Search

Since the development of radioimmunoassay for insulin, the diagnosis of insulinoma has been made easily. However, it has been assumed that insulinoma is heterogenous in the histological structure as well as in clinical findings. Therefore, the present study was performed to investigate the insulin response to various stimuli and to evaluate the various insulin response tests in 19 patients with insulinoma. The fasting blood glucose was 19 to 90 mg/100 ml in insulinoma and 81 +/- 5 (mean +/- S.D.) mg/100 ml in normal controls. Plasma insulin (IRI) in insulinoma ranged from 10 to 255 microU/ml, while in the control it was 14 +/- 9 microU/ml. However, insulin/blood glucose ratio increased in insulinoma (0.2-11.2) compared with the normal control (0.18 +/- 0.11). In oral glucose tolerance tests, plasma IRI increased and reached peak levels of 48-244 microU/ml, remaining elevated in most cases. In the intravenous tolbutamide test, plasma IRI increased conspicuously to 82-1,330 microU/ml and hypoglycemic coma was provoked in 54%. Plasma IRI was elevated in the intravenous glucagon test and reached the peak levels of 85-400 microU/ml, which exceeded those of the control group. Plasma IRI increased to more than 100 microU/ml after arginine infusion and formed bizarre curves. There were no correlations between plasma IRI response to various stimuli and malignancy, type of B-granule or insulin content of insulinoma tumors. It is concluded that fasting plasma IRI, insulin/glucose ratio, tolbutamide test and glucagon test are highly valuable for the diagnosis of insulinoma.

Two patients with advanced metastatic insulinoma are presented. Both failed to respond to streptozotocin and were refractory to other forms of chemotherapy. Therapy with adriamycin (Patient No. 1) and adriamycin plus 5-fluorouracil (Patient No. 2) led to significant decreases in insulin levels or measurable disease and both patients had clinical remissions lasting 5 to 7 months. Adriamycin may be effective in patients with advanced insulinoma unresponsive to conventional chemotherapy.

A patient with severe watery diarrhea and a non-beta islet cell carcinoma of the pancreas producing five hormones (secretin, serotonin, enteroglucagon, vasoactive intestinal polypeptide, and pancreatic glucagon) is described. We have demonstrated massive pancreatic hypersecretion to be a major factor in this patient's diarrhea. Possible inter-relationships of the actions of five hormones present in excess in the patient are discussed.

Rats injected with streptozotocin and nicotinamide developed grossly visible islet cell tumors of the pancreas. During i.v. glucose tolerance tests, two populations of tumor-bearing rats were identified: fast responders exhibited significantly lower plasma glucose and markedly elevated plasma immunoreactive insulin (IRI) levels relative to those of the controls. In slow responders, the plasma glucose level was significantly elevated up to 2 h after glucose injection, and the plasma IRI level was lower than that of the controls. During in vitro perfusions with glucose at 300 mg/dl (16.7 mM), tumor-bearing pancreata of fast responders released elevated levels of IRI and immunoreactive somatostatin (IRS); after tumor removal, glucose-stimulated release of these hormones returned to control levels. However, during similar perfusions of pancreata from slow responders, the IRI and IRS release did not decrease after tumor removal, suggesting that the nontumorous pancreatic islets rather than the gross tumors of the slow-responder group were the source of the glucose-stimulated hormone release. These studies demonstrate that gross tumors in the two responder subgroups differ in their glucose-stimulated hormone release.

Calcitonin has been shown to affect calcium handling within cells thus impairing insulin secretion and glucose tolerance in healthy subjects. In the present study we investigate the effects of calcitonin on basal and nutrients-induced plasma glucose and insulin levels variations in healthy subjects (n = 10) and in patients affected by islet cell tumor (n = 6). In healthy subjects calcitonin markedly decreased basal and nutrients-induced plasma insulin levels while in patients with islet cell tumor this calcitonin-mediated effect was lost. So we conclude that the lack of calcitonin effect upon insulin secretion in patients with insulinoma is probably due to the autonomous insulin secretion characterizing islet cell tumor.

