A comparison of techniques to assess cervicovaginal irritation and evaluation of the variability between two observers

Abstract

Background: Colposcopy is used to evaluate effects of new vaginal products on cervicovaginal epithelium as part of the US Food and Drug Administration-mandated product approval process, yet few aspects of its use have been investigated.

Objectives: To determine the effect of the colposcopic examination itself on the number and type of findings seen, to compare colposcopy with the AviScope hand-held device and the naked eye and to compare the findings reported by two examiners during a single visit.

Study design: Fourteen healthy women volunteered for five paired examinations in random order: (1) naked eye inspection plus colposcopy done twice by a single examiner; (2) naked eye inspection plus AviScope examination, then naked eye inspection plus colposcopy by a single examiner; (3) Examination 2 repeated with the order of device reversed; (4) naked eye inspection plus colposcopy done by two examiners; (5) Examination 4 repeated with the order of examiner reversed. The colposcopic examinations were done per published standards but were limited to the areas visible without manipulation of the speculum.

Results: Length of colposcopic examination averaged 7 min. The number of colposcopic findings found when the examination was done twice by the same clinician was not statistically different (p = 0.12), suggesting that the examination itself did not induce findings. More findings were seen using magnification than naked eye. A similar number of findings were seen by AviScope compared to the colposcope (p = 0.99), but clinically significant findings were "undercalled" or "overcalled" by the AviScope. A weighted kappa score of the "worst" colposcopic finding was 0.32 (SE 0.10, p = 0.00), indicating moderate agreement between examiners.

Conclusions: The colposcopic examination is not burdensome nor does it induce findings. If naked eye observation were used alone in practice, these data suggest that half the colposcopically detected findings would be missed. Using the naked eye observation for screening would minimally reduce the number of magnified observations carried out. For detecting epithelial changes, the colposcope seems to be the most sensitive technique, followed by the AviScope.