Abstract
Impact: This study explores the use of measuring plasma biomarkers at exacerbation of chronic obstructive pulmonary disease (COPD), providing insight into the underlying pathogenesis of these important events.
Rationale: The use of measuring C-reactive protein (CRP) to confirm exacerbation, or to assess exacerbation severity, in COPD is unclear. Furthermore, it is not known whether there may be more useful systemic biomarkers.
Objective: To assess the use of plasma biomarkers in confirming exacerbation and predicting exacerbation severity.
Methods: We assessed 36 biomarkers in 90 paired baseline and exacerbation plasma samples from 90 patients with COPD. The diagnosis of exacerbation fulfilled both health care use and symptom-based criteria. Biomarker concentrations were related to clinical indices of exacerbation severity. Interrelationships between biomarkers were examined to gain information on mechanisms of systemic inflammation at exacerbation of COPD.
Measurements and Main Results: To confirm the diagnosis of exacerbation, the most selective biomarker was CRP. However, this was neither sufficiently sensitive nor specific alone (area under the receiver operating characteristic curve [AUC], 0.73; 95% confidence interval, 0.66–0.80). The combination of CRP with any one increased major exacerbation symptom recorded by the patient on that day (dyspnea, sputum volume, or sputum purulence) significantly increased the AUC to 0.88 (95% confidence interval, 0.82–0.93; p < 0.0001). There were no significant relationships between biomarker concentrations and clinical indices of exacerbation severity. Interrelationships between biomarkers suggest that the acute-phase response is related, separately, to monocytic and lymphocytic–neutrophilic pathways.
Conclusions: Plasma CRP concentration, in the presence of a major exacerbation symptom, is useful in the confirmation of COPD exacerbation. Systemic biomarkers were not helpful in predicting exacerbation severity. The acute-phase response at exacerbation was most strongly related to indices of monocyte function.
The C-reactive protein (CRP) assay is widely performed in the assessment of many acute medical conditions including exacerbations of chronic obstructive pulmonary disease (COPD). However, CRP concentrations are elevated in most infective, inflammatory, and neoplastic pathologies (1), and the relevance in a particular disease of a high CRP level is often less well defined. This is particularly true for conditions such as COPD, where much of the morbidity, mortality, and health care costs are associated with acute-on-chronic deteriorations, with the presence of elevated systemic inflammation even in the stable state (2).
Acute-on-chronic deteriorations of respiratory health in COPD are termed exacerbations. Exacerbations are important events in the natural history of COPD, contributing to the decline in lung function (3) and impairment to health status (4), and responsible for many of the hospital admissions (5) and therefore health care costs and mortality (6) associated with the disease. It is now accepted that respiratory viruses and pathogenic bacteria cause most exacerbations and the presence of bronchial infection would be expected to cause a rise in CRP. However, as noted above, COPD is associated with raised systemic inflammation in the stable state (2) and therefore the relevance of an elevated CRP concentration at exacerbation is unclear.
The definition of exacerbation in COPD remains controversial (7): whether reliant solely on symptoms or also requiring an element of health care use, criteria remain subjective, either on the part of the patient or the physician. This creates difficulty, both in clinical practice and research, in reliably distinguishing exacerbations from day-to-day symptom variation. There is therefore the need for a “biomarker” to provide objective confirmation of exacerbation. Furthermore, a useful biomarker might not only differentiate stable disease from exacerbation, but may also predict the severity of such events. CRP is widely measured at exacerbation but the use of CRP as an exacerbation biomarker is unclear. There are no previous studies of other systemic biomarkers in this context.
The primary aim of this study was to assess the use of plasma biomarkers in the confirmation of COPD exacerbation. We have performed the largest and most comprehensive study of plasma biomarkers at exacerbation of COPD ever reported. Thirty-six candidate molecules were assessed in paired baseline and exacerbation plasma samples from 90 patients, tested against a “gold standard” definition of exacerbation that met both health care use and prospective symptom-based criteria. We also examined relationships between biomarker concentrations and exacerbation severity, and investigated interrelationships between exacerbation biomarkers to gain information about the underlying pathogenic mechanisms of systemic inflammation during such events.
