Comparative outcomes of GLP-1 receptor agonists in glycemic and weight control

GLP-1RAs improve blood sugar control in T2DM, with tirzepatide being most effective for weight loss in obesity. In normal-weight patients, semaglutide and liraglutide are effective, with liraglutide offering a safer profile for reducing the risk of hypoglycemia.

Type 2 diabetes mellitus (T2DM) is a surging global health challenge, driven by insulin resistance and progressive β-cell impairment. The International Diabetes Federation predicts that over 780 million people worldwide will be affected by 2045, emphasizing the urgent need for effective therapeutic strategies. Glucagon-like peptide-1 (GLP-1), an intestinal hormone released after food intake, enhances insulin secretion and supports β-cell function to maintain glucose balance. Glucagon-like peptide-1 receptor agonists (GLP-1RAs) exploit this pathway to lower blood glucose by stimulating insulin release and suppressing glucagon.

According to the American Diabetes Association (ADA) and European Association for the Study of Diabetes (EASD), GLP-1RAs can be used both as add-on therapy to lifestyle modification and as an alternative when metformin monotherapy does not provide sufficient glycemic control. These agents are categorized by their duration of action into long-acting agents—such as semaglutide, tirzepatide, liraglutide, dulaglutide, albiglutide, and once-weekly exenatide (EQW)—and short-acting agents like twice-daily exenatide (EBID) and lixisenatide.

Evidences from clinical trials indicate that GLP-1RAs not only improve glycemic control but also aid weight reduction more successfully than traditional antidiabetic therapies. However, direct comparisons among these agents and against traditional antidiabetic drugs remain limited. Clarifying their relative efficacy and safety is essential for precision therapy, enabling clinicians to tailor treatments to patient profiles and optimize outcomes in T2DM.

Objective

Researchers evaluated and compared the clinical efficacy and safety of GLP-1 RAs, while also assessing their performance against traditional antidiabetic therapies through the largest network meta-analysis (NMA) to date.

The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Network Meta-Analyses (PRISMA-NMA) framework to ensure methodological rigor and transparency.

Literature search

A comprehensive systematic search was conducted across PubMed, Cochrane Library, Embase, Web of Science, and Chinese databases including China National Knowledge Infrastructure (CNKI), WangFang, and VIP (China Science and Technology Journal Database) up to October 2024.

Database searches were performed via a combination of terms such as: “Glucagon-Like Peptide-1 Receptor Agonists,” “Exenatide,” “Semaglutide,” “Dulaglutide,” “Liraglutide,” “Albiglutide,” “Tirzepatide,” “Lixisenatide,” “Type 2 Diabetes Mellitus,” and “Randomized Controlled Trials.”

Moreover, reference lists of all the eligible studies were manually screened to explore more suitable studies. Two independent reviewers executed the literature search and study selection, and any discrepancies were cleared via discussion and consensus.

Inclusion criteria

Studies were deemed eligible if they fulfilled the below conditions:

• Designed as RCTs.
• Conducted in adults aged 18 years or more with T2DM.
• Compared GLP-1 RAs with conventional antidiabetic medications or placebo.
• Documented at least 1 outcome pertinent to this analysis.
• Had an intervention duration of 2 months or longer.
• Published in English or Chinese.
• Provided adequate data for calculating relative risk (RR), odds ratio (OR), or mean difference (MD).

Two researchers independently screened and examined studies; disagreements were solved by consensus.

Exclusion criteria

• Publication type exclusions: Letters, replies, comments, opinions, conference abstracts, awarded grants, trial registry records, and errata.
• Study design exclusions: Reviews, meta-analyses, case reports, non-randomized controlled trials, and studies involving adolescents, children, or animals.
• Language/publication status exclusions: Unpublished studies or studies not in English or Chinese.
• Data availability exclusions: Studies lacking necessary data or where data could not be converted into a format suitable for NMA.

Search and study characteristics

• A comprehensive search identified 12,074 relevant articles across PubMed (n = 1596), Embase (n = 3496), Cochrane Library (n = 4639), Web of Science (n = 2265), and Chinese databases (n = 78).
• After removing duplicates, 8523 titles and abstracts were screened for eligibility.
• 64 full-text articles involving 25,572 patients were ultimately included for the NMA as presented in Figure 1.

Data extraction and handling

• Extracted data included study characteristics, participant demographics, baseline glycemic and metabolic parameters, intervention details, and outcomes.
• For continuous outcomes—including glycosylated hemoglobin (HbA1c), fasting plasma glucose (FPG), body weight, body mass index (BMI), blood pressure (systolic and diastolic), and lipid profiles (total cholesterol, low-density lipoprotein cholesterol [LDL-C], high-density lipoprotein cholesterol [HDL-C])—the authors recorded the number of participants per treatment arm along with mean values and standard deviations (SD) or standard errors (SE).
• For dichotomous outcomes, data were collected on the total number of participants and those experiencing safety events (such as nausea, vomiting, diarrhea, dyspepsia, constipation, decreased appetite, nasopharyngitis, headache, elevated lipase, and hypoglycemia).
• When continuous variables were reported as medians with interquartile ranges, they were converted to mean ± SD using validated statistical methods. Where necessary, corresponding authors were contacted to obtain missing or incomplete data.
• Traditional antidiabetic drugs were aggregated into a single compared group to enhance statistical power.

