This site reported not long ago that the entire world literature on these drugs in dialysis patients was 388 people across 19 studies. That is now out of date. A single matched cohort at a large dialysis organization has followed 2,492 patients who began in-center hemodialysis while on a GLP-1 drug, against an equal number who did not [1].
The results point the same way as the rest of this literature. Hospitalization ran 9% lower (incidence rate ratio 0.91, 95% CI 0.87–0.96) and mortality 17% lower (IRR 0.83, 95% CI 0.73–0.95). Both are incidence rate ratios rather than hazard ratios, because hospitalization recurs and the model is counting episodes rather than time to a first one. Hospitalization as an outcome, and what it captures, is discussed further in the Crohn's disease cohort.
The reason this question was open at all is worth stating precisely, because it is easy to assume it was already answered. The FLOW trial showed semaglutide reducing kidney and cardiovascular outcomes in chronic kidney disease with type 2 diabetes — but it studied people whose kidneys were failing, not people whose kidneys had failed. End-stage disease on dialysis is a different physiological state, a different drug-clearance situation and a different population, which is why an entire literature of 388 patients existed alongside a major positive trial.
What this cohort supplies is scale rather than certainty. Thirteen times the previous evidence base, pointing in a plausible direction, from data an organization already held — and with a baseline imbalance large enough that the honest summary is encouraging rather than established. A randomized trial in dialysis patients remains the thing that would settle it, and the same structure of argument applies to every observational result on this site, including the sickle cell cohort and the kidney disease evidence.
For anyone in this situation the practical point is narrow: the question is now one a nephrologist can discuss with actual numbers rather than with an admission that nothing has been studied. That is a meaningful change even though nothing has been proved.