Endoscopy Practice and Sedation: Fasting and Periprocedural Antithrombotics

0:00Welcome to Board Pearls. This is episode one of four of the Endoscopy Practice and Sedation chapter, in the Endoscopic Procedures module. This episode is pre-procedure preparation and periprocedural antithrombotic management: fasting, the grid of procedure bleeding risk against patient thromboembolic risk, warfarin and DOAC holds, the narrowing of bridging, antiplatelet management after coronary stenting, and reversal in active bleeding.

0:26Start with the question the whole pre-procedure framework is built to answer: what in the patient could turn a routine sedation into an aspiration event or a thromboembolic catastrophe. Fasting and antithrombotic management are the two preventable causes behind the most common adverse events in endoscopy, and everything else about the morning of the procedure flows from those two answers. Fasting first, because it's the simplest mechanism and its exceptions tell you the most. The fasting rule is the two-two-six-eight numbers: clear liquids up to two hours before, breast milk up to four, infant formula or a light meal up to six, and a fatty or heavy meal up to eight. The rule applies to all sedation depths, from minimal sedation to general anesthesia, because aspiration risk doesn't respect the line between conscious sedation and deeper levels, since anyone sedated has impaired airway reflexes and gastric content can pass into an unprotected airway regardless of how much propofol it took to get there.

1:25The intervals aren't arbitrary, they're gastric emptying kinetics expressed as a clock: clear liquids leave the stomach within one to two hours, so a two-hour window empties them, while fats slow emptying through hormone-mediated antral relaxation, which is why a fatty meal needs eight hours and a light meal only six. The same physiology explains allowing clears up to two hours rather than enforcing nothing by mouth from midnight, because prolonged fasting just dehydrates the patient with no aspiration benefit since the clears were gone anyway.

1:57The exceptions are the teaching, because any condition that already delays gastric emptying makes the calendar interval insufficient, since the stomach the rule assumes is not the stomach the patient has: gastroparesis, achalasia, gastric outlet obstruction, retained contents on a GLP-1 receptor agonist, and pregnancy after the first trimester each hold content past the promised window, so each warrants full-stomach management with rapid sequence induction and a cuffed tube when sedation depth is required, or procedural delay when urgency permits.

2:27The GLP-1 exception is the new one and worth pausing on, because the practical guidance shifted recently. The class includes semaglutide, tirzepatide, liraglutide, and dulaglutide, and the mechanism is delayed gastric emptying through receptors on gastric smooth muscle plus central appetite suppression, so retained content can persist for hours to days after a normal fast. The earlier reflex was a uniform hold of the drug, but the current multi-society guidance is risk-stratified, with patients who have elevated-risk features getting a clear liquid diet for at least twenty-four hours before the procedure while continuing the drug, and holding reserved for shared-decision cases, where the elevated-risk features are active GI symptoms like nausea or distention, the dose-escalation phase when the receptor effect is climbing, and weekly dosing that sustains receptor activation.

3:20Pregnancy testing before elective endoscopy in women of reproductive age is institutional policy at most academic centers, because radiation-using procedures, sedation agents with incomplete teratogenicity data, and the procedural decision itself can all be modified by pregnancy status, and while pregnancy isn't an absolute contraindication when endoscopy is necessary, it changes the risk-benefit math and triggers fetal monitoring and obstetric coordination.

3:44Now the antithrombotic problem, which is structured, because every periprocedural anticoagulant or antiplatelet decision sits at the intersection of two things: the bleeding risk of the planned procedure and the thromboembolic risk of the patient. Holding the drug exposes the patient to thromboembolism, continuing it exposes them to procedural bleeding, and the grid tells you which exposure dominates.

4:09On the procedure side, endoscopic procedures are classified by whether the intervention produces a wound that exposes a clinically significant vessel: low-bleeding-risk procedures rarely do, like diagnostic endoscopy with biopsy, diagnostic colonoscopy with cold biopsy, endoscopic ultrasound without aspiration, stent exchange, capsule endoscopy, argon plasma coagulation, and Barrett ablation, while high-bleeding-risk procedures reliably do, like polypectomy of larger polyps, sphincterotomy, fine-needle aspiration of cysts, ampullectomy, mucosal resection, submucosal dissection, feeding-tube placement, variceal therapy, cystgastrostomy, and large-diameter dilation, and the mechanism that organizes the list is whether the injury can reach an artery or arteriole, so once you hold that, you stop needing to memorize the categories.

5:03On the patient side, thromboembolic risk for warfarin patients is stratified by the consequences of interrupting: high risk includes a mechanical mitral valve, an older-design mechanical aortic valve, a stroke or TIA within three months, a very high stroke-risk score, rheumatic valve disease, a prior clot on therapeutic warfarin, a recent venous thromboembolism within three months, and severe inherited thrombophilia; moderate risk includes a bileaflet mechanical aortic valve with one added risk factor, an intermediate stroke-risk score without prior stroke, a clot in the three-to-twelve-month window, recurrent clots, and active cancer; and low risk includes uncomplicated nonvalvular atrial fibrillation, a distant clot beyond a year, and a bileaflet mechanical aortic valve alone. The strata differ by an order of magnitude in the consequences of an event, because mechanical mitral valve thrombosis is essentially fatal while uncomplicated atrial fibrillation has an annual stroke risk a brief interruption barely changes, and that asymmetry is why bridging and continuation decisions look so different across strata.