Some of the results of this study have been previously published in the form of abstracts (8, 9).
The following is an abridged version of Methods; see the online supplement for the full version.
The samples analyzed were obtained from 90 patients enrolled in the London COPD Study. The cohort is as previously described (3, 4). Inclusion and exclusion criteria are reported in the online supplement. Patients included in this analysis were those in whom paired stable and exacerbation plasma samples were both available. The cohort has been the subject of numerous previous publications, including analysis of interleukin (IL)-6 and CRP from serum samples obtained at the same visits as the plasma samples reported in this analysis (10–13).
Plasma samples were obtained from all 90 patients at baseline and exacerbation. Each patient contributed one baseline and one exacerbation sample. The patients were trained to record daily, on diary cards, their morning peak expiratory flow (PEF) rate and any change in respiratory symptoms. At the time of any change in symptoms, patients were instructed to contact the study team and attend the research clinic for clinical assessment, sampling, and confirmation of exacerbation, using the criteria described below. Plasma samples at exacerbation were therefore obtained early in the time course of the event, before the initiation of any additional treatment. At baseline visits the diary cards were carefully examined to exclude the presence of exacerbation, and to assess the degree of day-to-day symptom variation, also as described below.
The study had approval from the local research ethics committee, and all participants provided written, informed consent.
Two definitions of exacerbation were employed in this analysis. Exacerbations were initially confirmed on the basis of our previously reported symptom-based criteria, adapted from that of Anthonisen and coworkers (14), and now validated against important outcome measures in COPD including the rate of lung function decline (3), airway inflammatory markers (15), and quality of life (4). The basis of our definition is the recording by patients of changes in respiratory symptoms. Patients record symptoms only when they are new or worse than usual. Symptoms recorded were categorized as major (dyspnea, sputum volume, or sputum purulence) and minor (cough, wheeze, sore throat, or coryza). An exacerbation was defined as the onset of two or more new or worsening symptoms, on two or more consecutive days, with at least one symptom being major. All symptoms, whether major or minor, were binary coded as present/increased over baseline (1) or absent/not increased (0) and summed to yield a Symptom Count that was calculated daily. The Symptom Count provides an assessment of exacerbation severity (and day-to-day symptom variability) as described further below.
The attending physician judged that antibiotics and/or corticosteroids were required in all 90 exacerbations and therefore all 90 exacerbations met both symptom-based and health care use definitions of exacerbation. All 90 patients at baseline visits were judged by the attending physician to be in a stable state and not to require additional therapy, and although individual symptoms may have been recorded at these visits (reflecting baseline symptom variability in COPD), these did not meet our definition of exacerbation.
Exacerbation severity was assessed using the magnitude of change at exacerbation onset, and recovery time, for both PEF and Symptom Count. Further detail of this methodology is included in the online supplement.
The selected biomarkers were those that had best distinguished patients with COPD from matched control subjects in a previous unpublished analysis from a separate cohort (GlaxoSmithKline, data on file). The study team selected additional candidate molecules to give 36 analytes in total. Five milliliters of venous blood was collected into a citrate Vacutainer (BD Diagnostics, Franklin Lakes, NJ). The tube was centrifuged for 10 min at 2,000 rpm, at 4°C, and the plasma was decanted and stored at –70°C for later analysis of candidate biomarkers. A SearchLight proteome array for the 36 analytes was manufactured by Pierce Biotechnology, Inc. (Woburn, MA), and validated at the GlaxoSmithKline Human Biomarker Center (Research Triangle Park, NC). Before assay, frozen plasma samples were thawed and transferred to ABgene 2DCypher tubes (ABgene, Epsom, UK). The sample was diluted with SearchLight sample diluent to a concentration appropriate to maximize the dynamic range. Further detail of this methodology is described in the online supplement.
