The diagnosis of significant rectocele is currently made on the basis of cinedefecographic findings. Clinical examination alone will only allow assessment of the presence but not the significance of a rectocele. Therefore, the aim of this study was to determine if anal manometric findings can predict the significance of a rectocele. All patients with a diagnosis of constipation and rectocele confirmed on cinedefecography between 1992 and 1998 were retrospectively reviewed. Significant rectocele was defined as the presence of three of the following five parameters: rectocele >4 cm in diameter as measured during the evacuatory phase of cinedefecography, rectal and/or vaginal symptoms present for longer than 12 months, persistence of rectal or vaginal symptoms for at least four weeks, despite increased dietary fiber (up to 35 g/day), need for rectal and/or vaginal digitation or perineal support maneuvers for rectal evacuation. Statistical analysis was performed using the Mann-Whitney test and Fisher's exact test. A logistic regression model with stepwise selection was used to determine significant prognostic factors. A total of 305 patients (31 men) with rectocele, with a median age of 68 years (range, 12-89) were identified. Of these, 89 (29.2%) had significant rectoceles. There was no difference in the frequency of significant and non-significant rectoceles with respect to gender or age. However, patients with a significant rectocele compared to those with a non-significant rectocele had higher median first sensation volume (45 vs. 30 ml, p=0.0005), median capacity (160 vs. 120 ml, p<0.0001), and median compliance (10 vs. 8 ml H(2)O/mmHg, p=0.05). Calculations based on a logistic regression model determined that with a first sensation of 100 ml, a capacity of 400 ml, and a compliance of 50 ml/mmHg, the probability of a significant rectocele would be 85%. In conclusion, anal manometric findings may be useful in predicting significant rectocele in constipated patients.