Hypoglycemia in adults without diabetes mellitus: Determining the etiology

INTRODUCTION — Hypoglycemia is an uncommon clinical problem in individuals without diabetes mellitus. It can occur in the fasting or postprandial state, and some individuals experience both fasting and postprandial hypoglycemia. Fulfillment of Whipple's triad supports the presence of pathologic rather than physiologic hypoglycemia.

Whipple's triad comprises the following:

●Symptoms consistent with hypoglycemia (table 1)

●A low plasma glucose concentration measured by laboratory assay when symptoms are present

●Resolution of hypoglycemic symptoms after the plasma glucose level is raised

For individuals with evidence of a hypoglycemic disorder (ie, confirmed fulfillment of Whipple's triad or high clinical suspicion based on home glucose monitoring), a thorough clinical history and physical examination can help identify the most likely cause(s) of the hypoglycemia and inform the next steps in evaluation. Subsequent evaluation depends on whether an insulin- or insulin-like growth factor (IGF)-dependent cause of hypoglycemia is suspected as the underlying etiology. If either etiology is suspected, additional laboratory measurements should be obtained during a spontaneous or provoked episode of hypoglycemia (table 1).

This topic will review the etiologic evaluation of hypoglycemia in individuals without diabetes. An overview of the causes of hypoglycemia and their clinical manifestations are discussed separately (table 2 and table 3), as is the diagnostic evaluation to establish the presence of a hypoglycemic disorder in adults without diabetes. The evaluation and management of hypoglycemia in patients with drug-treated diabetes mellitus are also reviewed separately. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis" and "Hypoglycemia in adults with diabetes mellitus".)

INITIAL ASSESSMENT TO DETERMINE THE ETIOLOGY

History and physical examination — For individuals with evidence of a hypoglycemic disorder (ie, confirmed fulfillment of Whipple's triad or high clinical suspicion based on home glucose monitoring), the patient's overall health status and a thorough clinical history can help identify the most likely cause(s) of hypoglycemia and determine the next steps in evaluation (algorithm 1). (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Causes of hypoglycemia'.)

Although hypoglycemia may be classified in a variety of ways, a clinically useful strategy is based on whether the patient has underlying illness or is seemingly healthy [1,2]. The context in which hypoglycemia occurs (eg, in an inpatient versus outpatient setting), the presence of comorbid conditions, and current medication use all inform this distinction. The major causes of hypoglycemia based on this classification are listed in the table (table 2). Once clinical suspicion for a hypoglycemic disorder is high, a detailed clinical and biochemical evaluation is usually required for healthy individuals, whereas in ill or medicated patients, the cause of hypoglycemia may be readily recognized as a consequence of underlying illness or its treatment [1].

Healthy individuals — In individuals who have no underlying medical conditions, take no medications, and feel otherwise well, the most likely etiologies of hypoglycemia are endogenous hyperinsulinism or factitious hypoglycemia (table 2). (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Causes of hypoglycemia'.)

These individuals should undergo additional, supervised testing to confirm the presence of hypoglycemia (if not previously confirmed with laboratory testing) and distinguish between these potential etiologies. (See 'Supervised testing' below.)

Ill or medicated individuals — In individuals with comorbid conditions who have hypoglycemia, the cause often can be determined based on clinical history alone. Clinical history may indicate a non-insulin-mediated etiology such as cortisol deficiency, critical illness, kidney or liver dysfunction, or a previously diagnosed mesenchymal tumor. In such individuals, initial evaluation should focus on the suspected underlying cause (algorithm 1). Hospitalized patients can present unique diagnostic challenges. In such patients, hypoglycemia is often multifactorial and iatrogenic. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Causes of hypoglycemia'.)

Assess for adrenal insufficiency — Although adrenal insufficiency is not the most common cause of hypoglycemia in ill individuals, this etiology should always be considered first because acute cortisol deficiency can be life-threatening if untreated. Weight loss, weakness, hyponatremia, and fatigue, particularly in the setting of a personal or family history of autoimmunity, suggest adrenal insufficiency as the underlying cause of hypoglycemia. If due to cortisol deficiency, hypoglycemia virtually always will be accompanied by other signs and symptoms supporting this etiology; these are reviewed in detail elsewhere. (See "Clinical manifestations of adrenal insufficiency in adults".)

In patients with suspected adrenal insufficiency, diagnostic testing is the next step in evaluation and is reviewed in detail elsewhere. (See "Diagnosis of adrenal insufficiency in adults".)

In individuals with adrenal insufficiency due to hypopituitarism, concurrent hypothyroidism and/or growth hormone deficiency may further contribute to hypoglycemia. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Illness and comorbid conditions'.)

Assess medications and alcohol use — Medications and excessive alcohol ingestion are common causes of hypoglycemia. Alcohol intake (eg, binge drinking) should be assessed, and all prescription and nonprescription medications should be reviewed in detail. Inspection of pills and bottles is important to identify prescription or dispensing errors. In hospitalized patients, careful examination of the medical and drug administration records is critical. Patients also should be queried about any home medications or supplements that they have continued while hospitalized. Medications other than those used to treat diabetes can cause hypoglycemia (table 3). (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Medications and alcohol' and "Screening for unhealthy use of alcohol and other drugs in primary care", section on 'Screening tests'.)

If medication- or alcohol-related hypoglycemia is possible based on clinical history, the suspected agent should be withdrawn. If the hypoglycemia resolves, this etiology is confirmed, and no additional evaluation is needed. If hypoglycemia persists, additional evaluation is warranted and reviewed immediately below.

