Skip to content
GLP Loss
← Research
Evidence

GLP-1s and Dialysis: What the Labels Claim and What Was Tested

Four labels in one class give four different answers about end-stage kidney disease. The reassuring one rests on eight dialysis patients — and drug exposure peaks in moderate impairment, not at the bottom of the range.

Owen Castellanos9 min read
Exposure does not rise as the kidney failsSingle-dose tirzepatide 5 mg, 45 subjects by renal groupNormal function, n=14referenceMild, n=8no changeModerate, n=8+25 to 29%Severe, n=7no changeDialysis, n=8no changeThe peak sits in the middle of the range, not at the end.Eight dialysis patients are the basis of the labeled sentence.No outcome trial in this class has enrolled anyone on dialysis.

Most people who ask this question are not asking it casually. They are on maintenance dialysis, they have been told a body-mass index ceiling stands between them and a transplant waiting list, and a weight-loss drug is aimed squarely at the obstacle. The evidence behind the answer is far smaller than the confidence in the labels suggests.

The trial most often invoked has nothing to say here. The semaglutide kidney outcome trial set an estimated glomerular filtration rate of 25 as the floor of its lower entry route, so nobody with worse function than that, and nobody receiving dialysis, was randomized into it.[1]What it did find, for the population it did enroll, is in the kidney disease article and the trial explainer.

Four labels, four different sentences

Read on DailyMed on September 15, 2026, the class does not speak with one voice about severe impairment. The Ozempic label states that no dose adjustment is recommended for patients with renal impairment, and that in subjects with renal impairment including kidney failure, no clinically relevant change in semaglutide pharmacokinetics was observed. The Zepbound and Mounjaro labels use the stronger formulation: in subjects with renal impairment including end-stage renal disease, no change in tirzepatide pharmacokinetics was observed — adding an instruction to monitor renal function in patients reporting adverse reactions that could lead to volume depletion.

Two labels in the same class say something different. Saxenda’s renal section consists of a single sentence: there is limited experience with the drug in patients with mild, moderate and severe renal impairment, including end-stage renal disease. The Trulicity label recommends no dose adjustment including in end-stage renal disease and then instructs, in the same section, to use the drug with caution in those patients.

The most interesting label is the one that says nothing. The Wegovy label — the obesity-indication semaglutide product, and the closest branded analog to what most readers of this site are buying — has no renal impairment subsection at all in its specific-populations chapter. Its population pharmacokinetic analysis reports no clinically meaningful difference in exposure by renal impairment, and defines that term as an eGFR of 30 or above and below 90. The obesity label’s coverage stops two full stages short of dialysis.

Where the pharmacokinetic claim comes from

The semaglutide sentence rests on a single-dose study in 56 subjects split across five renal function groups, each given 0.5 mg subcutaneously with plasma sampled to 480 hours. Exposure in mild and moderate impairment and in end-stage renal disease was similar to normal function. In severe impairment, mean exposure was 22% higher, and the 95% confidence interval for that ratio, 1.02 to 1.47, exceeded the prespecified limits of 0.70 to 1.43. Adjusting for sex, age and body weight brought every comparison back inside those limits, and hemodialysis did not appear to affect the pharmacokinetics.[2]

The tirzepatide sentence rests on 45 subjects given a single 5 mg dose: 14 with normal function, 8 mild, 8 moderate, 7 severe, and 8 with end-stage renal disease requiring dialysis. Every 90% confidence interval for the exposure and peak-concentration ratios spanned unity except a 25% to 29% increase in area under the curve in the moderate group, and there was no significant relationship between exposure and eGFR at all.[3]

Read together, those two studies say something the labels do not. Drug exposure in this class is not ordered by kidney function. In one program the only group that moved was severe impairment; in the other it was moderate; in neither was it the dialysis group. A reader who assumes worse kidneys mean more drug has the mental model backwards, and a reader who takes “no change including ESRD” as a large finding is reading a sentence built on eight people.

The randomized evidence, counted

A 2025 Cochrane review searched to September 2024 for randomized trials of GLP-1 receptor agonists in people with diabetes and chronic kidney disease at any stage. It included 42 studies and 48,148 participants, at a median study age of 66 and a median follow-up of 26 weeks. Six studies enrolled people at CKD stages 1 to 2, eleven at stages 3 to 5, and one study enrolled people on dialysis.[4]

That review also carries the result that complicates the enthusiasm. Against placebo, the class probably reduced all-cause death, at a risk ratio of 0.85 (95% CI 0.74 to 0.98, moderate certainty), and probably reduced three-point major adverse cardiovascular events, at 0.84 (95% CI 0.73 to 0.98). But on kidney failure defined as starting dialysis or receiving a transplant, the pooled risk ratio was 0.86 (95% CI 0.66 to 1.13), graded as probably little or no effect, and the composite kidney outcome was 0.89 (95% CI 0.78 to 1.02).[4]

So the class lowers the chance of dying in chronic kidney disease and has not been shown to lower the chance of reaching dialysis. Those two sentences are both from the same review and they are routinely quoted one at a time.

