Serum gastrin in Zollinger-Ellison syndrome: I. Prospective study of fasting serum gastrin in 309 patients from the National Institutes of Health and comparison with 2229 cases from the literature

Abstract

The assessment of fasting serum gastrin (FSG) is essential for the diagnosis and management of patients with the Zollinger-Ellison syndrome (ZES). Although many studies have analyzed FSG levels in patients with gastrinoma, limited information has resulted from these studies because of their small size, different methodologies, and lack of correlations of FSG levels with clinical, laboratory, or tumor features in ZES patients. To address this issue, we report the results of a prospective National Institutes of Health (NIH) study of 309 patients with ZES and compare our results with those of 2229 ZES patients in 513 small series and case reports in the literature. In the NIH and literature ZES patients, normal FSG values were uncommon (0.3%-3%), as were very high FSG levels >100-fold normal (4.9%-9%). Two-thirds of gastrinoma patients had FSG values <10-fold normal that overlap with gastrin levels seen in more common conditions, like Helicobacter pylori infection or antral G-cell hyperplasia/hyperfunction. In these patients, FSG levels are not diagnostic of ZES, and gastrin provocative tests are needed to establish the diagnosis. Most clinical variables (multiple endocrine neoplasia type 1 status, presence or absence of the most common symptoms, prior medical treatment) are not correlated with FSG levels, while a good correlation of FSG values was found with other clinical features (prior gastric surgery, diarrhea, duration from onset to diagnosis). Increasing basal acid output, but not maximal acid output correlated closely with increasing FSG. Numerous tumoral features correlated with the magnitude of FSG in our study, including tumor location (pancreatic > duodenal), primary size (larger > smaller) and extent (liver metastases > local disease). In conclusion, this detailed analysis of FSG in a large number of patients with ZES allowed us to identify important clinical guidelines that should contribute to improved diagnosis and management of patients with ZES.

Figure 1.

Upper gastrointestinal (GI) endoscopy, magnetic resonance imaging (MRI) and somatostatin receptor scintingraphy (SRS) result in the ZES patient in the case presentation. Upper panel. The upper GI endoscopy showed Barrett esophagus is present. Middle panel. The MRI in this image showed a 2 cm lesion compatible with a pancreatic endocrine tumor. On other images two additional lesions (1.8 and 2.5 cm) were seen. Lower panel. The SRS performed after the injection on ¹¹¹In penetreotide shows 3 lesions in the pancreatic head/duodenal region.

Figure 2.

Frequency of fasting serum gastrin levels in 309 Zollinger-Ellison patients seen at the National Institutes of Health (NIH) and 2229 Zollinger-Ellison patients from the literature. Gastrin levels are expressed as fold of the upper limit of the normal range. For NIH patients, the upper limit of the normal is 100 pg/ml and for the literature patients, the upper normal value indicated in the manuscript was used. Data are expressed as cumulative percent at the indicated gastrin level. The cumulative percentage of patients with fasting serum gastrin in the normal range or in the ranges 1.1-2.5 fold normal, 2.6-5, 5.1-7.5, 7.6-10, 10.1-12.5, 12.6-15, 15.1-20, 20.1-30, 30.1-50, 50.1-100, 100.1-1000 and greater than 1000-fold increase were determined and plotted for both the NIH and literature patients. The * indicates a significant difference at the given interval between NIH and literature patients (p<0.001 for **, p<0.02 for *). References for literature data prior to 1976 are ^{,,,,,,,,,,,,,–,,,,,,,,,,,,–,–,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,}. For the years 1977-1979 data are from ^{,,–,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,–,–,,,,,,,}. For the years 1980-1983 data are from ^{,,,,,,,,,,,,,,,,,,,,,,,–,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,}. For the years 1984-1987 data are from ^{,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,}. For the years 1988-1992 data are from ^{,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,}. For the years 1993-1995 data are from ^{,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,–,,,,,,–,,,,,,,,,,}. For the years 1996-1999 data are from ^{,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,–,,,,,,,,,,,}. For the years 2000-2005 data are from ^{,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,}.

Figure 3.

Correlation between basal acid output (BAO) and fasting serum gastrin level (upper panels) and between maximal acid output (MAO) and fasting serum gastrin levels (lower panels). Data from 238 ZES patients seen at NIH and 753 ZES patients from the literature without prior gastric surgery are shown. Literature data are from publications listed in Figure 2. Each point represents data from one patient. Indicated are the regression line and the correlation coefficient (r) using a least-squares analysis.

Figure 4.

Correlation between tumor size and fasting serum gastrin level in NIH and literature ZES patients. Upper Panels. The size of the largest tumor in NIH patients with ZES is correlated with the fasting gastrin level in all patients (left) or only those without liver metastases (right). Middle and bottom panels. Correlation between the size of the primary tumor and fasting serum gastrin level in NIH ZES patients (middle) and ZES patients from the literature (bottom) in all patients (right panels) or only those without liver metastases (left panels). Fasting serum gastrin levels are expressed as the log of the fold increase over the upper limit of the normal value. Literature data are from publications listed in Figure 2. Each point represents data from one patient. Indicated are the regression line and the correlation coefficient (r) using a least-squares analysis.

Figure 5.

Effect of primary tumor location, primary tumor size or tumor extent on fasting serum gastrin levels in ZES patients from NIH and from the literature. Fasting serum gastrin levels are expressed as the median fold increases in patients with the indicated variable of tumor size, location or extent. Other location refers to patients with an established primary tumor in a nonduodenal, nonpancreatic location. Primary tumor size is expressed as the diameter of the largest primary tumor. Localized disease was defined as occurring when no liver or distant metastases were found. literature data are from publications listed in Figure 2. Patient numbers from the NIH and literature with the indicated variable are shown in parenthesis.