Assess for alternative causes

Insulin- or insulin-like growth factor-mediated causes — Insulin- and insulin-like growth factor (IGF)-mediated causes of hypoglycemia are uncommon. The presence of any of the following suggests the possibility of an insulin- or IGF-mediated cause, and the patient should undergo supervised testing to further assess the etiology (see 'Supervised testing' below):

●History of bariatric surgery – Patients who have undergone gastric bypass surgery may have underlying postprandial hyperinsulinemic hypoglycemia. This etiology is supported by a history of predominantly postprandial hypoglycemia. If post-gastric bypass patients report hypoglycemia after high-carbohydrate meals, the first step in evaluation should be dietary modification. (See "Bariatric operations: Late complications with subacute presentations", section on 'Postprandial hyperinsulinemic hypoglycemia'.)

●Access to insulin or other glucose-lowering agents – Patients should be queried about any household members or other close contacts who are receiving medical therapy for diabetes. For patients without a personal history of diabetes, the presence of a household member with diabetes raises the possibility of access to medications that can cause hypoglycemia. (See "Factitious hypoglycemia", section on 'Etiology of factitious hypoglycemia'.)

●Recurrent, severe episodes of neuroglycopenic symptoms – Recurrent neuroglycopenic symptoms suggest underlying insulinoma, and patients may report a progressive increase in the frequency and severity of symptoms over time. Patients with insulinoma also may report weight gain due to the need for frequent eating to prevent or treat hypoglycemic episodes. (See "Insulinoma", section on 'Clinical features'.)

●History of malignancy – A history of malignancy suggests nonislet cell tumor hypoglycemia, a paraneoplastic syndrome that usually arises from tumor production of IGF-2. (See "Nonislet cell tumor hypoglycemia".)

Critical illness or organ dysfunction — Sepsis, other critical illness, or acute or chronic kidney or liver dysfunction can be the primary etiology of hypoglycemia or play a contributory role. However, these causes of hypoglycemia are a diagnosis of exclusion.

A careful clinical history should be obtained to verify that hypoglycemia developed in parallel with the illness or kidney or liver impairment. The concurrent evolution of hypoglycemia with either illness or organ dysfunction corroborates this etiology. If hypoglycemia developed before the onset of critical illness or organ dysfunction or if uncertainty otherwise persists, supervised testing to assess for other causes of hypoglycemia is usually warranted. (See 'Supervised testing' below.)

Etiology not apparent — If the clinical assessment does not identify an underlying cause of hypoglycemia, supervised testing to evaluate for insulin- or IGF-mediated causes of hypoglycemia should be performed. (See 'Supervised testing' below.)

SUPERVISED TESTING

Importance of supervised evaluation — Supervised testing (supervised fast or mixed meal test) should be performed in individuals with evidence of a hypoglycemic disorder in whom an insulin- or insulin-like growth factor (IGF)-mediated cause is suspected or those in whom the underlying cause is uncertain. This evaluation should only be performed in individuals in whom clinical suspicion is high for an underlying hypoglycemic disorder (ie, those with evidence of all three criteria for Whipple's triad or clinical history and fingerstick blood glucose measurements that raise suspicion for pathologic hypoglycemia). Supervised testing, including the decision to proceed with such testing, should be performed by an endocrinologist. The diagnostic evaluation for a hypoglycemic disorder must precede supervised testing and is reviewed separately. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Diagnostic evaluation'.)

Whipple's triad comprises the following:

●Symptoms consistent with hypoglycemia (table 1)

●A low plasma glucose concentration measured by laboratory assay when symptoms are present

●Resolution of hypoglycemic symptoms after the plasma glucose level is raised

When clinical suspicion for a hypoglycemic disorder is based on home glucose monitoring, the purpose of supervised testing is to document Whipple's triad and establish the mechanism of hypoglycemia [3]. If Whipple's triad has already been established through laboratory glucose measurement, then a supervised test may be necessary to establish the mechanism of hypoglycemia. Laboratory tests obtained during hypoglycemia specifically assess the etiologic role of insulin or an IGF. Thus, supervised testing should be performed only if these are the most likely etiologies of hypoglycemia or no specific etiology can be discerned from the clinical history and physical examination. (See 'Initial assessment to determine the etiology' above.)

Choice of test

●Testing during unprovoked hypoglycemia – Occasionally, supervised testing may be performed during a spontaneous episode of hypoglycemia that occurs under medical supervision. This scenario is usually limited to the inpatient setting. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Medications and alcohol'.)

Patients who are fortuitously observed during a symptomatic episode and found to have concurrent hypoglycemia should have the following blood tests performed at the time of hypoglycemia:

•Glucose level

•Insulin level

•C-peptide level

•Beta-hydroxybutyrate (BHB) level

•Proinsulin level

•Oral hypoglycemia agent (sulfonylureas and meglitinides) screen

(Related Lab Interpretation Monograph(s): "Low glucose in adults".)

These laboratory tests can confirm the presence of a hypoglycemic disorder and help establish whether the hypoglycemia is mediated by insulin or an IGF. The interpretation of test results is the same for spontaneous hypoglycemia and hypoglycemia provoked through a supervised fast. (See 'Interpretation of supervised fast results' below.)

●Testing to provoke hypoglycemia – If a patient is not evaluated during a spontaneous episode of symptomatic hypoglycemia, provocative testing is required. The choice of provocative test depends on both clinical history and logistical considerations (algorithm 2). If symptoms occur primarily in the fasting state, testing should be performed during a supervised fast. If the patient has a history of exclusively postprandial symptoms, a mixed meal test should be performed. The choice of appropriate test is important, as patients with hypoglycemic disorders that cause only postprandial hypoglycemia can have normal findings on a 72-hour fast. If the timing of hypoglycemia in relation to meals is uncertain, we proceed with a supervised fast. (See 'Supervised fast for fasting hypoglycemia' below and 'Mixed meal test for postprandial hypoglycemia' below.)

