OBJECTIVE—The pathogenesis of diabetic peripheral neuropathy (DPN) is poorly understood. We have recently reported a significant reduction in spinal cord cross-sectional area at the stage of clinically detectable DPN. In this study, we investigated whether spinal cord atrophy occurs in early (subclinical) DPN.

RESEARCH DESIGN AND METHODS—Eighty-one male type 1 diabetic subjects, 24 nondiabetic control subjects, and 8 subjects with hereditary sensory motor neuropathy (HSMN) type 1A underwent detailed clinical and neurophysiological assessments. Diabetic subjects were subsequently divided into three groups based on neuropathy severity (19 with no DPN, 23 with subclinical DPN, and 39 with clinically detectable DPN). All subjects underwent magnetic resonance imaging of the cervical spine and cord area measurements at disc level C2/C3.

RESULTS—Mean corrected spinal cord area index (SCAI) (corrected for age, height, and weight) was 67.5 mm [95% CI 64.1–70.9] in diabetic subjects without DPN. Those with subclinical (62.4 mm [59.5–65.3]) and clinically detectable DPN (57.2 mm [54.9–59.6]) had lower mean SCAIs compared with subjects with no DPN (P = 0.03 and P < 0.001, respectively). No significant difference was found between diabetic subjects without DPN and nondiabetic control subjects (69.2 mm [66.3–72.0], P = 0.47). Mean SCAIs in subjects with HSMN type 1A (71.07 mm [65.3–76.9]) were not significantly different from those for nondiabetic control subjects and diabetic subjects without DPN. Among diabetic subjects, SCAI was significantly related to sural sensory conduction velocities and the Neuropathy Composite and Symptom Scores.

CONCLUSIONS—Spinal cord involvement occurs early in DPN. There is also a significant relation between reduction in SCAI and neurophysiological assessments of DPN.