Nine eligible studies (four for delays and five for resistance issues) were included in the meta-analysis from the 770 articles originally identified in the database search. The mean duration of delayed diagnosis and treatment of pulmonary TB in the fluoroquinolone prescription group was 19.03 days, significantly longer than that in the non-fluoroquinolone group (95% confidence interval (CI) 10.87 to 27.18, p less than 0.001). The pooled odds ratio of developing a fluoroquinolone-resistant M. tuberculosis strain was 2.70 (95% CI 1.30 to 5.60, p=0.008). No significant heterogeneity was found among studies in the meta-analysis.
Empirical fluoroquinolone prescriptions for pneumonia are associated with longer delays in diagnosis and treatment of pulmonary TB and a higher risk of developing fluoroquinolone-resistant M. tuberculosis.