Hoy WE et al. (2006) How many glomerular profiles must be measured to obtain reliable estimates of mean glomerular areas in human renal biopsies? J Am Soc Nephrol 17: 556–563

Glomerular size has important clinical and prognostic implications, but gold-standard methods of volume estimation are labor-intensive and incompatible with common biopsy techniques. Using standard core biopsy tissue, Hoy et al. have determined the number of glomerular profiles that need to be measured to yield a reliable estimate of average size.

All centrally located glomeruli on slides of archived biopsy tissue from patients with nondiabetic renal disease (n = 384) were measured using traditional stereologic point-counting, to calculate the 'true individual mean' for each biopsy. These calculated averages were used to determine 'true population means' for groups of biopsies. Increasing numbers of randomly selected glomeruli were then measured to generate 'random sample means'.

Random sample means for individual biopsies correlated best with true means when ≥10 profiles were measured without regard to sclerosis, or when ≥8 nonsclerosed profiles were measured. The median number of glomerular profiles per biopsy was seven; when only nonsclerosed profiles were considered, the median number was four.

True population means could be reliably predicted by measuring as few as five randomly selected profiles per biopsy in a group of 30 biopsies; however, the reliability of this estimate improved as the number of renal biopsies included in the analysis increased. The authors recommend that, when estimating mean glomerular areas for a population of biopsies, all available biopsies should be included irrespective of the number of profiles in each.