Telomerase has been found to be reactivated in a majority of cancers but is inactive in most somatic cells. Our principal goal was to determine the potential use of the telomeric repeat amplification protocol (TRAP) assay as marker for malignancy in cytological effusions. The simple selection criterion was the cytological diagnosis, and routine samples were classified into malignant (58 samples) and nonmalignant (233 samples). Of the malignant samples, 44/58 (76%) were positive by TRAP assay. Of the 14 telomerase-negative cytology-positive samples, RNA integrity was poor in 9, indicating suboptimal sample conservation for molecular analysis. In 3 of the remaining 5 samples with a negative TRAP assay, a high number of malignant cells was observed, and these cells might have been telomerase-negative. Thus, the sensitivity of TRAP assay for the presence of malignant cells was about 76%. In the cytologically nonmalignant effusions, the presence of telomerase activity was observed in 24% (55/233). Of these, 6% were highly suspicious for malignancy, 9% were doubtful, and 9% were cytologically nonmalignant effusions confirmed by a follow-up of 12 mo or more. According to these data, the specificity of the TRAP assay to detect tumor cells in effusions ranged only between 82-91%. Our results indicate that, although the TRAP assay is positive in 6-15% of putative malignant effusions, the relatively high number of TRAP false-negative and false-positive cases renders this test unsuitable for routine diagnostic purposes