1998 clinical practice guidelines for the management of diabetes in Canada

Supplement to CMAJ 1998;159 (8 Suppl)

© 1998 Canadian Medical Association

In 1995, an international expert committee working under the auspices of the American Diabetes Association (ADA) was established to review the National Diabetes Data Group's (NDDG's) 1979 classification and diagnostic criteria for diabetes⁴¹ (adapted by the CDA in 1982) in view of more recent reported findings. This review culminated in a new classification and set of diagnostic criteria⁴² that were adopted by the ADA and that are likely to be adopted by the World Health Organization (WHO). After reviewing the new classification and all of the relevant published reports, our Expert Committee concluded that it was also appropriate for Canada to adopt these changes.

Definition of diabetes

Diabetes mellitus is a metabolic disorder characterized by the presence of hyperglycemia due to defective insulin secretion, insulin action or both. The chronic hyperglycemia of diabetes mellitus is associated with significant long-term sequelae, particularly damage, dysfunction and failure of various organs — especially the kidney, eye, nerves, heart and blood vessels.

Classification

The major changes from the earlier NDDG classification and WHO recommendations are summarized as follows (Table 3):

• The new classification proposes elimination of the terms "insulin-dependent diabetes mellitus" (IDDM) and "non-insulin-dependent diabetes mellitus" (NIDDM), but retention of "type 1" and "type 2" diabetes (using Arabic rather than Roman numerals). The IDDM–NIDDM terminology was based on treatment rather than pathogenesis and caused considerable confusion. People with any form of diabetes may require insulin treatment at some stage of their disease. Such use of insulin does not, in itself, classify the patient.
• Type 1 diabetes encompasses diabetes that is primarily a result of pancreatic beta-cell destruction and that is prone to ketoacidosis. This form includes cases due to an autoimmune process and those for which the etiology of beta-cell destruction is unknown.
• Type 2 diabetes may range from predominant insulin resistance with relative insulin deficiency to a predominant secretory defect with insulin resistance.
• A wide variety of relatively uncommon conditions are listed under "other specific types" (Table 3). These consist mainly of specific genetically defined forms of diabetes or diabetes associated with other diseases or drug use.
• The classification of gestational diabetes mellitus (GDM) remains unchanged. GDM refers to glucose intolerance with onset during pregnancy.⁴³

Diagnostic criteria

Previously used diagnostic criteria (level of fasting venous plasma glucose >=7.8 mmol/L) lacked sensitivity; a significant proportion of people in whom a diagnosis of diabetes would have been made based on glucose level 2 h after a 75-g glucose load never received this test and, thus the diagnosis was not made.^42,44 The diagnostic threshold of 11.1 mmol/L used for the 2-h sample in the oral glucose tolerance test (OGTT) was based on the risk of microvascular diabetic complications developing; the arbitrarily chosen fasting plasma glucose (FPG) threshold of 7.8 mmol/L actually defined a greater degree of hyperglycemia than did the 2-h plasma glucose (2hPG) threshold of 11.1 mmol/L.⁴²

A recent re-evaluation of population studies suggests that an FPG level of 7.0 mmol/L correlates most closely with a 2hPG level of >=11.1 mmol/L and best predicts the development of microvascular disease.^42,45,46 The lowering of the FPG diagnostic level from 7.8 to 7.0 mmol/L ensures that both the FPG and 2hPG define a similar degree of hyperglycemia and risk for microvascular disease. It also permits the diagnosis of diabetes to be made on the basis of a commonly available test — the FPG.

Although below the diabetic thresholds, FPG levels between 6.1 and 7.0 mmol/L are abnormally high; people with FPG levels in this range are considered to have "impaired fasting glucose" (IFG).⁴² Although they do not have the diabetes-associated risk for microvascular disease, they and people with "impaired glucose tolerance" (IGT) have a higher risk for the development of diabetes mellitus and cardiovascular disease than the general population. Preventive strategies involving lifestyle changes and increased frequency of screening for diabetes mellitus should be a priority for these people.⁴⁴ The long-term outcome and economic ramifications of identification of these new subgroups have yet to be assessed.

Recommendations

6. The specific fasting plasma glucose (FPG) level used to diagnose diabetes should be reduced from 7.8 to 7.0 mmol/L. [Grade A, Level 1^42,45,46]
7. The term "impaired glucose tolerance" (IGT) is retained, but now depends only on measurement of plasma glucose 2 h after a 75-g glucose load (2hPG). [Grade D, consensus]
8. The term "impaired fasting glucose" (IFG) should be established to identify another intermediate stage of abnormal glucose homeostasis. [Grade D, consensus]
9. Both IGT and IFG indicate a need for annual testing and attention to associated risk factors and lifestyle changes. [Grade D, consensus]

The diagnostic criteria for diabetes and the glucose thresholds for other diagnostic categories are summarized in Table 4 and Table 5. These criteria are based on venous sample methods in the laboratory. Although the frequency distributions of hemoglobin A_1c (HbA_1c) levels in some studies have characteristics similar to those obtained from FPG and 2hPG tests,^45,46 the lack of standardization of the HbA_1c test precludes its use in the diagnosis of diabetes.