The one trial, and what it found

The single randomized dialysis study assigned 24 patients with type 2 diabetes and dialysis-dependent end-stage renal disease, alongside 23 controls with normal kidney function, to 12 weeks of double-blind liraglutide titrated to 1.8 mg daily or placebo. Twenty patients with ESRD and 20 controls completed, which means the entire randomized dialysis experience in this drug class amounts to roughly ten people who actually received drug.[5]

Its primary endpoint pointed the opposite way from the single-dose studies above. Dose-corrected plasma trough liraglutide concentration was 49% higher in the end-stage group than in controls (95% CI 6 to 109, P = 0.02), and initial nausea and vomiting occurred more often among liraglutide-treated ESRD patients than among controls (P < 0.04). Body weight fell by 2.4 ± 0.8 kg in the ESRD group, which did not reach significance (P = 0.22), against 2.9 ± 1.0 kg in controls, which did (P = 0.03). The authors concluded that reduced doses and a prolonged titration period may be advisable.[5]

Liraglutide is not semaglutide or tirzepatide. It is dosed daily, it is heavily albumin-bound, and its handling in kidney failure need not generalize. That caveat cuts both ways: the class-wide reassurance in those labels also rests on molecules studied one at a time, in single doses, in groups of eight. The slow-titration logic in the titration article is the practical response to both readings.

What the observational record adds, and what it costs

A retrospective cohort followed 76 patients with CKD stage 4 or worse, including dialysis, who started semaglutide; 96% had type 2 diabetes. Mean weight fell from 106.2 to 101.3 kg (P < .001), about 4.6% of body weight, and mean HbA1c from 8.0% to 7.1% (P < .001), with 16% of the diabetic patients discontinuing insulin. Nearly two-thirds reported no adverse effects, and adverse effects were nonetheless the leading reason for stopping, at 37.0% of discontinuations. The median treatment duration was 17.4 months with an interquartile range of 0.43 to 48.8 — a lower quartile under two weeks.[6]

A single-center chart review of 36 dialysis patients prescribed semaglutide specifically to improve transplant candidacy reported a mean total body weight loss of 7.8% (SD 6.1), and 48.2% of those who had been ineligible on body-mass index reached waitlist activation. Treatment-limiting gastrointestinal side effects occurred in 16.7%, and discontinuation was far more common on peritoneal dialysis than on hemodialysis — 30.8% against 8.7%.[7]

The most recent multicenter series is the one that shows the size of the field. It included 17 patients with type 2 diabetes and end-stage renal disease over a median 1,187 days. HbA1c changed by −0.31 ± 2.57 percentage points (P = 0.630) and body weight by −12.63 ± 24.03 kg (P = 0.074) — neither statistically significant, both with standard deviations larger than the effect. Acute kidney injury occurred in 57.1% of the non-dialysis patients in the cohort, and severe hypoglycemia in two patients overall.[8]

That acute kidney injury figure is the one to carry away from an observational series. In someone with residual kidney function, the dehydration route is real, and the labeled monitoring instruction exists for it — the practical side is in the dehydration article. In someone anuric on maintenance hemodialysis the warning means something different, because there is no remaining filtration to injure, while interdialytic volume swings and blood pressure on the machine remain very much live concerns.

Why dialysis is not simply severe kidney disease

Three things change at the point maintenance dialysis starts, and none of them appears in a pharmacokinetic ratio. Fluid status is managed externally, so a week of vomiting shows up as a missed target weight rather than as a creatinine rise. Gastroparesis is common in long-standing diabetic kidney disease, and it interacts with a drug class that slows gastric emptying on purpose — the mechanics are in the gastroparesis article. And insulin requirements fall as intake falls, which in a population with unpredictable clearance is a hypoglycemia problem, not a convenience; the insulin article covers the adjustment.

The nutritional direction is the one that makes clinicians most cautious. Low body weight predicts worse survival on dialysis, so deliberately inducing weight loss in that population is a trade rather than a plain good, and it is defensible mainly when the goal is a transplant the patient is otherwise locked out of. That is exactly the setting the 36-patient series studied, and it is the setting where nearly half of those patients got onto a list.[7]

What a cash-pay checkout cannot resolve

Every figure above describes branded product at labeled doses under nephrology supervision. Most cash-pay sellers dispense compounded semaglutide or tirzepatide, which is not FDA-approved and which the FDA does not review for safety, efficacy or quality before it is dispensed — see what a compounded vial contains and the sellers on the compounded semaglutide board. A compounded vial also carries no renal-impairment section of its own, because it carries no FDA-reviewed label at all.

