Background <p>Objective and easily applicable biomarkers for diabetic polyneuropathy (DPN) are warranted. Circulating nerve-specific proteins have emerged as valuable biomarkers for central nervous system disease but few of these have been tested in peripheral neuropathy. Glial Fibrillary Acidic Protein (GFAP) is highly expressed in non-myelinating Schwann cells while UCH-L1 is a neuron expressed stress protein not previous analyzed in DPN. In this pilot study, we explore serum GFAP and UCH-L1 levels in patients with/without DPN and controls.</p> Methods <p>Persons with DPN (<i>n</i> = 28), without DPN (<i>n</i> = 31), and controls (<i>n</i> = 30) were evaluated in a cross-sectional design. Sural nerve conduction (velocity and amplitude) was evaluated by NC-stat DPNCheck™ and quantitative sensory testing of cold detection and pain was performed. GFAP and UCH-L1 levels were compared across study groups and the unadjusted correlation with nerve assessments evaluated.</p> Results <p>Serum GFAP were lower in persons with DPN (20.9 ± 10.9 pg/ml) than in persons without DPN (26.2 ± 14.1 pg/ml) (<i>p</i> = 0.04) or controls (31.7 ± 26.0 pg/ml) (<i>p</i> = 0.02). GFAP levels were not different in persons without DPN and controls (<i>p</i> = 0.61). UCH-L1 levels were not different between study groups (<i>p</i> = 0.48). GFAP levels correlated with cold pain threshold (Rho= −&#xa0;0.320, <i>p</i> = 0.02) but failed to reach significance for cold detection (Rho= −&#xa0;0.236, <i>p</i> = 0.09). No correlation was observed between GFAP and nerve amplitude (<i>p</i> = 0.58) or conductivity (<i>p</i> = 0.86).</p> Conclusion <p>Serum GFAP levels are reduced in persons with DPN compared to persons without DPN and controls. Reduced serum GFAP levels may be associated with reduced markers of small nerve fiber damage obtained from quantitative sensory testing in people with diabetes.</p>