September 8, 1997

Cumulative Glycemic Exposure and Microvascular Complications in Insulin-Dependent Diabetes MellitusThe Glycemic Threshold Revisited

Author Affiliations

From the Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh (Drs Orchard and Forrest), and the Division of Nephrology and Endocrinology, Children's Hospital of Pittsburgh (Drs Ellis and Becker), Pittsburgh, Pa.

Arch Intern Med. 1997;157(16):1851-1856. doi:10.1001/archinte.1997.00440370091009

Background:  The development of microvascular insulindependent diabetes mellitus (IDDM) complications has been shown to be related to both duration of diabetes and the degree of glycemic exposure. However, controversy exists as to whether there is a threshold of glycemic exposure, below which there is minimal risk. Furthermore, there are few data describing the relationship of total glycemic exposure (duration x degree) to complications rates—a potentially useful research and clinical tool.

Objectives:  To determine a cumulative glycemic exposure variable that combines the effect of both degree and duration of hyperglycemia and to evaluate this variable in terms of its relation to microvascular complications. The association between cumulative glycemic exposure and complication risk was also examined to evaluate whether there was a threshold effect.

Methods:  A total of 353 patients with IDDM who had completed the first 6 years of follow-up in the Pittsburgh Epidemiology of Diabetes Complications Study were included in this analysis. These subjects had a mean age of 27.9 years, and the mean duration of the disease was 19.4 years. Subjects were examined at baseline (cycle 1) and then biennially (cycle 2, cycle 3, and cycle 4) for diabetes complications. Total glycosylated hemoglobin (HbA1) was measured at each cycle. A cumulative glycemic exposure variable, named A1months, was calculated by multiplying the number of HbA1 units above normal at each cycle by the number of months between the midpoints of the preceding and succeeding cycle intervals.

Results:  The mean number of A1months experienced at the time of diagnosis of proliferative retinopathy (914), microalbuminuria (952), overt nephropathy (1043), and distal symmetrical polyneuropathy (1043) did not vary by duration of diabetes. Thus, approximately 1000 A1months were needed (on average) for the advanced complications to develop. Although the risk for developing proliferative retinopathy rose gradually as A1months increased, a more abrupt increase in the risk was seen (again at approximately 1000 A1months) for microalbuminuria (odds ratio, 6.9; 95% confidence interval, 2.5-19.1), overt nephropathy (odds ratio, 6.5; 95% confidence interval, 2.0-21.7), and distal symmetrical polyneuropathy (odds ratio, 6.5; 95% confidence interval, 2.4-17.8). Nonetheless, complications developed in the majority of cases at glycemic exposures below 1000 A1months. The cumulative glycemic exposure variable A1months does not predict complications any better than its component variables (duration and HbA1). Furthermore, formal statistical testing failed to show a definitive threshold for any complication.

Conclusions:  Although A1months does not enhance prediction of complications, it may be a useful summary measure of glycemic exposure for both patients and physicians. However, although subjects with 1000 A1months or more appear to be at increased risk of developing most microvascular complications, because the majority of complications arise in subjects with less than this exposure, this threshold value should only be considered a minimal goal. For example, our data suggest that for most microvascular complications to develop, it would take, on average, 83 years with an HbA1 unit at 1% above normal, 42 years at 2% above normal, 28 years at 3% above normal, 21 years at 4% above normal, and 18 years at 5% above normal.Arch Intern Med. 1997;157:1851-1856