© 2006 American Thoracic Society doi: 10.1164/rccm.200607-922ED
Biomarkers in Acute Exacerbation of Chronic Obstructive Pulmonary DiseaseAmong the Blind, the One-Eyed Is KingUniversity Hospital, Basel, Basel, Switzerland Exacerbations in chronic obstructive pulmonary disease (COPD) are associated with impaired quality of life, decline in lung function, poorer outcome, and consume enormous health care resources. The distinction of day-to-day variations of symptoms from exacerbations is challenging. Recording of symptoms by patients and their interpretation by physicians may be prone to subjective interobserver variability, suggesting that more objective criteria of disease activity are needed. Pulmonary biomarkers have been extensively assessed in COPD (1). However, tissue sample analysis requires invasive procedures and the high variability of measurements affects the clinical utility of noninvasive approaches, such as determination of biomarkers in breath condensate. Therefore, finding a biomarker in the blood that defines exacerbations would be a highly useful clinical advance. In this issue of the Journal (pp. 867–874), Hurst and colleagues assessed the value of 36 biomarkers measured in blood samples of 90 patients with COPD before and during an exacerbation (2). Two findings of this study merit our attention. First, for routine use, 35 of these 36 markers do not appear capable of enlightening our clinical judgment. These results may seem disappointing but are important to be reported. Why has there been no change in the level of most biomarkers during the exacerbation? Have the authors chosen the wrong markers? The latter is unlikely, as many of the tested biomarkers have been shown to play a role in inflammation and tissue remodeling in COPD. Therefore, despite considerable impact on quality of life and prognosis, exacerbations seem to represent more a local than systemic problem. Systemic inflammation in COPD might be more related to the disease itself as to acute exacerbations. Second, among the investigated biomarkers, C-reactive protein (CRP) was the most effective for differentiating exacerbation from day-to-day symptom variation. CRP is an acute-phase reactant routinely used in Europe for the diagnosis of infectious diseases. An additional role of CRP in the characterization of low-grade inflammation has been put forward, not only in cardiovascular risk assessment but also for the prognosis of COPD (3). As a mediator, binding of CRP to exposed ligands in damaged cells may lead to complement-mediated exacerbation of tissue injury (4). Diagnostic accuracy of CRP for acute infections is suboptimal, with impaired specificity in patients with severe inflammation and lower sensitivity due to a protracted response and blunted levels when immunosuppressive therapies are administered, particularly corticosteroids (5). On the other hand, the routine use of CRP is facilitated by low costs, easy availability, and extensive clinical experience. As mentioned by Hurst and colleagues, an association between CRP and exacerbations has been reported earlier (6), but the present article contains a more comprehensive analysis on the diagnostic value of CRP in well-characterized patients. Sensitivity of CRP at 5 mg/L was 74.4% and specificity for the confirmation of exacerbation was 57.5%. This seems unacceptably low for a "good" marker. The combination of CRP with any major exacerbation symptom improved the diagnostic accuracy. However, this approach to improve the area under curve (AUC) by combining CRP with symptoms is scientifically questionable considering that symptoms have been used to define the exacerbation. On the basis of these observational data, we think it would be premature to monitor CRP in patients with COPD without exacerbations, despite the proven correlation with overall mortality (3). First, physiologic fluctuations of highly sensitive CRP (hsCRP) are substantial, which makes individual therapeutic decisions difficult. In women, hsCRP levels vary up to threefold during the menstrual cycle (7). Patients with severe COPD are often treated with steroids that markedly suppress CRP (5). Conversely, steroid withdrawal may increase hsCRP levels regardless of an exacerbation. Let's assume that CRP in a patient with COPD is 4 mg/L and doubles to 8 mg/L. This level might urge the physician to initiate steroid and antibiotic treatment, in a "preemptive" strike to avoid more severe exacerbation and/or disease progression. Such an approach is obviously of interest but has to be prospectively examined in a randomized study to prove that "early treatment" improves long-term outcome. The value of CRP for the diagnosis of exacerbations could be judged through a different approach than only looking at sensitivity and specificity. Exacerbations are triggered mainly by viruses. Around 20% of patients with exacerbations are estimated to benefit from antibiotics because only the acquisition of new strains of bacteria is believed to cause bacterial exacerbations (8). CRP levels are highest in bacterial infection (6, 9). Therefore, the fact that a considerable number of patients show normal levels of CRP during exacerbations does not necessarily mean that CRP is a poor biomarker. Low levels of CRP might just reflect that some patients have no evidence of bacterial infection, and exacerbations might be caused by viruses or concomitant heart failure requiring different treatment. Weis and Almdal recently showed that 30% of 166 patients with an exacerbation of COPD showed normal CRP values, whereas median CRP was 45 mg/L in 51 patients reporting increased sputum purulence and 97 mg/L in 51 patients with a concomitant infiltrate (9). The clinical usefulness of a cutoff value of a biomarker depends on the underlying disease and the potential decision for treatment (10). In 243 patients with lower respiratory tract infection, we found that the sensitivity of CRP at a cutoff of 50 mg/L was 93.8% to detect bacterial infection requiring antibiotic treatment (11). Thus, among the blind—despite a moderate sensitivity and specificity—CRP seems to be a one-eyed biomarker for defining an exacerbation. Using one eye in the right direction may still help us to make correct decisions for the treatment of patients with COPD exacerbations. Any observational study is biased by the choice of the "gold standard," and for COPD exacerbation this gold standard does not exist. Accordingly, observational studies might overestimate the diagnostic value of a marker. Interventional studies, in which therapy is guided by cutoff ranges of a biomarker and in which efficacy is the outcome measure, have the potential to provide more clinically relevant data. Conceptually, the level of a surrogate marker like CRP mirrors only the likelihood for a disease. Regrettably, Hurst and colleagues did not use their observational experience to suggest specific cutoff ranges for future intervention studies. For procalcitonin, this has been done, and attempts to guide antibiotic therapy in lower respiratory tract infections have shown promising results (12, 13). Results of a recently completed intervention study performed in COPD exacerbations show that antibiotic therapy can safely be guided using a defined cutoff for procalcitonin (14). The time has arrived to move beyond the observational reporting of "promising" biomarkers. Specific cutoff ranges must be proposed and intervention studies conducted. Only this will open both our eyes, and reveal if CRP or other biomarkers can really help us in the routine care of patients with COPD. FOOTNOTES Conflict of Interest Statement: Neither author has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
