Journal Logo

The Speed of Sound

The Speed of Sound

US Can Identify Which GI Bleed Patients to Discharge

Butts, Christine MD

doi: 10.1097/01.EEM.0000697740.90551.05
    ultrasound, GI bleed:
    A parasternal view of the heart, perfect for assessing ventricular function, particularly the “kissing” walls, which indicate that the walls of the LV (stars) are nearly touching each other in systole.

    Can ultrasound help to determine which GI bleed patients are going to go bad? It seems far-fetched, but a recent study considered this very idea. (J Ultrasound Med. 2020;39[2]:279.)

    Adult ED patients who had reported GI bleeding in 24 hours before presentation were enrolled, but unstable patients, those with prior blood loss or recent ischemic arterial disease, and pregnant patients were excluded. All enrolled patients underwent a history and physical and had blood work done. This information was then used to calculate a predictive score, either a Glasgow-Blatchford or Rockall score for upper GI bleeding or a Velayos score for lower GI bleeding. At least 30 minutes before IV fluids were initiated, patients had an ultrasound that evaluated the IVC for collapsibility and the heart for evidence of hypovolemia and to calculate left ventricular outflow VTI (velocity time integral). These parameters were repeated after a passive leg raise.

    The authors primarily looked at measures of early adverse events, such as rebleeding, need for a blood transfusion, hypotension, hypoxia, and changes to parameters such as hemoglobin or creatinine. Late adverse events were also considered, including death within 30 days. As a secondary outcome, they looked at how including the ultrasound findings into the calculated traditional risk score improved its predictive performance.

    A lot of patients had adverse outcomes, although the most common one was decreased hemoglobin. Finding “kissing” LV walls had a significantly increased odds ratio for any adverse event, and an IVC with greater than 50 percent collapsibility after a passive leg raise had a significantly increased odds ratio for an early adverse effect. The addition of the ultrasound measures to the calculated score increased the sensitivity of upper GI bleeding scores by nine percent and decreased false-negatives by around four percent for both. For lower GI bleeds, the ultrasound measures (specifically IVC collapsibility after a passive leg raise, a hyperdynamic LV, and a small IVC) increased the sensitivity by nearly five percent for early adverse effects and by nearly 30 percent for late adverse events.

    A long axis view of the IVC. The proximal section, to the left of the image, can be evaluated for a change in size (collapsibility) when the patient sniffs in or with respiration.

    Is this helpful? Some of the parameters (for example, the VTI) require more time and effort, but looking at the IVC, even with a passive leg raise, is fairly easy. And it's intuitive. In patients with a collapsible IVC after a passive leg raise, you're seeing a clear sign that they've lost volume. Many of the patients in this study were rated as high risk by traditional scoring. Ultrasound doesn't add a ton to deciding whether to admit these patients, but it might give some guidance on resuscitation. In those patients with lower scores (particularly in the lower GI bleed group, where most were at intermediate risk), taking a few minutes to check the heart and IVC might make you more confident in discharging them home.

    Dr. Buttsis the director of the division of emergency ultrasound and a clinical assistant professor of emergency medicine at Louisiana State University at New Orleans. Follow her on Twitter@EMNSpeedofSound, and read her past columns at

    Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.