Although nobody has ever actually seen one [1], an elevated blood sugar is a very real thing to the patient with diabetes, to the physician or other medical person helping that patient deal with the disease, and to the epidemiologist trying to unravel the intricate mysteries of disordered blood sugars in populations. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus has now published a report, the result of 2 years of intensive labor, that lays out new criteria for assigning meaning to blood sugars [2]. Why did the group undertake this work? Why now? And why should we care?
The last time an expert group in the United States put forth guidelines for diagnosing diabetes in individual patients and classifying populations as normal and abnormal with respect to blood sugar levels was 1979-nearly 20 years ago. In that year, the National Diabetes Data Group formally recognized two principal abnormal metabolic states: insulin-dependent diabetes mellitus, or type 1 diabetes, and non-insulin-dependent diabetes mellitus, or type 2 diabetes, while also recognizing gestational diabetes, malnutrition-related diabetes, and other types. The group provided specific diagnostic and classification schemes based on the best information then available, using the predictive-that is, diagnostic-power of blood sugar levels. This scheme was adopted in 1980 by the World Health Organization (WHO) Expert Committee on Diabetes and later by the WHO Study Group on Diabetes Mellitus.
As suggested by others [3, 4], the most obvious reason for developing new criteria was to take advantage of the new and better information about the predictive power of blood sugar levels that has become available during the past 20 years. Although the blood sugar level itself has long been appreciated as a predictor of disease risk, blood sugar levels fluctuate substantially, from hour to hour and day to day, which has always caused major sampling problems. The recognition nearly two decades ago that glycosylated hemoglobin, most prominently hemoglobin A1c, provides a cumulative or "integral" measure of blood sugar over time has made it possible to avoid many of the problems of "differential" blood sugar sampling, that is, sampling at a single moment. The Expert Committee was not yet prepared to substitute the hemoglobin A1c level for the blood sugar level as a diagnostic criterion, largely because of difficulties in standardizing its measurement. The Committee did, however, find a statistical relation between glycosylated hemoglobin and blood sugar levels as predictors of risk in both individual patients and populations. The Expert Committee also decided that the time had come to rationalize the nomenclature for diabetes, arguing that the classification of diabetes should now reflect its underlying genetic, environmental, and pathophysiologic mechanisms, rather than empirical therapy, that is, the patient's need (or lack of need) for exogenous insulin. The impressive advances in understanding these disease mechanisms during the past 20 years now make "mechanism-based" diagnosis a realistic possibility in some few patients, although this rigorous approach is not yet a practical reality for most.
The Committee believed that the time had come to abandon the oral glucose tolerance test except in special circumstances, such as pregnancy, because the experience of the past two decades tells us it has been little used, even with individual patients, and is difficult to implement in population studies. The small amount of information lost thereby was thought to be more than compensated for by the increased number of persons who could be tested under a new and simplified testing scheme.
Finally, the evidence accumulated in the past 20 years convinced the Committee that the predictive powers of the two diagnostic thresholds for diabetes chosen in 1979-that is, a fasting plasma glucose level of 140 mg/dL and a 2-hour postprandial glucose level of 200 mg/dL in an oral glucose tolerance test-are not equivalent. By the time a patient's fasting glucose level is greater than 140 mg/dL, the 2-hour postprandial glucose level is almost always greater than 200 mg/dL; in contrast, only about one quarter of persons (without previously known diabetes) whose 2-hour postprandial glucose level exceeds 200 mg/dL have a fasting glucose level that exceeds 140 mg/dL. The time had come to choose a single best threshold.
Thus motivated, the Committee proposed the following key changes in the criteria for diagnosing and classifying diabetes.
First, the diagnosis in individual patients should now depend on 1) finding any one of three abnormalities: symptoms of diabetes plus a casual plasma glucose level of 200 mg/dL (11.1 mmol/L), a fasting plasma glucose level of 126 mg/dL [7.0 mmol/L], or a 2-hour plasma glucose level of 200 mg/dL during an oral glucose tolerance test using a glucose load that contains the equivalent of 75 g of anhydrous glucose dissolved in water; and 2) confirming the diagnosis by finding any one of the three abnormalities again by repeated testing on a different day. For population studies, a fasting plasma glucose level of 126 mg/dL is recommended as the single disease-defining measure.