The role of prostaglandins in endocrine diarrheagenic syndromes was evaluated by measuring peripheral concentration of immunoreactive PGE and PGF in patients with non-endocrine diarrhea as well as those with the Zollinger-Ellison (Z-E) syndrome, MCT, carcinoid tumors and the WDHA syndrome. In 21 normals, PGE and PGF levels averaged 272 +/- 18 and 119 +/- 14 pg/ml, respectively. Twenty eight patients with diarrhea of non-endocrine origin (mainly inflammatory bowel disease) had levels indistinguishable from normal, i.e. 353 +/- 25 and 77 +/- 37 pg/ml, respectively. Among 29 patients with the Zollinger-Ellison syndrome (mean gastrin 6127 +/- 3267 pg/ml) only 2 had significantly elevated PGE levels; mean PGE levels, 382 +/- 32 pg/ml, were not significantly different from normal and did not correlate with either diarrhea or the serum gastrin concentration. In contrast, 18 of 22 patients with carcinoid tumors (mean blood serotonin concentration 1655 +/- 604 ng/ml; mean urinary excretion of 5 HIAA 66.8 +/- 16.7 mg/day) had elevated peripheral concentrations of PGE. The mean PGE level (1367 +/- 245 pg/ml) was significantly elevated (P less than 0.001). Nonetheless PGE levels did not correlate with diarrhea, blood concentrations of serotonin, or urinary indole excretion. MCT (mean serum calcitonin 24.5 +/- 6.3 ng/ml) was similarly associated with consistent (18/19) elevation in peripheral concentrations of PGE (mean 1922 +/- 541 pg/ml; P less than 0.001). Inthis syndrome, PGE levels were higher in patients with diarrhea and in those with markedly elevated serum thyrocalcitonin levels. Finally, 8 of 21 patients with the WDHA syndrome had increased levels of PGE. Although 13 of 17 patients had high levels of VIP (mean 8133 pg/ml), 2 patients had hyperprostaglandinemia in the face of normal peripheral concentrations of VIP. In one patient the serum PGE level was elevated prior to resection of the primary pancreatic neoplasm (9939 pg/ml) as well as the subsequent extirpation of a solitary hepatic metastasis (1063 pg/ml); following each procedure the diarrhea abated and the PGE level returned to normal. In none of these syndromes were mean PGF levels elevated. The study has documented hyperprostaglandinemia in some endocrine diarrheagenic syndromes and validated the usefullness of measurements of PGE in patients with unexplained diarrhea.

Islet cell tumors produce a spectrum of syndromes. Intensive investigations of these tumors have enhanced our understanding of cellular origin, physiology and biochemistry of the islet hormones. Biochemical studies on the hormones are helpful in the diagnosis and treatment planning of the islet tumors. For example, insulinoma and glucagonoma can be diagnosed more readily by demonstration of proinsulin and proglucagon-like components, respectively, in the blood. Similarly, measurement of vasoactive intestinal polypeptide is not only useful in the diagnosis, but also in the follow-up of patients with pancreatic cholera syndrome. This mini-review examines these and other clinico-biochemical correlates seen in islet tumors.

This paper describes an inhibitory effect of propranolol on insulin secretion in rats with pancreatic islet cell tumors which have been induced by streptozotocin (65 mg/kg body weight) and nicotinamide (500 mg/kg). Following glucose ingestion (3 g/kg), propranolol (4 mg/kg) was injected into the tumor-bearing rats. Plasma insulin decreased paradoxically despite an increase in blood glucoses. In contrast, propranolol did not suppress insulin secretion in normal rats. The drug was found to have no effect on glucagon secretion in either experimental or control animals during glucose load. This may suggest that the experimentally induced insulinoma in hypersensitive to propranolol for inhibiting insulin secretion.

Five cases of somatostatinoma are reported, four being primarily located in the pancreas and one in the duodenum. The diagnosis was based upon the histological and immunochemical characteristics of tumoral and metastatic tissue. A marked clinical heterogeneity was noted: one patient presented with gallstones, steatorrhea, and diabetes, two patients suffered from severe hypoglycemic attacks, and two cases were admitted for obstructive jaundice. This varying symptomatology was related to differences in the circulating levels of biologically active somatostatin and to a variable cellular composition of the tumor. In all cases, a basal and/or tolbutamide-induced hypersomatostatinemia was measured. It is concluded that the clinical and hormonal features of the earlier defined somatostatinoma syndrome are no requisite for the diagnosis of somatostatinoma; the analysis of plasma somatostatin immunoreactivity might lead to a higher detection rate of this endocrine tumor.

