Anticoagulating a Patient the Old Framework Called 'Auto-Anticoagulated'
For years, advanced liver disease was treated as its own kind of blood thinner, a condition that made real anticoagulation seem redundant. Current evidence says his atrial fibrillation doesn't care what his liver is doing, and neither does his stroke risk.
R.H., a 66-year-old man, worked the docks for over thirty years before retiring, and still meets his old crew for coffee most mornings out of habit as much as friendship. He has Child-Pugh B cirrhosis from decades of heavy alcohol use, now eleven years sober, complicated by portal hypertension with small esophageal varices found on a surveillance endoscopy last year, never bled, currently on a non-selective beta-blocker for primary variceal prophylaxis. He has hypertension and a remote history of peripheral vascular disease, both stable on current management, and no history of falls or bleeding of any kind since his cirrhosis diagnosis. He was found to have atrial fibrillation on a routine ECG two weeks ago during a pre-colonoscopy clearance visit, asymptomatic, new-onset by history.
His CHA2DS2-VASc score clearly indicates anticoagulation, and the instinct to treat his liver disease itself as protective — the older framework that considered advanced liver disease a kind of built-in, auto-anticoagulated state — has been directly overturned by more recent evidence. Advanced liver disease is now recognized as a genuinely prothrombotic condition on balance, not merely a bleeding disorder, and current International Society on Thrombosis and Haemostasis guidance explicitly recommends standard-dose direct oral anticoagulants in Child-Pugh A or B cirrhosis with atrial fibrillation, following the same cardiology guideline recommendations used for patients without liver disease — it is only in Child-Pugh C that the evidence becomes genuinely inadequate to guide the decision. His varices complicate the bleeding side of the calculation, but they are a reason for careful management, including continued variceal prophylaxis, not a reason to withhold anticoagulation he would otherwise clearly warrant.
At the hepatology/cardiology co-management consultation
I want to name the outdated assumption directly before we talk about his specific case: advanced liver disease used to be treated as a kind of natural anticoagulation, protective against clot. That framework has been overturned — cirrhosis is now understood as a genuinely prothrombotic state on balance, and current ISTH guidance recommends standard-dose DOACs in Child-Pugh A or B cirrhosis with AF, following the same recommendations as patients without liver disease.
His CHA2DS2-VASc of 3 would indicate anticoagulation in anyone, and I agree the liver disease itself shouldn't be read as protective. My question is specifically about his varices — small, never bled, on prophylaxis — and whether that changes anything about drug choice or monitoring intensity even if it doesn't change the decision to anticoagulate at all.
It changes monitoring, not the decision. Apixaban standard dose, given his preserved renal function and the class's generally favorable GI-bleeding profile compared to older agents; continued non-selective beta-blocker for variceal prophylaxis unchanged, and closer surveillance endoscopy scheduling going forward rather than deferring it now that he's also on an anticoagulant.
Apixaban 5 mg twice daily started. Non-selective beta-blocker continued unchanged for variceal prophylaxis. Surveillance endoscopy interval shortened given the added anticoagulation.
Agreed without real disagreement once the outdated 'auto-anticoagulated' framing was set aside explicitly; the standing note for his ongoing care:
Any future decompensation moving him from Child-Pugh B toward C would need to trigger a real reassessment of this decision, since current guidance's confidence specifically covers A and B and becomes genuinely uncertain beyond that point.