Impact of GLP-1 receptor agonist-based therapies on cardiovascular and renal outcomes in diabetic and non-diabetic patients with CKD

Abstract

Background

The effect of glucagon-like peptide-1 (GLP-1) receptor agonists-based therapies on cardiovascular and renal outcomes has not been systematically reviewed across baseline kidney function groups. We conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) with GLP-1 Receptor Agonists (RAs) in patients with and without chronic kidney disease (CKD).

Methods

We performed a PubMed/Medline search of randomized, placebo-controlled, event-driven outcome trials of GLP-1 RAs versus placebo in patients with and without diabetes from inception to January 2025. CKD was defined as an estimated glomerular filtration rate (eGFR) < 60 ml/min/1.73m2. The primary outcome was major adverse cardiovascular events (MACE). Secondary outcomes included hospitalization for heart failure, CKD progression, cardiovascular and all-cause mortality. The relative risk (RR) was estimated using a random-effects model.

Results

Nine RCTs were included with a total of 75,088 patients, including 17,568 with eGFR < 60 ml/min/1.73m2. Use of an GLP-1 RA in patients with CKD was associated with a lower incidence of MACE (RR 0.84; 95% CI 0.74–0.95; P 0.006) and of CKD progression (RR 0.85, 95% CI 0.77–0.94; P 0.002), compared with placebo. There was no differential treatment effect of GLP-1 RA on these endpoints by CKD status at baseline.

Conclusions

GLP-1 RAs offer substantial cardiovascular and renal protection in patients with CKD. These findings support their use in CKD patients and confirms that these therapies may be continued as kidney function declines.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13098-025-01831-4.

Introduction

Glucagon-like peptide-1 (GLP-1) receptor agonists-based therapies function as glucose-lowering agents by mimicking the action of incretins, stimulating insulin secretion following oral glucose ingestion [1]. They are recognized for their efficacy in reducing blood glucose levels, facilitating modest weight loss, and lowering blood pressure. In addition, prescription and utilization of GLP-1 Receptor Agonists (RAs) are steadily increasing, reflecting their growing role in diabetes and overweight/obesity management [2]. Large randomized clinical trials (RCTs) have demonstrated their significant cardiovascular and renal benefits in individuals with type 2 diabetes [3, 4].

Assessing the impact of GLP-1 RAs on cardiovascular and renal outcomes based on the presence or absence of chronic kidney disease (CKD) at baseline remains an important question in order to confirm whether their therapeutic effect is independent of kidney function.

This study aims to conduct a systematic review and meta-analysis of RCTs to evaluate the impact of GLP-1 RAs on hard clinical endpoints, stratified by CKD status, among diabetic and non-diabetic patients administered GLP-1 RAs versus placebo control.

Methods

This systematic review and meta-analysis were carried out and presented following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [5, 6].

Search strategy and selection criteria

We aimed to identify all randomized, placebo-controlled, event-driven outcome trials of GLP-1 RAs or dual glucose-dependent insulinotropic polypeptides (GIP/GLP-1) RAs versus placebo control. Trials including participants with type 1 diabetes or individuals < 18 years of age were excluded. Inclusion of patients with CKD was required. CKD was defined as an eGFR < 60 ml/min/1.73m2. RCTs had to be peer reviewed manuscripts with a minimum follow-up of 6 months. At least one of the following cardiovascular or renal outcomes had to be reported: cardiovascular death, hospitalization for heart failure, major adverse cardiovascular events (MACE), renal death, or CKD progression.

We searched PubMed up until January 2025 for RCTs to identify potentially eligible studies, using the following search terms: ((((((myocardial infarction) OR (stroke or cerebrovascular accident)) OR (heart failure or cardiac failure)) OR (death OR mortality)) OR (“Cardiovascular Diseases“[Mesh])) OR (kidney failure)) AND ((lixisenatide or exenatide or liraglutide or semaglutide or dulaglutide or albiglutide or efpeglenatide or tirzepatide) OR (glucagon-like peptide-1 receptor[Mesh])).