Data are presented as means and standard deviation (SD) or as medians and interquartile range (IQR) as appropriate. All statistical tests, unless otherwise stated, employed Wilcoxon matched-pair signed rank tests or Spearman rank correlations. The use of plasma biomarkers in confirming exacerbation was investigated by receiver operating characteristic (ROC) analysis. A Bonferroni correction was applied for the analysis of relationships between individual markers, and for the correlations between marker concentrations and indices of exacerbation severity. Full Bonferroni correction would be highly conservative for an analysis of changes in biomarker concentration between baseline and exacerbation, potentially missing real differences. In the latter analysis we have therefore taken a p value of less than 0.01 as statistically significant, balancing type I and type II errors. The similarity between changes in marker concentrations at exacerbation is illustrated in a dendrogram, as described further in the online supplement and in Figure 4.
The funding source had no role in the sample collection, data analysis, interpretation of data, or decision to submit this article for publication. The study design, selection of analytes, and analysis of the biomarkers were performed in association with GlaxoSmithKline and the report was reviewed by three authors employed by GlaxoSmithKline for important intellectual content.
Baseline clinical characteristics of the 90 patients included in this analysis are reported in Table 1, which demonstrates that the cohort has moderately severe COPD with a median (IQR) FEV1 of 1.00 (0.74–1.33) L or 43.9 (27.5–56.8)% predicted. Most of the patients were taking inhaled corticosteroids with a median beclomethasone-equivalent dose of 1,000 (800–2,000) μg/d. There were no significant relationships between the dose of inhaled steroid and the concentrations of inflammatory markers at baseline. Four percent of the patients were taking long-term oral prednisolone, too small for a meaningful comparison of subjects who were and were not taking such therapy.
Mean | SD | |
|---|---|---|
| Age, yr | 70.1 | 8.2 |
| FEV1/FVC, % | 46.9 | 14.1 |
| PaO2, mm Hg | 67.6 | 7.7 |
| Median | IQR | |
| FEV1, L | 1.00 | 0.74–1.33 |
| FEV1, % predicted | 43.9 | 27.5–56.8 |
| FVC, L | 2.24 | 1.73–3.04 |
| PaCO2, mm Hg | 42.5 | 39.2–46.5 |
| Smoking, pack-years | 45 | 29–59 |
| Exacerbations, per yr | 2.61 | 1.86–3.67 |
| BMI, kg m−2 | 24.7 | 20.8–29.4 |
| Inhaled corticosteroid dose* | 1,000 | 800–2,000 |
| % | ||
| Current smoking | 27.8 | |
| Maintenance oral steroids | 4.0 | |
| Ischemic heart disease | 20.5 | |
| Cardiac failure | 10.3 | |
| Hypertension | 35.1 |
Concentrations of the 36 plasma biomarkers, at baseline and exacerbation, and the magnitude of change between these two clinical states, are reported in Table 2. The significant differences observed between baseline and exacerbation for CRP, IL-6, myeloid progenitor inhibitory factor-1 (MPIF-1), pulmonary and activation-regulated chemokine (PARC), adiponectin (ACRP-30), and soluble intercellular adhesion molecule-1 (s-ICAM-1) are illustrated in Figure 1 (using p < 0.01, as described above).
Figure 1. Candidate plasma biomarkers present at significantly different concentrations between baseline and exacerbation in chronic obstructive pulmonary disease (COPD; p < 0.01). n = 90. Box plots represent median, interquartile range, and range with outliers plotted separately. ACRP30 = adiponectin; CRP = C-reactive protein; ICAM-1 = soluble intercellular adhesion molecule-1; MPIF-1 = myeloid progenitor inhibitory factor-1; PARC = pulmonary and activation-regulated chemokine.
