Semaglutide reduces cardiovascular events in type 2 diabetes: a systematic review and meta-analysis highlighting enhanced benefits in chronic kidney disease

Abstract

Background

Cardiovascular disease (CVD) disproportionately affects type 2 diabetes (T2D) patients with chronic kidney disease (CKD). Although semaglutide shows cardiometabolic benefits, its efficacy across CVD endpoints and high-risk CKD subgroups remains underexplored.

Methods

We systematically reviewed PubMed, Embase, Cochrane, Web of Science (through September 2025) for randomized trials and observational studies evaluating oral semaglutide versus placebo in T2D. Data synthesis employed hazard ratios (HRs) and mean differences (MDs) with 95% confidence intervals (CIs).

Results

Across 17 studies (n = 40, 632 patients), semaglutide significantly reduced the risk of cardiovascular (CV) death by 23% [HR 0.77, 95% CI (0.68, 0.88)], major adverse cardiovascular events (MACE) by 18% [HR 0.82, 95% CI (0.75, 0.90)], expanded MACE by 22% [HR 0.78, 95% CI (0.70, 0.85)], nonfatal myocardial infarction (MI) by 18% [HR 0.82, 95% CI (0.68, 0.92)], and nonfatal stroke by 32% [HR 0.68, 95% CI (0.56, 0.83)]. The reductions in CV events were most pronounced in T2D patients with CKD (24–37% risk reduction vs. 13–35% in T2D alone or with CVD). Semaglutide also significantly improved modifiable CV risk factors, reducing systolic/diastolic blood pressure (SBP/DBP) decreased by 8.02/3.71 mmHg, and low-density Lipoprotein (LDL) decreased by 12.62 mg/dL, while increasing high-density lipoprotein (HDL) by 1.44 mg/dL. However, semaglutide had no significant effect on estimated glomerular filtration rate (eGFR) and urinary albumin-to-creatinine ratio (UACR).

Conclusion

Semaglutide confers robust cardiovascular protection in T2D, with amplified benefits in patients with comorbid CKD. It confers a collective advantage by simultaneously addressing multiple risk factors including blood pressure and lipoprotein, positioning it as a comprehensive treatment strategy for high-risk populations. This analysis is the first to synthesize evidence across CKD subgroups, supporting prioritizing semaglutide in T2D patients with CKD. Further research should clarify mechanisms underlying CKD-specific benefits and refine risk-stratified treatment algorithms.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40001-025-03241-8.

Introduction

Type 2 diabetes (T2D) is a chronic, progressive metabolic disorder affecting over 537 million people globally as of 2021, with projections indicating a rise to 643 million by 2030 and 783 million by 2045 [1–3]. Beyond its direct metabolic consequences, T2D significantly elevates the risk of microvascular and macrovascular complications, particularly cardiovascular disease (CVD)—the leading cause of morbidity and mortality in this population [4, 5]. CVD manifestations in T2D include ischemic heart disease, heart failure, stroke, coronary artery disease, and peripheral artery disease, collectively contributing to mortality in over 50% of affected individuals [6]. Atherosclerosis, the primary driver of CVD in T2D, underlies 85.8% of cases in the 34.8% of T2D patients globally with established CVD [7]. This risk is compounded by comorbidities such as chronic kidney disease (CKD), present in approximately 40% of individuals with diabetes [8]. The coexistence of CKD triples the risk of cardiovascular (CV) death compared to T2D alone, highlighting a critical intersection of metabolic and renal pathophysiology [9]. Globally, CVD accounts for one-third of all deaths, cementing its status as the foremost cause of mortality worldwide [10]. This alarming burden has intensified focus on integrated care strategies that address both glycemic control and cardiorenal protection in T2D management.

In response to the growing burden of diabetes and its complications, researchers have developed novel oral glucose-lowering therapies, including glucagon-like peptide-1 receptor agonists (GLP-1 RAs) [11], which are particularly compelling due to their multifactorial effects, which include improvements in glycemia, body weight, blood pressure, and dyslipidemia, alongside notable anti-inflammatory properties [12]. These agents are now prioritized in recent guidelines for their dual role in managing hyperglycemia and mitigating cardiorenal risks, offering benefits that extend beyond glycemic control [13]. Among GLP-1 RAs, semaglutide stands out as the only agent currently available in both subcutaneous and oral formulations for T2D management [14]. Clinical trials and previous meta-analyses have demonstrated its efficacy in lowering blood glucose, promoting weight loss, and reducing the risk of major CV events such as heart attack, stroke, or cardiovascular death in participants living with T2D [15–18].

