ENETS Consensus Guidelines for the Standards of Care in Neuroendocrine Tumors: Biochemical Markers

Neuroendocrine tumors (NETs) originating from the midgut may result in functional symptoms due to the secretion of various peptides and hormones and most notably 5-hydroxytryptamine (5-HT) or serotonin. This is a tryptophan-derived biogenic amine involved in smooth muscle contraction, blood pressure regulation and both peripheral and central nervous system neurotransmission. Approximately 2% of dietary tryptophan is converted into serotonin. Serotonin is synthesized and stored in enterochromaffin cells of the gastrointestinal tract (80% of total body serotonin content), in dense granules of platelets (storage only) and in the serotoninergic neurons of the central nervous system. The urinary breakdown metabolite of serotonin is 5-hydroxyindole acetic acid (5-HIAA) which is particularly useful in the diagnosis and follow-up of NETs with carcinoid syndrome. Serum measurements of serotonin are possible in these patients; however, large individual variation makes them unreliable for diagnosis and in follow-up. Universally, 5-HIAA is the most frequently performed assay in the clinical setting of the carcinoid syndrome.The overall sensitivity and specificity of urinary 5-HIAA in the presence of the carcinoid syndrome is of the order of 70 and 90%, respectively [1, 2]. Midgut carcinoids are most liable to produce the carcinoid syndrome with 5-HIAA elevation, thus attesting to a high specificity (>90%) in this setting. Fore- and hindgut NETs produce less serotonin than midgut tumors [1, 3]. The sensitivity is lower in patients with midgut carcinoid tumors without the carcinoid syndrome. Urinary 5-HIAA levels may also depend on tumor volume and may be normal in patients with non-metastatic carcinoid tumors. Levels may be normal even in the presence of the carcinoid syndrome, particularly in subjects without diarrhoea; however, this is a rare event [4]. In functional midgut tumors, discriminating performances may vary depending on whether the cut-offs are high or low. Meijer et al. [1] demonstrated that a low level 5-HIAA cut-off value (2.8 mmol/mol creatinine) yielded 68% sensitivity and 89% specificity, whereas a higher cut-off (6.7 mmol/mol creatinine) improved specificity to 98% at the expense of a lower sensitivity (52%). Thus, in order to confidently exclude a carcinoid tumor, a low-level cut-off value may be preferred; to confirm the presence of a carcinoid tumor, a high-level cut-off value is better. Some patients with the carcinoid syndrome excrete non-hydroxylated indole acids, not measured as 5-HIAA. There appears to be an inconstant correlation between 5-HIAA level and the clinical severity of the carcinoid syndrome; this may be related to a fluctuating release of serotonin from tumors such that the correlation may not be reliable. Recent data have examined 5-HIAA as a prognostic factor in these patients: while interesting data have emerged, the expert group felt that data have not confirmed 5-HIAA levels to be a consistently reliable prognosticfactor in this disease. To illustrate this, two studies including 256 and 139 patients with midgut carcinoid tumors showed that while elevated 5-HIAA levels were predictive of poor outcome at univariate analysis, this did not remain significant at multivariate analysis [5, 6]. In another study examining 76 patients, those with persistent moderately increased urinary 5-HIAA levels (≤20 mmol/mol creatinine) had a more favourable outcome compared to those with greatly elevated levels [7]. A further study in a mixed tumor group including 119 patients (53 of midgut origin) interestingly found high 5-HIAA to be an independent survival factor [8].While several assays are available to measure 5-HIAA (thin-layer chromatography, enzyme immunoassay, gas chromatography, gas chromatography-mass spectrometry) [9,10,11,12], the use of high-performance liquid chromatography (HPLC) is most frequently employed. HPLC with electrochemical detection is currently recommended; however, automated assays [13] or those using mass spectrometry [12] may be available in some laboratories. Liquid chromatography tandem mass spectrometry assay appears to be a rapid assay with little necessity for repeat analyses because of chromatographic interference or dilutions [12]. A further automated method with on-line solid-phase extraction and HPLC and fluorometric detection has recently been shown to have increased precision and faster throughput compared to the manual solvent extraction method [14]. Whatever technique is used, it should be performed in accredited laboratories.Conditions for Optimal Assay (fig. 1)Urine should be collected and measured in plastic containers. Acid should be added to ensure sterility and hence stability. The sample should be stored in a refrigerator until analysis. All the urine passed over 24 h should be collected into the container, preferably by using a measuring jug. Collecting should be started at a defined time point following urination, and after that urine should be collected