Titles and abstracts of all identified articles were independently screened by AS, SF and KG. When required, full-text manuscripts were reviewed to identify potentially relevant studies, as well as reference lists of all selected studies and available meta-analyses, to search for any additional qualifying studies.

Data synthesis and analysis

All relevant baseline characteristics and study outcomes were independently extracted by AS, SF and KM, using a standardized digital extraction form, including treatment effects in patient subgroups defined by the presence or absence of CKD. Any discrepancies in data extraction or risk-of-bias assessment were resolved by consensus.

Cardiovascular outcomes of interest were MACE (including cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke), each of the MACE components, hospital admission for heart failure and death from any cause. Two kidney outcomes were examined: new onset macroalbuminuria and a composite renal outcome (including a combination of development of new-onset macroalbuminuria, decline in estimated glomerular filtration rate (eGFR), progression to end-stage kidney disease, or death attributable to kidney causes, as reported in each study). The primary outcome was MACE. A detailed definition of clinical outcomes in each trial included in this meta-analysis is depicted in Table 1.

Table 1.

Outcome definitions

Study	Definition of MACE	composite Renal outcome	All-cause mortality (Yes/No)	CV death (Y/N)	Heart failure (Y/N)	CV death or heart failure (Y/N)	Non-fatal stroke (Y/N)	All strokes (Y/N)
AMPLITUDE-O	Death from CV or unknown causes, nonfatal myocardial infarction or nonfatal stroke	Macroalbuminuria, doubling creatinine, ESKD, renal death	Y	Y	Y
EXSCEL	Death from CV causes, nonfatal myocardial infarction or nonfatal stroke	Macroalbuminuria, ≥ 40% eGFR decline, RRT, renal death	Y
FLOW	Death from CV causes, nonfatal myocardial infarction or nonfatal stroke	ESKD (GFR < 15), RRT, renal death	Y
Harmony Outcomes	Death from CV causes, myocardial infarction or stroke	Not reported	Y	Y		Y		Y
LEADER	Death from CV causes, nonfatal myocardial infarction or nonfatal stroke	Macroalbuminuria, doubling creatinine, GFR ≤ 45, RRT, renal death	Y	Y	Y		Y	Y
PIONEER 6	Death from CV or unknown causes, nonfatal myocardial infarction or nonfatal stroke	Not reported	Y	Y	Y		Y
REWIND	Death from CV or unknown causes, nonfatal myocardial infarction or nonfatal stroke	Macroalbuminuria, ≥ 30% GFR decline, RRT	Y	Y			Y	Y
SELECT	Death from CV causes, nonfatal myocardial infarction or nonfatal stroke		Y	Y	Y	Y	Y
SUSTAIN-6	Death from CV causes, nonfatal myocardial infarction or nonfatal stroke	Macroalbuminuria, doubling creatinine, GFR ≤ 45, RRT, renal death	Y	Y	Y		Y

Outcome definitions. MACE, major adverse cardiovascular event; CV, cardiovascular; ESKD, end-stage kidney disease; GFR, glomerular filtration rate (in ml/min/1.73m2)

We compared treatment effects in patients with and without CKD. CKD was defined as an eGFR < 60 ml/min/1.73m2 in most studies. In the AMPLITUDE-O trial, patients with an eGFR as high as 71.5 ml/min/1.73m2 were included in the CKD group [7].

Statistical analysis

The principal summary measure was the relative risk (RR) with associated 95% confidence intervals. The pooled RR for each outcome was estimated using a random-effects model with the standard DerSimonian & Laird approach [8]. Treatment effect modification by CKD status was explored. Results were presented in a forest plot.

The Cochran’s Q test and Higgins I² index was used to quantify heterogeneity and assess inconsistency. Heterogeneity was considered to be low, moderate, or high if I² was less than 25%, 26–50%, or greater than 50%, respectively.