The role of human pancreatic-polypeptide in endocrine tumours of the pancreas is reviewed. Pancreatic-polypeptide may be involved in 3 different ways: 1. In cases with pure PP producing tumours. 2. In mixed endocrine tumours containing PP cells. 3. In cases with PP cell hyperplasia in normal pancreatic tissue associated with endocrine pancreatic tumours as VIP-omas, insulinomas, and glucagonomas. PP does not seem to serve as a general marker for endocrine tumours of the pancreas, but PP determinations are useful in patients wbith watery diarrhoea syndromes, because such syndromes may be associated with tumours that contain PP cells. Large molecular forms of PP occur in plasma from patients with endocrine tumours and high PP concentrations, but may also be found in other groups of patients. It is suggested that an atropin-suppression test could be of diagnostic value in revealing patients with increased serum concentrations of PP from other causes than vagal stimulation of normal PP cells.

Plasma pancreastatin (PST)-like immunoreactivity in normal subjects and patients with various diseases was estimated by a RIA, using antiserum raised against a synthetic C-terminal peptide of human PST deduced from the sequence of human chromogranin-A. The mean level +/- SEM was 13.2 +/- 0.6 pmol/L in normal subjects, but was significantly higher in patients with chronic renal failure (526.7 +/- 48.5). An immunoreactive form corresponding to a human PST-like sequence [human chromogranin-A-(250-301)] and a larger form were detected by gel filtration of plasma from these patients, suggesting accumulation of the larger molecular form in these patients. A significant increase in PST-like immunoreactivity was also found in patients with liver cirrhosis (20.8 +/- 3.0 pmol/L), but not in patients with noninsulin-dependent diabetes mellitus, chronic pancreatitis, or pancreatic cancer. Elevated levels were found in 16 of the 21 patients with small cell lung carcinoma examined. High levels were also found in 3 of 11 patients with islet cell tumor.

A newborn infant with severe hypoglycaemia and nesidioblastosis was subjected to subtotal pancreatectomy without any sign of improvement. In spite of very low plasma levels of glucose (i.e. less than 1 mmol/l) plasma insulin concentrations were high (i.e. greater than 700 pmol/l). Plasma proinsulin was considerably enhanced comprising 43% of the total insulin immunoreactivity. Plasma glucagon concentrations were normal. Postoperatively normal to subnormal plasma glucose levels could only be maintained by treatment with frequent meals, diazoxide and intramuscular injections of a long-acting glucagon preparation. With time, signs of mental retardation became obvious.

A synthetic human C-peptide analogue has been used as a common standard for the comparison of insulin C-peptide measurements in seven assay systems in six laboratories. Even in terms of this common standard there was statistically significant numerical heterogeneity between laboratories for estimates of the C-peptide content of the same plasma samples. However, the consistency in ranking order of estimates of C-peptide in the plasma samples between laboratories suggests that laboratories are in most cases measuring at least similar immunoreactive constituents and that a reference plasma might prove useful in comparing results between laboratories. Until a more suitable reference material is available, the synthetic analogue, 64 formyllysine C-peptide, in ampoules coded 76/561, will be made available for research purposes.

It has now been well established that the ulcerogenic syndrome associated with non-beta islet cell tumors of the pancreas is due to excess gastrin release and consequent marked gastric acid hypersecretion. The clinical manifestations may be similar to, but are often more severe and recurrent than, common peptic ulcer. The diagnosis of gastrinoma in patients with this clinical syndrome can be established principally by demonstration of fasting hypergastrinemia, and by application of provocative tests with measurement of serum gastrin in response to intravenous calcium infusion, intravenous secretin injection and feeding of a standard meal. Gastrinomas are usually located within the pancreas, are often multifocal and metastatic, but may be primary in non-pancreatic sites. There is substantial heterogeneity in the molecular forms of circulating and gastrinoma gastrin. Although hypergastrinemia is characteristic of gastrinoma, serum gastrin levels may be increased in disorders other than gastrinoma. Techniques are available to document the presence of gastrin in islet cell tumors and, thereby, to establish these as gastrinomas.

Four patients with metastatic glucagonoma and one patient with metastatic diarrheogenic islet cell carcinoma of the pancreas were treated with dimethyltriazenoimidazole carboxamide (DTIC), 250 mg/M2 daily for five days repeated every four weeks. All patients responded clinically and chemically in one or more ways by a reduction in plasma glucagon levels, improved glucose tolerance, decreased measureable tumor, weight gain, and resolution of necrolytic migratory erythema and diarrhea. This experience and other cases from the literature call for the investigation of DTIC as the initial therapy in metastatic islet cell carcinoma of the pancreas and as being of possible benefit in other tumors of neuroendocrine origin.

