Long-term kidney outcomes of semaglutide in obesity and cardiovascular disease in the SELECT trial

Obesity is a risk factor for decline in glomerular filtration rate (GFR) and increased albuminuria, and even in individuals without diabetes it increases the risk of chronic kidney disease (CKD; eGFR < 60 ml min⁻¹ 1.73 m²). Evidence from large cohorts and Mendelian randomization supports a causal link between higher body mass index and CKD risk. Proposed mechanisms include hyperfiltration, inflammation, oxidative stress, increased tubular sodium reabsorption, and activation of the renin–angiotensin–aldosterone system, with pathological features such as ectopic kidney lipid accumulation and obesity-related glomerulopathy. An ideal therapy for obesity-related CKD would reduce body weight and directly affect kidney injury pathways. Glucagon-like peptide-1 receptor agonists (GLP-1RAs) may have kidney-protective effects. Prior cardiovascular outcome trials in type 2 diabetes suggest GLP-1RAs improve composite kidney outcomes, but benefits in people with overweight/obesity without diabetes were unclear. The SELECT trial showed a 20% reduction in major adverse cardiovascular events with semaglutide 2.4 mg weekly in adults with pre-existing cardiovascular disease and BMI ≥27 kg m⁻² without diabetes. This pre-specified analysis evaluates semaglutide’s effects on kidney outcomes in SELECT.

Secondary analyses from large cardiovascular outcome trials of GLP-1RAs in type 2 diabetes have shown improvements in composite kidney outcomes. A meta-analysis across ELIXA, LEADER, SUSTAIN-6, EXSCEL, REWIND and AMPLITUDE-O reported a 21% reduction in a composite of macroalbuminuria, substantial kidney function decline, kidney replacement therapy, or kidney death, with trial-specific effects ranging from 12% to 36% reductions. Exploratory analyses in STEP 2 (semaglutide 1.0 and 2.4 mg) in people with diabetes and overweight/obesity showed significant reductions in UACR. Post-hoc analyses of tirzepatide (SURPASS-4) suggested potential kidney benefits. The present work extends these observations to individuals with overweight/obesity without diabetes. Contextual frameworks include evidence linking reductions in eGFR slope and UACR with lower long-term risks of kidney failure. Mechanistic literature suggests GLP-1RAs may exert direct renal effects including glomerular hemodynamic changes, anti-inflammatory and antioxidative actions, natriuresis/diuresis, and modulation of RAGE expression.

SELECT was a randomized, double-blind, placebo-controlled, event-driven trial comparing once-weekly subcutaneous semaglutide 2.4 mg versus placebo added to standard of care in 17,604 adults with established cardiovascular disease and overweight/obesity (BMI ≥27 kg m⁻²) without type 2 diabetes. Key inclusion criteria: age ≥45 years; established CVD (prior myocardial infarction, prior ischemic or hemorrhagic stroke, or symptomatic peripheral artery disease). Key exclusions: history of type 1 or type 2 diabetes; HbA1c ≥6.5%; end-stage kidney disease or need for dialysis; recent (within 60 days) MI, stroke, unstable angina hospitalization, or transient ischemic attack; NYHA class IV heart failure. Participants were randomized 1:1 to semaglutide (dose-escalation over 16 weeks to 2.4 mg weekly) or placebo. The main pre-specified 5-component kidney endpoint was time to first occurrence of: death from kidney disease; initiation of chronic kidney replacement therapy (dialysis or transplantation); onset of persistent eGFR <15 ml min⁻¹ 1.73 m²; persistent ≥50% reduction in eGFR from baseline; or onset of persistent macroalbuminuria. Persistent was defined as two or more qualifying measurements ≥4 weeks apart. Additional kidney outcomes included alternative composite endpoints (for example, excluding macroalbuminuria or including cardiovascular death), time to specified thresholds of eGFR decline (30%, 40%, 50%, 57%), change in eGFR from baseline to week 104, annualized eGFR slope (acute, chronic, total), and change in UACR to week 104. Outcomes were assessed overall and by baseline eGFR and UACR subgroups. Safety included treatment-emergent adverse events and adjudicated acute kidney failure. Laboratory assessments were performed centrally. Serum creatinine was measured at screening and at weeks 20, 52, 104, 156, 208, and end of treatment; eGFR was calculated using CKD-EPI 2009 creatinine equation, with sensitivity analyses using CKD-EPI 2021 equation (data not shown). UACR was measured from single urine samples at randomization and at weeks 20, 52, 104, 156, 208, and end of treatment; confirmatory testing was required for macroalbuminuria (>300 mg g⁻¹) and for eGFR events. European participants had additional labs at weeks 4, 8, 12, 16. All components of the main kidney composite and acute kidney failure were adjudicated by an independent, blinded committee. Efficacy analyses followed intent-to-treat (full analysis set) and on-treatment definitions. Time-to-event outcomes were analyzed with Cox proportional hazards models (Exact or Efron method for ties). Longitudinal continuous endpoints (eGFR, UACR) used mixed models for repeated measures (MMRM) with scheduled visits as fixed effects and treatment-by-visit interactions; REML estimation was used. eGFR slopes (acute: baseline–week 16; chronic: week 20–end; total: baseline–end) were estimated via linear random effects models including time, treatment, and time-by-treatment interactions with random patient effects. Exploratory correlation and mediation analyses evaluated relationships between eGFR change and changes in body weight, systolic blood pressure, and HbA1c at week 104. Significance was two-sided at 5% without multiplicity adjustment. SAS v9 was used.