Despite advancements in GLP-1 receptor agonist (GLP-1 RA) therapies, evidence for cardiorenal benefits in patients with T2D and CKD, established CVD, or high CVD risk has historically been limited [19]. However, recent landmark trials such as the FLOW study (2024) have demonstrated that semaglutide reduces the risk of major kidney events by 24% and major adverse CV events (MACE) by 18% in T2D patients with CKD, suggesting a paradigm shift in managing CV risk [20]. Current guidelines, including the 2023 European Society of Cardiology recommendations, prioritize GLP-1 RAs with proven CV benefits such as semaglutide for T2D patients with atherosclerotic CVD or high CV risk [21]. While these agents are now recognized for their dual cardiometabolic and renal protective effects, earlier meta-analyses lacked dedicated renal outcome trials to fully assess their efficacy in high-risk CKD/CVD populations. For instance, a 2021 meta-analysis of eight CV outcome trials showed a 14% reduction in MACE but did not specifically evaluate renal outcomes in advanced CKD cohorts [22]. Therefore, the current meta-analysis was conducted to assess the effect of semaglutide on CV outcomes in T2D patients with and CKD, established CVD, or high CVD risk.

Methods

Protocol and registration

This systematic review and meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [23] to evaluate the impact of semaglutide on cardiovascular (CV) outcomes in patients with T2D (Appendix 1). The study protocol was prospectively registered with the International Prospective Register of Systematic Reviews (PROSPERO; registration ID: CRD420251002785).

Search strategy

Two investigators (Yanni Yao and Nairong Liu) independently searched four databases—PubMed, Cochrane Library, Embase, and Web of Science—to identify relevant articles published up to September 10, 2025. The search strategy combined Medical Subject Headings (MeSH) terms and free-text terms related to semaglutide. Free-text terms were derived from the MeSH term “semaglutide” and supplemented by reviewing published meta-analyses, reviews, and targeted searches in PubMed and Embase. The key search terms included ‘semaglutide’, ‘NN9535’, ‘NN9924’, ‘Ozempic’, ‘Wegovy’, and ‘Rybelsus’. These terms were combined using the Boolean operator “OR”. No restrictions on language or publication date were applied. To minimize publication bias, additional searches included: (1) Abstracts from the American Diabetes Association (ADA) and European Association for the Study of Diabetes (EASD) annual meetings (up to 2024); (2) International clinical trial registries (ClinicalTrials.gov, EU Clinical Trials Register); (3) Pharmaceutical manufacturers’ trial databases.

Inclusion and exclusion criteria

This systematic review included studies meeting the following criteria: (1) population: Adult patients diagnosed with T2D, with or without CKD and/or elevated CVD risk; (2) intervention: Treatment with semaglutide or placebo; (3) outcomes: Primary outcomes included CV death, major adverse cardiovascular events (MACE, a composite of CV death, nonfatal myocardial infarction (MI), or nonfatal stroke), expanded MACE [MACE components plus coronary revascularization, hospitalization for unstable angina or heart failure (HF)], nonfatal MI, nonfatal stroke, and hospitalization for HF. Secondary outcomes included CV risk factors (e.g., systolic/diastolic blood pressure (SBP/DBP), Low/high Density Lipoprotein (LDL/HDL), estimated glomerular filtration rate (eGFR), and urine albumin-to-creatinine ratio (UACR)); (4) study design: Randomized controlled trials (RCTs), prospective or retrospective observational studies; RCTs are valued for their high internal validity in establishing causal efficacy of semaglutide under controlled conditions. Observational studies are included to provide evidence on effectiveness in broader, more heterogeneous real-world populations, to capture longer-term outcomes beyond typical RCT durations, and to assess safety signals that may be too rare to detect in RCTs; (5) language: Studies published in English.

The exclusion criteria were as follows: (1) patients were not diagnosed with T2D; (2) studies with incomplete, inconsistent, or unextractable outcome data; (3) non-primary research (e.g., reviews, meta-analyses, case reports, conference abstracts, book chapters, letters) or animal studies.; (4) non-English publications.

Study selection and data extraction

Two investigators (Yanni Yao and Nairong Liu) independently conducted the study selection process in three phases: First, all records retrieved from the databases were imported into EndNote. Duplicates were removed using automated deduplication followed by manual verification. Second, remaining records underwent dual-reviewer screening to assess alignment with eligibility criteria. Third, Articles passing initial screening were evaluated in full text to confirm compliance with inclusion/exclusion criteria. Disagreements were resolved through discussion and, if necessary, adjudicated by a third independent reviewer (Yanjie Guo).