[More] [Minimize]Marker | Units | Baseline Median (IQR) | Exacerbation Median (IQR) | Median (% change) | p Value (Wilcoxon) |
|---|---|---|---|---|---|
| CRP | mg/L | 4.0 (2.0–12.0) | 15.6 (4.5–74.0) | +185 | < 0.001 |
| IL-6 | pg/ml | 1.55 (0.94–3.07) | 3.25 (1.48–6.12) | +66 | < 0.001 |
| MPIF-1 | pg/ml | 734 (574–944) | 901 (72–1,237) | +18 | < 0.001 |
| PARC | pg/ml | 1.1 (0.8–1.5) × 105 | 1.3 (0.9–1.7) × 105 | +10 | 0.002 |
| ACRP-30 | pg/ml | 1.5 (0.9–2.3) × 107 | 1.6 (1.1–2.6) × 107 | +11 | 0.001 |
| s-ICAM-1 | pg/ml | 4.8 (3.7–5.9) × 105 | 5.0 (4.1–6.4) × 105 | +6 | 0.003 |
| Amphiregulin | pg/ml | 8.85 (5.86–11.40) | 8.89 (5.87–12.30) | 0 | 0.936 |
| BDNF | pg/ml | 4,958 (4,202–8,018) | 4,921 (4,114–6,821) | −6 | 0.256 |
| β-NGF | pg/ml | 2.11 (0.78–4.46) | 2.47 (0.94–4.16) | 0 | 0.743 |
| ENA-78 | pg/ml | 1,152 (752–1,934) | 1,038 (814–1,751) | −4 | 0.358 |
| Eotaxin-2 | pg/ml | 1,435 (944–2,052) | 1,315 (978–1,911) | −7 | 0.029 |
| Erb-B2 | pg/ml | 3,248 (2,827–4,002) | 3,271 (2,703–4,121) | 0 | 0.746 |
| Fibronectin | pg/ml | 3.0 (2.4–3.8) × 108 | 3.0 (2.2–4.0) × 108 | +1 | 0.734 |
| IFN-γ | pg/ml | 1.2 (0.5–3.0) | 1.6 (0.7–3.1) | −1 | 0.808 |
| IL-1β | pg/ml | 0.68 (0.40–0.95) | 0.69 (0.45–1.10) | −1 | 0.343 |
| IL-1Ra | pg/ml | 61.1 (42.2–89.6) | 71.2 (51.0–130.0) | +10 | 0.014 |
| IL-2Rγ | pg/ml | 26.9 (22.3–33.5) | 29.6 (22.5–36.7) | +6 | 0.059 |
| IL-8 | pg/ml | 2.5 (1.7–4.9) | 2.6 (1.9–4.9) | −2 | 0.642 |
| IL-12 p40 | pg/ml | 7.0 (4.9–10.7) | 7.7 (4.4–14.9) | +2 | 0.385 |
| IL-15 | pg/ml | 0.85 (0.60–1.10) | 0.95 (0.63–1.30) | +9 | 0.065 |
| IL-17 | pg/ml | 5.5 (2.9–9.6) | 6.4 (3.5–9.8) | +8 | 0.351 |
| IP-10 | pg/ml | 152 (106–234) | 192 (137–252) | +20 | 0.011 |
| ITAC | pg/ml | 13.3 (0.8–13.9) | 19.6 (1.6–39.2) | +1 | 0.232 |
| MCP-1 | pg/ml | 444 (370–534) | 439 (359–537) | −3 | 0.842 |
| MIP-1β | pg/ml | 57.8 (41.3–80.0) | 59.6 (40.0–85.7) | −1 | 0.933 |
| MMP-9 | pg/ml | 39,597 (25,802–85,597) | 33,542 (23,284–56,772) | −6 | 0.689 |
| MPO | pg/ml | 11,417 (7,921–21,237) | 13,280 (8,483–22,613) | −4 | 0.575 |
| Prolactin | pg/ml | 1,028 (710–1,499) | 928 (689–1,298) | −3 | 0.103 |
| RANTES | pg/ml | 32,968 (20,641–58,916) | 36,687 (22,341–52,693) | +10 | 0.161 |
| L-selectin | pg/ml | 7.6 (6.5–8.6) × 105 | 7.6 (6.5–8.5) × 105 | −1 | 0.485 |
| TGF-α | pg/ml | 2.5 (1.3–4.5) | 3.0 (1.5–4.8) | +7 | 0.308 |
| TIMP-1 | pg/ml | 63,979 (50,733–76,540) | 64,685 (51,946–75,712) | +1 | 0.734 |
| TNF-α | pg/ml | 1.9 (0.7–3.5) | 1.8 (0.9–3.7) | +6 | 0.249 |
| TNFR1 | pg/ml | 601 (445–773) | 654 (507–926) | +11 | 0.011 |
| TNFR2 | pg/ml | 1,580 (1,267–2,228) | 1,627 (1,249–2,394) | +3 | 0.177 |
| VEGF | pg/ml | 0.01 (0.01–0.62) | 0.01 (0.01–0.38) | 0 | 0.725 |
