#ICU Evidence #001: Dobutamine during (mostly) abdominal surgery for septic shock: who actually needs it? A new multicenter RCT from China randomized 584 patients with septic shock undergoing surgical source control, predominantly abdominal, to intraoperative dobutamine + norepinephrine vs norepinephrine alone. The headline is impressive: AKI within 7 days fell from 34.7% to 22.8%. But it is worth looking under the hood: → Creatinine-only AKI: 15.5% vs 18.7% → Urine-output-only AKI: 0.7% vs 6.8% → Both criteria: 6.6% vs 9.2% A substantial part of the AKI separation came from urine-output-defined AKI. That doesn't invalidate the finding, but it is particularly interesting when the intervention is a hemodynamically active drug that may influence renal perfusion and urine output. The physiology is interesting too. With dobutamine, lactate was lower at 1 h (2.3 vs 2.8 mmol/L) and at the end of surgery (2.2 vs 2.6), ScvO₂ was higher at the end of surgery (80.4% vs 77.6%), and capillary refill time was shorter. Norepinephrine requirements were also lower: mean 0.11 vs 0.15 μg/kg/min and maximum 0.17 vs 0.23 μg/kg/min. So dobutamine clearly did something physiologically. But here is my problem: only 82/584 patients (14%) underwent advanced hemodynamic monitoring. Among those with paired measurements, cardiac index changed +0.5 vs −0.2 L/min/m² from baseline to the end of surgery (between-group difference +0.72 L/min/m²). Who actually needed an inotrope? Low cardiac output? Septic cardiomyopathy? Vasoplegia with preserved/high CO? Persistent hypoperfusion despite adequate flow? The trial randomized the drug—not the physiology. And here is a thought experiment: what if, instead of routinely adding dobutamine, we added a fixed dose of vasopressin? Could we also produce a favorable renal signal through an entirely different hemodynamic mechanism? Possibly. VANISH previously suggested a renal signal with vasopressin, including less kidney replacement therapy, despite a neutral primary outcome. My point is not that vasopressin would necessarily have been better. It is that improving a renal endpoint with a vasoactive drug does not necessarily tell us which hemodynamic abnormality we corrected or whether that abnormality was present in the first place. Another interesting observation concerns arrhythmias. Despite more tachycardia with dobutamine (41.0% vs 29.9%), atrial fibrillation was similar (2.8% vs 2.4%), as was ventricular tachycardia (0.7% in both groups). Even mean heart rate at the end of surgery was nearly identical: 97.4 vs 96.2 beats/min. Perhaps we sometimes overestimate dobutamine's arrhythmogenic contribution in septic shock? Or perhaps the underlying illness, concurrent norepinephrine exposure, and lower NE requirements with dobutamine matter more than we appreciate. With so few arrhythmic events and such short exposure, we cannot answer that question, but I find the observation intriguing. Then there are the outcomes patients care about most. ICU mortality was 13.1% vs 15.6%, in-hospital mortality 17.6% vs 20.7%, and kidney replacement therapy within 7 days 9.0% vs 10.9%. None of these differences was statistically significant. A personal note: I don't think it is realistic to expect every critical care trial to demonstrate a mortality benefit. Mortality has too many determinants, and adequately powering every ICU trial for mortality would often be impractical. And there is an additional consideration here: dobutamine was administered only intraoperatively, for a median of 155 minutes. Is it realistic to expect a few hours of improved hemodynamics to translate into a measurable mortality benefit weeks later? Perhaps not. Conversely, preventing kidney injury during a particularly vulnerable period could still have meaningful clinical consequences. So I would not dismiss this trial because mortality was unchanged. But when an intervention substantially improves an intermediate endpoint, I still want to see some indication that the benefit extends toward hard, patient-centered outcomes (dialysis, duration of organ support, ICU stay, functional recovery, or survival). My take: Interesting physiology, a striking AKI signal, and potentially reassuring short-term arrhythmia data. But this does not yet convince me to routinely add dobutamine without knowing the hemodynamic phenotype I'm treating. And credit to the investigators: this is a good, clinically provocative randomized trial addressing a difficult question. Medical Twitter will always find something to criticize -including me 😄- but good studies should make us think. This one does. One final thought: I wish we had more direct participation from Chinese clinicians and researchers on Medical Twitter. China produces important clinical research, and discussions like this would be even better if the investigators themselves could join us. We are missing valuable perspectives. #JAMA #FOAMed #FOAMcc jamanetwork.com/journals/jam…