The three most important pancreatic tumors are carcinoma of the exocrine pancreas (pancreatic adenocarcinoma of duct origin), islet cell tumors producing insulin (insulinoma), and islet cell tumors producing gastrin (gastrinoma) associated with the Zollinger-Ellison (Z-E) syndrome.

Exocrine adenocarcinoma. Carcinoma of the pancreas as a descriptive term usually refers to an adenocarcinoma of the exocrine pancreatic ducts, which comprises 90%–95% of pancreatic carcinomas. At the time of diagnosis about 65% are located in the head of the pancreas, about 20% in the body, about 5% in the tail, and about 10% are relatively diffuse. At the time of initial diagnosis, pain (usually abdominal) is present in about 75% of patients, weight loss in 50%–60%, bowel symptoms in 20%–30%, an abdominal mass in 5%–50%, and thrombophlebitis or thromboembolism (traditionally highly associated with pancreatic carcinoma) in 5%–10% (more common with body or tail lesions). At the time of diagnosis there is said to be local invasion or lymph node spread in about 25% of cases and distant metastases in about 60% (although literature ranges for distant metastases vary widely, depending on whether the metastases are overt or occult).

Laboratory findings include anemia in 25%–50% of cases, stool occult blood in as many as 50% of tumors in the pancreatic head, fasting hyperglycemia in about 20%, and oral glucose tolerance test abnormality in about 50% (literature range, 20%–81%, depending on criteria used). Jaundice is present in about 65% (seen in 45%–95% of pancreatic head tumors but much less common in those from the body and tail). Alkaline phosphatase level is elevated in most patients with jaundice and about one third of those without jaundice. Serum amylase level is elevated in only about 10% of patients.

Cancer antigen 19-9 (CA 19-9) and carcinoembryonic antigen (CEA) levels are elevated in a substantial percentage of patients with pancreatic carcinoma. However, at least 10%–20% of patients have normal levels of these tumor markers, and levels of both markers are elevated in a significant percentage of other tumors and nonmalignant conditions. Therefore, neither test is currently being widely used for screening, diagnosis, or therapy of pancreatic carcinoma.

Upper GI series is reported to be about 50% sensitive alone and about 80% sensitive with hypotonic duodenography for detecting carcinoma of the pancreas head. Duodenal tube drainage with secretin stimulation has an overall sensitivity of 30% (literature range, 10%–90%), which is increased to 50% (20%–84%) when cytologic study is performed on the pancreatic duodenal secretions. Ultrasound has overall sensitivity of 80% (68%–94%), with about 10%–15% of the studies attempted being technically unsatisfactory. CT scan averages about 80% overall sensitivity (60%–92%), with average sensitivity probably even better with the newest-generation scanners. Endoscopic retrograde choledochopancreatography (ERCP) is 80% sensitive (54%–96%), whereas ERCP combined with duct aspiration cytology is reported to be 85% sensitive. ERCP technical failures occur in about 15%–20% of cases. Selective angiography (celiac and superior mesenteric artery) detects 60% of carcinomas, whereas catheterization of pancreatic vessels (“superselective angiography”) may detect about 90%. Percutaneous transhepatic cholangiography is occasionally needed when jaundice is present and the other methods fail to yield a diagnosis or cannot be used.

Zollinger-Ellison syndrome. The Z-E syndrome is caused by a gastrin-producing nonbeta islet cell (G-cell) tumor of the pancreas (gastrinoma). There are multiple tumors in 70% of cases (literature range, 55%–80%). Two thirds (range, 60%–100%) of the tumors are malignant. Most gastrinomas originate in the pancreas, but occasionally they are found in an “ectopic” location (alone or in addition to the pancreas). Thus, about 10%–13% occur in the duodenum, and, rarely, they may arise in the stomach. Within the pancreas, the majority are in the head or the tail. Their microscopic appearance is similar to that of an insulinoma or carcinoid. About 10%–40% of patients with gastrinomas also have other endocrine tumors (most commonly a parathyroid adenoma); association with the MEN I syndrome is fairly common.

The major components of the Z-E syndrome are listed in the box on this page. Approximately 50%–60% of Z-E syndrome ulcers (31%–75%) are located in the proximal duodenum, which is the usual site for peptic ulcers; 25% (range, 20%–42%) are found in the distal duodenum or the jejunum; and 10% (range, 8%–15%) are found in the stomach. Multiple ulcers occur in 10%–20% of patients. About 10% (range, 7%–15%) have no ulcer. Diarrhea is found in 30%–35% of patients (range, 16%–75%). The degree of diarrhea is variable, but severe chronic diarrhea with hypokalemia is typical. Diarrhea is the only symptom of the Z-E syndrome in 7%–10% of cases. Steatorrhea occurs in 40% of cases (range, 38%–66%). However, many patients initially have symptoms very similar to ordinary peptic ulcer without the classic features of Z-E syndrome.