MIXED MEAL TEST FOR POSTPRANDIAL HYPOGLYCEMIA — For individuals who experience exclusively postprandial hypoglycemia (ie, symptoms that occur within five hours after eating), the mixed meal test should be performed for further diagnostic evaluation.

Mixed meal test protocol — For a mixed meal test, the patient consumes a non-liquid meal and is then observed for up to five hours [3]. The macronutrient content of the meal should mimic that which usually leads to symptoms in the individual patient. However, patients who have undergone Roux-en-Y gastric bypass (RYGB) should have a test meal that is appropriate for post-gastric bypass patients; they should avoid high carbohydrate intake and caloric intake in liquid form [4,5]. (See "Bariatric operations: Late complications with subacute presentations", section on 'Dumping syndrome'.)

Protocol — Institutions have different protocols for the collection of venipuncture samples during a mixed meal test. In some centers, venous blood samples are collected prior to ingestion of the meal and once the patient's typical hypoglycemic symptoms develop. In other institutions, samples are collected prior to ingestion of the meal and every 30 minutes thereafter until one of the test endpoints is met. With either protocol, venous blood samples are collected by venipuncture for measurement of glucose, insulin, and C-peptide levels at each time point.

Test endpoints — The mixed meal test is terminated when either of the following endpoints is met:

●Patient develops typical hypoglycemic symptoms

●Five hours elapse

When either endpoint is met, a final blood collection should be performed with an additional sample drawn for an oral hypoglycemic agent (sulfonylureas and meglitinides) screen. After final blood samples are collected, the patient should be given food (eg, approximately 15 g of rapidly available carbohydrate). Carbohydrate administration will both correct hypoglycemia if present and verify whether symptoms resolve with reversal of hypoglycemia (the third criterion of Whipple's triad). Another plasma glucose level should be obtained by venipuncture once symptoms improve to verify reversal of hypoglycemia.

Blood samples from each time point are sent for measurement of glucose concentrations. The final samples collected for insulin and C-peptide levels and the oral hypoglycemic agent screen should be analyzed only if collected when plasma glucose is <65 mg/dL (3.3 mmol/L) and the patient has hypoglycemic symptoms. Some UpToDate contributors analyze the samples for insulin and C-peptide for each time point at which glucose is <65 mg/dL (3.3 mmol/L), whereas other contributors analyze these samples only once at the glucose nadir. The interpretation of laboratory results is discussed below. (See 'Interpretation of mixed meal test results' below.)

●No symptoms elicited during the test – If symptoms do not occur during the mixed meal test, the patient can be provided with additional laboratory requisitions for glucose and insulin measurements to be performed at the time of symptoms. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Assessment in symptomatic individuals'.)

In post-gastric bypass patients with prior hypoglycemic symptoms who do not develop symptoms during the supervised test, dietary modification (ie, consuming foods similar to those used in the mixed meal test) may be sufficient to alleviate symptoms. (See "Bariatric operations: Late complications with subacute presentations", section on 'Dumping syndrome'.)

●No diagnostic role for oral glucose tolerance test – An oral glucose tolerance test (OGTT) or a meal that is solely in liquid form (eg, Ensure, Boost) should not be used in the evaluation of postprandial hypoglycemia because misleading results may be obtained. These tests are particularly likely to produce erroneous results in patients with prior upper gastrointestinal surgery and disordered gastric emptying [6]. (See "Evaluation of postprandial symptoms of hypoglycemia in adults without diabetes", section on 'Postprandial syndrome'.)

Interpretation of mixed meal test results — Standards for the interpretation of the mixed meal test are not well established, and the interpretation of test results obtained during a supervised fast are generally applied to the mixed meal text. However, the interpretation of insulin levels in the postprandial state is more difficult than in the fasting state as circulating concentrations represent the net sum of both insulin secretion and clearance. Insulin has a half-life of approximately 5 minutes in individuals with normal kidney function. Therefore, postprandial insulin levels may reflect the clearance rate rather than the rate of secretion in response to nutrient ingestion. Nevertheless, irrespective of clearance rates, an appropriately suppressed insulin level (<3 microU/mL [20.8 pmol/L]) can exclude hyperinsulinism as the cause of hypoglycemia. (See 'Insulin level <3 microU/mL' below.)

The interpretation of laboratory data obtained during hypoglycemic symptoms is as follows:

●Glucose ≥65 mg/dL – If glucose is ≥65 mg/dL (3.6 mmol/L) at the time of symptoms, a hypoglycemic disorder is unlikely. Other causes of postprandial symptoms should be investigated. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Laboratory measurement (venipuncture) thresholds'.)

●Glucose <65 mg/dL – If glucose is <65 mg/dL (3.6 mmol/L) at the time of symptoms, and symptoms improved when the plasma glucose was raised, a hypoglycemic disorder is likely. A suppressed insulin level (<3 microU/mL [20.8 pmol/L]) suggests non-insulin-mediated causes of hypoglycemia (eg, alcohol, medications [other than antihyperglycemic agents], toxins, malnourishment, cortisol deficiency). The differential etiology also includes surreptitious administration of an insulin analog that is not detected by the insulin assay. (See 'Insulin level <3 microU/mL' below.)

A nonsuppressed insulin level (≥3 microU/mL [20.8 pmol/L]) suggests insulin-mediated causes of hypoglycemia (eg, postprandial hyperinsulinemic hypoglycemia, factitious hypoglycemia [surreptitious use of insulin, sulfonylureas, or meglitinides], insulinoma). The C-peptide level enables differentiation between endogenous and exogenous hyperinsulinism. The evaluation of insulin-mediated causes of hypoglycemia is largely the same for the mixed meal test and supervised fast and is discussed below. (See 'Insulin level ≥3 microU/mL' below.)

