OBJECTIVE To determine the efficacy of urinary BTA-TRAK as a marker in monitoring superficial transitional cell carcinoma of the bladder and to compare urine cytology with urinary sediment testing. METHODS 700 consecutive determinations using BTA-TRAK to monitor unselected patients that had undergone surgery for transitional cell carcinoma of the bladder were analyzed. Cystoscopy, urinary sediment and urine cytology were performed during follow-up. Urography was performed yearly or when tumor of the upper urinary tract was suspected. (positive cytology or hematuria with no bladder tumor). Cystoscopy was performed a few days after determination of BTA-TRAK and voiding urine cytology and urinary sediment analyses (considered positive when microhematuria was observed) were both requested. RESULTS Of the 700 determinations, 95 (13.6%) were urothelial carcinomas (93 bladder, 2 upper urinary tract) that had been discovered during patient monitoring. Of the 93 bladder tumors, 39 were Ta (37 TaG1 and 2 TaG2), 29 T1 (4 T1G1, 20 T1G2 and 5 T1G3), 5 Tis and 20 muscle-infiltrating tumors (progression from T2-4 during monitoring). The sensitivity of urine cytology to detect urothelial tumor was 41.1% and the specificity was 97.3%. The urine cytologies were negative in 48.4% and inflammatory in 9.5% of the tumors. The sensitivity was 19% in low grade tumors. The sensitivity of urinary sedimentation testing to detect urothelial tumor (microhematuria) was 40% and the specificity was 96.7%. When associated with pyuria, it was considered to be a urinary infection or urothelial inflammatory condition, which was observed in 10.6% of the cases. Considering the proposed normal reference value for BTA-TRAK (< or = 14 U/ml), we have found a sensitivity of 62.1% and a specificity of 68.4%. A logistic regression model was developed, including BTA-TRAK, urinary sedimentation and cytology, to identify the independent variables that are useful for tumor detection during follow-up of superficial carcinoma of the bladder in this series. The combination of three variables showed an odds ratio of 18.5 (8.9-38.5) for urinary cytology, 11.8 (5.9-23.5) for urinary sedimentation and an odds ratio for BTA-TRAK that did not fall within the equation. CONCLUSIONS Although overall the sensitivity of BTA-TRAK is higher than that of urine cytology and urinary sedimentation testing, it provides no additional information than that obtained from the combination of urine cytology and urinary sedimentation testing in the detection of tumor recurrence during monitoring for superficial bladder cancer.