Two investigators (Yanni Yao and Nairong Liu) independently extracted data using a standardized template, capturing the following: (1) Basic information of the included studies: Title, first author, publication year, study design (e.g., RCT, cohort), country, and population profile (e.g., T2D with/without CKD): (2) basic characteristics of the research subjects, including Age, sex/gender distribution, sample size per group, and diagnostic criteria for T2D and comorbid conditions; and (3) Primary and secondary outcomes as predefined (e.g., CV death, MACE, expanded MACE, nonfatal MI, nonfatal stroke, hospitalization for HF, SBP/DBP, LDL/HDL, eGFR, and UACR). Additionally, semaglutide administration route (oral/subcutaneous), dosage, and treatment duration were also extracted. All discrepancies were resolved through structured group deliberation, with persisting disagreements systematically escalated to an independent methodology expert (Yanjie Guo) for final arbitration.

Quality assessment

To evaluate the risk of bias in both prospective and retrospective studies included in our review, two researchers (Yanni Yao and Nairong Liu) utilized the Newcastle–Ottawa Scale (NOS) [24] to assess the quality of each article. The NOS offers a structured and systematic framework for evaluating research quality, which aids in identifying the strengths and weaknesses of the studies, thereby enhancing the credibility of our review process. The NOS assigns scores based on three key areas: selection of study subjects, comparability between groups, and the measurement of exposure factors. It consists of three sections: selection for the study group, comparability of the groups, and exposure assessment. The total score ranges from 0 to 9, with higher scores indicating better quality. Specifically, scores of 7 to 9 denote high quality, 4 to 6 indicate moderate quality, and 3 or below suggest poor quality [25]. Any discrepancies in scoring were resolved through discussions among the authors to reach a consensus.

The risk of bias in the RCT trials was evaluated using updated Cochrane tool (RoB2) [26] by two independent assessors (Yanni Yao and Nairong Liu). This tool is structured into five domains, assessing potential biases in the following aspects of the studies: randomization process, interventions, outcome description, outcome measurements, and selection of reported results. The articles were classified into three categories based on their bias: low risk of bias, some concerns, and high risk of bias.

Statistical analysis

Statistical analysis was conducted using Stata software (version 15.0; StataCorp, College Station, TX). The pooled hazard ratio (HR) and mean difference (MD) with 95% confidence interval (CI) were utilized to evaluate the combined effect sizes of the included studies. Heterogeneity among studies was assessed using the I-squared statistic (I²), which indicates the percentage of total variability across the pooled studies attributable to heterogeneity. If the I² value exceeded 50%, a random-effects model was applied to calculate the pooled effect size. Otherwise, a fixed-effects model was used. Results were deemed statistically significant if the p-value was less than 0.05. To assess the robustness of the meta-analysis findings, sensitivity analyses were conducted by iteratively removing individual studies to evaluate their impact on the overall results to evaluate the stability and reliability of the conclusions. Publication bias was evaluated using a funnel plot and Egger's test. A p-value less than 0.05 was considered indicative of significant publication bias.

Results

Literature search results

A total of 11,775 records were identified through database searches, with four additional records sourced from manual searches of reference lists. After removing 4487 duplicates, 7288 records underwent initial screening. Of these, 1559 records were excluded as non-primary research (e.g., reviews, meta-analyses, conference abstracts). Title/abstract screening of the remaining 5729 records led to the exclusion of 5744 articles that did not meet inclusion criteria. Full-text review of the 252 potentially eligible articles revealed 235 exclusions due to irrelevant outcomes or inaccessible full texts. 17 studies were ultimately included in the final analysis. The selection process is summarized in the PRISMA flow diagram (Fig. 1).

Fig. 1.

Flow diagram for search and selection of eligible studies included in the meta-analysis

Basic characteristics of the included studies

A total of 17 studies were included [15, 19, 27–41]. The characteristics of the included studies are presented in Table 1. The period of publication of the articles selected ranged between 2016 [15] and 2024 [29–32, 34, 35, 37], and the patients’ age varied from 61.0 years old [32] to 72.6 years old [33], The total number of patients participating in the studies included in this meta-analysis was 40, 632. The number of patients in those studies ranged between 61 [32] and 6374 [38]. Four studies originated from the SUSTAIN 6 trial [36, 38, 40, 41]. Despite this common source, they were included separately due to their distinct patient characteristics and outcomes. For example, Verma et al. [40] focused on CV events in a general T2D population [40], whereas Shaman et al. [36] and Verma et al. [41] studied patients with T2D at higher CVD risk and reported different outcomes (see Table 1 for details) [36, 41]. Similarly, four studies from the FLOW trial were included because they each reported different outcomes of interest [30, 31, 34, 35]. Among the included studies, four studies included T2D patients [27, 32, 37, 40], another four studies included T2D at higher CVD [36, 38, 39, 41], seven studies included patients with T2D combined with CVD [28–31, 33, 34], and the remaining two studies included patients with T2D had established CVD, CKD, or both [15, 19]. Depending on the treatment ways of semaglutide, two studies evaluated the impact of oral semaglutide [19, 32], and the rest of all studies received subcutaneous semaglutide [15, 27–31, 33–41]. The more detailed treatment plans (dosage and treatment duration) are shown in Table 1.

