Objective: To evaluate the clinical utility of the polymerase chain reaction (PCR) in the diagnosis of infections due to Mycobacterium tuberculosis.
Design: Clinical specimens were assayed by PCR for the presence of the insertion element IS6110, a DNA sequence unique to the M tuberculosis complex of organisms. The PCR results were then correlated with acid-fast bacilli (AFB) smears, cultures, pathology, and clinical histories.
Setting: Bellevue Hospital, a large municipal teaching hospital.
Patients: Inpatients on the Bellevue Chest Service.
Measurements and results: Sixty-five patients were evaluated. The PCR for M tuberculosis was positive in 37 patients and negative in 28. When correlated with smears, cultures, pathology, and clinical history, the sensitivity of PCR for a diagnosis of active tuberculosis (TB) was 100 percent. However, the specificity for a diagnosis of active TB was only 70 percent, as the PCR assay was positive in a number of patients with only prior, treated TB, or asymptomatic tuberculous infection. For a diagnosis of any TB infection (active, treated, or asymptomatic), sensitivity of PCR was 87.5 percent and specificity was 90 percent.
Conclusions: The PCR assay for TB is extremely sensitive, but it lacks specificity for a diagnosis of active TB. Its role in clinical practice will likely be limited to well-defined situations, such as HIV-positive patients with intrathoracic adenopathy, and it may be most useful in excluding active TB from consideration in selected patients. Given the cost of the assay and the labor intensity it requires, it should not be part of the routine initial evaluation of patients with suspected pulmonary TB.