Does Ozempic Protect Your Kidneys? Only on the Diabetes Label

Ozempic carries an FDA-approved indication to protect kidney function, and it applies only to adults who have both type 2 diabetes and chronic kidney disease.¹ The same molecule sold for weight loss carries no kidney indication at all.²

Both labels also warn about acute kidney injury, and that looks like a contradiction without being one, because the two things happen on different timescales for different reasons.

What did the FLOW trial measure?

The FLOW trial asked whether semaglutide slows kidney decline in people who already have kidney disease. The label describes it as a "randomized, double-blind, placebo-controlled, event driven trial" under the registry identifier NCT03819153, and reports that "A total of 3,533 patients were randomized to receive OZEMPIC 1 mg once weekly or placebo."¹

Its primary endpoint was a composite: time to first occurrence of "a sustained decline in eGFR of ≥50%, sustained eGFR <15 mL/min/1.73m2" plus "chronic renal replacement therapy, renal death, CV death."¹ Sustained carries a specific definition on the label: "2 consecutive measurements ≥28 days apart."¹

Median follow-up ran 41 months, so this is a long trial by the standards of anything else on this drug class.¹

Event driven deserves a sentence, because it changes what the duration means. An event-driven trial runs until a target number of endpoint events accumulates instead of until a fixed date, so the 41 months is an outcome of how fast events arrived instead of a protocol decision made in advance. In a population selected for existing kidney disease, events arrive.

How large was the effect?

The composite fell from 23.2% on placebo to 18.7% on semaglutide, a hazard ratio of "HR 0.76 [95% CI 0.66, 0.88], p=0.0003."¹ In counts, 410 of 1,766 placebo patients had an event against 331 of 1,767 on semaglutide.¹

Subtracting the label's own percentages gives a difference of 4.5 percentage points over a median 41 months. The label gives no number needed to treat and that subtraction is mine, not the label's, so treat it as arithmetic on published figures instead of a reported result.

This is a large effect by the standards of this site. The gallbladder signal on the same drug class runs to 27 extra cases per 10,000 treated years, and this runs to roughly 45 fewer events per 1,000 patients over three and a half years, in the opposite direction. Size and direction both differ, and they are measured in populations selected for entirely different reasons.

A hazard ratio of 0.76 also means something narrower than a 24% risk reduction, the form it usually gets reported in. It describes the rate at which events accumulated over the whole follow-up, in a population where roughly one in five placebo patients hit an endpoint. Applied to somebody whose baseline risk is much lower, the same ratio moves far fewer people.

Which parts of the composite actually moved?

Two of the five components reached significance on their own, and they are not the two most people would assume.

All six rows come from the label's own Table 10.¹ The two components whose intervals exclude 1.0 are the 50% eGFR decline and cardiovascular death.

Component analyses inside a composite carry a standard caveat. Each one has fewer events than the composite, so each has less power. An interval crossing 1.0 on a component is therefore a weaker statement than the same interval on a primary endpoint would be, and none of these components was the trial's registered primary. The label reports them without individual p values.¹

Did dialysis and kidney failure rates fall?

On their own figures, dialysis and kidney failure moved in the same direction as the composite but carried confidence intervals that cross 1.0. Chronic renal replacement therapy came in at a hazard ratio of 0.84 with an interval of 0.63 to 1.12, and sustained eGFR below 15 at 0.80 with an interval of 0.61 to 1.06, and both intervals cross 1.0.¹

Renal death was identical in both arms, at five events each, with a hazard ratio of 0.97 and an interval running from 0.27 to 3.49 that says almost nothing on ten events across 3,533 patients.¹

Nothing here means those endpoints did not improve. The point estimates all sit below 1.0 and point the same way as the composite. What the figures do not support is a claim that semaglutide was shown to prevent dialysis, because the component measuring dialysis did not separate from placebo on its own.

