Patient holding a monitoring device to her finger

You start to feel shaky and tingly. Sweat beads up on your skin. Suddenly you realize you are very hungry. These symptoms represent the response of the autonomic nervous system to low blood glucose and are vital to the perception of hypoglycemia. If the brain becomes deprived of glucose, symptoms progress to sensations of warmth, weakness, confusion, loss of consciousness, and death. Healthy people without diabetes rarely experience severe hypoglycemia. Their bodies effectively regulate falling glucose levels by decreasing insulin secretion and increasing glucagon and epinephrine release. For many people with diabetes, however, hypoglycemia is a common and feared adverse effect of antihyperglycemic therapy, as well as a major limiting factor in managing their disease.

The landmark Diabetes Control and Complications Trial demonstrated that intensive insulin therapy (mean hemoglobin A1c, HbA1c 7.2%) significantly reduced the development of microvascular complications—retinopathy, nephropathy, and neuropathy—in subjects with type 1 diabetes when compared with conventional insulin therapy (mean HbA1c 9.1%), but concurrently increased the risk of severe hypoglycemia by threefold (1).

Though the nature of the relationship between hypoglycemia and mortality is controversial, the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial revealed an association between severe hypoglycemia and increased mortality in type 2 diabetes (2). The Action in Diabetes and Vascular Disease study of over 11,000 patients with type 2 diabetes similarly revealed an association between severe hypoglycemia and various adverse clinical outcomes, such as increased risk of major macrovascular events, major microvascular events, and death (3). In addition, a large retrospective observational study of hospital admissions among Medicare beneficiaries from 1999 to 2011 revealed that hypoglycemia has surpassed hyperglycemia as a cause of hospitalization for older adults in the United States (4).

Diabetic patients who experience hypoglycemia frequently fear these episodes and consequently might suffer from high levels of anxiety. Hypoglycemia is clearly a significant cause of physical and psychosocial morbidity for individuals with diabetes.

Despite these ill effects, many barriers exist to preventing, detecting, and treating hypoglycemia. This review examines the clinical and laboratory challenges associated with identifying clinically significant hypoglycemia in diabetic populations.

Defining Clinically Significant Hypoglycemia

Defining clinically significant hypoglycemia is not straightforward. The symptoms and signs of hypoglycemia are nonspecific and variable among people with diabetes (5). In addition, many people with diabetes lose the ability to recognize the symptoms that warn of developing hypoglycemia.

Hypoglycemia unawareness is most common in individuals with more aggressive therapy regimens. Measuring glucose concentrations can provide more objective evidence of hypoglycemia, but defining hypoglycemia on a strictly biochemical basis also has its limitations.

In clinical practice guidelines, Standards of Medical Care in Diabetes, the American Diabetes Association (ADA) outlines three categories of hypoglycemia (6). Level 1 encompasses glucose concentrations between 54 and 70 mg/dL. Results in this category should alert patients, families, and/or caregivers that treatment with fast-acting carbohydrates is needed.

In people without diabetes, a glucose concentration of 70 mg/dL is generally considered to be the threshold that triggers neuroendocrine responses to hypoglycemia. While this threshold is fairly reproducible among nondiabetic individuals, glycemic thresholds are much more dynamic in those with diabetes due to the impaired glucose counter-regulatory responses described above. Patients with poor glycemic control, for example, tend to have higher glycemic thresholds for symptom onset relative to those with tighter glycemic control. Within a given patient, this glycemic threshold will decrease over time with repeated episodes of hypoglycemia.

Level 1 hypoglycemia is a fairly conservative biochemical definition for patients with diabetes. In fact, preprandial glucose targets of 70–130 mg/dL were recommended by ADA until 2015, when the lower end of the target range was raised to 80 mg/dL. However, according to a consensus statement from ADA and the American Association of Clinical Endocrinologists (AACE), glucose concentrations below 70 mg/dL often predict imminent Level 3 hypoglycemia in hospitalized patients and, therefore, should trigger adjustment of antihyperglycemic therapy to prevent further episodes of hypoglycemia independent of the severity of symptoms (7).

Level 2 hypoglycemia, defined by a glucose concentration less than 54 mg/dL, represents clinically significant hypoglycemia requiring immediate action to resolve. In 2017, the International Hypoglycemia Study Group (IHSG) sought to identify a level of hypoglycemia with serious clinical and health-economic consequences in patients with diabetes (8). This was a herculean undertaking, as there is considerable intra- and inter-individual variation in glycemic thresholds for hypoglycemia symptoms and counter-regulatory responses among people with diabetes, especially those treated with certain antihyperglycemic agents, such as insulin and sulfonylureas.