For study quality assessment (performed by AS & CT), the second version of the Cochrane Risk-of-bias tool for RCTs (RoB2) was used [9].

The GRADE approach was used to rate confidence in effect estimates [10]. The initial rating for RCTs was high and decreased in presence of serious inconsistency, indirectness, imprecision, risk of bias, or when publication bias was likely.

We used the risk in the placebo group, along with the pooled relative risk for overall patients at long-term follow-up from the systematic review, to calculate the absolute effect estimates in our evidence summaries. A funnel plot was created to evaluate publication bias.

Statistical analyses were performed using Stata (version 18 SE; College Station, TX) and meta package in R Studio (R Foundation for Statistical Computing) (See Table 2 and 3).

Table 2.

Study characteristics

Study Year	Drug, dose	Number of participants	Mean follow-up, years	Age, years	Women (%)	Diabetes (%)	SGLT2 inhibitors AT Baseline (%)	ACEI-ARB at baseline (%)	MRA at baseline (%)	Baseline eGFR < 60*	Lowest eGFR included (ml/min/1,73m2)	uACR > 30 mg/g (%)	Established CV disease (%)	History of heart failure (%)
AMPLITUDE-O 2021	Efpeglenatide	4 076	1.8	65 ± 8	1 344 (33)	100	206 (15)	3 262 (80)	NA	1 287 (32)*	25	1977 (48.5)	3 650 (90)	737 (18)
EXSCEL 2017	Exenatide 2 mg weekly	14 691	3.2	62 ± 9	5 603 (38)	100	77 (0.1)	11 788 (80)	912 (6)	3 177 (22)	30		10 782 (73)	2 389 (16)
FLOW 2024	Semaglutide 1.0 mg weekly	3 533	3.4	67 ± 9	1 069 (30)	100	550 (16)	3367 (95)	257 (7)	2 813 (80)	25	3 424 (96.9)	1838 (52)	675 (19)
Harmony Outcomes 2018	Albiglutide 30/50 mg weekly	9 463	1.5	64 ± 7	2 894 (31)	100	575 (6)	7726 (82)	NA	2 222 (23)	30		9 463 (100)	1 922 (20)
LEADER 2016	Liraglutide 1.8 mg daily	9 340	3.8	65 ± 7	3 337 (36)	100	NA	7741 (83)	505 (5)	2158 (23)	N/A	3 422 (37)	7 598 (81)	1 667 (18)
PIONEER 6 2019	Oral semaglutide 14 mg daily	3 183	1.3	66 ± 7	1 007 (32)	100	305 (10%)	NA	NA	875 (27)	30	N/A	2 695 (85)	388 (12)
REWIND 2019	Dulaglutide 1.5 mg weekly	9 901	5.4	66 ± 7	4 589 (46)	100	NA	8068 (81)	NA	2199 (22)	15		3 114 (31)	853 (9)
SELECT 2023	Semaglutide 2.4 mg weekly	17 604	3.3	62 ± 9	4 872 (28)	0	0	13 116 (75)	NA	1 898 (11)	N/A		17 604 (100)	4 286 (24)
SUSTAIN-6 2016	Semaglutide 0.5 or 1.0 mg weekly	3 297	2.1	65 ± 7	1 295 (39)	100	5 (0.2)	2753 (84)	194 (5.9)	939 (28)	N/A	1 304 (40)	2 735 (83)	777 (24)

Study characteristics. Results are presented as mean (standard deviation) or number (percentage). ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; MRA, Mineralocorticoid receptor antagonists; NA, not available; uACR, urine albumin-cretinine ration; eGFR, estimated glomerular filtration rate (in ml/min/1.73m2); CV, cardiovascular. *eGFR < 71.5 ml/min/1.73m2 for the AMPLITUDE-O trial

Table 3.