The honest position is narrow. Neither of the two dominant molecules requires a dose adjustment in kidney failure on the evidence available, and that evidence is single-dose pharmacokinetics in eight to a dozen people per group. Weight comes off in dialysis patients in observational series, at rates between 4.6% and 7.8%, and roughly one in six stops because of gastrointestinal effects. Nobody has run an outcome trial. Starting one of these drugs on dialysis is a decision for the nephrology team that manages the prescription, the dry weight and the transplant listing — and the criteria used to assess how carefully a seller screens for any of that are in the methodology.

Frequently asked

Do GLP-1 drugs need a dose adjustment on dialysis?
The labels say no. The Ozempic label states that no clinically relevant pharmacokinetic change was observed in renal impairment including kidney failure, and the tirzepatide labels say the same for end-stage renal disease. Those sentences rest on single-dose studies with eight dialysis patients in the tirzepatide program and a comparable handful in the semaglutide one.
Did the FLOW kidney trial include dialysis patients?
No. Its lower entry route required an estimated glomerular filtration rate of 25 or above, so nobody with worse function than that was randomized, and nobody on dialysis was enrolled at all. Its results describe slowing albuminuric kidney disease in people with type 2 diabetes, not treating established kidney failure.
How much randomized evidence exists in dialysis?
One trial. A 2025 Cochrane review of GLP-1 receptor agonists in chronic kidney disease and diabetes included 42 studies and 48,148 participants, of which exactly one enrolled people on dialysis. That study randomized 24 patients with end-stage renal disease to liraglutide or placebo for 12 weeks, and about ten of them received drug.
Does drug exposure go up as kidney function falls?
Not in a straight line. In the tirzepatide renal study the only group whose confidence interval excluded no change was moderate impairment, at a 25% to 29% higher exposure, while severe impairment and dialysis matched normal function. In the semaglutide study the outlier was severe impairment at 22% higher, and end-stage renal disease again matched normal.
How much weight do dialysis patients lose on semaglutide?
In observational series, between roughly 4.6% and 7.8% of body weight. A cohort of 76 patients with stage 4 or worse kidney disease lost about 4.6%, and a chart review of 36 dialysis patients seeking a transplant lost a mean 7.8%, with 48.2% of those previously ineligible on body-mass index reaching waitlist activation. Treatment-limiting gastrointestinal effects occurred in about one in six.
Why would someone on dialysis be prescribed one at all?
Usually to get onto a transplant list. Many programs set a body-mass index ceiling for listing, and weight loss is the only route past it. That framing matters because low body weight otherwise predicts worse survival on dialysis, so deliberate weight loss is a trade rather than a plain benefit, and the decision belongs to the nephrology and transplant team.

Sources

  1. [1] Perkovic V, Tuttle KR, Rossing P, et al. (2024). Effects of Semaglutide on Chronic Kidney Disease in Patients with Type 2 Diabetes. N Engl J Med. PMID 38785209
  2. [2] Marbury TC, Flint A, Jacobsen JB, et al. (2017). Pharmacokinetics and Tolerability of a Single Dose of Semaglutide, a Human Glucagon-Like Peptide-1 Analog, in Subjects With and Without Renal Impairment. Clin Pharmacokinet. PMID 28349386
  3. [3] Urva S, Quinlan T, Landry J, et al. (2021). Effects of Renal Impairment on the Pharmacokinetics of the Dual GIP and GLP-1 Receptor Agonist Tirzepatide. Clin Pharmacokinet. PMID 33778934
  4. [4] Natale P, Green SC, Tunnicliffe DJ, et al. (2025). Glucagon-like peptide 1 (GLP-1) receptor agonists for people with chronic kidney disease and diabetes. Cochrane Database Syst Rev. PMID 39963952
  5. [5] Idorn T, Knop FK, Jørgensen MB, et al. (2016). Safety and Efficacy of Liraglutide in Patients With Type 2 Diabetes and End-Stage Renal Disease: An Investigator-Initiated, Placebo-Controlled, Double-Blind, Parallel-Group, Randomized Trial. Diabetes Care. PMID 26283739
  6. [6] Long JJ, Sahi SS, Lemke AI, et al. (2024). The Use of Semaglutide in Patients With Renal Failure-A Retrospective Cohort Study. Endocr Pract. PMID 39025300
  7. [7] Wade FG, Lentine KL, Turk D, et al. (2025). Bridging the Gap to Waitlist Activation: Semaglutide's Weight Loss Efficacy and Safety in Patients With Obesity on Dialysis Seeking Kidney Transplantation. Clin Transplant. PMID 41075262
  8. [8] Alkhezi O, Alfehaid L, Alanezi A, et al. (2026). Real-world use of semaglutide in patients with type 2 diabetes and end-stage renal disease: a multicenter retrospective cohort study. Front Endocrinol (Lausanne). PMID 41970978

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence