ABSTRACT
OBJECTIVE: To evaluate the use in non-sputum samples of a commercial molecular amplification kit (LCx MTB, Abbott Diagnostica) (LCx) for the diagnosis of tuberculosis. MATERIAL AND METHOD: Ninety-nine non-sputum samples from the same number of patients (bronchoalveolar, pleural and ascitic fluid, fecal samples, blood cultures, biopsies from different sites, cerebrospinal fluid, urine and gastric juices) and 14 sputum samples (10 from patients clinically suspected of having tuberculosis and 4 from patients diagnosed of tuberculosis and undergoing appropriate treatment for at least one month). All samples were LCx processed according to the manufacturer's instructions. The reference diagnosis was obtained by the Löwestein-jensen method and when results were inconsistent, we took into account the degree of clinical suspicion, response to treatment and histology. RESULTS: Seven of the 99 samples were positive by the LCx technique, and 6 of the 7 were also LJ positive; 1 could not be evaluated because of culture contamination. One LJ positive culture was LCx negative. Only one sample was positive by Ziehl-Neelsen (ZN) staining. Ninety-two samples were LCx negative, with 91 showing no growth at all. Sensitivity was 86% and specificity 98%. Atypical mycobacteria were detected in 4 cases, all of which were LCx negative. CONCLUSIONS: Diagnosis of tuberculosis by applying the LCx system to various types of samples other than sputum is simple, rapid, sensitive and specific.