The distinction matters commercially as well as clinically, because preventing dialysis is the claim with the most weight behind it in a conversation about kidney disease, and it is the one component here whose interval includes no effect. Anybody citing FLOW for dialysis prevention is citing the composite and attributing it to a part that did not reach significance.

Why does cardiovascular death sit inside a kidney endpoint?

Cardiovascular death sits inside the composite because people with advanced kidney disease tend to die of cardiovascular causes before their kidneys fail, and trials in this population count it for that reason. Leaving it out would let a drug look good on kidney endpoints by letting patients die of something else first.

The consequence for reading this trial is specific. Cardiovascular death contributed 169 placebo events against 123 on semaglutide, the second largest absolute gap of any component, at a hazard ratio of 0.71.¹ A meaningful share of the composite's movement is therefore a cardiovascular result sitting inside a kidney composite.

That is a legitimate trial design and it does change what the headline means. The label's indication language reflects this exactly, covering "sustained eGFR decline, end-stage kidney disease, and cardiovascular death" instead of kidney outcomes alone.¹

Credit where it belongs: the indication is worded more carefully than most coverage of it. Anyone reading Section 1 learns that cardiovascular death is part of what the drug is approved to reduce, and that is exactly what the trial measured. The flattening happens downstream of the label, not on it.

What happened to death rates overall?

All-cause death fell, on the hardest endpoint in the trial and the one least often mentioned. The label reports 279 deaths among 1,766 placebo patients, 15.8%, against 227 among 1,767 on semaglutide, 12.8%, at a hazard ratio of 0.80 with an interval of 0.67 to 0.95 and a p value of 0.0104.¹

Major adverse cardiovascular events moved too, and that composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke came in at 254 events, 14.4%, against 212, 12.0%, with a hazard ratio of 0.82, an interval of 0.68 to 0.98 and a p value of 0.0289.¹

A trial that set out to measure kidney decline and came back with fewer deaths from any cause is a stronger result than its own primary endpoint. That result also complicates the simple reading, because if people on semaglutide were less likely to die of anything, some of the kidney composite's movement reflects patients surviving to be measured and not kidneys holding up.

One figure the label withholds belongs here too, since it states that Ozempic reduced the annual rate of change in eGFR without publishing a number for it.¹ That rate is the measure a nephrologist would want, because it describes the slope of decline instead of the crossing of a threshold, and the label reports the direction without the magnitude.

Who was in the trial?

Everybody in FLOW had established kidney disease and diabetes, and the entry criteria were narrow. Entry required an eGFR of 25 to 75 mL/min/1.73m2 with a urine albumin-to-creatinine ratio above 100 mg/g and below 5000 mg/g, an HbA1c of 10% or less at screening, and background RAAS blockade unless it was contraindicated or not tolerated.¹

Set against the general population taking these drugs, those criteria are restrictive: an eGFR ceiling of 75 excludes normal renal function. An albumin-to-creatinine ratio above 100 mg/g means measurable protein leaking into urine. Background RAAS blockade means these patients were already on standard kidney protection and semaglutide was added on top.

The trial therefore answers a narrow question well, saying what happens when semaglutide is added to existing care in people whose renal function is already failing and who have diabetes. Our guide to the GLP-1 side effect data covers the broader population.

The RAAS blockade requirement deserves particular attention, because those drugs are the existing standard for slowing diabetic kidney disease, so FLOW tested semaglutide as an addition to treatment and not as an alternative to it. A result that holds on top of standard care is a stronger finding than one obtained instead of it, and it is also not a license to substitute.

Does the kidney indication apply to Wegovy?

The kidney indication does not apply to Wegovy, whose indications cover cardiovascular risk reduction, weight reduction and noncirrhotic MASH with moderate to advanced fibrosis, none of which mentions kidney disease, eGFR or renal risk reduction.²

No kidney outcome trial appears in Wegovy's clinical studies either. The trial reported there is the cardiovascular outcomes study, and eGFR appears only twice on that label: once in pharmacokinetics, noting no clinically significant differences by renal impairment down to an eGFR of 30, and once in a baseline table recording that 10% of participants had moderate renal impairment.²

Same molecule, two labels, and the kidney protection sits on one of them. A person taking semaglutide for weight loss is taking a drug whose kidney benefit has been demonstrated in a population they are probably not in.