Signs and Symptoms Suggestive of the Zollinger-Ellison Syndrome
Intractable or recurrent peptic ulcer(s)
Multiple peptic ulcers or ulcers in unusual locations
Recurrent or marginal ulcer after complete vagotomy or partial gastric resection
Chronic diarrhea
Gastric acid hypersecretion

The Z-E syndrome has been divided into two types. Type I is caused by so-called G-cell hyperplasia, an increase in number or activity of gastric antrum G cells without gastrinoma tumor. Type II is due to gastrinoma of the pancreas or duodenum.

Gastric Analysis. The Z-E syndrome is usually accompanied by gastric hypersecretion and hyperacidity, which some consider an integral part of the syndrome. According to the old gastric analysis method using Topfer’s reagent as a pH indicator, now considered outmoded, basal gastric secretion quantity more than 200 ml/hour and basal gastric acid secretion more than 100 mEq/L/12 hours were considered suggestive of Z-E syndrome. According to the currently recommended gastric analysis methods using a pH meter or methyl red pH indicator, basal (1-hour) acid secretion greater than 10 mEq of hydrochloric acid (HCl)/hour and a ratio of basal acid output (BAO) to maximal acid output (MAO) of 0.4 or greater raise the question of Z-E syndrome; a BAO of 15 mEq/hour and a BAO/MAO ratio of 0.6 or greater are very suggestive of Z-E syndrome (although not pathognomonic) (Table 33-10).

Gastric analysis in Zollinger-Ellison syndrome

Table 33-10 Gastric analysis in Zollinger-Ellison syndrome*

Serum gastrin. Serum gastrin assay is the method of choice for diagnosis. Fasting gastrin levels are elevated in more than 95% of gastrinomas. Serum gastrin levels more than 5 times the upper limit of the reference range (1,000 pg/ml or 1,000 ng/L) are virtually diagnostic of Z-E syndrome. Some believe that gastric analysis can therefore be omitted in these patients. However, a few patients with gastrinoma have basal gastrin levels that are within reference range, and 50% have levels that are only mildly or moderately elevated and overlap with values found in certain other conditions associated with elevated serum gastrin levels. These other conditions include diseases associated with hypochlorhydria or achlorhydria, such as atrophic gastritis and pernicious anemia (if the antrum is not severely affected), after vagotomy, in patients with retained antrum following gastrojejunostomy, in uremia, and possibly in chronic hypercalcemia. In one series, about 60% of patients with elevated serum gastrin level had hypochlorhydria or achlorhydria as the cause of the elevated gastrin level. Because of this, some investigators recommend gastric analysis in patients with mild or moderate gastrin elevations. Food ingestion has also been reported to produce a significant temporary increase in serum gastrin level. A high-protein meal is said to increase serum gastrin levels 2-5 times baseline values. In addition, some patients with peptic ulcer have mild or moderate serum gastrin elevation that overlaps with those occasional gastrinoma patients who have values that are not markedly elevated.

Gastrin stimulation tests. Since overlap in gastrin values may occur between gastrinoma and other conditions when fasting gastrin is less than 1,000 pg/ml (1,000 ng/L), stimulation tests have been devised to assist differentiation (see the box on this page). The original standard procedure was calcium infusion. Patients with gastrinomas more than double the baseline values, whereas patients with ulcers fail to do so. Patients with pernicious anemia, however, frequently respond to calcium infusion. Also, calcium infusion can produce cardiac problems, especially in patients with renal or cardiac disease. Secretin stimulation appears to be replacing calcium infusion as the confirmatory procedure of choice. Secretin seems to be a little more sensitive than calcium and appears to differentiate better between gastrinomas and other causes of elevated serum gastrin. In one study, 6% of patients reached the peak at 2 minutes, 69% at 5 minutes, 20% at 10 minutes, and 5% at 15 minutes. However, about 5%–10% of Z-E patients with fasting gastrin elevated but less than 1,000 pg/ml had negative secretin tests. In these patients, calcium infusion may be helpful, since about one third have diagnostic results with calcium. It has been reported that pernicious anemia patients do not respond to secretin stimulation.

Gastrin Stimulation Tests in Zollinger-Ellison Syndrome
Secretin (2 units/kg IV bolus): Baseline; 2,5,10,15 minutes postsecretin. Peak response of over 200 pg/ml (200 ng/L) over baseline occurs in 87%–100% of Z-E syndrome patients.
Calcium Infusion (10% calcium gluconate infusion; 5 mg calcium/kg/hr for 3 hrs): Baseline; post dose 120, 150, and 180 minute specimens. Increase over 395 pg/ml (395 ng/L) occurs in over 95% of Z-E syndrome patients; increase over 3 times baseline occurs in over 85%. Response to calcium is less specific than response to secretin.

Primary gastrinoma localization. Ultrasound is reported to demonstrate 21%–28% of gastrinomas, CT is said to detect 35%–60% (range, 18%–80%), and selective angiography can locate 35%–68%.