SUPERVISED FAST FOR FASTING HYPOGLYCEMIA — Healthy individuals do not have neuroglycopenic symptoms after a prolonged fast because of hormonally mediated increases in glucose production and/or lipolysis and ketone body production. During fasting, the energy substrate for organ function partially shifts from glucose to free fatty acids. Gluconeogenesis accounts for approximately 50 percent of glucose production after an overnight fast and for almost all glucose production after ≥42 hours of fasting [7]. A prolonged fast will result in hypoglycemia only if the ability to maintain normoglycemia is compromised due, for example, to an excess of insulin, which inhibits both endogenous glucose production and ketogenesis.

The purpose of the supervised fast is to provoke the homeostatic responses that prevent glucose from falling to concentrations that cause hypoglycemic symptoms. Increased release of glucagon, epinephrine, and, to a lesser degree, growth hormone and cortisol are the most important components of this response. These hormonal responses begin well before the onset of symptomatic hypoglycemia. (See "Physiologic response to hypoglycemia in healthy individuals and patients with diabetes mellitus".)

Choice of setting — The choice of setting depends on patient characteristics and institutional resources. For some patients who report consistent symptoms after a short period of food withdrawal (eg, 8 to 12 hours), we initially choose the outpatient setting for the supervised fast. Prolonged observation in the office or clinic after an overnight fast may result in an episode of symptomatic hypoglycemia [3] and avoids the logistical challenges and inconvenience of an inpatient fast. If this approach fails to elicit hypoglycemia but clinical suspicion remains high, the patient should undergo a supervised 72-hour fast in the inpatient setting. For other patients, particularly those at risk for hypoglycemia-related morbidity (eg, older adults, those with comorbidities), we initially proceed with the inpatient 72-hour fast. The fast should be performed only under the supervision of experienced personnel. A detailed approach to the supervised fast is reviewed below. (See 'Protocol' below.)

Maximal duration of the fast — The outpatient fast typically lasts up to 20 to 24 hours (depending on when the patient initiates the overnight fast and the duration of clinic hours), whereas the inpatient fast lasts up to 72 hours. In either setting, the fast is terminated prior to maximal duration if any of the test endpoints is met. (See 'Test endpoints and completing the fast' below.)

Fasting for a maximum of 72 hours has been the standard test for the diagnosis of insulinoma. A 48-hour maximum has been proposed as a shorter alternative based on a report of 127 patients with insulinoma, all of whom had screening glucose values <45 mg/dL (2.5 mmol/L), in which adequate diagnostic information was obtained in all patients within 48 hours [8]. The fast was terminated because of neuroglycopenic symptoms and biochemical hypoglycemia (plasma glucose ≤45 mg/dL [2.5 mmol/L]) in 43 percent of patients by 12 hours, 67 percent by 24 hours, and 95 percent by 48 hours. The rationale for extending a fast beyond 48 hours arises from a series of 205 patients with surgically confirmed insulinoma, in which 7 percent of patients who underwent an extended supervised fast did not develop hypoglycemia with neuroglycopenic symptoms within 48 hours of fasting [9].

Protocol — The following protocol is similar to that used at the author's institution (Mayo Clinic, Rochester, MN) [10]. The fast can be initiated at home, usually after the evening meal, and continued the next day in a supervised setting (ie, hospital or outpatient clinic).

A careful record should be kept of all procedures, collected blood samples, plans for laboratory analyses, glucose values, and any symptoms experienced by the patient. Blood samples and laboratory slips must be carefully labeled, particularly with the exact time of collection, and sample labeling information should also be recorded on a flow sheet. The subsequent interpretation of laboratory results will require these details. (See 'Interpretation of supervised fast results' below.)

Procedures during the fast

●Prior to arrival at the hospital or clinic – The patient should be instructed to begin fasting by 8 PM the night prior to arrival at the hospital or clinic. After the fast begins, the patient may consume only beverages that are calorie- and caffeine-free. Nonessential medications or supplements should be discontinued.

●Upon arrival to the hospital or clinic – The following steps should be performed upon the patient's arrival to the hospital or clinic:

•The date and time of initiation of the fast should be carefully recorded, as should any medications, supplements, or beverages the patient has consumed since the fast began.

•Blood sample collection tubes should be labeled and maintained close at hand. The correct procedures for handling and processing the samples should be determined in advance.

•Glucagon (1 mg for intravenous injection) should be ordered and available for administration at the termination of the fast. (See 'Glucagon challenge to corroborate the etiology' below.)

●During the supervised fast – During the fast, the patient may continue to consume beverages that are calorie- and caffeine-free. The patient should remain moderately active during waking hours (eg, ambulating to the restroom or through the hallway).

•Frequency of glucose measurement – Glucose should be measured by either laboratory assay or fingerstick and glucose meter every two hours and any time the patient experiences symptoms suggestive of hypoglycemia. Although glucose meters are less accurate than laboratory measurement of venous glucose, delays in the availability of laboratory results often limit the utility of laboratory measurements when frequent samples are obtained. If laboratory glucose measurements can be obtained without delay, these are preferable to values obtained through fingerstick and meter.

Once the glucose concentration falls below 70 mg/dL (3.9 mmol/L), glucose measurement and monitoring for neuroglycopenic symptoms should be performed every 30 minutes. If glucose is measured through laboratory assay, each blood sample should be sent to the laboratory immediately after collection.

