Early vasopressors are superior to fluid boluses for septic shock--at preventing pulmonary edema? (ARISE FLUIDS Trial)
The debate continues, if you really want it to
Sepsis resuscitation decisions exist in a tension between two competing instincts: give fluids aggressively to restore perfusion, or limit initial fluid boluses in favor of early vasopressors to avoid potentially harmful fluid overload.
In the 1999 EGDT trial (Rivers NEJM), by repeatedly bolusing patients to a central venous pressure goal, the intervention arm received five liters of crystalloid (and the controls three-and-a-half) in the first six hours, with both groups exceeding +13 liters infused in the first 72 hours. Perhaps unsurprisingly, more than 60% required mechanical ventilation.
With its likely spurious signal of a mortality benefit in the goal-directed arm, this trial encouraged a generation of intensivists to continue the longstanding practice (which predated Rivers) of infusing unnecessarily large volumes of fluids into septic patients in the hope of restoring and maintaining perfusion.
Fifteen years later, the Process, Arise, and Promise trials (c. 2014) refuted the Rivers “goal-directed” package of dobutamine, transfusion, and central venous catheterization, without challenging the dominant fluid-forward paradigm of sepsis care. Patients in those trials received ~2 liters before randomization and another 2-3 liters in the first six hours (~3.5 to 5.5 liters total in 6 hours, generally).
Yet those three trials also showed clinical practice moderating toward less-aggressive fluids, with (e.g.) patients in Process and Arise “only” receiving an additional 4.5 to 5 liters from 6 to 72 hours (compared to ~8 liters in Rivers). There was no difference in mortality between arms in any of the three trials.
This set the stage for two large randomized trials testing a more restrictive (i.e., vasopressors-forward) approach to fluid resuscitation in sepsis after an initial bolus of ~2-3 liters. Neither showed a significant difference between the two fluid strategies:
CLASSIC (NEJM 2022, n=1,554, Europe): identical 90-day mortality (very high at ~42%), with only 500 mL (restrictive) vs 1,300 mL (liberal) infused on the first day after randomization (plus ~3 liters in each group before enrollment);
CLOVERS (NEJM 2023, n=1,563, U.S.): 90-day mortality 14% (restrictive) vs 14.9% (liberal), patients received ~2 liters before randomization, then 500 mL (restrictive) vs. 2,300 mL (liberal) in the first 6 hours; only 37% of liberal-arm patients required vasopressors vs 59% (restrictive) on the first day
Not to be outdone, the Australasian critical care consortium ANZICS recently published results from the ARISE FLUIDS trial, which will not extinguish the controversy over this deceptively simple-seeming clinical question.
ARISE FLUIDS Trial
At 51 sites in Australasia, 1,000 adults presenting to the ED receiving a diagnosis of sepsis, high lactate and hypotension after ≤1,000 mL I.V. fluid (but ≤2,000 mL) were randomized to six hours of either a restrictive (vasopressor-forward) or liberal fluid strategy (+1,000 mL then 500 mL boluses, 30 mL/kg in 3 hours advised, plus vasopressors after that or when deemed fluid-unresponsive).
There was meaningful separation: after the ~1,500 mL pre-randomization, restricted patients received only an additional ~500 mL in the first 6 hours vs. ~1,500 mL in the liberal arm (~2 L vs. ~3 L), with a ~1.1 L difference at 24 hours (~2,640 mL vs ~3,740 mL).
Fluid restriction led to higher vasopressor use in the first 24 hours (87% vs 68%)—about an additional one in five patients requiring vasopressors in the restricted arm.
No difference in the primary outcome: both arms had an identical number of 76 days alive and out of the hospital through day 90.
Mortality at 28 days (12.9% vs 10%) and 90 days (16.4% vs 14.4%) was nominally higher in the vasopressor group (non-significant).
There was eightfold more pulmonary edema in the liberal group: 5% vs. 0.6%, statistically significant. Pulmonary edema was identified by clinician opinion, based on hypoxemic respiratory failure, chest radiography, or other supporting findings. Clinicians were unblinded to treatment assignment, but were usually caring for study patients well past the six-hour mark, and thus not involved directly in the trial.
Despite this increase in clinician-identified pulmonary edema, the rate of mechanical ventilation was equal in both groups (15%).
Vasopressor-forward patients required ICU admission more often (77% vs 68%). Understandable, as many hospitals require vasopressors to be managed in ICUs.
