Real-time RT-PCR has great advantages for estimating transcript levels in a variety of situations. These include relative rapid assay times (hours), reliability and ease of performing replicate analyses. In contrast, competitive PCR is a very labor-intensive procedure requiring a few days to generate useful data. We compared the same samples from CML patients by both methods. Importantly, we used the Bcr-Abl junction plasmid DNA, which is used as a competitor in the manual competitive PCR assay, to generate a standard curve for the real-time assay. This permitted reporting the real-time data as the number of BCR-ABL transcripts per microg of total RNA, which is the same format used for the competitive PCR assay. In this study, a total of 435 peripheral blood and marrow samples from 285 CML patients were analyzed by RT-PCR; these patients were undergoing therapy by STI-571, interferon, and bone marrow transplantation treatment. Most samples also had assay values for the Philadelphia chromosome (Ph), FISH and Western blotting for the Bcr-Abl oncoprotein. Our findings indicated that the real-time assay was less sensitive than the manual competitive RT-PCR assay (t = 5.118; P < 0.001). Of interest, the transcript levels in cell line mixtures with various ratios of K562/KG-1 (BCR-ABL positive/negative) cells were also significantly higher with the competitive RT-PCR assays than real-time RT-PCR, except for levels of BCR-ABL below 200 transcripts per microg of RNA. In both patient and cell line experiments, dividing the BCR-ABL transcripts by the total ABL transcripts virtually eliminated the difference between real-time BCR-ABL transcript values and quantitative competitive BCR-ABL transcript values, indicating that both BCR-ABL and ABL transcripts were underestimated by the real-time assay. In addition, the increased sensitivity of the nested, competitive RT-PCR was readily apparent in patients with minimal residual disease, which by the real-time were negative in the majority of patients but were positive by nested, competitive RT-PCR in 44.6% (n = 29) of samples analyzed (n = 65). These findings indicate that real-time RT-PCR, when normalized for the total ABL transcripts, can be used to monitor CML patients during therapy, but we suggest that nested, competitive RT-PCR be used to determine BCR-ABL/ABL transcript ratios at low transcript values or especially when real-time analyses are negative.