Was the FLOW trial stopped early, and does that matter?

Yes. FLOW was stopped at a prespecified interim analysis after a median 3.4 years of follow-up, roughly one year before the projected end. Trials stopped early for benefit tend to overestimate treatment effects, meaning the reported 24% reduction may be an upper-bound estimate.

Did most FLOW participants take an SGLT2 inhibitor?

No. Only about 15% of participants were on an SGLT2 inhibitor at baseline. Since SGLT2 inhibitors are now standard of care for diabetic CKD, the trial's background therapy does not reflect current clinical practice.

Why is cardiovascular death included in a kidney endpoint?

Including CV death increases event rates and statistical power. It also captures the reality that CKD patients often die from cardiovascular causes before reaching dialysis. However, it makes it harder to isolate the purely renal benefit of semaglutide.

Does FLOW prove semaglutide protects kidneys structurally?

Not definitively. The early eGFR dip seen with semaglutide suggests a hemodynamic component. Whether the drug provides true structural nephroprotection beyond hemodynamic effects is unresolved.

Can FLOW results apply to people without diabetes?

No. FLOW enrolled only patients with type 2 diabetes. There is no evidence from this trial supporting semaglutide for kidney protection in non-diabetic CKD.

Who funded the FLOW trial?

Novo Nordisk funded the trial, supplied the drug, and participated in design, data collection, analysis, and manuscript preparation. An independent steering committee and separate statistical group provided oversight.

What about patients with very low eGFR?

Patients with eGFR below 25 mL/min/1.73 m² were excluded. The benefit of semaglutide in advanced CKD approaching dialysis is unknown.

How does FLOW compare with SGLT2 inhibitor kidney trials?

FLOW tested semaglutide vs. placebo, not vs. an SGLT2 inhibitor. DAPA-CKD and CREDENCE tested SGLT2 inhibitors in similar populations. No head-to-head trial between these drug classes for kidney outcomes exists.

Is semaglutide FDA-approved for CKD?

As of mid-2024, semaglutide does not carry an FDA-approved indication specifically for CKD or kidney protection. Its approved indications cover type 2 diabetes and obesity.

What is the biggest unanswered question after FLOW?

Whether semaglutide provides additional kidney benefit when added to an SGLT2 inhibitor. This combination reflects modern standard of care, but FLOW was not powered to answer it.

Honest Criticisms and Limitations of the FLOW Trial

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Parameter | Detail | |---|---| | N | 3,533 | | Intervention | Subcutaneous semaglutide 1.0 mg weekly | | Comparator | Placebo (both on top of standard care including max-tolerated RAS blockade) | | Duration | Median 3.4 years (stopped early at a prespecified interim analysis) | | Primary endpoint | Composite of onset of kidney failure (dialysis, transplant, or sustained eGFR <15), sustained ≥50% eGFR decline from baseline, or renal or cardiovascular death | | Key result | HR 0.76 (95% CI 0.66, 0.88), 24% relative risk reduction (Perkovic et al., NEJM 2024) |

Why a Limitations Page Matters

Positive trials attract headlines. Criticism gets buried in supplementary appendices and post-publication letters. For clinicians deciding whether to prescribe semaglutide specifically for kidney protection in their patients with type 2 diabetes (T2D) and chronic kidney disease (CKD), the caveats matter as much as the topline hazard ratio. This page catalogs the methodological, statistical, and practical limitations of FLOW, drawing on the primary publication, its supplementary appendix, and subsequent editorial commentary.

1. Early Stopping and Effect Size Inflation

The Data Safety Monitoring Board recommended stopping FLOW at its second prespecified interim analysis after a median follow-up of 3.4 years, roughly one year before the projected close. Trials stopped early for benefit are well documented to overestimate treatment effects. A 2010 meta-epidemiological study (Bassler et al., JAMA 2010) found that truncated RCTs showed relative risk reductions approximately 29% larger than comparable non-truncated trials. The reported 24% reduction in the kidney composite may therefore represent the upper bound of the true effect.

Early stopping also meant fewer total primary endpoint events (331 semaglutide vs. 410 placebo). With a smaller event pool, subgroup analyses lose power quickly, which limits the granularity of conclusions about which patients benefit most. The confidence interval (0.66, 0.88) is wide enough that the true effect could be as modest as 12%.

