NO EVIDENCE FOR AXONAL ATROPHY IN HUMAN DIABETIC POLYNEUROPATHY

In rats with streptozin-induced diabetes mellitus, the caliber of dist al myelinated fiber (MF) axons in relation to the number of myelin lam ellae is smaller than in controls. This finding usually has been attri buted to axonal atrophy, but shrinkage or maldevelopment has also been considered. For human diabetic polyneuropathy (DP), axonal atrophy ha s been assumed by some investigators, but convincing evidence has not been demonstrated. We morphometrically evaluated transverse sections o f 33 sural nerves from carefully evaluated diabetic patients greater t han or equal to 30 years old without (8 patients) or with (25 patients ) DP and compared them with 24 nerves from healthy subjects greater th an or equal to 30 years old. Nerves from diabetic patients and control s were obtained under identical conditions and processed and evaluated in the same way, using an observer blind to the disease condition. Us ing computer digitization of electron micrographs, we evaluated the ax onal area, perimeter, index of circularity, number of myelin lamellae; and frequency of adaxonal sequestration of 50.4 (mean) +/- 5.8 (SD) M F per sural nerve for healthy subjects and diabetic patients greater t han 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 regress ion lines of nerves of healthy subjects for large MFs-the most reliabl e 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 si gnificantly 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 axo nal atrophy occurs in human DP. For small ME or all ME some significan t differences in regression lines of In axonal area or IC on number of lamellae were found, but these changes are probably explained by even ts of remyelination and axonal regeneration, which can affect these re lationships and are known to occur in DP.