Table 1.

hubfdkj cbd vc

Studies	Design	Patients type	Drug dose	Treatment duration	Group		Age, years		Gender, male (%)		Outcomes
					Treatment, n	Control, n	Treatment	Control	Treatment	Control
Pantanetti et al. [32]	Prospective	T2D	14 mg, oral	6 months	Semaglutide, 61	–	61.0	–	61.0	–	SBP, DBP, LDL, HDL
Berra et al. [27]	Retrospective	T2D	1 mg, subcutaneous	12 months	Semaglutide, 459		63.9	–	62.7	–	SBP, DBP, LDL, HDL, eGFR
Verma et al. [40]	Retrospective	T2D	1 mg, subcutaneous	26 months	Semaglutide, 3297		65.0	–	60.7	–	CV death, MACE, Expanded MACE
Sivalingam et al. [37]	RCT	T2D	1 mg, subcutaneous	6.5 months	Semaglutide, 30	Placebo, 30	70.5	69.4	80.0	76.7	HbA1c, UACR, SBP, eGFR
Shaman et al. [36]	RCT	T2D at higher CVD	1 mg, subcutaneous	26 months	Semaglutide, 821	Placebo, 1648	65.0	–	60.7	–	eGFR
Verma et al. [41]	RCT	T2D at higher CVD	1 mg, subcutaneous	26 months	Semaglutide, 1648	Placebo, 1649	65.0	–	60.7	–	CV death, MACE, Expanded MACE
Verma et al. [39]	RCT	T2D at higher CVD	1 mg, subcutaneous	26 months	Semaglutide, 460	Placebo, 2380	–	–	–	–	MACE, Expanded MACE, Nonfatal MI, Nonfatal stroke
Strain et al. [38]	RCT	T2D at higher CVD	1 mg, subcutaneous	26 months	Semaglutide, 3196	Placebo, 3178	65.2	65.5	64.8	64.2	Nonfatal stroke
Kobo et al. [29]	RCT	T2D and CVD	1 mg, subcutaneous	26 months	Semaglutide, 640	Placebo, 1648	65.8	–	60.7	–	CV death, MACE, Expanded MACE, Nonfatal MI, Nonfatal stroke
Pe’rez-Belmonte et al. [33]	Retrospective	T2D and CVD	1 mg, subcutaneous	12 months	Semaglutide, 136	–	72.6	–	54.4	–	HbA1c, BMI, SBP, DBP, LDL, HDL, eGFR
Bueno et al. [28]	Retrospective	T2D and CKD	1 mg, subcutaneous	12 months	Semaglutide, 122	–	65.0	–	62.0	–	HbA1c, BMI, UACR, SBP, DBP, LDL, eGFR
Mahaffey et al. [30]	RCT	T2D and CKD	1 mg, subcutaneous	26 months	Semaglutide, 1767	Placebo, 1766	67.3	67.0	69.2	67.3	Nonfatal stroke
Perkovic et al. [34]	RCT	T2D and CKD	1 mg, subcutaneous	26 months	Semaglutide, 1767	Placebo, 1766	66.6	66.7	70.6	68.9	MACE
Pratley et al. [35]	RCT	T2D and CKD	1 mg, subcutaneous	26 months	Semaglutide, 1767	Placebo, 1766	66.6	66.7	70.6	68.9	CV death
Mann et al. [31]	RCT	T2D and CKD	1 mg, subcutaneous	26 months	Semaglutide, 1767	Placebo, 1766	66.6	66.7	70.6	68.9	Nonfatal MI
Marso et al. [15]	RCT	T2D and CKD/CVD	1 mg, subcutaneous	26 months	Semaglutide, 822	Placebo, 825	64.7	64.4	63.0	61.5	CV death, MACE, Expanded MACE, Nonfatal MI, Nonfatal stroke
Husain et al. [19]	RCT	T2D and CKD/CVD	14 mg, oral	22 months	Semaglutide, 1591	Placebo, 1592	66.0	66.0	68.1	68.6	CV death, MACE, Expanded MACE, Nonfatal MI, Nonfatal stroke

CV death: cardiovascular CV death; MACE: major adverse cardiovascular events; MI: myocardial infarction. SBP: systolic blood pressure; DBP: diastolic blood pressure; LDL: low-density lipoprotein; HDL: high-density lipoprotein; eGFR: estimated glomerular filtration rate; UACR: urinary albumin-to-creatinine ratio

Risk of bias assessment

The risk of bias in the five included studies [27, 28, 32, 33, 40]was evaluated. Based on the NOS criteria, four studies were classified as high-quality [27, 28, 32, 33], and one study was rated as medium-quality [40]. Detailed results are presented in Table 2.