Wegovy's label does carry one relevant observation in its adverse reactions section, noting that renal adverse reactions occurred more frequently in adults with a history of renal impairment.² That points the opposite way from protection, describing who is more likely to have a kidney problem on the drug instead of who benefits.

Does it apply to Zepbound or tirzepatide?

Tirzepatide has no kidney indication and no renal outcome trial behind it on the Zepbound label. Zepbound's indications cover weight reduction and moderate to severe obstructive sleep apnea in adults with obesity, with no kidney indication, and its clinical studies section reports four weight-reduction studies and two sleep apnea studies.³

Its renal content runs the other way, with section 8.6 stating that "No dosage adjustment of ZEPBOUND is recommended for patients with renal impairment," and the pharmacokinetics section reports that "Renal impairment does not impact the pharmacokinetics of tirzepatide."³ Those are statements about how the drug behaves in damaged kidneys, not about protecting them.

The labeled kidney benefit therefore belongs to one drug, at one indication, from one trial. Our comparison of semaglutide and tirzepatide covers what else divides them.

Why do the same labels warn about kidney injury?

Volume depletion is a separate mechanism from chronic kidney decline, and all three labels say so in their own headings. Wegovy carries subsection 5.5, "Acute Kidney Injury Due to Volume Depletion," and Zepbound carries subsection 5.3 under the same heading.²³

Ozempic's subsection 5.6 reports the same thing: "There have been postmarketing reports of acute kidney injury, in some cases requiring hemodialysis," and adds that "The majority of the reported events occurred in patients who experienced gastrointestinal reactions."¹

The mechanism sits in that second sentence: vomiting and diarrhea cause dehydration, dehydration drops perfusion to the kidneys, and an organ deprived of blood flow fails acutely. None of that is a direct effect on renal tissue. Our guide to dehydration on retatrutide covers the same pathway on a different molecule.

How often does acute kidney injury happen?

Only one of the three labels puts a number on it. Acute kidney injury appears as an adverse reaction on the Zepbound label at 0.5% against 0.2% on placebo.³ Neither semaglutide label attaches a frequency to its postmarketing reports.

A postmarketing report and a trial adverse reaction rate are different kinds of evidence. Postmarketing reports have no denominator, so they establish that something happened without establishing how often. A trial rate has a denominator and a comparator, so the 0.5% against 0.2% figure is the more informative of the two.

Acute renal failure and worsening chronic renal failure appear again in Zepbound's postmarketing section, sometimes requiring hemodialysis, so the signal sits in both places on that label.³

One detail in that listing is easy to pass over. Worsening chronic renal failure appears alongside acute failure, so the postmarketing picture includes people whose existing kidney disease deteriorated and not only people whose kidneys failed suddenly. The label records both without separating them.³

Is protection and injury a contradiction?

Protection and injury are not a contradiction, and collapsing them is the most common error on this topic. Chronic renal decline unfolds over years and is what FLOW measured across 41 months. Acute kidney injury from dehydration unfolds over days and is what the postmarketing reports describe.

A drug can slow the first while occasionally causing the second. The two effects sit on different timescales, run through different mechanisms, and appear in different sections of the same document.

What the labels ask for in response is monitoring, not avoidance. All three direct clinicians to monitor renal function in patients reporting reactions that could lead to volume depletion, with Ozempic and Zepbound both flagging initiation and dose escalation as the period of concern.¹³

What does this mean for somebody taking it for weight loss?

The kidney protection has not been demonstrated in that population and the kidney injury warning applies to everybody. Those two facts pull in opposite directions and both come from the labels.

