Insulinoma resection facilitated by the artificial endocrine pancreas

Intra-operative blood glucose and serum insulin concentrations in the surgical management of insulinoma

The surgical management of insulinoma is frequently complicated by difficulty in pre- and intra-operative localization of the tumour. An early and reliable post-operative indicator of successful surgery would therefore be useful in the management of this condition. In a prospective, controlled study serial measurements of blood glucose and serum insulin concentrations were performed in 5 patients with insulinoma during surgical removal of the tumour. Results were compared with 5 patients, matched for age and body weight, undergoing abdominal surgery for non-malignant disease. Serum insulin (mean +/- s.e.m.) was significantly elevated in the insulinoma patients at the start of surgery (32.1 +/- 3.1 vs 6.1 +/- 2.2 mU/l; P less than 0.01). A significant fall (P less than 0.01) to levels comparable with the control patients (7.5 +/- 1.8 vs 10.0 +/- 3.3 mU/l) occurred following removal of the tumour. We conclude that serial intra-operative measurements of serum insulin concentration are technically simple to perform and provide useful retrospective corroboration of successful surgery in patients with insulinoma.

Glucose turnover in insulinoma--a case report

To define glucose flux in a state of chronic endogenous insulin excess, a patient with an insulinoma was studied. Plasma glucose, insulin (IRI), glucagon (IRG) and glucose turnover ([3-3H]glucose infusion) were measured before and after insulinoma resection in the postabsorptive state (PA), during a glucose infusion adjusted to attain euglycemia (before insulinoma resection only) and following an intravenous glucagon bolus (1 mg). Before insulinoma resection, plasma glucose was 55 mg/dl, glucose production (Ra) and disappearance (Rd) were equal (1.6 mg/kg/min) and glucose clearance was elevated (2.8 ml/kg/min) in PA. When glycemia was raised with a glucose infusion to 77 mg/dl, Rd did not change; in contrast Ra dropped to zero. Plasma IRI and IRG concentrations were 0.7 ng/ml and 110 pg/ml respectively before glucose infusion and remained constant throughout. After resection of the insulinoma, glycemia in PA was 103 mg/dl, Ra and Rd were increased slightly to 1.9 mg/kg/min while the metabolic clearance of glucose was decreased by 25% (2.1 ml/kg/min). Glucagon stimulation pre- and postinsulinoma resection resulted in significant increases in glycemia and IRI. We conclude that hypoglycemia with insulinoma is a consequence of decreased glucose production and increased glucose clearance. Hepatic sensitivity to small increments in glycemia is markedly enhanced so as to fully suppress endogenous glucose production at euglycemic levels in the absence of any change in IRI and IRG. The mechanisms controlling hepatic Ra in insulinoma appear different from normal.

[Diagnosis and surgery of an insulinoma using a glucose-controlled insulin infusion system]

A 71-year-old woman with insulinoma was studied. Preoperatively, using the glucose controlled insulin infusion system (GCIIS) for glucose clamping at various blood glucose levels, autonomous insulin production was demonstrated and intravenous glucose needs for maintenance of normoglycaemia were evaluated. The results of a somatostatin suppression test, guided by the GCIIS, supported the postulation of a well differentiated beta cell adenoma with reduced storage capacity. These assumptions were later confirmed by histochemical and ultrastructural investigations. Hypoglycaemia during surgery was avoided by means of the GCIIS. Upon clamping of the plasma glucose at 90 mg/dl, 15.5 g dextrose had to be given until resection of the tumour. Immediately thereafter, a sharp rise in plasma glucose to 140 mg/dl together with a need for 4.1 U insulin showed that the insulinoma tissue had been removed completely.

To what extent does the artificial pancreas facilitate the surgery of preoperatively not localized insulinomas?

The patient study has been performed in order to evaluate the usefulness of the artificial pancreas in the surgical management of previously not localized insulinomas. In the 4 patients studied, blood glucose was maintained both overnight and during surgery up to a preselected individualized level in order to avoid hypoglycemia. During surgery, only one patient required dextrose infusion. The continuous intrasurgical monitoring of blood glucose in the 4 cases examined showed that: (1) anesthesia induction, surgical incision and viscera mobilization were accompanied by a rise in blood glucose (10.30 mg/dl), reaching the highest levels 30-40 min after the start of the operation; (2) adenoma manipulation was followed by a drop in blood glucose (10.40 mg/dl), reaching the lowest level after 30-40 min; (3) adenoma resection was followed by a rise in blood glucose (25-40 mg/dl), particularly evident after 30-40 min. It is concluded that the artificial pancreas is certainly useful during surgery of insulin-producing tumors, allowing continuous monitoring of glycemia and avoiding dangerous blood glucose excursions: however, when the insulinoma is not identified during surgery, the periods elapsing between the surgical phases and the blood glucose changes observed can be too prolonged to ensure successful conservative serial pancreatectomy in all cases.

