The case for computed tomographic angiography for initial management of lower gastrointestinal hemorrhage

Abstract

Lower gastrointestinal hemorrhage remains a common disease, frequently presenting with acute life-threatening symptoms. Although prompt detection and treatment are imperative, it is difficult to diagnose lower gastrointestinal hemorrhage in an accurate and efficient manner. Most available modalities are time consuming. Computed tomographic angiography of the abdomen and pelvis, on the other hand, has the unique capability of rapidly detecting whether life-threatening hemorrhage is occurring and accurately localizing it, thus facilitating definitive treatment. We present a case in which computed tomographic angiography was invaluable in the detection and subsequent empirical transarterial embolization of a lower gastrointestinal hemorrhage and offer evidence as to why it should be a first-line tool in the management of these patients.

Technological advances have made computed tomography an indispensible tool in the diagnosis of many diseases, both acute and nonacute. In a stable patient, colonoscopy remains the diagnostic procedure of choice for lower gastrointestinal hemorrhage, but computed tomographic angiography (CTA) also may render an etiologic diagnosis (1). The main diagnostic advantage of CTA in the setting of acute life-threatening lower gastrointestinal hemorrhage is its speed. It is faster than the more traditional alternatives: colonoscopy, technetium 99m-labeled red blood cell scanning, and conventional angiography. Lower gastrointestinal hemorrhage is intermittent in nature and therefore elusive to detection. If detected, however, CTA yields valuable anatomic information that may allow for treatment via conventional angiography even in the absence of detection of the hemorrhage at the time of the conventional angiogram. We present a case in which a patient was successfully treated using this approach.

CASE PRESENTATION

A 66-year-old woman with a history of end-stage renal disease, type 2 diabetes mellitus, and peripheral artery disease presented with rectal bleeding for 12 hours with a hemoglobin of 7.2 g/L. She was given 2 units of packed red blood cells. CTA of the abdomen and pelvis showed active extravasation within the rectum eccentric to the right side of the lumen, localizing the bleeding source to the right inferior rectal artery (Figure 1). Superselective conventional angiography of the right inferior rectal artery showed no active extravasation (Figure 2). Three-vessel mesenteric angiography also disclosed no extravasation. Empirical superselective transarterial embolization of the right inferior rectal artery was performed with Gelfoam. The patient maintained a stable hemoglobin after embolization throughout her hospital course.

Figure 1.

Axial CTA image of the pelvis showing hyperdense contrast material pooling within the right aspect of the rectal lumen (arrow), indicating active hemorrhage.

Figure 2.

Digital subtraction angiographic frame showing no active extravasation of contrast during an angiographic run from the right inferior rectal artery (single arrow). The active extravasation would be indicated by pooling of contrast in the region indicated by the double arrows; however, none is seen. Rectal mucosal enhancement is noted at the terminus of the artery between the arrows, a normal angiographic finding.

DISCUSSION

The ability of CTA to detect active extravasation of blood into the colon at rates below 0.4 mL/min has been established (2). This rate is slower than the accepted lower limit of detection for conventional angiography (1, 2), and indeed some series show improved detection over conventional angiography (2). There is substantial evidence in the English literature that CTA of the abdomen and pelvis is a safe, fast, accessible, and accurate diagnostic modality for assessing patients presenting with acute lower intestinal bleeding (1, 3). The positive predictive value has been reported as high as 95% with high concordance with standards of reference (3). The detection rate of CTA for active extravasation into the colonic lumen ranges from 26% to 100% (3–6). This is not significantly different from the range of sensitivities for technetium 99m-labeled red blood cell scanning (20%–90%) (6, 7). Furthermore, it is well known that red blood cell scanning produces poor localization of the bleeding (7). There is well-documented, significant, but unsurprising discordant localization of bleeding when comparing CTA and technetium 99m-labeled red blood cell scanning (6). However, given that the sensitivity is similar between the two studies, it is reasonable to implement the study that better localizes the extravasation in the acute setting.

Colonoscopy has been regarded by many sources as the standard for localization of lower gastrointestinal bleeding; however, it is relatively insensitive (70%) in the acute setting during active extravasation, particularly when bleeding is high volume and may be mixed with fecal material if a prep could not be obtained prior to the exam (1, 5, 7–9). Indeed, a formal bowel preparation may inappropriately delay care in the acute setting (1). Furthermore, colonoscopic intervention has a relatively high failure rate at 32% (9).

By comparison, a potential disadvantage of CTA is the contrast load that is potentially additive with that administered at conventional angiography. However, this is potentially offset by the more accurate anatomical information attained at CTA, directing the angiographer to the precise location of the extravasation and thus shortening the intervention (3). Furthermore, even though conventional angiography is often negative due to the intermittent nature of lower gastrointestinal bleeding, the anatomic information afforded by a positive CTA allows for superselective catheterization of the target area to more precisely localize the bleeding site, and in selective cases, empirical embolization. At least two studies have proposed an algorithm that places CTA as the initial evaluation tool for management of acute lower gastrointestinal bleeding (1, 6, 8). Our experience is corroborative that this is perhaps the most appropriate first diagnostic step in managing lower gastrointestinal hemorrhage.