Intermittent Hemodialysis Fell Short in Critically Ill Acute Kidney Injury
In critically ill adults with acute kidney injury who required renal-replacement therapy, initial treatment with intermittent hemodialysis did not match continuous therapy on the prespecified noninferiority framework and carried a higher 90-day mortality signal. The multicenter trial enrolled patients across intensive care settings and compared two established dialysis strategies used at the bedside when kidney failure becomes severe enough to demand extracorporeal support. The central clinical question was not whether dialysis works — it does — but which mode better preserves survival and kidney recovery in patients already facing systemic instability.
A total of 554 patients were randomized, with 272 assigned to intermittent hemodialysis and 282 to continuous renal-replacement therapy. After exclusions from the intention-to-treat population, 535 patients remained for analysis: 260 in the intermittent group and 275 in the continuous group. The primary endpoint was a major adverse kidney event by day 90, defined as death, dependence on renal-replacement therapy, or a serum creatinine concentration more than 25% above baseline. By that composite measure, 78.1% of patients in the intermittent arm and 71.6% in the continuous arm reached the endpoint, yielding an adjusted absolute risk difference of 6.28 percentage points, with a 95% confidence interval from -1.22 to 13.79.
The trial stopped early after sequential interim analyses showed a sustained signal of higher mortality with intermittent hemodialysis than with continuous renal-replacement therapy. In an ICU population already exposed to profound hemodynamic stress, an early mortality divergence is not a trivial statistical event. It signals a clinically meaningful difference in how the two approaches interact with circulatory reserve, solute shifts, and fluid removal in patients who may have little physiologic margin left.
Mortality Divergence Was the Most Concerning Signal
Death by day 90 occurred in 164 patients, or 63.1%, in the intermittent group and in 147 patients, or 53.5%, in the continuous group. Kaplan-Meier mortality estimates at day 90 were 63.3% and 53.6%, respectively, with a hazard ratio of 1.30 and a 95% confidence interval of 1.04 to 1.63. In practical terms, the mortality curves separated enough to raise concern before the trial could continue to full completion. The direction of effect favored continuous renal-replacement therapy, and the investigators chose to halt the study once the mortality signal persisted across interim looks.
The biologic rationale is plausible. Intermittent hemodialysis removes fluid and solute over a shorter interval, which can impose abrupt osmotic and hemodynamic stress on already fragile patients. Continuous therapy spreads ultrafiltration and clearance across the day, a profile that better preserves intravascular stability when vasopressor dependence, shock, or multiorgan failure are part of the clinical picture. Even when both modalities correct uremia and electrolyte disorders, the tempo of therapy may matter when arterial pressure and organ perfusion are barely sustainable.
The composite kidney endpoint moved in the same direction, though the confidence interval crossed unity. That pattern does not erase the mortality signal; rather, it suggests that death drove much of the between-group separation. For bedside decision-making, survival often dominates the renal-recovery conversation in the acute ICU phase, especially when the patient is too unstable to tolerate rapid shifts in volume or blood pressure.
Hypotension Was Common in Both Arms, but It Did Not Explain Away the Difference
Hypotension remained the most common complication potentially related to acute kidney injury or renal-replacement therapy. It occurred in 179 patients, or 68.8%, in the intermittent group and in 183 patients, or 66.5%, in the continuous group. Those rates were high in both arms, reflecting how often shock physiology coexisted with kidney failure in this cohort. A simple reading would suggest that hypotension alone cannot fully account for the higher mortality in the intermittent arm, even if intermittent treatment can create sharper hemodynamic swings during treatment sessions.
Critically ill patients with dialysis-requiring acute kidney injury are not a homogeneous population. Some are recovering from sepsis, some from postoperative complications, some from cardiopulmonary failure, and some from multifactorial organ dysfunction that still has not stabilized. In such patients, the renal-replacement strategy is rarely an isolated nephrology choice; it is part of a larger hemodynamic package that includes vasopressor use, fluid balance, ventilation, and organ perfusion. Continuous therapy may offer a physiologic advantage precisely because it is less disruptive to that package.
The study design strengthens the relevance of the findings. It was multicenter, randomized, and open-label, and it tested the initial dialysis strategy rather than a rescue maneuver after deterioration. That framing mirrors real ICU practice, where clinicians must choose the first modality before the trajectory is fully clear. In that setting, an early mortality penalty attached to intermittent hemodialysis deserves serious attention.
What the Trial Changes at the Bedside
For critically ill adults with acute kidney injury who meet indications for renal-replacement therapy, these results do not support starting with intermittent hemodialysis when continuous therapy is available and feasible. The trial did not show noninferiority for the primary kidney endpoint, and the mortality signal tilted against intermittent treatment. In the context of an ICU population with more than half of patients dying by day 90 in both groups, even a modest relative increase in mortality carries substantial clinical weight.
The findings also sharpen the importance of matching modality to physiology. Continuous renal-replacement therapy appears better aligned with patients who cannot tolerate abrupt ultrafiltration or rapid solute shifts. Intermittent hemodialysis still has operational advantages in some hospitals, especially where staffing, equipment, or access to continuous circuits are limited, but this trial places a higher evidentiary burden on choosing it as the initial approach in unstable patients.
The trial leaves several practical questions unresolved. The population was critically ill and treatment-naive with respect to the randomized modality, but the exact subgroup most likely to benefit from one approach over the other is still not defined here. The article also does not supply a granular breakdown of which causes of acute kidney injury or which hemodynamic profiles drove the observed mortality signal. Those details would matter in deciding whether the advantage of continuous therapy is universal or concentrated in the sickest patients. Until those distinctions are clarified, the strongest message is straightforward: in this trial, intermittent hemodialysis as the initial strategy came with more deaths and did not earn noninferior status against continuous renal-replacement therapy.