Improving the clinical prediction of detrusor overactivity by utilizing additional symptoms and signs to overactive bladder symptoms alone

Introduction and hypothesis

We attempted to improve the accuracy of the clinical diagnosis of detrusor overactivity (DO) by using other significant clinical parameters in addition to overactive bladder (OAB) symptoms alone.

Methods

One thousand one hundred and forty women attending for their initial urogynecological assessment, including urodynamics, due to symptoms of pelvic floor dysfunction, underwent a comprehensive clinical and urodynamic assessment. Multivariate logistic regression analysis of a wide range of clinical parameters was used in order to determine a model of factors most accurately predicting the urodynamic diagnosis of DO. Data were separated according to women without DO; women with DO. The analysis involved the stepwise building of an optimal clinical model for predicting DO.

Results

In multivariate analysis, the OAB symptoms of urgency incontinence, urgency and nocturia (not frequency) were significantly associated with DO. Their prediction of DO was not particularly accurate (sensitivity 0.64; specificity 0.67). The addition of other significant clinical parameter, i.e. absent symptoms of stress incontinence; lower parity (0–1); no signs of prolapse, to the diagnostic model, resulted in marginally improved accuracy (area under the ROC curve increased from 0.70 to 0.74).

Conclusions

Overactive bladder symptoms alone are not accurate in predicting DO. Adding other significant clinical parameters to the model resulted in a small statistical advantage, which is not clinically useful. An accurate clinical diagnosis of DO in women would appear to remain elusive.