until the same time point the next day (a precise 24-hour collection). Written instructions should be handed out including food and medication precautions (fig. 1).Intraindividual variation of 5-HIAA is also possible and this variation may be high, therefore two consecutive 24-hour collections should be performed and the mean value of these two can be taken, especially when the collection is required for diagnosis. A single specimen may be sufficient for follow-up purposes. Certain co-morbidities or associated disorders may have effects on the concentration of 5-HIAA. Falsely low 5-HIAA levels may be encountered in patients with renal impairment and those on haemodialysis. In addition, 5-HIAA may be increased in untreated patients with malabsorption, who have increased urinary tryptophan metabolites. Such patients include those with gluten-sensitive enteropathy (coeliac disease), tropical sprue, Whipple disease, intestinal stasis and cystic fibrosis (chronic intestinal obstruction) [1, 15]; plasma 5-HT, but not urinary 5-HIAA, have been elevated in diarrhoea-predominant irritable bowel syndrome [16]. A small number of normal individuals may have elevated urinary 5-HIAA and therefore other objective findings should be used in conjunction with tumor marker analysis to support the diagnosis of a carcinoid [17]. The following food substances are rich in dietary tryptophan and therefore patients should abstain from these for 3 days prior to urinary collection: plums, pineapples, bananas, eggplants (aubergines), tomatoes, avocados and walnuts [2, 18, 19]. Certain medications may increase or decrease 5-HIAA urinary levels as follows:– Increased: acetanilide, phenacetin, glyceryl guaiacolate (found in many cough syrups), methocarbamol, and reserpine, cisplatin, fluorouracil, melphalan, rauwolfia.– Decreased: chlorpromazine, heparin, imipramine, isoniazid, levodopa, monoamine oxidase inhibitors, methenamine, methyldopa, phenothiazines, promethazine, and tricyclic antidepressants, chlorophenylalanine, corticotrophin, guanfacine, imipramine, isocarboxazid, isoniazid, levodopa, MAO inhibitors, moclobemide, octreotide.Patients are frequently treated with somatostatin analogues and these are known to decrease levels of 5-HIAA; where possible, assays for diagnostic purposes should be made in patients not on somatostatin analogues, while in the follow-up setting, comparisons should be performed in patients on stable or comparable doses.NETs secreting insulin are termed insulinomas and are almost exclusively intrapancreatic in nature. Excessive insulin secretion leading to hypoglycaemia usually results in a combination of neurologic (diplopia, blurred vision, confusion, abnormal behavior and amnesia, seizures, coma, etc.) and autonomic (sweating, weakness, hunger, tremor, nausea, feelings of warmth, anxiety, palpitations) symptoms. Symptoms are usually related to the degree of insulin-induced hypoglycaemia but may be non-specific. Hypoglycaemia-induced clinical signs are classically present in the early morning preprandial phase or may be exercise-induced. The diagnosis is suggested in the presence of: (1) symptoms of hypoglycaemia; (2) glucose <2.2 mmol/l (40 mg/dl; others use a threshold of <3 mmol/l, 50 mg/dl), and (3) relief of symptoms with administration of glucose [20]. This is known as Whipple’s triad. The 72-hour fast is the gold standard for diagnosing insulinoma and relates to the integrity of patients’ endogenous suppression of insulin in the face of hypoglycaemia. The fast attests to autonomous insulin secretion and the failure of appropriate insulin suppression in the presence of hypoglycaemia. Factitious hypoglycaemia secondary to exogenous use of insulin is suspected on the finding of high (often very high) serum insulin in combination with suppression of C-peptide. Sulphonylureas and related insulin secretagogues result in a clinical picture similar to patients with insulinoma and may be diagnosed by a positive drug screen [21]. An overall approach to diagnosing and managing insulinoma has been provided elsewhere in a recent consensus statement [22].This test has been verified as the gold standard in establishing a biochemical diagnosis of insulinoma [23]. Patients should be hospitalized in a specialist unit experienced in performing the test. A 72-hour period is universally recognized as the most appropriate duration [22] although some groups have proposed a shorter fast of 48 h [24, 25]. Symptoms appear within 12 h for one third of patients, 80% within 24 h, 90% with 48 h and approaching 100% within 72 h [26]. Absolute values of glucose and insulin are the most important variables and any measurable insulin is abnormal when blood glucose drops to 2.5 mol/l (45 mg/dl). Assays used for the determination of insulin, pro-insulin, C-peptide and β-hydroxybutyrate may vary but should be performed in accredited laboratories. Very occasionally, an insulinoma is only revealed by hypoglycaemia induced by a mixed meal rather than fasting.A detailed description of the fast should be provided to all patients with an information card to help in symptom identification. Patients should stay off all foods except for plain water, black tea or coffee and essential medications (particularly hypoglycaemic agents, e.g., sulphonylureas).The timing of the 72-hour fast is not critical – some teams prefer to perform the test early in the week when staffing levels may be higher and avoiding prolonging the test into the weekend. An oral glucose tolerance or mixed meal test can be performed before the fast. The patient should be monitored in a supervised environment and fasting should be accompanied by an intravenous line.