Despite these challenges, IHSG argued that a consensus definition of clinically significant hypoglycemia would offer many benefits, including standardized reporting of hypoglycemia in clinical trials and the ability to robustly compare the effectiveness of various interventions aimed at preventing hypoglycemia.

To that end, IHSG evaluated two glucose thresholds as potential candidates: 50 mg/dL and 54 mg/dL. These two values were selected because both cause defective glucose counter-regulation and impaired hypoglycemia awareness in diabetic individuals and neither occur under physiological conditions in people without diabetes. The generally accepted threshold for onset of neuroglycopenic symptoms in people without diabetes is 50 mg/dL. The results of the ACCORD and Outcome Reduction with Initial Glargine Intervention (ORIGIN) trials did not definitively support that hypoglycemia below 54 mg/dL was associated with increased mortality. Of note, both of these trials excluded subjects with known risk factors for severe hypoglycemia, including those with insulin-dependent diabetes in the ORIGIN trial and those with frequent or recent serious hypoglycemic events in the ACCORD trial.

While these large randomized controlled trials did not provide strong evidence supporting a specific hypoglycemia threshold of either 50 mg/dL or 54 mg/dL, a few smaller studies offered interesting insights. In one study of 12 patients with long-duration type 1 diabetes, a failure to recognize one’s own hypoglycemia when plasma glucose concentrations were less than 54 mg/dL (3 mmol/L) significantly increased the risk of severe hyperglycemia (9).

Two other studies of patients with insulin-dependent type 2 diabetes found that interstitial glucose concentrations less than 54 mg/dL were associated with cardiac arrhythmias (10, 11). Based on this limited evidence, IHSG recommended that glucose concentrations below 54 mg/dL were sufficiently low to indicate serious, clinically significant hypoglycemia. At the end of 2017, ADA, AACE, and multiple other prominent diabetes groups adopted this consensus definition of clinically significant hypoglycemia as part of a larger initiative to identify clinically meaningful outcomes other than HbA1c in type 1 diabetes (12).

The third and final category of hypoglycemia, Level 3 or severe hypoglycemia, is defined not by a numeric threshold but as an event: serious cognitive and/or physical impairment requiring external assistance for reversal by administration of rapid-acting glucose or glucagon. The risk factors for developing severe hypoglycemia are complex and multifactorial, but include older age, sulfonylurea and/or insulin therapy, intensive glycemic control, social determinants of health, and a variety of co-morbid health conditions.

Accurately and rapidly quantifying glucose concentrations is an essential tool for identifying hypoglycemic episodes, especially in individuals not experiencing symptoms, and preventing the increased morbidity and mortality associated with severe hypoglycemia. Many preanalytical and methodological variables hinder the utility of glucose monitoring for this purpose.

Methods for Glucose Monitoring

Multiple methods for glucose monitoring are now in use, comprising glucose meters for self-monitoring of blood glucose (SMBG), devices for continuous glucose monitoring (CGM) in interstitial fluid, and automated assays performed in clinical laboratories. The most common methods used in clinical laboratories and at patients’ bedsides involve enzymatic reactions coupled with either spectrophotometric or electrochemical detection.

Automated clinical chemistry analyzers employ hexokinase or glucose oxidase (GO) to measure glucose in serum or plasma; these methods are accurate, precise, and relatively standardized. Glucose meters widely used in at-home and hospital settings quantify glucose concentrations in whole blood, typically capillary, utilizing either GO or glucose dehydrogenase (GDH). These devices have two main advantages: rapid turnaround times and use of minimal blood volumes. While the quality of glucose meters has improved over the past 30 years, their performance does not approach that of central laboratory methods in terms of accuracy and precision.

Moreover, the analytical performance varies significantly between glucose meters. A recent study compared the accuracy of 17 point-of-care glucose meters from nine manufacturers using nearly 350 contrived samples covering a wide range of glucose (20–440 mg/dL) and hemoglobin (7.0–16.5 g/dL) concentrations (13). The overall mean deviations of glucose meter measurements relative to a reference method varied nearly fourfold, from 5.6% to 20.8%. The accuracy of all devices, including the Nova Biomedical StatStrip cleared by the Food and Drug Administration (FDA) for use in critically ill patients, deteriorated significantly for hypoglycemic values less than 70 mg/dL; the mean deviations exceeded 25% for nearly half of the glucose meters evaluated.

The impact of meters’ poor analytic performance at low glucose concentrations can be exacerbated by additional factors, such as operator variability and the inappropriate use of capillary specimens in patients with impaired peripheral circulation. For these reasons, hypoglycemia indicated by a glucose meter should be confirmed by a central laboratory method using an arterial or venous sample, especially if the patient’s clinical picture is inconsistent with this finding.

