GI Bleeding:

UGIB and LGIB are often very concerning for the housestaff caring for the patient, as these entities can be life threatening disorders, especially in the elderly patient with cardiac and possibly other coexisting diseases.Treat this phonecall as a consult, not an order for a study.This is a complicated, multistep process.

1st step:Calm the house officer and take the patientís past medical and past surgical and medication history.Demand this information in full, as this will be important to give to the attending when you call him or her, and when considering the patient for angio.Write this information down.

When obtaining the history and physical information, make sure to inquire about the coagulation parameters and CBC (with trends).Frequently, these will be supertherapeutic and the cause of the patientís bleeding.When these are corrected, the bleeding will often stop.Supertherapeutic parameters also limit angio therapy options and effectiveness.

Inquire about any known contrast allergies or renal failure or other contraindications to angiography.

Can the patient give consent for angiography?If not, is there someone else available to give consent?(If there is no family to consent, 2 physicians outside of the department can consent.)

If the patient has been passing blood per rectum, but not at alarming rates continuously, you can relay to the housestaff that the blood which was passed (causing all this franticness) could have been in the large intestine since the last bowel movement.That said, the nuclear medicine Ė angiography pathway is designed to detect active bleeding, happening at the time of the examination.

At night, the study we perform is a Tc 99m sulfur colloid scan per Dr. Bader.This does not require drawing the patientís blood for the examination and is slightly quicker to do.It also does not interfere with a planned tagged RBC study the next day.

The advantages of Tc-SC imaging is that the agent is rapidly cleared from the intravascular space by the reticuloendothelial system. The circulating half-life in patients with normal liver function is between 2-3 minutes, and by 15-20 minutes there is effectively none left in the blood. This permits clear visualization of extravasated isotope at bleeding site. The technique is very sensitive and can detect bleeding rates as low as 0.05 to 0.1 cc/min. (Angiography realistically requires a bleeding rate of about 1.5 to 3 ml/min. to be detected. Angiography will detect GI bleeding in 65% of patients when the bleeding rate exceeds 1 ml/min.)

Disadvantages Tc-SC: Prominent liver and spleen activity may obscure the bleeding site. The imaging time is markedly limited (probably no more than 10 minutes). Given the characteristic intermittent nature of GI bleeding, unless the patient is actively bleeding at time of study, the site of hemorrhage will not be detected. Reports have indicated that the sensitivity Tc-SC imaging for GI bleeding is about 30% of the labeled RBC technique. Because of this, to improve sensitivity, some centers will administer the agent in 2 mCi doses, every 10 minutes for an hour.

Alternative: Technetium labeled RBC's: (Dose 20 mCi)

The main advantage of using labeled RBC's is that the agent remains in the intravascular space for 24 (up to 48) hours, thus it is excellent for imaging intermittent or slow bleeds. Liver, spleen, kidneys and bladder (due to excretion of non-erythrocyte bound tech and labeled hemoglobin fragments) are also visualized, but to a lesser degree and thus will probably not obscure the bleeding focus.

The labeled RBC exam cannot detect bleeding rates as low at the Tc-SC exam, however, bleeding rates as low as 0.1 to 0.5 cc/min. and sensitivities of over 90% have been reported for this procedure (the tagged RBC exam has a sensitivity of 91% and a specificity of 95% of GI bleeding [Datz, p. 349]). Continuous imaging is recommended because it is often difficult to localize the bleeding site with intermittent imaging. In vivo labeling of the RBC's is also not recommended as free tech will be excreted by the gastric mucosa and pass into the small bowel and colon. Continuous gastric suction may be required if the in vivo labeling technique is used. In vitro tagging results in significantly less amounts of free technetium. Presently, many centers employ the use of the ultratag kit for labeling the RBC's prior to GI bleed imaging. †††††††††††††††††††††††††††††††††††††††††††††††††††† from AuntMinnie.com

The purpose of the nuclear medicine examination is to find patients likely to have active bleeding during angiography.There is no need to do angio on a pt who is not bleeding by NM exam as demonstrated by their relative sensitivities.

This last point brings up the need for surgical and cardiology/internal medicine consultation.Surgery needs to direct the care toward angiography if the patient is not in need of emergent surgery.GI bleeding may require emergent surgery, and it would be bad to have such a patient under a gamma camera for an hour if that is the case.If the patient is unstable and still needs a GI bleeding study, a nurse should accompany the patient to NM.Surgery also needs to know about the patient for any angiographic complications that might arise during diagnosis or treatment in angio.

Cardiology and internal medicine input is helpful to clear the patient for vasopressin therapy.Vasopressin could potentially have cardiac side effects, and the patient may have known ischemic CAD.In this setting, one might choose to use embolization to arrest active bleeding in angio.

Finally, there is literature to support emergent endoscopy in actively bleeding patients.Below are given two examples.Medicine may want to consult gastroenterology for a case. 

Paper on GI Bleeding from University of Washington:Margaret Shuhart, M.D. , Kris Kowdley, M.D., and Bill Neighbor, M.D.

http://www.uwgi.org/cme/cmeCourseCD/ch_07/ch07txt.htmsee section 2.2.2

Urgent colonoscopy for the diagnosis and treatment of severe diverticular hemorrhage.From Jensen, M.D., Machicado, M.D., Jutabha, M.D., Kovacs, M.D. at the Digestive Disease Research Center, UCLA

http://content.nejm.org/cgi/content/abstract/342/2/78