All 90 exacerbations fulfilled both our prospective, validated symptom-based definition (3) and the need for additional therapy. Judged against either of these standards, using the area under the curve (AUC) and 95% confidence intervals (CI) of ROC analysis, the three best performing biomarkers were CRP (AUC, 0.73; 95% CI, 0.66–0.80), IL-6 (AUC, 0.67; 95% CI, 0.59–0.74), and MPIF-1 (AUC, 0.64; 95% CI, 0.56–0.72), all less than the accepted standard of AUC ⩾ 0.8 (16). The AUC for CRP was significantly greater than that for MPIF-1 (p = 0.03), but not for IL-6 (p = 0.11).
Illustrating the clinical implications of these data, the widely employed clinical CRP cutoff at 5 mg/L would be 74.4% sensitive and 57.5% specific for confirmation of exacerbation. Furthermore, to achieve 90% sensitivity, a CRP concentration of 2.3 mg/L would be only 29% specific, whereas at 90% specificity a CRP of 27.6 mg/L would be only 40% sensitive. Moreover, assessed by AUC, assay of these biomarkers in isolation would not provide a more useful confirmation of exacerbation than the presence of an increase in any one of the major exacerbation symptoms alone on that day (AUC, 0.83; 95% CI, 0.77–0.89). This is illustrated in Figure 2.
Figure 2. Receiver operating characteristic (ROC) curves for CRP alone, any one major symptom alone, and the combination of any one major symptom with CRP for the confirmation of exacerbation in COPD. Area under the receiver operating characteristic curve and 95% confidence interval: 0.73 (0.66–0.80), 0.83 (0.77–0.89), and 0.88 (0.82–0.93), respectively.
[More] [Minimize]Regarding the ability of combinations of biomarkers to confirm exacerbation, of the 630 biomarker pairs only the AUC for combinations of CRP with matrix metalloproteinase-9 (MMP-9), MPIF-1, and vascular endothelial growth factor (VEGF) were greater than that for CRP alone (respectively: AUC, 0.73; 95% CI, 0.66–0.81; AUC, 0.73; 95% CI, 0.60–0.81; and AUC, 0.73; 95% CI, 0.65–0.80). These results were not significantly different from CRP alone (p > 0.05), and still not greater than the commonly employed cutoff at AUC ⩾ 0.8 (16). Similarly, the best combination of three biomarkers (CRP, MMP-9, and MPIF-1) resulted in an AUC of 0.75 (95% CI, 0.67–0.82), also not significantly better than CRP alone.
Figure 2 also depicts the ROC curve for CRP and the presence or increase in one or more major (14) exacerbation symptoms at that visit, judged against the health care use definition. The AUC was 0.88 (95% CI, 0.82–0.93), significantly greater than that for either CRP (p < 0.0001) or symptoms alone (p = 0.004), and of a magnitude considered to be clinically useful (16). Therefore, in the presence of one major symptom recorded on that day, CRP ⩾ 8 mg/L would be 95% specific for exacerbation and 57% sensitive. A model involving CRP and one major symptom was also significantly better than using all 36 biomarkers in combination, in the absence of symptom assessment (AUC, 0.79; 95% CI, 0.73–0.86; p = 0.03).
