Impact of additional antibiotics on in-hospital mortality in tuberculosis isolated general bacteria: A propensity score analysis.

ABSTRACT

Whether or not additional antibiotics with anti-tuberculosis agents are required to treat bacterial co-infection with pulmonary tuberculosis is unclear. We aimed to assess the impact of additional antibiotics on mortality in pulmonary tuberculosis patients whose sputum cultures were positive for general bacteria as a surrogate definition of bacterial pneumonia. This study was a single-center retrospective cohort using a propensity score analysis. We included patients who were admitted for pulmonary tuberculosis and whose sputum cultures were positive for general bacteria. The mortality of patients who received additional antibiotics was analyzed after adjusting for other variables, including the propensity score predicting treatment with additional antibiotics. We assessed 68 and 55 tuberculosis patients treated with and without general antibiotics, respectively. Additional antibiotics tended to be administered to patients with a high level of C-reactive protein and neutrophil count, poor performance status, hypoxemia and hypoalbuminemia (C-statistics of area under receiver operating characteristic curve to the propensity score; 0.884, p < 0.001). In the multivariate analysis, advanced age and not the use of additional antibiotics was associated with in-hospital mortality. Additional antibiotics with anti-tuberculosis agents may not improve the prognosis of pulmonary tuberculosis patients whose sputum cultures were positive for general bacteria. Isolation of general bacteria does not equate to complication with bacterial pneumonia, so physicians should not administer general antibiotics to TB patients based solely on the results of sputum culture for general bacteria. A prospective study is needed to verify these results using a more accurate definition of pulmonary tuberculosis complicated with bacterial pneumonia.