For many years, it has been thought that positive culture of M. tuberculosis is a definitive diagnostic evidence of tuberculosis and cross-contamination of M. tuberculosis culture in clinical laboratories is rare. However recently introduced RFLP analysis has enabled us to identify a strain of M. tuberculosis, and many cases of the cross-contamination in clinical laboratories confirmed by RFLP analysis have been reported. In this report, we present the first case of the cross-contamination confirmed by RFLP in Japan. In our case, 5 patients without any personal link to each other were suspected based on clinical findings to have cross-contaminated results of M. tuberculosis culture. All their specimens were processed on the same day, and were smear negative and culture positive with only a small number of colonies (less than 8 colonies). The sputum from the suspected source of contamination processed on the same day was strongly positive for AFB smear and heavily culture positive. The RFLP patterns of these 6 patients were identical, so it was concluded that the positive cultures of the sputum from the 5 patients who were not expected to be culture positive on clinical findings were caused by the cross-contamination in our hospital laboratory. We review all the charts of patients with M. tuberculosis culture positive results in the same year of this case, but we didn't find no other cases suspected of the cross-contamination. Then we reviewed the literature of M. tuberculosis culture cross-contamination. The patterns of the cross-contamination are divided into two. One is associated with malfunction of a sampling needle in the BACTEC 460 system and the other associated with the initial processing of the specimens, mostly involving reagents such as NaOH solution. Cross-contaminated specimens are usually smear negative with only a few colonies (less than 5), and processed just after the source specimen of the contamination in most reported cases, but not in all. In almost half of them the cross-contamination results had significant influence on the clinical management. The frequency of the cross-contamination is estimated around 1% of the patients with M. tuberculosis culture positive results. For early detection of the cross-contamination, not only clinicians but also laboratory staffs have important role and close cooperation between them is mandatory. To prevent the contamination, it is advisable to process smear positive and probable culture positive specimens separately from others, and not to use a large same container of reagents for processing of different specimens.