Quality assessment and statistical analysis

• Study quality was determined with the aid of the Cochrane Risk of Bias tool, with conflicts solved by consensus. Analyses were executed via Review Manager 5.4, Stata 14.0, and R.
• Continuous outcomes were analyzed with mean differences, and dichotomous outcomes with risk ratios (95% confidence intervals [CI]). Heterogeneity was checked via the Cochran Q test and I² statistic to guide model selection (random vs. fixed-effects).
• Frequency-based NMA allowed indirect comparisons, visualized via network plots. Consistency was determined via the design-by-treatment interaction model and deviance information criterion.
• Surface under the cumulative ranking curve (SUCRA) probabilities ranked interventions; node-splitting and loop-specific analyses assessed inconsistencies.
• Publication bias was probed by funnel plots and Egger’s regression, with p < 0.05 deemed significant.

This ensured a rigorous, transparent, and comprehensive comparison of GLP-1RAs and traditional antidiabetic therapies, supporting clinically meaningful conclusions.

(a) HbA1c (53 trials, 21,486 patients)

• All GLP-1 RAs significantly reduced HbA1c vs. placebo.

1. Tirzepatide: −2.3% (95% CI: −2.7 to −1.9) – most effective.
2. Semaglutide: −1.5% (95% CI: −1.7 to −1.3).
3. Liraglutide: −1.2% (95% CI: −1.3 to −1.0).

• Compared with traditional antidiabetic drugs, tirzepatide, semaglutide, liraglutide, EQW, and dulaglutide showed significant reductions; lixisenatide and EBID showed no advantage.

• SUCRA rankings confirmed tirzepatide as the top performer, followed by semaglutide and liraglutide.

(b) FPG (41 trials, 17,621 patients)

• All agents except lixisenatide reduced FPG vs. placebo.
• Tirzepatide led with −3.1 mmol/L (95% CI: −3.8 to −2.4), followed by semaglutide (−2.0 mmol/L) and liraglutide (−1.6 mmol/L).
• Only tirzepatide and semaglutide markedly outperformed traditional antidiabetic drugs, while lixisenatide increased FPG.

(c) Secondary outcomes

• Body weight reduction: Tirzepatide (−10 kg) > semaglutide (−3.8 kg) > EBID (−2.8 kg).
• Changes in BMI, blood pressure, and lipid profiles were not statistically significant across GLP-1RAs.

(d) Safety

• Compared with placebo and traditional drugs, GLP-1RAs were associated with a higher incidence of gastrointestinal adverse events, such as nausea, vomiting, diarrhea, and reduced appetite.
• Hypoglycemia was elevated only with EBID and semaglutide vs. placebo, while liraglutide and lixisenatide reduced hypoglycemia risk relative to traditional drugs.
• No major differences were observed among GLP-1RAs for other safety outcomes.

(e) Consistency & reliability

• Node-splitting and loop-specific analyses showed no significant inconsistencies, supporting the robustness of the NMA.

By pooling evidence from multiple RCTs, this NMA provides a clear, evidence-based guide for clinicians, helping them personalize T2DM care and maximize patient outcomes.
Semaglutide stands out with two formulations: injectable and oral. The injectable version has a fatty acid side chain that shields it from enzymatic breakdown, giving it a half-life of 1 week. The oral tablet uses a special absorption enhancer for preventing degradation in the stomach. Despite these differences, both forms are equally efficient, though the oral version may boost patient adherence owing to its convenience.

For analysis, both formulations were treated as a single intervention. The review included 64 trials with over 25,500 volunteers. Long-acting GLP-1RAs consistently outperformed short-acting ones in minimizing HbA1c and FPG. Among them, tirzepatide, semaglutide, and liraglutide delivered the strongest improvements compared with other GLP-1RAs, traditional antidiabetic drugs, and placebo, confirming their superior glycemic benefits. When it comes to weight management, tirzepatide, semaglutide, EBID, and liraglutide led to considerable weight loss versus placebo and standard treatments.

Notably, tirzepatide and semaglutide were more potent than other GLP-1RAs. Tirzepatide, a dual glucose-dependent insulinotropic polypeptide (GIP) and GLP-1 RA approved in 2022, combines enhanced blood sugar control with marked weight reduction, making it particularly suitable for obesity management. Liraglutide, on the other hand, may be ideal for those who do not require weight loss, such as older adults or those at risk of sarcopenia. Regarding safety, semaglutide, dulaglutide, liraglutide, and tirzepatide were related to a higher incidence of gastrointestinal adverse effects.

EBID and semaglutide escalated the risk of hypoglycemia, whereas liraglutide and lixisenatide lowered it compared to traditional therapies. No major differences were noted among GLP-1RAs themselves. These findings suggest careful selection based on individual patient profiles: liraglutide may be safer for those prone to gastrointestinal issues or hypoglycemia, while tirzepatide is preferable for those needing weight reduction. In short, the choice of a GLP-1RA should consider blood sugar control, weight impact, risk of hypoglycemia, and gastrointestinal tolerance.