Table 2.

Risk of bias assessment for prospective and retrospective studies

Included studies	Study population selection	Comparability between group	Comparability between group	Levels
Pantanetti et al. [32]	☆☆☆	☆☆	☆☆☆	8☆
Berra et al. [27]	☆☆☆☆	☆☆	☆☆☆	9☆
Verma et al. [40]	☆☆☆	☆	☆☆	6☆
Pe’rez-Belmonte et al. [33]	☆☆☆	☆☆	☆☆	7☆
Bueno et al. [28]	☆☆☆☆	☆☆	☆☆	8☆

The risk of bias in the included RCTs [15, 19, 29–31, 34–39, 41] was assessed, and found that four studies raised concerns, specifically regarding bias arising from the randomization process and bias in the selection of reported results [15, 34, 39, 41]. Details can be seen in Fig. 2.

Fig. 2.

Results of methodological quality evaluation of included studies

Meta-analysis

CV outcomes

Of the included studies, 10 studies reported CV death [15, 19, 29–31, 34, 35, 39–41], 10 studies reported MACE [15, 19, 29, 31, 34, 39–41], 6 studies reported Expanded MACE [15, 19, 29, 39–41], 5 studies reported nonfatal MI [15, 19, 29, 31, 39], 6 studies reported nonfatal stroke [15, 19, 29, 30, 38, 39], and 3 studies reported hospitalization due to heart failure (hospitalization for HF) [15, 19, 39]. The heterogeneity among the studies was low (I² = 11.1%, p = 0.276), therefore, the pooled HR with a fixed-effects model showed that semaglutide treatment was associated with significant reductions in multiple CV outcomes, as shown in Fig. 3. The most pronounced benefit was observed for nonfatal stroke, with a 32% risk reduction [I² = 0%, HR 0.68, 95% CI (0.56, 0.83); p < 0.001]. Significant risk reductions were also observed for CV death [I² = 0%, HR 0.77, 95 CI% (0.72, 0.84), p < 0.001], expanded MACE [I² = 32.2%, HR 0.78, 95% CI (0.70, 0.85); p < 0.001], and both MACE and nonfatal MI [MACE, I² = 0%, HR 0.82, 95 CI% (0.76, 0.87), p < 0.001; nonfatal MI, I² = 0%, HR 0.82, 95 CI% (0.68, 0.99), p = 0.043]. In contrast, semaglutide showed no significant association with risk of hospitalization for HF [I² = 0%, HR 1.10, 95% CI (0.92, 0.83), p = 0.291].

Fig. 3.

Forest plot of effects of semaglutide on CV events

Additionally, depending on the patient subgroups among the included studies, we further conducted subgroup analyses for the effects of semaglutide on CV events, and the results are presented in Table 3. We found that the pooled meta-analysis suggested a numerically greater treatment benefit in patients with T2D and CKD compared to other clinical cohorts, including those with T2D alone, T2D and CVD, and T2D with both CKD and CVD. Specifically, semaglutide was associated with a significant 24% reduction in CV death [I² = 0.0%, HR 0.76, 95 CI% (0.69, 0.94), p < 0.001] and a 27% reduction in MACE [I² = 0.0%, HR 0.73, 95% CI (0.62, 0.86), p < 0.001] in patients with T2D and CKD. In contrast, a lower and non-significant reduction was observed in patients with T2D and CVD for CV death [I² = 0.0%, HR 0.97, 95% CI (0.73, 1.28), p = 0.807] and MACE [I² = 0.0%, HR 0.82, 95% CI (0.72, 0.93), p = 0.805], and no significant benefit was observed in the T2D and CKD/CVD subgroup for CV death [I² = 70.5%, HR 0.79, 95% CI (0.56, 1.11), p = 0.177] and MACE [I² = 70.5%, HR 0.76, 95% CI (0.62, 0.92), p = 0.177]. We observed similar patterns for expanded MACE and nonfatal stroke (Table 3). For example, the point estimate for the reduction in expanded MACE with semaglutide was numerically greater in patients with T2D and CKD [I² = 0.0%, HR 0.72, 95% CI (0.64, 0.82), p < 0.001] than in those with T2D alone [I² = 0.0%, HR 0.87, 95% CI (0.78, 0.97), p = 0.012]. Nonetheless, there was no difference for the reduction in nonfatal MI with semaglutide among these clinical cohorts [T2D and CKD, I² = 0.0%, HR 0.80, 95% CI (0.55, 1.16), p = 0.236; T2D and CVD, I² = 0.0%, HR 0.76, 95% CI (0.54, 1.07), p = 0.117; T2D and CKD/CVD, I² = 0.0%, HR 0.88, 95% CI (0.66, 1.19), p = 0.413].