GRADE

Outcome	Study group	GLP-1 RA vs. placebo	Absolute effect estimate	Quality of evidence
MACE	< 60	0.84 (95% CI 0.75–0.95)	24 fewer per 1000 (95% CI 8 fewer – 38 fewer)	MODERATEA Serious inconsistency
	≥ 60	0.85 (95% CI 0.80 –0.91)	14 fewer per 1000 (95% CI 8 fewer – 19 fewer)	HIGH
CV death	< 60	0.87 (95% CI 0.55–1.39)	11 fewer per 1000 (95% CI -34 fewer – 40 fewer)	VERY LOWB Serious imprecision & very serious inconsistency
	≥ 60	0.80 (95% CI 0.70–0.93)	9 fewer per 1000 (95% CI 3 fewer – 14 fewer)	HIGH
HF	< 60	0.90 (95% CI 0.60–1.33)	7 fewer per 1000 (95% CI -24 fewer – 29 fewer)	VERY LOWC Serious imprecision & very serious inconsistency
	≥ 60	0.91 (95% CI 0.77–1.08)	3 fewer per 1000 (95% CI -2 fewer – 7 fewer)	MODERATED Serious imprecision
All MI	< 60	0.87 (95% CI 0.68–1.13)	13 fewer per 1000 (95% CI -12 fewer – 33 fewer)	LOWE Serious imprecision & inconsistency
	≥ 60	0.95 (95% CI 0.85 –1.07)	3 fewer per 1000 (95% CI -4 fewer – 9 fewer)	MODERATED Serious imprecision
All strokes	< 60	0.81 (95% CI 0.38–1.74)	9 fewer per 1000 (95% CI -34 fewer – 28 fewer)	VERY LOWF Very serious imprecision & inconsistency
	≥ 60	0.88 (95% CI 0.64–1.20)	4 fewer per 1000 (95% CI -6 fewer – 11 fewer)	VERY LOWG Very serious imprecision & serious inconsistency
All cause mortality	< 60	0.88 (95% CI 0.66–1.17)	17 fewer per 1000 (95% CI -24 fewer – 48 fewer)	VERY LOWH Serious imprecision & very serious inconsistency
	≥ 60	0.84 (95% CI 0.73–0.95)	11 fewer per 1000 (95% CI 3 fewer – 19 fewer)	HIGH
Composite renal outcome	< 60	0.85 (95% CI 0.77–0.94)	24 fewer per 1000 (95% CI 9 fewer – 36 fewer)	HIGH
	≥ 60	0.84 (95% CI 0.75–0.94)	12 fewer per 1000 (95% CI 4 fewer – 18 fewer)	MODERATEI Serious inconsistency

A Serious inconsistency: I2 = 52.8%

B Serious imprecision: wide confidence interval that includes no difference. Very serious inconsistency: I2 = 84.7%

C Serious imprecision: wide confidence interval that includes no difference. Very serious inconsistency: I2 = 74.8%

D Serious imprecision: confidence interval includes no difference

E Serious imprecision: confidence interval includes no difference. Serious inconsistency: I2 = 56.1%

F Serious imprecision: wide confidence interval that includes no difference. Very serious inconsistency: I2 = 88.1%

G Very serious imprecision: wide confidence interval that includes no difference. Serious inconsistency: I2 = 71.8%

H Very serious imprecision: wide confidence interval that includes no difference. Very serious inconsistency: I2 = 77.5%

I Serious inconsistency: I2 = 51.9%

Results

Study characteristics

A total of 2537 articles were identified. After title and abstract screening, 42 articles were selected for full-text review. Nine trials (11 publications) with a total of 75,088 patients were included in this meta-analysis (Supplemental Fig. 1) [7, 11–20]. The overall risk of bias was assessed as low (“not serious”) for all the RCTs included in this meta-analysis (Supplemental Fig. 2). There was no major publication bias for any of the outcomes identified at the inspection of the funnel plots (Supplemental Fig. 5).

