FLOW vs AWARD-7 vs LEADER Renal Substudy: GLP-1 Receptor Agonists and Kidney Outcomes Compared

By
Published Updated Last reviewed
Medication safety clinical consultation image for FLOW vs AWARD-7 vs LEADER Renal Substudy: GLP-1 Receptor Agonists and Kidney Outcomes Compared

At a Glance

At a glance

| Feature | FLOW (Perkovic 2024) | AWARD-7 (Tuttle 2018) | LEADER Renal Substudy (Mann 2017) | |---|---|---|---| | Drug | Semaglutide 1.0 mg SC weekly | Dulaglutide 0.75 mg or 1.5 mg SC weekly | Liraglutide 1.8 mg SC daily | | N | 3,533 | 577 | 9,340 (full trial); renal substudy n=9,340 | | T2D required | Yes | Yes | Yes | | Baseline eGFR range (mL/min/1.73 m²) | 50, 75 at screening; UACR ≥300 mg/g | 15, 59 | ≥15 (no upper exclusion; modal 60, 90 at baseline) | | Baseline UACR | ≥300 mg/g (macroalbuminuria required) | Not mandated; mean ~900 mg/g in overt nephropathy subgroup | Broad range; ~27% macroalbuminuria | | Median follow-up | 3.4 years | 52 weeks | 3.84 years | | Primary endpoint | Composite kidney/CV death endpoint (kidney failure, ≥50% eGFR decline, kidney death, CV death) | HbA1c reduction at 52 weeks (metabolic; renal secondary) | CV death, nonfatal MI, nonfatal stroke (MACE; renal was pre-specified secondary) | | Primary result | HR 0.76 (95% CI 0.66, 0.88); p<0.001 | Dulaglutide 1.5 mg attenuated eGFR decline vs. insulin glargine (, 1.7 vs. , 4.1 mL/min/1.73 m²; p=0.005) | New macroalbuminuria HR 0.74 (95% CI 0.60, 0.91); eGFR decline ≥40% HR 0.57 (95% CI 0.34, 0.96) | | Background ACEi/ARB | ~97% | ~73% | ~83% | | Background SGLT2i | ~5% (protocol pre-dated wide SGLT2i adoption; permitted after protocol amendment) | <1% (2016 to 2018 enrollment) | <1% (2010 to 2016) | | Dropout / discontinuation | 16.5% discontinued study drug | 9.9% | 9.4% | | Key adverse events | GI (nausea 16%, diarrhea 12%); no excess AKI | GI events higher with dulaglutide; no excess AKI | GI events; no excess AKI |

Population Differences, and What They Mean for Generalizability

The three trials enrolled meaningfully different patients, and that gap matters when clinicians try to extrapolate a result from one study to a patient in front of them.

FLOW required an eGFR between 50 and 75 mL/min/1.73 m² at screening and macroalbuminuria (UACR ≥300 mg/g). This is a targeted, high-risk CKD population. Roughly 97% were on renin-angiotensin-aldosterone system (RAAS) blockade, which is the contemporary standard of care defined by KDIGO guidelines. The median baseline eGFR was approximately 62 mL/min/1.73 m², so this is largely CKD G3a-G3b with heavy proteinuria. These patients carry real near-term risk of kidney failure, which is exactly why the trial could detect a hard endpoint in 3.4 years.

AWARD-7 went lower on kidney function (eGFR 15 to 59 mL/min/1.73 m²), meaning a portion of enrolled patients had CKD G4. That population is closer to late-stage diabetic kidney disease. However, AWARD-7 was a glycemic non-inferiority trial; kidney outcomes were pre-specified secondary endpoints. The comparator was insulin glargine, not placebo, which is an important distinction. Any renal advantage for dulaglutide must be interpreted against an active control whose own renal effects are not neutral: insulin's sodium-retaining properties and the hypoglycemia risk it carries may themselves worsen eGFR trajectory. AWARD-7 is the only trial of the three with an active glycemic comparator.

The LEADER renal substudy enrolled a broad cardiovascular-risk population. The entry eGFR threshold was ≥15 mL/min/1.73 m², and a substantial proportion of participants had preserved or mildly reduced kidney function at baseline. Only about 27% had macroalbuminuria. This spectrum actually improves generalizability to early CKD and even to patients without CKD who happen to have elevated CV risk and diabetes, but it dilutes the ability to detect hard kidney endpoints because most participants were too early in their kidney disease trajectory to reach kidney failure during a 3.84-year trial.

The age distribution and cardiovascular burden also differed. FLOW participants had a mean age of approximately 66 years and near-universal established cardiovascular disease or risk factors, consistent with its dual kidney/CV composite endpoint. LEADER enrolled patients aged ≥50 years with established CV disease or aged ≥60 years with ≥1 CV risk factor, making it somewhat more heterogeneous. AWARD-7 was the smallest by far (N=577) and had the shortest follow-up (52 weeks), which makes it difficult to draw conclusions about hard outcomes like kidney failure or mortality.

