Diagnosis of Zollinger-Ellison syndrome: Increasingly difficult

Abstract

In the present paper the increasing difficulty of diagnosis of Zollinger-Ellison syndrome (ZES) due to issues raised in two recent papers is discussed. These issues involve the difficulty and need to withdraw patients suspected of ZES from treatment with Proton Pump Inhibitors (omeprazole, esomeprazole, lansoprazole, rabeprazole, pantoprazole) and the unreliability of many gastrin radioimmunoassays. The clinical context of each of these important issues is reviewed and the conclusions in these articles commented from the perspective of clinical management.

Key Words: Zollinger-Ellison syndrome, Gastrinoma, Hypergastrinemia, Secretin test, Serum gastrin, Gastrin, Neuroendocrine tumor

INVITED COMMENTARY ON HOT ARTICLES

Two recent papers by Poitras et al[1] and Rehfeld et al[2] call attention to two areas that are making it more difficult to diagnosis Zollinger-Ellison syndrome (ZES). In this short review after listing the papers and their abstracts, the importance of these two issues will be briefly commented on.

Background: General

ZES is a clinical syndrome due to the ectopic secretion of gastrin by a neuroendocrine tumor (gastrinoma), located primarily in the duodenum (60%-80%) or pancreas (10%-40%), resulting in gastric acid hypersecretion, which if left untreated results in refractory peptic ulcer disease, severe gastroesophageal reflux disease, diarrhea and finally death, primarily due to the complications of the refractory peptic ulcer disease[3-6]. The diagnosis of ZES, like the diagnosis of other ectopic hormonal pancreatic endocrine syndromes, historically requires the demonstration of inappropriate release of the hormone and evidence of hormonal hypersecretion[3,6-8]. In the case of ZES this requires the demonstration of inappropriate gastrin release by demonstrating fasting hypergastrinemia in the presence of gastric acid hypersecretion[3,6-10]. This is clinically most frequently accomplished by demonstrating the inappropriate presence of fasting hypergastrinemia when gastric fluid is acidic with a pH ≤ 2 is present or gastric hypersecretion is present [> 15 mEq/h basal acid output (no previous gastric acid reducing surgery), > 5 mEq/h (if gastric acid reducing surgery)][7,8,11-14]. The combination of fasting hypergastrinemia and elevated gastric acid secretion are required for ZES diagnosis because numerous unrelated conditions can cause one or the other of these alone[5,6,8,10,14]. The most frequent cause of fasting hypergastrinemia is physiological hypergastrinemia (also called appropriate hypergastrinemia) due to the presence of hypo/achlorhydric, which in normal individuals results in a reciprocal increase in gastrin release from the G cells of the gastric antrum causing hypergastrinemia[7,8,10,14,15]. This is most frequently due to chronic atrophic gastritis, commonly due to the presence of a Helicobacter pylori (H. pylori) infection that spares the antrum or due to pernicious anemia[7,8,10,14,15]. A second very frequent cause is the use of potent gastric acid anti-suppressant drugs such as proton-pump inhibitors (PPIs) (omeprazole, lansoprazole, pantoprazole, esomeprazole, rabeprazole), which is discussed in the next paragraph. Other less common causes of hypochlorhydric/achlorhydria include chronic renal failure and vagotomy, which can be distinguished by appropriate other laboratory/clinical investigations[5,8,10,13]. Similarly, the presence of gastric acid hypersecretion without hypergastrinemia can be seen in patients with idiopathic gastric acid hypersecretion and a few other uncommon conditions (mastocytosis, basophilic granulocytic leukemia)[8,16,17].

