Sputum bacteriology and clinical response to antibiotics in moderate exacerbation of chronic obstructive pulmonary disease

Background Presence of purulent sputum during an acute exacerbation of chronic obstructive pulmonary disease (AECOPD) is considered sufficient indication for starting empirical antibiotics. We investigated the relationship between detection of potentially pathogenic bacteria (PPB) using sputum culture or polymerase chain reaction (PCR) and clinical response and sought the risk factors for PPB growth. Methods In 342 outpatients with AECOPD, we compared detection rates of H. influenzae (HI) and S. pneumoniae (SP) using conventional sputum culture versus PCR. The utility of either technique to predict clinical cure or failure after effective antibiotics was assessed. The factors predicting positive sputum cultures were evaluated using logistic regression. Results Using sputum culture, 132 PPB were detected. The predominant bacteria were HI (40.9%) and SP (19.7%). Detection of HI or SP in sputum was higher using PCR than culture growth (60.8% vs 18.6%; P  < .001). Clinical response was not affected by the results of either technique. Independent risk factors for PPB isolation were Gram‐negative bacteria on sputum smear (OR 15.78, 95% CI 6.38‐39.06; P  < .001), sputum purulence (OR 2.31, 95% CI, 1.05‐5.11; P  = .04), body temperature (OR 0.16, 95% CI 0.05‐0.54; P  = .003), albumin level (OR 0.29, 95% CI 0.09‐0.88; P  = .03) and dyspnea grade (OR 0.51, 95% CI 0.27‐0.96; P  = .04). Conclusions Neither culture growth nor PCR positivity for HI or SP in sputum predicted clinical response to antibiotics; therefore, these tests are not necessary for outpatients with AECOPD. Examining Gram‐staining and purulence on sputum smear, however, was significant to predict PPB growth in sputum.