Effects of semaglutide with and without concomitant SGLT2 inhibitor use in participants with type 2 diabetes and chronic kidney disease in the FLOW trial

Type 2 diabetes (T2D) and chronic kidney disease (CKD) confer high risks of kidney failure and cardiovascular (CV) complications, underscoring the need to optimize glucose-lowering therapies to prevent death, kidney failure and CV events. Early GLP-1 RA and SGLT2 inhibitor trials, largely powered for CV outcomes, suggested kidney benefits; later dedicated kidney outcome trials showed substantial kidney protection with SGLT2i. The FLOW trial was the first kidney outcomes trial of a GLP-1 RA (semaglutide 1.0 mg once weekly) and demonstrated kidney, CV, and survival benefits versus placebo. Given potentially complementary mechanisms of GLP-1 RAs and SGLT2i, it has been hypothesized that combined therapy could further improve outcomes. This prespecified FLOW analysis evaluated efficacy and safety of semaglutide stratified by baseline SGLT2i use and assessed effects of SGLT2i initiated post-randomization.

Prior CV outcome trials of GLP-1 RAs and SGLT2i indicated CV benefits and suggested renal protection, though kidney outcomes were often secondary. Subsequent dedicated SGLT2i CKD trials (including patients with and without T2D) established primary kidney benefits. Observational and meta-analytic data suggest potential additive benefits of GLP-1 RA and SGLT2i combination on MACE and serious kidney events versus either alone. However, no randomized trials have directly tested the combination on hard kidney and CV outcomes in T2D with CKD. In SGLT2i kidney outcome trials, GLP-1 RA background use was rare (<5%), and subgroup analyses showed no modification of SGLT2i benefits by GLP-1 RA use. Conversely, in AMPLITUDE-O (GLP-1 RA vs placebo), stratified by SGLT2i use, overall benefits appeared independent of SGLT2i, with signals for heterogeneity in specific components. Short-term studies reported additive effects of the combination on surrogates (weight, BP, HbA1c). Current guidelines position SGLT2i as foundational therapy for T2D with CKD, and FLOW suggests semaglutide provides complementary protection.

Design: Prespecified subgroup analysis within FLOW, a randomized, double-blind, placebo-controlled, international, multicenter kidney outcomes trial. Participants were randomized 1:1 to once-weekly subcutaneous semaglutide 1.0 mg or matching placebo via a central system (June 2019–May 2021), stratified by baseline SGLT2i use (yes/no). SGLT2i use was permitted and could be initiated during the trial. Population: Adults (≥18 years) with T2D, HbA1c ≤10% (≤86 mmol/mol) and CKD defined by creatinine-based eGFR 25–75 ml/min/1.73 m2 with albuminuria thresholds (UACR >300 to <5,000 mg/g if eGFR ≥50; UACR >100 to <5,000 mg/g if eGFR 25 to <50), on stable maximally labeled/tolerated renin-angiotensin system inhibitor (RASi) therapy unless contraindicated/intolerant. Up to 20% could have eGFR ≥60. Key exclusions included prior GLP-1 RA within 30 days, certain kidney diseases (e.g., polycystic kidney disease, glomerulonephritis), advanced heart failure (NYHA IV), recent major CV events, active malignancy within 5 years (exceptions), dialysis, and other safety-related criteria. Intervention: Semaglutide 1.0 mg weekly vs matching placebo, in addition to standard of care, including RASi and allowed SGLT2i. Outcomes: Primary outcome was a five-component composite: sustained (≥28 days) ≥50% eGFR reduction from baseline; kidney failure (initiation of chronic dialysis, kidney transplantation, or sustained eGFR <15 ml/min/1.73 m2); or death due to kidney or CV causes. A kidney-specific four-component outcome excluded CV death. Confirmatory secondary outcomes (hierarchical): (1) total eGFR slope (annual rate of eGFR change), (2) time to first MACE (nonfatal MI, nonfatal stroke, CV death), and (3) all-cause death. Supportive outcomes included UACR, HbA1c, body weight, and safety. Serum creatinine was centrally measured; cystatin C was also measured at prespecified times to derive eGFRcys and post hoc renal endpoints without persistence confirmation requirements. Statistical analysis: Intention-to-treat. Time-to-event endpoints analyzed with stratified Cox models (stratified by SGLT2i use at baseline), with treatment and subgroup as fixed factors; cumulative incidence estimated by Aalen-Johansen with appropriate competing risks. Subgroup interaction tested (score test). Time-dependent Cox models assessed SGLT2i use initiated after randomization (as time-varying yes/no), with and without interaction with treatment. eGFR slopes analyzed via linear mixed-effects models with random intercepts/slopes and fixed effects for treatment, SGLT2i subgroup, time, and interactions; missing scheduled eGFR not imputed. Continuous endpoints (e.g., change to week 104) analyzed by ANCOVA with treatment by SGLT2i subgroup, baseline adjustment, multiple imputation for missing values (Rubin's rules); log-transformations applied per SAP for ratio outcomes. Two-sided P<0.05 considered significant; no multiplicity adjustment. Analyses performed in SAS 9.4. Trial conduct: Event-driven, designed for 90% power to detect 20% relative risk reduction in the primary outcome. A single interim analysis (~two-thirds of planned events) led the IDMC to recommend early completion for efficacy (October 2023). Median follow-up 3.4 years. Screening: 5,581; randomized: 3,533 (semaglutide n=1,767; placebo n=1,766). Baseline SGLT2i use: 550 (15.6%). Adherence to randomized treatment ~89% of planned time; vital status known in ~98.6%.

