Role of clinical judgment in the application of a nucleic acid amplification test for the rapid diagnosis of pulmonary tuberculosis.

BACKGROUND Several nucleic acid amplification (NAA) tests for Mycobacterium tuberculosis (MTB) have been licensed for the rapid diagnosis of active pulmonary tuberculosis (PTB) in respiratory secretions. There is uncertainty however regarding the practical application of these tests in clinical decision making. OBJECTIVE To evaluate the utility of the COBAS AMPLICOR assay (Roche Diagnostics; Singapore) for MTB as applied by specialists for the rapid diagnosis of PTB in the routine clinical setting. DESIGN A prospective study of consecutive patients suspected of PTB and tested with the AMPLICOR assay under the care of respiratory physicians. The final diagnosis was based on all relevant clinical information after at least 3 months of follow-up. Accuracy of the NAA test was compared with that of the initial expectant treatment. Expectant treatment was based on an integrated approach that incorporated clinical evaluation with results of direct smear and NAA tests. RESULTS The incidence of PTB in 168 patients was 32%. The basis for expectant treatment of PTB was positive smear result in 47%, clinical suspicion in 26%, and positive AMPLICOR result in 23%. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the AMPLICOR test were 77%, 100%, 99%, 90%, and 93%, respectively. In comparison, they were 96%, 97%, 94%, 98%, and 97%, respectively, for the integrated clinical approach. CONCLUSIONS In the rapid diagnosis of PTB, the clinical judgment of specialists augmented the utility of the NAA test: (1) specialists selected patients with high-to-moderate pretest probabilities, (2) they commenced treatment promptly on a positive NAA test result, and (3) they were willing to start treatment in some patients on the basis of high clinical suspicion despite negative smear and negative NAA test results.