Table 3.

Subgroup analyses for the effects of semaglutide on CV events

Patient subgroups	Numbers	Pooled meta-analysis		Heterogeneity		Egger’s test
		HR (95% CI)	p	I2	p
CV death		0.77 (0.72, 0.84)	< 0.001	0.0%	0.728	0.739
T2D	1	0.79 (0.67, 0.94)	0.007	0.0%	< 0.001
T2D and CKD	4	0.76 (0.69, 0.84)	< 0.001	0.0%	0.556
T2D and  CVD	3	0.97 (0.73, 1.28)	0.807	0.0%	0.899
T2D and CKD/CVD	2	0.79 (0.56, 1.11)	0.177	70.5%	0.066
MACE		0.82 (0.76, 0.87)	< 0.001	0.0%	0.776	0.002
T2D	1	0.87 (0.78, 0.97)	0.007	0.0%	< 0.001
T2D and CKD	2	0.73 (0.62, 0.86)	< 0.001	0.0%	0.839
T2D and CVD	2	0.82 (0.72, 0.93)	0.805	0.0%	1.000
T2D and CKD/CVD	2	0.76 (0.62, 0.92)	0.177	70.5%	0.757
Expanded MACE		0.79 (0.74, 0.85)	< 0.001	32.2%	0.194	0.060
T2D	1	0.87 (0.78, 0.97)	0.012	0.0%	< 0.001
T2D and CVD	3	0.72 (0.64, 0.82)	< 0.001	0.0%	0.422
T2D and CKD/CVD	2	0.76 (0.65, 0.89)	0.001	0.0%	0.561
Nonfatal MI		0.82 (0.68, 0.99)	0.043	0.0%	0.586	0.476
T2D and CKD	1	0.80 (0.55, 1.16)	0.236	0.0%	< 0.001
T2D and CVD	2	0.76 (0.54, 1.07)	0.117	0.0%	0.729
T2D and CKD/CVD	2	0.88 (0.66, 1.19)	0.413	55.8%	0.132
Nonfatal stroke		0.68 (0.56, 0.83)	< 0.001	0.0%	0.763	0.111
T2D and CKD	1	0.80 (0.55, 1.16)	0.236	0.0%	< 0.001
T2D and CVD	3	0.63 (0.47, 0.84)	0.002	0.0%	0.496
T2D and CKD/CVD	2	0.65 (0.43, 0.97)	0.033	0.0%	0.671

CV death: cardiovascular death; MACE: major adverse cardiovascular events; MI: myocardial infarction; HR: hazard ratio

Funnel plots (Fig. 1S) and Egger’s tests indicated no significant publication bias for most endpoints, including CV death (p = 0.858), expanded MACE (p = 0.223), nonfatal MI (p = 0.867), and nonfatal stroke (p = 0.310). A significant publication bias was detected for MACE (Egger’s p = 0.013). Trim-and-fill analysis imputed three studies and yielded an adjusted effect estimate [Fig. 2S, I2 = 35.3%, HR 0.79, 95% CI (0.74, 0.85)], which remained consistent with the primary analysis, suggesting the findings were robust despite potential bias. Sensitivity analyses further confirmed the stability of the results (Appendix 2, Fig. 3S).

Impact of semaglutide on CV risk factors

Of the included studies, 6 studies reported the effect of semaglutide on the main CV risk factors including mean SBP [15, 27, 28, 33, 37], DPB [15, 27, 28, 33], LDL [15, 27, 28, 33], HDL [15, 27, 33], eGFR [27, 28, 33, 36, 37], and UACR [28, 37]. Our analysis revealed substantial reductions in SBP [I² = 96.6%, MD = −8.02, 95% CI (− 12.27, − 3.77), p < 0.001], DBP [I² = 89.0%, MD = − 3.71, 95% CI (− 6.68, − 0.74), p = 0.014], and LDL [I² = 96.0%, MD = − 12.62, 95% CI (− 21.35, − 3.90), p = 0.005], and a significant mean increase in HDL [I² = 65.8%, MD = 1.44, 95% CI (0.08, 2.81), p = 0.038]. However, eGFR and UACR levels showed no significant variation (Fig. 4). Due to the limited numbers of the included studies, only Egger’s tests were conducted and the results no publication bias among the studies (SBP, p = 0.127; DBP, p = 0.801; HDL, p = 0.524; eGFR, p = 0.285) except LDL (p = 0.024). The trim-and-fill method imputed 2 studies, yielding an adjusted effect estimate for LDL [MD = − 18.99, 95% CI (− 20.67, − 17.72), p < 0.001], which remained consistent with the primary analysis.

