Individualized PCR strategies hamper comparability of molecular results between different laboratories in several fields of medicine. To harmonize BCR-ABL mRNA quantification an international multicenter trial involving 37 laboratories in 14 countries was initiated using 10 samples, each containing various dilutions (10, 2, 1 and 0.1%) of b3a2 or b2a2 BCR-ABL positive in normal leukocytes and negative controls. A novel control plasmid (pME-2) was designed for external calibration containing BCR-ABL and glucuronidase-beta (GUS) sequences. Median BCR-ABL/ABL ratios were 9.1, 1.8, 0.85 and 0.11% in b3a2 samples and 9.5, 1.6, 0.84 and 0.11% in b2a2 samples. Median BCR-ABL/GUS ratios were 3.4, 0.77, 0.37 and 0.042% in b3a2 samples and 2.8, 0.48, 0.29 and 0.031% in b2a2 samples. The coefficients of variation were 0.62 for ratios BCR-ABL/ABL and 1.03 for ratios BCR-ABL/GUS. Five of 37 evaluable participating laboratories (13%) detected low BCR-ABL copy numbers in negative control samples; one laboratory failed to detect BCR-ABL in a low-level sample. We conclude that the use of a common control plasmid does indeed improve comparability of BCR-ABL mRNA quantification results. However, further standardizing efforts like introducing a calibrator and regular control rounds are needed.