MACE

Included studies were limited to those using the 3-point MACE definition (CV death, non-fatal myocardial infarction, or stroke). GLP-1 RAs were associated with a significantly reduced MACE incidence compared to placebo in both CKD patients (RR 0.84; 95% CI 0.75–0.95; P 0.006) and non-CKD patients (RR 0.85, 95% CI 0.80–0.91; P < 0.001) (Fig. 1). Moderate heterogeneity was detected (I² = 37%). There was no interaction between CKD status and the effect of GLP-1 RAs on MACE outcomes.

Fig. 1.

Incidence of MACE. Incidence of major adverse cardiovascular events (MACE) with GLP-1 agonists compared with placebo in patients with and without chronic kidney disease (CKD). Results are stratified by CKD status. Data are presented as risk ratios (RR) with 95% confidence intervals (95%-CI). A similar incidence of MACE is identified with GLP-1 agonists compared with placebo in patients with and without CKD. A random effects model is used. Definition of MACE is detailed in Table 1

Renal outcomes

The definition of the composite renal outcome was variable across different studies, incorporating different outcomes or thresholds for eGFR decline (Table 1). GLP-1 RAs were associated with a significantly reduced incidence of the composite renal outcome in patients with CKD (RR 0.85, 95% CI 0.77–0.94; P 0.002) (Fig. 2) and without CKD (RR 0.84, 95% CI 0.75–0.94; P 0.003) (Fig. 2) compared to placebo. Moderate heterogeneity was observed (I² of 35.9%). No interaction between CKD status and the effect of GLP-1 RAs on composite renal outcome was observed.

Fig. 2.

Incidence of composite renal outcome with GLP-1 agonists compared with placebo in patients with and without chronic kidney disease (CKD). Results are stratified by CKD status. Data are presented as risk ratios (RR) with 95% confidence intervals (95%-CI). A similar incidence of composite renal outcome is identified with GLP-1 agonists compared with placebo in patients with and without CKD. A random effects model is used. Definition of the composite renal outcome is detailed in Table 1

Composite renal outcome 2, which includes composite renal outcome above plus incident macroalbuminuria, was explored in the EXSCEL study. They found that exenatide did not affect composite renal outcome 2 in CKD patients (RR 1.00, 95% CI 0.76–1.31; P 0.995) (Supplemental Table 1) compared to placebo. In patients without CKD, exenatide was associated with a numerically lower incidence of this outcome (RR 0.87, 95% CI 0.74–1.02; P 0.085) (Supplemental Table 1). However, no interaction was observed between CKD status and the effect of exenatide on incident macroalbuminuria [11].

The REWIND study demonstrated that dulaglutide was associated with a non-significant reduction in the incidence of new macroalbuminuria in CKD patients compared to placebo (RR 0.93, 95% CI 0.76–1.14; P 0.488) (Supplemental Table 1). In contrast, in patients without CKD, dulaglutide significantly reduced the incidence of new macroalbuminuria (RR 0.72, 95% CI 0.62–0.83; P < 0.001) (Supplemental Table 1) [17].

Cardiovascular death

When compared to placebo, GLP-1 RAs were associated with a non-significant reduction in cardiovascular death in CKD patients (RR 0.87, 95% CI 0.55–1.39; P 0.563), but the number of events was limited and there was significant heterogeneity between the two studies in this subgroup. GLP-1 RAs they significantly reduced it cardiovascular mortality in patients without CKD (RR 0.80, 95% CI 0.70–0.93; P 0.002) (Fig. 3). High heterogeneity was detected (I2 of 61.4%). No interaction between CKD status and the effect of GLP-1 RAs on cardiovascular death was observed.

Fig. 3.

