Upper Gastroin-Testinal Bleed
Kirsten Courtade, MD
Jonathan Bortinger, MD
The ED calls you for an admission. The patient is a 51-year-old woman who began vomiting 2 days ago after several of her coworkers had the same symptoms. Prior to presentation, she vomited a small volume of “coffee ground” emesis. In the ED, her vitals are HR 75 bpm, BP 122/60 mm Hg, RR 16/min, and oxygen saturation 100% on room air. Labs are remarkable for Hgb 12.5 g/dL, BUN 10 mg/dL, and Cr 0.8 mg/dL. After an antiemetic, she is able to tolerate oral intake and wants to be discharged home. The ED physician asks you if she needs to be admitted for further treatment or monitoring.
Which patients with upper gastrointestinal bleed (UGIB) can be discharged without admission?
Patients with a Glasgow-Blatchford score (GBS) score of 0 can be safely managed in the outpatient setting.
The GBS is a risk stratification tool for UGIB based on gender, presentation (syncope, melena), hemodynamic (HR, SBP), lab (BUN, Hgb), and past medical history (hepatic disease, cardiac disease). It has been validated to predict the need for blood transfusion, endoscopic treatment, or surgery for UGIB.1 A prospective study at four UK hospitals was designed to test the hypothesis that those with the lowest GBS (0) could be safely discharged from the ED without need for intervention (defined as blood transfusion, endoscopic treatment, or surgery) or
mortality.2 In phase 1 of the study, 676 patients presenting with UGIB were evaluated (prospectively for 3-12 months at three hospitals, retrospectively for 3 months at one hospital). None of the 105 (16%) patients with a GBS of 0 required intervention or died after inpatient monitoring. In phase 2, patients with GBS 0 were not hospitalized unless they had another indication for admission. Of 491 patients identified with UGIB, 123 had a GBS of 0, of whom, 84 were not hospitalized. Nonadmitted patients were followed with outpatient endoscopy or, in absence of endoscopy, by chart review or discussion with their family physician. Out of 123 patients with a GBS of 0, none required intervention. One of these patients died due to causes unrelated to their UGIB. Using GBS to identify the lowest risk patients could therefore avoid unnecessary admissions. Caveats include a somewhat small and homogenous population.
mortality.2 In phase 1 of the study, 676 patients presenting with UGIB were evaluated (prospectively for 3-12 months at three hospitals, retrospectively for 3 months at one hospital). None of the 105 (16%) patients with a GBS of 0 required intervention or died after inpatient monitoring. In phase 2, patients with GBS 0 were not hospitalized unless they had another indication for admission. Of 491 patients identified with UGIB, 123 had a GBS of 0, of whom, 84 were not hospitalized. Nonadmitted patients were followed with outpatient endoscopy or, in absence of endoscopy, by chart review or discussion with their family physician. Out of 123 patients with a GBS of 0, none required intervention. One of these patients died due to causes unrelated to their UGIB. Using GBS to identify the lowest risk patients could therefore avoid unnecessary admissions. Caveats include a somewhat small and homogenous population.
The 2012 American College of Gastroenterology (ACG) peptic ulcer guidelines note that discharge from the ED “may be considered” if GBS is 0.3 Subsequent studies4,5 suggest that patients with GBS ≤1 may also be good candidates for outpatient management due to low risk of mortality or need for intervention.
You advise the ED physician that the patient has a GBS score of 0 and is therefore at very low risk of adverse event or needing endoscopic therapy. You both agree that she can be discharged to follow-up with her primary care doctor.
A 70-year-old man with hypertension, prior stroke, and arthritis presents with melena and anemia. His vitals are stable. Labs are notable for Hgb 7.2 g/dL. You are suspicious that his bleed is from an upper source, but consider the possibility that this could in fact be a lower gastrointestinal bleed (LGIB). You wonder how you can determine that before the gastroenterology consultant arrives.
How accurate are signs, symptoms, and labs at distinguishing UGIB from LGIB?
Melena on examination and elevated blood urea nitrogen (BUN)/Cr ratio can be helpful in distinguishing an upper from lower source of bleeding. Nasogastric (NG) lavage with blood or coffee grounds, if present, makes UGIB much more likely, but false negatives are common.
A 2012 meta-analysis6 assessed published studies regarding the diagnostic accuracy of signs and symptoms to differentiate UGIB and LGIB. Studies were selected if they involved patients with apparent GIB presenting to the ED or requiring hospitalization. Exclusion criteria included studies of primarily inpatients or children. Twenty-five studies met inclusion criteria. Sensitivity, specificity, +LRs (likelihood ratios), and −LRs were calculated.
Several features reliably rule in UGIB, including melena detected on clinical examination or observed by a clinician (+LR 25, 95% CI 4-174), NG lavage demonstrating blood or coffee ground material (+LR 9.6, 95% CI 4.0-23.0), and BUN/Cr ratio >30 (+LR 7.5, 95% CI 2.8-12.0). Other elements of history and physical are moderately helpful in determining if the patient has an UGIB, including prior history of UGIB (+LR 6.2, 95% CI 2.8-14.0) or patient report of melena (+LR ranging 5.1-5.9 in two studies). Of note, the absence of these features is not diagnostically useful to rule out UGIB.
Factors decreasing the likelihood of UGIB include clots in stool (+LR 0.05, 95% CI 0.01-0.38) and prior LGIB (+LR 0.17, 95% CI 0.09-0.35). NG lavage without blood or coffee ground material is only marginally helpful for ruling out UGIB (+LR 0.58, 95% CI 0.49-0.70). The resultant risk of falsely ruling out UGIB has led the ESGE to recommend against routine NG lavage in UGIB (strong recommendation, moderate quality evidence).7
You visualize melena during your examination and his BUN/Cr ratio is 35, both strongly suggesting UGIB rather than LGIB. He is resuscitated and scheduled for esophagogastroduodenoscopy (EGD). Prior to admission, the ED physician started a continuous infusion of proton pump inhibitor (PPI). You consider transitioning this to IV bolus dosing.
Related posts:
Full access? Get Clinical Tree