Screening for gestational diabetes

GDM occurs in 2% to 4% of all pregnancies, and the diagnosis of GDM has implications for both the baby and mother. The established morbidity for the baby includes macrosomia (with the risk of fetal and maternal trauma during birth) and neonatal hypoglycemia;⁴⁷ other consequences are now rare. Although the value of diagnosing and treating GDM has been questioned,^48,49 recent cost–benefit analyses^50,51 have demonstrated the value of treating this condition primarily due to decreased costs for care of the newborn. The value of identifying a mother who is at high risk for later diabetes remains unproven; however, the incidence of postpartum diabetes mellitus, IGT and lipid abnormalities is elevated. Recognition of those at risk would allow application of preventive strategies, such as changes in nutrition and physical activity, that might help minimize the progression to more significant disease.

Due to its high prevalence and impact and because clinical criteria cannot reliably identify those with GDM, all pregnant women should be screened for GDM unless they are in a very low-risk group. Thus, screening should be carried out for all women over 25 years of age, as well as any woman under age 25 who is obese, belongs to an ethnic group predisposed to diabetes (e.g., Aboriginal people and people of Hispanic, Asian and African descent), has a family history or previous history of diabetes or has a history of giving birth to babies with a birthweight over 4 kg.^43,52 (Low-risk people include lean Caucasian woman under age 25 years, with no personal or family history of diabetes and no history of large babies.) The evidence regarding glucose levels for screening remains unclear;⁵² therefore, no changes were made in this area.

Recommendations

10. All pregnant women should be screened for gestational diabetes mellitus (GDM) between 24 and 28 weeks' gestation, with the exception of those in a very low-risk group. [Grade D, consensus]
11. The preferred screening test is measurement of plasma glucose level 1 h after a 50-g oral glucose load given at any time of day.
    • If the level at 1 h is >=7.8 mmol/L, a glucose tolerance test is warranted.
    • If the level at 1 h is >=10.3 mmol/L, then GDM can be diagnosed. [Grade B, Level 2⁵³]

Diagnosis of gestational diabetes

The worldwide diversity of criteria for the diagnosis of GDM continues to be problematic. The NDDG criteria were based on original data (100-g, 3-h oral glucose test) that predicted long-term risk of diabetes in the mother, and these levels of hyperglycemia were then found to correlate with neonatal morbidity. WHO criteria (using a 75-g, 2-h test) aim for uniformity with the nonpregnant state, but suggest treating IGT when found in pregnancy.⁵⁴

Two studies^55,56 involving over 4000 patients in total provide statistical normative data using a 75-g, 2-h glucose test. Given the ease of the 75-g, 2-h test in terms of less nausea, less time for the patient and cost savings, and in view of the normative data available in support of it, it should be recommended as the optimum test. The fasting level found in the 2 studies was slightly different, and that derived by Carpenter and Coustan⁵⁷ by the conversion of the original O'Sullivan and Mahan data⁵⁸ lies between these 2 values. Thus, it was felt that a fasting glucose threshold of 5.3 mmol/L, which best identifies a high risk for macrosomia,^57,59 should be used.

Finally, these recommendations are recognized as interim ones in the absence of clear evidence. For the next revision of the Canadian clinical practice guidelines, we hope that the criteria for the diagnosis of GDM will be based on firm outcome data.

Recommendations

12. GDM should be diagnosed by measuring FPG level and plasma glucose levels at 1 and 2 h after ingesting a 75-g glucose load. If 2 of the following 3 values are met or exceeded, a diagnosis of GDM is established. If only 1 value is met or exceeded, the diagnosis is impaired glucose tolerance of pregnancy:
    

    Fasting	>5.3 mmol/L
    1 h	>10.6 mmol/L
    2 h	>8.9 mmol/L

[Grade D, consensus]

In view of the common use of the 100-g OGTT during pregnancy, a 100-g glucose load may be used in carrying out a diagnostic test and measuring following values as recommended by the ADA.⁶⁰

Screening for type 2 diabetes

Approximately 3% to 5% of the general adult population has unrecognized type 2 diabetes.⁶¹ Tests for hyperglycemia can identify these people, and this may result in significant benefit because many of them will have or will be at risk for preventable diabetic complications.⁶² Routine testing for type 2 diabetes is, therefore, justifiable as a routine clinical activity in some, but not all, settings.⁶³ Thus, although the relatively low prevalence of diabetes in the general population makes it unlikely that mass screening will be cost–beneficial, testing for diabetes in people with risk factors for type 2 diabetes or with diabetes-associated conditions is likely to result in more good than harm and will lead to overall cost savings.⁸

Several widely available tests for hyperglycemia have been assessed in the context of diabetes screening. These include FPG and casual plasma glucose levels, 2hPG level during an OGTT, glycated hemoglobin level and glycosuria assessment.^62,63 The FPG is the most reliable of these tests, although each has advantages and disadvantages in terms of convenience, cost, assay standardization and reliable identification of people for whom further evaluation and treatment are worthwhile. Thus, routine use of OGTTs or measurement of insulin levels to identify people at high risk for type 2 diabetes is not necessary.