One generalizability concern cutting across all three trials: SGLT2 inhibitor use was negligible in all of them (under 5%), because enrollment preceded the era of routine SGLT2i prescribing in CKD. Current patients are increasingly on both drug classes simultaneously, and we do not have trial-level data on GLP-1 receptor agonist kidney benefits on top of established SGLT2i use. The KDIGO 2022 CKD guideline recommends SGLT2i as a first-line kidney-protective agent, and real-world patients today present with that layer already in place.

Methodology Differences

The single most important methodological divide is whether kidney outcomes were the primary or secondary target.

FLOW was explicitly powered and designed to detect a kidney composite outcome. Its primary endpoint was a composite of kidney failure (dialysis, transplant, or sustained eGFR <15), a sustained ≥50% eGFR decline from baseline, kidney-cause death, or cardiovascular death. Because CV death was included in this composite, the trial captures what KDIGO and most nephrologists consider the full burden of late CKD: patients with advanced kidney disease most often die of cardiovascular events before reaching dialysis. The trial was stopped early at a pre-specified interim analysis after crossing the superiority boundary.

AWARD-7's primary endpoint was HbA1c reduction. Kidney outcomes, including eGFR trajectory and UACR, were pre-specified secondary endpoints. This makes all AWARD-7 kidney findings exploratory in the strict frequentist sense; they were not adjusted for multiple comparisons in most published analyses.

The LEADER renal substudy used pre-specified secondary kidney outcomes from the parent cardiovascular outcomes trial (CVOT). New-onset macroalbuminuria (UACR crossing 300 mg/g), sustained ≥40% eGFR decline, and need for renal replacement therapy were each analyzed separately. No single composite kidney endpoint was specified as a primary renal outcome, and the trial was not powered for any individual kidney endpoint, which means the wide confidence intervals on hard renal outcomes must be taken seriously.

Blinding was double-blind and placebo-controlled in both FLOW and LEADER. AWARD-7 was open-label, comparing dulaglutide to insulin glargine. Open-label design in a metabolic trial raises concern about differential dropout and reporting bias, though eGFR measurement is objective and less susceptible to observer bias than patient-reported outcomes.

Background RAAS blockade differed materially. FLOW required near-universal ACEi/ARB use as part of optimized background therapy. LEADER's background ACEi/ARB rate (~83%) was high but not mandated. AWARD-7 had lower baseline RAAS use (~73%). If RAAS blockade modifies the renal benefit of GLP-1 receptor agonists (a plausible hypothesis, given that both reduce glomerular hypertension through different mechanisms), this differential background therapy complicates direct comparison.

Results, Matched

eGFR Decline Trajectory

All three trials found less eGFR decline in the GLP-1 receptor agonist arm, but the magnitude and statistical robustness vary considerably.

In FLOW, the rate of eGFR decline was approximately 1.16 mL/min/1.73 m² per year slower with semaglutide than placebo. This was measured over a 3.4-year follow-up in a population already losing eGFR rapidly. The total eGFR difference accumulated to a clinically meaningful margin by end of trial.

In AWARD-7, dulaglutide 1.5 mg reduced eGFR decline by 2.4 mL/min/1.73 m² relative to insulin glargine over 52 weeks (, 1.7 vs. , 4.1; p=0.005). This appears numerically larger than FLOW's annualized figure, but the comparison is confounded by the active-control design (insulin glargine may accelerate eGFR decline relative to placebo), the short follow-up, and the deeper CKD at baseline.

The LEADER renal substudy reported a significantly lower rate of new macroalbuminuria (HR 0.74) and a nominally significant reduction in sustained ≥40% eGFR decline (HR 0.57 to 95% CI 0.34, 0.96). The eGFR trajectory across the full trial also favored liraglutide, though the CKD-specific subgroup was not separately powered.

Composite Renal Endpoint Reduction

Only FLOW had a prospectively powered composite kidney endpoint, and the result was clear: HR 0.76 (95% CI 0.66, 0.88), corresponding to a 24% relative risk reduction. The number needed to treat to prevent one primary composite event was approximately 20 over 3.4 years.

AWARD-7 did not report a composite kidney endpoint as a pre-defined outcome. LEADER reported a post-hoc composite of kidney outcomes that was directionally favorable but was not the trial's primary or pre-specified secondary kidney endpoint.

We cannot quote a matched relative risk reduction across the three trials for a single composite definition because the composites are not the same. FLOW's composite included CV death; LEADER's kidney-specific analyses did not.

All-Cause and CV Mortality

FLOW's primary composite included CV death, and the overall results were significant. However, the trial was not powered separately for all-cause mortality, and the point estimate for all-cause mortality alone trended favorable but did not reach statistical significance as an isolated endpoint in the primary analysis.

LEADER (the parent trial) demonstrated a statistically significant reduction in CV death with liraglutide (HR 0.78 to 95% CI 0.66, 0.93) in the full CVOT population. Whether this mortality benefit is larger in the CKD subgroup is a hypothesis, not a finding, from the renal substudy.

