The Hallway Consult — Issue #12: DOAC Bleeding, Agitation, and Pediatric DKA Fluids

GI Bleed on DOACs: Not Every Patient Needs Reversal — And When You Do Reverse, Know What You're Actually Buying

Factor Xa inhibitors are everywhere now, and GI bleed is the most common site of major bleeding on a DOAC. That means this is not a rare toxicology-style edge case. It is routine emergency medicine. The hard part is that the ED question is not just "can I reverse the drug?" but "does reversal meaningfully help this specific patient?"

A 2024 Circulation subanalysis of the ANNEXA-4 cohort looked specifically at major GI bleeding in patients taking apixaban or rivaroxaban. The safety population was 109 patients, with 74 evaluable for hemostatic efficacy. Excellent or good hemostasis was reported in 82.4% of those evaluable patients. That is encouraging, but it comes with two important limits: there was no control group, and 11.9% of patients still died within 30 days. In other words, this is a signal that specific reversal can work, not proof that it improves the outcomes we care about most versus usual care.

Most DOAC-related GI bleeds are still won or lost with resuscitation, source control, and the patient's underlying physiology. Endoscopy remains the main intervention. What is still missing is randomized GI-bleed evidence showing that andexanet alfa improves transfusion needs, length of stay, or mortality. So the right read of the current literature is narrower: andexanet is a reasonable tool for severe factor Xa-associated GI bleeding, but it is not a mandate to reverse every patient just because apixaban or rivaroxaban is on the med list.

The ED translation is practical. If the patient is hemodynamically unstable, actively crashing through transfusion and resuscitation, or not getting to hemostasis, andexanet alfa is a defensible move if it is available. If it is not available, or cost and logistics make it unrealistic, 4-factor PCC remains the pragmatic backup many departments will actually use. But if the patient is stable, improving, and moving toward endoscopy, you do not need to let the presence of a DOAC hijack the case. Not every GI bleeder on a factor Xa inhibitor needs reversal.

Bottom line: In the 2024 ANNEXA-4 GI bleed subanalysis, andexanet alfa achieved excellent or good hemostasis in 82.4% of evaluable patients, but this was single-arm data with no control group and a 30-day mortality rate of 11.9% - so reversal should be reserved for truly severe or ongoing hemorrhage, not treated as an automatic response to every DOAC-associated GI bleed.

Eikelboom JW, et al. The Efficacy and Safety of Andexanet Alfa in Patients With Acute Gastrointestinal Bleeding While Taking Factor Xa Inhibitors: An ANNEXA-4 Subanalysis. Circulation. 2024;149(16):1315-1318.

Undifferentiated Agitation: Droperidol Is Back, and the Dose Conversation Has Changed

The undifferentiated agitated patient is one of the higher-stakes situations in emergency medicine. The person is dangerous to themselves, to staff, and sometimes to a clinical condition that will kill them if it goes unrecognized because they won't hold still. Whatever you give them needs to work fast.

Droperidol had an odd recent history in the United States: strong clinician loyalty, then years of limited availability, then a quiet return. The more important update is that the American College of Emergency Physicians stopped being coy about where it belongs. In its 2024 clinical policy on severe agitation, ACEP recommends droperidol plus midazolam for more rapid and effective control, and if you are going to use a single agent after verbal de-escalation fails, droperidol is the preferred choice. That is a meaningful shift in how the ED pharmacology conversation is framed.

The dose question matters too. For years, the reflex was simple: more agitation, more milligrams. But a 2025 ED cohort study comparing 5 mg versus 10 mg as the initial droperidol dose found that the 5 mg group needed rescue sedation less often. That does not mean 5 mg is magically stronger. It probably reflects, at least in part, that clinicians selected lower doses for less severe agitation. Still, it undercuts the idea that 10 mg should be the universal opening move.

The practical translation is straightforward. For most ED agitation presentations, 5 mg of droperidol is a reasonable starting point, with midazolam ready if speed or severity demands combination therapy. Intramuscular administration remains one of the biggest operational advantages when IV access is unrealistic. For the physically combative patient who is becoming dangerous, the literature increasingly supports moving to the combination early instead of waiting to see whether a single agent is enough.

Ketamine still has a role, especially in the truly explosive patient where immediate dissociation is the safest option. But if the goal is fast control without buying an airway, current expert guidance puts droperidol with midazolam in the lead position. The useful mindset change is this: dose to the patient in front of you, not to your own adrenaline. You can always escalate. You do not need to overshoot every case at minute one.

Bottom line: The 2024 ACEP clinical policy pushes droperidol plus midazolam to the front of the ED agitation algorithm, and newer comparative dosing data do not support reflexively starting every severely agitated patient at 10 mg of droperidol when 5 mg will often get you where you need to go.

American College of Emergency Physicians Clinical Policy Subcommittee on Severe Agitation. Clinical Policy: Critical Issues in the Evaluation and Management of Adult Out-of-Hospital or Emergency Department Patients Presenting With Severe Agitation. Ann Emerg Med. 2024;83:e1-30. // Cole JB, Glass S, Martel M, et al. Rescue Sedation after 5 mg or 10 mg of Droperidol as the Initial Treatment for Acute Agitation in the Emergency Department. J Emerg Med. 2025.

Pediatric DKA Fluids: Normal Saline Is No Longer the Unquestioned Default

Pediatric DKA has always had one fluid that felt automatic: normal saline. It is familiar, widely stocked, and built into years of protocols. The problem is that saline's chloride load was never harmless, especially in a disease state where metabolic acidosis is already doing enough damage on its own.

A 2025 double-blind randomized controlled trial in BMJ Open Diabetes Research & Care compared 0.9% saline with Ringer's lactate as the initial fluid in children with diabetic ketoacidosis. The practical finding was the one emergency physicians actually care about: Ringer's lactate was associated with earlier DKA resolution and less hyperchloremia. That is exactly the physiologic argument balanced-crystalloid advocates have been making for years, but this time it showed up in a pediatric DKA trial rather than in a general ICU fluid conversation.

That does not mean saline suddenly became malpractice. The bigger cerebral-edema trials already taught us that the dangerous myths around DKA fluids were overstated, and most kids will still do fine if saline is what gets hung first. But if two reasonable crystalloids are sitting in front of you and one of them appears to shorten the acidotic course while reducing iatrogenic hyperchloremia, that should matter. The old answer of "saline because that's what we've always used" is getting thinner.

The ED takeaway is not to obsess over decimal points in the bicarbonate. It is to choose the fluid that is least likely to worsen the chemistry you are trying to fix. In a child with DKA who needs volume expansion and ongoing deficit replacement, Ringer's lactate is now a very defensible first choice, especially if you are trying to avoid stacking hyperchloremic acidosis on top of ketoacidosis.

This is probably where practice is headed: away from blind saline loyalty and toward balanced crystalloids when there is a clear physiologic upside and no obvious safety penalty. Not every protocol has caught up yet. The literature has.

Bottom line: A 2025 double-blind randomized trial found that Ringer's lactate was associated with earlier DKA resolution and less hyperchloremia than normal saline in children, which makes the old reflexive saline-first approach harder to defend when a balanced crystalloid is available.

Agarwal A, Jayashree M, Nallasamy K, Dayal D, Attri SV. 0.9% Saline versus Ringer's lactate as initial fluid in children with diabetic ketoacidosis: a double-blind randomized controlled trial. BMJ Open Diabetes Research & Care. 2025;13(2):e004623.

The Hallway Consult is built for EM clinicians who want the useful version of the literature. Forward it to a colleague if it helped.

www.hallwaymedicine.com

— The Hallway Consult team