The acceptability of glucose meter analytical performance is typically adjudicated relative to guidelines put forth by FDA, the Clinical and Laboratory Standards Institute (CLSI), and/or the International Organization for Standardization (ISO). The CLSI guidelines for performance of point-of-care blood glucose meters (document POCT12-A3) states that 95% of glucose meter results should be within 12.5 mg/dL or 12.5% of reference method measurements (14). Updated in 2013, the ISO 15197 guidelines define acceptable glucose meter performance as achieving a bias no larger than 15 mg/dL or 15% relative to a reference method for 95% of measurements (15).

In 2016, FDA released two sets of guidelines pertaining to point-of-care glucose meters: 1) 95% of results must be within 12 mg/dL or 12% of the reference method; and 2) 98% of results must be within 15 mg/dL or 15% of the reference method (16). But meeting even the most rigorous of these requirements is not good enough for glucose meters to reliably identify Level 1 and Level 2 hypoglycemia as defined by ADA.

For example, given a true glucose concentration of 65 mg/dL, acceptable performance per FDA guidelines permits glucose meter measurements spanning 53–77 mg/dL, with 5% of measurements allowed to exceed this range. Therefore, classifications of Level 1 hypoglycemia, Level 2 hypoglycemia, and within recommended glucose targets of 80–130 mg/dL are all statistically possible outcomes for a glucose concentration of 65 mg/dL measured by a glucose meter. Though glucose meters are often used to detect clinically significant hypoglycemia, this is clearly not what they were designed to do.

As subject matter experts in assay methodological performance, clinical laboratorians must be proactive in collaborating with patient care teams to educate them about the limitations of point-of-care testing in the context of hypoglycemia, and to implement workflows to minimize the impact of these limitations on patient care. Furthermore, given the poor reliability of glucose meters in hypoglycemic ranges and their variable analytical performance overall, clinical trials investigating the prevention and/or detection of hypoglycemia or the impact of hypoglycemia on long-term outcomes should clearly specify the method(s) they use to measure glucose. Ideally, glucose meters should be avoided when accurate classification of Level 1 or Level 2 hypoglycemia is crucial.

PreAnalytical Considerations

Another challenge with glucose measurement is its notorious susceptibility to preanalytical error. One mechanism of error in hospitalized patients is contamination of samples drawn through central or intravenous lines concurrently used for infusions. Depending on the chemical composition of the infusion, measured glucose results might be falsely lowered, with the potential to falsely indicate hypoglycemia, or they might be falsely elevated, with the potential to mask true hypoglycemia.

Another source of preanalytical error is spurious hypoglycemia due to delayed sample processing. At room temperature, ongoing glycolysis by erythrocytes and leukocytes in whole blood removes glucose at an hourly rate of 5%–7%. The joint AACC-ADA guideline on laboratory testing in diabetes recommends that blood samples for glucose monitoring be immediately immersed in an ice slurry and analyzed within 30 minutes of collection (17). While not applicable for glucose monitoring at the point-of-care, these sample processing instructions also are not feasible in the vast majority of scenarios requiring glucose measurement by a central laboratory method, especially for samples collected at sites distant from laboratories. Consequently, there is a risk of spurious hypoglycemia if centrifugation is delayed.

A recent study illuminates the importance of preanalytical processing in diagnosing gestational diabetes mellitus (GDM) (18). The authors of this study compared mean glucose concentrations and overall rates of GDM diagnosis before and after a change in preanalytical processing of oral glucose tolerance test (OGTT) samples. Interestingly, the rate of GDM diagnosis almost doubled from 11.6% with delayed centrifugation to 20.6% with early centrifugation. Diagnosis of GDM based on fasting glucose concentrations experienced the most significant increase (5.9% to 13.4%). The authors called for standardized preanalytical blood sampling protocols for OGTTs during pregnancy and for disclosure of these protocols when GDM studies are published.

Though most recommendations about preanalytic factors impacting glucose measurement have been derived from studies seeking to detect and diagnose diabetes, these recommendations should also apply to studies on hypoglycemia. This is especially true for Level 1 and Level 2 hypoglycemia defined by single glucose concentration cutoffs of 70 mg/dL and 54 mg/dL, respectively. Patients with blood samples subjected to stringent preanalytical processing protocols, all other things being equal, will be classified less frequently as biochemically hypoglycemic than those subjected to more relaxed processing steps. Strong evidence is currently lacking about the impact of Level 1 and Level 2 hypoglycemia and their association with long-term outcomes. Future studies seeking to tackle these gaps in evidence should standardize preanalytical handling of blood samples in order to provide meaningful results.

