Kidney Outcomes With Glucagon-Like Peptide-1 Receptor Agonists Versus Other Glucose-Lowering Agents in People With Type 2 Diabetes: A Systematic Review and Meta-Analysis of Real-World Data

Abstract

Aims: Randomized placebo-controlled clinical trials showed that glucagon-like peptide-1 receptor agonists (GLP-1 RA) reduce kidney risk in patients with type 2 diabetes (T2D), prominently in those with chronic kidney disease. It is unclear whether these findings may apply to broader populations of patients with T2D treated in real-world settings and compared to active controls. We summarised real-world data of adverse kidney outcomes among patients with T2D initiating GLP-1 RA versus other glucose-lowering agents.

Materials and methods: We searched PubMed and Embase for observational cohort studies (April 2005-January 2025; PROSPERO CRD42023405356). Initiators of GLP-1 RA were compared to sodium-glucose cotransporter-2 inhibitors (SGLT2i), dipeptidyl-peptidase 4 inhibitors (DPP4i), sulfonylureas, or basal insulin. Outcomes included risks of albuminuria progression, ≥ 40 or ≥ 50% eGFR reduction from baseline, acute kidney injury (AKI), kidney-related hospitalizations, and end-stage kidney disease (ESKD), per data availability. We synthesised the data using inverse variance-weighted averages of logarithmic hazard ratios (HR)s in random-effect models.

Results: Thirty-one studies were eligible, encompassing 1,601,389 patients (mean age 49-78 years, 5%-64% women), with 21, 6, 5, and 1 of them using SGLT2i, DPP4i, basal insulin, and sulfonylureas as a comparator, respectively. Compared with SGLT2i, GLP-1 RA initiators had higher risks for AKI (HR [95% CI] 1.12 [1.05-1.20]), kidney-related hospitalizations (1.66 [1.01-2.73]), and ≥ 40% reduction in eGFR (1.40 [1.27-1.53]), without evidence for differences in risks of ≥ 50% eGFR reduction or ESKD. Compared to DPP4i, GLP-1 RA initiators had lower risks for experiencing ≥ 50% eGFR reduction (0.84 [0.76-0.92]), kidney-related hospitalizations (0.73 [0.65-0.83]), and ESKD (0.70 [0.63-0.78]). Similar benefits were observed when comparing GLP-1 RA to sulfonylureas. Compared to basal insulin, GLP-1 RA initiation was associated with a lower risk of albuminuria progression (0.89 [0.80-0.99]), with inconsistent data regarding possible benefits in reducing ESKD risk.

Conclusions: In patients with T2D, initiation of GLP-1 RA in real-world settings may be associated with improved kidney outcomes compared to DPP4i, sulfonylureas, and basal insulin, and worse kidney outcomes compared to SGLT2i.

Keywords: GLP‐1 RA; kidney outcomes; meta‐analysis; real world data; systematic review; type 2 diabetes.

FIGURE 1

PRISMA flow‐chart. The search was conducted via PubMed and Embase between April 1, 2005, and January 1, 2025. Eligible manuscripts included observational cohort studies examining kidney outcomes in patients with T2D initiating GLP‐1 RA versus other GLAs (SGLT2i, DPP4i, sulfonylureas, and basal insulin). DPP4i, dipeptidyl peptidase‐4 inhibitors; GLAs, glucose‐lowering agents; GLP‐1 RA, glucagon‐like peptide‐1 receptor agonist; SGLT2i, sodium‐glucose transporter 2 inhibitor; T2D, type 2 diabetes.

FIGURE 2

Meta‐analysis of kidney outcomes in patients with type 2 diabetes initiating GLP‐1 RA versus SGLT2i in real‐world studies. We assessed categorical kidney outcomes (new incident of macroalbuminuria, ≥ 40 or 50% reduction from baseline eGFR, doubling of serum creatinine, hospitalizations due to kidney events, AKI, and ESKD) among initiators of GLP‐1 RA versus initiators of SGLT2i. Data were synthesised using inverse variance‐weighted averages of logarithmic hazard ratios (HR)s in random‐ and fixed‐effect models, presented as HR [95% CI] including p‐value and studies heterogeneity (I²). We selected random models as the main analysis because of the heterogeneity that is common when combining real‐world studies. Four studies [14, 21, 22, 25] that assessed the risk of AKI and one study [31] that assessed the risk of ESKD were not synthesised in the meta‐analysis due to an overlap between the datasets (see Table 2). AKI, acute kidney injury; CI, confidence interval; eGFR, estimated glomerular filtration rate; ESKD, end‐stage kidney disease; GLP‐1 RA, glucagon‐like peptide‐1 receptor agonist; HR, hazard ratio; N, Population size; SGLT2i, sodium‐glucose transporter 2 inhibitor.

FIGURE 3

Meta‐analysis of kidney outcomes in patients with T2D initiating GLP‐1 RA versus DPP4i in real‐world studies. We assessed the risk for ESKD among initiators of GLP‐1 RA versus initiators of DPP4i. Data were synthesised using inverse variance‐weighted averages of logarithmic hazard ratios (HR)s in random‐ and fixed‐effect models. Presented as HR [95% CI] including p‐value and studies heterogeneity (I²). We selected random models as the main analysis because of the heterogeneity that is common when combining real‐world studies. CI, confidence interval; DPP4i, dipeptidyl peptidase‐4 inhibitors; ESKD, end‐stage kidney disease; GLP‐1 RA, glucagon‐like peptide‐1 receptor agonist; HR, hazard ratio; (N), Population size.

FIGURE 4

Kidney outcomes among patients with type 2 diabetes who initiated GLP‐1 RA versus other GLAs in real‐world studies. The colour represents the advantage/disadvantage of GLP‐1 RA versus the comparator: green, GLP‐1 RA advantage; red, comparator advantage; yellow‐no significant difference; grey, unavailable data. A hazard ratio (HR) with a 95% confidence interval is presented when available. Some data represent pooled HR from this meta‐analysis, while others correspond to individual studies, as noted. AKI, acute kidney injury; CKD, chronic kidney disease; DPP4i, dipeptidyl peptidase‐4 inhibitor; ESKD, end‐stage kidney disease; eGFR, estimated glomerular filtration rate; GLP‐1 RA, glucagon‐like peptide‐1 receptor agonist; HR, hazard ratio; SGLT2i, sodium‐glucose transporter two inhibitors; UACR, urinary albumin to creatinine ratio.