Diagnostic accuracy of immunofluorescence versus immunoperoxidase staining to distinguish immune complex-mediated glomerulonephritis and C3 dominant glomerulopathy.

Abstract

The technique used for classification of membranoproliferative glomerulonephritis (MPGN) has been changed from an electron microscopy-based to an immunofluorescence (IF)-based semiquantitative technique with immunoperoxidase (IP) staining as a backup option when IF is not possible. Since data on that matter is lacking, our aims were to study the interobserver variability, the correlation and the reclassification of MPGN based on these two techniques. We retrospectively analysed cases of type 1 MPGN. We repeated IF staining and performed IP staining for IgG, kappa, lambda, C3c and C4d in 35 renal biopsies, among which 19 biopsies had matched IP and IF samples. We observed substantial to near-perfect agreement among the seven observers for both IF and IP (W coefficients from 0.66 for IF lambda to 0.89 for IF C4d). Of the 19 cases with matched IP and IF samples, five (26%) turned out to have different diagnoses on IF and on IP. Also, the ability of C4d to discriminate immune complex-mediated glomerulonephritis (ICGN) from C3 glomerulopathy (C3G) was poor, with areas under the curve of 0.44 [95% confidence interval (CI) 0.24-0.63] and 0.66 (95% CI 0.50-0.81) for the receiver operating characteristic curves of IF and IP respectively. Limitations include the fact that no clinical data regarding complement activation were available. The diagnosis of ICGN versus C3GN depends on the immunochemical technique used. Also, the use of C4d failed to discriminate ICGN from C3G in our study. Further validation studies are required to avoid misdiagnosis based on kidney biopsy.