The accuracy of diagnosing pulmonary tuberculosis at a teaching hospital

Accuracy of the clinical diagnosis of tuberculosis and of the mycobacteriology laboratory test results was assessed in a teaching hospital by reviewing clinical and microbiologic data on patients from whom Mycobacterium tuberculosis had been recovered. Mycobacteria were isolated in 230 of 6,550 specimens (3.5%). Clinical data were available for 42 patients with tuberculosis, 20 of whom had significant underlying nonmycobacterial disease. Positive tuberculin skin tests wee recorded for 90% of the patients with no underlying disease and for 29% of the patients with underlying disease. Tuberculosis was not suspected initially in 16 of 32 patients with pulmonary disease, and had not been diagnosed by the time of discharge in 10 patients. Of all respiratory specimens from patients with cavitary disease, 57% of the acid-fast stains and 96% of the cultures were positive. In contrast, 32% of stains and 70% of cultures were positive from patients with noncavitary pulmonary tuberculosis. One false-positive acid-fast stain was observed during this study.