Nonvariceal Upper Gastrointestinal Hemorrhage: Probing Beneath the Surface

Abstract

Looking at the surface of the gastrointestinal (GI) tract has long been the foundation of practice in GI endoscopy. Indeed, many of our day-to-day management decisions are based largely on what the surface looks like, yet it is generally accepted that visual appearances alone can be quite subjective. For many decades, visual stigmata of recent hemorrhage (SRH) formed the basis from which the endoscopist decided whether or not to perform endoscopic hemostasis in patients with nonvariceal upper GI hemorrhage (NVUGIH). The so-called Forrest classification of SRH in peptic ulcer bleeding was published almost 4 decades ago.1 Current recommendations for endoscopic therapy in acute peptic ulcer hemorrhage reside entirely on the surface appearance of the bleeding ulcer.2,3 However, published evidence shows that there is significant interobserver variability in the identification of such SRH. When endoscopists were shown images or videoclips of bleeding peptic ulcers and asked to identify SRH, there was disagreement 25% of the time.4 Moreover, even among international experts, good agreement could only be reached when there was spurting blood.5 Furthermore, some published studies have shown that high-risk stigmata, such as a nonbleeding visible vessel (NBVV), may have uncharacteristic visual appearances (nonpigmented, pale, translucent, or pearl-colored protuberance) and thus may be misinterpreted by the endoscopist as low-risk stigmata.6,7 As Michael V. Sivak, Jr, MD, succinctly noted, “It is well-established that visual assessment alone of an ulcer in a patient with upper GI bleeding is inaccurate, especially with respect to ulcers with SRH. Consequently, the use of the Forrest classification as a guide to the need for endoscopic therapy is necessarily inaccurate” (personal communication, 2006). Indeed, misguided application of endoscopic therapy (either underor overuse) can potentially lead to undesired patient outcomes and to inappropriate use of health care resources. Conceptually, it may be more important to know whether a bleeding lesion still has active blood flow or whether such blood flow has ceased, for instance, as a result of spontaneous intravascular thrombosis. Thus, for any particular bleeding lesion that has stopped bleeding, the risk of recurrent bleeding would theoretically be reduced if active blood flow to the bleeding point had ceased. This theoretical concept has led to the development of novel imaging and nonimaging technologies to try and determine whether blood flow is present or absent beneath a bleeding lesion such as a peptic ulcer. Some of these new technologies may also allow for an assessment of certain physical characteristics of subsurface blood flow, such as relative velocity, and may provide information on anatomic and structural aspects of the surrounding tissue in relation to the bleeding lesion. This