Cardiovascular outcomes of semaglutide and tirzepatide in type 2 diabetes mellitus and obesity: a systematic review and meta-analysis Authors: Xiao-Yu Zhou, Qi Wu, Hao-lan Lu, Jia-Jie Xie, Yu-Fan Zhang, Yong Xu, Man Guo, Fang-Yuan Teng Published: 2026 ABSTRACT AbstractAimTo systematically evaluate cardiovascular outcomes associated with semaglutide and tirzepatide in adults with type 2 diabetes mellitus (T2DM), overweight or obesity.Materials and methodsWe systematically searched PubMed, Embase, Cochrane Library, Scopus, Web of Science, and ClinicalTrials.gov from inception to November 02, 2024, using standardized search terms for semaglutide and tirzepatide. The review was registered in PROSPERO (CRD420251008819).Results47 randomized controlled trials (RCTs) involving 59,352 participants were included. Semaglutide was associated with a reduced risk of major adverse cardiovascular events (MACE) (RR, 0.71; 95% CI [0.60–0.84]; p < 0.001), with reductions in cardiovascular death (RR, 0.81; 95% CI [0.72–0.93]; p = 0.002), myocardial infarction (RR, 0.70; 95% CI [0.61–0.80]; p < 0.001), and heart failure (RR, 0.60; 95% CI [0.41–0.87]; p = 0.007), but not stroke (RR, 0.86; 95% CI [0.73-1.00]; p = 0.056). For tirzepatide, a lower risk of myocardial infarction was observed (RR, 0.63; 95% CI [0.40-1.00]; p = 0.049), whereas the reduction in MACE did not reach statistical significance (RR, 0.79; 95% CI [0.54–1.17]; p = 0.245). Associations with cardiovascular death, heart failure, and stroke were also not statistically significant. Exploratory subgroup analyses suggested variation by treatment duration. Exploratory subgroup analyses suggested variation by treatment duration.ConclusionsSemaglutide showed a more consistent signal of cardiovascular benefit across the included trials, whereas cardiovascular outcome evidence for tirzepatide remained limited and uncertain. Because most available data were derived from separate drug-versus-control trials, the findings should not be interpreted as definitive comparative evidence favoring one agent over the other.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s13098-026-02163-7.Keywords: Semaglutide, Tirzepatide, Diabetes, Obesity, Cardiovascular benefit, Meta-analysis Introduction The metabolic epidemics of type 2 diabetes mellitus (T2DM) and obesity have reached alarming proportions, placing substantial pressure on healthcare systems worldwide [1]. According to the World Obesity Federation (2025), more than 40 million adult life-years are lost annually to obesity-related morbidity [2]. Accumulating evidence indicates that type 2 diabetes and obesity share common pathogenic pathways, which collectively accelerate the development of cardiovascular diseases [1]. Longitudinal cohort analyses demonstrate a threefold elevation in incidence rates of major adverse cardiovascular events (MACE) among patients with T2DM compared to their non-diabetic counterparts [3]. While glycemic control and weight management remain cornerstones of treatment, emerging evidence underscores the necessity of increasing attention being paid to cardiovascular outcomes when evaluating pharmacologic therapies. Notably, pharmacovigilance studies have revealed divergent cardiovascular outcomes among antidiabetic agents: rosiglitazone has been associated with an increased risk of myocardial infarction [4], whereas sodium-glucose cotransporter-2 inhibitors exemplified by empagliflozin exhibit significant cardioprotective properties, achieving a 38% relative risk reduction in cardiovascular death among high-risk populations [5]. These observations highlight the importance of evaluating cardiovascular endpoints in addition to metabolic efficacy when assessing novel antidiabetic therapeutics. In recent years, the therapeutic landscape for T2DM and obesity has been reshaped by the advent of two novel receptor-targeted agents: semaglutide, a long-acting glucagon-like peptide-1 receptor agonist (GLP-1 RA), and tirzepatide, a dual glucose-dependent insulinotropic polypeptide (GIP)/GLP-1 receptor agonist. Semaglutide has shown substantial efficacy in glycemic control, with HbA1c reductions of at least 1.5% [6]. In addition, it induces significant weight loss, with a mean bodyweight reduction of 11.40% (95% CI -12.51 to -10.29) and approximately 506 more individuals per 1000 achieving at least 10% weight loss compared to lifestyle modification alone [7]. Furthermore, semaglutide has been demonstrated to have benefits in cardiovascular risk reduction across landmark trials such as SELECT (20% cardiovascular disease risk reduction in obesity) [8], especially in reducing the incidence of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. However, tirzepatide has shown superior efficacy in HbA1c reduction (2.6%) and weight loss (12%–15%) compared to semaglutide in the SURPASS-2 trial [9]. Together, these findings have intensified interest in the comparative cardiovascular outcome profiles of the two agents. Despite the expanding clinical use of semaglutide and tirzepatide, direct comparative evidence for cardiovascular outcomes remains limited. Most randomized trials have evaluated each agent against a placebo or other comparators rather than against each other. Therefore, we conducted a systematic review and meta-analysis to summarize cardiovascular outcomes associated with semaglutide and tirzepatide in adults with