Terlipressin or Norepinephrine for HRS-AKI: Setting Decides As Much As Pharmacology
Two cirrhotic patients meet criteria for hepatorenal syndrome-acute kidney injury within the same week — one already in the ICU on norepinephrine, one on the ward being considered for terlipressin. The pivot isn't which drug is better in the abstract; it's whether the setting each patient is actually in changes the right answer.
Marcus T., a 61-year-old man with decompensated cirrhosis from longstanding alcohol use and a creatinine that had sat at 1.1 through every clinic visit of the past year — meaning the kidney now failing was, until this week, working normally — was admitted three days ago for worsening ascites and was started on norepinephrine in the ICU yesterday after his creatinine climbed from a baseline of 1.1 to 2.6 despite volume expansion and holding diuretics — a rise of that size over roughly 48 hours, with the two reversible causes already excluded, is what converts a number into an HRS-AKI diagnosis rather than a prerenal one still worth fluid-challenging. His MAP is 68 on the infusion already running, which is the detail that matters most for what comes next: he is not being started on a vasoconstrictor, he is already on one, and it is working. He has no other organ failure, and his intubation is for airway protection during a brief episode of hepatic encephalopathy that has since cleared rather than for respiratory failure — a distinction worth stating plainly, since an intubated cirrhotic patient reads as sicker than he is. The arterial line is in, the ICU bed is his, and the monitoring burden either drug would require is already being carried. His case isn't a question of whether to treat — the ICU team already has — it's whether to keep going with what's already infusing or switch to a drug built specifically for this diagnosis.
For Marcus, I'd keep norepinephrine going rather than switch him to terlipressin. He's already on it, already monitored, and trials run in ICU populations have found the two agents comparably effective for HRS-AKI renal response. Switching a stable infusion to a specialty-pharmacy drug for a patient already being watched the way terlipressin requires isn't caution — it's just added cost and a med-reconciliation risk for no real gain.
I'll grant the ICU-equivalence data here — that's real, and I'm not fighting to switch him. But it's worth naming why this case is easy: he's already got the exact monitoring level either drug would need. That won't be true for the next HRS-AKI patient we see.
Calling this settled misses that 'comparably effective in ICU populations' is doing real work in that sentence — the equivalence data doesn't automatically travel to a patient who isn't in an ICU.
Norepinephrine continued unchanged. Both physicians agreed the switch offered no real benefit given his existing monitoring — the one point of full agreement in the case.
Renata S., a 58-year-old woman with cirrhosis from hepatitis C, cured of her infection three years ago but left with decompensated disease — a reminder that eradicating the virus arrests the injury without undoing the architecture already built — and with a baseline creatinine of 0.9 documented as recently as her last outpatient panel, was admitted to the general medicine ward for a spontaneous bacterial peritonitis episode now three days into antibiotic treatment. Her creatinine has risen from a baseline of 0.9 to 2.1 over 72 hours despite albumin expansion and three days of appropriate antibiotics — and it is the failure to respond to treated infection plus albumin, not the absolute number, that meets HRS-AKI criteria here, since a creatinine of 2.1 in a woman her size represents a steeper fall in filtration than the same figure would in Marcus. What she does not have is any of the monitoring Marcus already has: no arterial line, no ICU bed currently available, standard ward telemetry only. Where Marcus's monitoring already exists, hers would have to be built around whichever drug she starts — and that difference, not anything about her kidneys specifically, is what the team is actually arguing about.
Renata is the actual disagreement. Terlipressin is the agent with the FDA-approved indication for HRS-AKI specifically, and CONFIRM's own composite renal-response endpoint is the largest trial base either drug has for this exact diagnosis. Starting her on norepinephrine on a general ward, off-label for this indication, isn't the conservative choice just because it's more familiar.
I'm not arguing the trial base — I'm arguing the room she's in. Norepinephrine needs a central line and pressor-level monitoring, which we can arrange on the ward tonight. Terlipressin's own label warns specifically about respiratory failure in patients without close monitoring, and 'close monitoring' means more than what a general medicine floor currently gives her.
Calling norepinephrine the safe default for her skips the actual question, which is whether a ward patient without an arterial line can be safely started on either vasoconstrictor without more monitoring than she currently has — and that's a monitoring problem, not an argument for one specific drug over the other.
The monitoring question is exactly where I'd land this. Terlipressin's respiratory-failure warning isn't diffuse risk — it concentrates in patients without close monitoring, which is precisely Renata's situation and precisely not Marcus's. For Renata, starting terlipressin tonight means either securing closer monitoring first or accepting a risk the label specifically warns is worse without it. Norepinephrine, needing a central line but not carrying that same labeled respiratory signal, may genuinely be the lower-risk start for her even though it isn't the disease-specific drug.
Started on norepinephrine via a newly placed central line rather than terlipressin, with a plan to reconsider terlipressin if an ICU bed opens or her respiratory status is reassessed as low-risk enough to proceed on the ward.
Not agreed: the hepatologist views this as a genuine undertreatment risk given terlipressin's stronger disease-specific trial base, and flagged it explicitly rather than let it read as full team consensus.