While the diagnosis of ZES sounds simple enough, in recent years it is becoming increasingly more difficult, due to a number of developments, and now, as pointed out in the above two papers, it is even further complicated. Let’s first consider the issues that were complicating the diagnosis of ZES prior to these two papers. First, acid secretion is now rarely measured and therefore generally not available, so alternatives to the classical basal acid output assessment were proposed. These include, in addition to the presence of fasting hypergastrinemia, an endoscopic measurement of gastric acid output[18], the use of either pH paper or a pH meter to establish the presence of an acidic pH in gastric fluid, assessment for the presence of other features of ZES such as the presence of a tumor on imaging or pathologic studies and the development of gastric fluid pH criteria , that when coupled with the presence of hypergastrinemia, support the diagnosis of ZES[1,11,14]. Second, the widespread use of PPIs is markedly complicating the ability to diagnosis ZES, because their use interferes with both needed assessments to establish the diagnosis of ZES: the measurement of fasting gastrin levels and the assessment of acid secretion[5,10,12,14,19]. This occurs because PPIs have a very long duration of action, with their gastric acid suppressive action lasting up to one week[13,20,21] which not only contributes to their marked effectiveness, but also makes it difficult to withdraw patients from these drugs to assess gastric secretion[1,5,14]. Furthermore, their potent antisecretory activity can lead to fasting hypergastrinemia due to the PPIs and thus mislead one into suspecting ZES[14,19,22]. The result of the PPIs is that they both mask the diagnosis of ZES, leading to delays in diagnosis, because they effectively control all presenting clinical symptoms, but also they can lead to a false diagnosis[3,5,14,19,23]. False diagnoses occur because long-term use of PPIs can cause fasting hypergastrinemia in 80%-100% of patients without ZES[24-27] with fasting gastrin values > 4 fold increased in 20%-25% of patients[24-27] and in some cases the gastrin levels are > 10-fold elevated into ranges that are frequently thought to reflect ZES[25,26].

Commentary paper

In Poitras et al[1] two patients are reported in whom ZES was suspected (later proven) after presenting with severe symptoms of gastroesophageal disease, subsequently treated with PPIs with symptom improvement and when the PPIs were withdrawn, both patients developed severe complications of peptic disease (patient No. 1, esophageal stricture requiring repeated dilatations; patient No. 2, intestinal perforation). It was proposed that PPI therapy should always be maintained and diagnostic evaluations be performed while taking PPIs[1]. This is a novel recommendation and would have a marked effect on the diagnostic approach to patients suspected of having ZES. This recommended approach differs from the approach recommended in recent consensus guidelines and by most authorities in recent reviews, wherein, it is recommended that acid antisecretory drugs have to be stopped at some point to establish the diagnosis of ZES. Specifically, both the North American Neuroendocrine Tumor Network guidelines[9] and the European Neuroendocrine Tumor Network’s guidelines[3,12], as well as a recent reviews of the diagnosis of ZES by a number of authorities[7,10,14,22,28], all recommend that PPIs need to be generally stopped to establish the diagnosis of ZES.

Is there any additional evidence to support this novel recommended diagnostic approach for ZES in paper 1[1]? Others have reported severe esophageal strictures in patients with ZES, since the time that antisecretory drugs were available, whose acid hypersecretion was not controlled[29-31] and in one large study[32], 8% (10/122) patients with ZES required repeated esophageal dilations, because of previous poor control of the acid hypersecretion. We don’t find additional cases in the literature to patient No. 1 described in this report[1], however, we have seen one patient with ZES who developed a severe long, esophageal stricture requiring a stent, because antisecretory medications were withdrawn at an outside hospital for diagnosis (unreported case). Similarly intestinal perforations have been reported in a number of patients with ZES since antisecretory drugs became available. Intestinal perforations were reported in 7% of patients with ZES in one large series (11/160 cases)[33] and in a number of other case reports[34,35], with the perforations occurring prior to the diagnosis in most cases. However, we have seen three cases of ZES patients who developed intestinal perforations when taken off of antisecretory drugs for diagnostic reasons[33,34] (unpublished 2 cases). One case occurred after a patient with suspected ZES reduced the PPI dose they were taking on their own because they developed constipation while taking the PPI and then presented with a duodenal perforation after 5 d of stopping the PPI in preparation for a secretin test[34]. Two other cases occurred prior to the use of PPIs early in our experience and we have seen no additional cases at National Institutes of Health (NIH) in the last 20 years in acid studies of more than 300 patients with ZES, the majority of who were taking PPIs prior to diagnosis. These latter results demonstrate that acid secretory studies can be safely carried out if proper precautions are taken, even if patients are taking PPIs, although it requires a center well versed in performing these studies, and a proven approach, one of which is discussed below, However, the query raised by Poitras et al[1] still remains, as to whether it is necessary to withdraw antisecretory medications in most patients with ZES to establish the diagnosis.

