Diabetes Classification

Are You Sure the Patient Has Diabetes?


Diabetes is defined as an elevated blood glucose of >=126 mg/dl or a hemoglobin A1c >=6.5% or a random blood glucose >=200 mg/dl with signs and symptoms of hyperglycemia. Diabetes is commonly associated with signs and symptoms of acute or chronic microvascular and macrovascular complications.

Classically, symptoms due to acute hyperglycemia include polyuria, polydipsia, polyphagia, weight loss, fatigue, increased infections, and blurred vision. Early in diabetes, there may also be an absence of symptoms during which time complications may develop. Signs are related to the effects of dehydration, catabolism, and metabolic changes including hypotension, weight loss, dehydration, and diabetic ketoacidosis (DKA). Symptoms and signs related to the complications of diabetes are due to tissue effects of long-term hyperglycemia (often in the context of hypertension, hyperlipidemia, obesity, and other coexisting metabolic abnormalities) and are related to specific organ damage commonly including eye, kidney, and nerve damage, as well as cardiovascular disease, and extending to nearly all tissues.

Diabetes is a very heterogeneous disorder marked by relative or absolute deficiency in insulin secretion and most often by decreased insulin action. Some forms of diabetes are autoimmune and may be associated with other autoimmune endocrine diseases others clearly linked to genetic (MODY) and syndromic disorders (Turner’s, Wolfram’s), while other forms have no known etiology. There are several traditional classifications based on our understanding of the cause of hyperglycemia. Although in many instances, individual patients do not neatly fit these classifications, appropriate classification may be important for identifying the pathogenesis, predicting treatment outcomes and potential complications. Future advances in diabetes care rely on identifying known and unknown etiologies of diabetes.

Type 1 diabetes

Definition: Type 1 diabetes accounts for ~10% of all diabetes cases and is due to a chronic progressive asymptomatic autoimmune destruction of beta cells over time resulting in a critical loss of beta cell function. This results in decreased insulin secretion followed by a rise in blood glucose. At the time of diagnosis, there is typically an absolute insulin deficiency and except for a limited duration honeymoon (<1 year), persists throughout life. The autoimmune destruction of beta cells is linked to certain HLA associations, which are involved in antigen presentation to the T cells. There are specific HLA types, such as DQA and DQB as well as DRB genes. Some HLA associations are protective against diabetes (DQB1*0602), while others predispose to diabetes. Not all individuals with predisposing HLA subtypes develop diabetes. Thus, it is assumed that the immunological background is permissive for the development of diabetes in the presence of a yet unidentified environmental trigger.

Usually patients with type 1 diabetes are normally insulin sensitive or only slightly insulin resistant and are commonly young, lean individuals. However, older individuals may also develop autoimmune type 1 diabetes. The spectrum of presentation of blood glucose can range from a mild elevation in blood glucose to symptomatic hyperglycemia and marked metabolic deterioration including DKA. The earliest metabolic alterations are a loss of first phase insulin response, then a rise in the 2-hour blood glucose after an oral glucose tolerance test and a gradual increase in the A1c within the normal range before the development of overt diabetes.

Untreated, type 1 diabetes has a high mortality and long-term complications are significantly decreased with sustained intensive glycemic control typically targeting an A1c <7%. The importance of correct classification is to predict the need for life-long insulin treatment for symptom and metabolic control and to prevent complications.

Signs and Symptoms: In adults and children, diabetes presents acutely with signs and symptoms of hyperglycemia including weight loss, polyuria, polydipsia and variable degrees of dehydration. Although classical, many do not present DKA with the fruity smelling breath of ketones. Patients may be also asymptomatic during the long prodromal phase of autoimmune beta cell destruction.

Laboratory Findings: Abnormal elevation of blood sugar is required for the diagnosis. In DKA, there is hyperglycemia, acidemia (arterial pH is <7.3), low bicarbonate (<15 mmol/L) and elevated serum ketones (beta-hydroxybutyrate, acetoacetate). In practice, testing for acetoacetate is readily available and is used as an adjunct to the diagnosis of DKA. Acetoacetate levels may be low in seriously ill volume-depleted and acidotic patients with DKA as the redox reaction favors beta hydroxybutyrate, not acetoacetate; treatment of DKA in these cases may result in a paradoxical increase in acetoacetate despite clinical and biochemical improvement of the patient as redox shifts back from beta-hydroxybutyrate to acetoacetate. Direct measurement of beta-hydroxybutyrate, while more reliable for diagnosis and following the treatment of DKA, may not be available in clinical labs and is not necessary to measure in typical cases. Other findings include dehydration, low serum sodium (resulting from intracellular fluid shifts), and hypertriglyceridemia (the effect of decreased insulin and hence decreased hepatic lipase activity). In outpatients, ketones may sometimes be measured at home by patients in cases where metabolic control is suspected to be deteriorating.