Incidence of cardiovascular death with GLP-1 agonists compared with placebo in patients with and without chronic kidney disease (CKD). Results are stratified by CKD status. Data are presented as risk ratios (RR) with 95% confidence intervals (95%-CI). A similar incidence of cardiovascular death is identified with GLP-1 agonists compared with placebo in patients with and without CKD. A random effects model is used

Myocardial infarction

GLP-1 RAs were associated with a numerically lower incidence of all myocardial infarction events in CKD (RR 0.87, 95% CI 0.68–1.13; P 0.302) (Supplemental Fig. 4) and non-CKD patients (RR 0.95, 95% CI 0.85–1.07; P 0.405) (Supplemental Fig. 4) compared to placebo, though these results were not statistically significant. Low heterogeneity was detected overall (I² of 5.6%). There was no interaction between CKD status and the effect of GLP-1 RAs on myocardial infarction.

The LEADER study showed that there was a non-significant decrease in the incidence of non-fatal myocardial infarction with liraglutide in both CKD (RR 0.76, 95% CI 0.57–1.01; P 0.058) and non-CKD patients (RR 0.94, 95% CI 0.78–1.13; P = 0.504) compared to placebo (Supplemental Table 1). Non-fatal myocardial infarction is included in MACE outcome [15].

Stroke

GLP-1 RAs were associated with a numerically lower incidence of all stroke events in patients with CKD (RR 0.81, 95% CI 0.38–1.74; P 0.587) (Supplemental Fig. 3) and without CKD (RR 0.88, 95% CI 0.64–1.20; P = 0.413) (Supplemental Fig. 3) compared to placebo. High heterogeneity was noted (I² of 75.1%). There was no interaction between CKD status and the effect of GLP-1 RAs on stroke.

The LEADER study found a significant decrease in incidence of non-fatal stroke in patients with CKD treated with liraglutide (RR 0.53, 95% CI 0.34–0.81; P 0.003), and a non significant increase in incidence in non-CKD patients (RR 1.07, 95% CI 0.84–1.37; P 0.575) (Supplemental Table 1) compared to placebo. Non-fatal stroke is included in MACE outcome [15].

Heart failure

When compared to placebo, GLP-1 RAs showed a non-significant reduction in heart failure in CKD (RR 0.90, 95% CI 0.60–1.33; P 0.593) (Fig. 4) and non-CKD patients (RR 0.91, 95% CI 0.77–1.08; P 0.271) (Fig. 4). Moderate heterogeneity was detected (I² of 33.4%). No interaction between CKD status and the effect of GLP-1 RAs on heart failure was found.

Fig. 4.

Incidence of heart failure with GLP-1 agonists compared with placebo in patients with and without chronic kidney disease (CKD). Results are stratified by CKD status. Data are presented as risk ratios (RR) with 95% confidence intervals (95%-CI). A similar incidence of heart failure is identified with GLP-1 agonists compared with placebo in patients with and without CKD. A random effects model is used

Expanded cardiovascular outcome

The LEADER study reported a significant reduction in an expanded cardiovascular outcome (CV death, nonfatal MI, nonfatal stroke, coronary revascularization, or hospitalization for unstable angina or heart failure) in patients with CKD treated with liraglutide compared to placebo (RR 0.77, 95% CI 0.67–0.89; P < 0.001) (Supplemental Table 1). In non-CKD patients, a non-significant reduction was observed compared to placebo (RR 0.94, 95% CI 0.86–1.03; P 0.199) (Supplemental Table 1) [15]. A differential treatment effect was identified (p-value for interaction of 0.02) suggesting a protective effect from GLP-1 RAs in the subgroup of patients with CKD at baseline.

Mortality

When compared to placebo, GLP-1 RAs were associated with a numerically lower incidence of mortality in patients with CKD (RR 0.88, 95% CI 0.66–1.17; P 0.378) (Fig. 5). The use of GLP-1 RAs was associated with a significant decrease is incidence of mortality in people without CKD compared to placebo (RR 0.84, 95% CI 0.73–0.95; P 0.007) (Fig. 5). High heterogeneity was detected (I² of 52.5%). There was no interaction between CKD status and the effect of GLP-1 RAs on mortality.

Fig. 5.

