Glycated Hemoglobin
Measurement of glycated hemoglobin is not a standard aspect of the clinical evaluation of hypoglycemia. Concentrations of glycated hemoglobin are statistically significantly lower in patients with insulinomas than in normal persons, but there is too much overlap between the two groups for this test to provide a diagnostic criterion.6
Oral Glucose Tolerance Test
The oral glucose tolerance test should not be used for the evaluation of hypoglycemia, because it is fraught with risk of misdi-agnosis. At least 10% of healthy persons have serum glucose nadirs below 50 mg/dl, and the results of the test do not correlate with serum glucose responses to a mixed meal (i.e., a meal containing a balance of proteins, carbohydrates, and fat).
Mixed-Meal Test
For persons with a history of neuroglycopenic symptoms within 5 hours after food ingestion, a mixed-meal test may be conducted. The test is considered to be positive if the patient experiences neuroglycopenic symptoms when a concomitant serum glucose level measures 50 mg/dl or less. A positive mixed-meal test does not provide a diagnosis, only biochemical confirmation of the history.
C-Peptide Suppression Test
C-peptide is formed during the conversion of proinsulin to insulin by the pancreatic beta cells. In the C-peptide suppression test, the patient fasts overnight and then receives an hour-long intravenous infusion of insulin, during which levels of serum glucose and C-peptide are measured.
Figure 1 During the 72-hour fast, levels of serum glucose are compared with serum levels of insulin (a), C-peptide (b), proinsulin (c), and p-hydroxybutyrate (d). At the end of the fast, 1 mg of glucagon is injected intravenously, and its effect on glucose levels is measured (e). Normal patients may have glucose levels that drop into the hypoglycemic range, so careful documentation of hypoglycemic symptoms is necessary.
In normal patients, hypo-glycemia from the exogenous insulin results in suppression of C-peptide production; patients with insulinomas have higher levels of C-peptide. When the likelihood of a hypoglycemic disorder is not high, a normal result on the C-peptide suppression test may preclude the need for a 72-hour fast. Interpretation of the C-peptide suppression test requires normative data appropriately adjusted for the patient’s body mass index and age.7
Intravenous Tolbutamide Test
In the past, serum glucose response to an intravenous injection of tolbutamide was used in the diagnosis of insulinoma. This test is potentially dangerous and is less accurate than other tests for insulinoma and, therefore, has been rarely used in recent years.
Conditions That Cause Hypoglycemia
The causes of hypoglycemia in healthy-appearing adults encompass the following conditions: insulinoma, factitious hypo-glycemia from insulin or sulfonylurea, noninsulinoma pancre-atogenous hypoglycemia syndrome (NIPHS), and insulin autoimmune hypoglycemia.
Insulinoma
Epidemiology
Between 1927 and 1986, 224 hypoglycemic patients underwent their first pancreatic exploration at the Mayo Clinic and were found to have insulinoma. Because of the relatively large number of cases of insulinoma treated at the Mayo Clinic, in comparison with other medical centers, and the comprehensive epidemiologic database that the Mayo Clinic maintains for Olm-sted County, Minnesota, it was possible to determine the population-based incidence of insulinoma, the risk of recurrence, and the survival in patients with insulinoma.15
The median age of the Mayo Clinic insulinoma patients was 47 years, with a range of 8 to 82 years; 59% were female. The incidence in Olmsted County was 4 cases per 1 million person-years. Of the 224 patients, 7.6% had multiple endocrine neoplasia type I (MEN I) and 5.8% had malignant insulinoma. The risk of recurrence was greater in patients with MEN I (21%) than in those without this condition (7%). Over a 45-year period, overall survival of the total cohort was similar to the expected survival (78% versus 81%).
Insulinomas have been found in pregnant patients and in patients with type 2 (non-insulin-dependent) diabetes mellitus. One case of insulinoma in type 1 (insulin-dependent) diabetes mellitus has been reported.16
Diagnosis
Laboratory tests Insulinoma is characterized by hypogly-cemia caused by elevated levels of endogenous insulin. Confirmation of the diagnosis requires exclusion of hypoglycemia from exogenous sources.
Localization Once a biochemical diagnosis of insulinoma has been made, the next step is localization. Success with the various modalities reflects local skill and experience. Great success has been seen with transabdominal ultrasonography and triple-phase spiral computed tomography. Magnetic resonance imaging and scintigraphy with indium-111 (In-111)-pente-treotide (OctreoScan) can also be used. I reserve endoscopic ul-trasonography for complex cases. Percutaneous transhepatic portal venous sampling has been abandoned even by its former proponents.
In patients whose tumor is not found by ultrasonography or CT, the selective arterial calcium stimulation test provides a means to both regionalize and confirm endogenous hyperinsu-linemia. This test involves serial injections of calcium into the splenic, gastroduodenal, and superior mesenteric arteries. Subsequent doubling of serum insulin concentrations in the right hepatic vein indicates hyperfunctioning beta cells in the part of the pancreas served by that artery.