– Absolute blood (venous) determinations should be performed at least 2–4 times per day and when the patient describes symptoms. The test interpretation should be made using laboratory blood glucose assays; bedside measurements can be used in the presence of clinical symptoms to determine if more definitive measurements should be made.– Blood should also be drawn for insulin measurement concurrently with glucose estimations, and assay for insulin and C-peptide when the hypoglycaemia is confirmed.– β-Hydroxybutyrate (or urinary ketones) should be measured at the end of the test in order to confirm the validity of the fast. A low level of hydroxybutyrate in the presence of hypoglycaemia confirms inappropriate insulin or insulin-like hormone secretion.– A urinary assay for sulphonylureas should be performed as a specific request.• Not all drugs are detected, e.g. repaglinide [21]; false positive results may also occur, e.g. on paracetamol.• The results need to be confirmed with the local laboratory.The endpoint of the test is documented hypoglycaemia.– Documented blood glucose levels ≤2.2 mmol/l (≤40 mg/dl; according to some <3 nmol/l, 50 mg/dl; levels may depend on age and sex);– Concomitant insulin levels >6 µU/l (≥36 pmol/l; ≥3 µU/l by ICMA);– A β-hydroxybutyrate level ≤2.7 mmol/l can be used as a surrogate marker to confirm the validity of the fast and inappropriate insulin suppression;– A glucagon test immediately after 72-hour fasting in patients without definite results has also been recommended;– Exercise test immediately after 72-hour fasting in patients without definitive results may be performed in a supervised setting;– Use of a ratio of insulin to glucose to aid in the diagnosis is not recommended.The diagnosis of Zollinger-Ellison syndrome (ZES) can be established by the demonstration of elevated fasting serum gastrin (FSG) in the presence of low gastric pH. FSG alone is not adequate to make the diagnosis of ZES because hypergastrinaemia can be seen in patients with achlorhydria associated with chronic atrophic fundus gastritis (e.g., pernicious anaemia) and in other conditions with hyperchlorhydria (e.g., Helicobacter pylori infection, gastric outlet obstruction, renal failure, antral G-cell syndromes, short bowel syndrome, retained antrum). In addition, the use of chronic proton pump inhibitors (PPIs) leads to high FSG levels and therefore gastrin provocative tests are needed to establish the diagnosis of ZES. Indeed, in a recent prospective analysis, up to two-thirds of gastrinoma patients were found to have FSG values <10-fold normal [27]. The gold standard is the secretin test [27,28,29,30,31]. This hormone, when given intravenously provokes an increase in serum gastrin and secondarily in gastric acid secretion. The most reliable data concerning the secretin test have emanated from the National Institute of Health studies in patients with sporadic and multiple endocrine neoplasia type-I (MEN1)-associated gastrinomas [27,28,29,30,31]. Recent consensus guidelines have described the criteria used for establishing the diagnosis of gastrinoma [31]; however, according to the expert committee, acid output studies are available to only a limited number of groups (including those experts’ groups). For the NIH group the secretin test was useful in diagnosing ZES regardless of the extent or locations of the tumor, the presence or absence of MEN1 or the level of FSG (less than or greater than 1,000 pg/ml) [28]. In patients with fasting gastrin <1,000 pg/ml, the sensitivity of the secretin test using the criterion delta (increase from prestimulation level) gastrin of ≥110 pg/ml was 93% (CI 76–99%) and for a delta gastrin of 200 pg/ml sensitivity was 85% (CI 66–96%) (p > 0.05) [28]. The same group recently reported their prospective experience on gastrin provocative tests in 293 patients from the NIH with ZES and compared with 537 ZES patients in the literature and 462 non-ZES patients (again from the literature) [30]. This group established a delta gastrin of ≥120 pg/ml in patients with <10-fold increase as having the highest sensitivity and specificity (94 and 100%, respectively) [30]. They also demonstrated the clear superiority of the secretin provocation test compared to the calcium test (94 vs. 62%). However, in ZES patients with a negative secretin test the calcium provocation test may be helpful [30]. The expert group noted that certain groups had difficulty in obtaining secretin, hindering accurate diagnosis.