Incidence of mortality with GLP-1 agonists compared with placebo in patients with and without chronic kidney disease (CKD). Results are stratified by CKD status. Data are presented as risk ratios (RR) with 95% confidence intervals (95%-CI). A similar incidence of mortality is identified with GLP-1 agonists compared with placebo in patients with and without CKD. A random effects model is used

Discussion

This meta-analysis evaluates the effects of GLP-1 RA-based therapies on cardiovascular and renal outcomes in both diabetic and non-diabetic patients with or without CKD. In this report, we demonstrate that GLP-1 RAs reduce the absolute risk of MACE and renal outcomes regardless of CKD status. Importantly, these therapies maintain their protective effects even in patients with an eGFR < 60 ml/min/1.73m2. There is no significant interaction with renal function, and the benefits of GLP-1 RAs remain consistent irrespective of the timing of patient inclusion in the trials.

GLP-1 RAs are well-documented for their ability to lower glucose levels, induce modest weight loss, and reduce blood pressure. Glucagon-like peptide-1 (GLP-1) is an incretin hormone that binds GLP-1 receptors, which are found in multiple organ systems, such as digestive, cardiovascular and central nervous systems. Glucose-lowering effects are mediated by the stimulation of insulin release and the suppression of glucagon secretion, with a low risk of hypoglycemia [21]. Beyond their metabolic effects, GLP-1 RAs also modulate inflammation, an increasingly important factor in the pathophysiology of diabetes [1, 3]. GLP-1 receptor activation plays a protective role against oxidative injury, which is particularly relevant to its cardiovascular and kidney protective effects [22, 23].

The cardioprotective effects of GLP-1 RAs were primarily attributed to improved glycemic control, but also result from direct actions on GLP-1 receptors located on cardiomyocytes and vascular endothelial cells, contributing to vasodilation and direct cardiac protection. Additionally, GLP-1 receptor expression in other tissues leads to natriuresis, weight reduction, improved lipid profiles, reduced plaque formation, and neurohormonal regulation [4]. Moreover, the significant results of the SELECT study on cardiovascular outcomes in a population of overweight or obese non-diabetic individuals confirm the clinical benefit of GLP-1 RAs beyond glycemic control [18]. The renal protective effects of GLP-1 RAs are observed through various mechanisms, including improved glucose control and blood pressure regulation [24]. These therapies have been shown to prevent the onset of macroalbuminuria and slow the progression of glomerular filtration rate (GFR) decline in diabetic patients [25]. Importantly, in patients with CKD, who have a higher baseline risk of cardiovascular and renal events, the benefits on cardiovascular and renal outcomes are not diminished even as kidney function declines [26]. The FLOW study further demonstrated the improvement in the primary renal outcome, defined as the time to first kidney failure (persistent eGFR < 15 mL/min/1.73 m² or initiation of chronic renal replacement therapy), ≥ 50% eGFR decline from baseline, or death from renal or cardiovascular causes, thus showing for the first time a beneficial impact of GLP-1 RAs on an even more clinically significant renal outcome [13]. The precise distribution of GLP-1 receptors within the kidney is still being explored, but they are thought to be located in vascular smooth muscle cells and immune cells. Although GLP-1 RAs are known to inhibit the sodium-hydrogen exchanger-3 in renal tubules, it remains unclear if their kidney-protective effects are mainly due to hemodynamic changes. Presently, immunomodulatory actions are believed to play a central role in the renoprotective effects of GLP-1 RAs [3, 26–28].

Additionally, from a mechanistic perspective, it remains to be determined whether the renal benefits of GLP-1 RAs are mediated through a direct action on the kidneys and/or indirectly, particularly via weight reduction.

It is worth noting that several cardio-renal benefits of GLP-1 RAs were present for both molecules with a shorter half-life and those with a prolonged half-life administered weekly.

