ABSTRACT
BACKGROUND: For the microbiological diagnosis of tuberculosis, most standard laboratory guidelines recommend the processing of multiple specimens. This fact added to the nonspecific clinical presentation of many cases, contribute to a low efficiency of the tuberculosis laboratory. The introduction of new procedures have increased the sensibility of the diagnosis and so the cost. The object of our work, was to asses the value of examining 3-6 specimens, in optimizing the diagnosis. METHODS: We retrospectively reviewed 327 specimens from 105 patients with culture-proven tuberculosis, between 1997-1998. All specimens were received in the laboratory within 1 month of the initial diagnosis. RESULTS: Of the 7694 specimens received for acid-fast smear and culture in that period of time, 519 (6.7%) were positive for mycobacteria, many of them without clinical relevance. The average number of specimens processed by patient was 3.1, however 22% of the patients with culture-proven tuberculosis had less than tree specimens submitted at the time of diagnosis and 19% had more than three. In 49.5% of the patients, all the smears were negative and the diagnosis was done by culture. With the first specimen submitted were diagnosed 76% of the patients, and with the second we reach 94.2%, finally, we needed 3 specimens for the diagnosis of 104 from the 105 cases (99%). CONCLUSIONS: With the better sensibility achieved in the detection procedures, the interest in the improvement of the efficiency in the tuberculosis laboratory should consider the contribution of the specimen number on the diagnosis. If studying only one specimen we could fail 25% of the early diagnosis, we have rarely found more than two specimens being of diagnostic value. As the time for detecting a positive culture has decrease with the use of liquid media and probes for identification, we would recommend to submit only two specimens for the routine diagnosis and see the patient again 1 moth later. If in this time, the laboratory has'nt obtain any positive result and the clinician still suspect a tuberculosis, a second set of specimens can be submitted. With this measures, laboratory that received a great number of specimens, could assume the work of small laboratory without increasing there work and budget, and the diagnosis of tuberculosis will be optimized.