Joint Program in Nuclear Medicine

Evaluation of GI Bleeding with Tc-99m RBC

Paulo A. Carvalho, MD PhD
J. Anthony Parker, MD PhD

February 13, 1996

Presentation

62 year old man with known hypertension.
status post cystectomy and ileoneobladder for a poorly differentiated transitional cell carcinoma of the bladder.
5th post operative day:

sepsis with positive blood culture (gram + cocci) was successfully treated with gentamycin, vancomycin and flagyl.
tachycardia and tachypnea with ECG demonstrating atrial flutter and questionable myocardial ischemia with no chest pain. A previous O2 saturation of 96% in room air was 93% with 2 L of mask O2. Lopressor / Mg++ were prescribed with a decrease in heart rate and reversal of ST-segment changes.

6th post operative day:

questionable bilateral calf tenderness, but a negative Homans' sign. After an episode of sudden tachypnea, a plain radiograph of the chest showed signs of chronic heart failure with no other significant abnormalities. A perfusion / ventilation lung scan demonstrated multiple mismatched segmental perfusion defects consistent with a high probability of pulmonary embolism. Anticoagulation therapy was started with improvement of acute symptoms.

8th post operative day:

IVC filter was placed below the renal veins.

9th post operative day:

hematochezia x 2. In the next 24 hours the patient had other bleeding episodes and received 5 units of fresh frozen plasma and 6 units of RBC (total volume of approximately 2 L). A scintigraphic study with Tc-99m RBC was requested to localize the bleeding site in the setting of anticoagulation after extensive pelvic surgery (higher probability of bleeding in the site of ileal anastomosis, although other sites could not be ruled out).

Imaging Technique

Tc-99m Red Blood Cells

Tc-99m RBCs remain in the vascular compartment allowing continuous monitoring of the whole gastrointestinal tract for a long period. The in vitro or the modified in vivo labeling of RBCs are the methods of choice, maximizing the likelihood of detecting intermittent gastrointestinal bleeding. A dynamic sequential planar image acquisition is obtained for 60 minutes allowing identification of the bleeding site and intraluminal motion of the labeled RBCs. Cinematic acquisition and display improves accuracy preventing incorrect localization or false-negative results due to rapid antegrade and retrograde peristaltic movements.

Tc-99m RBCs have the potential for false-positive readings due to misinterpretation of intravascular activity and the possibility of free pertechnetate accumulation. The sensitivity and specificity of this method are very high if the study is technically adequate and if imaging is long enough.

Tc-99m sulfur colloid

Rapid blood clearance of this tracer from circulation and into the macrophage / phagocytic system allows for increased detection at very low bleeding rates (0.05 to 0.1 ml/min. in animal experiments). This tracer, however, detects bleeding only up to 15 minutes after intravenous injection and areas adjacent to the liver may not be identified. It can be used when bleeding is extremely active or when a Tc-99m RBC study cannot be used.

Imaging Findings

Scintigraphic study demonstrated active bleeding in the right lower quadrant of the abdomen (arrow on left image), most probably corresponding to the terminal ileum. Based on this finding, an angiographic study was performed. (Arrow on the right image shows the ileoneobladder.)
Contrast angiography showed active contrast extravasation (shown by arrow) off a small branch of the ileocolic artery. This branch was successfully embolized with gelfoam. Another contrast injection showed no signs of further extravasation.

The patient was discharged on Coumadin and ß-blockers approximately 3 weeks after the initial surgical procedure.

Discussion

Scintigraphy does not replace either endoscopy or angiography. Both of these studies may accurately detect the bleeding site and potentially provide therapy. Scintigraphy is an adjunct to these methods due to the intermittent nature of GI bleeding and to the difficulty of endoscopic evaluation of acute/massive bleeding. The advantages of scintigraphy include its high sensitivity in low bleeding rates, the ability to provide continuous monitoring, being well tolerated, easy to perform and requires no patient preparation. In addition it provides prognostic information, once high-risk patients needing aggressive therapy usually show a very intense bleeding focus early in the course of the scan (mean bleeding rate of 0.4 ml/min. within one hour). Patients with negative scintigraphy rarely will need urgent surgery.

Patients with GI bleeding are usually classified as to the site of bleeding and the severity of blood loss. Upper-GI bleeding is more frequent and result in hospitalization of up to 300,000 patients annually in the United States. Endoscopy of the proximal GI tract is reported to have an overall diagnostic accuracy of more than 90% in identifying duodenal and gastric ulcers, gastric erosions, varices and Mallory-Weiss tears. Lower-GI tract bleeding is more frequent in the colon as opposed to the small bowel. Colonoscopy performed within 24 hours of hospital admission will confirm a colonic bleeding site in 68% to 77% of cases, being of limited value in acute hemorrhage. The most common causes of colonic bleeding include mucosal vascular malformations such as angiodysplasia, diverticula, or adenomatous neoplasms and polyps.

Angiography will locate GI bleeding sites in up to 65% of cases when hemorrhage occurs at a rate greater than 1 ml/minute. The major pitfall, however, is that bleeding must be occurring during contrast injection (20-30 seconds).

Aside from the necessary localization of the bleeding site, identification of high-risk patients is important because 85% of patients will stop bleeding spontaneously. The remaining 15% are usually elderly patients with at least one major organ-system disease, with mortality being greatest for hospitalized patients (70% for in-patients as opposed to 22% for outpatients).

Indication for aggressive therapy in the high-risk group includes loss of 30% of the estimated blood volume in the first 24 hours, a need for 1,500 ml of blood transfused per 24 hours to maintain stable blood pressure, any hemorrhage to the point of hypotension or shock and significant rebleeding during medical management.

In the present case, the patient1s GI bleed fulfilled the criteria for aggressive therapy and the Tc-99m RBC scintigraphy was properly employed to confirm the suspicion of bleeding in the site of a recent surgical anastomosis of the distal ileum. Furthermore, the identification of the bleeding site within the first hour also pointed to a more aggressive therapy. Angiography was the next logical step, being a less aggressive way to identify the vascular anatomy and attempt embolization. If, however, bleeding would have persisted, laparotomy would have been indicated.

References

1. Maurer AH. Gastrointestinal bleeding and cine-scintigraphy. Semin Nucl Med 996;28:43-50.

2. Datz FL. Gastrointestinal and hepatobiliary imaging. In: Datz FL. Nuclear Medicine. Handbooks in Radiology. Year Book Medical Publishers, Inc., Chicago,1988, pp. 171-175.

3. Alavi A and Worsley D. Scintigraphic detection and localization of gastrointestinal bleeding sites.In: Sandler MP, Patton JA, Coleman RE, et al. Diagnostic Nuclear Medicine. 3rd Edition, Williams & Wilkins, Baltimore, 1996, pp. 801-821.

4. Tumeh SS, Parker JA, Royal HD, Uren R, Kolodny GM. Detection of bleeding from angiodysplasia of the jejunum by blood pool scintigraphy. Clin Nucl Med 1983;8:127-128.

5. Markisz JA, Front D, Royal HD et al. An evaluation of Tc-99m- labeled red blood cell scintigraphy for the detection and localiza- tion of gastrointestinal bleeding sites. Gastroenterology 1982;83: 394-398.

6. Smith R, Copely DJ, Bolen FH. Tc-99m RBC scintigraphy: correlation of gastrointestinal bleeding rates with scintigraphic findings. AJR 1987;148:869-874.

Click here to go to Joint Program in Nuclear Medicine home page and Copyright notice.

J. Anthony Parker, MD PhD, Tony_Parker@bidmc.harvard.edu