Mortality Split Emerges Between Intermittent and Continuous Therapy in Acute Kidney Injury
In critically ill adults with acute kidney injury who required renal-replacement therapy, the choice between intermittent hemodialysis and continuous renal-replacement therapy was not clinically neutral. A multicenter, open-label, noninferiority trial found that intermittent hemodialysis failed to match continuous therapy on the prespecified kidney-centered composite and was associated with higher mortality by day 90.
The study randomized 554 patients, with 272 assigned to intermittent hemodialysis and 282 assigned to continuous renal-replacement therapy. The intention-to-treat population included 535 patients: 260 in the intermittent arm and 275 in the continuous arm. The trial was stopped early after sequential interim analyses repeatedly pointed toward higher mortality with intermittent hemodialysis. In a field where renal support is often chosen on the basis of hemodynamics, staffing, and ICU logistics, that signal carries immediate bedside weight.
A major adverse kidney event by day 90 occurred in 203 patients in the intermittent group and 197 patients in the continuous group, corresponding to 78.1% and 71.6%, respectively. The adjusted absolute risk difference was 6.28 percentage points, with a 95% confidence interval from -1.22 to 13.79. On the prespecified noninferiority framework, intermittent hemodialysis did not earn equivalence on the primary endpoint. The composite was broad by design, combining death, ongoing dependence on renal-replacement therapy, and persistent creatinine elevation above baseline, so the endpoint reflects both survival and incomplete renal recovery rather than a single narrow renal measurement.
Higher 90-Day Mortality Favors Continuous Renal Support
The most consequential divergence came from survival. Death by day 90 occurred in 164 patients in the intermittent arm and 147 patients in the continuous arm, or 63.1% versus 53.5%. Kaplan-Meier estimates at day 90 were 63.3% and 53.6%, and the hazard ratio for death was 1.30, with a 95% confidence interval of 1.04 to 1.63. That mortality gap is difficult to dismiss as a statistical artifact because the trial was stopped early only after repeated interim analyses continued to show the same direction of effect.
Intermittent hemodialysis produces rapid solute and volume shifts over a shorter treatment window. In unstable ICU patients, those shifts can challenge vascular tone, myocardial performance, and cerebral perfusion at the exact moment when multiorgan failure has already narrowed physiologic reserve. Continuous renal-replacement therapy, by contrast, distributes ultrafiltration and solute clearance over time, which is biologically more compatible with shock states, vasopressor exposure, and fragile perfusion. The mortality signal in this trial fits that physiologic logic, even though the study was not designed to prove a single mechanism.
The practical consequence is direct: when critically ill patients with acute kidney injury already meet criteria for renal support, continuous therapy should not be treated as merely a convenience-driven alternative. The observed survival difference suggests that treatment tempo may matter as much as the decision to start dialysis itself.
Hypotension Remained Common in Both Arms
Hypotension was the most common complication potentially related to acute kidney injury or renal-replacement therapy, occurring in 179 patients in the intermittent group and 183 patients in the continuous group, or 68.8% and 66.5%, respectively. The frequency was high in both groups, which is consistent with the severity of illness in this cohort and with the fact that renal-replacement therapy was being delivered to patients already sick enough to need ICU-level organ support.
The near overlap in hypotension rates should not be read as evidence that the two strategies were interchangeable. Bedside hemodynamic tolerance is only one part of the story. Intermittent treatment compresses fluid and solute removal into shorter sessions, while continuous therapy spreads the burden across the day. Even when hypotension is recorded at similar frequency, the cumulative physiologic stress may differ in ways that are not captured by a binary adverse-event tally. In a population with high baseline mortality, small differences in perfusion stress can plausibly translate into different survival curves.
The trial also leaves no room for casual assumptions that intermittent hemodialysis is the default safer choice simply because it has long been used in ICUs. The data move in the opposite direction. Continuous therapy held the advantage in the survival analysis, and the kidney composite did not favor intermittent treatment enough to offset that signal.
What the Trial Means for ICU Renal-Replacement Decisions
For clinicians, the key issue is not whether intermittent hemodialysis is effective in general; it clearly remains a powerful and widely used kidney replacement method. The issue is whether it should be the initial treatment of choice in critically ill adults with acute kidney injury who already have an indication for renal-replacement therapy. In this randomized trial, the answer was no. Intermittent hemodialysis was not shown to be noninferior to continuous renal-replacement therapy, and mortality was higher in the intermittent group.
The open-label design is unavoidable in a trial comparing two dialysis strategies that differ visibly in delivery, but it also means that the main findings rest on hard outcomes rather than blinded subjective assessments. The sample size was moderate, and the study ended early because interim analyses kept pointing in the same direction. Early stopping can limit precision, yet persistent mortality separation is still difficult to ignore when it appears across sequential analyses and is accompanied by a hazard ratio above 1.0.
For ICU teams, this trial reinforces a simple but consequential point: renal replacement is not merely about clearing creatinine or removing fluid. The method chosen can shape the patient’s hemodynamic burden, exposure to treatment-related instability, and possibly survival. In a patient with vasopressor dependence, shock physiology, or limited cardiopulmonary reserve, the more gradual profile of continuous therapy now has randomized evidence behind it in the initial-treatment setting.
The remaining uncertainty is not whether both modalities can support kidney failure. It is how best to match modality to the patient at the moment dialysis is first prescribed, when organ failure is active and the margin for physiologic stress is thin. This trial supplies a clear signal that, for critically ill patients with acute kidney injury and an indication for renal-replacement therapy, intermittent hemodialysis should not be assumed to be the equivalent starting strategy.