5:52With the grid in place, warfarin management writes itself from the pharmacokinetics. Warfarin is held five days before a high-bleeding-risk procedure, with an INR checked the morning of and a target below one and a half, and the five-day interval reflects prothrombin, the vitamin-K-dependent factor with the longest half-life at roughly sixty hours, so five days is about two half-lives, which is what it takes to drop its activity into the range that supports hemostasis. The morning-of INR check isn't a formality, because the early INR drop is driven by a different factor with a four-to-six-hour half-life that corrects much faster, so an INR that looks reassuring on day one or two can mislead you into proceeding before prothrombin has actually decayed, and the morning-of value confirms the slowest factor has corrected.

6:37For low-bleeding-risk procedures warfarin is generally continued, because the bleeding doesn't justify the thromboembolic exposure of a hold, and resumption after a high-risk procedure is the same day for low-thromboembolic-risk patients and within one to two days for most others, gated by hemostatic confidence rather than an arbitrary minimum, with daily INR monitoring guiding re-entry, while the cirrhotic with a baseline elevated INR is its own framework because that INR reflects synthetic dysfunction rather than anticoagulant effect.

7:06DOACs are easier because their half-lives are short and their effects don't require monitoring: apixaban and rivaroxaban, with half-lives in the eight-to-seventeen-hour range, are held one to two days for low-bleeding-risk procedures and two to three days for high-bleeding-risk ones, and the predictable kinetics are themselves the assurance of clearance.

7:26Dabigatran is the case where pharmacokinetics meets comorbidity, because most of it is renally cleared, so reduced kidney function prolongs the effective half-life, and the standard high-risk hold is about two days at a creatinine clearance of fifty or above, four days from thirty to fifty, and longer below thirty, with roughly half those intervals for low-risk procedures, so the same calendar hold leaves different residual drug depending on kidney function. DOAC resumption after high-risk procedures is gated by hemostasis and tends to fall within one to three days, and the decision is binary because there's no titration-by-assay analog to the warfarin INR, so you resume the prescribed dose or you don't.

8:05Now bridging, where the grid shifted most dramatically in the last decade. Bridging means giving low-molecular-weight heparin during the warfarin hold so the patient is anticoagulated for as much of the period as possible, and for years it was routine, until a large randomized trial of atrial fibrillation patients on warfarin settled it: bridging produced no reduction in arterial thromboembolism, with nearly identical event rates, while it clearly increased major bleeding, and a second trial extending the question to mechanical valve and high-risk atrial fibrillation patients found no bridging non-inferior for thromboembolism in the moderate strata while bleeding favored no bridging.

8:41The mechanism is the same logic as the patient side: lower-risk patients have such low absolute event rates during a brief hold that bridging can't meaningfully reduce them, while bridging itself adds a substantial bleeding burden through residual heparin at the procedure and during resumption, so the math doesn't favor it.

8:58Bridging is now reserved for high-thromboembolic-risk patients only, meaning a mechanical mitral valve, a stroke or TIA within three months, a very high stroke-risk score, a prior clot on therapeutic warfarin, a recent venous thromboembolism, or severe thrombophilia, while lower-risk patients including uncomplicated atrial fibrillation, distant clots, and a bileaflet aortic valve alone don't benefit and incur excess bleeding, so the fellow who reflexively bridges every warfarin patient is doing quantified harm.

9:28DOACs don't require bridging at all, because their short half-life means the patient is off anticoagulation only for a brief one-to-four-day hold, intrinsically shorter than the warfarin hold plus resumption window, so there's no gap long enough to bridge. Aspirin and P2Y12 management doesn't involve bridging either, because aspirin is generally continued through procedures and the brief P2Y12 hold is covered by that continued aspirin.

9:54The antiplatelet framework runs on a single asymmetry: stent thrombosis after antiplatelet interruption carries a forty-to-sixty-percent mortality, while most procedural bleeding can be managed endoscopically with the patient returning to baseline, so the risk math isn't symmetric and the recommendations favor continuation whenever feasible. Aspirin for secondary cardiovascular prevention is generally continued through endoscopy, including diagnostic procedures with biopsy, polypectomy of small-to-moderate polyps, and most therapeutic procedures, because the absolute increase in bleeding from continued aspirin is small and usually endoscopically manageable, while interruption exposes the patient to acute coronary syndrome and stroke through unopposed platelet activation, and if aspirin must be held for a particular high-risk procedure it's resumed the same day hemostasis is confirmed.

10:43Aspirin for primary prevention doesn't earn the same continuation, because a patient with no established cardiovascular disease doesn't gain enough to justify the bleeding, and modern practice is to deprescribe primary-prevention aspirin in patients with significant GI bleeding history.