2. Enrollment Biases and Population Narrowing

FLOW enrolled adults with T2D, an eGFR of 50 to 75 mL/min/1.73 m² with a urine albumin-to-creatinine ratio (UACR) of 300, 5 to 000 mg/g, or an eGFR of 25, 50 with a UACR of 100, 5,000. This captures a specific CKD phenotype: moderate-to-severe kidney disease with significant albuminuria.

Excluded populations limit how broadly the results apply:

| Excluded group | Clinical relevance | |---|---| | eGFR <25 mL/min/1.73 m² | The patients closest to dialysis, where treatment decisions are most urgent | | Non-albuminuric CKD | Up to 30 to 40% of T2D patients with CKD stage 3 have normoalbuminuria | | Type 1 diabetes | Different pathophysiology; results cannot be extrapolated | | Recent CV events (<60 days) | Acute-phase patients excluded, reducing generalizability to post-MI or post-stroke CKD | | Prior GLP-1 RA or SGLT2i use within 90 days | These are now first-line agents for diabetic kidney disease per KDIGO 2024 guidelines |

The 90-day washout for SGLT2 inhibitors is particularly notable. At baseline, only about 15% of FLOW participants were on an SGLT2 inhibitor. Since SGLT2 inhibitors are now considered standard of care for diabetic CKD based on trials like CREDENCE and DAPA-CKD (Heerspink et al., NEJM 2020), the FLOW population does not reflect contemporary best practice. Whether semaglutide adds kidney protection on top of established SGLT2 inhibitor therapy remains uncertain.

3. The Composite Endpoint Problem

FLOW's primary endpoint bundles five components: dialysis initiation, kidney transplant, sustained eGFR <15, sustained ≥50% eGFR decline, and renal or cardiovascular death. Composites increase statistical power but can obscure which components actually drive the result.

Breaking down the components from the primary publication:

| Component | Semaglutide (n) | Placebo (n) | Contribution to composite | |---|---|---|---| | Sustained ≥50% eGFR decline | Lower | Higher | Major driver | | Kidney failure (eGFR <15, dialysis, transplant) | Modestly lower | Modestly higher | Smaller absolute numbers | | Renal death | Very rare | Very rare | Minimal events | | Cardiovascular death | Lower | Higher | Significant contributor |

The inclusion of cardiovascular death in a "kidney" composite is a design choice, not a given. Cardiovascular death was a meaningful contributor to the overall signal. If the endpoint had been restricted to hard renal outcomes (dialysis, transplant, sustained eGFR <15), the effect size might differ. The sustained eGFR decline component, while clinically validated as a surrogate, is still a laboratory measurement rather than a clinical event the patient experiences.

4. Background Therapy and the SGLT2 Inhibitor Gap

By 2024, SGLT2 inhibitors had become the established renoprotective drug class for T2D with CKD, endorsed by the FDA label for dapagliflozin and international guidelines. FLOW was designed before SGLT2 inhibitor use became widespread in this population. The low baseline SGLT2 inhibitor use (~15%) means the trial essentially tested semaglutide vs. placebo in a treatment era that no longer exists.

A prespecified subgroup analysis stratified by baseline SGLT2 inhibitor use showed no significant interaction (p-interaction not significant), but the SGLT2 inhibitor subgroup was small and underpowered. The critical clinical question, whether adding semaglutide to an SGLT2 inhibitor provides incremental kidney benefit, cannot be answered by FLOW with confidence.

Novo Nordisk funded FLOW, provided the study drug, participated in trial design, data collection, data analysis, and manuscript preparation. The statistical analysis was performed by the sponsor. An independent academic steering committee oversaw the trial, and an independent statistical group verified the primary analysis.

This level of sponsor involvement is standard in large cardiovascular and renal outcome trials, but it warrants acknowledgment. Decisions about endpoint definitions, interim analysis timing, subgroup specifications, and the statistical analysis plan all shape results. The authors disclosed extensive financial relationships with Novo Nordisk and other pharmaceutical companies. None of this invalidates the findings, but it contextualizes the incentive structure behind the trial design and publication.

