Background: Prompt diagnosis of pulmonary tuberculosis (TB) is necessary to minimize the risk of transmission to patients and health care providers. We sought to identify factors that could facilitate establishment of the diagnosis at an urban hospital in the United States and to determine rates of antibiotic resistance.
Methods: Cases of pulmonary TB were reviewed, and epidemiological, clinical, and microbiological data were analyzed.
Results: Fifty-one patients with pulmonary TB were identified during a 5-year period. Five (10%) of the patients identified were positive for human immunodeficiency virus. Among the 5 US-born patients, 2 had no identifiable risk factors for TB. Classic signs, symptoms, and laboratory abnormalities were not reliable for identifying TB patients. Four patients (2 of whom were positive for human immunodeficiency virus) had initial chest x-rays read as normal by radiologists. Sputum smears and culture were negative in 24% and 14% of patients, respectively. Sputum samples collected every 8 hours were similarly sensitive for detecting TB compared with samples collected once per day. Drug resistance was frequent, including 1 case of multidrug-resistant and 2 cases of extreme drug-resistant TB. One-third of patients required treatment modification owing to a combination of toxicities or the susceptibility profile of their infecting organisms.
Conclusions: A high index of suspicion must be maintained to make the diagnosis of TB in the United States, given frequent lack of classically described signs or symptoms of infection. Our findings support the use of sputum testing every 8 hours, which requires substantially less in-patient resources than once-daily testing. Antibiotic resistance and drug toxicities commonly complicate treatment.