Mycobacterium tuberculosis typing: usefulness of DRE-PCR to confirm cross-contamination in the mycobacteriology laboratory of a general reference hospital for AIDS.

In this study two molecular typing methods, a simple double repetitive element PCR-based assay and the standardized restriction fragment length polymorphism (RFLP), were used to confirm cross-contamination in the mycobacteriology laboratory. Clinical specimens from 12 patients, submitted for acid-fast bacilli stain smear and processed for culture in Lowenstein-Jensen on the same day, resulted in positive bacterioscopy (+++) and confluent growth only for one of the patients. The specimens from all the other patients but two were smear-negative and culture-positive, with one or two colonies. None of them had clinical symptoms and radiological findings for active tuberculosis (TB). The suspicion of false-positive cultures arose when a health care worker who had had a PPD skin test conversion, claimed to be healthy and had no TB symptoms, was found to have a positive sputum culture. DRE-PCR demonstrated that all nine cultures typed belonged to one cluster, further confirmed by RFLP. Although DRE-PCR has been found to be poorly reproducible, it has enough discriminatory power to be useful for rapid epidemiological investigation in selected settings.