Visual Assessment of Vaginal Atrophy – valid objective, observational measures (VAVAvoom study)

Objectives: Vaginal atrophy (VA) is a significant component of the genitourinary syndrome of menopause, causing significant symptoms for patients as well as characteristic visual changes on pelvic exam. Objective measures exist for grading VA, including vaginal pH and the Vaginal Maturation Index (VMI), however these methods are often substituted with a subjective visual inspection on exam. There is currently no validated method of using visual characteristics to accurately grade VA. The objective of this study was to assess the accuracy of a dichotomous global visual assessment of vaginal atrophy (akin to current common practice) relative to the objective VMI. We also aimed to develop and validate a simple, translatable, and reliable method for the visual grading of vaginal atrophy at the time of pelvic exam, which would minimize subjectivity and facilitate standardized clinical and academic communication.

Methods: This was a prospective cohort study of 125 postmenopausal women (33 on vaginal estrogen therapy and 92 without) that underwent two independent and blinded dichotomous global visual assessments of VA and two blinded assessments of VA according to a detailed visual scorecard containing 6 common visual signs of VA (a score over 2 was considered atrophic) that was developed for the purpose of this study. Demographics, medical history, VMI, vaginal pH, and a modified Urogenital Atrophy Questionnaire containing 6 pertinent symptoms of VA were also collected. Global assessments and detailed visual scores were assessed for inter-rater reliability and correlated to valid objective measures (VMI and pH) and symptom scores to assess validity.

Results: The dichotomized global visual assessment had good inter-rater reliability (74%) and high sensitivity (100%), but greatly lacked specificity (16%) when compared to the gold standard VMI. The detailed visual scorecard maintained good sensitivity (92%) and showed improvement in specificity (48%) and inter-rater reliability (93%), with an OR of 10.50 (p=0.002). For individual visual signs, only the absence of physiologic discharge (OR 3.72, p=0.007) or the presence of vaginal abrasions/petechiae (OR 3.31, p=0.02) were significantly associated with an atrophic VMI, while other visual signs, vaginal pH, and urogenital symptoms were not reliable predictors. No patients on current vaginal estrogen therapy had an atrophic VMI despite varied visual scores.

Conclusions: These results suggest that although visual characteristics are often seen in VA, they may not be accurate predictors of atrophy as defined by the VMI. Using an objective visual scorecard offers some improvements in inter-rater reliability and accuracy compared to a simple global visual assessment for presence of atrophy. Patients on vaginal estrogen therapy tend to not have VA despite presence of visual characteristics. Despite the perception that symptoms are the most important factor in treating VA, urogenital symptoms often associated with VA did not show an association to atrophy as defined by the VMI.