Fig. 4.

Forest plot of effects of semaglutide on CV risk factors

Discussion

Our meta-analysis demonstrated that semaglutide significantly reduced 23% in CV death, 18% in MACE, 22% in expanded MACE, 18% in nonfatal MI, and 32% in nonfatal stroke in patients with type 2 diabetes (T2D). Notably, the reduction in CV events was more pronounced in patients with T2D and CKD compared to those with T2D alone, T2D with preexisting CVD, or T2D with both CKD and CVD. Additionally, semaglutide significantly improved key CV risk factors, including reductions in SBP/DBP by 8.02/3.71 mmHg, and LDL by 12.62 mg/dL, alongside favorable trends in HDL by 1.44 mg/dL. However, semaglutide had no significant effect on eGFR and UACR. To our knowledge, this is the first systematic review and meta-analysis to comprehensively evaluate semaglutide’s effects on a broad spectrum of CV endpoints and risk factors in patients with T2D, including subgroups with CKD and/or elevated CVD risk.

While previous meta-analyses have primarily focused on the effects of semaglutide on glycaemic control, weight loss, and blood pressure in T2D [42–46], its cardioprotective role is now increasingly recognized. A recent meta-analysis in patients without diabetes demonstrated significant benefits for CV death, nonfatal MI, and HF hospitalization [47]; however, it was unable to assess broader endpoints like MACE due to a limited number of trials. Our study builds upon this by demonstrating that the magnitude of CV death reduction with semaglutide is comparable between T2D patients and overweight or obese patients without T2D, extending to MACE, expanded MACE, and nonfatal stroke. The evidence for semaglutide's cardioprotection has evolved across formulations and trial designs. While the PIONEER CVOT for oral semaglutide did not achieve statistical significance for MACE reduction [19] [48], the subcutaneous formulation demonstrated a clear benefit in the SUSTAIN-6 trial [15], a finding consistent with the broader class effect of GLP-1RAs [49, 50]. This underscores the importance of considering formulation and trial power when interpreting outcomes.

A critical finding of our analysis is the significantly greater cardiovascular benefit observed in the subgroup of patients with T2D and comorbid chronic kidney disease (CKD). This pronounced effect can be attributed to their substantially higher baseline CV risk, driven by a confluence of pathophysiological pathways including endothelial dysfunction, oxidative stress, systemic inflammation [51], volume overload, hypertension [52, 53], hyperphosphatemia, and vascular calcification [54, 55]. Consequently, there is a greater "absolute room for improvement" in this population. In contrast, the effect may be attenuated in patients with preexisting CVD, where irreversible structural damage may limit risk reversal [56], or in those with both CKD and CVD, where advanced disease and competing risks may blunt the treatment effect [61]. This suggests that CKD represents a particularly modifiable risk factor, and early intervention with semaglutide may be highly effective in preventing secondary CV events in this group [57, 58]. Our results, confirming significant CV risk reduction—especially in high-risk CKD patients—support current guidelines from the ADA and EASD that recommend prioritizing GLP-1RAs with proven benefit in T2D patients with elevated CV risk [59, 60]. Definitive evidence regarding the efficacy of semaglutide in such populations is anticipated from the ongoing SOUL trial (NCT03914326), which is evaluating MACE prevention over 5 years in patients with T2D and established CVD or CKD [61].