The growth and metabolic effects of a transplantable radiation-induced rat insulinoma were examined in intact male and female New England Deaconess Hospital (NEDH) rats, and in parathyroidectomised or adrenalectomised male NEDH rats. Subscapular transplantation of insulinoma fragments in intact male rats consistently produced a highly vascularised encapsulated tumour associated with hyperphagia, hyperinsulinaemia and hypoglycaemia which progressed to fatal neuroglycopaenic coma by 30 +/- 0.8 days (mean +/- SEM) and 19 +/- 0.5 days for slow-growing and fast-growing tumour sublines respectively (P less than 0.001). In intact female rats transplanted with the slow-growing subline, the onset of hyperphagia was advanced by 4 days and the severity of hyperinsulinaemia and hypoglycaemia increased (21% and 36%; P less than 0.01 and P less than 0.001, respectively), resulting in a 10% decrease of survival time (P less than 0.05) and a 65% reduction of tumour weight (P less than 0.01). Transplantation of the fast-growing subline into parathyroidectomised male rats, which exhibited a 15-24% (P less than 0.05 - less than 0.01) decrease of plasma calcium, did not modify either the growth or metabolic effects of the tumour. In contrast, transplantation of this subline into adrenalectomised male rats decreased survival time by 32% (P less than 0.001) and reduced final tumour weight by 88% (P less than 0.02) without markedly affecting the onset or magnitude of the hyperinsulinaemia. These results indicate that the growth and metabolic effects of the transplantable NEDH rat insulinoma are modified by the presence of ovarian hormones and by adrenal hormones.(ABSTRACT TRUNCATED AT 250 WORDS)

A case of watery diarrhea, hypokalemia and hypercalcemia associated with an islet cell tumor was described. A 62-year old man exhibited frequent watery diarrhea and hypokalemia for two years. He had no peptic ulcer and serum gastrin level was normal. His serum calcium was abnormally high and serum phosphate was lowered. He had secretin-like activity in his plasma. Autopsy revealed a small islet cell tumor in the pancreas and several metastatic masses in the liver. Microscopic examination revealed the tumor cell was not beta, alpha nor D cells. By electron microscopy the secretion granules of the tumor cell resembled those of S, M and T cells. It was not possible to decide which of the tree cell types was responsible for the pancreatic cholera.

The serum levels of total immunoreactive insulin (IRI) and proinsulin-like component (PLC) in the fasting state and following the administration of insulin secretagogues in 5 patients with organic hyperinsulinism and age and sex matched normal subjects are reported. Diagnosis of organic hyperinsulinism could be established in all instances on the basis of the inappropriately high total serum IRI levels for the corresponding blood glucose values; such an abnormal relationship was not seen in normal subjects, and was further enhanced by insulin secretagogues. Unrestrained insulin secretion in organic hyperinsulinism was enhanced following the administration of glucose, tolbutamide, glucagon or amino acids; the last 2 stimuli are known to be ineffective in causing insulin secretion in the presence of hypoglycemia in normal subjects. Four patints had insulinomas and one probably had islet cell hyperplasia or abnormal function of islet cells. Chromatography of serum IRI to quantitate PLC is a useful adjunct to the diagnosis of organic hyperinsulinism as in the fasting state the proportion of PLC is always elevated, above the normal range of 5-22%. Following the administration of insulin secretagogues there was pronounced increase in total serum IRI in organic hyperinsulinism but the proportion of PLC generally decreased, suggesting thereby that mojor increase in IRI was due to release of stored granular IRI which is known to have a low proportion of PLC.

An early diagnosis of the Verner-Morrison syndrome will greatly enhance the chances of curative resection. There is a striking need for a simple diagnostic test. A number of suggestions have been made for the presumed hormone mediator of this syndrome. Numerous reports have led to a wide acceptance that vasoactive intestinal peptide (VIP) is the responsible mediator for the pharmacologic actions of this peptide are similar to the physiological characteristics noted in the watery diarrhea syndrome. A raised plasma VIP concentration, on the other hand, would suggest the presence of a tumour. These observations argue for the radioimmunoassay measurement of plasma VIP in a patient with the watery diarrhea syndrome.