Unfortunately, the evidence suggests, as concluded in the accompanying editorial[14] to paper No. 1[1], that to clearly establish the diagnosis of ZES, some appropriate assessment of gastric acid acidity/secretion is required after withdraw of PPIs in almost every patient. What is the evidence? First, it is both fortunate and unfortunate that PPIs are so effective in patients with ZES as well as patients with idiopathic peptic disease. It is fortunate because PPIs are very effective at controlling gastric acid secretion in these patients[5,7,36,37]. However, it is unfortunate for diagnostic purposes, because their effectiveness results in that fact that with PPIs, the usual doses used in idiopathic peptic disease are often effective also in ZES and result in hypo/achlorhydria in both ZES patients and in patients without ZES. In contrast with Histamine H₂-receptor antagonists, frequently 10-fold greater doses than used in treating patients with idiopathic peptic disease are required in ZES patients, with more frequent dosing to control the hypersecretion, however these high doses in ZES and the usual doses in patients without ZES, rarely result in hypo/achlorhydria[5,7,36,37]. Therefore, in most patients taking PPIs where ZES is suspected, the gastric pH will not be < 2 (the range required for diagnosis of ZES)[7,11], and therefore physiological and pathological hypergastrinemia can not be distinguished on the drug. Second, there is no feature of the clinical course that unequivocally allows one to establish the presence of ZES, with most patients currently presenting with idiopathic peptic ulcer disease or gastroesophageal reflux disease which is indistinguishable from that seen in non-ZES patients[7,14,35,38]. Third, when ZES is suspected, a fasting gastrin level is almost invariably the first diagnostic study performed[7-10,12,14,39]. Unfortunately, there is no absolute level of fasting hypergastrinemia alone that can distinguish a patient with ZES from a patient without ZES[6,39]. This will be covered in more detail below in the discussion of Rehfeld et al[2], however a few additional points will be made here. In the most common cause of fasting hypergastrinemia, chronic atrophic gastritis, fasting gastrin levels > 70 fold elevated have been reported and levels > 1000 or > 2000 ng/L are not uncommon[40-43], which is a similar finding in patients with pernicious anemia[44]. These values overlap with 80%-100% of patients with ZES in various series[39]. Similarly, in patients taking PPIs without ZES, which is the also one of the most common causes of hypergastrinemia, the PPIs can lead to various degrees of hypergastrinemia in different patients. Although, as pointed out in a number of studies, PPIs frequently lead to < 3 fold increase in fasting gastrin and in some studies do not increase the value out of a normal range[22,45,46], this finding can not be relied on in an individual patient. This conclusion is firmly supported by various studies which report 80%-100% of the patients without ZES in their studies treated with PPIs develop hypergastrinemia, 20%-25% > 4 fold elevated, and values > 1000 pg/mL are not uncommon[24-27,47]. These levels of PPI induced increases in fasting gastrin overlap with that seen in more than 60% of patients with ZES[6,39,48,49]. Fourth, alone, no absolute level of any other tumor marker such as a serum chromogranin A (CgA) level, can establish the diagnosis of ZES. CgA levels are elevated in 90%-100% of patients with ZES, which can be contributed to by both the gastrinoma and the gastrin induced enterochromaffin-like (ECL) cell hyperplasia which is almost always present[50-55]. However, the main problem is that PPIs or high doses of other antisecretory drugs, can