You are called urgently to resuscitation. Walking into the room, you recognize a familiar face. In fact, didn’t you discharge him home yesterday with epigastric pain that you attributed to his chronic alcohol use? Today’s diagnosis is not a mystery. There is already a puddle of blood on the floor, and he is quickly filling another emesis basin with bright red blood. A glance at the monitor reveals a heart rate of 135 and a blood pressure of 74/34. We need an approach to the resuscitation of massive GI bleeds…
My approach to the massive GI bleed
Get on personal protective equipment. Get your eyes covered. No matter how sick the patient appears, there is no excuse for getting blood in your eyes.
Call for help. We can resuscitate, replacing lost blood, but to turn the tides the bleeding needs to be stopped. Definitive control of this patient’s bleeding will require an endoscopist, an interventional radiologist, or a surgeon. Make sure everyone is aware early.
Be an emerg doc. Do all the stuff you can do blindfolded, like starting multiple large bore IVs, applying oxygen, and getting the monitors on. (Remind yourself to avoid the classic rookie mistake: make sure the blood pressure cuff is set to cycle frequently.)
Give blood. Don’t waste time on crystalloid fluid boluses. The patient is losing blood, they need blood replaced. Start with type O, and change to type specific as soon as possible. I won’t get into the debate about different transfusion ratios here. Just use whatever protocol your institution has, give lots of blood, and make sure you have more ready when needed. (Dworzynski 2012) Almost everyone has heard that we should be transfusing less and aiming for a lower hemoglobin target in GI bleeds. (Villanueva 2013) However, patients were excluded from that study if they had “massive exsanguinating bleeding” – so ignore the evidence here, and just get blood on board. I am not targeting a specific hemoglobin number in these sick patients. The lab results are too far behind. I am targeting vital signs and clinical tissue perfusion. That being said, I am not trying to achieve a normal blood pressure. Much like trauma, I don’t want to put pressure on any newly formed clots. Permissive hypotension is the plan, until definitive control of bleeding is achieved with endoscopy, interventional radiology, or surgery.
Reverse any known or potential coagulopathy. Although I am not certain that a balanced transfusion is necessary, GI bleed patients will often have an elevated INR, and early empiric administration of FFP with the packed cells definitely makes some sense. If the patient is on warfarin, I would give the 4 factor prothrombin complex concentrate complex that we stock immediately. (Dworzynski 2012) If the patient is on one of the novel anticoagulants – well, they do wear off eventually. For renal failure patients, a dose of DDAVP (0.4mcg/kg IV) is reasonable.
Update: In the initial version of this post, I suggested giving TXA, based on the evidence that was available at the time. However, the HALT-IT trial has since been published, and it is pretty clear that we shouldn’t be using TXA in these patients.
Control the airway. The classic approach to controlling the airway in critically ill patients has been immediate rapid sequence intubation. However, not taking the time to address hypotension, possible acidosis, and inadequate oxygenation prior to intubation puts the patient at higher risk of peri-intubation complications, including cardiac arrest. (Mosier 2015, Perbet 2015) Therefore, I am a huge fan of the concept of “resuscitation sequenced intubation” that I believe was coined by Dr. Richard Levitan. Resuscitate before you intubate. That is why, contrary to the classic ABCs, massive transfusion is started before airway control.
- You were supposed to put on PPE at the very beginning of this resuscitation. Why don’t you have it on? Cover up now!
- The goal should be perfect preoxygenation and apneic oxygenation. We know this airway is going to be a challenge, with blood potentially impairing our view and classic backups like video laryngoscopy being less helpful, so it is important to have as much oxygen on board as possible to give us time. Practically, this means putting nasal prongs on under a non-rebreather set at flush rate. Try to avoid positive pressure ventilation or bag valve mask use prior to intubation, as you don’t want to distend the stomach.
- Place an NG tube to empty the stomach (trying to limit regurgitation and aspiration). This step is controversial, but makes sense to me.
- If there is time, give metoclopramide 10mg IV. (It may increase stomach emptying and also may increase lower esophageal sphincter tone. I know that’s a lot of “mays”, but that’s all we get with these critically ill patients.)
- Position the patient and intubate with the head of the bed up at 45 degrees (again, trying to keep blood in stomach). However, if the patient starts vomiting during intubation, Trendelenburg positioning makes sense to try to limit aspiration of stomach contents.
- Make sure you have 2 working suctions (at least). Consider a set-up with a meconium aspirator attached directly to an endotracheal tube for suction. (Weingart 2011)
- I would perform a standard rapid sequence intubation, but with adjustments to the standard induction medication doses. This will be a future post here, but for now I would refer to EMCrit’s excellent post on how to intubate a patient with hypotension. Practically, this means I am using 0.5mg/kg of ketamine and 1.5mg/kg of rocuronium as my initial dosing.
- Have all your backup devices like an LMA, bougies, video laryngoscope, and a scalpel out and ready.
- There are a number of techniques that might help with epiglotoscopy and laryngoscopy, such as SALAD: “suction assisted laryngoscopy airway decontamination”
- Be prepared for a BP drop with intubation. The packed cells should already be flowing before you start. Ideally, I would have a pressor on a pump ready to start. More likely, I would be ready with a push dose pressor (covered in this anaphylaxis post).