AWARD-7's 52-week duration was far too short to make meaningful statements about mortality. No deaths attributable to kidney failure or CV events were reported as a notable outcome difference at that follow-up.

What the Trials Together Do and Do Not Establish

Taken together, the three trials make a credible, class-level argument that GLP-1 receptor agonists slow kidney disease progression in people with type 2 diabetes and CKD. The signal is directionally consistent across three different molecules (semaglutide, dulaglutide, liraglutide), three different trial designs, and three different follow-up durations. That consistency matters. A class effect is more plausible when it appears with structural analogs, and GLP-1 receptor agonists share anti-inflammatory, blood-pressure-lowering, and glomerular-hemodynamic mechanisms that are not purely glycemic.

What the trials do not establish is equally important to state clearly.

First, there is no head-to-head comparison of these three agents. FLOW's 24% composite reduction in a high-CKD population cannot be directly compared to LEADER's 26% reduction in new macroalbuminuria in a broader population, because the endpoints, populations, and background therapies differ. Ranking agents based on these numbers would be methodologically inappropriate.

Second, none of the trials were conducted in patients on concurrent SGLT2 inhibitors at meaningful rates. SGLT2 inhibitors themselves reduce composite kidney endpoints by approximately 37% in CKD (as demonstrated in CREDENCE and DAPA-CKD). Whether adding a GLP-1 receptor agonist on top of an SGLT2 inhibitor produces additive kidney protection, or whether the marginal benefit narrows significantly, remains an open empirical question.

Third, the mechanism of kidney protection is not resolved. The trials cannot separate glycemic improvement, blood pressure reduction, weight loss, direct anti-inflammatory effects on the glomerulus, and hemodynamic effects at the afferent arteriole. AWARD-7's insulin comparator arm partially controls for glycemia, and the persistence of an eGFR benefit suggests non-glycemic mechanisms are operating. But FLOW's design, against placebo in patients already on optimal background RAAS therapy, is the most rigorous test of whether GLP-1 receptor agonists add benefit beyond glycemic control, and it says yes.

Outstanding Questions for the Next Trial

Several gaps warrant prospective investigation.

Additive kidney protection with SGLT2i co-therapy. The most pressing clinical question is whether combining a GLP-1 receptor agonist with an SGLT2 inhibitor produces greater kidney protection than either alone, and at what magnitude. A factorial or add-on design in patients already on maximally tolerated SGLT2i therapy would answer this directly.

Patients with eGFR below 30 mL/min/1.73 m². FLOW excluded patients with eGFR <50 at screening. AWARD-7 enrolled down to eGFR 15 but was too small and short to reach hard endpoints in this subgroup. Safety and efficacy data in CKD G4 are insufficient.

Duration of benefit after GLP-1 receptor agonist discontinuation. FLOW's early stopping means we have limited insight into what happens to the trajectory after long follow-up. Does the preserved eGFR advantage persist, narrow, or reverse after the drug is stopped?

Non-diabetic CKD. All three trials required type 2 diabetes. GLP-1 receptor agonists are increasingly used in obesity without diabetes, and IgA nephropathy, hypertensive nephrosclerosis, and focal segmental glomerulosclerosis all share inflammatory pathways that GLP-1 receptor agonists might plausibly modify. This is currently unanswered.

Kidney outcomes by baseline albuminuria stratum. FLOW required macroalbuminuria. LEADER enrolled patients across the albuminuria spectrum. A pre-specified interaction analysis by baseline UACR category, conducted with adequate power in each stratum, would clarify whether benefit is concentrated in overt proteinuria or extends to microalbuminuria.

Frequently asked questions

›
›
›

References

  • Perkovic V, et al. Semaglutide and Kidney Outcomes in Patients with Type 2 Diabetes and Chronic Kidney Disease. N Engl J Med. 2024. https://pubmed.ncbi.nlm.nih.gov/38785350/
  • Tuttle KR, et al. Dulaglutide versus insulin glargine in patients with type 2 diabetes and moderate-to-severe chronic kidney disease (AWARD-7): a multicentre, open-label, randomised trial. Lancet Diabetes Endocrinol. 2018. https://pubmed.ncbi.nlm.nih.gov/29937267/
  • Mann JFE, et al. Liraglutide and Renal Outcomes in Type 2 Diabetes. N Engl J Med. 2017. https://pubmed.ncbi.nlm.nih.gov/28902528/
  • Perkovic V, et al. Canagliflozin and Renal Outcomes in Type 2 Diabetes and Nephropathy (CREDENCE). N Engl J Med. 2019. https://pubmed.ncbi.nlm.nih.gov/30990260/
  • Heerspink HJL, et al. Dapagliflozin in Patients with Chronic Kidney Disease (DAPA-CKD). N Engl J Med. 2020. https://pubmed.ncbi.nlm.nih.gov/32970396/
  • Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2022 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int. 2022. https://pubmed.ncbi.nlm.nih.gov/36007658/