There is general agreement that the best localization of insulinomas is achieved with intraoperative ultrasonography and careful mobilization and palpation of the pancreas by a surgeon experienced with insulinoma surgery. This approach has seen a 98% success rate in the identification of insulinoma. After this test, these patients go straight to surgery.
Management
The treatment of choice for insulinomas is surgical removal. Depending on the lesion, the surgery required may range from enucleation of the insulinoma to subtotal pancreatectomy. It is advisable for the surgery to be performed at an institution with expertise in the management of insulinoma.
Medical therapy is less effective than tumor resection, but the former can be used in patients who are not candidates for surgery, who refuse surgery, or whose surgery is unsuccessful. The most effective medication for controlling symptomatic hypoglycemia in these patients is diazoxide, which lowers insulin secretion. Diazoxide is given in divided doses of up to 1,200 mg daily. Side effects include edema, which may require high doses of loop diuretics, and hirsutism. Other medications for insulinomas include verapamil, phenytoin, and octreotide.
Factitious hypoglycemia
The term factitious (or factitial) has been used in medical parlance to imply covert patient activity. The consideration of such a possibility often changes the patient-physician relationship, leading the physician to feel deceived and the patient to feel mistrusted. However, the pejorative connotation with which factitious illness has been encumbered requires softening because some patients with factitious disease suffer through no fault of their own.
Epidemiology
Factitious hypoglycemia is more common in women and occurs most often in the third or fourth decade of life. Many of these patients work in health-related occupations.
Factitious hypoglycemia in patients with diabetes is probably more common than the incidence noted in published series.17 Confirmation of the diagnosis in these cases can be very difficult. When deprived of access to hypoglycemic agents, diabetic patients with factitious hypoglycemia become hyperglycemic.
Etiology
Factitious hypoglycemia results from the use of insulin or the use of sulfonylureas or meglitinides that stimulate insulin secretion. The most common form of factitious hypoglycemia is the covert self-administration of a hypoglycemic drug or insulin by a patient without diabetes or the inappropriate manipulation of hy-poglycemic drugs or insulin by a patient with diabetes. Less often, a parent may administer a hypoglycemic agent to a child; this is a form of child abuse.18 In all reported cases, the alleged perpetrator was the patient’s mother, who had ready access to insulin. Insulin has also been used to attempt suicide or homicide.19
There are increasing numbers of patients who, by taking a prescribed medication in good faith, incur hypoglycemia because a sulfonylurea was mistakenly dispensed.20 In most instances, confusion in dispensing the drug arose because of similarity in spelling between the intended medication and the sul-fonylurea. In some cases, however, the dispensing error was a result of negligence. On occasion, cases have arisen in which a nondiabetic person mistakenly takes hypoglycemic medication belonging to another member of the household.
Diagnosis
The possibility of factitious hypoglycemia should be considered in every patient undergoing evaluation for a hypoglycemic disorder, especially when the hypoglycemia has a chaotic occur-rence—that is, when it has no relation at all to meals or fasting. All medications should be identified; the assistance of a pharmacist is desirable. The practice of searching personal effects and labeling insulin with a traceable substance that can be detected in blood or urine is probably unacceptable in the current climate of patients’ rights.
The diagnosis of factitious hypoglycemia can usually be established by measuring serum insulin, sulfonylurea, and C-pep-tide when the patient is hypoglycemic. If a spontaneous episode of hypoglycemia is not observed, the patient should undergo a 72-hour fast. The results of the fast may be negative, however, should the patient not take the offending agent.
In a patient whose hypoglycemia results from covert use of a hypoglycemic agent, the agent will be present in the blood. A sensitive method such as liquid chromatography linked to mass spectroscopy should be used for the detection of sulfonylureas and meglitinides.
In insulin-related factitious hypoglycemia, the serum insulin level is high and the C-peptide level is suppressed, usually being close to the lower limit of detection. This observation applies both to nondiabetic patients and to those with type 2 diabetes. Patients with type 1 diabetes are characteristically severely insulin deficient and have low or undetectable serum concentrations of C-peptide. Although the C-peptide values in these patients cannot be further suppressed, confirmation that the values are low during a hypoglycemic episode eliminates any consideration of endogenous hyperinsulinism.
Management
Treatment of factitious hypoglycemia is simple: the patient stops taking the offending medication. The difficulty involved when medication is taken in error is identification of the drug. In the case of deliberate covert use, psychiatric referral is indicated.
Noninsulinoma pancreatogenous hypoglycemia syndrome
There have been cases of adults who do not have insulinomas but have hypoglycemia resulting from postprandial hypersecre-tion of insulin by pancreatic beta cells. Because of the unique clinical, diagnostic, radiologic, surgical, and histologic features of this disorder, it warrants designation as a new syndrome. We have termed it noninsulinoma pancreatogenous hypoglycemia syndrome, or NIPHS.21
Epidemiology
Like insulinoma, NIPHS affects patients across a broad age range—16 to 78 years, in one series—and causes severe neuro-glycopenia, with loss of consciousness and, in some cases, generalized seizures. Unlike insulinoma, NIPHS occurs predominantly in males (70%).
Pathophysiology and Pathogenesis
Histologic analysis of pancreatic tissue from patients with NIPHS shows cells budding off ducts, which is best seen by chromogranin A and insulin immunohistochemical staining. Islet cell hypertrophy is also evident. No gross or microscopic tumor has been identified on hematoxylin-eosin-stained sections in any NIPHS patients.
Whether islet hypertrophy, nesidioblastosis, or both are pathogenic in these patients is open to question, as is the case with persistent hyperinsulinemic hypoglycemia of infancy (PHHI). However, a role for some form of diffuse islet cell dysfunction appears well established in these cases. Whatever the pathologic process may be, it is nonfocal, yet it does not necessarily involve the entire pancreas uniformly.
The histologic findings in NIPHS are similar to those in PHHI. Although familial forms of PHHI may be associated with mutations in the Kir6.2 and SUR1 genes, analysis of these genes in NIPHS patients has not shown such mutations.22 However, these patients may have common mutations at another, as yet unspecified, locus.
Diagnosis
Clinical manifestations Symptoms of NIPHS occur primarily in the postprandial state 2 to 4 hours after eating. Although insulinoma patients may experience symptoms postprandially, they also have symptoms during food deprivation. It is extremely rare for insulinoma patients to have symptoms solely in the postprandial state.
Laboratory tests Patients with NIPHS have low serum glucose levels and elevated serum insulin levels in the postprandial period. Because of the short half-life of insulin, the criteria for hy-perinsulinemia used in the fasting state appear to apply in the postprandial state, as long as the low glucose level occurs more than 30 minutes from the peak postprandial insulin level. Supervised 72-hour fasts have shown normal results in patients with NIPHS, whereas a negative 72-hour fast in a patient with insuli-noma is a rare occurrence.
The selective arterial calcium stimulation test has shown positive results for patients with NIPHS.23,24 All radiologic localizing studies in patients with NIPHS (transabdominal ultrasonogra-phy, triple-phase CT, celiac axis angiography, and intraoperative ultrasonography) have been negative for insulinoma.
Management
Gradient-guided partial pancreatectomy has been effective in relieving symptoms in patients with NIPHS. The pancreas is resected to the left of the superior mesenteric vein when results of the selective arterial calcium stimulation test are positive only for the splenic artery, and the pancreas is resected to the right of the superior mesenteric vein when the test is positive for an additional artery. Fortunately, gradient-guided debulking of the pancreas can ameliorate the symptoms of NIPHS even in patients whose disease would appear to have involved the whole pancreas. In rats, the mechanism for this effect may be related to decreased insulin secretion, attributed to reduced glucose transporter GLUT2, in remnant pancreas after partial pancreatectomy.25 Unfortunately, recurrence of hypoglycemia after a few symptom-free years has developed in a few of the NIPHS patients.
Insulin autoimmune hypoglycemia
Epidemiology
IAH is an extraordinarily rare disorder that is observed primarily, although not exclusively, in persons of Japanese and Korean ethnicity. The disorder may occur at any age. IAH tends to be be self-limited in Asians, but it may be persistent in whites. There is no gender predilection. Many patients have an ongoing autoimmune disorder or a history of treatment with a sulfhydryl-containing drug such as antithyroid medication. No patients have had a history of exposure to insulin.14
Pathogenesis
IAH is characterized by the presence of autoantibodies to insulin or the insulin receptor. There is speculation that meal in-gestion in these patients may result in the unbinding of insulin from these antibodies. However, measurements of total insulin and free insulin have shown no postprandial alteration in their relative concentrations. The mechanism for the generation of insulin antibodies is unknown but may involve enhanced im-munogenicity resulting from an effect of the disulfide bond in drugs with a sulfhydryl component.
Diagnosis
Clinical manifestations Patients with IAH typically experience postprandial hypoglycemia resulting in neuroglycopenia. The symptomatic severity of IAS appears to vary greatly. Whites may become more seriously debilitated than Asians.
Laboratory tests Serum insulin levels are markedly elevated in IAH, because the insulin antibodies interfere with this assay. Values can be as high as 1,000 ^U/ml. Oddly, C-peptide levels are usually not suppressed. Insulin antibody titers are very high, higher than those seen in insulin-treated diabetic patients. The antibodies may bind only to human insulin or to both human insulin and beef and pork insulin. The antibodies may be poly-clonal or monoclonal, and they usually have characteristics similar to those that occur in patients with type 1 diabetes mellitus. It should be noted that very low titers of insulin antibodies may also be observed in healthy persons without hypoglycemia and occasionally in persons with insulinoma.
Management
Supportive treatment, such as frequent small meals, may be effective in IAH, especially for mild cases. For more severely affected patients, a variety of approaches have been tried, including glucocorticoids, immunosuppressants, plasmapheresis, oc-treotide, and diazoxide. Unfortunately, all these treatments usually fail. Use of partial pancreatectomy and splenectomy has led to amelioration but not complete resolution of symptoms.
