Morphometry of kidney biopsy: number of glomeruli and accuracy of morphometric measurements
Abstract
Five kidney biopsies with diffuse changes containing 16 - 48 glomeruli were analysed morphometrically. After excluding profiles with a diameter smaller than 60% of the diameter of the largest glomerular tuft in the biopsy the coefficient of variation between glomeruli was 12 - 19 per cent for the following parameters: number of nuclei per glomerular area, surface density of glomerular BM, and the volume density of capillaries. The volume density of mesangium showed a CV of 27%. The biopsy should contain 1 - 4 glomerular profiles if one would like to keep the relative standard error (CV/√n) of these parameters under 15%. The estimates were tested by letting a computer to sample a glomerular data file in samples of 1 - 10 glomeruli. The probabilities for deviations larger than 10 - 20% of the unbiased value were calculated. Cellular proliferation (nuclear density) in the glomeruli can be estimated in biopsies with 2 glomeruli with a probability of 12% for a deviation of more than 20%. In samples with 6 glomeruli no nuclear density figures deviated more than 20% of the unbiased estimate.