Diagnosis of pulmonary tuberculosis using PCR assays on sputum collected within 24 hours of hospital admission

There have been few studies evaluating the efficacy of polymerase chain reaction (PCR) testing in front-line clinical practice. We assessed the diagnostic yield of PCR prospectively in a blinded study of patients admitted to rule out tuberculosis and compared PCR results to a culture and clinical diagnosis of tuberculosis. Specimens were sent for routine smear, culture, and PCR analysis. Sputum sediments were submitted for PCR amplification of IS6110 sequences by an in-house assay and also the Roche Amplicor PCR assay targeting 16s ribosomal RNA genes. Eighty-five patients were enrolled: 27 patients had cultures positive for tuberculosis; 12 were smear-positive. PCR by both assays on the first specimen picked up all patients smear-positive on any specimen. A positive PCR on at least one of two specimens collected in the first 24 h was 85 and 74% sensitive and 88 and 93% specific for tuberculosis by the in-house and Roche techniques, respectively. Sensitivity in smear-negative patients was 73 and 53%, respectively. The in-house PCR detected 100% and Roche detected 95% of patients with more than paucibacillary (greater than 20 colonies) tuberculosis. We conclude that PCR may be a useful tool to evaluate patients for tuberculosis within the first hospital day.