We investigated the agreement between simple indirect immunofluorescence (IF) and bright-field immunohistochemistry (BFI) on free-floating sections for intraepidermal nerve fiber density (IENFD) quantification. Fifty-five healthy subjects and 63 patients with probable small fiber neuropathy (SFN) underwent two adjacent skin biopsies at the distal leg processed by IF and BFI technique. Agreement between IENFD pairs obtained by each method was assessed by Bland-Altman testing. The area under the curve of the receiving operating characteristics (ROC) curves was used to compare the discrimination ability. The diagnostic judgment was based on sex and age-adjusted normative values. IF and BFI showed good correlation (r = 0.81), with a ratio of about 2:1 and a mean difference of 5.5 ± 3.0 IENF per millimeter between paired measures, as demonstrated by linear regression and Bland-Altman test analyses. The square root transformation confirmed a Poisson distribution of the data and a fixed bias between IF and BFI measurements. The ROC curves analysis demonstrated a striking overlap between IF and BFI (0.83 and 0.82; p = 0.72). The diagnosis of SFN disagreed in only 6.7% of cases when the judgment was based on a difference of >1 IENF from 5% cut-off value. IF and BFI showed comparable diagnostic efficiency when referred to appropriate normative reference values.
Keywords: immunofluorescence; immunohistochemistry; intraepidermal nerve fiber; painful neuropathy; skin biopsy.
© 2015 Peripheral Nerve Society.