Ninety percent of children with new-onset diabetes have one or more pancreatic autoantibodies including those directed toward insulin, GAD-65, tyrosine phosphatase–like protein fragment 1A-2 or 1A-2β, and islet beta-cell zinc ion efflux transporter, ZnT8. In patients with clinically typical type 1 diabetes there is no practical benefit to measuring autoantibodies and screening for autoantibodies among the relatives of patients with type 1 diabetes is not recommended. On the other hand, if there is a clinical uncertainly about the type of diabetes, the presence of autoantibodies may help predict outcomes and guide treatment. It is not recommended to measure insulin or C-peptide levels or HLA subtypes as this will not change clinical management. In special situations such as testing for organ donor compatibility (HLA testing), these tests may be indicated.

Latent autoimmune diabetes of adults (LADA)

Definition: LADA is an autoimmune form of diabetes in adults who appear to have phenotype of type 2 diabetes. Unlike type 2 diabetes, there is early beta cell failure. These patients are initially treated with oral agents but in few years require insulin therapy.

Signs and Symptoms: In contrast to the frequent acute DKA presentation of typical autoimmune type 1 diabetes, two-thirds of patients with LADA present with typical signs and symptoms of acute hyperglycemia without DKA and clinically resemble type 2 diabetes. An important clue to LADA might be early failure of oral agents such as sulfonylureas in these patients.

Laboratory Findings: Hyperglycemia and presence of any one of the pancreatic autoimmune markers [glutamic acid decarboxylase (GAD 65) antibodies, Islet cell antibodies (ICA), Insulin autoantibodies (IAA) and Insulinoma- associated protein 2 (IA2)]. GAD 65 antibodies are sensitive and specific for LADA. Other useful markers include C-peptide and HLA typing.

Type 1B diabetes

Some cases of diabetes without autoantibodies present have marked hyperglycemia, DKA, and severe insulin deficiency. However, DKA and the hyperglycemic crisis may be transient with only occasional recurrences. During stable metabolic periods, patients can be treated with diet or oral antidiabetic agents. This is termed type 1B by the American Diabetes Association classification. We and others have termed this atypical diabetes, Flatbush diabetes, or ketosis-prone diabetes, as discussed here.

Type 2 diabetes

Definition: This accounts for 80-90% of all diabetes cases worldwide and is not associated with autoimmune markers. Type 2 diabetes is characterized by a relative insulin deficiency in the context of an increased insulin demand due to insulin resistance associated with increased adiposity (overweight to obese). This setting of hyperglycemia, insulin resistance, low adiponectin levels, and other associated metabolic abnormalities leads to increased cardiovascular disease risk. However, there is some variability in the insulin resistant phenotype and, depending on ethnicity, 15-50% of patients with type 2 diabetes may be normally insulin-sensitive with greater insulin deficiency than insulin-resistant type 2 diabetic patients. Such individuals have decreased intra-abdominal, hepatic, and pericardial fat and lower triglyceride levels and do not have an increase in cardiovascular risk factors, as do the insulin-resistant subtypes.

Signs and Symptoms: There may be NO identifiable presenting signs and symptoms of underlying type 2 diabetes. Because of this, patients may present primarily with evidence of such complications as eye, kidney, or nerve damage, peripheral vascular disease, myocardial infarction (MI), or stroke. Patients without known diabetes who present with diabetic retinopathy, cataracts, renal or vascular disease, MI, or stroke should be evaluated for diabetes. Symptoms of hyperglycemia include nocturia, polyuria, polydipsia, weight loss, blurred vision, and increased infections in those with or without end organ complications.

In the past decade, there have been numerous reports of DKA without precipitating factors in adult patients (and children) who are obese and after treatment appear to have type 2 diabetes. This is common among African American, African Caribbean, or Hispanic minorities but has also been reported worldwide. Indeed nearly half the patients with DKA who present to urban emergency departments have this form of diabetes. In contrast, the American Diabetes Association classification traditionally maintains that type 2 diabetes rarely presents with DKA and only when clear precipitating factors are present.