Conclusion

Hypoglycemia is a considerable barrier to glycemic control in people with diabetes. Even with recently established consensus definitions from ADA, clinically significant hypoglycemia is difficult to define biochemically due to significant variability in glycemic thresholds for hypoglycemia symptoms and counterregulatory responses among people with diabetes.

The problem for patient care can be compounded by the fact that glucose meters used at the point-of-care do not currently have sufficiently robust analytical performance to differentiate between Level 1 and Level 2 hypoglycemia. Consistent preanalytical sample processing is also important to properly classify a patient’s degree of hypoglycemia.

By promoting the appropriate use of glucose testing strategies for accurate and timely detection of hypoglycemia, clinical laboratorians play a crucial role in improving the quality of care for patients with diabetes.

References

  1. Nathan DM, Group DER. The diabetes control and complications trial/epidemiology of diabetes interventions and complications study at 30 years: Overview. Diabetes Care 2014;37:9-16.

  2. Bonds DE, Miller ME, Bergenstal RM, et al. The association between symptomatic, severe hypoglycaemia and mortality in type 2 diabetes: Retrospective epidemiological analysis of the accord study. BMJ 2010;340:b4909.

  3. Zoungas S, Patel A, Chalmers J, et al. Severe hypoglycemia and risks of vascular events and death. N Engl J Med 2010;363:1410-8.

  4. Lipska KJ, Ross JS, Wang Y, et al. National trends in U.S. hospital admissions for hyperglycemia and hypoglycemia among Medicare beneficiaries, 1999 to 2011. JAMA Intern Med 2014;174:1116-24.

  5. Cryer PE, Davis SN, Shamoon H. Hypoglycemia in diabetes. Diabetes Care 2003;26:1902-12.

  6. American Diabetes Association. Glycemic targets: Standards of medical care in diabetes-2020. Diabetes Care 2020;43:S66-76.

  7. Moghissi ES, Korytkowski MT, DiNardo M, et al. American Association of Clinical Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control. Diabetes Care 2009;32:1119-31.

  8. International Hypoglycaemia Study Group. Glucose concentrations of less than 3.0 mmol/l (54 mg/dl) should be reported in clinical trials: A joint position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2017;40:155-7.

  9. Cranston I, Lomas J, Maran A, et al. Restoration of hypoglycaemia awareness in patients with long-duration insulin-dependent diabetes. Lancet 1994;344:283-7.

  10. Chow E, Bernjak A, Williams S, et al. Risk of cardiac arrhythmias during hypoglycemia in patients with type 2 diabetes and cardiovascular risk. Diabetes 2014;63:1738-47.

  11. Pistrosch F, Ganz X, Bornstein SR, et al. Risk of and risk factors for hypoglycemia and associated arrhythmias in patients with type 2 diabetes and cardiovascular disease: A cohort study under real-world conditions. Acta Diabetol 2015;52:889-95.

  12. Agiostratidou G, Anhalt H, Ball D, et al. Standardizing clinically meaningful outcome measures beyond HbA1c for type 1 diabetes: A consensus report of the American Association of Clinical Endocrinologists, the American Association of Diabetes Educators, the American Diabetes Association, the Endocrine Society, JDRF International, the Leona M. and Harry B. Helmsley charitable trust, the Pediatric Endocrine Society, and the TLD exchange. Diabetes Care 2017;40:1622-30.

  13. Ekhlaspour L, Mondesir D, Lautsch N, et al. Comparative accuracy of 17 point-of-care glucose meters. J Diabetes Sci Technol 2017;11:558-66.

  14. Point-of-care blood glucose testing in acute and chronic care facilities; approved guideline -- third edition. CLSI document POCT12-A3. Wayne, PA: Clinical and Laboratory Standards Institute 2013.

  15. In vitro diagnostic test systems – Requirements for blood-glucose monitoring systems for self-testing in managing diabetes mellitus. ISO 15197. International Organization for Standardization 2013.

  16. Self-monitoring blood glucose test systems for prescription point-of-care use. Guidance for Industry and Food and Drug Administration Staff. United States Food and Drug Administration 2016.

  17. Sacks DB, Arnold M, Bakris GL, et al. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Clin Chem 2011;57:e1-47.

  18. Potter JM, Hickman PE, Oakman C, et al. Strict preanalytical oral glucose tolerance test blood sample handling is essential for diagnosing gestational diabetes mellitus. Diabetes Care 2020;43:1438-41.

Anna E. Merrill, PhD, DABCC, is a clinical assistant professor of pathology and associate director of the clinical chemistry laboratory at the University of Iowa Hospitals & Clinics in Iowa City. +Email: anna-merrill@uiowa.edu