Our study has several limitations. First, we categorized eGFR as a dichotomous variable due to insufficient data for stratification by creatinine, albuminuria, or detailed eGFR ranges. The definitions of cardiovascular and renal outcomes varied across studies, which may have influenced the comparability of results. Many studies lacked stratification by CKD status, and few reported data based on albuminuria, an essential component of CKD definition. Certain outcomes, such as myocardial infarction or stroke, were evaluated in only a few studies, and in some instances, outcomes like non-fatal myocardial infarction or stroke were assessed in just one study. All included studies enrolled diabetic patients, except for the SELECT study, which included 17,000 non-diabetic patients but only 11% with CKD. Therefore, future prospective studies evaluating GLP-1 RAs in patients with obesity and CKD without diabetes are needed to better assess the benefits of GLP-1 RAs in populations for whom this treatment is indicated. It should also be noted that currently available data from clinical trials on patients treated with GLP-1 RAs who have advanced CKD, particularly those with an eGFR < 30 mL/min, remain very limited and limits the generalizability of our findings in this population. Since treatment with GLP-1 RAs is possible in patients with an eGFR < 30 mL/min, it would be valuable to have future data on this specific population to better evaluate whether a cardio-renal benefit exists in these individuals, whose numbers are steadily increasing. Currently, there is no available data on cardio-renal outcomes based on the presence or absence of CKD in RCTs evaluating the efficacy of the dual GLP-1/GIP RA, tirzepatide, particularly in the SURPASS and SURMOUNT programs [29, 30]. Therefore, such data will be of significant interest. This molecule has already demonstrated significant benefits in terms of weight reduction and glycemic control. It also appears to be safe in patients with advanced CKD, including those on dialysis, by promoting fat mass reduction without compromising patient stability [31–33]. However, the currently available data remain very limited, based on small cohorts and not derived from prospective randomized controlled trials. Therefore, further studies are needed in the future.

Despite these constraints, our meta-analysis has several unique strengths. It includes data from 17 568 patients with CKD across 9 trials with diverse inclusion criteria, representing a broad patient population. The absence of significant heterogeneity for major renal and cardiovascular outcomes and the consistency of results across various analyses further strengthen our findings.

An important aspect of our analysis is the consistency of the benefits observed across different patient populations. GLP-1 RAs demonstrated a consistent reduction in both cardiovascular and renal events independently of diabetes or CKD status. Moreover, the effectiveness of GLP-1 RAs did not vary based on the timing of patient inclusion in the studies. Whether patients were enrolled at earlier or later stages of CKD, the protective effects on cardiovascular and renal outcomes remained. Future meta-analyses using patient-level data could provide a more granular understanding of how renal function impacts the safety and efficacy of GLP-1 RAs. Such studies could incorporate standardized definitions for renal outcomes, such as the composite of 40% GFR decline, ESKD, or renal death, as primary endpoints, which would improve comparability. Moreover, future research should aim to include more ethnically diverse populations, especially those from underserved areas.

In conclusion, GLP-1 RAs offer substantial cardiovascular and renal protection, making them a valuable treatment option for patients with CKD. These findings support their use in CKD patients with eGFR as low as 60 ml/min/1.73 m², and confirms that these therapies may be continued as kidney function declines, until the need for renal replacement therapy arises.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Karim Gariani conceptualised the study, wrote the methodology, undertook investigation and formal analysis, visualised the data, reviewed and edited the manuscript. Aliona Siniukovich undertook investigation and formal analysis, visualised the data and wrote the original draft, and reviewed and edited the manuscript. Christoforos Travlos was the second reviewer for the purposes of screening and full-text evaluation, undertook investigation and formal analysis, visualised the data, reviewed and edited the manuscript. Sarah Freedman undertook investigation and formal analysis, visualised the data, reviewed and edited the manuscript. Thomas Mavrakanas reviewed and edited the manuscript. Karim Gariani and Thomas Mavrakanas are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Funding

Open access funding provided by University of Geneva. Open access funding provided by University of Geneva. The project was supported by the Swiss National Science Foundation (#10005333).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics and consent to participate

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.