Two Labels Say the Same Thing, Two Drugs Behave Differently: DOAC Choice in Mild Cirrhosis
Both drugs' labels say the same thing at the boundary he's approaching: avoid beyond this point. The actual exposure data behind that shared warning tells two different stories about which drug gets there faster.
C.O., a 58-year-old man, manages a family-owned vineyard, work that keeps him outdoors through most of the growing season and gives him, by his own description, more physical labor than most people his age take on voluntarily. He was found to have compensated cirrhosis two years ago, attributed to longstanding fatty liver disease, currently Child-Pugh class A with well-preserved synthetic function; he has no history of variceal bleeding or ascites, and his only other medical history is hypertension, controlled on a single low-dose agent. He was diagnosed with atrial fibrillation on a routine ECG three weeks ago, asymptomatic, found incidentally during a pre-surgical clearance visit for an unrelated knee procedure.
His CHA2DS2-VASc score is 1 — hypertension alone, since at 58 he is below the age band that scores points — which puts him in the group for whom anticoagulation is reasonable rather than mandated, and after discussion he chose to treat. Child-Pugh A places him below the threshold where either rivaroxaban or edoxaban's label formally says to avoid use — both direct their caution at Child-Pugh B and C. What the shared label language doesn't convey is that the underlying pharmacokinetic data behind it isn't symmetric: in subjects with Child-Pugh B impairment, rivaroxaban's drug exposure roughly doubles, a real, mechanistically concerning increase paralleled by greater factor Xa inhibition, while edoxaban's exposure at the same degree of impairment is essentially unchanged, even slightly decreased. Neither drug has outcome-trial data in patients with clinically meaningful hepatic impairment, since both pivotal atrial fibrillation trials excluded significant liver disease — so the choice between them, for a patient currently at the boundary rather than past it, rests on which drug's exposure curve is doing something concerning as liver function trends toward that boundary, not on which one the label technically still permits today.
At the pre-surgical/new-AF consultation
His CHA2DS2-VASc is 1 — hypertension only, no age points at 58 — so treating him is reasonable rather than obligatory, and he wants to treat. Child-Pugh A puts both rivaroxaban and edoxaban technically within label. I don't think the label distinguishes meaningfully between them here — either would be a reasonable standard choice.
I'd push back on treating them as interchangeable, even within the same labeled boundary. The actual exposure data at Child-Pugh B shows rivaroxaban's AUC roughly doubling, while edoxaban's is essentially flat, even slightly lower. He's compensated now, but cirrhosis trends in one direction more often than not — I'd rather start him on the drug whose exposure curve isn't already climbing steeply at the next stage he could plausibly reach.
That's a fair distinction, and I hadn't weighted the asymmetry in the exposure data as heavily as the shared label language. Edoxaban 60 mg daily. His creatinine clearance of 78 mL/min matters in both directions here — high enough that no renal dose reduction applies, and low enough to clear the other constraint people forget, since edoxaban is not to be used for atrial fibrillation at all above a clearance of 95 mL/min, where it was less effective than warfarin. Closer monitoring of his liver function at routine intervals rather than only if he develops overt decompensation.
Edoxaban 60 mg daily started. Liver function panel scheduled at routine intervals going forward, with an explicit plan to reassess anticoagulant choice if his Child-Pugh classification changes.
Agreed without real disagreement once the exposure data was named directly — both physicians noted this distinction is not something the shared label language communicates on its own, and is worth flagging any time a hepatic-impairment DOAC decision is being made near the labeled boundary rather than clearly on one side of it.