Morphometry of dermal nerve fibers in human skin

We aimed to assess the innervation density of dermal nerves in human skin biopsies by bright-field immunohistochemistry. The size of dermal area where nerve length was quantified was validated in 30 skin biopsy sections (5 controls and 5 patients with small-fiber neuropathy [SFN]). It was obtained dividing an area of 200-μm depth from the dermal-epidermal junction into 4 equal portions. The length of dermal nerves (DNFL) was measured into 150 sections (25 controls and 25 patients with SFN) and values per millimeter of epidermis (DNFL/mm) and dermal area (DNFL/mm2) were obtained. Age- and gender-matched normative values of intraepidermal nerve fiber (IENF) density were used as gold standard to calculate the performance of dermal nerve morphometry. Patients showed significantly lower DNFL (1.96 mm ± 0.96 SD), DNFL/mm (0.65 ± 0.29 SD), and DNFL/mm2 (3.75 ± 1.7 SD) than controls (DNFL 3.52 mm ± 1.31 SD, 5th percentile 2.05; DNFL/mm 1.25 ± 0.39, 5th percentile 0.71; DNFL/mm2 7.07 ± 2.41 SD, 5th percentile 3.95). Sensitivity, specificity, and percentage of individuals correctly classified were 75.8%, 73.9%, and 74.8% for DNFL, 75%, 80%, and 77.7% for DNFL/mm, and 75.8%, 80.2%, and 78.1% for DNFL/mm2. Receiver operator characteristic area analysis confirmed the excellent discrimination (0.8-0.9) between patients and controls. Dermal nerve morphometry significantly correlated with IENF density. Spearman rank correlation demonstrated good agreement for interobserver analysis (0.87-0.89), and between DNFL and IENF densities (0.71-0.73; p < 0.0001). We provided a reliable method to quantify the innervation density of dermal nerves that might improve the diagnostic yield of skin biopsy.