increase CgA levels in patients without ZES, within a few days of use, which is thought secondary to the PPI-induced hypergastrinemia causing gastric ECL proliferative effects[27,56,57]. The PPI induced increases in CgA can are usually < 4 fold, but can be up to 40-fold, which overlaps with values seen in > 90% of patients with ZES, as well as seen in numerous, non-ZES conditions[24,27,53,54,57-59]. Fifth, the recommended establishment of a diagnosis of ZES in a hypergastrinemic patient by other methods, as proposed in paper No. 1[1], such as by attempting to establish the presence of a neuroendocrine tumor (primarily by imaging studies) or by establishing the presence of a gastrinoma, is unlikely to be successful in many patients and may lead, in fact, to false diagnoses. It is likely to be unsuccessful in many patients because < 30% of patients with ZES at presentation at the current time, have liver metastases that possibly could be biopsied and the diagnosis of a gastrinoma established by immunohistochemistry[13,60,61]. Even this does not secure the diagnosis, because other pancreatic neuroendocrine tumors [non-ZES primitive neuroectodermal tumors (pNETs)] can stain occasionally positively for gastrin but not be associated with ZES or the portion of the neuroendocrine tumor biopsied may not show gastrin staining[62-69]. Furthermore, localization of a primary gastrinoma, even with the use of increasingly sensitive methods such as somatostatin receptor imaging or endoscopic ultrasound studies, misses most small duodenal tumors, which are present in 60%-80% of patients with ZES[70-75]. Furthermore, cross-sectional imaging studies (computed tomographic scanning, magnetic resonance imaging, trans-abdominal ultrasound) will be negative in > 60% of all patients with duodenal tumors and thus not be able to assist in suggesting the presence of a pNET in the majority of suspected cases, especially those patients being examined early in their course[71,72,76,77]. Lastly, sensitive methods such as somatostatin receptor scintigraphy can lead to false positive localization results and in one prospective study[78,79], 12% of all possible pancreatic endocrine tumor localizations were false positive. These include the presence of an accessory spleen, gallbladder retention, an abscess, various inflammatory processes, inadequate bowel cleansing, thyroid disease, various granulomatous lung diseases, and other neuroendocrine tumors/proliferations such as gastric ECL proliferation or other gastrointestinal-neuroendocrine tumors like gastric carcinoids tumors[78-81]. Therefore, one cannot conclude that localization on a somatostatin receptor scintigraphy study in a patient with hypergastrinemia necessarily equates to localization of a gastrinoma and establishment of the diagnosis of ZES. Fifth, besides the assessment of fasting gastrin levels and gastric fluid pH, various gastrin provocative tests (primarily after secretin, occasionally after glucagon)[14,48,49,82,83] are widely used in the diagnosis of ZES. The secretin test in particular has been well studied and in one recent detailed analysis of 537 patients with ZES in the literature as well as 293 NIH patients with ZES prospectively studied, the secretin test was shown to have a sensitivity of 94% with a specificity of 100% using a criterion of ≥ 120 ng/L increase post secretin[49]. Couldn’t the secretin provocative test be used while the patient is taking PPIs to circumvent the need to stop the PPI? Unfortunately, the answer is no, because a recent study[84] reports a false positive secretin test in a patient without ZES taking PPIs. Furthermore, another study demonstrated if patients are achlorhydric, false secretin positive tests can occur[85].

