SETTING: A laboratory cross-contamination event was suspected because Mycobacterium tuberculosis was unexpectedly detected at a high incidence in the cultures of several clinical specimens at the National Hospital Organization, Tokyo National Hospital, Japan. OBJECTIVE: To describe a case of Mycobacterium tuberculosis laboratory cross-contamination. DESIGN: We reviewed the medical records of 20 patients whose clinical specimens were suspected to have been contaminated by Mycobacterium tuberculosis. Variable number of tandem repeat analysis with 15 loci, the Japan Anti-Tuberculosis Association-12, and three additional hyper-variable loci, was performed to identify the cross-contamination event. RESULTS: The clinical, laboratory, and variable number of tandem repeat data revealed that the cross-contamination had possibly originated from one strongly positive specimen, resulting in false-positive results in 11 other specimens, including a case treated with anti-tuberculosis drugs. CONCLUSION: Clinical and laboratory data must be re-evaluated when cross-contamination is suspected and variable number of tandem repeat analysis should be used to confirm cross-contamination. Furthermore, original isolates should be stored appropriately, without sub-culturing and genotyping should be performed at the earliest possible for better utilization of variable number of tandem repeat for the identification of cross-contamination.