Prediction of findings at defecography in patients with genital prolapse.

OBJECTIVE Defecography may be useful in pre-operative assessment of patients with genital prolapse. Defecography is an invasive and embarrassing procedure for patients and little effort has been made to optimalise selection criteria for defecography. This study investigated whether discrimination of high and low probability of abnormal defecography is possible based on the quantified findings from patient history, pelvic examination and a validated questionnaire. DESIGN Prospective observational study. SETTING Three teaching hospitals in The Netherlands. POPULATION Eighty-two patients undergoing surgical correction of uterine prolapse Stages 2-4. METHODS A history and pelvic examination were obtained from all patients. A validated questionnaire was used to assess the presence of defecation and micturition symptoms. Using multivariate logistic regression analyses with receiver operating characteristic curves, a diagnostic model to predict the presence of an abnormal defecography was systematically constructed and validated. MAIN OUTCOME MEASURE Presence of abnormal finding at defecography. RESULTS The most important predictors for abnormal defecography were prolapse of the posterior vaginal wall, history of abdominal or pelvic surgery and the presence of constipation. With these variables, a prediction rule could be constructed which predicted the prevalence of an abnormal defecography (area under curve = 0.73; 95% CI 0.61-0.83). CONCLUSIONS This study shows that a diagnostic model based on findings obtained from a non-invasive workup can accurately predict the presence of an abnormal defecography. Such a model provides the possibility to tailor the request for defecography to the individual patient.