Commentary on paper

In the study by Rehfeld et al[2], Seven of 12 tested commercial kits inaccurately measure plasma concentrations of gastrin; these assays used antibodies with inappropriate specificity that were insufficiently validated. Misdiagnosis of gastrinoma based on lack of specificity of assays for gastrin results in ineffective or inappropriate therapy for patients with ZES.

Background: An accurate assessment of fasting serum gastrin levels (FSG) is central to the diagnosis of ZES and the diagnosis cannot be made without it. This is especially true, because this is the initial study that leads to the suspicion of ZES, in most cases[2,7,9,12,83]. An assessment of FSG is generally used as the initial study not only because of its convenience and widespread availability, but also because it is elevated in almost all patients with ZES, except for a few specific situations. In a review of 2229 cases of ZES from the literature[39], FSG levels were elevated in 97% of patients and in 309 patients with ZES seen at the NIH it was elevated in 99.3% of patients[39]. These results and others[6,86,87] demonstrate that normal FSG levels are uncommon overall in patients with ZES however, in a few small subgroups with active disease, normal values are not uncommon. This includes patients with active ZES who had had unsuccessful curative resection of a gastrinoma previously, who have ZES with multiple endocrine neoplasia type 1 (MEN1) post successful parathyroidectomy for hyperparathyroidism, in patients being treated with somatostatin analogues or occasionally in patients after various anti-tumor treatments (chemotherapy, chemo-embolization, etc.)[86-96]. Although many physicians think FSG levels are massively elevated in ZES and easy to distinguish from other disorders, unfortunately, as pointed out above, this is not the case. Chronic atrophic gastritis, other hypo/achlorhydric conditions, chronic renal failure and the use of PPIs can cause FSG levels that overlap with those seen in most patients with ZES. Furthermore, in 60%-65% of ZES patients when initially diagnosed, the FSG levels are < 10-fold elevated and these levels overlap with a number of other conditions, some of which are much more frequent than ZES, which can also be associated with gastric acid hypersecretion, such as seen in ZES[5,13,39,48,97]. This FSG level (i.e., > or < 10 × increased) is pointed out because the existing criteria used for the diagnosis of ZES in most guidelines and reviews is divided on the basis of the elevation of FSG[3,12,14,38,39]. If the FSG is > 10-fold elevated (usually > 1000 ng/L) and the gastric fluid pH < 2, a diagnosis of ZES is established after ruling out a possible retained antrum syndrome by history[12,14,38,39,98]. On the other had if the FSG is < 10-fold increased and the gastric fold pH < 2, other conditions need to be excluded including H. pylori infection, antral G cell syndromes, gastric outlet obstruction and renal failure[3,7,9,38,39]. In these cases a gastric analysis with determination of basal acid output, and a secretin provocative test is recommend, or if secretin is not available, a glucagon provocative test has been recommended[7,12,49,82,87]. These latter provocative tests involve the assessment of serum gastrin before and after a secretin or glucagon injection, so an accurate assessment of serum gastrin is essential to their correct interpretation.

Specific comments: Rehfeld et al[2] report that in many cases that the gastrin laboratory assays being used to assess FSG levels are not giving accurate results. Only 5 of the 12 commercial assays examined accurately assessed FSG levels with the others giving FSG values either too high or too low and thus their results could lead to an over- or under-diagnosis of ZES, based on the FSG levels reported[2]. While gastrin provocative test results (secretin, glucagon) were not assessed, a similar result would be expected with these and thus the result would not be dependable in most cases. This report raises a number of problems for physicians trying to diagnose and treat patients with ZES. First, it demonstrates that FSG levels should not be compared from different laboratories using different assays unless some validation is performed. Second and more important, this report raises a real dilemma for the practicing clinician, because it raises the question of whether he can rely on FSG values reported to him by the laboratory he uses. There is no simple solution to this dilemma. The lists of laboratories assessed in this paper[2] can be consulted to see if the one used for the blood samples sent by for the clinician’s patients(s) are on this list. Because the diagnosis of ZES has such significance for any patient and alternative approach as discussed in a section below is to refer the patient to center with known expertise, or to contact them and find out which laboratory in their area they recommend to assess FSG, and confirm the results using this laboratory.