FLOW enrolled people with an eGFR of 25 to 75 and measurable albuminuria who already had diabetes and were already on RAAS blockade.¹ Somebody with normal kidney function taking semaglutide for weight reduction shares none of those entry criteria, and the drug they are prescribed carries no renal indication.²

The acute injury warning has no such restriction, appearing on all three labels and tied to gastrointestinal reactions instead of to any baseline renal status, and gastrointestinal reactions are what these drugs reliably produce. Our guide to the GLP-1 pills covers how route affects those reactions.

What has not been measured?

Four gaps sit in the evidence, and each one is specific enough to name.

No renal outcome trial exists for semaglutide in people without diabetes. FLOW required type 2 diabetes at entry, so whether the same protection appears in obesity without diabetes is untested, not unlikely.¹

No renal outcome trial exists for tirzepatide at all, and the Zepbound label reports none, so the question of whether this is a class effect or a semaglutide effect has no trial answer.³

Dialysis prevention is unproven on the component figures, with the chronic renal replacement therapy interval running 0.63 to 1.12, so the trial did not demonstrate that endpoint individually even though the composite moved.¹

The frequency of acute kidney injury on semaglutide is also unquantified. Both semaglutide labels describe postmarketing reports without rates, and only the tirzepatide label gives a trial figure.¹²³

Common questions

Does Ozempic protect the kidneys?

What did the FLOW trial find?

Did FLOW show semaglutide prevents dialysis?

Which FLOW components reached significance?

Why was cardiovascular death in a kidney trial?

Who qualified for the FLOW trial?

Does Wegovy have a kidney indication?

Does Zepbound protect the kidneys?

Can Ozempic cause kidney damage?

How common is acute kidney injury on these drugs?

Is kidney protection and kidney injury a contradiction?

Should kidney function be monitored on these drugs?

Does the kidney benefit apply without diabetes?

What dose did FLOW use?

Is there a kidney outcome trial for tirzepatide?

How big was the FLOW benefit in absolute terms?

Sources

1. Novo Nordisk. OZEMPIC (semaglutide) injection, for subcutaneous use. US prescribing information, revised 05/2026. Read at source from the FDA label document. Quoted for the Section 1 kidney indication, the FLOW trial design and registry identifier in Section 14.3, the randomization and median follow-up, the full entry criteria, the primary composite definition and the definition of sustained, every row of Table 10 including the hazard ratios and confidence intervals for all five components, the MACE and all-cause death secondary results, and the Section 5.6 postmarketing acute kidney injury text. The label states that Ozempic reduced the annual rate of change in eGFR without giving a figure, and that absence is cited above as an absence. The FLOW primary publication was not read; every FLOW figure here comes from the label. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18 ↩19 ↩20 ↩21 ↩22 ↩23 ↩24 ↩25 ↩26 ↩27
2. Novo Nordisk. WEGOVY (semaglutide) injection, for subcutaneous use, and WEGOVY tablets. US prescribing information, revised 05/2026. Read at source from the FDA label document. Cited for the complete absence of any kidney indication, the absence of a kidney outcome trial in its clinical studies, subsection 5.5 "Acute Kidney Injury Due to Volume Depletion," and both eGFR mentions, appearing in pharmacokinetics and in a baseline population table. The clinical studies text visible on the read covered Section 14.1 only, with later subsections present as headings, so the absence of a kidney trial is established from the indications and the visible text, not from a complete read of Section 14. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8
3. Eli Lilly. ZEPBOUND (tirzepatide) injection, for subcutaneous use. US prescribing information, revised 02/2026. Read at source from the FDA label document. Cited for the absence of a kidney indication, the contents of its clinical studies section, subsection 5.3 "Acute Kidney Injury Due to Volume Depletion," the Section 6.1 adverse reaction rate of 0.5% against 0.2% on placebo, the Section 6.2 postmarketing listing of acute renal failure and worsening chronic renal failure, and the Section 8.6 and 12.3 statements on renal impairment and pharmacokinetics. eGFR is not mentioned anywhere on this label. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12