Welcome to PICU Doc On Call, A Podcast Dedicated to Current and Aspiring Intensivists.

I’m Pradip Kamat

I’m Rahul Damania, a third-year PICU fellow.

I’m Kate Phelps, a second-year PICU fellow and we are all coming to you from Children’s Healthcare of Atlanta – Emory University School of Medicine, joining Pradip and Rahul today. Welcome to our episode, where will be discussing gastrointestinal bleeding.

Kate: Let’s start with a case:

A 4-year-old, previously healthy male presents to the emergency room after a large, bloody stool at home. He notably had an episode of dark emesis and an episode of blood-tinged emesis on the day prior. In triage, he is altered and unable to answer questions coherently. Initial vital signs are temperature 36.1 C, RR 24, HR 146, BP 110/54. Point-of-care labs show hemoglobin to be 5.1 with hematocrit 15. His venous blood gas is reassuring against respiratory disease, and he is in no respiratory distress. Further labs are sent and a massive transfusion protocol is initiated before transfer to the PICU. Before arrival in the PICU, he receives two aliquots of RBCs, 1 aliquot of FFP, and 1 aliquot of platelets. Additional labs are sent from the PICU, post-transfusion. His post-transfusion hemoglobin is 8.8. Other labs are notable for normal MCV, elevated total bilirubin to 4.1 (with direct component 3.4), and elevated AST and ALT to 309 and 495 respectively.

Rahul: To summarize key elements from this case, this patient has:

  • An undifferentiated gastrointestinal bleed with both hematemesis and hematochezia.
  • He has symptomatic anemia, as evidenced by tachycardia
  • Altered mental status.
  • He is initially stabilized via transfusion of several blood products and liver function labs are shown to be very abnormal — which we will get more into later!

PK: Let’s get into important parts of the history and physical. Kate, can you tell me what some key history items in this patient are — and what are some areas to make sure to touch on when a patient has a GI bleed?

Kate: Yeah! I’d love to.

First – in our patient, some important elements are his rather acute onset. His parents mention he has had one day of bleeding symptoms – first with emesis yesterday, with components of old, partially digested blood, as well as some fresh blood. Second, he has a frankly bloody stool at home. Given his clinical instability, history taking was probably limited at first, so it’s important to ask follow-up questions and really dig into the case after stabilization!

I like to put my questions about gastrointestinal bleeding into buckets based on the questions I need to answer. I need to answer: is this active bleeding or old blood? Is this slow, insidious bleeding or fast, life-threatening bleeding? Is this an upper GI bleed or a lower GI bleed? Bright red blood in emesis tells us that bleeding is active, whereas coffee-ground or dark emesis tells us that, while recent, the blood has been partially digested in the stomach and may not be ongoing. Similarly, melena (dark, tarry stool), tells us blood has come through the colon. While coffee-ground emesis and melena don’t rule out an active bleed, they do tell us the bleeding may be slower, as large volume, active bleedy is irritating to the stomach and gastrointestinal tracks and moves through the system quickly.

The next question I want to answer is: what is the cause of this bleed? Easy bruising, petechiae and mucosal bleeding may point to a coagulation disorder. Abdominal cramping, frequent stooling, and weight loss may point to inflammatory bowel disease. Past medical history, family history, and a thorough review of systems are key here.

Rahul: Yeah, that’s great! Let’s talk about your question of upper GI vs lower GI bleed.

First, a definition: an upper GI bleed is bleeding that occurs above the ligament of Treitz — which is ligamentous tissue that supports the end of the duodenum and beginning of the jejunum at their junction. While not 100% specific, some symptoms that point to an upper GI bleed are: hematemesis, coffee-ground or dark emesis, and melena. Symptoms that lend themselves to the diagnosis of a lower GI bleed are hematochezia (bright red blood in the stool) and melena (which may represent a more bleed more proximal to Treitz). However, with a brisk, heavy upper GI bleed — say from the duodenum — patients can also have hematochezia.

