UCSF

Context

Atrophic vaginitis is a commonly reported subset of Papanicolaou test results that are negative for intraepithelial lesion or malignancy, but interpretive criteria overlap with atrophic changes and other entities, hindering concordance among observers.

Objectives

To report on the participant concordance from 2000 to 2009 in the College of American Pathologists Interlaboratory Comparison Program in Gynecologic Cytopathology, with a reference interpretation of atrophic vaginitis, and to investigate cytologic features of good and poorly performing slides to identify criteria useful in the interpretation of atrophic vaginitis.

Design

We summarized 18 302 responses from the program for slides with a reference interpretation of atrophic vaginitis. We randomly selected 18 Papanicolaou test results (3 conventional, 4 SurePath, and 11 ThinPrep) from good and poor performers for prospective, blinded criteria scoring for the following features: abundance of neutrophils, more than 100 degenerating parabasal cells, more than 25% necrotic background, more than 100 pseudoparakeratotic cells, and the presence of stripped or streaked nuclei, histiocytes, and superficial or intermediate squamous cells.

Results

Most Papanicolaou test results (>90%) with a specific reference interpretation of atrophic vaginitis were categorized as negative. Cytotechnologists are more likely than pathologists are to label it negative for intraepithelial lesion or malignancy (NILM) and are equally likely to mistake it for a high-grade lesion. Degenerating parabasal cells, pseudoparakeratosis, and necrotic background are associated with atrophic vaginitis (P  =  .001) on Papanicolaou. Abundant neutrophils (>100 per ×400 field) are also significantly correlated (P  =  .01).

Conclusions

Exact concordance to atrophic vaginitis is less than 90%. Most of the discrepancies are negative results for intraepithelial lesion or malignancy. Advanced atrophic features are as significant as neutrophils are to the interpretation of atrophic vaginitis.