Artificial pancreas as an aid during insulinoma resection

Blood glucose control by an artificial pancreas facilitated the intraoperative management of two patients with insulin-secreting pancreatic islet cell tumors (one had a large adenoma, the other diffuse islet-cell hyperplasia). The artificial pancreas is a glucose-controlled insulin and dextrose infusion system uniquely suited for the control of blood glucose concentrations during surgery. The artificial pancreas produced a stable glycemic baseline before surgery, prevented hypoglycemia during palpation of the pancreas and manipulation of the tumor, and indirectly amplified small decreases in the blood sugar level with large increases in the rate of dextrose infusion. The advantages and limitations of continuous glucose monitoring during surgery and the use of this instrument in locating occult insulin-secreting tumors are reviewed.

Glucose-clamp by artificial pancreas in the study and management of a patient with insulinoma

Four euglycemic glucose-clamp studies by artificial pancreas (Biostator, Miles) have been performed during prolonged fast before and after pharmacological treatment in a patient with insulinoma. In the basal state a high glucose infusion rate (8.9 g/h) was unable to achieve the preselected blood glucose plateau of 80 mg/100 ml. The plasma insulin levels during this first glucose-clamp were comprised between 18 and 50 microunits/ml. On the first day of diazoxide treatment (300 mg/die),. the glucose infusion rate decreased to 6.4 g/h, without variation in plasma insulin level, thus suggesting a diazoxide effect independent of the inhibition of insulin secretion. After 7 days of diazoxide treatment, a further reduction of glucose infusion (5.8 g/h), together with a lowering of plasma insulin levels (7-18 microunits/ml) was observed. Both in the basal state and during diazoxide treatment a circadian pattern of glucose requirement was noted, with lower glucose need and plasma insulin levels during the night. Surgery was undertaken with glucose-clamp by artificial pancreas; blood glucose level being higher than the preselected value of 80 mg/100 ml, simple monitoring of glycemia was performed. The excision of a single adenoma was followed by a substantial rise in blood glucose 20 min later. An additional glucose-clamp, performed 3 months after surgery showed a dramatic fall of the glucose infusion rate (2.9 g/h) needed to achieve the preselected blood glucose plateau, confirming the completeness of the intervention. This experience shows that glucose-clamp by artificial pancreas may be of great value in the study as well as in the pharmacological and surgical treatment of patients with insulinoma.

Treatment of diabetes mellitus by devices

A very important aspect of diabetes mellitus is whether or not normalization or near-normalization of blood glucose and/or other metabolites and hormones may reduce or eliminate the chronic complications of this disease. To answer this question and to provide a more "physiologic" approach to insulin administration, a constellation of devices have reached the stage of clinical investigation. These include small portable pump systems that can provide variable rates of insulin infusion via the subcutaneous intravenous or intraperitoneal routes. In addition, bedside artificial "beta cells" having the capability of providing insulin infusions, with the rate varying as a function of continuous glucose measurements, are available for short-term studies. Under development are implantable continuous infusion devices and implantable glucose sensors that could in the future lead to a miniaturized implantable glucose-controlled insulin administration system.

Use of an extracorporeal glucose monitor for the diagnosis and surgical treatment of an insulinoma

An extracorporeal glucose monitor (EGM) has been used in a patient with an insulinoma. This system, which continuously measures blood glucose, can be used in combination with a feedback-controlled glucose infusion to maintain glycaemia at any preselected level. It prevents the risk of severe hypoglycaemia during diagnostic tests (total fast, tolbutamide and insulin tolerance tests) as well as during surgery. It also helps in determining whether all the insulin-secreting tumour tissue has been removed. The EGM is thus helpful in the diagnosis and surgical treatment of insulinomas.

Insulinoma: clinical manifestations, diagnosis and treatment. The significance of the prolonged fasting test and of the fasting blood glucose-insulin relationship

This paper describes the clinical manifestations and the significance of the fasting blood glucose-insulin relationship and the prolonged fasting test in the diagnosis of insulinoma. Pre-and postoperative results are reported. In addition, a description is given of the angiographic demonstration and the results of the surgical removal of the insulinomas.

