Heparin Resistance on VV-ECMO: Replace the Cofactor, or Change the Drug
A single patient, four days into VV-ECMO, no longer responding to escalating heparin. The disagreement isn't whether he needs anticoagulation — it's whether the actual problem is a fixable cofactor deficiency or a reason to abandon heparin altogether.
J.K. supervises a night shift at a regional distribution warehouse and had never spent a night in a hospital before this admission, according to the wife who has been sleeping in the family lounge since he was cannulated. A rapidly necrotizing bacterial pneumonia — Staphylococcus aureus grew from his initial sputum and blood cultures — put him on VV-ECMO on his third ICU day after conventional ventilation failed to maintain adequate oxygenation despite a chest tube already draining a developing empyema. That tube matters to every decision that follows: an inflamed, actively draining pleural space on full systemic anticoagulation is a bleeding surface, and it sets a ceiling on how aggressively anyone can chase a therapeutic level. For the first three days his heparin infusion tracked therapeutic anti-Xa levels without much trouble, running at a dose proportionate to his 88-kilogram frame. Over the past eighteen hours that has changed — the infusion has been escalated repeatedly without reaching a therapeutic level, and today's circuit check found early fibrin deposition forming on the oxygenator membrane, visible enough on inspection that the perfusionist flagged it before the morning labs even came back.
His antithrombin activity, drawn this morning once the resistance pattern became clear, came back at 42% — low, and low enough to explain the resistance on its own, since heparin's anticoagulant effect depends entirely on antithrombin as a cofactor; without enough of it, even an escalating heparin dose can't produce the expected effect. His platelet count has also been trending down over the same 48 hours, which is why HIT antibody testing was sent this morning, result still pending. His inflammatory markers remain markedly elevated — a CRP over 300 — consistent with the kind of ongoing consumptive, pro-thrombotic state that can drive both his falling antithrombin and his falling platelets simultaneously, independent of any single cause. What makes this genuinely hard is that the evidence favoring the alternative is thin in a specific way. The comparison most often cited — Rabinowitz's ASAIO meta-analysis, in which heparin-treated ECMO patients had roughly twice the odds of circuit thrombosis as bivalirudin-treated ones — pools six retrospective cohorts, not a single randomized trial. Nobody in this room has ever seen that question answered properly, and his oxygenator is already laying down fibrin while they decide.
A resistant circuit, a pending antibody test
I'd give antithrombin concentrate and continue heparin rather than switch drug classes. His resistance has a clean mechanistic explanation — antithrombin at 42% — and heparin needs antithrombin as a cofactor to work at all. Replace what's actually deficient and heparin should become effective again at a reasonable dose, without taking on bivalirudin's unfamiliar monitoring and added cost before we've even tried fixing the mechanism we already understand.
Antithrombin deficiency in critical illness is common, and ECMO itself lowers antithrombin, but replacement often doesn't fully correct heparin resistance — consumption tends to keep outpacing what's given, especially with active circuit clotting already underway. I'd switch to bivalirudin directly. Rabinowitz's meta-analysis in ASAIO pooled six retrospective ECMO cohorts and found heparin recipients had about twice the odds of circuit-related thrombosis — odds ratio 2.05 — and I'd flag what that is before anyone treats it as more: retrospective cohorts, not trials, with all the selection bias that carries. It also sidesteps a HIT question we haven't answered yet, since bivalirudin inhibits thrombin directly and needs no antithrombin at all, rather than treating around a heparin problem while his platelet count keeps falling.
I take the cofactor mechanism seriously — I just don't think naming the mechanism means correcting the level reliably fixes the resistance at the bedside, especially with his platelets already trending the direction they are.
I'd come down on your side of it, but not for the reason you gave. The meta-analysis is six retrospective cohorts and I wouldn't move a patient on that alone. What decides it is that heparin's entire mechanism runs through a cofactor he is measurably short of and actively consuming, while bivalirudin inhibits thrombin directly and needs no cofactor at all — so switching removes the failure mode rather than chasing it. Repleting antithrombin treats the number; changing the drug treats the dependency.
Where I'd push back on both of you is the monitoring, which neither of you has mentioned and which will undo whichever drug you pick. aPTT is unreliable in a consumptive state like his. If we stay on heparin it has to be anti-Xa; if we move to bivalirudin it has to be an ecarin-based assay. Switching drug classes without fixing that just trades one poorly-controlled anticoagulation state for a different poorly-controlled one.
Agreed: send the HIT antibody panel (already sent this morning) and switch to bivalirudin now, dosed against a bivalirudin-specific therapeutic range using an ecarin-based assay rather than aPTT, given the unreliability of aPTT in his current consumptive state.
Not agreed:
What to do if the HIT panel returns negative. The critical care physician would switch back to heparin — cheaper, more familiar to the team, and with the antithrombin question still potentially correctable — while the hematologist would keep him on bivalirudin for the remainder of his ECMO course regardless of the antibody result, given that today's resistance and early circuit clotting happened on heparin whatever their ultimate cause turns out to be.