Diesen Beitrag gibt es auch auf Deutsch.
What has changed in dialysis medicine over the last seven years? What was the biggest innovation?
Answer: there are flat-screen TVs on the walls now.
The joke holds a true core. Haemodialysis as we practise it today is more than 50 years old in its basic design. Four hours on a machine three times a week, a fistula in the arm, large volumes of water and plastic. At the same time, we have been hearing about wearable dialysis, implantable kidneys and pig kidneys for years. Which of these are close, and which are far away?
A new Roadmap in Nature Reviews Nephrology takes stock. 29 authors from nephrology, engineering, stem cell research, transplant surgery and patient advocacy reviewed progress since the first Kidney Health Initiative Roadmap of 2018, drawing among other things on NIDDK workshops held in 2024.
What the paper is and is not
The paper is an expert overview with recommendations, not a systematic review and not a trial. It contains no new data. Its central message appears in the key points: despite remarkable experimental progress, few advances have reached the clinic, and none is yet in routine clinical practice.
Filtration and tubule: the lab is ahead of the clinic
Any complete kidney replacement needs an artificial glomerulus. Current dialysis membranes separate by size, with a cut-off around 30 to 50 kDa. The native glomerulus also separates by charge, stays clog-free for a lifetime and remains highly permeable to water. Ultrathin silicon membranes come close but are rigid, prone to clogging under flow and usually lack a negative charge.
The tubule is harder still. Dialysis replaces filtration only. Reabsorption, acid-base balance and urine concentration are missing. Kidney organoids from stem cells now form segmented nephrons but remain at the maturity of the first or second trimester.
Uraemic toxins: targeting the gut
One section covers solutes poorly removed by dialysis, above all protein-bound toxins such as indoxyl sulfate and p-cresyl sulfate. Many originate from bacterial metabolism in the colon. The oral adsorbent AST-120 lowered indoxyl sulfate in two large trials but did not slow CKD progression. This shows the core problem: for many of these solutes, there is no proof lowering them benefits patients.
Making dialysis mobile and sustainable
Current machines consume large amounts of water, electricity and plastic and tie patients to a centre. Systems regenerating dialysate through sorbents in a closed loop are in early clinical testing. They still weigh several kilograms. The bottleneck is compact urea removal. One data point from the paper: in a 2025 study, 68% of surveyed US patients preferred home dialysis.
Vascular access: still the weak point
Here the paper is at its most explicit. Arteriovenous fistulas show primary failure rates of up to 60% in some populations and take months to mature. Grafts become infected more often, catheters most often. Materials have barely changed in decades. Patients name cannulation as their main concern. The authors call for access devices with built-in sensors, cannulation aids and mandatory outcome reporting, which exists neither in the USA nor in Europe for access surgery.
The implantable artificial kidney
Instead of rebuilding a healthy kidney in full, the authors define a minimum viable product. It should be fully implanted with no external connections, no larger than 750 ml, operate for at least two years and provide continuous small-solute clearance of 10 to 15 ml/min. For comparison, healthy kidneys provide about 120 ml/min. In the authors‘ view, potassium, volume and osmolality could still be managed with medication and diet. Long-term blood compatibility of the membrane, power supply and a regulatory pathway for such hybrid devices remain unsolved.
Pig kidneys and stem cells
Xenotransplantation is the most advanced field. Genetically modified pig kidneys have been transplanted into several living recipients. One recipient remained dialysis-free for 271 days before graft removal and later received a human kidney. In 2025, the FDA cleared the first clinical trial. Open questions include immunosuppression, risks from porcine viruses and long-term function.
Stem cell-derived kidneys are further away. Organoids lack ureters, their own vasculature and parts of the stroma, and after transplantation into mice some formed cartilage-like or Wilms tumour-like tissue. The authors expect first clinical steps within the next 5 to 10 years.
Patients as co-developers
A full section calls for involving patients from the start. The SONG-HD survey found patients rate the ability to travel and dialysis-free time higher than health professionals, who prioritise mortality and hospitalisation.
Limitations
This is an opinion and overview paper with a strongly US-centred view of regulation and reimbursement. Time frames such as 5 to 10 years are the authors‘ expectations, not evidence-based forecasts. Many authors have conflicts of interest: several founded or lead companies developing dialysis devices, membranes, uraemic toxin therapies or xenografts, and others hold patents on the described approaches. Keep this in mind when judging how close a technology appears.
What this means for practice
Nothing changes for patient care today. The paper does show where the bottlenecks are, and some of them are in our hands.
- Treat vascular access as a quality topic in its own right: record outcomes, train cannulation, audit primary failure.
- Offer home dialysis actively. Many patients want it, and new devices target exactly this setting.
- When wearable or implantable kidneys are announced, ask about operating lifetime, blood compatibility and regulatory status.
I am happy to discuss which of these fields will reach the ward first on LinkedIn.
Source: Winkler MKH, Ash SR, Barraclough KA, et al., Himmelfarb J. An integrative roadmap for a new era of kidney replacement therapies. Nat Rev Nephrol 2026, published online ahead of print. doi:10.1038/s41581-026-01146-1.
Source: DOI: 10.1038/s41581-026-01146-1 · Original article