Holding Apixaban for a Colonoscopy: Why Bridging Isn't the Question It Used to Be
High-risk AF, a scheduled polypectomy, and an instinct — carried over from the warfarin era — to protect a high-risk patient with a bridging anticoagulant. The pharmacology of the drug he's actually on argues that instinct doesn't transfer.
M.F., a 74-year-old man, has coordinated his neighborhood's community garden for eleven years — organizing plot assignments, running the spring seed swap, mediating the occasional dispute over whose tomatoes are shading whose peppers. He has AF with a genuinely elevated thromboembolic risk profile, including a TIA two years ago, and has been on apixaban 5mg twice daily since. He also has mild-to- moderate CKD, eGFR 51. A screening colonoscopy last year found a polyp large enough to warrant removal rather than simple biopsy, and that procedure is now scheduled — a moderate-to-high bleeding-risk procedure by endoscopic classification, not a routine surveillance look. The question the team is working through is how to manage his apixaban around it: how long to hold it beforehand, and whether any bridging anticoagulant belongs in the plan given how much his thromboembolic risk exceeds an average AF patient's.
Periprocedural bridging with a low-molecular-weight heparin was standard practice for high-risk warfarin patients for years, built around a real pharmacologic gap: warfarin's anticoagulant effect takes days to establish and days to reverse, leaving genuine periods of under- or over-anticoagulation around an interruption that a fast-acting bridge could smooth over. The BRIDGE trial tested that practice directly in warfarin patients and found bridging did not reduce thromboembolism and significantly increased major bleeding — though its own population, with a mean CHADS2 score of 2.3, underrepresented patients with very high risk scores, and excluded mechanical heart valves entirely. Apixaban's pharmacokinetics are fundamentally different from warfarin's: its onset and offset are measured in hours, not days, so there's no equivalent gap in protection during a brief interruption for a bridge to actually cover. About a quarter of an apixaban dose is cleared renally — far less than dabigatran's roughly 80%, but enough that M.F.'s reduced kidney function lengthens its elimination somewhat beyond what a patient with normal renal function would experience. Resumption timing runs the other way from interruption timing: for a procedure carrying real bleeding risk, standard practice is to wait 48 to 72 hours and confirm hemostasis before restarting, not to resume the next day.
Pre-procedure anticoagulation planning
I'd hold apixaban for a standard interval based on his renal function and the procedure's bleeding risk, resume it once hemostasis is confirmed, and skip bridging entirely. Apixaban doesn't have warfarin's slow onset — there's no equivalent gap in protection during resumption for a bridge to actually cover.
This isn't a claim that his elevated risk doesn't matter — it's that bridging doesn't address a gap that exists with this specific drug.
Given his TIA history and elevated risk score, I'd want bridging genuinely considered, not dismissed out of hand — that profile is exactly what bridging was historically meant to protect.
Checking BRIDGE's actual population as I say this — that trial was warfarin patients, not DOAC patients, and even within it, high-CHADS2 patients like him were underrepresented. Combined with apixaban's fast on/off pharmacokinetics not creating the gap bridging exists to cover in the first place, I don't think bridging is actually the right lever here for his risk.
Separate from the bridging question, I'd extend the standard hold interval for him specifically. His eGFR of 51 means apixaban's elimination half-life runs longer than in a patient with normal renal function — the usual interval assumes a clearance rate he doesn't have, which matters directly for a moderate-to-high bleeding-risk procedure like a polypectomy.
Agreed: hold apixaban 3 days before the colonoscopy given his renal function, proceed without bridging, and resume 48 to 72 hours afterward once hemostasis is confirmed, barring any bleeding complication.
Not agreed, and carried forward explicitly rather than smoothed over:
All three voices agreed this would prompt a genuine reassessment of periprocedural management for his risk category going forward — not proof today's reasoning was wrong, but a real reason to revisit it.
Was treated as the actual basis for ruling out bridging — not because his underlying risk concern was dismissed, but because the specific tool he first reached for turned out not to fit the specific drug M.F. is on.
M.F.'s elevated thromboembolic risk was never in dispute — what the debate actually resolved was that bridging, a warfarin-era tool, doesn't address a gap that exists with his current anticoagulant.