatment regimens based on patients’ personal tolerability, comorbidities and dose titration strategies, rather than selecting drugs simply according to efficacy rankings.

In the meta-regression analysis of this study, diabetes duration and baseline HbA1c level were identified as important effect modifiers influencing treatment effects. The analysis results indicated that patients with longer disease duration and higher baseline glycated hemoglobin levels achieved a more significant absolute reduction in glycated hemoglobin after treatment with incretin-based medications, suggesting that patients with a heavier baseline disease burden may obtain greater net glycemic-lowering benefits. However, it is noteworthy that such patients often have a prolonged disease course, a heavier burden of complications, and may be accompanied by multiple cardiovascular and metabolic abnormalities simultaneously. They may also have lower gastrointestinal tolerance to high-dose regimens, which may affect treatment persistence and overall benefits. Therefore, when extrapolating the efficacy conclusions of this study to different clinical subgroups, full consideration should be given to the heterogeneity of baseline characteristics.

Existing systematic reviews and meta-analyses are often restricted to limited number of agent types or focus on single outcomes, thereby failing to address the guideline-recommended multi-target therapeutic goals or adequately support individualized treatment strategies for patients with multiple comorbidities. This network meta-analysis has distinct strengths. First, it boasts an exceptionally comprehensive scope, solving problems such as the single research index and few types of drugs in similar studies. Second, stratified subgroup analyses of dosages, treatment courses, and combination therapies were conducted to address individualized clinical needs, clarifying efficacy-safety balancing strategies. Third, outcome measures spanned glycemic control, weight regulation, lipid metabolism, cardiovascular events, and renal function, enabling a holistic assessment of drug benefits. Fourth, strict adherence to PRISMA 2020 and PRISMA-NMA guidelines, PROSPERO registration, rigorous bias risk assessment, and robustness verification collectively ensure high result reliability. Notwithstanding these merits, the study has several limitations. The paucity of RCTs for some IBTs restricts analyzable indicators, and scarce head-to-head trials between specific agents necessitate heavy reliance on indirect evidence. The predominantly adult cohort excludes pediatric, adolescent, and severely comorbid patients, limiting conclusion generalizability. Additionally, data on rare adverse events and long-term efficacy stability beyond 2 years remain insufficient; subjective outcomes (e.g., quality of life) were not assessed, constraining comprehensive clinical decision support. Notably, unanalyzed agents do not equate to ineffectiveness but reflect evidence gaps warranting further trials.

Clinical trial data for some novel drugs remain relatively limited. In such cases, high SUCRA rankings should be interpreted with caution. As a probabilistic ranking indicator, the stability of SUCRA is highly dependent on the quality and precision of the evidence included in the analysis. Small sample sizes, large variances or sparse network structures in primary studies can render ranking results vulnerable to random fluctuations, model assumptions and potential biases, which may lead to inflated rankings and overestimation of the actual efficacy of drugs. Therefore, for novel drugs with insufficient supporting evidence, SUCRA analysis results can only serve as exploratory conclusions rather than definitive judgments.

Given the exploratory nature of network meta-analysis, and the fact that most comparisons rely on indirect evidence, surrogate endpoints or lack of direct head-to-head trials, the drug treatment recommendations based on different comorbidities presented in this study should be regarded as hypothesis-generating preliminary findings instead of definitive clinical guidance. For rankings derived from limited evidence or a small number of studies, the seemingly superior efficacy of certain drugs may not be robust and needs to be verified in well-powered randomized controlled trials with direct comparisons.

