Digital electron microscopic examination of human sural nerve biopsies

Diabetic peripheral polyneuropathy is characterized by axonal degeneration and regeneration as well as by Schwann cell and microvascular changes. These changes have been described at both the light (LM) and the electron microscopic (EM) levels; however, EM has not been applied to large clinical trials. Our goal was to adapt the rigorous techniques used for quantifying human biopsies with LM image analysis to accommodate ultrastructural analyses. We applied digital image capture and analysis to the ultrastructural examination of axons in sural nerve biopsies from diabetic patients enrolled in a multicenter clinical trial. The selection of sural nerve biopsies was based on the quality of specimen fixation, absence of physical distortion, and nerve fascicle size (> or =100,000; < or =425,000 microm2). Thin sections were collected on formvar-coated slot grids, stabilized with carbon and scanned on a Phillips CM100 transmission electron microscope. Digital images were captured with a Kodak Megaplus 1.6 camera. A montage was constructed using software derived from aerial mapping applications, and this virtual image was viewed by EM readers. Computer-assisted analyses included identification and labeling of individual axons and axons within regenerating clusters. The average density of regenerating myelinated axon clusters per mm2 was 65.8 +/- 5.1, range of 0-412 (n = 193). These techniques increase the number of samples that may be analyzed by EM and extend the use of this technique to clinical trials using tissue biopsies as a primary endpoint.