– The secretin test is performed to confirm a biochemical diagnosis of gastrinoma. The test may be repeated during the follow-up after curative surgery. FSG should be performed prior to secretin test; if FSG >1,000 pg/ml a secretin test is not necessary. When FSG lies between 200 and 1,000 pg/ml, a secretin test should be performed;– The following conditions should also be documented:○ Absence of fundic atrophic gastritis• Antral and fundic biopsies (± serology for anti-parietal and intrinsic factor antibodies);• 24-hour pH-metry (loss of diurnal pH course); basal acid output is recommended pre and post secretin where possible; BAO >15 mmol/h is highly suggestive of diagnosis of ZES; a random pH analysis during upper gastrointestinal endoscopy was also suggested (this requires further evaluation);• Helicobacter pylori testing;○ Other conditions leading to high FSG should be considered, including: gastric outlet obstruction, renal failure, antral G-cell syndromes, short bowel syndrome, retained gastric antrum.– If possible, PPIs should be interrupted 10 days to 2 weeks prior to the test (PPIs for 2 weeks can be replaced by H2 blockers); interruption of H2 blockers for approximately 48 h prior to test, however, interruption of all anti-secretory medications should be individually adapted and patients should be warned of re-apparition of symptoms and should have sufficient anti-secretory medications to start should they become symptomatic; certain patients may have to be hospitalized during anti-secretory therapy withdrawal;– Heparinized vacutainers are used and should be labeled and placed in ice.– Patient fasting overnight, 12–14 h.– Site indwelling i.v. cannula.– Kabi-secretin (2 U/kg body weight) is given by intravenous bolus.– Serum gastrin ○ baseline measured at –15, and –1 min before test;○ 2, 5, 10, 15, 20, and 30 min after secretin.– Samples stored on ice to effects of the secretin test include gastrin at least 200 pg/ml any time during the test is as – The NIH has recently a delta gastrin of ≥120 pg/ml as having a high sensitivity and specificity (94 and 100%, respectively) A is an acid with that is present in the dense granules of most cells The of at least and and other hormones and are is also from tumors including midgut and hindgut gastrointestinal some tumors, and NETs and other amine and tumors. has also been used as an marker in NETs and is recognized as the most has been recognized as a serum as it is in tumors with the and peptides that are present in the granules and can be elevated in both and of elevated is related to tumor and is almost universally elevated in patients with gastrinoma is high in NETs from midgut and in tumor and tumor volume may the level of and interpretation may also depend on the assay used in has been found to be and moderately in the of studies to this was found to be more than in all of a large mixed patient performances have been limited in low-level cut-offs due to the with very high levels of serum are found the setting of NETs with the of patients on gastric acid especially PPIs or those with of in the diagnosis of NETs on the tumor and have been reported in patients with the of the used and the cut-off values was found to be more than high urinary 5-HIAA levels in patients with midgut vs. respectively) et al. demonstrated a significant positive between the serum levels of and the tumor mass in however, the between high and low tumor volume may be to This study also confirmed as high were found in all patients with although small in and tumor volume In a mixed of patients with increased levels were found in and of patients with or disease, respectively the prognostic value of in patients with has not been confirmed to of may in the following renal disease, untreated and or which can to of atrophic gastritis with anti-secretory especially PPIs of gastrin levels provokes of the cells of the and these cells are to In patients with elevated and et al. demonstrated that the can be by without of the A more recently described of was due to the presence of which can to and may be present in up to of the normal in the may be by using a recognized standard for assay is not available and in assay may assays for measurements of and have been The of assays is by the presence of several peptides from and other and according to A can with the use of in the of used vs. and include enzyme detection and in of have also to measurement of in plasma has greater sensitivity for the diagnosis of than the measurement of et al. compared the available in a group of patients and found to vary between and while were for all A recent prospective between two and found a clinical results were with a of variation in a recent study at detection as to and assays A further prospective analysis to be a marker in patients with however, with limited diagnostic using a cut-off of for and for for discriminating between and patients yielded only and respectively) and and respectively) of of of of of of of and of of and and of and of for of of and and of of of and of of and of and of of of Institute of of of and