Acute Lower GI Bleeding in a Patient on Apixaban With a Recent Drug-Eluting Stent
A 76-year-old woman presents with 36 hours of maroon stools and one syncopal episode. Heart rate is 112, blood pressure 94/58 supine, hemoglobin 7.1 g/dL, platelets 210,000, INR 1.1. She takes apixaban 5 mg twice daily for atrial fibrillation and is on aspirin plus clopidogrel after a drug-eluting stent placed 7 months ago. Her last apixaban dose was 15 hours ago. She has COPD and a prior myocardial infarction. Nasogastric lavage is clear. She remains hypotensive after 2 L of crystalloid and 1 unit of packed red cells. The admitting team requests colonoscopy today, but the consultant says she cannot be prepped.
1. Rank the next steps and justify your choice of imaging.
Read the worked answer
Resuscitation comes first - large-bore access, blood products rather than crystalloid alone, correction of coagulopathy, and ICU-level monitoring. Hold apixaban and aspirin/clopidogrel pending hemostasis, and discuss reversal: for life-threatening hemorrhage on a factor Xa inhibitor, andexanet alfa or 4-factor PCC is reasonable, but her 7-month-old drug-eluting stent makes thrombosis risk real. That trade-off needs cardiology input and should be documented.
CT angiography is the right localizing test here. It detects bleeding at roughly 0.3–0.5 mL/min, is fast, does not require bowel preparation, and - critically - tells the interventionalist which vessel to target, converting a blind catheter search into a focused one.
Urgent colonoscopy is the wrong move in this patient: unprepped colonoscopy in an actively bleeding, hypotensive patient has low diagnostic yield, poor visualization, and a higher perforation risk. It becomes appropriate after resuscitation and prep in a stabilized patient.
A clear NG lavage lowers but does not exclude an upper source; keep that in the differential if CT is negative.
2. CT angiography shows active extravasation in a sigmoid branch of the inferior mesenteric artery. What are your technical options and their trade-offs?
Read the worked answer
Superselective microcatheter embolization is the procedure of choice. Advance a microcatheter beyond the marginal artery into the offending vasa recta and embolize as distally as possible.
Embolic agent:
- Coils - precise, visible, no risk of downstream particle migration; preferred for a discrete bleeding point.
- Liquid agents (NBCA, Onyx) or particles - useful for a bleeding field or a catheter position that will not allow coils; higher risk of inadvertent distal ischemia if not truly superselective.
Bowel ischemia is the main risk, generally quoted in the single digits with superselective technique, and rises sharply with nonselective embolization. Avoid proximal IMA or SMA embolization; if the bleeding cannot be isolated, that is a surgical conversation.
Vasopressin infusion is an alternative for diffuse small-vessel bleeding, but it requires ICU monitoring and carries a cardiac ischemia risk that is unattractive in a patient with COPD and prior MI - I would not choose it here.
If no extravasation is seen at angiography, provocative angiography (heparin, vasodilator, or tPA) can be considered before abandoning the procedure, and repeat CTA is reasonable if bleeding recurs.
3. When do you restart her antithrombotic therapy?
Read the worked answer
This is a shared decision, not a protocol lookup, and it should be framed as a three-way tension: rebleeding risk, stent thrombosis, and stroke risk from atrial fibrillation.
- Rebleeding risk is highest in the first 24–72 hours, especially with a coagulopathy and a large transfusion requirement.
- Stent thrombosis risk is highest in the first weeks to months after implantation; at 7 months, the hazard has fallen substantially, and a brief interruption is generally tolerated.
- Stroke risk from a few days off apixaban in atrial fibrillation is small but not zero.
A workable plan after confirmed angiographic hemostasis: resume aspirin within 24–72 hours, then restart clopidogrel or apixaban after roughly 3–7 days once hemoglobin is stable and there is no recurrent bleeding, with cardiology and GI agreement. If bleeding recurs, stop again rather than pushing through.
Document the reasoning, the specific restart dates, and who agreed - this is exactly the scenario where an undocumented decision looks like an oversight on review.