Feedback-controlled dextrose infusion during surgical management of insulinomas

Through the use of a feedback-controlled dextrose infusion system, we obtained continuous monitoring of the blood glucose level in a patient undergoing surgery for multiple pancreatic beta-cell tumors. In addition, this device infused dextrose at variable rates to maintain a predetermined euglycemic level of 90 mg/dl. Before locating the source of excessive insulin production, the maximum dextrose infusion rate of 400 mg/min was required; but after extirpation of multiple insulinomas, the dextrose-infusion rate declined whereas the blood glucose level rose above the preselected level. These results emphasize the usefulness of a monitoring and infusion system not only in protecting the patient from the hazard of hypoglycemia under anesthesia but also as an aid in determining whether all insulin-secreting tumors have been removed.

Use of a glucose controlled insulin infusion system (artificial beta cell) to control diabetes during surgery

An artificial beta cell has been used to achieve and maintain a preset plasma glucose concentration in five diabetic patients undergoing surgery. These subjects were compared to control groups of normal subjects receiving either saline or glucose, and diabetics receiving glucose intraoperatively. Hyperglycaemia during surgery was seen in normals (mean plasma glucose +/- SEM: 185 +/- 16 mg/dl) and, to a greater degree, diabetics (247 +/- 36 mg/dl) receiving glucose. Insulin and C-peptide levels did not increase during 2 hours of operation in any of the control groups, suggesting beta cell suppression during surgery. As C-peptide levels declined similarly in normal subjects whether they received saline or glucose, the hyperglycaemia seems to be due to an inability to use exogenous glucose. This is confirmed by a correlation of maximal plasma glucose to glucose infusion rate (r = 0.78, p less than 0.01). The artificial beta cell was able to achieve the same plasma glucose after 2 hours of operation (128 +/- 21 mg/dl) as normal subjects receiving saline (110 +/- 7 mg/dl). The artificial beta cell proved to be a safe, convenient and effective way of monitoring and controlling the hyperglycaemia seen in diabetic patients undergoing surgery.

Blood glucose regulation using an open-loop insulin delivery system in pancreatectomized dogs given glucose infusions. I. Portal square waves

This study characterizes the glycaemic and insulin responses of a group of 5 anaesthetized dogs to a portal glucose infusion of 10 mg/kg/min before and after pancreatectomy. Insulin was administered intraportally to the pancreatectomized dogs according to a simple preprogrammed waveform composed of a constant basal rate of 0.35 +/- 0.02 mU/kg/min which was increased to 2.00 mU/kg/min at the time of the 60 minute glucose challenge. When this square waveform was applied the glycaemic response was similar to that seen in the normal controls in the baseline and challenge periods. Blood glucose concentration differed significantly (p less than 0.05) only from 20 to 100 minutes after the end of the challenge when it was higher by 20 +/- 1 mg/dl. Insulin levels were not significantly different from controls. It may be concluded that normoglycaemia and normoinsulinaemia can be maintained by a simple constant rate of portal insulin delivery while the blood glucose response to a glucose infusion can be ostensibly normalized without hyperinsulinaemia simply by enhancing insulin delivery during the challenge. The feasibility of this approach implies that with further development of the preprogrammed waveforms and with a greater understanding of their characteristics portable insulin delivery systems may be realized which accomodate more physiological challenges. The portal route for insulin delivery may however be necessary if peripheral hyperinsulinism is inappropriate.

Continuous blood glucose analysis in vitro and in vivo

A practical method for continuous blood glucose analysis in vivo is described. Using an interference-free enzyme reagent in a modified Auto-Analyzer whole blood glucose concentration can be monitored continuously and interpreted in terms of the actual plasma glucose concentration. The method uses a novel technique for preheating the sample diluent without introducing additional time delays, consumes whole blood at a rate of 0.05 ml/min, and demonstrates a transport delay of 148 s. Other improvements include an average baseline drift of -0.11 mg/dl/h and a mean change in sensitivity of -0.4% after 8.5 h. In vitro glucose recovery studies comparing whole blood to the corresponding plasma samples show the method is precise (101.3 +/- 0.7%), linearly proportional (slope of 1.007 +/- 0.012) and highly correlated (greater than 0.998) over the range of 0 to 500 mg/dl, with reference to a Beckman glucose analyzer. In vivo applications are presented to show that this method is suitable for us in systems such as the 'artificial endocrine pancreas'.