Beware the Big Bleed

Contributed by readers. Edited by Sheldon Jacobson, MD

A 76-year-old man was brought to the emergency department (ED) by his wife after he experienced a near-syncopal episode at home during the night. He was in his usual average health when, after arising to have a bowel movement, he became dizzy and had to lie on the floor for a time, unable to get up. Not wishing to awaken his wife, he remained supine for about a half-hour and then crawled back to bed. Several hours later, he was able to get out of bed but still felt weak and dizzy. His concern about this continued state of debility prompted him to visit the ED.

The patient's medical history was positive for alcoholism and vague liver "trouble." He claimed to have been sober for seven years. A retired welder, he smoked a pack of cigarettes a day. His only medications were terazosin for symptoms of prostatism and antacids for chronic dyspepsia.

On arrival in the ED, the patient was in mild distress. His skin was slightly cool and diaphoretic. His pulse was 104 and regular; blood pressure, 110/50 mm Hg; and respiratory rate, 16. Pulse oximetry on room air revealed 92% oxygen saturation. The results of a pulmonary examination indicated decreased breath sounds, prolonged expiration, and an increased AP diameter of the chest--all compatible with chronic obstructive pulmonary disease (COPD). The results of a cardiac examination revealed only regular tachycardia, and those of an abdominal examination were unremarkable, except for slight epigastric tenderness. Results of a rectal examination showed dark stool that was 4+ heme positive. The patient had one episode of emesis of clear liquid in the ED. An ECG revealed low voltage and nonspecific ST-segment and T-wave abnormalities.

After 500 cc of intravenous saline was administered within the first half-hour, the patient's blood pressure rose to 126/60 mm Hg and his pulse rate slowed to 80. Initial laboratory data included a hematocrit of 36% and normal WBC and platelet counts. The blood urea nitrogen (BUN) level was 38 mg/dL and the creatinine level was 1.4 mg/dL. The partial thromboplastin time (PTT) was 28 seconds and the international normalized ratio (INR) was 1.4.

After repeating the rectal examination, the gastrointestinal consultant came to the conclusion that the stool color was actually "dark green" and therefore did not indicate true melena. He believed that since the patient's vomitus was clear and the stool green, emergency endoscopy was not indicated. He also believed that the patient was dehydrated and that the terazosin he had been taking had blunted his vasomotor reflexes. In addition, now that the patient was looking and feeling better, he could undergo an elective outpatient work-up.

After receiving two liters of intravenous fluids, the patient had no significant orthostatic hypotension. Accordingly, he was discharged on a regimen of iron supplementation and H2-blocker therapy. Arrangements were made for a follow-up gastrointestinal work-up through his primary care physician.

Just as the nurse was about to remove the patient's intravenous tube, he vomited up a large amount of dark red liquid and clotted blood. He became diaphoretic, and his blood pressure dropped to 80/50 mm Hg. The patient was given a rapid fluid infusion along with a packed red-cell infusion, and the gastrointestinal bleeding team was recalled.

The patient again vomited copious amounts of blood. The medical team administered nasal oxygen, inserted a femoral line--with difficulty--and infused additional blood under pressure. The general surgery service on call that day were paged immediately, but the intern was available, who went to the ED to assist the staff. The patient began having frequent ventricular premature contractions, and his T waves were now inverted in leads I, aVL, and V1-3.

Because the situation had become desperate, the invasive radiology team was asked to attempt immediate mesenteric arteriography with the aim of identifying the bleeding site so that it could be embolized. When the gastrointestinal team arrived, they were unable to perform upper endoscopy because the patient was too unstable. Before being rushed to the radiology suite, the patient underwent intubation and mechanical ventilation.

During the intubation procedure, the patient's blood pressure could not be measured. After administration of fluids, the systolic reading returned once again to a level between 70 and 80 mm Hg, but to no avail. Before arteriography could be performed, the patient suffered cardiac arrest in the radiology department. Resuscitation efforts were unsuccessful. The postmortem examination revealed copious amounts of blood in the small and large bowel and a duodenal ulcer that had penetrated posteriorly into the pancreaticoduodenal artery.

COMMENT

This patient had essentially died from a major arterial hemorrhage that was not identified and treated with the haste and alacrity required to alter the unfortunate outcome. Although it is true that in several older studies early endoscopy was not found to affect patient outcome, these studies were conducted before endoscopic hemostatic techniques were perfected. Even in the closely scrutinized cases in those older studies, had there been a subgroup analysis of patients who presented with near-syncope, hypotension, and 4+ heme positive stools, certainly early endoscopy would have proved helpful in identifying the bleeding site, treating the bleeding vessel (if identified), and contributing to the determination of the proximate and ultimate prognosis.

The emergency physician and gastrointestinal consultant did not appreciate that this patient had sustained a major gastrointestinal hemorrhage at home and that the bleeding had slowed because hypotension was present. That there was no blood in the gastric juice--which incidentally was minimally bile stained--simply means that there was no reflux of duodenal contents across the pylorus. The finding does not rule out upper gastrointestinal bleeding. Having minimized the bleeding, the physicians were unprepared to deal with a major arterial gastrointestinal exsanguinating hemorrhage. The final mechanism of this patient's death could have been a myocardial infarction, as elderly patients can have occult coronary artery disease and little cardiac reserve to deal with extreme blood loss.

The best response in a clinical scenario such as this is to assume the worst and not downplay the bleeding. The patient should be examined immediately by both the general surgeon and the endoscopist, who should not be let off so easily: The patient's vital signs were not measured early in his ED course, nor was a second hematocrit obtained after he underwent hydration. Significant abnormalities in these findings might have convinced the physicians of the seriousness of the situation.

Once an exsanguinating gastrointestinal hemorrhage has occurred, immediate surgery is the only viable solution. If logistical or other factors make this option impossible, some temporizing measures can be tried in addition to massive blood transfusion: intravenous administration of octreotide, which is a somatostatin analog that reduces mesenteric perfusion; perfusion of the bleeding vessel or proximal tributary with vasopressin or octreotide or embolization of the vessel with autologous clots or thrombogenic material; and temporary balloon tamponade of the thoracic aorta to restore blood volume. If the blood is coming from the esophageal varices, a Sengstaken-Blakemore tube can be passed into the esophagus and inflated to tamponade the varices.

The aphorism that patients in the ED are usually sicker than they look is proven here once again.