At exacerbation, PEF decreased by a median (IQR) of 14.2 (27.1 to –1.4) L/min and had recovered to baseline in 6 (1 to 13.5) d. The median (IQR) rise in Symptom Score at onset was 2.85 (2.00 to 3.00) and the Symptom Score had recovered to baseline over a median period of 11 (6.5 to 17) d. Three of the 90 exacerbations (3.3%) resulted in hospitalization.
We examined relationships between the concentration of systemic biomarkers at exacerbation, and the change in concentration from baseline to exacerbation, with the four indices of exacerbation severity: the magnitude of change at onset, and recovery time in PEF and Symptom Count. When corrected for multiple comparisons, there were no statistically significant relationships for any of the 36 analytes.
We investigated interrelationships between the concentrations of plasma biomarkers at exacerbation, using Spearman rank correlations, fully corrected for multiple comparisons. This resulted in a matrix of 630 correlations. We were particularly interested in relationships involving the acute-phase proteins CRP and IL-1 receptor antagonist (IL-1Ra) given reported associations between systemic inflammation and cardiovascular morbidity (17). The interrelationships at exacerbation involving CRP and IL-1Ra are summarized in Figure 3, and described below. There were no other significant correlations involving these biomarkers with any of the remaining 25 analytes. Figure 3 illustrates, as expected, that there was a significant correlation between plasma IL-6 and CRP concentrations (rho = 0.60, p < 0.001). The concentration of IL-6 was also associated with the concentrations of two further groups of proteins: the monocyte marker MPIF-1 and, separately, a network of seven interrelated proteins (comprising IL-1Ra, IL-8, IL-12, IL-15, IL-2 receptor γ [IL-2Rγ], nerve growth factor-β [β-NGF], and amphiregulin) associated with lymphocyte and granulocyte function.
Figure 3. Analysis of relationships between exacerbation concentrations of the 36 biomarkers resulted in a matrix of 630 correlations. This analysis was fully Bonferroni corrected. We wished to examine associations of the acute-phase response at exacerbation, and the data presented here demonstrate all the analytes significantly associated with the acute-phase reactants interleukin-1 receptor antagonist (IL-1Ra), IL-6, and CRP. Concentrations of eight other biomarkers were related, either directly or indirectly, to IL-1Ra, IL-6, and CRP, as illustrated, whereas concentrations of the other 25 biomarkers were not significantly correlated. In addition to demonstrating which relationships were present, the lines also indicate the strength of the relationships (solid line, r2 ⩾ 0.5, interrupted line, r2 < 0.5). bNGF = nerve growth factor-β.
[More] [Minimize]Regarding changes in the concentrations of markers between baseline and exacerbation, only the changes in MPIF-1 and IL-1Ra remained significantly associated with the change in IL-6 (rho = 0.78, p < 0.0001, and rho = 0.53, p < 0.0001, respectively). As expected, the increase in IL-6 remained significantly associated with the increase in CRP (rho = 0.63, p < 0.0001). Relationships between the changes in concentrations of biomarkers between baseline and exacerbation are illustrated in Figure 4. The dendrogram is constructed to graphically present similarities in the change in concentration with exacerbation of the 36 biomarkers. Conceptually, the position of each biomarker is defined by its correlation with the other biomarkers: the closest or most highly correlated are joined first, and this connection has the shortest horizontal distance on the dendrogram. Once joined, the average position of the pair is used in linking with the next most closely related biomarker. This process is repeated until all the biomarkers are linked. Thus, for example, the changes in concentration at exacerbation for IL-6, MPIF-1, and CRP were closely correlated, as were the changes in myeloperoxidase and MMP-9, but the changes observed between these two groups were not closely related.