T2DM, overweight, or obesity. Our primary aim was to synthesize the available cardiovascular outcome evidence for each drug within its own trial framework. Any apparent contrast between the two agents should therefore be interpreted cautiously. We also explored whether outcome patterns differed according to treatment duration, age, obesity status, baseline glycemic level, renal function, comparator type, and drug-specific dose categories. Methods The protocol for this systematic review and meta-analysis was developed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guideline [10] and prospectively registered in PROSPERO (CRD420251008819). Search strategy A comprehensive search was conducted in PubMed, Embase, Scopus, Web of Science, Cochrane Central Register of Controlled Trials (CENTRAL), and ClinicalTrials.gov from database inception to 02 November 2024. The search strategy was developed according to the PICOS framework (Population, Intervention, Comparitor, Outcome, and Study design) and combined free-text keywords with controlled vocabulary terms, including Medical Subject Headings, using Boolean operators. To maximize sensitivity, the search was kept intervention-focused rather than restricted to dedicated cardiovascular outcome trials, and eligibility for cardiovascular outcomes was determined during full-text review on the basis of extractable cardiovascular event data. Detailed search strategies for each database are provided in Supplementary Table S1-S6. Study selection Following PRISMA recommendations, duplicate records were removed before screening. Two reviewers independently screened titles and abstracts, reviewed full-text articles for eligibility, and determined final study inclusion according to predefined criteria. Disagreements were resolved through discussion, with unresolved cases adjudicated by a senior investigator. The detailed study selection process is summarized in Fig. 1. PRISMA Flow Diagram Eligibility criteria Eligible studies were English-language randomized controlled trials with an intervention duration of at least 26 weeks in adults (≥ 18 years) with type 2 diabetes, overweight, or obesity, in which subcutaneous tirzepatide or semaglutide was evaluated against placebo or active comparators, and sufficient cardiovascular event data were available for extraction. Comorbid conditions such as chronic kidney disease or heart failure were recorded as baseline clinical characteristics when present. Studies were excluded if they were single-arm trials, had an intervention duration of less than 26 weeks, were non-randomized designs (e.g., reviews, case reports, observational studies, or crossover trials), or were published in languages other than English. Detailed inclusion and exclusion criteria are provided in Supplementary Table S7. Data collection Two reviewers independently extracted data using piloted forms. Extracted variables included study characteristics, baseline participant characteristics, intervention details, and cardiovascular outcomes. Disagreements were resolved through discussion, with adjudication by a third investigator when necessary. When both journal publications and ClinicalTrials.gov records were available, peer-reviewed publications were used as the primary source, whereas registry records were used only to supplement missing or unclear information. If discrepancies between sources could not be reconciled, the affected data were not combined. Primary outcome In this meta-analysis, MACE was operationalized as a composite of cardiovascular death, myocardial infarction, heart failure-related events, and stroke. Heart failure-related events included broader trial-reported heart failure events and hospitalization for heart failure when separately reported. The primary outcome was the incidence of MACE, and secondary outcomes included its individual components. Detailed operational definitions are summarized in Supplementary Table S8. Risk-of-bias assessment Risk of bias was assessed independently by two reviewers using the Cochrane Risk of Bias 2.0 (RoB 2.0) tool [11] across five domains: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result. Each study was judged as low risk of bias, some concerns, or high risk according to the RoB 2.0 algorithm, and disagreements were resolved through discussion, with adjudication by a senior investigator when required. For open-label or less conventional trial designs, particular attention was given to bias due to deviations from intended interventions, whereas outcome measurement was judged in light of the objective ascertainment of cardiovascular events. Data analysis All statistical analyses were performed using Stata/MP 18.0 (StataCorp, College Station, TX, USA). Pairwise meta-analyses were conducted using the command, and study-level event data were pooled as risk ratios (RRs) with 95% confidence intervals (CIs). RRs were selected because study-level event count data were reported more consistently than hazard ratios across the included trials; accordingly, pooled estimates should be interpreted as cumulative trial-level risk comparisons rather than formal time-to-event effect estimates. All primary pooled analyses were performed using a random-effects model, given the expected clinical and methodological heterogeneity across the included trials [12, 13]. Between-study variance was estimated with the DerSimonian–Laird