- Primary outcome: Overall semaglutide reduced the composite kidney/CV death outcome by 24% vs placebo (HR 0.76; 95% CI 0.66, 0.88). Among baseline SGLT2i users: 41/277 (14.8%) vs 38/273 (13.9%), HR 1.07 (95% CI 0.69, 1.67; P=0.755). Among non-users: 290/1,490 (19.5%) vs 372/1,493 (24.9%), HR 0.73 (95% CI 0.63, 0.85; P<0.001). No significant interaction by SGLT2i use (Pinteraction 0.109). - Kidney-specific four-component outcome (excluding CV death): With SGLT2i use HR 1.18 (95% CI 0.71, 1.98); without SGLT2i HR 0.75 (95% CI 0.61, 0.90; P=0.003); Pinteraction 0.100. - eGFR slopes (creatinine-based): Annual eGFR decline was slower with semaglutide vs placebo. Between-group differences: +0.75 ml/min/1.73 m2/year (95% CI −0.01, 1.50) in SGLT2i users and +1.25 (95% CI 0.91, 1.58) in non-users; Pinteraction 0.237. - MACE: Overall HR 0.82 (95% CI 0.68, 0.98). No interaction by SGLT2i use (Pinteraction 0.741). Component outcomes showed no effect modification (P=0.303 to 0.911). - All-cause death: Overall HR 0.80 (95% CI 0.67, 0.95). No interaction by SGLT2i use (Pinteraction 0.901). - Supportive outcomes: At week 104, semaglutide lowered UACR vs placebo by 24% (95% CI 4%, 39%) with SGLT2i and by 34% (95% CI 26%, 40%) without (Pinteraction 0.279). HbA1c change from baseline favored semaglutide: −8.2 vs +1.2 mmol/mol (with SGLT2i) and −9.7 vs −1.0 mmol/mol (without) (Pinteraction 0.606). Body weight change favored semaglutide: −6.1 vs −1.1 kg (with SGLT2i) and −6.3 vs −1.5 kg (without) (Pinteraction 0.746). - Safety: Serious adverse events were similar between semaglutide and placebo within both SGLT2i subgroups (approximately 48–54%); no new safety concerns for combining semaglutide with SGLT2i. - Initiation of SGLT2i during trial: Among participants using SGLT2i at baseline or who initiated during study, HR for primary outcome with semaglutide vs placebo was 0.88 (95% CI 0.66, 1.17). Among those never using SGLT2i, HR was 0.70 (95% CI 0.59, 0.82); Pinteraction 0.169. Time-dependent Cox (SGLT2i as time-varying covariate): HR 0.75 (95% CI 0.65, 0.86). With SGLT2i-by-treatment interaction, HRs were 0.92 (95% CI 0.64, 1.33) in SGLT2i users during study and 0.72 (95% CI 0.61, 0.84) in never-users. - Post hoc cystatin C analyses: Five-component primary outcome HRs were 0.74 (95% CI 0.47, 1.16) with SGLT2i and 0.70 (95% CI 0.60, 0.82) without; interactions across renal outcomes were non-significant (Pinteraction 0.799–0.983). eGFRcys slope differences: +0.92 (95% CI 0.16, 1.68) with SGLT2i and +1.55 (95% CI 1.21, 1.88) without (Pinteraction 0.142). Week-104 eGFR changes favored semaglutide by ~3.2–3.7 ml/min/1.73 m2 with no interaction (Pinteraction 0.686 creatinine-based; 0.901 cystatin C-based).

Semaglutide’s benefits on kidney function decline, MACE, and all-cause mortality were consistent irrespective of baseline SGLT2i use, suggesting independent and potentially complementary mechanisms. Although in the smaller subgroup using SGLT2i at baseline, the composite kidney outcomes did not reach statistical significance, interaction testing did not indicate heterogeneity by SGLT2i use. The more pronounced effect of semaglutide on eGFR slopes and UACR across subgroups supports sustained renal benefits that may require longer follow-up or larger samples of combined therapy users to translate into hard kidney endpoints. Observational data and meta-analyses suggest potential additivity of GLP-1 RAs plus SGLT2i, and the pattern in FLOW is compatible with additive or at least non-antagonistic effects, particularly for CV and survival outcomes. These findings reinforce semaglutide as an effective therapy alongside foundational SGLT2i and RASi in T2D with CKD.

In FLOW, semaglutide reduced major kidney disease events, slowed eGFR decline, and lowered risks of MACE and all-cause death, with no evidence that these benefits are modified by concomitant SGLT2i use. Given the substantial and complementary benefits of GLP-1 RAs and SGLT2i and acceptable safety of their combination, concurrent therapy may be considered for patients with T2D and CKD. Future dedicated randomized trials directly comparing SGLT2i, GLP-1 RA, and their combination on kidney and CV outcomes are warranted to quantify potential additive effects.

- Limited power for interaction analyses due to low baseline SGLT2i use (15.6%) and fewer kidney events occurring within the trial timeframe among SGLT2i users. - Differential, uncontrolled initiation of SGLT2i post-randomization (more frequent in placebo), potentially diluting between-group differences. - Median follow-up of 3.4 years may be insufficient to detect differences in hard kidney endpoints within the small combined-therapy subgroup; eGFR and UACR benefits were clearer. - Post hoc cystatin C analyses, while consistent, were exploratory and did not require persistence confirmation for eGFR thresholds. - No multiplicity adjustment across multiple secondary and subgroup analyses.