Patients in the liberal group also required about one additional day of mechanical ventilation and renal replacement therapy—although these were not statistically significant.
Central lines were largely unnecessary. Most patients received vasopressors through a peripheral IV, without any major safety signal reported in the trial.
What to Take From Arise-Fluids?
Arise-Fluids “settles” the controversy of fluids-versus-vasopressors for septic shock by again demonstrating their general equivalence when all patients’ outcomes are combined.
But partisans can find favor in its data for their preferred tilt in either direction:
Restrict Fluids! After an initial fluid bolus of one or two liters for patients without volume depletion, restricting fluids and relying on vasopressors to maintain perfusion is reasonable. 30 cc/kg was always an arbitrary figure; over-resuscitation causes pulmonary edema and possibly adds time on the ventilator or dialysis circuit. Iatrogenic volume overload is an under-recognized cause of disability and incomplete recovery from critical illness, which Arise-Fluids did not measure.
Bolus Away! Liberal fluid resuscitation is generally superior from a practical clinical standpoint. A fluid-forward approach reduces the need for vasopressors, which in turn can reduce the need for ICU admission, conserving resources and improving care delivery system-wide. Pulmonary edema is not a hard clinical outcome and did not increase the need for mechanical ventilation either in Arise-Fluids or in Clovers, the most comparable major trial. The trend toward faster weaning from mechanical ventilation in the restrictive group in Arise-Fluids was not significant. Further, the numerically higher mortality in the restrictive arm is a potential warning of the limits of safe fluid restriction.
Fluids vs Vasopressors: A False Dichotomy
U.S. clinicians have appropriately chafed at policymakers’ attempts at national standardization of sepsis care via the governmental “quality” measure called SEP-1; this resistance has been in large part because of its mandate for providing three, four, or five-liter boluses to severely obese patients with sepsis who may also have concomitant cardio-pulmonary-renal failure.
Subsequent iterations of the rule permit clinicians to deliver smaller fluid boluses they consider beneficial, with documentation of their clinical reasoning.
The Arise-Fluids and Clovers trials together support the safety of just such an individualized, patient-centered approach to resuscitation and hemodynamic management. Deferring vasopressors for aggressive fluids for tolerant or volume-depleted patients; limiting fluid boluses and favoring vasopressors in patients with apparent impending respiratory failure or significant cardio-renal disease: the complexity of sepsis and each patient’s physiology necessitate thoughtful tailoring of their plan, rather than blindly following a trial result.
Rather than picking a winner in the fluids-vs-vasopressors debate, the Arise-Fluids trial reaffirms the truth that the best approach to sepsis resuscitation is the one tailored to the patient in front of you.
References
ARISE-FLUIDS trial. Vasopressors or Fluids in Early Septic Shock. New England Journal of Medicine. Published online June 11, 2026. doi:https://doi.org/10.1056/nejmoa2516225
Meyhoff TS, Hjortrup PB, Wetterslev J, et al. Restriction of Intravenous Fluid in ICU Patients with Septic Shock. New England Journal of Medicine. 2022;386(26):2459-2470. doi:https://doi.org/10.1056/nejmoa2202707
The National Heart, Lung, and Blood Institute Prevention and Early Treatment of Acute Lung Injury Clinical Trials Network. Early Restrictive or Liberal Fluid Management for Sepsis-Induced Hypotension. New England Journal of Medicine. 2023;388(6). doi:https://doi.org/10.1056/nejmoa2212663
These trials may be most useful when they loosen a doctrine rather than create a new one. For years the danger was that sepsis resuscitation became morally attached to volume: giving more fluid felt like doing more resuscitation. ARISE-FLUIDS and CLOVERS make that reflex harder to defend, without handing clinicians a replacement reflex to apply blindly.
The useful residue may be proportionality. A liter has a different meaning in the patient who is volume-depleted than in the patient already carrying respiratory or cardio-renal vulnerability. The trial does not remove judgment; it makes the judgment harder to outsource to a rule.
That is why equivalence may be less boring than it sounds. When two strategies look broadly similar in the aggregate, the bedside question shifts from which strategy won to which patient has the least margin for the wrong excess. In sepsis, restraint and rescue are not opposite virtues. The work is knowing which one the physiology is asking for now.
They just can’t let go of this religious belief, despite multiple randomized trials showing that liberal fluid strategy is not harmful.