Second, for the reasons mentioned above, the expressions insulin-dependent diabetes mellitus and non-insulin-dependent diabetes mellitus are no longer considered acceptable. The terms type 1 diabetes and type 2 diabetes are retained, however, because these categories capture clinically important differences in pathophysiology: The former refers to patients whose metabolic lesion is absolute insulin deficiency, the latter to patients (the great majority) in whom insulin is present but in insufficient amounts to accommodate a simultaneous increase in insulin need (that is, insulin "resistance").
Third, gestational diabetes continues to be recognized as an important clinical diagnosis. For now, the Committee continues to recommend that, for most women, a screening procedure be performed between 24 and 28 weeks of gestation. The process starts with a 50-g oral glucose load, and women with a 1-hour postprandial glucose level that exceeds 140 mg/dL on this initial screen need to have a full oral glucose tolerance test with a 100-g oral load. Screening is deemed unnecessary in women in certain low-risk groups; at the same time, the prevalence of diabetes is so high in certain other populations that women in those groups should probably proceed directly to the full oral glucose tolerance test.
Fourth, the important category of impaired glucose tolerance has been retained. The sole diagnostic criterion for the condition is newly defined as a fasting plasma glucose level greater than 110 mg/dL but less than 126 mg/dL, for which the inelegant term impaired fasting glucose has been assigned. Patients in this group may have a slightly increased risk for developing the chronic microvascular and macrovascular complications of diabetes but, perhaps more important, are at an increased risk for developing overt diabetes over time.
Finally, the term malnutrition-related diabetes has been dropped from the lexicon. The evidence that diabetes can be directly related to protein deficiency is simply insufficient to support this as a separate clinical entity.
We should care about these changes primarily because they are reasonable and helpful, both physiologically and clinically.
Thus, classification of disease in western medicine has generally moved from the use of broad, outward manifestations, such as fever, rash, or arthritis, toward groupings based on specific causes and pathophysiologic changes. It was time for diabetes to do the same, although the journey down that road has just begun.
Linking diagnostic threshold values to the risk for microvascular and macrovascular complications, both directly, by using prospective population blood sugar data, and indirectly, by using the more stable surrogate measure of glycosylated hemoglobin, is likewise sensible.
Aligning the diagnostic power of fasting and postprandial glucose values by shifting the diagnostic threshold for the fasting level downward is reasonable. The overall effect of relying primarily on a new and lower level of fasting glucose on the apparent prevalence of diabetes is not yet clear, however. Postprandial glucose levels are more sensitive to the presence of the disease than are fasting levels because even small decreases in insulin secretory reserve impair the ability to handle a glucose load, while fasting insulin secretory capacity is the last to go. On the one hand, therefore, minimizing the importance of postprandial glucose levels should result in a loss of diagnostic sensitivity. On the other hand, dropping the fasting glucose diagnostic threshold from 140 to 126 mg/dL should make the diagnostic process more sensitive. Will these two changes balance each other out? Only time and experience will tell, but, interestingly, the Committee estimates that the net change will be a small decrease in the overall prevalence of diabetes, largely because the more sensitive oral glucose tolerance test, which has been used to determine prevalence in populations, will no longer be the standard.
Anything that simplifies the testing procedures and criteria without, apparently, losing a substantial amount of diagnostic information also makes good sense.
Of course, the good and rational guidance on diagnosis and classification provided by the Report is just the starting point for many important clinical issues. For one, the Report does not explicitly address the vexing question of blood sugar elevations in patients stressed by acute illness. Because even severe illness is usually not accompanied by blood sugar elevations, it is probably reasonable to infer that patients whose blood sugar levels are increased when very ill are probably at higher-than-average risk for future diabetes, even if the blood sugar level returns to normal when the illness has resolved.
For another, the Report does not (and cannot) go deeply into the risks and benefits in various patient groups, such as the likelihood that the various complications will develop; the efficacy (and effectiveness) of current therapies in delaying or preventing complications; and the costs, both financial and emotional, involved in testing and treatment. It does, however, underscore the insidious and potentially serious consequences of even mild, asymptomatic diabetes, giving sensible guidance on identifying the risk factors that should encourage earlier or more frequent testing in asymptomatic persons not known to have diabetes.
Copies of the Report are available from Richard Kahn, MD, American Diabetes Association, 1660 Duke Street, Alexandria, VA 22314. Further information is also available on the World Wide Web (http://www.diabetes.org).
Top
References
Article
