Nonenzymatically glycosylated proteins gradually form fluorescent cross-linked protein adducts--a process termed "browning." The rate of this reaction increases with the glucose concentration. Assaying for the presence of browning products in long-lived proteins should therefore provide information on long-term metabolic control. We measured collagen-linked fluorescence typical for nonenzymatic browning in skin-biopsy specimens from 41 subjects with longstanding Type I diabetes and from 25 controls. Fluorescence correlated with age and (weakly) with the duration of diabetes. Mean age-adjusted fluorescence values were twice as high in diabetic subjects as in control subjects (P less than 0.0001) and increased with the severity of retinopathy, nephropathy, and arterial and joint stiffness. The correlation was significant for retinopathy (r = 0.42; P less than 0.01), arterial stiffness (r = 0.41; P less than 0.01), joint stiffness (r = 0.34; P less than 0.05), and the sum of all complications (r = 0.47; P less than 0.01). Fluorescence also correlated with systolic (r = 0.42; P less than 0.01) and diastolic (r = 0.36; P less than 0.05) blood pressures. If one can assume that the fluorescence results from a browning product of glucose, our data suggest that there is an overall correlation between the severity of diabetic complications and cumulative glycemia over many years.

Nonenzymatic glycosylation (glycation) of collagen was measured by boronate affinity chromatography in skin biopsies from 41 type I diabetics with mean duration of diabetes of 25 yr (range 20-40 yr) and from 25 age-matched controls. Mean level of Amadori products was significantly increased in diabetic [7.85 +/- 1.78 (SD) nmol/mg collagen] versus control subjects [3.34 +/- 1.06 (SD) nmol/mg collagen, P less than .001] but did not correlate with age, diabetes duration, or severity of retinopathy, nephropathy, arterial stiffness, and joint stiffness. Similarly, mean collagen content per biopsies was 42% increased in diabetic versus control subjects (P less than .001) but did not correlate with age, diabetes duration, or severity of complications. A weak but positive correlation between glycohemoglobin level and glycation of skin collagen was observed. These results indicate that Amadori products cannot explain by themselves the pathogenesis of diabetic complications unless individual tissue response to glycation is different in subjects with and without complication. They do not exclude a role for the late stages of the Maillard reaction, nonenzymatic browning, in the formation of some of these complications.

The age-related changes in collagen-linked fluorescence (browning) were investigated in skin from subjects with long-standing type I diabetes. Overall browning rates were 2A times higher in diabetics than in controls (P < 0.02) and slope intercept accurately reflected the mean age of onset of diabetes (11.6 vs. 11.2 yr), suggesting that the browning process has the attributes of a biological clock. Browning rates were not different in controls and diabetics without retinopathy (P > 0.05) but were 2.4 (P < 0.05) and 2.7 (P < 0.01) times increased in the presence of background and proliferative retinopathy, respectively. Compared with subjects with retinopathy, individual browning rates since onset of diabetes decreased with advancing age in subjects free ofretinopathy (P < 0.001). Extrapolation revealed that they would become identical to that of nondiabetic subjects by the age of 66A yr.These results suggest the presence of a mechanism that controls the browning rate ofcollagen in diabetics who do not develop retinopathy.