Atrial Fibrillation: Same Diagnosis, Divergent Path
Two patients, both newly diagnosed with atrial fibrillation in the same week. One is a straightforward guideline case. For the other, a single medication already on the chart changes the entire calculus.
A 68-year-old man, a retired postal carrier with well-controlled hypertension and type 2 diabetes on lisinopril and metformin, presents to his primary care physician after several days of intermittent palpitations. An ECG confirms atrial fibrillation; he reports no prior episodes and has no documented history of the arrhythmia. Renal function has not yet returned but his baseline creatinine on record from six months ago was normal.
His hypertension and diabetes are both well controlled and neither complicates today's decision — no CYP3A4 or P-glycoprotein inducers or inhibitors on his medication list, no extremes of age or weight forcing a dose adjustment, nothing to second-guess once the score clears the threshold. Once his renal function returns and confirms what his six-month-old baseline already suggests, starting a DOAC is close to a formality.
A CHA₂DS₂-VASc of 3 clears the threshold for oral anticoagulation without much room for debate — in a man, a score of 2 or more is already a Class I indication, and we're a point past that. The question here isn't whether to anticoagulate. It's simply which agent, and at what dose.
I'd start with a direct oral anticoagulant rather than warfarin. The major trials — and ARISTOTLE in particular for apixaban — showed at least comparable stroke prevention with meaningfully lower rates of major bleeding and intracranial hemorrhage compared to warfarin, without the burden of INR monitoring. Nothing about this man's presentation argues against it.
Agreed, and the dosing question resolves itself once you check it against the criteria rather than assume it. Apixaban's dose reduction to 2.5 mg twice daily requires meeting at least two of three: age 80 or older, weight 60 kg or under, or serum creatinine 1.5 mg/dL or higher. He meets none of the three, let alone two. Full-dose apixaban, 5 mg twice daily, is the correct starting point — nothing in his medication list induces or inhibits the P-gp or CYP3A4 pathways apixaban depends on for clearance.
No real disagreement among the three of them. The primary care physician confirms the indication, the cardiologist confirms the agent, the pharmacologist confirms the dose — each checking a different part of the same decision rather than debating it. He is started on apixaban 5 mg twice daily, with routine follow-up once renal function results return.
A 74-year-old woman, a retired seamstress with hypertension and peripheral vascular disease, has been on a multidrug regimen for chronic osteomyelitis of the tibia for several weeks already — the infection followed a minor foot ulcer that never fully healed, an unsurprising course given her vascular disease. Her regimen is anchored by rifampin — chosen specifically for its bone and biofilm penetration, a property few other antibiotics share — with an expected total course measured in months rather than weeks. New atrial fibrillation was found incidentally on routine monitoring during this same follow-up visit, not from any cardiac symptom she reported.
Her CHA₂DS₂-VASc score is higher than Patient A's, which would ordinarily make her anticoagulation decision the easier one — a stronger indication, running on the same guideline logic. What actually complicates it has nothing to do with her heart at all. Rifampin is not simply another medication on her list; it is one of the most potent inducers of the CYP3A4 and P-glycoprotein pathways every direct oral anticoagulant depends on for its plasma exposure, and her infection requires it for as long as her clinical response demands — months, not the days or weeks a drug interaction might otherwise be timed around. A decision that took five minutes in Patient A's case now depends on whether an entire drug class can be trusted to actually work in a body already primed to clear it faster than the label assumes.
Score of 4, clearly anticoagulate. My instinct is the same as it would be for anyone else her age — a DOAC, apixaban most likely, adjusted if needed for her weight.
Before anyone settles on an agent, I want to flag the rifampin, because it isn't incidental to this decision and it isn't going away soon. Whatever anticoagulant she goes on has to coexist with it for the duration.
Then a DOAC isn't a dosing question, it's an exclusion. Apixaban's AUC falls by roughly half with concurrent rifampin. Rivaroxaban's does too. Dabigatran is P-gp dependent for absorption alone and loses even more — exposure reductions upward of 65 percent have been reported. Edoxaban is affected the same way through P-gp induction, and its own labeling recommends against the combination. There's no dose-adjustment path out of this: we have no reliable, standardized way to measure a DOAC's anticoagulant effect and correct for it in real time.
Warfarin is the one agent left standing — not because it escapes the interaction, but because we can see it. INR gives us a number to titrate against, which is the one thing a DOAC can't offer her right now. I'd start at a standard low, weight-adjusted dose, check INR twice weekly at first, and expect to climb the dose over one to two weeks as induction fully establishes. And we need to flag now that the reverse will happen once her antibiotic course ends — warfarin requirements will fall as induction resolves, over roughly the same window.
I don't dispute the pharmacology, but I want us to be honest about what we're asking of her. We're proposing to add twice-weekly blood draws and a drug with a narrow therapeutic window in a patient whose weight alone already puts her in a higher bleeding-risk bracket than Patient A. That's not a reason to avoid anticoagulation — her stroke risk is real and higher than his. But it's a reason to ask whether warfarin is genuinely the only door open, or whether it's the only door open as long as the antibiotic regimen is fixed. I'd want infectious disease to tell us plainly whether a non-rifamycin alternative was ever seriously on the table for her.
Warfarin, not a DOAC, is where the group lands for now — on that point, cardiologist, infectious disease, and pharmacologist all agree once the mechanism is on the table. What doesn't get settled is the geriatrician's question: whether the antibiotic regimen itself was ever genuinely reconsidered, or whether it was treated as fixed the moment osteomyelitis therapy began. Neither side moves.