Laboratory Findings: If the patient is asymptomatic, a fasting plasma blood glucose >=126 mg/dl (7 mmol/L) or an A1c value >=6.5% under nonstressed conditions establishes the diagnosis of diabetes. The blood sugar may be markedly elevated with variable evidence of dehydration (depending upon the degree of hyperglycemia), hyperosmolarity, and the absence of autoimmune markers.

Atypical diabetes

Definition/Description: This increasingly recognized heterogeneous entity, termed Flatbush diabetes, type 1.5 diabetes, or ketosis-prone diabetes, is clinically characterized by transient DKA. When a patient presents with DKA to the emergency department, it is not possible to classify them. Their identity becomes clear over time with observation of the response to treatment. As with type 1 diabetes, insulin is required to treat any DKA. Adults are frequently newly diagnosed and without evidence of precipitating factors. Despite presenting with DKA, which classically implies a life-long requirement for insulin, after a brief period of treatment (up to ~3-6 months), patients no longer require insulin and can be treated with diet or oral antidiabetic agents. Thus, when evaluating a patient presenting with DKA it is important to first treat using insulin, fluids, and electrolytes as appropriate. Once metabolism is stabilized, over the ensuing weeks and months, it is important to appropriately classify the patient in order to predict the need for future long-term insulin requirements. The underlying pathophysiology is a marked, but transient, decrease in insulin’s effect to prevent ketogenesis and insulin resistance demonstrated at the level of muscle, fat, and liver. Initially described in adults, children, and adolescent minorities and those of African origin, this has also reported among Europeans and Asians.

A related variant–typically in obese children and adults–is a presentation of acute marked hyperglycemic crisis without DKA. After initial management, insulin secretion may improve to the point of not needing insulin and can be treated with oral antidiabetic medications or diet alone. Such near normoglycemic remissions may be prolonged up to 15 years. Thus, there is a spectrum of clinical presentations ranging from asymptomatic to symptomatic type 2 diabetes and from hyperglycemic crisis to DKA.

The classification of adults presenting with DKA can be based on clinical symptoms and signs, degree of obesity, residual insulin secretory capacity, and presence of autoimmune markers (GAD 65, IA-2). The ADA classifies diabetes with autoimmune markers and insulin deficiency as type 1a. Those with DKA without autoimmune markers and intermittent insulin deficiency are termed type 1b or idiopathic. This does not provide sufficient basis for future treatment for a condition that appears to be increasing in frequency. When classified based on obesity, lean patients require insulin therapy while obese patients can often transition to oral medications or diet therapy. Another more complex, but perhaps physiologically relevant classification merges a functional assessment of adequate beta cell insulin secretion in response to iv glucagon (beta cell negative or positive) with the presence or absence of autoimmune markers (A negative or positive). When analyzing all patients who come to the emergency department with DKA, there are 4 possible combinations or groups: (1) A pos-B neg (frequency 17%, this is typical of type 1 diabetes), (2) A neg-B neg (frequency 22%), (3) A pos-B pos (frequency 11%), and (4) A neg-B pos (frequency 50%; 50% are new onset; this is atypical, Flatbush diabetes, or ketosis-prone type 2 diabetes).

The advantage of this classification is the ability to predict the requirement for insulin at least in near term and describe the pathogenesis. Patients who are beta cell positive based on an adequate fasting and a stimulated C-peptide measured 1-3 weeks after DKA have a very high likelihood (50%) of not needing insulin. The presence or absence of antibodies does not determine the need for insulin. The absence of insulin secretion, regardless of the antibody status was predictive of future recurrence of DKA and need for insulin treatment.

Signs and Symptoms: Typical of hyperglycemia and marked insulin deficiency. These include polyuria, polydipsia, weight loss, dehydration, and DKA. Often this is the initial presentation of diabetes in both children and adults who may be obese or lean, and frequently adolescents and children have acanthosis nigricans.

Laboratory Findings: An elevated blood glucose and anion gap ketoacidosis are almost invariable. Evidence of dehydration, low serum sodium (proportional to resulting from intracellular fluid shifts), and hypertriglyceridemia are frequently present. There is an absence of known antibodies directed toward epitopes of GAD-65, islet cell, tyrosine phosphatase 1A-2 or 1A-2β, and ZnT8. Once the patient has been stabilized for several weeks (2-3 weeks) after presenting with DKA, the functional beta cell reserve may be assessed in order to predict the need for future insulin therapy. A C-peptide may be measured in the fasting state and 6 and 10 minutes after an intravenous dose of 1 mg of glucagon. Low levels of C-peptide suggest the need for insulin. The presence of antibodies does not suggests the need for insulin in the near term but long term data is unavailable.

