Deaths Halt Trial of Hemodialysis for Acute Kidney Injury

— ICU patients who received initial intermittent hemodialysis faced higher mortality

Intermittent hemodialysis failed to prove noninferior to continuous renal-replacement therapy for critically ill patients with acute kidney injury and was associated with a higher incidence of death, according to findings from the randomized ICRAKI trial.

In the intention-to-treat analysis of 535 patients, a major adverse kidney event by day 90 -- the trial's primary endpoint -- occurred in 78.1% of the intermittent group versus 71.6% of the continuous group, reported Stéphane Gaudry, MD, PhD, of Avicenne Hospital in Bobigny, France, and colleagues.

But the open-label trial's data and safety monitoring board halted the trial early after a higher mortality signal turned up in the intermittent group. By day 90, deaths occurred in 63.1% of that arm versus 53.5% of the continuous arm, with Kaplan-Meier mortality estimates suggesting a 30% higher risk of death with the approach (HR 1.30, 95% CI 1.04-1.63).

"Although early stopping was justified on ethical grounds, it may have reduced the precision of treatment-effect estimates and led to overestimation of the magnitude of the observed effect on mortality," Gaudry's team noted in the New England Journal of Medicineopens in a new tab or window.

Most of the study participants were in the intensive care unit (ICU) for medical rather than surgical reasons, with about half experiencing acute respiratory distress syndrome at baseline and over 70% receiving invasive mechanical ventilation at the time of ICU admission. Common coexisting conditions included hypertension (50%), type 2 diabetes (28%), chronic kidney disease (19%), and congestive heart failure (11%).

The researchers called the higher mortality in the intermittent hemodialysis group "unexpected" based on prior literature and "not readily explained."

"Post hoc analyses suggested greater hemodynamic stress with intermittent hemodialysis," they acknowledged. "These exploratory findings are hypothesis generating and do not establish a causal mechanism."

The research team also noted that earlier trialsopens in a new tab or window were conducted before trial-proven criteria for initiating renal-replacement therapy were established, whereas ICRAKI enrolled only patients with documented indications. "This shift in the target population may partly explain the differences between our findings and those of earlier studiesopens in a new tab or window," Gaudry and co-authors suggested.

In an accompanying editorialopens in a new tab or window, Paul Palevsky, MD, of the Veterans Affairs Pittsburgh Healthcare System and University of Pittsburgh School of Medicine, explained that while randomized trials comparing the two modalities have "been challenging" to perform, earlier small-to-moderately sized trials have largely shown no difference in survival.

Palevsky detailed the distinct trade-offs of each approach. Intermittent hemodialysis corrects fluid and electrolyte imbalances and removes toxins rapidly, and a single machine can treat multiple patients successively. However, this "can result in rapid fluid and electrolyte shifts that may contribute to hemodynamic instability."

Conversely, continuous renal-replacement therapy provides lower solute clearance and slower volume removal, "theoretically causing less hemodynamic stress," noted Palevsky.

Although clinical practice guidelines currently suggest continuous therapy for hemodynamically unstable patients, Palevsky suggested ICRAKI's findings raise a broader question. "The real question is not which specific method is superior, but how best to integrate the use of these multiple therapies to improve patient outcomes," he wrote.

"In some healthcare systems, virtually all renal-replacement therapy in the ICU is provided as continuous renal-replacement therapy, with patients transitioning to intermittent therapy when they are ready to transfer out of the ICU; in other systems where both intermittent and continuous therapy are available, continuous renal-replacement therapy is used in patients with greater hemodynamic instability," Palevsky explained. "In comparing these strategies, the difference does not hinge on the initial method of renal-replacement therapy."

Originally planned for 1,000 patients, the multicenter trialopens in a new tab or window enrolled 554 participants across ICUs in France from October 2023 to July 2025. The median duration of renal-replacement therapy was 5 days in the intermittent group and 7 days in the continuous group.

Assigned therapies were maintained for at least 4 consecutive days unless kidney recovery occurred. Thereafter, clinicians could switch treatment modalities based on usual clinical practice. "The trial therefore does not establish which method is preferable for treatment after the first 3 days, when the risk-benefit trade-offs may differ," the study authors explained.

Treatment switching occurred in 21.9% of the intermittent group and 42.5% of the continuous group. Hypotension was the most common complication in both arms, occurring in 68.8% of intermittent and 66.5% of continuous therapy patients.

Prespecified sensitivity and subgroup analyses were consistent with the main analyses.