OK to summarize, when we think of GI bleeding, first stratify your patient into slow vs. fast bleeding, identify whether it is upper or lower GI bleeding, and dive deeper into an underlying cause after your patient is stabilized.

Pradip: Relatively little data exists about the prevalence of GI bleeds in the PICU. In a study by Chaibou, et al., they reported that approx 10% of PICU children have upper GI bleeding with only 1/5 of those with UGIB having clinically relevant bleeding (characterized by significant hemoglobin drop, need for transfusion, hypotension, multi-organ failure, or death). Incidence of lower GI bleeding is even less well characterized in current available evidence.

Kate: Thanks, Pradip. Given our patient’s symptoms, I would be most concerned for an upper GI bleed, given the bloody emesis — but a significant one if it’s leading to hematochezia.

Rahul: Yeah, that’s exactly what I was thinking, KP. Pradip, in the literature we see they mention that NG saline lavage can be used diagnostically to help confirm if bleeding is occurring in the upper GI tract vs a pulmonary source. Further, NG lavage has been advocated as a therapeutic practices, however, this may be outdated now as we push for more timely endoscopy. In fact, studies show: ice water lavage is not recommended; this older practice does not slow bleeding and may induce iatrogenic hypothermia, particularly in infants and small children.

Kate: Ok, let’s back up for a second — let’s talk about red flag symptoms! ABCs should always come first for every patient who arrives anywhere in the hospital. In this patient, concerning symptoms in this scenario, are his tachycardia and his altered mental status. These symptoms tell us that anemia is symptomatic and likely more acute. Hypotension and tachycardia indicate that bleeding is significant enough to cause hypovolemia. Altered mental status indicates that the brain is hypoxic, in this case, due to inadequate hemoglobin. Other red flags symptoms in GI bleeding include: orthostatic changes, delayed capillary refill and other signs of poor perfusion, currant jelly stools (which may indicate bowel ischemia), and of course anything that points to a large volume of blood in emesis or stool (for example, “the whole toilet bowl was red”) — as these may precede hypotension. Rahul will fill us in later about how to treat patients with red flag symptoms!

Absolutely, the identification of hypovolemic shock is essential in GI bleeding. Notice subtle data trends and optimize O2 delivery. Please check out our prior episode entitled Oxygen Content & Delivery.

Pradip: To switch gears, tell me how you think about the differential in patients with bleeding?

Kate: Sure, the differential will be different for upper vs lower but will also be relevant to the age of the patient. The differential for clinically relevant GI bleeding in an infant includes hemorrhagic disease of the newborn (in those who did not receive Vit K at birth), necrotizing enterocolitis, and Hirschprung’s enterocolitis (which interestingly can occur after repair), and volvulus. For children >1 year, the differential includes esophageal varices, gastric or duodenal ulcers, volvulus, intussusception, Meckel’s diverticulum, Mallory Weiss tears, IgA vasculitis, hemolytic uremic syndrome, and several infectious etiologies. Adolescents and young adults have a similar differential but now we begin to think more about inflammatory bowel disease and NSAIDs. Of course, there is a lot of overlap between school-age children and adolescents. In the oncology population, we have to think about graft-versus-host disease and typhlitis.

Rahul: So really — the differential is broad. Let’s talk about initial and ongoing work up to narrow our differential.

Initial labs should include a complete blood count, a comprehensive metabolic panel with a fractionated bilirubin, coagulation studies, and — perhaps most importantly — a type & screen! Initial imaging might include a two-view abdominal X-ray to evaluate for obstruction or perforation. Ultrasound can help rule in intussusception. Later imaging might include CT with angiography or even MRI.

Remember when it comes to liver function tests: alkaline phosphatase and GGT give us info about the biliary ducts, AST and ALT tell us about hepatocellular function, and albumin and PT/INR give us info about hepatic synthetic function.

Pradip: Great — now let’s get into treatment.