•Additional blood sample collection – Once the glucose level is <70 mg/dL (3.9 mmol/L), additional blood samples should be collected with each glucose measurement for insulin, C-peptide, proinsulin, and beta-hydroxybutyrate (BHB) concentrations. These additional samples are not usually sent for laboratory analysis until one of the test endpoints is met. However, some UpToDate contributors analyze all samples collected with each glucose measurement. (See 'Test endpoints and completing the fast' below.)

If the patient experiences neuroglycopenic symptoms but the glucose is >70 mg/dL (3.9 mmol/L) by fingerstick and meter check, a venous glucose concentration should be measured by laboratory assay and additional blood collected for measurement of insulin, C-peptide, proinsulin, and BHB concentrations. The fast should continue with serial blood sample collection until a test endpoint is met.

•Symptom monitoring – Careful questioning and testing for subtle symptoms or signs of neuroglycopenia should be conducted repeatedly when the glucose is near or in the hypoglycemic range. Simple tests of cognitive function may be indicated (eg, basic arithmetic questions, counting backwards from 100). Ending the fast in the absence of symptoms or signs of hypoglycemia jeopardizes the ability to discriminate between physiologic and pathologic hypoglycemia. A low plasma glucose is a necessary, but not sufficient, finding for the diagnosis of hypoglycemia.

Test endpoints and completing the fast

●Test endpoints – The fast is ended when one of the following occurs:

•For an inpatient fast, 72 hours have elapsed; for an outpatient fast, clinic hours have ended.

•Signs or symptoms of hypoglycemia are present, and plasma glucose by laboratory assay is <55 mg/dL (3 mmol/L) [1,3].

•Signs or symptoms of hypoglycemia remain absent, and plasma glucose by laboratory assay is ≤45 mg/dL (2.5 mmol/L).

For patients who do not develop hypoglycemic symptoms during a supervised fast, a lower glucose threshold for terminating the test is used because glucose values in the hypoglycemic range (eg, 40 to <60 mg/dL [2.2 to 3.3 mmol/L]) can reflect either normal physiology or underlying pathology [11]. Plasma glucose concentrations fall below 50 mg/dL (2.8 mmol/L) in some healthy individuals and remain >50 mg/dL in occasional patients with insulinoma [8-10].

●Completing the fast – Three steps are performed at the end of the fast:

•Collection of blood samples for glucose, insulin, C-peptide, proinsulin, and BHB concentrations and an oral hypoglycemic agent (sulfonylureas and meglitinides) screen. The venous glucose concentration should be measured immediately, and the remaining tests should be sent only if a sufficiently low glucose value (ie, <55 mg/dL [3 mmol/L]) is verified. (See 'Interpretation of supervised fast results' below and 'Test endpoints not achieved during the supervised fast' below.)

•Administration of glucagon 1 mg intravenously with measurement of the plasma glucose concentration 10, 20, and 30 minutes after glucagon administration. The results of the glucagon challenge are used to help distinguish between insulin- or insulin-like growth factor (IGF)-mediated causes and other etiologies of hypoglycemia. (See 'Glucagon challenge to corroborate the etiology' below.)

•Provision of food to the patient. The patient should be monitored carefully for the reversal of hypoglycemic symptoms with food consumption. Reversal of symptoms with correction of hypoglycemia meets the third criterion of Whipple's triad. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Criteria for a hypoglycemic disorder'.)

Interpretation of supervised fast results — Interpretation of the data obtained at the end of a supervised fast is largely the same as that performed for a spontaneous hypoglycemic episode or a mixed meal test (table 4 and algorithm 3) [10,12]. These laboratory data will help distinguish hyperinsulinism (endogenous or exogenous) from other causes of hypoglycemia but will not identify all potential causes. Thus, additional testing may be required to verify the etiology.

Laboratory data are reliably interpretable only if plasma glucose is <55 mg/dL (3 mmol/L) at the end of the fast. The corresponding insulin concentration helps determine whether the low glucose resulted from hyperinsulinism or another cause. When plasma glucose is in the hypoglycemic range, the normal fasting reference ranges for plasma insulin, C-peptide, proinsulin, and BHB concentrations cannot be applied; rather, the etiologic evaluation of hypoglycemia requires unique diagnostic thresholds for each of these analytes. The interpretation of these laboratory results is shown in the table and algorithm (table 4 and algorithm 3) and described immediately below. The recommended diagnostic criteria for plasma insulin, C-peptide, proinsulin, and BHB levels, described below, have sensitivities and specificities of >90 and >70 percent, respectively, for the diagnosis of insulinoma [13]. The occurrence of hypoglycemia with associated symptoms and signs, in conjunction with laboratory measurements, usually enables distinction among the various causes of hypoglycemia (table 2) [10].

Insulin level ≥3 microU/mL — In patients with a low venous glucose concentration (ie, <55 mg/dL [3.0 mmol/L]), a plasma insulin concentration ≥3 microU/mL (20.8 pmol/L; measured by immunochemiluminometric assay [ICMA]) indicates insulin-mediated hypoglycemia. The absolute insulin value should be used to determine whether the hypoglycemia is insulin mediated, as ratios of insulin-to-glucose or of glucose-to-insulin are not helpful for establishing hyperinsulinemia [14]. When the insulin level is unsuppressed, a rise in glucose level >25 mg/dL (1.4 mmol/L) after glucagon challenge further supports an insulin-mediated etiology of hypoglycemia. (See 'Glucagon challenge to corroborate the etiology' below.)

The plasma C-peptide level distinguishes endogenous from exogenous hyperinsulinemia, and measurement of the plasma proinsulin level can help corroborate this distinction (figure 1) [15-17].

C-peptide level <0.2 mmol/L — In patients with an insulin-mediated cause of hypoglycemia, a C-peptide level <0.2 mmol/L (0.6 ng/mL) suggests hypoglycemia due to exogenous insulin administration. A plasma proinsulin level <5 pmol/L also supports an etiology of exogenous hyperinsulinism. These C-peptide and proinsulin values exclude an endogenous origin of hyperinsulinism.

