Telomerase activity is present in most human malignant tumors, whereas it is generally not detectable, with some exceptions, in normal cells. Therefore, it represents a potential tool for tumor detection. In the present study, telomerase activity was determined in urine from 79 healthy individuals and 121 previously untreated bladder cancer patients using a highly sensitive telomeric repeat amplification protocol (TRAP) assay and the results were expressed as arbitrary enzymatic units (AEU). This approach enabled us to identify cutoff values characterized by high sensitivity (from 75% to 93%) and specificity (from 72% to 92%). Moreover, analysis as a function of gender showed a higher accuracy of TRAP assay in males (93% sensitivity and 90% specificity at the cutoff of 50 AEU) than in females. This sensitivity was confirmed in patients with nonassessable or negative cytology. In women, morphological and immunocytochemical determinations using a monoclonal antibody (anti-hTERT) recently developed in our laboratory showed a large fraction of immunoreactive inflammatory or nonbladder cells, which may justify the false-positive TRAP results. In conclusion, this assay represents an important noninvasive diagnostic tool to detect bladder cancer also in patients with negative or nonassessable urine cytology and with low-grade and early-stage lesions.