OBJECTIVES: The objective of this prospective study is to establish an appropriate cutoff value of urinary CYFRA 21.1 assay and to assess its utility combined with voided cytology and/or haemoglobin dipstick in the follow-up of patients with superficial bladder cancer. METHODS: From December 2000 to November 2003, 446 patients in follow-up for superficial bladder cancer (Ta-T1) after transurethral resection of the bladder (TURB) were included in a prospective study. Voided urine specimens were collected 7-14 d before cystoscopy and/or TURB for CYFRA 21.1 (one sample), haemoglobin dipstick (one sample), and cytology (three samples). All samples were processed for CYFRA 21.1 and haemoglobin dipstick according to manufacturer instructions. A control group (n=185) was obtained from patients in follow-up after transurethral resection of superficial disease (without recurrences within the following 6 mo). There were 125 recurrent transitional tumours detected by cystoscopy (34 TaG1; 53 TaG2/T1G1-2; 23 Ta-1G3/Tis, and 15 T2-4). Receiver operator characteristic (ROC) curves were constructed and cutoff values were chosen. Sensitivity, specificity, PPV (positive predictive value), NPV (negative predictive value), and their 95% confidence intervals were calculated. RESULTS: ROC curve analysis based on the previously reported cutoff value of 4ng/ml for CYFRA 21.1 demonstrated a sensitivity and specificity of 43% and 68%, respectively. At a cutoff value of 1.5ng/ml, sensitivity was 73.8% with a low specificity (41%). Further lowering of the cutoff point below 1.5ng/ml did not demonstrate a significant increase in sensitivity. Therefore, this value was chosen as the most sensitive CYFRA 21.1 cutoff point during the rest of the study. Specificity increased when all the patients treated with pelvic radiotherapy or with UTI, urethral catheterisation, and intravesical instillations within 3 previous months were not included in our analysis. CYFRA 21.1 plus cytology and the combination of CYFRA 21.1, cytology, and haemoglobin dipstick demonstrated the highest overall sensitivities, and detected 91.3% of Ta-1G3 tumours and 93.3% of T2-4 tumours. However, there were one muscle-invasive tumour, two T1G3/Tis, three T1G2, and nine T1G1 neoplasms with negative combination of cytology and CYFRA 21.1 (1,5ng/ml). All these tumours were smaller than 2cm in size; most were single tumours. Nevertheless, there were 16 tumours larger than 0.5cm (0.5-2cm), and multiple neoplasms were endoscopically detected in 14 patients. Similar results were obtained through the combination of CYFRA 21.1 (cutoff: 1.5ng/ml), cytology, and haemoglobin dipstick. CONCLUSIONS: In our experience the low sensitivity of urinary CYFRA 21.1, even using lower cutoff values and/or a combination with cytology and/or haemoglobin dipstick, makes its application not very useful as a surveillance tool for superficial bladder carcinoma.