Anticoagulating a Circuit in a Patient Who Cannot Clot: Citrate or Heparin for CRRT
A single patient with decompensated cirrhosis and hepatorenal-pattern AKI needing CRRT. The disagreement is whether an anticoagulant metabolized by the liver is a genuine contraindication, or whether the alternative brings its own real risk in a patient who cannot afford to bleed.
J.R., a 52-year-old former line cook, has not worked a kitchen in three years — not since his cirrhosis, the eventual result of two decades of heavy drinking he has now been sober from for four years, first put him in the hospital with variceal bleeding. He has been listed for liver transplant for eight months, and his admission this week, for worsening ascites and confusion, was meant to be a routine large-volume paracentesis and diuretic adjustment. Instead his creatinine has climbed steadily over five days despite volume optimization, meeting criteria for hepatorenal syndrome, and today, with urine output now under 100 mL over twelve hours and a rising potassium, the team has decided he needs continuous renal replacement therapy. His wife, who has sat with him through every admission this year, has already asked the one question everyone in the room is thinking — whether this changes his transplant timeline — and been told, honestly, that it depends partly on how the next several hours go.
His INR is 2.4 and his platelet count is 38,000 — not a laboratory artifact of liver disease but a real, functional bleeding risk that shapes every decision from here, including which anticoagulant, if any, keeps his CRRT circuit from clotting. Regional citrate anticoagulation works by chelating ionized calcium within the extracorporeal circuit itself, preventing coagulation there while calcium is replaced systemically to protect the patient — a mechanism that, on its surface, looks tailor-made for someone who cannot tolerate systemic anticoagulation. But citrate is metabolized primarily in the liver, converted to bicarbonate as part of the Krebs cycle — and a MELD-Na of 29 describes a liver that may well not clear it fast enough, risking citrate accumulation, worsening ionized hypocalcemia, and a rising total-to-ionized calcium ratio that itself becomes a new problem to monitor. Heparin, the more familiar circuit anticoagulant, avoids that specific accumulation risk entirely — but adds systemic anticoagulation on top of a patient whose own liver has already built in a coagulopathy that makes any additional bleeding risk a genuinely dangerous one.
ICU bedside, before the CRRT circuit is primed
My default for any patient with a real bleeding risk is regional citrate — it keeps the anticoagulant effect confined to the circuit and spares the patient systemically, which is exactly the property he needs given his platelet count and INR. The RICH trial, Zarbock's randomized comparison of regional citrate against systemic heparin in critically ill AKI patients on CRRT, found longer filter life with citrate and no survival difference, with more hypocalcemia as the price paid for it. I do want to name the real caveat directly rather than treat citrate as risk-free here: his liver clears citrate, and a failing liver may not do that fast enough, which can cause citrate accumulation and worsening hypocalcemia if we are not watching closely.
That caveat is the one I want to press on. Citrate accumulation in advanced liver failure is not a rare edge case — it is a described, real complication. And RICH is precisely the wrong trial to reassure us here: a MELD-Na of 29 describes hepatic failure those cohorts are thinnest on, which makes the one adverse signal it did report, hypocalcemia, land harder in him than in its own enrolled population rather than softer. It shows up as a rising total-to-ionized calcium ratio, worsening metabolic acidosis, and can be difficult to distinguish from the coagulopathy and metabolic derangement he already has from his underlying disease. I am not arguing for heparin instead — I understand the bleeding-risk logic against it — I am arguing that citrate needs closer, more frequent monitoring here than in a patient with normal hepatic clearance, not that it should be avoided.
If his total-to-ionized calcium ratio starts climbing, that is the specific signal to switch away from citrate before it becomes a bigger problem, not a reason to have avoided it from the start.
Agreed, and I will set the monitoring interval for total and ionized calcium at every four hours rather than the standard six for exactly that reason. I still think heparin is the worse choice here on balance — his coagulopathy already puts him close to a real bleeding event, and adding systemic anticoagulation on top of an INR of 2.4 and a platelet count of 38,000 is a harder risk to walk back than a monitored, reversible citrate accumulation.
The hepatologist's caution is well taken and is now built into the monitoring plan, not set aside — the actual disagreement, if there ever was one, was about vigilance during citrate use, not about which agent to start with.
Agreed: regional citrate anticoagulation started, with total-to-ionized calcium ratio and ionized calcium monitored every four hours rather than the standard interval, given his hepatic impairment.
Both physicians agreed explicitly that citrate accumulation, if it develops, is a signal to switch strategy rather than evidence the initial choice was wrong — a threshold (a rising total-to-ionized ratio) was set in advance rather than left to be judged in the moment.