10:59The P2Y12 inhibitors are held five to seven days before high-risk procedures with aspirin continued throughout, and the interval isn't drug clearance, it's platelet turnover, because clopidogrel and prasugrel bind platelets irreversibly for the cell's lifetime so inhibition recovers only as new platelets are produced over seven to ten days, while ticagrelor is reversible with a shorter washout of three to five days, and the rule lets the procedure proceed with aspirin-mediated platelet function intact while the P2Y12 effect has cleared. Other agents follow their own holds: dipyridamole two to three days, cilostazol two days, ticlopidine ten to fourteen days, and vorapaxar five to thirteen days.

11:40The antiplatelet-after-stenting rule is where the asymmetric math is most acute: elective procedures are deferred entirely within the first thirty days after a bare-metal stent and within six months after a drug-eluting stent, with that six-month figure revised down from a year with newer stents, because endothelialization is incomplete in the early window and stent thrombosis from premature interruption carries the mortality that organizes the whole framework, so after the early window a P2Y12 hold for a high-risk procedure becomes acceptable as long as aspirin continues, while urgent procedures take precedence when clinically necessary because an actively bleeding ulcer can't wait the full window.

12:14When a procedure must proceed inside the early window, cangrelor offers a true bridge, an intravenous P2Y12 inhibitor with a half-life of three to six minutes, so the patient is transitioned from oral P2Y12 to cangrelor beforehand, the cangrelor is stopped for the brief procedure window, and oral therapy is restarted immediately after, with platelet function recovering within an hour of stopping the infusion, which makes it a real bridge in the way warfarin bridging fails to be, because the pharmacokinetics actually deliver continuous coverage to the moment of incision and restart promptly.

12:41That leaves reversal in active bleeding, tested by mechanism, each agent matched to its target. For life-threatening warfarin-related bleeding, four-factor prothrombin complex concentrate is preferred over fresh frozen plasma, and the reason is volume, because correcting an INR of three to four with plasma takes four to eight units, roughly one to two liters, which risks circulatory overload in a patient already being volume-resuscitated for shock, whereas the concentrate achieves comparable correction in about fifty milliliters, so the volume problem disappears, and vitamin K is added at five to ten milligrams intravenously for sustained correction beyond the half-life of the infused factors, though it's not the acute agent because vitamin-K-driven synthesis takes twelve to twenty-four hours.

13:25For dabigatran-related bleeding, the targeted agent is idarucizumab, a monoclonal antibody fragment whose affinity for dabigatran is far greater than thrombin's, so it sequesters circulating drug into a non-functional complex, dosed at five grams intravenously given as two infusions of two and a half grams each, with hemodialysis the alternative when it's unavailable because dabigatran is small and minimally protein-bound and therefore dialyzable, which is itself a board point since the other DOACs are not.

13:45For factor Xa inhibitor reversal the framework has just shifted, because the targeted decoy agent was recently withdrawn from the US market after postmarketing data showed increased thromboembolic events and thrombosis-related death, so four-factor prothrombin complex concentrate is now the practical reversal agent for life-threatening apixaban, rivaroxaban, and off-label edoxaban bleeding, and these agents are not dialyzable because they're more than eighty-five percent protein-bound. And aspirin and P2Y12 effect can't be pharmacologically reversed, because platelet transfusion has limited efficacy in the actively drug-bound patient since the circulating drug binds the transfused platelets too, so meaningful reversal requires waiting for platelet turnover.

14:27Pull it together and the framework is the grid plus the asymmetry. The grid sets every periprocedural decision at the intersection of procedure bleeding risk and patient thromboembolic risk, so only the intersection of high bleeding risk with high thromboembolic risk justifies bridging, and only low bleeding risk justifies continuation across any thromboembolic risk. The asymmetry organizes the antiplatelet decisions specifically, because the cost of an unrecognized stent thrombosis dwarfs the cost of endoscopically manageable bleeding, which is why aspirin stays on through most procedures, why DOAC holds scale to half-life and kidney function, why bridging narrowed dramatically, why elective procedures defer for thirty days after a bare-metal stent and six months after a drug-eluting one, and why reversal pairs idarucizumab with dabigatran and four-factor concentrate with everyone else now that the factor Xa decoy is gone from the US market.

15:23The next two episodes pick up where the patient is anticoagulated correctly and the stomach is empty enough to proceed, and turn to sedation itself. The first covers the depth levels and physical-status classification and the agents, propofol versus midazolam and fentanyl and the alternatives. The second covers the safety apparatus, capnography and the difficult airway, the reversal agents, and sedation in the cirrhotic, the elderly, and the patient with sleep apnea.

15:51For the full chapter, the practice vignettes, and the topic-tagged question bank, head to board pearls dot com. You'll find the rest of the series on Apple Podcasts, Spotify, or wherever you listen to podcasts. That brings us to the end of episode one of four of chapter twenty-eight, and I'll see you in the next one.