- Population: 8,803 randomized to semaglutide 2.4 mg; 8,801 to placebo; median follow-up 182 weeks. Baseline characteristics, including kidney function and albuminuria status, were balanced. Just over one-fifth had eGFR <60 ml min⁻¹ 1.73 m² or UACR ≥30 mg g⁻¹. - Treatment exposure and completion: Permanent discontinuation occurred in 26.7% (semaglutide) vs 23.6% (placebo); treatment exposure time 82.5% vs 87.7%, respectively. Trial completion rate 96.9%. - Main 5-component kidney composite endpoint: Incidence 1.8% (semaglutide) vs 2.2% (placebo); HR 0.78 (95% CI 0.63, 0.96); P=0.02 (in-trial). On-treatment HR 0.75 (95% CI 0.59, 0.94); P=0.01. Reduction driven by lower onset of macroalbuminuria and persistent ≥50% eGFR decline; other components were sparse. - Subgroups: No significant treatment interactions across prespecified subgroups. Example: Baseline eGFR ≥60 ml min⁻¹ 1.73 m², HR 0.64 (0.48, 0.85); baseline eGFR <60, HR 0.97 (0.70, 1.34). Among those on ACEi/ARB, HR 0.74 (0.58, 0.94) vs 0.92 (0.60, 1.42) in those not on ACEi/ARB (interaction P=0.39). - eGFR change at week 104 (MMRM): Overall, semaglutide −0.86 vs placebo −1.61 ml min⁻¹ 1.73 m²; estimated treatment difference (ETD) 0.75 (95% CI 0.43, 1.06); P<0.001. By baseline eGFR: <60 ml min⁻¹ 1.73 m²: +5.28 vs +3.09; ETD 2.19 (1.00, 3.38); P<0.001. ≥60: −1.62 vs −2.20; ETD 0.57 (0.26, 0.89); P<0.001. - eGFR slopes: Total slope per year −0.78 (semaglutide) vs −1.17 (placebo); between-group difference 0.39 ml min⁻¹ 1.73 m² per year (95% CI 0.30, 0.48); P<0.001. Chronic slope difference 0.29 (0.18, 0.40); P<0.001. Acute period (European subset) showed an initial greater decline with semaglutide (acute slope difference −1.33 ml min⁻¹ 1.73 m² per year; 95% CI −2.68, 0.02; P=0.05), with convergence by week 20. - UACR at week 104: Overall net treatment benefit −10.7% (95% CI −13.2, −8.2); P<0.001. By baseline UACR: <30 mg g⁻¹: −8.1% (−10.6, −5.6); P<0.001. 30–<300: −27.2% (−35.3, −18.1); P<0.001. ≥300: −31.4% (−54.9, 4.3); P=0.08. By baseline eGFR: <60 ml min⁻¹ 1.73 m²: −13.1% (−22.1, −3.1); P=0.01; ≥60: −10.9% (−13.5, −8.4); P<0.001. - Other kidney composites: A 5-component endpoint excluding incident macroalbuminuria but including cardiovascular death had HR 0.82 (95% CI 0.69, 0.97); P=0.02. Additional composites excluding macroalbuminuria were directionally favorable but not significant due to few events. - Correlation/mediation: Little correlation between within-person eGFR change and changes in body weight, systolic blood pressure, or HbA1c. Exploratory mediation suggested 81% (95% CI 41.3, 120) of eGFR change might be attributable to weight change, with wide imprecision. - Safety: More adverse events leading to discontinuation with semaglutide (16.6%) vs placebo (8.2%), predominantly gastrointestinal. Discontinuations were more common in those with baseline eGFR <60 in both arms. No excess of adjudicated acute kidney failure with semaglutide, irrespective of baseline eGFR.