method. Heterogeneity was assessed using Cochran’s Q statistic, degrees of freedom, the corresponding p value, I², and τ², with detailed statistics provided in Supplementary Table S9. Trials with zero events in both arms were excluded from the corresponding RR-based analyses, whereas trials with zero events in one arm were retained. Sensitivity analyses were conducted by sequentially omitting individual studies. Small-study effects were assessed by visual inspection of funnel plots and, when at least 10 studies were available for a given outcome, by Begg’s rank correlation test and Egger’s regression test. Exploratory subgroup analyses were performed according to treatment duration, baseline BMI, age, comparator type, baseline glycemic level, dose, and baseline eGFR. Detailed subgroup definitions are provided in Supplementary Table S10. Results Search results and study characteristics The search identified 18,529 records from bibliographic databases and 106 records from ClinicalTrials.gov. After duplicate removal, 8,584 records remained for screening, of which 8,369 were excluded on the basis of titles and abstracts (Fig. 1). A total of 215 full-text articles were assessed for eligibility, and 168 were excluded, mainly because of duplicate reports (n = 84), irrelevant outcomes (n = 44), non-randomized controlled trials (n = 24), and improper study design (n = 7). The remaining reasons for exclusion were intervention duration that was too short (n = 4), other intervention (n = 3), other language (n = 1), and other population (n = 1). Ultimately, 47 randomized controlled trials (RCTs), involving 59,352 eligible participants and intervention durations ranging from 26 to 104 weeks, were included in the systematic review and meta-analysis [8, 9, 14–58]. The included studies evaluated semaglutide or tirzepatide primarily against placebo, other GLP-1 receptor agonists, or insulin-based regimens. All trials used a parallel-group design. Most were multinational, whereas eight were conducted in a single country (including three in the United States, three in Japan, one in Germany, and one in China). Intervention durations were categorized into three groups: 26–52 weeks (n = 21), 52–104 weeks (n = 20), and > 104 weeks (n = 6). Duration-based subgroup analyses were defined according to intervention duration rather than recruitment period or overall trial timeline. The pooled study population included 44.5% female participants, with a mean age of 58.0 years (SD 12.0), a mean HbA1c of 7.1% (SD 1.5), fasting plasma glucose of 10.4 mmol/L (SD 14.0), a mean body weight of 94.7 kg (SD 24.8), and a mean BMI of 33.5 kg/m² (SD 6.6). Full demographic and trial characteristics are summarized in Table 1 and Supplementary Tables S11 and S12. Table 1Baseline characteristics of randomized controlled clinical trials and participants includedStudy, yearCohort name or NCTPopulation descriptionNo. of participantsIntervention durationAge (mean, SD), yearsNo. (%) of maleBMI (mean, SD), kg/m2Weight (mean, SD), kgFPG (mmol/l)HbA1c (mean, SD), %InterventionsControlsOverall risk of biasMarso SP,2016 [14]NCT01720446(SUSTAIN 6)T2DMor ≥ 50 years with CVD3297104 weeks64.6 (7.4)2002 (60.7)32.8 (6.2)92.1 (20.6)NR8.7 (1.5)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWPlacebo 0.5 mgPlacebo 1.0 mgLow riskAhrén B,2017 [15]NCT01930188(SUSTAIN 2)T2DM123156 weeks55.1 (10.0)620 (50.6)32.5 (6.2)89.5 (20.3)9.4 (2.3)8.1 (0.9)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWSitagliptin 100 mg Po. QDLow riskAroda VR,2017 [16]NCT02128932(SUSTAIN 4)T2DM108230 weeks56.5 (10.4)574 (53.0)33.0 (6.5)93.5 (21.8)NR8.2 (0.9)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWInsulin glargine QDLow riskDavies M,2017 [17]NCT01923181T2DM63226 weeks57.9 (11.1)88 (13.9)31.7 (4.4)91.3 (16.9)9.6 (2.6)7.8 (0.5)Semaglutide 1.0 mg QWPlacebo Po. QDLow riskSorli C,2017 [18]NCT02054897(SUSTAIN 1)T2DM38730 weeks53.7 (11.3)210 (54.3)32.9 (7.7)91.9 (23.8)NR8.05 (0.9)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWPlaceboSome concernsAhmann AJ,2018 [19]NCT01885208(SUSTAIN 3)T2DM83148 weeks56.6 (15.8)447 (55.3)33.8 (12.9)95.8 (37.1)NR8.3 (1.2)Semaglutide 1.0 mg QWExenatide ER 2.0 mg QWSome concernsKaku K,2018 [20]NCT02207374T2DM60156 weeks58.5 (10.3)430 (71.5)26.4 (4.7)71.5 (15.5)8.9 (2.0)8.1 (0.9)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWAdditional OAD QWLow riskLingvay I,2018 [21]NCT02461589T2DMBMI 25–4070626 weeks56.7 (9.9)380(53.8)32.8 (4.4)94.3 (17.6)9.5 (2.6)8.1 (0.8)Semaglutide 0.05 mg QDSemaglutide 0.1 mg QDSemaglutide 0.2 mg QDSemaglutide 0.3 mg QDSemaglutide MTD QDLiraglutide 0.3 mg QDLiraglutide 0.6 mg QDLiraglutide 1.2 mg QDLiraglutide 1.8 mg QDPlacebo QDSome concernsO’Neil PM,2018 [22]NCT02453711BMI ≥ 3095752 weeks47.0 (17.0)338 (35.3)39.3 (12.7)111.5 (43.4)5.4 (2.1)5.5(0.7)Semaglutide 0.05 mg QDSemaglutide 0.1 mg QDSemaglutide 0.2 mg QDSemaglutide 0.3 mg QDSemaglutide 0.4 mg QDSemaglutide 0.3 mg FE QDSemaglutide 0.4 mg FE QDLiraglutide 3.0 mg QDPlacebo QDSome concernsPratley RE,2018 [23]NCT02648204(SUSTAIN 7)T2DM120140 weeks55.5 (10.6)662 (55.1)33.5 (6.8)95.2 (22.6)9.7 (2.5)8.2 (0.9)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWDulaglutide 0.75 mg QWDulaglutide 1.5 mg QWLow riskRodbard HW,2018 [24]NCT02305381(SUSTAIN 5)T2DM39730 weeks58.8 (16.8)222 (56.1)32.2 (8.0)91.7 (29.5)8.6 (4.8)8.4 (1.1)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWPlaceboLow riskLingvay I,2019 [25]NCT03136484(SUSTAIN 8)T2DM78852 weeks56.6 (10.9)424 (53.8)32.3 (6.8)90.2 (22.6)9.4 (2.7)8.3 (1.0)Semaglutide 1.0 mg QWCanagliflozin 300 mg Po. QDLow riskCapehorn MS,2020 [26]NCT03191396(SUSTAIN 10)T2DM57730 weeks59.5 (10.2)327 (56.7)33.7 (6.8)96.9 (21.3)9.9 (2.4)8.2 (1.0)Semaglutide 1.0 mg QWLiraglutide 1.2 mg QDLow riskDavies M,2021 [27]NCT03552757(STEP 2)BMI ≥ 27 T2DM121068 weeks55.0 (11.0)181 (15.0)35.9 (6.4)99.9 (22.5)8.5 (2.3)[396]8.1 (0.8)Semaglutide 1.0 mg QWSemaglutide 2.4 mg QWPlaceboLow riskDel Prato S,2021 [28]NCT03730662(SURPASS-4)T2DM2002104 weeks63.6 (8.6)1246 (62.2)32.6 (5.5)90.3 (18.7)9.5(2.8)8.5 (0.9)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWInsulin glargine QDSome concernsFrías JP,2021 [9]NCT03987919(SURPASS-2)T2DM187940 weeks56.6 (10.4)882 (47.0)34.2 (6.9)93.7 (21.9)9.6 (2.9)8.3 (1.0)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWSemaglutide 1.0 mg QWLow riskJi L,2021 [29]NCT03061214(SUSTAIN China)T2DM86830 weeks53.0 (10.8)369 (42.5)27.8 (4.9)76.4 (15.8)9.2 (2.4)8.1 (0.9)Semaglutide 0.5 mg QWSemaglutide 1.0 mg QWSitagliptin 100 mg Po. QDLow riskLudvik B,2021 [30]NCT03882970(SURPASS-3)T2DM144452 weeks57.4 (10.0)802 (55.5)33.5 (6.1)94.3 (20.1)9.4 (2.6)8.2 (0.9)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWInsulin degludec QDLow riskNewsome PN,2021 [31]NCT02970942BMI ≥ 2532072 weeks55.0 (10.6)126 (39.3)35.7 (6.4)98.3 (21.6)NR7.3 (1.2)Semaglutide 0.1 mg QDSemaglutide 0.2 mg QDSemaglutide 0.4 mg QDPlacebo QDLow riskRosenstock J,2021 [32]NCT03954834(SURPASS-1)T2DM47840 weeks54.1 (11.9)247 (51.7)31.9 (6.6)85.9 (19.8)8.5 (2.2)7.9 (0.9)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWPlaceboSome concernsRubino D,2021 [33]NCT03548987(STEP 4)BMI ≥ 3080368 weeks46.7 (12.0)169 (21.0)34.4 (7.0)96.1 (22.6)4.9 (0.4)5.4 (0.3)Semaglutide 2.4 mg QWPlaceboLow riskWadden TA,2021 [34]NCT03611582(STEP 3)BMI ≥ 3061168 weeks46.0 (13.0)116 (19.0)38.0 (6.8)105.8 (22.9)93.9 (9.5)5.7 (0.3)Semaglutide 2.4 mg QWPlaceboLow riskWilding JPH, 2021 [35]NCT03548935(STEP 1)BMI ≥ 30196168 weeks46.3 (12.7)508 (25.9)37.9 (6.6)105.3 (21.9)NR5.7 (0.3)Semaglutide 2.4 mg QWPlaceboLow riskDahl D,2022 [36]NCT04039503(SURPASS-5)T2DMBMI ≥ 2347540 weeks60.7 (10.0)264 (55.6)33.4 (6.1)95.2 (21.6)9.0 (2.8)8.3 (0.9)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWPlaceboLow riskEli Lilly and Company,2022NCT04235959T2DM2452 weeks56.3 (5.4)11 (45.8)NRNRNRNRTirzepatide 10 mg QWTirzepatide 15 mg QWPlaceboLow riskGarvey WT,2022 [38]NCT03693430(STEP 5)BMI ≥ 30304104 weeks47.4 (11.0)68 (22.4)38.6 (6.9)106.1 (21.9)5.3 (0.6)5.7 (0.4)Semaglutide 2.4 mg QWPlaceboLow riskHeise T,2022 [39]NCT03951753T2DM11728 weeks61.9 (6.9)86 (73.5)31.3 (4.3)94.7 (14.2)7.3 (1.5)7.8 (0.6)Tirzepatide 15 mg QWSemaglutide 1.0 mg QWPlaceboLow riskInagaki N,2022 [40]NCT03861052(SURPASS J-mono)T2DM63652 weeks56.6 (10.3)481 (75.6)28.1 (4.4)78.2 (14.5)8.9 (1.9)8.2 (0.9)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWDulaglutide 0.75 mg QWLow riskJastreboff AM,2022 [41]NCT04184622(SURMOUNT-1)BMI ≥ 28253972 weeks44.9 (12.5)825 (32.5)38.0 (6.8)104.8 (22.1)5.4 (0.6)5.6 (0.4)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWPlaceboSome concernsKadowaki T,2022 [42]NCT03811574(STEP 6)BMI ≥ 2740168 weeks51.0 (11.0)253 (63.1)31.9 (4.3)87.5 (15.2)6.2 (1.5)6.4 (1.2)Semaglutide 1.7 mg QWSemaglutide 2.4 mg QWPlaceboLow riskKellerer M,2022 [43]NCT03689374(SUSTAIN 11)T2DM174852 weeks61.2 (9.4)894 (51.1)31.5 (5.5)87.9 (18.2)NR8.6 (0.7)Semaglutide 1.0 mg QWInsulin aspart 4U TIDLow riskRubino DM,2022 [44]NCT04074161(STEP 8)BMI ≥ 2733868 weeks49.1 (13.2)73 (21.6)37.5 (6.8)104.5 (23.8)5.5 (0.3)5.3 (0.6)Semaglutide 2.4 mg QWLiraglutide 3.0 mgPlaceboLow riskEli Lilly and Company,2023NCT04844918(SURMOUNT-J)BMI ≥ 2722572 weeks50.8 (10.7)133 (59.1)NRNRNRNRTirzepatide 10 mg QWTirzepatide 15 mg QWPlaceboSome concernsFrias JP,2023 [46]NCT04982575T2DM9232 weeks58.0 (9.0)59 (64.1)35.5 (6.3)105.7 (24.1)9.6 (2.7)8.4 (0.8)Semaglutide 2.4 mg QWCagrilintide 2.4 mg QWLow riskGao L,2023 [47]NCT04093752(SURPASS-AP-Combo)T2DMBMI ≥ 2391740 weeks54.1 (11.4)507 (55.9)27.9 (4.0)76.6 (14.5)NR8.7 (1.0)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWInsulin glargine QDSome concernsGarvey WT,2023 [48]NCT04657003(SURMOUNT-2)BMI ≥ 2793872 weeks54.2 (10.6)462 (49.3)36.1 (6.6)100.7 (21.1)NR8.0 (0.9)Tirzepatide 10 mg QWTirzepatide 15 mg QWPlaceboLow riskKosiborod MN,2023 [49]NCT04788511(STEP-HFpEF)BMI ≥ 30 withheart failure52952 weeks69.0 (9.6)274 (51.8)37.0 (5.7)105.1 (21.0)NR6.8 (1.0)Semaglutide 2.4 mg QWPlaceboLow riskLincoff AM, 2023 [8]NCT03574597(SELECT)≥ 45 years BMI ≥ 27 with CVD17,064104 weeks61.6 (8.8)12,732 (74.6)33.3 (5.0)96.6 (17.7)NR5.8 (0.3)Semaglutide 2.4 mg QWPlaceboLow riskNovo Nordisk A/S, 2023NCT05486065T2DM24540 weeks52.8 (8.3)125 (51.0)NRNRNRNRSemaglutide 2.0 mg QWSemaglutide 8.0 mg QWSemaglutide 16.0 mg QWPlaceboSome concernsRosenstock J,2023 [51]NCT04537923(SURPASS-6)T2DM142452 weeks58.8 (9.7)291 (20.4)33.2 (5.3)90.5 (18.1)8.8 (3.1)8.8 (1.0)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWInsulin lispro TIDLow riskWadden TA,2023 [52]NCT04657016(SURMOUNT-3)BMI ≥ 3057972 weeks45.6 (12.2)215 (37.1)NRNRNRNRTirzepatide MTD QWPlaceboSome concernsAronne LJ,2024 [53]NCT04660643(SURMOUNT-4)BMI ≥ 3067052 weeks48.5 (12.5)197 (29.4)30.5 (6.4)85.2 (21.1)4.7 (0.4)5.1 (0.3)Tirzepatide MTD QWPlaceboSome concernsKosiborod