6. Follow-Up Duration and Durability

A median 3.4 years of follow-up, truncated by early stopping, leaves open questions about long-term outcomes. CKD progression to dialysis often unfolds over 5 to 15 years. Whether semaglutide's renoprotective signal persists, attenuates, or strengthens beyond three years is unknown. There is no planned open-label extension with renal endpoints.

The eGFR slope data showed a characteristic initial dip in the semaglutide group (a hemodynamic effect similar to what is seen with RAS blockade and SGLT2 inhibitors) followed by a slower chronic decline. Whether this hemodynamic component, rather than true structural nephroprotection, accounts for a portion of the benefit remains a topic of debate in nephrology commentary.

7. Generalizability to Non-Diabetic CKD

FLOW enrolled only patients with T2D. Semaglutide's kidney effects in non-diabetic CKD, including IgA nephropathy, FSGS, or hypertensive nephrosclerosis, are entirely unknown from this trial. The metabolic and weight-loss effects of GLP-1 receptor agonists may confer renal benefit through glucose control, weight reduction, and blood pressure lowering, mechanisms less relevant in non-diabetic kidney disease. Extrapolation beyond the enrolled population would be speculative.

8. Adverse Events and Tolerability Bias

Gastrointestinal side effects (nausea, vomiting, diarrhea) were significantly more common with semaglutide, consistent with the semaglutide prescribing information. Discontinuation rates due to adverse events were higher in the semaglutide group. Patients who tolerate GLP-1 receptor agonists may differ systematically from those who discontinue them. Survivors in the semaglutide arm, the ones who stayed on treatment, could represent a healthier or more adherent subpopulation, subtly biasing the per-protocol results.

The intention-to-treat analysis mitigates this concern, but real-world effectiveness may fall short of trial efficacy if a larger proportion of patients discontinue outside the structured trial environment.

9. Post-Publication Commentary

Editorials and letters following the FLOW publication raised several recurring themes:

The "how much is kidney vs. CV" question. Cardiovascular death's inclusion in the primary composite means the trial cannot fully disentangle cardiac from renal protection.
Hemodynamic vs. structural benefit. The early eGFR dip mirrors RAS and SGLT2 inhibitor effects, raising the question of whether semaglutide provides true nephroprotection or primarily a hemodynamic effect that slows measured GFR decline.
Cost and access. Semaglutide costs roughly $900, $1,000/month in the US without insurance coverage for a CKD indication. Whether payers will cover it for renal protection when cheaper SGLT2 inhibitors exist is a practical barrier.
Missing head-to-head data. No trial has compared semaglutide directly against an SGLT2 inhibitor for kidney outcomes. FLOW compared against placebo, not against the current standard of care.

What FLOW Does and Does Not Prove

FLOW demonstrates that semaglutide 1.0 mg weekly reduces a composite kidney-cardiovascular endpoint by 24% compared with placebo in patients with T2D and CKD with albuminuria, on a background of RAS blockade but limited SGLT2 inhibitor use.

FLOW does not demonstrate that semaglutide adds kidney benefit on top of SGLT2 inhibitors. It does not prove structural nephroprotection independent of hemodynamic effects. It does not apply to non-diabetic CKD, advanced CKD with eGFR <25, or non-albuminuric diabetic kidney disease.

Frequently asked questions

›

References

Perkovic V, Tuttle KR, Rossing P, et al. Effects of semaglutide on chronic kidney disease in patients with type 2 diabetes. N Engl J Med. 2024;391(2):109-121. PubMed
Heerspink HJL, Stefánsson BV, Correa-Rotter R, et al. Dapagliflozin in patients with chronic kidney disease. N Engl J Med. 2020;383(15):1436-1446. PubMed
Bassler D, Briel M, Montori VM, et al. Stopping randomized trials early for benefit and estimation of treatment effects: systematic review and meta-regression analysis. JAMA. 2010;303(12):1180-1187. PubMed
KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of CKD. Kidney Int. 2024;105(4S):S117-S314. PubMed
Dapagliflozin (Farxiga) prescribing information. U.S. Food and Drug Administration. FDA Label
Semaglutide (Ozempic) prescribing information. U.S. Food and Drug Administration. FDA Label