Additionally, our analysis confirms that semaglutide administration induces a significant reduction in blood pressure, with a more pronounced effect on systolic (SBP: − 8.02 mmHg) than diastolic pressure (DBP: − 3.71 mmHg). This effect is of critical importance for CV risk management in T2D, given the high prevalence of hypertension and its well-established role in driving CV event rates within this population [62]. The blood pressure-lowering properties of GLP-1 receptor agonists are consistent across clinical studies. For instance, the SUSTAIN 6 trial documented substantial reductions, underscoring the potential for cardiovascular benefits independent of glycemic control [63]. These effects are likely mediated through multiple pathways, including promoted weight loss, improved endothelial function, and reduced arterial stiffness [63]. In addition to its effects on blood pressure, semaglutide therapy led to significant improvements in lipid metabolism. We observed a notable reduction in LDL cholesterol (− 12.62 mg/dL), a primary driver of atherosclerosis and a key therapeutic target for reducing CV risk [64]. This finding aligns with prior meta-analyses confirming the efficacy of GLP-1 receptor agonists in improving the lipid profile [65, 66], and likely contributes to the overall cardioprotective effect of this drug class. Our analysis identified a modest but significant increase in high-density lipoprotein (HDL) cholesterol levels (+ 1.44 mg/dL). However, the clinical interpretation of this specific finding remains complex. Mechanistically, HDL facilitates reverse cholesterol transport, removing excess cholesterol from arterial walls and shuttling it to the liver for excretion—a process that may attenuate atherosclerosis progression [67, 68]. Additionally, HDL exerts anti-inflammatory and antioxidant effects, stabilizing vascular endothelium and mitigating plaque instability, both critical factors in reducing CV events [69–71]. The clinical utility of HDL is best interpreted in the context of other metabolic markers, as evidenced by observational data and the fact that pharmacologically raising HDL has not consistently improved clinical outcomes [72, 73]. Consequently, current management guidelines rightly prioritize LDL reduction and holistic risk factor modification over isolated increases in HDL [74].

This study has several notable limitations. First, while 17 studies were included in the analysis, several had small sample sizes (e.g., [32, 33]), which introduce imprecision and reduce the statistical power to detect true associations, potentially inflating the risk of overestimating effect sizes. Consequently, subgroup analyses with few studies should be interpreted with caution, and observed trends must be validated in future dedicated trials. Second, meta-analyses of CV risk factors were constrained by the limited number of eligible studies addressing these endpoints, reflecting gaps in the published literature; this paucity of data weaken the robustness of conclusions regarding semaglutide’s ancillary metabolic benefits. Third, the observed reductions in CV outcomes were predominantly driven by large trials such as SUSTAIN and FLOW, raising questions about generalizability to broader populations and underscoring the need for further independent replication. While these trials are high-quality, their dominance in the pooled analysis highlights a potential bias toward specific patient cohorts and trial conditions. Fourth, significant clinical heterogeneity was present across trials, including variations in outcome definitions, semaglutide dosage and formulation (e.g., 14 mg oral vs. 1.0 mg subcutaneous), and treatment duration (ranging from 6 to 26 months). These differences compromise the comparability and synthesis of results, despite the use of random-effects models to address statistical heterogeneity. Fifth, the inclusion of both RCTs and observational studies introduces potential bias due to differences in design, particularly confounding and selection bias in observational data, which may distort pooled effect estimates. Finally, while Egger’s test and trim-and-fill analysis did not suggest substantial publication bias, the possibility of undetected bias remains, especially given the limited power of these tests when few studies are included. Future large-scale randomized trials across diverse populations are warranted to confirm these benefits, clarify optimal dosing and treatment duration, and elucidate mechanisms underlying risk reduction, particularly in understudied subgroups.

Conclusion

This meta-analysis demonstrates that semaglutide significantly reduces CV events in patients with T2D including CV death, MACE, expanded MACE, nonfatal MI, and nonfatal stroke. Notably, the magnitude of CV risk reduction was greatest in patients with comorbid T2D and CKD, surpassing benefits observed in those with T2D alone, preexisting CVD, or combined CKD/CVD. Semaglutide also improved critical modifiable CV risk factors, including SBP, DBP, LDL, and HDL, suggesting pleiotropic mechanisms beyond glycemic control. While previous meta-analyses have established the cardiovascular benefits of semaglutide, this review extends the current evidence by providing a comprehensive evaluation across diverse CV end points and high-risk clinical subgroups, with a specific focus on CKD. These findings underscore semaglutide’s potential as a multifaceted therapeutic strategy for T2D patients, particularly those with concurrent CKD, who face disproportionately elevated CV risk. Further research is warranted to elucidate mechanisms underlying subgroup-specific benefits and optimize risk stratification in clinical practice.

Author contributions

All authors contributed to the study conception and design. Writing—original draft preparation: [Yanni Yao]; Writing—review and editing: [Yanni Yao, Nairong Liu, and Yanjie Guo]; Conceptualization: [Yanni Yao, Siyan Chen, and Meng Sun]; Methodology: [Yanni Yao, Siyan Chen, and Meng Sun]; Formal analysis and investigation: [Yanni Yao and Nairong Liu]; Resources: [Yanni Yao]; Supervision: [Yanjie Guo], and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

No specific grant was received for this research.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Formal ethics committee approval was not required for this study, as it involves the secondary analysis of publicly available, de-identified data from previously published clinical trials. All original trials included in our analysis have been documented to have adhered to the Good Clinical Practice guidelines and/or the principles of the Declaration of Helsinki.

Competing interests

The authors declare no competing interests.