MN,2024 [54]NCT04916470(STEP-HFpEF DM)BMI ≥ 30 with heart failure61652 weeks69.5 (8.9)343 (55.7)36.9 (5.7)102.8 (20.0)NR6.8 (1.0)Semaglutide 2.4 mg QWPlaceboLow riskLoomba R,2024 [55]NCT04166773(SYNERGY-NASH)BMI 27–5019052 weeks54.4 (11.3)81 (42.6)36.1 (6.1)99.8 (21.5)NR6.5 (1.1)Tirzepatide 5 mg QWTirzepatide 10 mg QWTirzepatide 15 mg QWPlaceboSome concernsMcGowan BM,2024 [56]NCT05040971(STEP 10)T2DM20752 weeks53.0 (11.0)60 (29.0)40.1 (6.9)111.6 (22.2)5.9 (0.6)5.9 (0.3)Semaglutide 2.4 mg QWPlaceboLow riskNovo Nordisk A/S, 2024NCT03596450T2DM1278104 weeks57.4 (11.1)692 (54.2)NRNRNRNRSemaglutide MTD QWStandard of CareSome concernsPerkovic V,2024 [59]NCT03819153(FLOW)T2DM3533104 weeks66.6 (9.0)2464 (69.7)32.0 (6.3)89.6 (20.5)NR7.8 (1.3)Semaglutide 1.0 mg QWPlaceboLow riskNCT, National Clinical Trial; NO., number of total participants; BMI, body mass index; FPG, fasting plasma glucose; HbA1c, glycosylated haemoglobin; T2DM, type 2 diabetes mellitus; PO, oral intake; NR, not report; QD, once daily; QW, once weekly; TID, thrice daily; OAD, oral antidiabetic drug (dipeptidyl peptidase-4 inhibitor, biguanide, sulphonylurea, glinide, α-glucosidase inhibitor or thiazolidinedione); ER, extended release; FE, fast (2-weekly) dose escalation; MTD, maximum tolerated doseIntervention duration refers to the randomized treatment period used for subgroup classification in this review Baseline characteristics of randomized controlled clinical trials and participants included Marso SP, 2016 [14] NCT01720446 (SUSTAIN 6) T2DM or ≥ 50 years with CVD Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Placebo 0.5 mg Placebo 1.0 mg Ahrén B, 2017 [15] NCT01930188 (SUSTAIN 2) Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Aroda VR, 2017 [16] NCT02128932 (SUSTAIN 4) Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Davies M, 2017 [17] Sorli C, 2017 [18] NCT02054897 (SUSTAIN 1) Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Ahmann AJ, 2018 [19] NCT01885208 (SUSTAIN 3) Kaku K, 2018 [20] Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Lingvay I, 2018 [21] T2DM BMI 25–40 Semaglutide 0.05 mg QD Semaglutide 0.1 mg QD Semaglutide 0.2 mg QD Semaglutide 0.3 mg QD Semaglutide MTD QD Liraglutide 0.3 mg QD Liraglutide 0.6 mg QD Liraglutide 1.2 mg QD Liraglutide 1.8 mg QD Placebo QD O’Neil PM, 2018 [22] Semaglutide 0.05 mg QD Semaglutide 0.1 mg QD Semaglutide 0.2 mg QD Semaglutide 0.3 mg QD Semaglutide 0.4 mg QD Semaglutide 0.3 mg FE QD Semaglutide 0.4 mg FE QD Liraglutide 3.0 mg QD Placebo QD Pratley RE, 2018 [23] NCT02648204 (SUSTAIN 7) Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Dulaglutide 0.75 mg QW Dulaglutide 1.5 mg QW Rodbard HW, 2018 [24] NCT02305381 (SUSTAIN 5) Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Lingvay I, 2019 [25] NCT03136484 (SUSTAIN 8) Capehorn MS, 2020 [26] NCT03191396 (SUSTAIN 10) Davies M, 2021 [27] NCT03552757 (STEP 2) 8.5 (2.3) [396] Semaglutide 1.0 mg QW Semaglutide 2.4 mg QW Del Prato S, 2021 [28] NCT03730662 (SURPASS-4) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Frías JP, 2021 [9] NCT03987919 (SURPASS-2) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Ji L, 2021 [29] NCT03061214 (SUSTAIN China) Semaglutide 0.5 mg QW Semaglutide 1.0 mg QW Ludvik B, 2021 [30] NCT03882970 (SURPASS-3) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Newsome PN, 2021 [31] Semaglutide 0.1 mg QD Semaglutide 0.2 mg QD Semaglutide 0.4 mg QD Rosenstock J, 2021 [32] NCT03954834 (SURPASS-1) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Rubino D, 2021 [33] NCT03548987 (STEP 4) Wadden TA, 2021 [34] NCT03611582 (STEP 3) Wilding JPH, 2021 [35] NCT03548935 (STEP 1) Dahl D, 2022 [36] NCT04039503 (SURPASS-5) T2DM BMI ≥ 23 Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Eli Lilly and Company, 2022 Tirzepatide 10 mg QW Tirzepatide 15 mg QW Garvey WT, 2022 [38] NCT03693430 (STEP 5) Heise T, 2022 [39] Semaglutide 1.0 mg QW Placebo Inagaki N, 2022 [40] NCT03861052 (SURPASS J-mono) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Jastreboff AM, 2022 [41] NCT04184622 (SURMOUNT-1) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Kadowaki T, 2022 [42] NCT03811574 (STEP 6) Semaglutide 1.7 mg QW Semaglutide 2.4 mg QW Kellerer M, 2022 [43] NCT03689374 (SUSTAIN 11) Rubino DM, 2022 [44] NCT04074161 (STEP 8) Liraglutide 3.0 mg Placebo Eli Lilly and Company, 2023 NCT04844918 (SURMOUNT-J) Tirzepatide 10 mg QW Tirzepatide 15 mg QW Frias JP, 2023 [46] Gao L, 2023 [47] NCT04093752 (SURPASS-AP-Combo) T2DM BMI ≥ 23 Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Garvey WT, 2023 [48] NCT04657003 (SURMOUNT-2) Tirzepatide 10 mg QW Tirzepatide 15 mg QW Kosiborod MN, 2023 [49] NCT04788511 (STEP-HFpEF) BMI ≥ 30 with heart failure Lincoff AM, 2023 [8] NCT03574597 (SELECT) ≥ 45 years BMI ≥ 27 with CVD Semaglutide 2.0 mg QW Semaglutide 8.0 mg QW Semaglutide 16.0 mg QW Rosenstock J, 2023 [51] NCT04537923 (SURPASS-6) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW Wadden TA, 2023 [52] NCT04657016 (SURMOUNT-3) Aronne LJ, 2024 [53] NCT04660643 (SURMOUNT-4) Kosiborod MN, 2024 [54] NCT04916470 (STEP-HFpEF DM) BMI ≥ 30 with heart failure Loomba R, 2024 [55] NCT04166773 (SYNERGY-NASH) Tirzepatide 5 mg QW Tirzepatide 10 mg QW Tirzepatide 15 mg QW McGowan BM, 2024 [56] NCT05040971 (STEP 10) Perkovic V, 2024 [59] NCT03819153 (FLOW) NCT, National Clinical Trial; NO., number of total participants; BMI, body mass index; FPG, fasting