Exogenous hyperinsulinism is caused by factitious hypoglycemia (eg, accidental or surreptitious insulin administration). The evaluation of factitious hypoglycemia is reviewed separately. (See "Factitious hypoglycemia", section on 'Laboratory tests'.)

C-peptide level ≥0.2 mmol/L — Patients with insulin-mediated hypoglycemia and a plasma C-peptide level ≥0.2 mmol/L (0.6 ng/mL) have an endogenous origin of hyperinsulinism. The corresponding threshold for plasma proinsulin that is diagnostic of endogenous hyperinsulinism is ≥5 pmol/L (figure 2).

In patients in whom the plasma glucose fell below 45 mg/dL (2.5 mmol/L) during the supervised fast, a threshold C-peptide concentration of ≥0.2 nmol/L (0.6 ng/mL) distinguished patients with insulinoma from healthy individuals without any overlap in C-peptide levels between groups; all patients with insulinoma had values above this threshold, and all healthy individuals with glucose values in the hypoglycemic range had values below this threshold [15].

The differential etiology of endogenous hyperinsulinism includes insulinoma, oral hypoglycemic agent-induced hypoglycemia, insulin autoimmune hypoglycemia, and nesidioblastosis/islet cell hypertrophy (table 2). The results of the oral hypoglycemic agent screen will confirm or exclude this potential etiology.

●Positive oral hypoglycemic agent screen – A positive oral hypoglycemic agent screen confirms the etiology of medication-induced hypoglycemia. Sulfonylureas or meglitinides are present in the plasma only in oral hypoglycemic agent-induced hypoglycemia [18]. This cause of hypoglycemia may be due to medication error or surreptitious medication administration and warrants careful inspection of all the patient's prescriptions, pill bottles, and tablets, as well as inquiry regarding any access to oral hypoglycemic medications. (See "Factitious hypoglycemia".)

●Negative oral hypoglycemic agent screen – In patients with hypoglycemia due to endogenous hyperinsulinism who have a negative oral hypoglycemic agent screen, the degree of elevation in plasma C-peptide and proinsulin levels may be helpful for suggesting either insulinoma or insulin autoimmune hypoglycemia as the underlying etiology.

•Marked elevation in C-peptide and proinsulin levels – Marked elevation in C-peptide and proinsulin levels (eg, >10 times the diagnostic threshold) can support an underlying etiology of insulin autoimmune hypoglycemia, which is caused by antibodies to endogenous insulin or to the insulin receptor. In patients with suspected insulin autoimmune hypoglycemia, autoantibodies to insulin and the insulin receptor (when possible) should be measured to distinguish insulin autoimmune hypoglycemia from other causes of endogenous hyperinsulinism (table 2 and table 4 and algorithm 3) [19]. Insulin antibodies do not need to be drawn during hypoglycemia.

In insulin autoimmune hypoglycemia, hypoglycemia and associated symptoms can occur postprandially, fasting, or in both states. In patients with insulin autoantibodies, the presumptive mechanism underlying hypoglycemia is that insulin secreted in response to a meal initially binds to the circulating antibodies, which block insulin action and cause hyperglycemia. Subsequently, insulin disassociates from the antibodies in an unregulated fashion, causing hyperinsulinemia and hypoglycemia. In patients with antibodies to the insulin receptor, hypoglycemia occurs due to antibody activation of the receptor.

If antibody titers are negative, the differential etiology is the same as that for mild to moderate elevations in C-peptide and proinsulin levels, reviewed immediately below.

•Mild to moderate elevation in C-peptide and proinsulin levels

-Predominantly fasting hypoglycemia – Patients with predominantly fasting hypoglycemia who have mild to moderate elevations in C-peptide and proinsulin levels most likely have underlying insulinoma. Higher levels of insulin, proinsulin, and C-peptide (though still below those seen in insulin autoimmune hypoglycemia) may suggest the presence of a malignant rather than benign insulinoma [20], and the proinsulin level may be disproportionately elevated in less differentiated tumors. This disproportionate elevation in proinsulin level is therefore one of the features that may indicate the need for increased surveillance after insulinoma resection. (See "Insulinoma", section on 'Post-treatment surveillance'.)

With insulinoma, the BHB level is low (≤2.7 mmol/L) due to insulin-mediated suppression of ketogenesis. Once endogenous hyperinsulinism has been established as the cause of hypoglycemia, further evaluation is warranted to localize the insulinoma. Localizing studies should not be performed until endogenous insulin-mediated hypoglycemia has been demonstrated and a screen for oral hypoglycemic agents is negative. (See "Insulinoma", section on 'Tumor localization'.)

-Postprandial hypoglycemia – Mild to moderate elevations in plasma insulin, C-peptide, and proinsulin values are found with noninsulinoma pancreatogenous hypoglycemia syndrome (NIPHS), a form of endogenous hyperinsulinemia due to islet hypertrophy and nesidioblastosis (the neodifferentiation of islet of Langerhans cells from pancreatic duct epithelium). An unusual feature of this disorder is that hypoglycemia occurs postprandially, two to four hours after a meal. Fasting hypoglycemia, characteristic of insulinoma, is rare in this disorder. Nesidioblastosis has also been described in patients with post-gastric bypass postprandial hyperinsulinemic hypoglycemia. Selective arterial calcium stimulation can help distinguish insulinoma from NIPHS as the etiology of hypoglycemia. (See "Insulinoma", section on 'Selective arterial calcium stimulation' and "Noninsulinoma pancreatogenous hypoglycemia syndrome".)