Recommended approach to diagnosis of ZES (based on points raised in papers 1, 2)

First, it is essential to realize that patients with untreated gastric acid hypersecretion with ZES can develop complications rapidly and that this needs to be adequately treated before trying to establish a diagnosis, especially by stopping PPIs. There is no urgency in establishing the diagnosis. Therefore if the patient has active peptic ulcer disease or symptoms and the diagnosis of ZES is suspected, a FSG should be drawn and the acid hypersecretion adequately controlled [our initial starting dose is equivalent to omeprazole 60 qd[99,100], or if complicated disease (presence of MEN1, Billroth 2 surgery, or severe gastroesophageal reflux symptoms), we start with the equivalent of omeprazole 40 bid][37,101-103] and the patient should undergo an upper gastrointestinal endoscopy. We start with a higher PPI dose to make sure the acid hypersecretion is initially well controlled and then later it can be reduced in many patients[37,101]. During this endoscopy, gastric pH can be measured and also the size of gastric mucosal folds noted because 92% of ZES patients have prominent gastric folds[35]. Most patients can be satisfactorily treated by this PPI dose initially, however some require higher doses and therefore it is best to assess the control of acid hypersecretion on PPI[37,99,104], however, only a few specialty centers have this capability, and thus most use control of symptoms to monitor effectiveness of treatment. If the patient has symptoms of gastroesophageal peptic disease or active disease on endoscopy they should be treated for 8-12 wk until symptom free and then the upper gastrointestinal endoscopy repeated to make sure any peptic disease is resolved before attempting to establish the diagnosis by stopping PPIs. During this time the reliability of the FSG assay used needs to be explored both by reviewing the laboratories in paper No. 2[2] and contacting some group well versed in your area with the diagnosis of ZES that uses gastrin assays regularly. If the FSG is elevated it, of course this could be due to the PPI the patient was taking when the blood sample was drawn, which is the situation in most cases, so that the diagnosis remains unclear at this point, particularly if the gastric pH was > 2. At this point many authorities recommend that consideration be given to referring the patient to a group in your area well versed in the diagnosis of ZES[7,9,12,14]. If not possible, then after the repeat endoscopy shows healing of mucosal disease and the patient is asymptomatic, one can consider PPI withdrawal for diagnosis. Our approach is similar to outlined recently[14] and briefly, is to carefully consult with the patient about the need to keep in close contact, then to substitute an H₂ receptor antagonist (usually ranitidine 450-600 every 4 to 6 h) for the PPI for 3-5 d and then stop the ranitidine for 24 h allowing liberal use of antacids. Then on the day of the test we measure FSG × 2, and measure gastric pH and acid output. We usually perform a secretin test at the same time if our index of suspicion is high[49]. This circumvents the need to take the patient off of PPIs again at a later time if the secretin test is deemed necessary. If a gastric/esophageal pH probe is available the assessment of gastric pH and FSG at early times can be done with an attempt to document a pH < 2 with an elevated FSG.

One could ask at this point why go through all of this and why is it so important to establish the diagnosis of ZES correctly in most patients. The primary reason is that the diagnosis of ZES requires special treatment and the treatment must be continued life-long if the patient is not cured[7,9,105]. If the patient is not cured, life-long PPI treatment will be required, the doses may be different than that usually used in patients with idiopathic peptic disease, periodic assessment for the presence of MEN1 will be needed because 20%-30% of ZES patients have this and its diagnosis may not be initially evident and treatment directed at the gastrinoma, which are malignant in 60%-90% of cases, must be considered[13,35,38,106,107]. The latter include periodic assessment of tumor location/extent with imaging studies, consideration of surgical resection which is recommended in ZES patients in whom MEN1 is not present, life expectancy is good and no serious surgical contra-indications are present[7,71,108-110]. Therefore a diagnosis of ZES markedly effects clinical management and thus needs to be well established.