No evidence for axonal atrophy in human diabetic polyneuropathy

In rats with streptozin-induced diabetes mellitus, the caliber of distal myelinated fiber (MF) axons in relation to the number of myelin lamellae is smaller than in controls. This finding usually has been attributed to axonal atrophy, but shrinkage or maldevelopment has also been considered. For human diabetic polyneuropathy (DP), axonal atrophy has been assumed by some investigators, but convincing evidence has not been demonstrated. We morphometrically evaluated transverse sections of 33 sural nerves from carefully evaluated diabetic patients greater than or equal to 30 years old without (8 patients) or with (25 patients) DP and compared them with 24 nerves from healthy subjects greater than or equal to 30 years old. Nerves from diabetic patients and controls were obtained under identical conditions and processed and evaluated in the same way, using an observer blind to the disease condition. Using computer digitization of electron micrographs, we evaluated the axonal area, perimeter, index of circularity, number of myelin lamellae; and frequency of adaxonal sequestration of 50.4 (mean) +/- 5.8 (SD) MF per sural nerve for healthy subjects and diabetic patients greater than or equal to 30 years old. The regression lines of the natural log (In) of axonal area on number of myelin lamellae of diabetic patients (with or without DP) were not significantly different from the regression lines of nerves of healthy subjects for large MFs-the most reliable group in which to recognize atrophy. Likewise, the regression lines of index of circularity (IC) (an index that is decreased with atrophy or shrinkage) on number of myelin lamellae for large fibers was not significantly different between the disease and control groups. The rate of adaxonal sequestration was not significantly higher in DP than in healthy subjects. These results do not support the hypothesis that axonal atrophy occurs in human DP. For small ME or all ME some significant differences in regression lines of In axonal area or IC on number of lamellae were found, but these changes are probably explained by events of remyelination and axonal regeneration, which can affect these relationships and are known to occur in DP.