Maturity Onset Diabetes of the Young (MODY)

Another uncommon class of diabetes is called maturity onset diabetes of the young (MODY). It is estimated to account for about 1-2% of all diabetic patient population. See details below.

Gestational diabetes mellitus (GDM)

With the increasing overall prevalence, there are several classifications of diabetes in a pregnant woman: type 1, type 2, and diabetes diagnosed during pregnancy usually in the 2nd and 3rd trimester. The classification of GDM recently changed when The Hyperglycemia and Adverse Pregnancy Outcomes study (HAPO) demonstrated a continuous relationship between hyperglycemia (within the “normal” range) at 24-28 weeks and adverse maternal and fetal outcomes in women who were not diabetic at the start of the pregnancy.

GDM is reported in 2-25% of pregnancies and traditionally described as any diabetes first identified during pregnancy. Typically GDM resolves after the birth of the child. GDM is a risk factor for increased rates of adverse maternal and fetal/neonatal outcomes. For the neonate, these include macrosomia, increased percent fat, shoulder dystocia, organomegaly (if diabetes is present early in pregnancy), neonatal hypoglycemia, respiratory problems, hyperbilirubinemia, increased mortality, and an increased risk of adult metabolic diseases including diabetes and cardiovascular disease. For the mother, these include preeclampsia and caesarean section. Physiological adaptations in pregnancy, such as increased placental lactogen, ACTH, growth hormone, low adiponectin, and obesity, cause insulin resistance, ensuring availability of nutrients for the developing fetus. When the maternal blood sugar rises, plasma insulin rises to compensate. In individuals predisposed to diabetes, this compensation fails and diabetes develops. Maternal blood glucose, which crosses the placenta (insulin does not), gives rise to fetal overnutrition, increased insulin, and other hormones and may result in neonatal hypoglycemia in the post partum state.

Current increased rates of diabetes in the population as well as increased rates of maternal age, obesity, and prior GDM have increased diagnosed and undiagnosed GDM in pregnancy. Thus, it is important to refine the classification of GDM into pregestational or overt diabetes (present before pregnancy) versus diabetes with the pregnancy (usually the last 2 trimesters). Overt or pregestational diabetes should be assessed during the first prenatal visit, ideally as early as possible in gestation; in practice, this often depends on when the patient first comes for prenatal care and may be months into gestation. Patients with diagnosed type 1 or type 2 diabetes clearly do not need testing. The criteria for the classification of pregestational diabetes are the same as for nonpregnant persons. These include a fasting plasma glucose >=126 mg/dl or A1c >=6.5% or random glucose >=200 mg/dl with signs and symptoms of hyperglycemia such as polyuria and polydipsia. If there is no diabetes, then the patient should be screened for diabetes between 24 and 28 weeks’ gestation.

The traditional approach is a 2-step process endorsed by ACOG (American College of Obstetrics and Gynecology): a positive screening 1-hour glucose value of >130 mg/dl (or 140 mg/dl) after a 50-gram glucose challenge is followed by a diagnostic 3-hour 100-gram OGTT. If any 2 of 4 glucose values are above the targets, gestational diabetes is diagnosed. The targets for the fasting plasma glucose, 1-hour, 2-hour, and 3-hour plasma glucose levels are 95, 180, 155, and 140 mg/dl, respectively. These values are linked to the risk of developing maternal diabetes post partum and not to perinatal complications. In 2008, the HAPO trial first showed a relationship between maternal hyperglycemia based on a one step 2-hour 75 gram OGTT and maternal/fetal adverse effects. The relationship was a continuous one without a threshold, so the cut-off values for the OGTT diagnosis of gestational diabetes were set by expert consensus as those that resulted in a 75% increase in maternal or fetal adverse effects. In pregnant woman without known diabetes, any value of the fasting blood glucose, 1-hour or 2-hour glucose >=92, 180 and 153 mg/dL respectively be sufficient for GDM diagnosis on OGTT testing. The IADPSG (International Association of Diabetes and Pregnancy Study Group), the ADA, and others have agreed to use this but not the ACOG. Despite increasing the numbers of patients with GDM to 18%, 2 studies have shown improved maternal and fetal outcomes when treating pregnant women with such mild hyperglycemia. These treatments were mostly diet and lifestyle modification.

What Else Could the Patient Have?

Type 1 diabetes

A young lean patient presenting with DKA, a history of diabetes, DKA and auto-antibodies, is consistent with classical type 1 auto-immune diabetes and will require life-long insulin. Alternatively, a patient presenting with DKA, young or adult, especially if newly diagnosed, could have transient insulin requirement and represent atypical diabetes, ketosis-prone (A- B+ form), or Flatbush diabetes. Patients presenting with hyperglycemic crisis (marked hyperglycemia but no DKA, frequently hyperosmolar) may also have a transient insulin requirements. Patients with new-onset diabetes, minority ethnicity, presence of beta cell insulin secretion soon after stabilization of the acute episode, obesity, and transient insulin requirements are likely to have atypical diabetes, whose later clinical course is that of type 2 diabetes. See details in “Atypical Diabetes.” The presence or absence of islet-associated antibodies in such phenotypes does not invariably suggest future insulin dependency at least after 6 months.

Patients presenting acutely with DKA, especially if Asian, could also have “fulminant diabetes.” While the blood glucose is elevated, the A1c is normal, suggesting a sudden metabolic decompensation. The underlying etiology is unclear but appears to include a viral prodrome, altered T cell function, and evidence of autoimmunity. Most reports are from Asia, and only one patient with this form of diabetes has been reported to date in the United States.

Young patients who carry the diagnosis of type 1 or type 2 diabetes may also have MODY (see later)

Type 2 diabetes

Patients with the phenotype of type 2 diabetes could have LADA (latent autoimmune diabetes of adults). The prevalence varies among different populations (4-10% in Caucasians, 5.9% in Chinese population). It is less common among African-Americans and other US minorities. Among all type 2 diabetic patients of all ages in United Kingdom Prospective Diabetes Study (UKPDS), LADA was prevalent among 10% of patients. LADA is a slowly progressive autoimmune disease with beta cell failure. The diagnosis is suggested by early loss of glycemic control on oral hypoglycemic agents. To standardize the definition and diagnosis, Immunology of Diabetes Society (IDS) has proposed 3 diagnostic criteria.

a) Age>30 years at diagnosis

b) Presence of circulating islet cell autoantibodies (at least one), most sensitive GAD 65 antibodies

c) Lack of insulin requirement for at least 6 months after diagnosis

A clinical screening tool developed in Australia and validated in 102 Caucasian patients with type 2 diabetes found that the presence of 2 of 5 clinical criteria gave a 90% sensitivity and 75% specificity of classifying a patient as LADA. The 5 clinical criteria are: onset <50 years, lean (BMI <25), acute symptoms at onset (polyuria, polydipsia, and unintentional weight loss), personal history of autoimmune disease, or a family history of autoimmune disorders. However none of the above diagnostic criteria are perfect and are prone to physician bias (for e.g. when to initiate insulin therapy) and these should be replicated in studies.

Studies have shown that patients with LADA have phenotypic differences according to bimodal GAD65 antibodies titers. Patients with high titer are closer to type 1 diabetes patients (e.g. lean, early age of diagnosis ~ 48 years) while those with low GAD65 antibodies titer resemble type 2 diabetic patients (obese/overweight, metabolic syndrome, late age of diagnosis ~ 51 years). There is also correlation with the type of HLA susceptibility and GAD antibodies. Certain diabetes-susceptible haplotypes (e.g. DQA1*03-DQB1*0302, DRB1*03-DQB1*0201) are higher in LADA patients compared to those with type 2 diabetes. Even more, these haplotypes are seen in higher frequency in LADA patients with high GAD titers compared to those with low GAD titers and type 2 diabetes. Patients with high GAD 65 antibodies titers also have other autoimmune disease markers, more commonly thyroid antibodies (anti-TPO antibodies).

Patients who are clinically diagnosed as having type 1 or type 2 diabetes could have another much less common type known as MODY (maturity-onset diabetes of the young) (about 1-2% of all patients with diabetes). These are a heterogeneous group of more than 10 gene mutations affecting pancreatic beta cell function and have a varied phenotype. MODY usually first occurs during adolescence or early adulthood. However, sometimes it remains undiagnosed until later in life. Online Mendelian Inheritance in Man (OMIM) now lists 14 forms of MODY. Despite its nonautoimmune etiology, autoimmune markers may be present in MODY. Thus, the presence of GAD65 does not rule out its diagnosis. MODY occurs in young people with an age of onset <25 years, some degree of insulin independence, and a strong (multigenerational) family history of diabetes reflecting its autosomal dominant mode of inheritance.

The most common forms of MODY are transcription factor mutations of the genes of hepatic nuclear factor 1A (HNF1A) and hepatic nuclear factor 4A (HNF4A) and glucokinase (GCK). HNF1A and GCK mutations account for 70% of MODY. HNF1A and HNF4A both are progressive disorders and are associated with complications of diabetes. GCK mutations produce a mild elevations of blood glucose (the set point for glucose sensing is altered) and are not associated with diabetic complications. Despite being infrequent, establishing the correct diagnosis and classification is important in selecting effective treatment, predicting the degree of complications, and providing genetic counseling.

Traditionally, perhaps because of the high cost of genetic testing, criteria for performing genetic testing was limited to an age <25 years and a parental history of diabetes. While this minimizes the chance of excessive negative testing and thus high cost, one may miss patients who do not fit the traditional description. A study of a cohort of both type 1 and type 2 patients using an “extended criteria for testing” concluded that all patients diagnosed up to the age of 30 years, with C-peptide–positive diabetes (random C-peptide >0.2 nmol/L, a glucagon-stimulated C-peptide >0.2 nmol/L performed 3 years after diagnosis) regardless of BMI, family history, GAD65 positivity, or evidence of metabolic syndrome should be tested for MODY gene mutations of HNF1A, HNF4A, and GCK. The other mutations are rare and include genes encoding the Kir6.2 subunit of the beta cell KATP channel. These are of great interest in understanding the spectrum of the etiology of diabetes.

Other disorders that are associated with hyperglycemia include exocrine pancreatic diseases (pancreatitis, pancreatic cancer, hemochromatosis, cystic fibrosis, and fibrocalcific disease), endocrinopathies (hyperthyroidism, Cushing’s syndrome, pheochromocytoma, and others), and medication use (glucocorticoid, thiazides, nicotinic acid, pentamidine, thyroid hormone, beta adrenergic drugs, diazoxide). There are other genetic syndromes, which also include hyperglycemia, however, the primary disorder is usually apparent, including Turner, Down, Klinefelter, and Wolfram syndromes and myotonic dystrophy.

Management and Treatment of the Disease


In this section, the treatment for diabetes will focus on atypical variants with a brief comment on the other forms for the sake of completeness. Every patient with diabetes must be assessed for the treatment of blood pressure (goal 140/80) and cholesterol (LDL cholesterol <100 mg /dl and HDL levels for men and women of <40 and <50 mgdl); also, patient must not smoke and must increase physical activity and adhere to a moderate diet.

Type 1 diabetes

The treatment of type 1 diabetes is insulin. There are numerous insulins available with varying pharmacokinetics administered either as multidose or via continuous infusion.


In the initial stage of the disease, oral agents such as sulfonylureas can be used. However, though sulfonylureas increase insulin secretion, no beta cell protective effects have been seen. Thiazolidinediones (TZD’s) might be useful as well. There are no clinical trials with metformin and there seems to be little benefit obtained from a peripheral insulin sensitizer when the basic pathology is beta cell dysfunction. The use of incretin based therapy has not been tested in these patients.

The mainstay of management is Insulin therapy. It seems that there is a variable time period before one starts insulin therapy in LADA patients. In UKPDS study, it has been shown that in patients younger than 35 years at diagnosis, 84% of GAD positive status required insulin in next 6 years where as at older age of diagnosis (>55 years), the insulin was required in only 34% of GAD positive patients.

Type 2 diabetes

The cornerstone for the treatment of type 2 diabetes is dietary modification and increased physical activity. Pharmacological therapy initially involves oral antidiabetic agent, and in long-term type 2 diabetes, insulin therapy may be required for glycemic control.

Atypical diabetes

When patients are first admitted to the hospital with either marked hyperglycemic crises or DKA, insulin treatment is required. At this stage it is generally not possible nor necessary to classify them as all require immediate insulin even though this need may be transient. Clinical clues to their long-term need for insulin include a long history of insulin use, prior DKA, young age, and lean body habitus, but these are not invariable. Patients should be discharged home on insulin; we use NPH and regular twice a day OR basal bolus insulin (peakless combined with short-acting analogue insulins with meals). They are instructed to check and log fingerstick blood glucoses and maintain ADA targets for capillary blood glucose with a fasting values of 70-130 mg/dl and postprandial values of 140 to 180 mg/dl.