Hypoglycemia coexisting with very low plasma immunoreactive insulin concentrations was found in a 78-year-old woman with an insulinoma. However, the absolute proinsulin levels during hypoglycemia were elevated (0.9 to 1.4 ng/ml), accounting for 66.5% of the total circulating immunoreactive insulinlike material. The raised serum proinsulin concentrations together with an abnormal proinsulin:insulin ratio proved to be of critical importance in establishing the diagnosis of an islet-cell tumor in this patient.

In 13 patients with insulinoma, human pancreatic polypeptide (hPP) was measured under basal conditions, after a mixed meal, and during a fast of up to 72 h. Significant negative correlatons between hPP and plasma-glucose levels were found both under basal conditions and during fasting. Patients with insulinoma and basal plasma-glucose values above 60 mg/dl had hPP values similar to those in control subjects, whereas insulinoma patients with plasma-glucose values under 60 mg/dl had significantly higher hPP values (323 versus 115 pg/ml). The mean hPP response to a mixed meal was similar in patients with insulinoma and in control subjects. hPP values during symptomatic hypoglycaemia were higher in insulinoma patients than controls fasted for comparable times (751 versus 171 pg/ml). Under conditions of euglycaemia, hPP is not a reliable marker for insulinoma.

Blood amino acid concentrations were determined in the postabsorptive state in nine patients with insulin excess (functioning insulinomas), nine juvenile-type diabetics with insulin deficiency (diabetic ketosis due to insulin withdrawal), six juvenile diabetics in moderate metabolic control, and five healthy control subjects. Blood branched-chain amino acid (BCAA) levels were elevated in diabetic ketosis and decreased in patients with insulinomas. Blood concentrations of BCAA were significantly correlated to blood glucose levels, and in diabetics they were also correlated to blood ketone bodies, serum free fatty acids, and glycerol levels. These data indicate an inverse relationship between circulating effective insulin levels and blood BCAA concentrations. It is suggested that blood levels of BCAA might represent an indicator of insulin-dependent alterations of protein metabolism.

Nine patients with insulinoma were studied in order to investigate glucagon levels in the fasting state and the response of plasma glucagon to tolbutamide and arginine. Fasting plasma glucagon levels were within the normal range in all patients except two cases with malignant insulinoma. Although there was no correlation between blood glucose and plasma glucagon, a significant correlation between plasma glucagon and plasma insulin was observed. No detectable changes were found in glucagon levels during tolbutamide injection. In almost all patients except one an exaggerated response of plasma glucagon was demonstrated during arginine infusion test.

Based on the literature and on experience and on experience with 23 cases the clinical diagnosis of insulinomas is reviewed. In clinical praxis the diagnostic steps can usually be simplified to a demonstration of fasting, symptomatic hypoglycaemia. Hence, we have used the 72-hour fast as the single diagnostic test for insulinomas during the last 5 years. Pro-insulin measurements may be a valuable supplementary tool. Other diagnostic tests are probably of little importance.

Islet beta cell adenomata were induced in rats by combined treatment with nicotinamide and streptozotocin. Three weeks after treatment marked alterations in glucose tolerance were noted in animals which later exhibited large beta cell tumors. Eight months after treatment, the rats known to have beta cell tumors on the basis of marked hypoglycemia and later confirmed by autopsy showed variable response to a glucose load. Some tumor-bearing rats showed fast response to glucose load, their blood sugar levels were elevated moderately and returned to normal or below normal levels rapidly; these animals are described as having "fast-acting tumors". Rats with "slow-acting tumors" responded sluggishly to a glucose load; their blood glucose pattern was similar to that of subdiabetic animals. Animals with beta cell tumors exhibited elevated serum insulin levels 30 min after glucose administration. Insulin biosynthesis by beta cell adenomata was demonstrated by in vitro incorporation of [14C]leucine into proinsulin and insulin. In the small number of tumor samples studied, a stimulatory effect of glucose on insulin biosynthesis was observed.

