Goblet cell density estimate differences in impression cytology samples varies with different magnification of images

To assess the impact of using different microscope magnifications for the goblet cell density (GCD) estimates from conjunctival impression cytology (CIC) samples from healthy individuals METHODS: In a prospective study, CIC specimens were collected from the superior bulbar conjunctiva (12 o'clock, 5mm from limbus) of 20 adult subjects (average age 22 years) onto Millicell-CM membranes and Giemsa stained. A region from each CIC filter containing reasonably high numbers of goblet cells was imaged by light microscopy at a final magnification of 400X and then the same region assessed at 200X and then 100X. The images were enlarged, the goblet cells marked and counted and GCD values/sq mm calculated.

The mean GCD estimates at 400X magnification, 200X and 100X were 644±180, 405±72 and 365±81 cells/sq mm respectively, and these values were statistically different (p<0.001).

As a result of non-uniform distribution, a strategy to select a 400X high power microscope field (HPF) that appears to include a moderate number of goblet cells will have a probability (by at least 20:1) that the GCD estimates will likely be higher compared to those at 200X or 100X, and the probability for higher GCD values is at least 15:1 comparing assessments made at 200X to 100X. Investigators should use only one magnification, with that of a medium power field (200X final magnification) likely being the most useful.