BACKGROUND:Rational prescription of antibiotics in acute exacerbations of COPD (AECOPD) requires predictive markers. We aimed to analyze whether markers of systemic inflammation can predict response to antibiotics in AECOPD. METHODS:We used data from 243 exacerbations out of 205 patients from a placebo-controlled trial on doxycycline in addition to systemic corticosteroids for AECOPD. Clinical and microbiologic response, serum C-reactive protein (CRP) level (cutoffs 5 and 50 mg/L), and serum procalcitonin level (PCT) (cutoffs 0.1 and 0.25 μg) were assessed. RESULTS:Potential bacterial pathogens were identified in the majority of exacerbations (58%). We found a modest positive correlation between PCT and CRP (r = 0.46, P < .001). The majority of patients (75%) had low PCT levels, with mostly elevated CRP levels. Although CRP levels were higher in the presence of bacteria (median, 33.0 mg/L [interquartile range, 9.75-88.25] vs 17 mg/L [interquartile range, 5.0-61.0] [P = .004]), PCT levels were similar. PCT and CRP performed similarly as markers of clinical success, and we found a clinical success rate of 90% in patients with CRP ≤ 5 mg/L. A significant effect of doxycycline was observed in patients with a PCT level < .1 μg/L (treatment effect, 18.4%; P = .003). A gradually increasing treatment effect of antibiotics (6%, 10%, and 18%), although not significant, was found for patients with CRP values of ≤ 5, 6-50, and > 50 mg/L, respectively. CONCLUSIONS:Contrary to the current literature, this study suggests that patients with low PCT values do benefit from antibiotics. CRP might be a more valuable marker in these patients.