Five hypoglycaemic hyperinsulinaemic patients (three with proven benign insulinoma, one with proven metastasizing insulinoma, one with probable insulinoma not found at surgery) were treated with propranolol for a variable time ranging from two weeks to one year. Three patients showed favourable clinical results and a significant increase of the mean basal blood glucose level was found while two patients showed no improvement of the frequency of neuroglycopenic episodes and no significant increase of their mean blood glucose level. No patient showed a significant decrease in mean basal IRI concentration. A decrease of insulinaemic responses was observed during oral and intravenous glucose tolerance tests, a prolonged fast, and tolbutamide and glucagon tests performed in some patients. The results suggest that propranolol may induce in certain patients an improvement of basal clinical status through not understood effects (probably hepatic), which leave the peripheral concentrations of insulin unchanged, whereas inhibition of insulin secretion may represent the main way by which the improvement of metabolic situation during physiological or pharmacological stimulation may have been achieved.

Effects of somatostatin on fasting and arginine-or tolbutamide-stimulated insulin release were studied in four patients with insulinoma. Somatostatin (bolus or bolus + infusion) reduced fasting insulin values in all patients; insulin response to tolbutamide was partially reduced in two patients; somatostatin bolus impaired the insulin response to arginine. Fasting glucagon levels and glucagon response to arginine were also reduced by somatostatin. These results indicate the potential usefulness of somatostatin in the diagnosis of insulinoma even if its effect on insulin is only partial.

During hypoglycemia induced by an infusion of porcine insulin, impaired suppression of endogenous insulin secretion as measured by C-peptide was demonstrated in 11 of 12 patients with insulinoma. During hypoglycemia (plasma glucose less than or equal to 40 mg/dl) the mean C-peptide immunoreactivity (CPR) of normal subjects was less than or equal to 1.2 ng/ml, whereas 11 of 12 insulinoma patients had a mean CPR of larger than or equal to 1.9 ng/ml. One patient showed normal CPR suppression by these criteria but may have shown impaired CPR suppression for glucose less than or equal to 30 mg/dl. Impaired CPR suppression during insulin-induced hypoglycemia may prove to be a useful test for insulinoma.

Immunohistochemical studies of neuron-specific enolase were performed on pancreatic tissues from patients with insulinoma, nonfunctioning islet cell tumor, chronic pancreatitis, and pancreatic adenocarcinoma, and from 5 normal patients. The concentration of neuron-specific enolase was also measured in the sera of patients and in the pancreatic tissue, and the tissues were stained for carbohydrate antigen 19-9 by immunohistochemical techniques. Neuron-specific enolase was localized in nerve fibers, normal islet cells, and islet cell tumors; its concentration was elevated only in the tissue of islet cell tumors and in serum from patients with insulinoma. In the pancreatic tissue of pancreatitis or pancreatic adenocarcinoma, various changes in acini and islets were present. The altered islets stained clearly for neuron-specific enolase and could easily be distinguished from altered, unstained acini in cases of pancreatitis or pancreatic adenocarcinoma. Islets in the pancreatic tissue remained intact with various morphologic changes, although acini had degenerated severely. Carbohydrate antigen 19-9 was localized in all the carcinoma cells in the pancreatic tissue and in some of the normal pancreatic ducts. No cells were simultaneously immunostained by anti-neuron-specific enolase and anti-carbohydrate antigen 19-9 antibodies. Thus, neuron-specific enolase is a good marker for islet cell tumor, and is valuable for examining islets in pancreas with various disorders both alone and in combination with other tumor markers.

A patient with the Watery-Diarrhoea syndrome and episodic hypercalcaemia is reported. Plasma levels of vasoactive intestinal peptide (VIP) were elevated, and an islet cell adenoma of the pancreas was removed following which VIP levels decreased and diarrhoea ceased. During a hypercalcaemic episode, serum parathyroid hormone (PTh) levels were suppressed indicating the hypercalcaemia was independent of PTh and probably due to a direct action of VIP on calcium turnover.

A 58-year-old white woman with known metastatic glucagonoma had widespread necrolytic migratory erythema characteristic of the glucagonoma syndrome. She did not respond to conventional chemotherapy with streptozocin. After one course of dacarbazine therapy, she had remission of the glucagonoma clinically with regression of tumor metastases as defined by liver scanning. After 10 months and additional courses of dacarbazine therapy, she remains in clinical remission. Plasma glucagon levels have decreased although they remain at two to four times the upper limit of normal. On several occasions there was resolution of this patient's rash after intravenous glucose in the absence of supplemental amino acids. We conclude that dacarbazine is an effective mode of chemotherapy for malignant glucagonoma.