In order to examine the prognostic role of routine urodynamic testing in prostatism, a prospective study was undertaken. The patients were selected for prostatic surgery exclusively by means of non-urodynamic criteria, but in addition an extensive, blinded urodynamic work-up was performed, including uroflowmetry, pressure-flow study combined with stop-flow test and water cystometry. The study comprised 139 patients consecutively referred for prostatism. Six months postoperatively all patients were reexamined. Previously the preoperative urodynamic data from each modality were analysed separately. In an attempt to increase the prognostic information, all data were computed simultaneously using linear discriminant analysis. Several preoperative variables appeared valuable to discriminate between patients with favourable and unfavourable postoperative outcome, respectively. However, the results of classification employing even the best combinations of variables did not result in more than 88% of the patients being correctly classified. Furthermore, especially the patients with unfavourable outcome generally were misclassified. Several reasons for this lack of efficacy are briefly mentioned, including the great overlap in preoperative urodynamic data between patients with favourable and unfavourable postoperative outcome and the relatively low postoperative failure rate.