Insulin level <3 microU/mL — If the venous glucose is low (<55 mg/dL [3 mmol/L]) and the corresponding insulin level is <3 microU/mL (20.8 pmol/L), the hypoglycemia is unlikely to be mediated by insulin. However, measurement of the plasma insulin level may be erroneous or misleading in some circumstances. Plasma insulin measured by ICMA may be artifactually low due to hemolysis after blood sample collection. Some insulin assays detect only human insulin, whereas others detect both human insulin and synthetic insulin analogs. Consequently, depending on the assay employed, an insulin level <3 microU/mL (20.8 pmol/L) may not exclude exogenous insulin as a cause of hypoglycemia. Interpretation of the plasma insulin level therefore requires that the clinician know which insulin assay is used and whether the assay measures human insulin, insulin analogs, or both. If exogenous insulin administration is suspected, blood samples obtained during supervised testing at the time of hypoglycemia should be used to measure specific synthetic insulin analogs (eg, glargine, detemir, lispro, aspart). (See "Factitious hypoglycemia", section on 'Insulin measurements'.)

When hypoglycemia is not mediated by insulin, the BHB level is used to determine whether ketogenesis is suppressed through an insulin-independent pathway (algorithm 3).

●Suppressed ketogenesis (BHB level ≤2.7 mmol/L) – A low BHB level indicates that ketogenesis is suppressed, either by insulin or an IGF. Patients with a low venous glucose level, appropriately suppressed insulin secretion, and suppressed ketogenesis may have IGF-mediated hypoglycemia, also referred to as nonislet cell tumor hypoglycemia. This etiology of hypoglycemia is a paraneoplastic syndrome in which a tumor produces IGF-2 or its precursor pro-IGF-2. IGF-2 and pro-IGF-2 can lower glucose levels and suppress ketogenesis through activation of the insulin receptor, and hypoglycemia further may be potentiated by the high metabolic needs of the tumor. Such tumors are rarely occult, and IGF-related hypoglycemia in patients with known malignancy and large tumor burden can be identified by measuring free IGF-2, pro-IGF-2, and IGF-1 (to determine the ratio of IGF-2 to IGF-1). Values for these measurements that are above the upper limit of the laboratory reference range confirm the diagnosis. A healthy individual with hypoglycemia is unlikely to have nonislet cell tumor hypoglycemia. (See "Nonislet cell tumor hypoglycemia", section on 'Diagnostic approach'.)

If the IGF levels are not elevated and there is no evidence of malignancy, the proinsulin level may be useful in identifying an alternative etiology.

•Elevated proinsulin level (≥5 pmol/L) – Very rarely, insulinoma can present without overt hyperinsulinemia (ie, an insulin level <3 microU/mL [20.8 pmol/L]). In such cases, the proinsulin level is elevated, and the tumor is sometimes referred to as a proinsulinoma [21,22]. Therefore, if ketogenesis is suppressed and the proinsulin level is elevated, radiologic studies for tumor localization are a reasonable next step in evaluation. (See "Insulinoma", section on 'Tumor localization'.)

•Suppressed proinsulin level (<5 pmol/L) – Suppressed ketogenesis without hyperinsulinism also can occur with exogenous insulin administration if the insulin analog administered is not detected by the clinical laboratory insulin assay. Therefore, if the proinsulin level is suppressed, samples obtained during supervised testing should be used for measurement of insulin analogs that are not detected by the clinical assay. (See "Factitious hypoglycemia", section on 'Insulin measurements'.)

●Unsuppressed ketogenesis (BHB level >2.7 mmol/L) – A high BHB level indicates ketogenesis. A progressive rise in BHB with prolonged fasting, particularly after the 18-hour time point, excludes underlying insulinoma [23]. All patients with insulinoma have plasma BHB values ≤2.7 mmol/L at the end of the fast, whereas healthy individuals have values above this threshold (figure 3). In patients with evidence of ketogenesis, the presence or absence of neuroglycopenic symptoms (eg, dizziness, drowsiness, weakness, confusion, or delirium) during supervised testing can be helpful for differentiating between pathologic and physiologic hypoglycemia.

•With associated neuroglycopenic symptoms – Patients with a low venous glucose level (<55 mg/dL [3 mmol/L]) and concurrent neuroglycopenic symptoms, appropriately suppressed insulin secretion, and evidence of ketogenesis do not have an insulin- or IGF-mediated cause of hypoglycemia. Further evaluation should focus on other etiologies of hypoglycemia; these alternative etiologies include cortisol deficiency, chronic malnutrition, kidney or liver disease, and certain medications. The underlying cause is often suggested by clinical history and physical examination (table 2 and table 3). (See 'Initial assessment to determine the etiology' above.)

•Without associated neuroglycopenic symptoms – Patients with a venous glucose level ≥40 to <55 mg/dL (2.2 to 3 mmol/L), appropriately suppressed insulin secretion, and evidence of ketogenesis who do not experience attendant, neuroglycopenic symptoms are unlikely to have pathologic hypoglycemia. Glucose levels between 40 to 60 mg/dL (2.2 to 3.3 mmol/L) can reflect normal physiology, particularly in young, healthy individuals.

However, even in the apparent absence of neuroglycopenic symptoms, a venous glucose value <40 mg/dL (2.2 mmol/L) should raise clinical suspicion for pathologic hypoglycemia. In such patients, non-insulin- or IGF-mediated causes of hypoglycemia should be evaluated further (table 2). A glucose value in this range is not usually consistent with normal physiology, and subtle neuroglycopenic symptoms may be missed during provocative testing. (See 'Initial assessment to determine the etiology' above.)

