Abstract

Morphometric estimates of mean glomerular volume (MGV) have clinical implications, over and above histologic data. However, MGV estimation is time-consuming, could waste tissue sections and requires expertise limiting its utility in retrospective clinical studies. MethodsWe evaluated MGV using both plastic and paraffin-embedded tissue from control and FSGS mice (n=10 each) using the gold-standard Disector/Cavalieri technique (Vglom-Cav) and other reported techniques [2- or 3-profile technique, Weibel-Gomez method (W-G)]. Within Vglom-Cav we examined the precision of MGV estimation while using MGVs obtained from 5- or 10-individual glomeruli measurements vs the true mean (20 glomeruli). ResultsIn both FSGS and controls, we identified an acceptable precision of 10-glomerular sampling vs true MGV within Vglom-Cav technique [88 (79-94) % of MGV obtained were within 10% of the true MGV]. The 5-glomerular sampling was less precise [70 (56, 81) % of MGV obtained were within 10% of true MGV]. In plastic based techniques, 2- or 3-profile MGVs showed greater concordance with Vglom-Cav, than W-G MGV. The new 3-profile technique offered incremental benefit to the existing 2-profile method (improved Lins concordance in control and FSGS animals). We observed a consistent reduction of Vglom values within control animals (52+/-0.06%) in paraffin-embedded tissue (vs corresponding methods in plastic) demonstrating a clear shrinkage artefact due to tissue processing. FSGS glomeruli showed significantly less and more variable shrinkage artefact likely due to glomerular fibrosis. ConclusionWe report the precision of 5- or 10-glomerular sampling for MGV estimation using controls and FSGS animals. We demonstrate and quantify the shrinkage bias in MGV during tissue processing for paraffin-embedding that also differentiated control animals and FSGS. Our findings have implications for experimental studies using glomerular morphometry.