The preliminary findings are summarized as follows: Among obese patients, retatrutide exhibits the best weight loss effect, yet direct comparisons with other dual or triple agonists are lacking. In terms of dyslipidemia, the rankings indicating that liraglutide increases HDL, visepegenatide reduces LDL-C and TG, and exenatide lowers TC are all based on limited evidence, with visepegenatide having the weakest evidence base and its results deemed exploratory. For hypertension management, the findings that orforglipron reduces systolic blood pressure and visepegenatide decreases diastolic blood pressure are solely derived from indirect comparisons and surrogate endpoints, and require confirmation via dedicated direct antihypertensive studies. In patients with concomitant ASCVD, tirzepatide demonstrates superior benefits in MACE, while hard endpoint data for many other drugs remain inadequate. Among patients with CKD, the conclusions that pegylated lixisenatide improves eGFR and exenatide reduces UACR come from studies with small sample sizes and wide confidence intervals, and should be recognized as preliminary results. Regarding impaired insulin function, the benefits of tirzepatide on HOMA-β and liraglutide on HOMA-IR still need further validation using clinically relevant indicators. For patients with multiple comorbidities, tirzepatide or 45 mg orforglipron may be considered for their broad-spectrum efficacy, though direct comparisons with combination therapy regimens are unavailable. Clinicians shall combine the above exploratory findings with patients’ individual tolerability, treatment history and the strength of direct evidence for each drug. Dose titration, monitoring of gastrointestinal tolerance and lifestyle interventions remain essential components of clinical management. There is an urgent need to conduct head-to-head randomized controlled trials focusing on clinically relevant endpoints in the future to validate or refute these exploratory rankings. Furthermore, this study further defines the applicable populations for different incretin-based therapies. It is suggested that subsequent guidelines incorporate stratified recommendations for various preparations and clarify the preferred indications of each drug according to comorbidity types, so as to improve the clinical practicality of the guidelines. In terms of health policies, the medical insurance reimbursement policies for IBTs should be further optimized to improve drug accessibility, especially for oral IBTs and dual-receptor or triple-receptor agonists, so that more T2DM patients can receive high-quality glucose-lowering treatment. Meanwhile, efforts should be made to strengthen the construction of a monitoring system for adverse drug reactions of IBTs, establish a national adverse reaction database, and achieve early identification and warning of rare adverse reactions. In addition, more investment should be increased in clinical research on IBTs to support studies on special populations, long-term efficacy, and pharmacoeconomics, fill evidence gaps, and provide a more comprehensive evidence-based basis for the precise treatment of T2DM.

The findings of this network meta-analysis are generally consistent with major international guidelines for the management of T2DM, including the ADA 2025, EASD/ADA 2025 Consensus, IDF 2025 and CDS 2024. All these guidelines endorse the multi-dimensional benefits of IBTs in glycemic control, weight management and cardiorenal protection. Moreover, the efficacy of top-ranked agents such as tirzepatide provides quantitative evidence for the “cardiorenal priority” strategy recommended in the guidelines. Nevertheless, this study presents discrepancies from existing guidelines. It reveals that the high SUCRA rankings of several novel agents including orforglipron and retatrutide are derived from limited indirect evidence, suggesting that guidelines should interpret purely statistical rankings with caution. Additionally, the dose-subgroup analyses in this study, such as the favorable balance between benefits and tolerability of 45 mg orforglipron, as well as the substantial heterogeneity in gastrointestinal tolerability across different IBTs, call for the inclusion of more refined dose stratification and individualized tolerability recommendations in future guidelines.

5 Conclusion

In conclusion, this large-scale and comprehensive network meta-analysis confirms that IBTs yield significant multi-dimensional benefits in the treatment of T2DM, with marked heterogeneity in efficacy and safety among different drugs. Dosage, treatment duration and combination regimens exert important modulating effects on clinical outcomes. Nevertheless, the drug rankings and comparative results derived from this analysis are mainly based on indirect evidence and surrogate endpoints such as glycated hemoglobin, body weight and blood pressure. Some drugs are even evaluated relying on sparse data or limited head-to-head trials. Accordingly, the aforementioned rankings should be regarded as exploratory findings rather than definitive conclusions. In clinical practice, any preference for a specific IBT based on these results needs to be balanced against direct evidence, patients’ comorbidities, drug tolerability and shared decision-making. There is an urgent need for high-quality, long-term head-to-head randomized controlled trials focusing on hard clinical endpoints including MACE, mortality and renal failure, especially for newly developed drugs. Until such evidence becomes available, treatment decisions should prioritize a comprehensive assessment of the overall weight of both direct and indirect evidence and strictly adhere to the principle of individualized treatment.

