Stem Cell Therapy for Diabetic Kidney Disease: A Phase 2 Trial Worth Watching; But Not Yet Acting On

Andrew Kowalski, MD, FASN

A summary and commentary on: Čižman B, Butler EL, Stavas J, et al. A Randomized Clinical Trial of Kidney Autologous Cell Therapy in Diabetic Kidney Disease. Clin J Am Soc Nephrol. Published January 2, 2026. doi:10.2215/CJN.0000000969

So if you have not heard, there has been some excitement in the world of kidney disease lately! A article came out in 1/2026 discussing a possible, and promising, use of stem cell therapy to address diabetic kidney disease.

So why are we excited about stem cells?

Stem cells are special cells, more like “baby cells,” that have not matures yet. They can make copies of themselves and develop into other kinds of cells in the body, ANY kind of cell in the body. We already use them in medicine, they already play an important role in blood and bone marrow transplantation, and they may soon help repair damaged tissues, reduce harmful inflammation, and improve how we model diseases and test drugs. The near future of stem-cell medicine is promising, especially in regenerative medicine and immune regulation, but most new uses still require careful clinical trials to confirm safety and real benefit.

And that is what makes this article so exciting!

A recently published phase 2 trial in CJASN (Clinical Journal of the American Society of Nephrology) is generating genuine interest in nephrology circles, and pushing what our abilities are. The study tests rilparence, an autologous (cells or blood that have come from the same person who will also receive them), kidney-derived cell therapy harvested by biopsy and then injected directly into the kidney cortex in patients with diabetes and chronic kidney disease. It is a genuinely novel approach in a field where most therapies slow progression, but do not plausibly repair anything.

What Was Studied?

The trial enrolled 53 participants across five US sites, of whom 49 received at least one injection and were included in the analysis. This was a moderate-to-advanced CKD population: mean baseline eGFR was 33 mL/min/1.73 m² (CKD Stage IIIb) and median UACR was 421 mg/g (A3).

• 78% had type 2 diabetes

• 22% had type 1

Background kidney-protective therapy was reasonably solid for the study period

• 80% were on ACEi or ARB

• 37% on SGLT2 inhibitors

• 39% on GLP-1 receptor agonists

The cohort was predominantly White (88%) and male (69%).

Participants were randomized into one of two active dosing strategies, not to placebo.

Cohort 1: received scheduled bilateral treatment; one rilparencel injection into each kidney, approximately three months apart, with 96% ultimately receiving both doses.

Cohort 2: received an initial injection, and a second dose only if a sustained eGFR decline or rise in UACR occurred; 60% ended up receiving a second injection

• median interval of 11 months between doses

The primary efficacy endpoint was the within-patient change in eGFR slope from the pre-injection period to the post-injection period:a design choice that carries important implications, discussed below.

What the Trial Found

The headline result comes from Cohort 1. Annual eGFR slope improved from -5.84 to -1.27 mL/min/1.73 m²/year, a difference of +4.57 mL/min/1.73 m²/year (95% CI 1.95 to 7.18). Meaning that the change in function improved. That is a large enough signal to make most nephrologists stop and look twice.

In Cohort 2, slope improved from -3.40 to -1.71 mL/min/1.73 m²/year, a difference of +1.70 (95% CI -0.24 to 3.63), directionally favorable, but not definitively so based on the confidence interval.

Absolute mean (average) eGFR change from baseline after the last injection remained modest: Cohort 1 lost approximately 0.8 to 2.2 mL/min/1.73 m² through 18 months

Cohort 2 approximately 2.1 to 2.9.

Hard kidney endpoint event counts were low and limited further interpretation:

• Sustained ≥40% eGFR decline occurred in 17% of Cohort 1 and 8% of Cohort 2

• Progression to eGFR <15 or kidney replacement therapy occurred in 17% and 16% (C1, C2), respectively

• Albuminuria was not a convincing strength of the therapy, and no consistent effect on urine albumin excretion emerged.

Safety: Not Trivial, But Acceptable

Across 87 injections:

• 31 procedure-related treatment-emergent adverse events occurred in 16 participants (33%)

• 14 product-related events in 6 participants (12%)

• Three participants experienced six biopsy-related serious adverse events

  • Subcapsular kidney hematoma (Blood collection under the top layer of the kidney)

  • AKI

  • Hematuria with hydronephrosis (Blood int he urine with kidney enlargement probably from blockage due to a blood clot)

  • One participant had an injection-related subcapsular hematoma.

There were no product-related serious adverse events and no procedure or product-related deaths. The investigators note that biopsy-related event rates were comparable to those reported in similar CKD populations undergoing native kidney biopsy. Even so, this is not a pharmacologically simple intervention, it requires kidney biopsy plus cortical injection, and the safety discussion is procedural in nature, not just pharmacologic.

Where the Trial Is Strong — and Where It Struggles

The trial’s most compelling feature is what it is testing. A regenerative, autologous cell approach is a fundamentally different study than another hemodynamic or metabolic modifier, and the magnitude of slope improvement in the scheduled bilateral cohort is large enough to sustain attention. The investigators also enrolled a genuinely high-risk population on relatively contemporary background therapy, so the signal did not arise against an undertreated comparator.

The most significant limitation is structural. This was not a parallel-group, placebo-controlled trial. The primary efficacy claim rests on comparing each participant’s historical pre-treatment eGFR slope with their post-treatment slope, using data that included chart review. That design is vulnerable to numerous changes when calculating the results and all the authors acknowledge explicitly.

The sample was also small, clinical events were few, the two cohorts had meaningful baseline differences (notably higher UACR in Cohort 1), and the study was conducted at five US sites with a predominantly White male population, limiting generalizability.

Several authors are affiliated with ProKidney, the company developing rilparencel, and sponsorship bias must be taken seriously when interpreting a favorable phase 2 signal.

The Bottom Line

The authors conclude that bilateral kidney administration of rilparencel was associated with slowing CKD progression and that the safety profile was acceptable.

What this paper does usefully establish is that dose intensity and timing appear to matter, that any meaningful effect appears to operate on eGFR trajectory rather than albuminuria, and that the approach is procedurally feasible, however invasive. If the phase 3 trial confirms benefit, the eventual niche will almost certainly be high-risk DKD patients with progressive GFR loss despite contemporary therapy, not broad first-line use. The biopsy-and-injection logistics alone make rilparencel a very different kind of intervention than adding finerenone or semaglutide. For now, it is worth watching closely.​​​​​​​​​​​​​​​​

Exciting times ahead!