Vasopressor Selection in Septic Shock with Emerging Kidney Injury
A single patient, six hours into septic shock resuscitation with rising norepinephrine requirements and early oliguric kidney injury. The disagreement is which second agent, if any, actually protects the kidney rather than just the blood pressure number.
T.M., a 58-year-old man, has driven long-haul freight for the better part of three decades, sleeping in his cab more nights than not and, by his wife's account on the phone, treating his blood pressure pills about the way he treats his DOT physicals — something to catch up on eventually. He had been “running hot” for two days, she said, before he stopped answering his phone at a rest stop and a fellow driver called it in. He arrived febrile to 39.8°C and hypotensive, and a CT scan found an 8mm stone obstructing his right ureter with hydronephrosis behind it — the infection has a source, and for now it still has one, since urology has staged his decompression for once he is stable enough to tolerate sedation.
Six hours into resuscitation, his norepinephrine requirement has climbed to 35 mcg/min — a dose high enough that the drug holding up his pressure is now plausibly also throttling blood flow to the kidney it is meant to be protecting through perfusion alone. His urine output has run under 0.5 mL/kg/hr for eight hours, and his creatinine, an estimated 1.0 mg/dL on arrival given no prior labs on file, is 2.3 mg/dL tonight — a real, fast-moving injury layered onto an unresolved source of infection, not a chronic baseline finally being noticed. His lactate sits at 4.1 and has not meaningfully cleared. The obstruction stays in place for now, which is its own argument for caution: whatever vasopressor strategy the team settles on tonight has to hold up for hours more before decompression, not just clear the next pressure check. He has no history of nephrolithiasis and no known baseline kidney disease of any kind — part of why tonight reads as a genuine acute departure rather than a chronic problem finally catching up with him. Twelve hours ago he had, by every available account, two ordinary kidneys; tonight one of them sits behind an undrained obstruction and both are downstream of a pressor dose the team is no longer certain is doing them any favors.
At the bedside, six hours into resuscitation
At 35 mcg/min his own vasopressor is now plausibly doing some of the damage his kidneys are showing tonight. I want vasopressin added now, not after we have squeezed another 10 mcg/min out of norepinephrine. I want to state VANISH accurately rather than in my own favor: it randomized vasopressin against norepinephrine as an early, not rescue, agent, and its primary kidney-failure-free-days endpoint did not favor vasopressin — the trend actually ran the other way. What it did show was less renal replacement therapy in the vasopressin arm, 25.4% against 35.3%, and RRT is the endpoint I am actually trying to keep him away from tonight. Adding a second mechanism now lets us flatten the norepinephrine curve instead of climbing it further.
And I will not lean on VASST's subgroup, because he is not in it: that mortality signal came from patients randomized on under 15 mcg/min of norepinephrine, and he is at 35. The argument is about sparing further catecholamine-driven vasoconstriction starting now, while there is still kidney left to spare — not about a subgroup that describes someone else.
You are right that his catecholamine burden is already a real cost, and I am not arguing to keep climbing norepinephrine instead. But vasopressin is still a nonselective vasoconstrictor systemically — it does not target renal perfusion specifically, it just moves some of the load off alpha-1 receptors. Angiotensin II is the one agent in this class with a renal-specific mechanistic argument: at the efferent arteriole, AT1-receptor agonism raises glomerular filtration pressure the same way withdrawing an ACE inhibitor does, rather than simply constricting systemically like the other two.
ATHOS-3's own post hoc analysis, patients already on renal replacement therapy at enrollment, found improved survival and a higher rate of RRT liberation with angiotensin II over placebo. He is not on RRT yet, so I will say plainly that applying that finding to him tonight is extrapolation, not a population match — but his trajectory, oliguric, KDIGO stage 2 and climbing, with an unrelieved source, is heading toward exactly that population if nothing changes.
I want to name ATHOS-3 as precisely as it was studied: a sicker population than T.M., higher baseline severity scores, many already on RRT, and its own safety data recorded real arterial and venous thromboembolic events on the drug. Reaching for it tonight, in a patient whose actual source of infection is still sitting undecompressed inside him, adds a second agent's own risk on top of an already unstable picture, for a renal argument that is still one step ahead of where he actually is.
And I will grant the point he just made against himself: he is not claiming VASST's subgroup covers this patient, and he named VANISH's primary endpoint as the negative it was. That leaves the catecholamine-sparing case plus a real reduction in RRT — thinner than it first sounded, but resting on the whole of two trials rather than on one post hoc slice, and it does not foreclose angiotensin II later — it just does not reach for the least-studied option first, in the patient furthest from the population it was actually tested in.
Agreed: vasopressin started as a second agent tonight, norepinephrine held rather than further escalated, with a plan to reassess after urologic decompression. Angiotensin II was not started, but was explicitly named as the next step if he progresses to renal replacement therapy rather than left as an unstated possibility.
Not agreed: whether angiotensin II should have gone first given his already-oliguric trajectory, or only once RRT is actually reached. The nephrologist's mechanistic case was accepted as reasonable but not acted on tonight — the emergency medicine physician's read that reaching for the least-studied agent in the patient furthest from ATHOS-3's actual population added avoidable risk carried the room, without either side treating the other's reasoning as wrong.