Statements

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

Author contributions

XW: Data curation, Formal Analysis, Investigation, Methodology, Project administration, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. YY: Formal Analysis, Software, Validation, Visualization, Writing – review and editing. XC: Formal Analysis, Investigation, Writing – review and editing. XL: Conceptualization, Project administration, Validation, Writing – review and editing. YL: Conceptualization, Data curation, Funding acquisition, Resources, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This article has special funding support from the Shandong Provincial Health Commission Project (No. 2024YZ002).

Acknowledgments

We would like to express our gratitude to the teachers and students from the Department of Pharmacy, Shandong Provincial Qianfoshan Hospital, for their assistance in data extraction and technical guidance on data analysis for this study.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphar.2026.1846714/full#supplementary-material

Glossary

  • T2DM

    type 2 diabetes mellitus

  • GLP-1RA

    glucagon-like peptide-1 receptor agonist

  • GIP

    glucose-dependent insulinotropic polypeptide

  • GCGR

    glucagon receptor

  • GLP-1/GCGR

    GLP-1/glucagon receptor agonist

  • GIP/GLP-1RA

    GIP/GLP-1 receptor agonist

  • GIP/GLP-1/GCGR

    GIP/GLP-1/glucagon receptor agonist

  • IBTs

    incretin-based therapies

  • ASCVD

    atherosclerotic cardiovascular disease

  • CKD

    chronic kidney disease

  • MASLD

    metabolic dysfunction-associated steatotic liver disease

  • MASH

    metabolic dysfunction-associated steatohepatitis

  • ADA

    American Diabetes Association

  • RCTs

    randomized controlled trials

  • PRISMA

    Preferred Reporting Items for Systematic Reviews and Meta-Analyses

  • NMA

    network meta-analysis

  • PROSPERO

    International Prospective Register of Systematic Reviews

  • RoB 2.0

    Cochrane Risk of Bias Tool for Randomized Trials (version 2.0)

  • CINeMA

    Confidence in Network Meta-Analysis

  • SUCRA

    Surface Under the Cumulative Ranking

  • MCMC

    Markov Chain Monte Carlo

  • PSRF

    Potential Scale Reduction Factor

  • MD

    mean difference

  • OR

    odds ratio

  • CrI

    credible interval

  • SD

    standard deviation

  • HbA1c

    glycated hemoglobin

  • FPG

    fasting plasma glucose

  • SBP

    systolic blood pressure

  • DBP

    diastolic blood pressure

  • HDL

    high-density lipoprotein

  • LDL

    low-density lipoprotein

  • TC

    total cholesterol

  • TG

    triglyceride

  • ACM

    all-cause mortality

  • NFS

    non-fatal stroke

  • NFM

    non-fatal myocardial infarction

  • CVM

    cardiovascular mortality

  • MACE

    major adverse cardiovascular events

  • eGFR

    estimated glomerular filtration rate

  • UACR

    urinary albumin/creatinine ratio

  • FIL

    fasting insulin level

  • HOMA-β

    homeostasis model assessment of β-cell function

  • HOMA-IR

    homeostasis model assessment of insulin resistance and

  • EASD

    European Association for the Study of Diabetes

References

Summary

Keywords

adverse reactions, cardiorenal protection, efficacy, glucagon-like peptide-1 receptor agonists, network meta-analysis

Citation

Wu X, Yang Y, Cui X, Li X and Li Y (2026) Efficacy and safety of incretin-based therapies in patients with type 2 diabetes mellitus: a network meta-analysis based on clinical trials. Front. Pharmacol. 17:1846714. doi: 10.3389/fphar.2026.1846714

Received

03 April 2026

Revised

28 May 2026

Accepted

28 May 2026

Published

18 June 2026

Volume

17 - 2026

Edited by

Bernd Rosenkranz, Fundisa African Academy of Medicines Development, South Africa

Reviewed by

Theodoros Androutsakos, National and Kapodistrian University of Athens, Greece

Hai-Anh Ha, Da Nang University of Medical Technology and Pharmacy, Vietnam

Updates

Copyright

© 2026 Wu, Yang, Cui, Li and Li.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Yan Li,

†

ORCID: Wu Xichao, orcid.org/0009-0000-6298-7558; Li Yan, orcid.org/0000-0002-3966-1610

Disclaimer

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Outline

Figures

Cite article

Copy to clipboard


Export citation file


Share article

Share on WeChat

Scan with WeChat to share this article

Article metrics