Rahul: As Kate eluded to earlier, if any red flag symptoms are present, we need to think about resuscitation and stabilization. Initial stabilization for patients should include attention to the airway, breathing, and circulation. For serious upper GI bleeds, intubation should be considered for repeated bloody emesis, to control the airway and prevent aspiration. Hypotension can be initially managed with judicious fluid resuscitation to temporize but should be followed by blood products as soon as possible. Most hospital centers have a massive transfusion protocol, so consider this in hemorrhage states before you have signs of end-organ hypo-perfusion! Kate, can you touch on additional specific treatment for ongoing bleeding?

Kate: Yeah – we really have two avenues for intervention: medical and surgical. Medical treatment can be tailored to the etiology but can include an IV proton pump inhibitor (or PPI) as first like during workup, followed by an octreotide infusion. Rahul, can you tell us about octreotide before I continue?


  • Octreotide is a Long-acting somatostatin-analog: that reduces splanchnic blood flow and inhibits gastric acid secretion.
  • Dosing: an initial bonus of 1 mcg/kg followed by a maximum infusion of 10 mcg/kg/hr, which can be titrated down as bleeding improves and resolves.
  • Side effect: hyperglycemia as we inhibit the effects of insulin.

Kate: Perfect- thanks! An additional medication sometimes used in GI bleeding is vasopressin, though octreotide has been shown to be as efficacious and does not carry the same daunting side effect profile. Most management strategies have shifted to using octreotide over vasopressin. If intermittent PPI dosing plus octreotide doesn’t control bleeding, a continuous infusion of a proton pump inhibitor can be considered though no data has shown this.

Let’s summarize the medical therapies, PPI, octreotide, and in some cases vasopressin.

Pradip: For surgical intervention, we’re first talking about upper endoscopy (esophagogastroduodenoscopy) or a colonoscopy — which can be both diagnostic and therapeutic. Endoscopy should ideally occur after hemodynamic stabilization but within 12 hours of admission for variceal bleeding and within 24 hours of admission for non-variceal bleeding in the case of upper gastrointestinal sources. Endoscopic interventions may include: adhesive cyanoacrylate applied to the bleeding lesion, band ligation applied to varix, injection sclerotherapy, and epinephrine injection, among other things. Interventional radiology may be able to perform arterial embolization.

Kate: I think this is the perfect point to follow up with our case, initial labs point toward normocytic anemia biliary duct obstruction without coagulopathy. During the hospital admission, bleeding stabilized after the initial massive transfusion. EGD showed acute clot formation near the ampulla of Vater in the duodenum. The eventual MRI showed a choledochal cyst with arterial erosion leading to the acute hemorrhage. An angiogram and percutaneous biliary drain placement were accomplished with IR.

OK Kate, do you mind summarizing our takeaways for today?

Kate: Key objective takeaways:

  1. Clinical relevant GI bleeds are uncommon in the PICU, but the skills to stabilize are crucial in the setting of a life-threatening hemorrhage.
  2. The differential for a GI bleed is broad but can be narrowed through careful and thorough history taking, physical examination, laboratory data, and imaging.
  3. Endoscopy should occur in a timely fashion in the setting of clinically significant upper GI bleeding.
  • For more reading, information can be found in:
  • Pediatrics in Review, “Gastrointestinal Bleeds” by Baker, et. al in the October 2021 edition.
  • Chapter 95 of the most recent edition of Fuhrman & Zimmerman’s Pediatric Critical Care, with sections on many of the differential diagnoses, included today.

This concludes our episode on GI hemorrhage. We hope you found value in our short, case-based podcast. We welcome you to share your feedback, subscribe & place a review on our podcast! Please visit our website picudoconcall.org which showcases our episodes as well as our Doc on Call management cards. PICU Doc on Call is hosted by myself Dr. Pradip Kamat, and my dream cohosts Dr. Rahul Damania and Dr. Kate Phelps. Stay tuned for our next episode! Thank you!