plasma glucose; HbA1c, glycosylated haemoglobin; T2DM, type 2 diabetes mellitus; PO, oral intake; NR, not report; QD, once daily; QW, once weekly; TID, thrice daily; OAD, oral antidiabetic drug (dipeptidyl peptidase-4 inhibitor, biguanide, sulphonylurea, glinide, α-glucosidase inhibitor or thiazolidinedione); ER, extended release; FE, fast (2-weekly) dose escalation; MTD, maximum tolerated dose Intervention duration refers to the randomized treatment period used for subgroup classification in this review Quality control and small-study effects of included studies Risk of bias was assessed using the Cochrane RoB 2.0 tool, with the results summarized in Figure S1 and Supplementary Table S13. Overall, 33 of the 47 included trials were evaluated at low risk of bias, whereas 14 were judged as having some concerns. All trials were rated as low risk in the domains of randomization, missing outcome data, measurement of the outcome, and selection of the reported result. The main source of concern was bias due to deviations from intended interventions, primarily in open-label or less conventional trial designs. Accordingly, although many studies used blinded procedures, the risk-of-bias profile should not be interpreted as representing a uniformly blinded evidence base. Detailed heterogeneity statistics for each pooled analysis are provided in Supplementary Table S9. Sensitivity analyses based on sequential omission of individual studies showed that the pooled estimates were generally stable. Primary outcomes: MACE In this systematic review and meta-analysis, MACE was analyzed under a harmonized framework comprising cardiovascular death, myocardial infarction, and stroke, whereas heart failure-related events were evaluated separately. In pooled analyses of 47 randomized trials, semaglutide was associated with a reduced risk of the primary composite cardiovascular outcome (RR, 0.71; 95% CI [0.60–0.84]; p < 0.001), whereas the pooled estimate for tirzepatide did not reach statistical significance (RR, 0.79; 95% CI [0.54–1.17]; p = 0.245) (Fig. 2). The p-values reported here refer to treatment effects rather than between-study heterogeneity. Overall risks of MACE in T2DM or obesity patients treated with semaglutide or tirzepatide compared with controls in all trials Heterogeneity for the MACE analysis was low to moderate for semaglutide (I² = 38.2%; p for heterogeneity = 0.016), low for tirzepatide (I² = 15.3%; p = 0.278), and low to moderate overall (I² = 30.7%; p = 0.025). Deconstruction of the composite outcome showed that semaglutide was associated with reduced risks of cardiovascular death (RR, 0.81; 95% CI [0.72–0.93]; p = 0.002) (Figure S2), myocardial infarction (RR, 0.70; 95% CI [0.61–0.80]; p < 0.001) (Fig. 3), and heart failure-related events (RR, 0.60; 95% CI [0.41–0.87]; p = 0.007) (Figure S3), whereas the association with stroke did not reach statistical significance (RR, 0.86; 95% CI [0.73-1.00]; p = 0.056) (Figure S4). For tirzepatide, a lower pooled risk estimate was observed for myocardial infarction (RR, 0.63; 95% CI [0.40-1.00]; p = 0.049) (Fig. 3), whereas associations with cardiovascular death (RR, 0.67; 95% CI [0.41–1.08]; p = 0.100) (Figure S2), heart failure-related events (RR, 0.60; 95% CI [0.32–1.13]; p = 0.115) (Figure S3), and stroke (RR, 1.16; 95% CI [0.66–2.05]; p = 0.613) were not statistically significant (Figure S4). Forest plot of available comparisons of semaglutide or tirzepatide and controls for myocardial infarction in all trials Subgroup analyses Exploratory subgroup analyses were conducted across seven domains: age, BMI, HbA1c, eGFR, dose category, intervention duration, and comparator type. These analyses were performed to explore potential sources of heterogeneity rather than to provide confirmatory evidence. Detailed subgroup definitions are provided in Supplementary Table S8. Funnel plots, Begg’s and Egger’s tests, and subgroup sensitivity analyses are shown in Supplementary Figures S5–S20. MACE In exploratory subgroup analyses, semaglutide was associated with a lower MACE estimate at 52–104 weeks (RR, 0.55; 95% CI [0.34–0.89]; p = 0.014) (Fig. 4A). Within semaglutide dose categories, the > 0.5 mg group showed a lower pooled risk estimate than the ≤ 0.5 mg group. Lower pooled risk estimates were also observed in participants aged > 60 years, in those with overweight or obesity, and across baseline glycemic subgroups. For tirzepatide, the ≥ 104-week subgroup showed a borderline estimate (RR, 0.72; 95% CI [0.52-1.00]; p = 0.053) (Fig. 4B). Factors and risks of MACE in RCTs of semaglutide (A) or tirzepatide (B) use Cardiovascular death For semaglutide, lower pooled risk estimates for cardiovascular death were observed in the ≥ 104-week subgroup (RR, 0.81; 95% CI [0.72–0.92]; p = 0.002) and in the > 0.5 mg dose category (RR, 0.81; 95% CI [0.72–0.93]; p = 0.002), participants aged > 60 years (RR, 0.81; 95% CI [0.71–0.92]; p = 0.001), those with overweight or obesity (RR, 0.81; 95% CI [0.72–0.93]; p = 0.002), and in HbA1c > 6.5% subgroup (RR, 0.78; 95% CI [0.64–0.94]; p = 0.009) (Figure S21A). For tirzepatide, subgroup estimates for cardiovascular death did not reach statistical significance (Figure S21B). Myocardial infarction For semaglutide, lower pooled risk estimates for myocardial infarction were observed in the ≥ 104-week subgroup (RR, 0.70; 95% CI [0.61–0.80]; p < 0.001), the > 0.5 mg dose category (RR, 0.70; 95% CI [0.61–0.80]; p < 0.001), participants aged > 60 years (RR, 0.70; 95% CI [0.61–0.80]; p < 0.001), those with overweight or obesity (RR, 0.70; 95% CI [0.61–0.80]; p < 0.001), and across baseline glycemic subgroups (HbA1c > 6.5%: RR, 0.74; 95% CI [0.59–0.92]; p = 0.007; HbA1c ≤ 6.5%: RR, 0.68; 95% CI [0.57–0.80]; p < 0.001). (Fig. 5A). For tirzepatide, a lower pooled risk estimate for myocardial infarction was observed in the subgroup with HbA1c > 6.5% (RR, 0.60; 95% CI [0.37–0.96]; p = 0.032) and in the comparator subgroup (RR, 0.34; 95% CI [0.13–0.92]; p = 0.034); whereas other subgroup estimates did not reach statistical significance (Fig. 5B). Factors and risks of myocardial infarction in RCTs of semaglutide (A) or tirzepatide (B) use Heart failure For semaglutide, lower pooled risk estimates for heart failure-related events were observed at 52–104 weeks (RR, 0.22; 95% CI [0.14–0.36]; p < 0.001) and in the > 0.5 mg dose category (RR, 0.59; 95% CI [0.41–0.84]; p = 0.003). The ≤ 0.5 mg category showed a higher pooled risk estimate than the > 0.5 mg category (≤ 0.5 mg: RR, 1.42; 95% CI [1.01–1.99]; p = 0.043). Variation in pooled estimates was also observed according to age (> 60 years: RR, 0.59; 95% CI [0.38–0.93]; p = 0.022), body weight category (RR, 0.61; 95% CI [0.41–0.90]; p = 0.013), and in HbA1c > 6.5% subgroup (RR, 0.51; 95% CI [0.29–0.90]; p = 0.021) (Fig. 6A). For tirzepatide, subgroup estimates for heart failure-related events did not reach statistical significance (Fig. 6B). Factors and risks of heart failure in RCTs of semaglutide (A) or tirzepatide (B) use Stroke In exploratory subgroup analyses of stroke outcomes, the > 0.5 mg semaglutide category was associated with a lower stroke risk estimate than the ≤ 0.5 mg category (RR, 0.61; 95% CI [0.41–0.90]; p = 0.013) (Figure S22A). These findings should be interpreted within the subgroup framework and do not alter the non-significant overall pooled stroke estimate for semaglutide. Tirzepatide exhibited evidence across prespecified stroke subgroups remains limited and did not show statistically significant differences (Figure S22B). Discussion This meta-analysis synthesized evidence from 47 RCTs involving 59,352 participants to evaluate cardiovascular outcomes associated with subcutaneous semaglutide and tirzepatide in adults with type 2 diabetes, overweight, or obesity. Given that the available evidence was derived predominantly from separate drug-versus-control trials, the present analysis should not be interpreted as providing definitive comparative evidence. Within the currently available evidence base, semaglutide showed a more consistent signal of cardiovascular benefit, particularly for MACE and several secondary outcomes, whereas the cardiovascular outcome evidence for tirzepatide remained more limited and less precise. In addition, non-significant findings for tirzepatide should be interpreted cautiously, as they may reflect limited event numbers and imprecision rather than definitive evidence of no effect. Accordingly, any apparent contrast between the two agents is best regarded as indirect and exploratory. Existing systematic reviews have reported cardiovascular outcome data for semaglutide and tirzepatide, and some have compared tirzepatide with other GLP-1 receptor agonists [59–62]. However, direct comparative evidence between semaglutide and tirzepatide remains limited, and prior reviews have not fully synthesized the available cardiovascular outcome evidence for these two agents within a single framework [61, 63–65]. In this context, the present study extends the existing literature by integrating cardiovascular outcome evidence for semaglutide and tirzepatide within one review and by exploring whether pooled estimates varied according to intervention duration, baseline characteristics, dose categories, and comparator type. Potential biological explanations for the cardiovascular effects of semaglutide and tirzepatide have been discussed in prior experimental and clinical literature [3, 66–69]. For semaglutide, proposed pathways include improved glycemic control, clinically meaningful weight reduction, and broader cardiometabolic effects that may reduce vascular and myocardial stress [14, 70, 71]. Tirzepatide may confer similar cardiometabolic benefits through combined GIP/GLP-1 receptor agonism, particularly via marked effects on glucose lowering and body weight reduction. From a clinical perspective, these mechanisms provide biologically plausible support for cardiovascular benefit with both agents [9, 41, 72–74]. However, the present study was based on pooled cardiovascular event counts rather than mechanistic endpoints and was not designed to determine whether any apparent differences between semaglutide and tirzepatide reflect true pharmacologic differences, differences in the maturity of the evidence base, or differences in trial design and follow-up [69]. Therefore, mechanistic considerations should be interpreted as background biological plausibility rather than as direct explanations for the comparative patterns observed in this meta-analysis. The more consistent cardiovascular signal observed for semaglutide in the present review may therefore reflect both biologically plausible cardiometabolic effects and the fact that semaglutide currently has a larger and more mature evidence