At this point an NG or OG tube is an essential piece of equipment. I don’t use it as a diagnostic tool. I use it as a monitor. Even before the vital signs change, you will know the patient has started to rebleed if the suction fills with bright red blood. I don’t worry about passing an NG in patients with varices. (Lopez-Torres 1973 and Ritter 1988)
As an aside, forget the classic question: Is it upper or lower? Ignore it. Really, just forget it. If you just assume the source is upper in any GI bleed patient who is hemodynamically unstable, you will be in good shape. Upper bleeds are more common and more deadly. The classic NG is not helpful diagnostically. (Witting 2004) And if you are wrong? It doesn’t matter, because the massive lower GI bleed patient needs the exact same treatment: resuscitation and then definitive control of their bleeding with interventional radiology, surgery, or endoscopy. However, a quick assessment to make sure you aren’t dealing with massive hemoptysis or epistaxis is reasonable.
One important question: Does the patient have a history of aortic disease? An aortoenteric fistula is a rare but incredibly deadly cause of GI bleed. Unlike other etiologies that are generally best served by endoscopy or interventional radiology, the patient with an aortoenteric fistula needs a surgeon as soon as possible. (Goralnick 2014)
Get direct control of the bleeding. After getting the resuscitation started and controlling the airway, it is essential to get the patient somewhere to definitely control the bleeding. Variceal bleeding, in particular, rarely stops on its own and needs a rapid intervention. The first step is usually endoscopy, but with massive hemorrhage, the endoscopist may just be faced with a sea of red. If endoscopy fails, interventional radiology can place a TIPS (transjugular intrahepatic portosystemic shunt) in the case of variceal bleeding, or potentially embolize the bleeding artery. (Ramaswamy 2014, Orloff 2009) The best plan is to have all three services on board from the beginning, working together to ensure the patient ends up in the right place.
Is there a chance this is a variceal bleed secondary to cirrhosis? If so, give an empiric prophylactic dose of antibiotics (cetriaxone 1gram IV will do). It decreases mortality, as well as rebleeding and infections. (Chavez-Tapia 2010)
Depending on your local resources, you may be required to temporize the patient until help arrives, during transfer, or if one of the definitive management strategies fails. The final, temporizing step in the dying patient with a GI bleed is balloon tamponade, which is covered in this accompanying post.
You will notice that proton pump inhibitors are not mentioned at all. There is no role for PPIs is the emergency management of GI bleeds. Every study looking at them has been negative, and one was stopped early for increased mortality. I review these studies in a little more depth in this post.
I also don’t routinely use octreotide or any somatostatin analogue. This might be more controversial, but seems to be supported by the evidence. According the the most recent Cochrane analysis (Gøtzsche 2008) there is no reduction in mortality, which is really what we are trying to accomplish. There may be a decrease in the amount of blood transfused by half a unit, but I would tend to agree with the authors that “it is doubtful whether this effect is worthwhile.” If specifically requested, or if I happen to have an open IV line that I don’t need to use for anything else (doubtful), I would consider starting is as an adjunct. The dose is 50mcg IV bolus followed by a 50mcg/hr drip. However, if I need the line for red blood cells, FFP, or really anything else, they will take priority. Terlipressin is an agent that people outside of North America probably know really well. We don’t have it in Canada, so I haven’t reviewed it. It apparently has stronger evidence than octreotide, but I have not ready the studies myself, so I am unable to comment.
Other FOAMed Resources
Chavez-Tapia NC, Barrientos-Gutierrez T, Tellez-Avila FI, Soares-Weiser K, Uribe M. Antibiotic prophylaxis for cirrhotic patients with upper gastrointestinal bleeding. The Cochrane database of systematic reviews. 2010. PMID: 20824832
Dworzynski K, Pollit V, Kelsey A, Higgins B, Palmer K, . Management of acute upper gastrointestinal bleeding: summary of NICE guidance. BMJ (Clinical research ed.). 344:e3412. 2012. PMID: 22695897
Goralnick E and Meguerichian DA. Chapter 30. Gastrointestinal bleeding. In: Marx JA et al. eds. Rosen’s Emergency Medicine, 8e. Philadelphia: Elsevier Saunders; 2014.
Gøtzsche PC, Hróbjartsson A. Somatostatin analogues for acute bleeding oesophageal varices. The Cochrane database of systematic reviews. 2008. PMID: 18677774
Lopez-Torres A, Waye JD. The safety of intubation in patients with esophageal varices. The American journal of digestive diseases. 18(12):1032-4. 1973. PMID: 4543409
Orloff MJ, Isenberg JI, Wheeler HO. Randomized trial of emergency endoscopic sclerotherapy versus emergency portacaval shunt for acutely bleeding esophageal varices in cirrhosis. Journal of the American College of Surgeons. 209(1):25-40. 2009. PMID: 19651060
Perbet S, De Jong A, Delmas J. Incidence of and risk factors for severe cardiovascular collapse after endotracheal intubation in the ICU: a multicenter observational study. Critical care (London, England). 19:257. 2015. PMID: 26084896 [free full text]
Ritter DM, Rettke SR, Hughes RW, Burritt MF, Sterioff S, Ilstrup DM. Placement of nasogastric tubes and esophageal stethoscopes in patients with documented esophageal varices. Anesthesia and analgesia. 67(3):283-5. 1988. PMID: 3278651
Weingart SD, Bhagwan SD. A novel set-up to allow suctioning during direct endotracheal and fiberoptic intubation. Journal of clinical anesthesia. 23(6):518-9. 2011. PMID: 21783351
Witting MD, Magder L, Heins AE, Mattu A, Granja CA, Baumgarten M. Usefulness and validity of diagnostic nasogastric aspiration in patients without hematemesis. Annals of emergency medicine. 43(4):525-32. 2004. PMID: 15039700
Morgenstern, J. Massive GI bleed management, First10EM, May 23, 2016. Available at: