Introduction and hypothesis: This study was undertaken to investigate the relationship between symptoms of prolapse and International Continence Society Pelvic Organ Prolapse Quantification (ICS POP-Q) measurements in order to establish optimal cutoffs for predicting prolapse symptoms using receiver operator characteristic (ROC) statistics.
Methods: This was a retrospective study using 764 archived data sets of patients seen for symptoms of lower urinary tract and pelvic floor dysfunction between March 2011 and November 2012. Main outcome measure was symptoms of prolapse. Explanatory parameters were Ba, C, and Bp as defined by the ICS POP-Q. Patient age, body mass index (BMI), previous hysterectomy or incontinence/prolapse surgery, and vaginal parity were tested for a confounding effect on the relationship between ICS POP-Q measurements and symptoms of prolapse.
Results: Optimal cutoffs for predicting prolapse symptoms were defined as follows: Ba = -0.5 (sensitivity 69 %, specificity 71 %), C =-5 (sensitivity 67 %, specificity 64 %), Bp = -0.5 (sensitivity 63 %, specificity 62 %). ROC statistics resulted in an area under the curve of 0.768 for Ba [confidence interval (CI) 0.729-0.807), for C of 0.724 (CI 0.672-0.776), and for Bp of 0.686 (CI 0.639-0.733).
Conclusion: Our findings suggest that the ICS POP-Q staging system requires revision. Prolapse of the anterior and posterior vaginal wall of < -1 should probably be regarded as normal. On the other hand, stage 1 uterine prolapse as currently defined seems highly relevant.