Guiding antibiotic treatment with inflammatory biomarkers in COPD? Another brick in the wall

Acute exacerbations of chronic obstructive pulmonary disease (COPD) often prompt initiation of empirical antibiotic treatment, although in many cases a bacterial pathogen cannot be detected, and viruses may indeed account for a large proportion of episodes. Indeed, data supporting the effectiveness of the broad antibiotic utilisation at exacerbation not requiring intensive care are insufficient. Herein, personalising antibiotic treatment based on a patient's individual risk for bacterial infection has great potential to improve antibiotic stewardship efforts to encourage judicious and correct usage of these agents and mitigate the emergence of multidrug-resistant pathogens, one of the most urgent threats to global health and directly linked to antibiotic overuse. Integration of host response markers, which correlate with bacterial infection, into the overall assessment and clinical care of patients with acute exacerbation of COPD has high potential to improve individual antibiotic decisions. Among such promising host response markers in acute exacerbation of COPD, procalcitonin (PCT), a marker specific to bacterial infections, and C-reactive protein (CRP), a more general inflammatory marker with high sensitivity, have generated most interest. Several randomised trials have found PCT to result in a significant reduction of antibiotic usage with a similar resolution of clinical symptoms in patients with sepsis and respiratory infections, including acute exacerbation of COPD [1, 2]. In fact, a meta-analysis based on individual data of 1252 patients with COPD exacerbation found PCT guidance to result in a significant reduction in antibiotic initiation (72% versus 43%) and antibiotic exposure (5.3 versus 3.1 days) with no difference in mortality (4% versus 3%) or risk of treatment failure (17% versus 17%) [3]. Still, a recent trial investigating COPD patients needing intensive care treatment did not report a significant effect of PCT on antibiotic usage, and non-inferiority of PCT in regard to clinical outcomes could not be demonstrated [4]. CRP has been used successfully to direct antibiotic treatment in primary care studies [5, 6], and observational research suggested CRP may be more suitable to direct antibiotic treatment in acute exacerbation of COPD compared to PCT [7]. Still, randomised trials looking at the effect of CRP for guiding antibiotic decisions in patients with acute COPD exacerbation have been largely missing.

In this issue of the European Respiratory Journal, Prins et al. [8] report the result of a trial in 220 patients with acute exacerbation of COPD, comparing antibiotic usage according to a CRP-based strategy with a strategy based on Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines. Subjects assigned to the CRP group were treated with antibiotic for 7 days if CRP on admission or after 24 h was ≥50 mg·L⁻¹. Subjects in the GOLD group were also treated with the same antibiotic regimen if they reported increased sputum purulence in combination with increased dyspnoea and/or increased sputum volume, as propagated by the GOLD guidelines. The results showed a significant reduction in antibiotic initiation of more than 30% (from 46.2% to 31.7%) associated with the CRP algorithm. There was no difference in clinical outcomes in the short term and after 1-year of follow up, yet the study did not have the statistical power to assess safety of this approach and type II error is possible. Also of note, in 40 patients who were not treated initially, antibiotics were prescribed due to treatment failure. This deviation from the protocol, however, was well-balanced between groups.

Decisions regarding antibiotic use in an individual patient with acute exacerbation of COPD is complex and should be based on several considerations including the pretest probability for bacterial infection, which is based on clinical examination, severity of the disease and results from microbiological tests, the severity of acute episode and results of host response markers [9]. In the context of a low risk situation and a low pretest probability for bacterial infections, a low level of host response markers aids in ruling out bacterial infection and empirical antibiotic therapy can be avoided. In some patients, this strategy may not play out and clinical assessment and retesting is mandatory. While there was data on the use of PCT to direct antibiotic decisions from several trials, Prins et al. [8] have now provided important study results implying that CRP may also be helpful to direct treatment. While results are most convincing regarding the lowering of antibiotic exposure, proof of safety using CRP is still pending. Ultimately, a head to head trial will be needed to compare a CRP versus a PCT strategy taking into account effects on antibiotic treatment, patient outcomes and costs [6].

There is a strong interest in the medical community in reducing unnecessary antibiotic exposure. The urgency in rationalising and individualising antibiotic decisions is particularly meaningful in a situation in which antibiotics can be withheld in the vast majority of the cases, such as for exacerbations of COPD [10]. Integration of host response markers into treatment decision, now also including CRP, has shown promising results. These markers, however, should not be used as a substitute for good clinical practice, but rather be part of the overall assessment of a patient. Decisions pertaining to the initiation and cessation of antibiotic treatment remains strongly dependent on an assessment of all available clinical and diagnostic parameters, including a thorough patient assessment and the severity of the illness. Still, host response markers remain the best line of defence against diagnostic uncertainty and antibiotic overuse. Further research is needed to compare markers head to head, and to explore their optimal combination with pathogen-directed tests.