the three diagnostic glucose tests for diabetes, all of which have limitations in hospitalized patients, the most recently added is HbA1c.5 Despite its advantages, a number of cautions still pertain to the reliability and accuracy of this test in the acutely ill population, especially when critical illness is superimposed on chronic comorbidities.11 There are numerous clinical scenarios in which the HbA1c may be falsely high or more commonly low and therefore not actually reflect the glycemic history that usually relates to changes in red blood cell survival times (Table 2.2). This obviously makes the HbA1c difficult to utilize as a diagnostic tool in the hospital setting without careful consideration. In one study, just treating an iron-deficiency anemia can lower the HbA1c from 10.1% to 8.2% in a population with diabetes and from 7.6% to 6.2% in a population without diabetes.12
Cardiac valvulopathies and valve replacements with both aortic or mitral valves can cause a microhemolysis resulting in a falsely low HbA1c.11 Thus, despite the fact many hospitals now require admission HbA1c measurements on patients with and without known diabetes entering the hospital, the test has inherent problems, resulting in the potential for misdiagnosis and mismanagement. Nonetheless, a significantly high HbA1c level (e.g., >8.0%) in the context of hyperglycemia (>180–200 mg/dL) makes the diagnosis of diabetes highly probable.
Table 2.2—Etiologies of Falsely High or Low HbA1c Levels
Falsely high
—Iron deficiency (with or without anemia)
—Anemia
—Hemoglobinopathies
—Race: African American, Hispanic, Asian
Falsely low
—Hemolysis
—Reticulocytosis
—Hemoglobinopathies
—Posthemorrhage or post-transfusion
—Drugs: iron, erythropoietin, dapsone
—Uremia
—Splenomegaly
Plasma glucose is another test that can be used for the diagnosis of diabetes. In the outpatient setting, a fasting glucose of 126 mg/dL or higher or a 75-g oral glucose tolerance test with a 2-h glucose ≥200 mg/dL confirms the diagnosis of diabetes.5 It is recommended that two such tests are performed in the absence of unequivocal hyperglycemia. The third diagnostic test using glucose is a random plasma glucose >200 mg/dL with classic symptoms of hyperglycemia (polyuria, polydipsia).5 In addition to a significantly high HbA1c, only this last diagnostic test can be used definitively to confirm the diagnosis of diabetes for the hospitalized patient.
Consider the patient with an HbA1c of 6.8%, anemia, and renal insufficiency who is admitted with pneumonia and fasting and postprandial glucose levels in the 130–140 mg/dL and 200–220 mg/dL range, respectively. This patient may or may not meet the criteria for diabetes once discharged from the hospital. Nonetheless, the inpatient strategy for the treatment of this patient’s hyperglycemia should be to meet the goals for optimal inpatient glycemic control and should not be influenced by diagnostic ambiguity. It is critical that “stress hyperglycemia versus diabetes” be included on the discharge problem list so both the patient and the outpatient care team appreciates the specific diagnosis clarifications to be investigated once the acute illness has resolved.
In addition to the diagnostic categories of diabetes and stress-induced hyperglycemia, the Expert Committee on Diagnosis and Classification of Diabetes now recognizes a significant group of patients who are at increased risk of developing future diabetes. The term “prediabetes” is used to describe these individuals with impaired fasting glucose or impaired glucose tolerance (Table 2.3).5 Although these patients do not meet the diagnostic criteria for diabetes, their glucose values are too high to be considered normal, and numerous prospective studies have shown a strong association between HbA1c and progression to diabetes. In the hospital setting, these patients can develop significant hyperglycemia and are at increased risk for complications while hospitalized and subsequently for cardiovascular disease.
Table 2.3—Categories of Increased Risk for Diabetes (Prediabetes)*
Fasting plasma glucose 100–125 mg/dL (impaired glucose tolerance)
OR
2-h plasma glucose in the 75-g oral glucose tolerance test 140–199 mg/dL
OR
HbA1c 5.7–6.4%
*For all three tests, risk is continuous, extending below the lower limit of the range and becoming disproportionately greater at higher ends of the range.
Classification
The goal of classifying a patient with a particular type of diabetes in the hospital setting should be to provide useful information about the pathogenesis, natural history, genetics, and phenotype of their disease to optimize safe and appropriate treatments, monitoring, education, patient expectations, and quality of life. Additionally, proper classification of hospitalized patients with hyperglycemia assists in appropriate transitions of aftercare.
Recent classifications have broadly distinguished the types of diabetes into two groups—autoimmune (T1D) and nonautoimmune (T2D)—with all other types being classified in an “other” category. The other category includes monogenic, gestational, pancreatic, steroid-induced, HIV-associated, hepatitis C–associated, polycystic ovarian syndrome–related, and endocrinopathy-associated (acromegaly and Cushing’s syndrome) diabetes. This general schema is useful despite considerable overlap in classic phenotypic presentation in each major class and will guide the knowledgeable health-care provider to make prudent decisions on whom to consider for more specific assignment. In addition to sophisticated testing, it is highly useful to obtain accurate and detailed histories of presentation and family history to advise further evaluation.
Type 1 Diabetes
T1D accounts for ~5–10% of diabetes and is the result of cellular-mediated autoimmune destruction of the pancreatic β-cells,13 resulting in moderate to severe insulin deficiency. It classically but not invariably manifests with acute and severe symptoms of hyperglycemia, dehydration, and ketoacidosis. Although the presence of autoantibodies assists in identifying autoimmune versus nonautoimmune diabetes, these antibodies usually but not always disappear over a variable amount of time. The most common antibody in the adult population is glutamic acid decarboxylase 65 (GAD65).14 Other antibodies that are quickly becoming commercially available include antibodies to tyrosine phosphatase IA-2 and zinc transporter 8 (ZnT8). Traditional islet cell antibodies (ICA) generally are not used because of the assay’s subjectivity. Insulin autoantibodies rarely are seen in adults (although they cross-react with antibodies from exogenous insulin). T1D has strong human leukocyte antigen (HLA) associations, which may be either predisposing or protective in most cases.
Although severe insulin deficiency and the tendency to ketosis and acute onset of symptoms are the hallmarks of T1D, the time of progression to absolute insulin deficiency is variable. Particularly in adults with newly diagnosed T1D, residual endogenous insulin secretion may still be present decades after the diagnosis15 and appears to be protective to the complications of the disease.16 This is significant to the health-care provider in the hospital setting because the measurement of c-peptide, while helpful in some circumstances, does not necessarily differentiate
