The reliability of routine BCR-ABL RT-nested-PCR was evaluated in 1453 B-lineage ALL or hybrid leukemia at initial diagnosis by RT-nested-PCR. All BCR-ABL-positive (n = 642) and 176 BCR-ABL-negative samples underwent a second RT-PCR. In 518 patients, karyotyping and/or FISH was compared to the BCR-ABL status. The second RT-PCR revealed in 155/642 initially positive samples a divergent result (153 BCR-ABL-negative, two other transcripts) that in most cases turned out to be caused by contaminations in the first RT-nested-PCR. Confirmatory RT-PCR detected 2/176 false negative first RT-nested-PCR results. Thirty-nine specimens remained ambiguous despite different RT-PCR approaches. As far as cytogenetic evaluation and FISH is available (n = 23), the majority but not all patients with an ambiguous RT-PCR result were Ph-negative (n = 18). RT-nested-PCR and cytogenetics yielded in 346 of 383 evaluable samples a concordant result. Differing results are given and account in part to the lower sensitivity of karyotyping. Taken together, confirmed RT-PCR detected BCR-ABL fusion transcripts consistently in 487 out of 1453 ALL samples (c-ALL: 43%, pre-B ALL: 34%, pro-B ALL: 5%, B-ALL: 0%, hybrid leukemia: 5/11). Since false positive initial RT-nested-PCR data were frequent, either confirmatory second RT-PCR or FISH analysis is warranted to guarantee sensitive and reliable results of utmost clinical relevance.