Allocation to semaglutide 2.4 mg weekly in adults with overweight/obesity and established CVD but without diabetes reduced the risk of a pre-specified kidney composite endpoint by 22%, primarily via reductions in macroalbuminuria and persistent ≥50% eGFR decline. Benefits were consistent across key subgroups with no significant interactions, and an alternative composite including cardiovascular death also showed an 18% risk reduction. Continuous endpoints demonstrated modest but statistically significant preservation of eGFR overall and a larger benefit among participants with baseline eGFR <60 ml min⁻¹ 1.73 m², alongside favorable effects on total and chronic eGFR slopes. UACR increased less overall with semaglutide and fell substantially more among patients with baseline albuminuria. These magnitudes of eGFR slope preservation and UACR reduction are aligned with surrogate thresholds predictive of long-term kidney protection. The findings extend prior evidence of kidney benefits with GLP-1RAs observed in people with type 2 diabetes to a population without diabetes, suggesting class and dose effects may confer renoprotection beyond glycemic control. Mechanisms may include weight loss, hemodynamic changes, anti-inflammatory and antioxidative effects, natriuresis/diuresis, and modulation of RAGE, though mediation analyses were underpowered and imprecise. An initial transient eGFR decline with semaglutide resembles the early dip seen with RAS blockers and SGLT2 inhibitors and may signal favorable long-term trajectories. Interpretation of creatinine-based eGFR in the context of weight loss warrants caution due to potential changes in muscle mass and indexing, underscoring the future value of cystatin C measurements. Overall, results support semaglutide as a therapy with beneficial kidney effects in individuals with overweight/obesity and high cardiovascular risk without diabetes.

In SELECT, once-weekly semaglutide 2.4 mg significantly improved kidney outcomes versus placebo in adults with overweight/obesity and established CVD without diabetes, reducing a 5-component kidney composite by 22%, attenuating eGFR decline (and improving eGFR among those with baseline eGFR <60 ml min⁻¹ 1.73 m²), and lowering UACR. These benefits were consistent across subgroups and accompanied by favorable eGFR slope differences. The results provide the first robust evidence of kidney benefits of a GLP-1RA in a non-diabetic population with overweight/obesity. Future research should clarify mechanisms (including the role of weight loss vs direct renal effects), incorporate cystatin C-based assessments, and evaluate long-term kidney outcomes in dedicated trials across CKD severities and etiologies, including confirmation from the FLOW trial and mechanistic studies.

- Event numbers for some hard renal outcomes were small, limiting power for certain composite endpoints and subgroup contrasts. - Formal mediation analyses for the main composite were underpowered; the exploratory mediation of eGFR change by weight change had wide confidence intervals. - Creatinine-based eGFR may be confounded by weight loss–related changes in muscle mass and by indexing to body surface area, potentially affecting interpretation of eGFR changes; cystatin C was not measured. - Early eGFR timepoints (acute slope) were available only in European participants, limiting generalizability of acute-phase findings. - P values were not adjusted for multiplicity across multiple endpoints and subgroups. - Higher discontinuation rates with semaglutide (mainly gastrointestinal) could influence on-treatment analyses, although intent-to-treat findings were consistent.