Predictors of Success for Pelvic Floor Muscle Training in Pelvic Organ Prolapse

Abstract

Background Although the effectiveness of pelvic floor muscle training in women with prolapse has been demonstrated in several studies, there seem to be subgroups of responders and nonresponders. Objective The objective of this study was to identify factors that predict treatment success in women receiving pelvic floor muscle training for prolapse. Design The design was a secondary analysis of data from 2 randomized controlled trials comparing conservative prolapse treatments. Methods After 12 months, 172 women subjectively assessed treatment success ("better") or failure ("the same" or "worse"). Potential predictors were identified by a literature search and by consultation with experts in the field of urogynecology and pelvic floor muscle training. The relationship between potential predictors and treatment success was explored using logistic regression analysis. Results Treatment was successful in 94 women (55%) and unsuccessful in 78 women (45%). The presence of ≥1 indicators of obstetric trauma (eg, high birth weight, episiotomy, perineal laceration during vaginal delivery, forceps delivery, or vacuum extraction) (odds ratio = 4.4; 95% CI = 1.6-12.0) and younger age (odds ratio = 0.94 per year; 95% CI = 0.9-1.0) independently predicted treatment success. The area under the receiver operating characteristic curve for the final model was 0.65 (95% CI = 0.57-0.74), and the model explained 11.7% of the variance. Limitations Although attempts were made to include all relevant predictors, the selection or operationalization of variables could have been incomplete or insufficient. Conclusions Identifying women who have prolapse and are likely to benefit most from pelvic floor muscle training is of great importance to clinical practice. In this study, 2 factors that independently predicted favorable outcomes were identified with this management approach. However, further research is needed to identify other predictive factors and to validate a new model in another population.