Objective: Both bacteria and viruses may cause acute exacerbations of chronic obstructive pulmonary disease (AECOPD). The objective of this study was to identify readily available clinical parameters to discriminate between them.
Methods: During a winter period all consecutive patients with an AECOP who were hospitalized in a non-ICU general ward were prospectively enrolled. In addition to blood tests, cultures of spontaneous or induced sputum samples, and genome detection of respiratory viruses in nasopharyngeal swab samples using multiplex RT-PCR assays were obtained. Only patients with positive microbiological results (bacteria, virus, or both) were eventually included. Mixed infections (bacteria plus viruses) were categorized into the bacterial group due to therapeutic implications (ie, need for antibiotics). Demographic and routine clinical and analytical information was collected.
Results: A total of 127 AECOPD patients out of 213 initially evaluated met inclusion criteria and were classified as having bacterial (70, 55.1%) or viral (57, 44.9%) infection. Although no single variable was useful to identify bacteria, the combination of serum C-reactive protein >70 mg/L (2 points), >1 day of symptoms (1.5 points), and a blood neutrophil count >9,500 x109/L (1 point) into a scoring system reached an AUC of 0.80 (95% CI=0.73-0.88) for bacterial etiologies. With this model, scoring 0 or 1 point significantly reduced the probability of a bacterial infection (likelihood ratio negative of 0.2), whereas summing up 2.5 points or more increased it sufficiently to be clinically meaningful (likelihood ratio positive >3.7). Viral infections resulted in fewer hospitalization days (78.9% of patients spent ≥3 days in hospital vs 95.7% of those with bacterial infections; P=0.008).
Conclusion: A simple and easy to obtain score system can help clinicians in the decision of prescribing antibiotics in AECOPD patients.
Keywords: C-reactive protein; COPD; respiratory infections.
© 2022 Ruiz-González et al.