GLP-1 Therapies in ESRD: 2026 Meta-Analysis

A 2026 systematic review and meta-analysis pooled 19 studies involving 388 people with end-stage renal disease (ESRD) who received a GLP-1 receptor agonism-based therapy. The authors reported reductions in body weight with semaglutide at three, six and 12 months. They also found substantial uncertainty around several estimates and called the results hypothesis-generating rather than definitive evidence for treatment decisions^[1].

No new randomized trial was conducted. Instead, the review combined published studies in a population defined by an estimated glomerular filtration rate below 15 mL per minute per 1.73 square metres or renal replacement therapy. Most participants, 91.5%, were receiving renal replacement therapy^[2]. This is a narrow population. Its clinical context is more complex than that of general obesity or type 2 diabetes trials.

What the ESRD meta-analysis examined

The review made change in body weight its primary outcome. Secondary outcomes covered body mass index, glycaemia and safety. At baseline, the pooled participants had a mean weight of 97.7 kg, a mean body mass index of 34.4 kg/m² and mean glycated haemoglobin of 7.26%^[2]. Those averages describe the assembled evidence base, not a profile that determines whether an individual resembles every included study.

Treatment exposure was uneven across medicines, with semaglutide accounting for 266 of the 388 patients, liraglutide for 66 and dulaglutide for 47. Only nine received tirzepatide^[2]. The imbalance is stark. The abstract did not provide a pooled tirzepatide-specific efficacy or safety estimate, so a class-level title should not be mistaken for equally developed evidence for each medicine.

The clearest estimates concerned semaglutide

For semaglutide, pooled mean weight change was minus 3.09 kg at three months, minus 3.68 kg at six months and minus 7.00 kg at 12 months. The respective 95% confidence intervals were minus 5.95 to minus 0.24 kg, minus 5.71 to minus 1.65 kg and minus 11.38 to minus 2.61 kg^[1]. These are averages across included reports, not a forecast for a particular patient.

Heterogeneity, expressed as I², was 58.1% for the three-month weight estimate and 79.2% at 12 months, compared with 0% at six months^[2]. I² describes how much variation in a pooled analysis may reflect differences among studies rather than sampling error; high heterogeneity makes a single summary estimate less portable across settings, but it does not identify which study feature caused the variation.

The semaglutide BMI estimates followed a similar time pattern, but the three-month confidence interval crossed zero. Glycated haemoglobin fell by a pooled 0.50 percentage points at six months with a confidence interval from minus 1.20 to 0.20, which also crossed zero. At 12 months, the pooled reduction was 0.75 percentage points with a confidence interval from minus 1.07 to minus 0.43^[2]. The time-specific estimates should remain separate rather than being summarized as one guaranteed trajectory.

Liraglutide and tirzepatide need a narrower reading

For liraglutide, pooled weight change was minus 2.24 kg at three months and minus 4.50 kg at six months. Both confidence intervals crossed zero: minus 7.73 to 3.26 kg at three months and minus 11.26 to 2.27 kg at six months^[1]. The abstract did not report pooled dulaglutide or tirzepatide weight estimates. It therefore does not support a claim that all four medicines produced the semaglutide result.

Nine tirzepatide recipients were present somewhere in the included evidence, but that small count is not a treatment-effect estimate^[2]. That is only a count. It cannot establish effectiveness, frequency of adverse events or comparative performance in ESRD. Readers interested in the drug's broader evidence can use our tirzepatide research overview, while keeping this renal evidence boundary intact.

What the pooled safety figures show

The abstract reported semaglutide safety proportions rather than a randomized comparison with an untreated or active-control group. Total adverse events occurred in a pooled 44% of patients, with a 95% confidence interval from 24% to 66%. Gastrointestinal adverse events occurred in 39%, and discontinuation because of adverse events occurred in 9%^[1]. Wide intervals indicate limited precision.

Nausea was estimated at 36%, but its confidence interval ranged from 3% to 91% and heterogeneity was 73.6%. Vomiting was estimated at 13%, with a confidence interval from 2% to 53%. Hypoglycaemia was estimated at 9%, with an interval from 0% to 82%^[2]. Precision is poor. Those ranges are too broad for individual risk prediction and should not be read as proof that semaglutide caused every recorded event.

ESRD can alter the clinical consequences of fluid loss, nutrition changes and glucose-lowering treatment. The meta-analysis does not turn pooled rates into instructions for starting, stopping or adjusting a prescription. Medication decisions in renal replacement therapy require the treating kidney and metabolic care teams. Our semaglutide safety summary offers general context, not individualized renal guidance.

Evidence limits and the next research step

The public PubMed record provides the abstract and bibliographic identity but no record-level PMCID. The publisher route did not expose inspectable full text during this review. Details absent from the abstract, including the design and quality of each included study, cannot be reconstructed safely. The abstract also does not report a randomized pooled comparator for the headline weight estimates.

The authors themselves described the findings as hypothesis-generating and called for prospective randomized controlled trials^[1]. A stronger study would need enough ESRD participants for medicine-specific estimates and prespecified safety ascertainment. It would also need a clear comparator so that changes over time can be separated from treatment effects and background clinical care.

The defensible takeaway is limited: published ESRD studies, dominated by semaglutide exposure, produced pooled weight-loss estimates and imprecise safety proportions. The synthesis supports further study in a population often underrepresented in trials. It does not establish equal evidence across the GLP-1 class, a tirzepatide benefit, or a treatment plan for an individual receiving dialysis.

Medical disclaimer: This article is for educational and journalistic purposes only and does not constitute medical advice. Peptides discussed may be classified as prescription medicines or research chemicals depending on your jurisdiction. Always consult a qualified healthcare professional before using any peptide product. PeptideMethods.com does not sell, distribute, or facilitate the sale of any peptide product.