Recommendations

13. Mass screening for type 2 diabetes in the general population is not recommended. [Grade D, consensus]
14. Testing for diabetes using a FPG test should be performed every 3 years in those over 45 years of age. [Grade D, consensus]
15. More frequent or earlier testing (or both) should be considered in those with additional risk factors for diabetes, i.e.,
    • a first-degree relative with diabetes
    • member of high-risk population (e.g. Aboriginal people, Hispanic, Asian and African descent)
    • obesity
    • a low level of high-density lipoprotein (HDL) cholesterol (¾0.9 mmol/L) or an elevated fasting level of trigylcerides (>2.8 mmol/L). [Grade D, consensus]
16. Annual testing should be considered in those with one or more of the following more predictive risk factors (irrespective of the above factors), such as
    • history of IGT or IFG
    • presence of complications associated with diabetes mellitus
    • history of GDM or baby with birthweight over 4 kg
    • presence of hypertension
    • presence of coronary artery disease (CAD). [Grade D, consensus]

Preventing type 2 diabetes

Prospective cohort studies have identified historical, physical and biochemical variables that are associated with subsequent type 2 diabetes. These variables include older age, certain ethnic backgrounds, obesity (especially central obesity), physical inactivity, a history of GDM, overt coronary artery disease, high fasting insulin levels and IGT.^64,65,66 Randomized studies testing behavioural modification and sulfonylureas in the prevention of type 2 diabetes in people with IGT have followed these observations.^67,68,69,70 To date, no effective treatments have been identified, with the exception of a program of diet and exercise that yielded a clinically significant absolute risk reduction (about 25%) in the rate of diabetes over 6 years in a trial conducted in Da Qing, China.⁶⁹

A large, randomized trial of behaviour modification (Diabetes Prevention Trial) being carried out by the National Institutes of Health should confirm and extend the generalizability of the Da Qing study.⁶⁹ In the meantime, in view of the promising results of that study and the accepted value of weight control, diet and exercise in reducing cardiovascular risk, these activities should be promoted. This recommendation can be made irrespective of one's risk for type 2 diabetes, including risk according to laboratory parameters. Ongoing preventive trials will also assess the efficacy and safety of other interventions including intensified lifestyle change, metformin, troglitazone and acarbose.^65,66,71,72 However, until these trials are completed, the use of pharmacologic treatments to prevent type 2 diabetes remains experimental.

Recommendation

17. In those at increased risk, a program of weight control through diet and regular exercise is recommended and may prevent type 2 diabetes. [Grade B, Level 1⁶⁹]

Preventing type 1 diabetes

The common form of type 1 diabetes is marked by immune-mediated loss of pancreatic beta-cells. This process is incited by an interaction between genetic and environmental factors. During the asymptomatic phase, ongoing autoimmune destruction of the pancreatic beta-cells occurs. This can be identified reliably in first-degree relatives by screening for immune abnormalities, such as islet-cell antibodies, and later by metabolic abnormalities, namely reduced first-phase insulin secretion measured by an intravenous glucose tolerance test.

Two strategies to prevent the disease are, therefore, possible: altering environmental factors in people who are genetically at risk (primary prevention)⁷³ or modifying the immune process in people with subclinical beta-cell loss identified by positive screening tests (secondary prevention).^74,75 Randomized trials of primary prevention (i.e., removal of cow's milk protein from infant feeds) and secondary prevention (i.e., nicotinamide, oral insulin and parenteral insulin) have been initiated. However, at present no preventive measures for the disease are known to be effective and safe. In the case of primary prevention, the strength of evidence for a causal link between early exposure to cow's milk protein and subsequent type 1 diabetes is not adequate to recommend that cow's milk be routinely proscribed from infant feeds (although breast feeding can be recommended given its other benefits).

Clinically important benefits of screening for prodromal type 1 diabetes are not established, and such screening carries the potential for negative psychosocial effects (specifically, informing otherwise well people that they are at increased risk for a debilitating chronic disease may do more harm than good).

Recommendation

18. Attempts to prevent type 1 diabetes — either by manipulating environmental factors or by treating people at high risk — are experimental and should be confined to formal research projects. [Grade D, consensus]