Qualitative RT-PCR methods used for monitoring minimal residual disease (MRD) in APL patients fail to predict relapse in up to 25% of patients in remission. We report here the development and evaluation of a highly sensitive (10(-5) and 10(-6) with one round and two rounds of PCR, respectively) competitive RT-PCR method to quantitate the PML-RARalpha fusion transcripts. PML-RARalpha transcript's levels were normalised to 10(5) copies of ABL transcript. Serial BM and PB samples from 16 patients with APL and t(15;17) were examined. Presentation samples from three patients (three BM, one PB) showed levels in the range of 0.7 x 10(6)-3.5 x 10(6) and 1.2 x 10(5) molecules in BM and PB samples respectively. Serial quantitation of MRD in both BM and PB samples showed significantly lower levels of PML-RARalpha transcripts in remission, although the majority of samples remain positive for the PML-RARalpha transcripts even those in long-term remission (up to 94 months). Levels of PML-RARalpha in remission samples were up to 2 x 10(2) and up to 5.2 x 10(1) molecules in BM and PB respectively. BM and PB samples taken from two patients 2-4 months before relapse showed significantly higher levels of PML-RARalpha transcripts (1.2 x 10(4) molecules in BM; 3.5 x 102, 1.2 x 10(2) and 1.2 x 10(3) in PB). The same samples, when tested with a standard qualitative RT-PCR for the amplification of PML-RARalpha (with a sensitivity of 10(-4)) produced negative results. This indicates that the qualitative methods would not have predicted relapse in these patients. Our data show that quantitating PML-RARalpha transcripts with a sensitive method may provide a superior approach for monitoring MRD in APL and identifying patients at high risk of relapse.