When kidneys fail in the ICU, the first dialysis method may matter

On 14 October 2026 the New England Journal of Medicine published ICRAKI, a randomised trial in French intensive care units, sponsored by the Paris public hospital network (AP-HP) and funded by the French Ministry of Health. It enrolled critically ill adults whose kidneys had failed suddenly (acute kidney injury) badly enough that they needed a machine to take over the work of filtering the blood. Each patient was assigned at random to one of two ways of starting that support: intermittent haemodialysis, which cleans the blood in sessions of a few hours, or continuous renal-replacement therapy, which runs slowly around the clock.

The trial set out to show that sessions were "not worse", and could not

The main question was whether intermittent haemodialysis was no worse than continuous therapy on a combined kidney outcome by day 90: death, still needing dialysis, or blood creatinine more than 25% above the patient's baseline. Of 554 people randomised, 535 were analysed (260 in the intermittent group, 275 in the continuous group). The combined outcome occurred in 78.1% with intermittent sessions and 71.6% with continuous therapy, an adjusted difference of 6.28 percentage points (95% CI −1.22 to 13.79). That interval was too wide to support the "not worse" claim, so the trial's primary goal was not met. The trial reports a difference in outcomes; it does not establish the mechanism behind it.

Deaths by day 90 were more frequent with intermittent sessions

The trial was stopped early: repeated interim analyses showed a sustained signal of higher mortality with intermittent haemodialysis. By day 90, 164 people (63.1%) had died in the intermittent group and 147 (53.5%) in the continuous group. Kaplan–Meier estimates were 63.3% and 53.6%, a hazard ratio of 1.30 (95% CI 1.04 to 1.63). Low blood pressure, the most common complication potentially related to the kidney injury or the dialysis, occurred in 68.8% and 66.5% of the two groups, respectively.

These numbers also show how ill these patients were. More than half of those in the continuous group died within three months, so the comparison is between two ways of supporting people at very high risk, not a choice that makes kidney failure in intensive care safe.

An earlier French trial found no difference in survival

This is not the first test of the question. In a trial published in The Lancet in 2006, 360 patients with acute kidney failure as part of multiple-organ failure in 21 French intensive care units were randomised to the same two approaches, with strict protocols to make intermittent sessions gentler. Survival at 60 days was 32% with intermittent haemodialysis and 33% with continuous therapy, with no difference at any time point. The two trials report different survival findings, and their different designs matter when interpreting them.

Where the new result stops

Mortality was not the trial's primary outcome, and the trial was stopped early. The confidence interval for the hazard ratio runs from a 4% to a 63% higher hazard of death with intermittent sessions, so the size of the difference is uncertain even if its direction holds. Doctors and patients knew which treatment was used (an open-label design), and the trial ran in one country's hospitals. The result concerns only critically ill adults with acute kidney injury who already need dialysis; it says nothing about people on long-term dialysis for chronic kidney disease or about kidney injury that does not need a machine.

If someone close to you is in intensive care with kidney failure, the choice of dialysis method depends on their blood pressure, other organs and what the unit can provide. It is reasonable to ask the care team which approach they are using and why; discuss it with the treating doctor.

This material is for information only and does not replace a consultation with a doctor. Nothing in it is a recommendation to start, stop or change treatment.

Sources

• Gaudry S, et al. Intermittent or Continuous Renal-Replacement Therapy in Acute Kidney Injury. N Engl J Med. 2026. PMID 42873070

• ICRAKI — IHD Versus CRRT for Severe Acute Kidney Injury in Critically Ill Patients. ClinicalTrials.gov. NCT06032884

• Vinsonneau C, et al. Continuous venovenous haemodiafiltration versus intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial. Lancet. 2006. PMID 16876666