Patients are seen in the outpatient clinic in 1-3 weeks, and the insulin doses are adjusted as needed. Many of these patients may have transient insulin requirement. There are 2 approaches to tapering insulin successfully based on (1) clinical parameters and (2) adequate beta cell function. The latter is determined by measuring the fasting and glucagon stimulated C-peptide sometime after 2 weeks of hospitalization. C-peptide is measured on a fasting blood sample and 5 and 10 min after 1 mg of intravenous glucagon is given; an OGTT can also be used but is much more time consuming. Within the first 2 weeks of hospitalization, beta cell function is highly suppressed and thus testing should be done at some interval after establishing glycemic control to eliminate the effects of glucose toxicity.

1. Clinical parameters suggesting the likelihood of successful discontinuation of insulin are new-onset diabetes, lack of multiple prior admissions for DKA, lack of established diagnosis of type 1 diabetes, overweight/obese, glucose levels at goal while on insulin, and minority status. Age has not been sufficiently specific to count on. These parameters are not inviolate and reflect clinical experience. For example, patients who are lean may also have transient insulin requirements, especially males who are muscular. Insulin should not be tapered or discontinued in patients with a clear diagnosis of type 1 diabetes.

2. Beta cell function suggesting the likelihood of successful discontinuation of insulin has been validated up to 1 year. If the fasting C-peptide is <1 ng/ml and the stimulated peak C-peptide is <1.5 ng/ml, the patient is classified as β- and there is a very low likelihood of discontinuing insulin; if these values are >1.0 or >1.5 ng/ml, respectively, the patient is classified as β+. While these values (using a Linco assay) do not predict discontinuing insulin, they identify the subset of patients who have the potential to do so. Data derived from OGTT also suggested that higher mean C-peptide–to–glucose levels ratios predicted successful tapering; however, when treating a patient, this must be individualized. Indeed, it is not always possibly to predict the rate and extent of recovery of an individual soon after the onset of the acute event as this may vary considerably among patients.

Tapering insulin: Regardless of whether the approach is based on empirically lowering the insulin or first assessing C-peptide as a measure of adequate beta cell function, if the blood glucoses are within the target range, we taper the patient’s insulin gradually (beginning with ~30% but this decrease should be individualized). This process is continued with monthly visits until patients are able to discontinue insulin. During the process, patients monitor their fingerstick capillary blood glucoses and, if the blood glucoses increase above 200 mg/dl, are instructed to test their urine for ketones. Usually, 3-4 months are required to taper patients, although some patients report successful abrupt discontinuation of insulin without increases in blood glucoses. The key is to monitor frequently and adjust insulin as needed. If after a period off insulin, the blood glucose rises, patients are started on oral antidiabetic agents, generally starting with metformin (link to outpatient insulin therapy). An important clinical point is that patient may relapse to hyperglycemia; therefore, patients should be informed and have appropriate follow-up care.


Treatment of MODY patients should be on the specific type of genetic mutation. Patients are initially considered to have type 1 or type 2 diabetes and are treated with insulin since they develop diabetes in youth. Patients with HNF1-α and HNF4-α may do very well with sulfonylurea treatment. However, they may also have diabetes as severe as type 1. Patients with the glucokinase mutation (which encodes an enzyme in the glucose-sensing mechanism of the pancreatic beta cell) may do well with diet treatment alone. Other mutations are rare and treatment should be individualized.

Gestational diabetes

Treatment of diabetes in pregnancy is based upon established glycemic goals designed to decrease maternal and fetal adverse effects. In practice, whether the patient is identified at the first prenatal visit (overt or pregestational diabetes) or during oral glucose tolerance testing during the 24th-28th week of pregnancy, traditional treatment is insulin. NPH and regular insulin do not cross the placenta and have a long safety record. Recently the use of analogue insulins have proven safe and effective. These analogues include detemir insulin (long acting) and lispro and aspart (short acting) insulins. In patients with milder cases of diabetes, metformin and glyburide (do cross the placenta) are increasingly used to achieve glycemic targets in the US and studies have not shown significant adverse effects.