base for cardiovascular outcomes. The more consistent cardiovascular signal observed for semaglutide in the present review may therefore reflect both biologically plausible cardiometabolic effects and the fact that semaglutide currently has a larger and more mature cardiovascular outcome evidence base [8, 14, 66]. By contrast, cardiovascular outcome evidence for tirzepatide remains more limited, and firmer conclusions will require additional dedicated cardiovascular outcome studies and adequately powered direct comparative trials [60, 75]. In exploratory subgroup analyses, lower pooled risk estimates for heart failure-related events were observed for semaglutide during 52–104 weeks of follow-up and in the higher semaglutide dose category. However, these findings should be interpreted cautiously and should not be taken as evidence of a definitive duration- or dose-dependent mechanistic effect [66]. This analysis has several limitations. First, the available evidence was derived predominantly from separate semaglutide-versus-control and tirzepatide-versus-control trials, so any apparent between-drug differences remain indirect and may be influenced by differences in study populations, protocols, and comparator conditions. Second, rare cardiovascular outcomes may have been affected by the exclusion of double-zero-event trials and by the instability inherent to rare-event meta-analyses, particularly for cardiovascular death, heart failure, and stroke. Third, despite the large pooled sample, several endpoint-specific and subgroup-specific analyses, especially those involving tirzepatide, may still have been underpowered because of limited event numbers; accordingly, non-significant findings should be interpreted as reflecting uncertainty rather than definitive evidence of no effect. Finally, clinical and methodological heterogeneity across trials, including variation in populations, composite outcome definitions, follow-up duration, and multiple exploratory subgroup comparisons, may reduce exchangeability and limit the interpretability of pooled estimates. In addition, participants in randomized trials of semaglutide and tirzepatide are often highly selected and may not fully reflect routine care populations with multimorbidity, frailty, variable adherence, and differing background therapies, which may further limit external validity. Conclusion In adults with T2DM or overweight/obesity, the currently available randomized evidence suggests that semaglutide is associated with a more consistent signal of cardiovascular benefit across the included trials, whereas cardiovascular outcome evidence for tirzepatide remains limited and uncertain. Because most available data derive from separate drug-versus-control studies, these findings should not be interpreted as definitive evidence that semaglutide is superior to tirzepatide. Further adequately powered head-to-head trials and dedicated cardiovascular outcome studies are needed to clarify the comparative cardiovascular effects of these agents. Supplementary material Supplementary material 1. Acknowledgements None. Abbreviations Confidence intervals Glucagon-like peptide-1 receptor agonist Glucose-dependent insulinotropic polypeptide Heart failure with preserved ejection fraction Heart failure with reduced ejection fraction Major adverse cardiovascular events Preferred reporting items for systematic reviews and meta-analyses Risk ratios Randomized controlled trials Type 2 diabetes Author contributions CRediT authorship contribution statement: Xiao-Yu Zhou : investigation, software, data curation, writing-original draft, writing-review & editing; Qi Wu : data curation, methodology, software; Hao-lan Lu : software, data curation, writing-review & editing; Jia-Jie Xie: software, data curation, writing-review & editing; Yu-Fan Zhang : investigation, methodology; Yong Xu : resources, validation; Man Guo : conceptualization, resources, funding acquisition, software, supervision, writing-review & editing; Fang-Yuan Teng : funding acquisition, supervision, resources, writing-original draft, writing-review & editing. All authors read and approved the final manuscript. Funding This work was supported by the National Natural Science Foundation of China (grant numbers: 82300911, 32201056), Sichuan Province Science and Technology Program (application number: 2025ZNSFSC1622), the Luzhou Science and Technology Program (grant number: 2025LZXNYDJC27), the Scientific Research Funding of Southwest Medical University (grant number: 2025LCYXZX33), and 2026 Undergraduate Innovation Training Project of Southwest Medical University (to Xiao-Yu Zhou). Data availability The study was based on previously published studies and publicly available sources. The extracted study-level data sheets and analytic code supporting the findings of this review are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xiao-Yu Zhou, Qi Wu and Hao-lan Lu contributed equally. Contributor Information Yong Xu, Email: xywyll@swmu.edu.cn. Man Guo, Email: guoman613@163.com. Fang-Yuan Teng, Email: tengfangyuan383@swmu.edu.cn. References Associated Data Data Availability Statement The study was based on previously published studies and publicly available sources. The extracted study-level data sheets and analytic code supporting the findings of this review are available from the corresponding author upon reasonable request.