Two reverse transcription-polymerase chain reaction methods to detect the AML1/ETO rearrangement in the M2 subtype of acute myeloid leukaemia those of Downing et al. (Blood 1993; 81: 2860-5) and Satake et al. (Br J Haematol 1995; 91: 892-8) were evaluated. Bone marrow samples, one at diagnosis and two in complete remission from a patient with M2 subtype of acute myeloid leukaemia, with t(8;21), were analysed using both methods. The Kasumi-1 cell line was used as a positive control and a patient with M3 subtype of acute myeloid leukaemia as a negative control. To confirm the feasibility of Satake's method a group of 35 patients with subtypes of acute myeloid leukaemia at diagnosis were studied. The method of Downing requires Southern blotting and hybridization with a specific probe because it often generates non-specific amplification products. By contrast, the method of Satake yields only a single amplification product, using one single round of PCR in samples at diagnosis, or two rounds in complete remission samples. The sensitivity of this method allows the detection of a single Kasumi-1 cell in 10(6) normal cells. The AML1/ETO rearrangement was observed in 5 of the 35 cases of acute myeloid leukaemia at diagnosis (14.3%) and in 3 of the 14 cases of M2 subtype of acute myeloid leukaemia (21.4%). The two remaining positive cases corresponded to the acute myeloid leukaemia subtypes M4 and M6. The results indicate that the method of Satake better meets the requirements of the clinical laboratory due to its greater simplicity, specificity, sensitivity and feasibility, thus making it more appropriate for use in diagnosing and monitoring minimal residual disease.