Glucagon challenge to corroborate the etiology — In patients with suspected insulin- or IGF-mediated hypoglycemia, the glucagon challenge at the end of the supervised fast can help corroborate these etiologies. A suppressed BHB level and a vigorous plasma glucose response (>25 mg/dL [1.4 mmol/L]) to intravenous glucagon strongly support insulin or an IGF as the cause of hypoglycemia (table 4) [13,16]. In very rare cases in which a prolonged fast fails to elicit symptomatic hypoglycemia, glucagon challenge can induce insulin-mediated hypoglycemia and reveal an underlying insulinoma [24].

Insulin and the insulin-like action of IGFs inhibit glycogenolysis, and, therefore, these mediators permit retention of glycogen stores within the liver. In contrast, after prolonged fasting, healthy individuals will have depleted virtually all hepatic glycogen stores. As a result, patients with insulin- or IGF-mediated hypoglycemia respond to 1 mg of intravenous glucagon (a potent glycogenolytic agent) with an increase in hepatic glucose production and a robust rise in plasma glucose (>25 mg/dL [1.4 mmol/L]) 20 to 30 minutes after glucagon administration. Healthy individuals will not respond as vigorously to intravenous glucagon and will exhibit an increment in plasma glucose ≤25 mg/dL (1.4 mmol/L) (figure 4).

Test endpoints not achieved during the supervised fast — The supervised fast occasionally may be terminated before one of the endpoints is met. This may occur due to patient or provider anxiety during the fast or if the decision to terminate the fast was based on fingerstick glucose levels rather than laboratory measurements.

If the fast was continued for ≥18 hours, we send a blood sample collected at the end of the fast for measurement of the serum BHB level (algorithm 4). We also evaluate the glucose response to glucagon challenge at the end of the fast.

If the BHB level is >2.7 mmol/L and the glucose response to glucagon challenge is blunted (ie, increase ≤25 mg/dL), an underlying hypoglycemic disorder is unlikely. In such patients, no further evaluation is generally required.

The fast was not diagnostic and should be repeated in case of any of the following:

●Insufficient duration of the fast (ie, <18 hours).

●No evidence of ketogenesis (ie, serum BHB level ≤2.7 mmol/L at the end of the fast).

●Hepatic glycogen stores are not depleted (ie, glucose increment >25 mg/dL [1.4 mmol/L] after glucagon challenge). (See 'Glucagon challenge to corroborate the etiology' above.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Hypoglycemia in adults".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics” and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Low blood sugar in people without diabetes (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Initial assessment to determine the etiology – For individuals with evidence of a hypoglycemic disorder (ie, confirmed fulfillment of Whipple's triad or high clinical suspicion based on home glucose monitoring), the patient's overall health status and a thorough clinical history can help identify the most likely cause(s) of the hypoglycemia and inform the next steps in evaluation (table 2). (See 'History and physical examination' above.)

•Healthy individuals – In individuals who have no underlying medical conditions, take no medications, and feel otherwise well, the most likely etiology of hypoglycemia is endogenous hyperinsulinism or factitious hypoglycemia (table 2). Supervised testing should be performed to further evaluate these potential etiologies. (See 'Healthy individuals' above and 'Supervised testing' above.)

•Ill or medicated individuals – In individuals with underlying illness or comorbid conditions, the cause of hypoglycemia often can be determined based on clinical history alone (algorithm 1). In such patients, medications and alcohol use are the most common causes of hypoglycemia (table 3). (See 'Ill or medicated individuals' above.)

●Supervised testing – The purpose of supervised testing is to document Whipple's triad and establish the mechanism of hypoglycemia. If Whipple's triad has already been established, then a supervised test may be necessary for etiologic evaluation. This evaluation can be performed during an episode of fortuitously observed spontaneous hypoglycemia but usually requires provocative testing to elicit hypoglycemia (algorithm 2). (See 'Importance of supervised evaluation' above and 'Choice of test' above and "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis", section on 'Diagnostic evaluation'.)

Whipple's triad comprises the following:

•Symptoms consistent with hypoglycemia (table 1)

•A low plasma glucose concentration measured by laboratory assay when symptoms are present

•Resolution of hypoglycemic symptoms after the plasma glucose level is raised

●Mixed meal test for postprandial hypoglycemia – For individuals who experience exclusively postprandial hypoglycemia, the mixed meal test should be performed for further diagnostic evaluation. For this test, the patient consumes a non-liquid meal and is then observed for up to five hours [3]. If a low glucose level (<65 mg/dL [3.6 mmol/L]) is accompanied by hypoglycemic symptoms, additional blood samples should be sent for measurement of insulin and C-peptide levels, as well as an oral hypoglycemic agent screen. (See 'Mixed meal test for postprandial hypoglycemia' above.)

●Supervised fast for fasting hypoglycemia – For individuals who experience predominantly fasting hypoglycemia, a supervised fast should be performed. The fast may be performed in the outpatient or inpatient setting. During the fast, if the glucose level is <55 mg/dL (3 mmol/L) by laboratory measurement, additional blood samples should be sent for measurement of insulin, proinsulin, C-peptide, and beta-hydroxybutyrate (BHB) levels, as well as an oral hypoglycemic agent screen. (See 'Supervised fast for fasting hypoglycemia' above.)

●Interpretation of the supervised fast results – Laboratory test results help discriminate between insulin- and non-insulin-mediated causes of hypoglycemia but will not identify all potential etiologies of hypoglycemia (table 4 and algorithm 3). Additional testing may be required to verify the cause. (See 'Interpretation of supervised fast results' above.)

If the fast was terminated before test endpoints were achieved but was continued for ≥18 hours, findings may be sufficient to exclude an underlying hypoglycemic disorder (algorithm 4).

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges F John Service, MD, PhD (deceased), who contributed to earlier versions of this topic review.