Figure 4. Dendrogram demonstrating similarities between the magnitude of changes in concentration between baseline and exacerbation for the 36 plasma biomarkers. As explained in text, the position of each biomarker is defined by its correlation with the other biomarkers: the closest or most highly correlated are joined first, and this connection has the shortest horizontal distance on the dendrogram. Once joined, the average position of the pair is used in linking with the next most closely related biomarker. This process is repeated until all the biomarkers are linked. Thus, for example, the changes in concentration at exacerbation for IL-6, MPIF-1, and CRP were closely correlated, as were the changes in myeloperoxidase (MPO) and matrix metalloproteinase-9 (MMP-9), but the changes observed between these two groups were not closely related. For definitions of abbreviations, see Table 2 footnote.
[More] [Minimize]This study has demonstrated that plasma CRP, in the presence of one or more recorded major exacerbation symptoms, is able to reliably differentiate exacerbation of COPD from day-to-day symptom variation, judged against either a validated symptom-based definition or the need for additional exacerbation therapy. CRP (or any other of the plasma analytes alone) was neither sufficiently sensitive nor specific to be a useful biomarker in the absence of symptom assessment. CRP was the most useful of the 36 biomarkers assessed. Plasma biomarkers were not useful in predicting the clinical severity of exacerbation. Interrelationships between plasma biomarkers at exacerbation suggest that the acute-phase response is related, separately, to both monocytic and granulocytic–lymphocytic indices, with the strongest relationship for the monocyte marker MPIF-1.
There is ongoing debate about how exacerbation of COPD should be defined (7). In practical terms, an exacerbation represents an acute deterioration in symptoms that is beyond the patient's usual day-to-day variation. However, this definition requires subjective assessment by both the patient and the physician, and there is a need in both the clinic and the context of clinical trials for an objective method of confirming exacerbation: a “biomarker.” A useful biomarker might also reflect exacerbation severity or etiology. We have previously developed a symptom-based definition of exacerbation that has been validated against important outcome measures in COPD including airway inflammatory markers (15), quality of life (4), and rate of decline in FEV1 (3). This method requires daily recording of changes in symptoms by patients, and careful interpretation of data, such that it may be regarded as rather complex and expensive. However, data collected on the basis of this definition have allowed us to determine that up to 50% of exacerbations go unreported to health care professionals, and therefore that definitions of exacerbation based solely on health care use criteria underestimate the true impact of these events (4).
Because of these difficulties in objectively assessing symptom changes, there has been great interest in the development of a biomarker that is present at different concentration during exacerbations from that present in stable disease. Accepted definitions of COPD now include reference to the presence of airway inflammation, but airway inflammatory markers are difficult to assess. In general, bronchoalveolar lavage and endobronchial biopsy are too invasive, and sputum and exhaled breath condensate are too complex to analyze, producing results that are inconsistent across studies. Both approaches are therefore impractical in the clinical setting. Attention has consequently turned to systemic markers and, indeed, it is now widely accepted that COPD is associated with increased systemic inflammation compared with control subjects (2), and that there is further up-regulation of systemic inflammation at the time of exacerbation. Markers reported to be higher in blood during exacerbation compared with the baseline state include CRP (12, 18–21), IL-8 (19), tumor necrosis factor-α (22), leptin (22), endothelin-1 (23), eosinophil cationic protein (24), myeloperoxidase (24), fibrinogen (25), IL-6 (12, 25), α1-antitrypsin (20, 21, 26), and leukotriene E4 (27). However, such findings have not necessarily been replicated between studies, and no studies have specifically reported the use of such assessment in the confirmation of exacerbation. Interestingly, we have reported that serum IL-6 and CRP concentrations at exacerbation are correlated with selected markers of airway inflammation, and are higher in the presence of a bacterial pathogen (12). This suggests that assay of systemic inflammatory markers may to some extent reflect inflammatory load in the airway, although the precise source of systemic inflammation remains to be confirmed.
Clearly, there is the need for a plasma biomarker in COPD that will not only aid in distinguishing stable disease from exacerbation, but might also have prognostic value. Fortuitously, given the ready availability of CRP assay, CRP was the most useful biomarker of the 36 candidate molecules assessed. In part, this may reflect previous findings that CRP is an exceptionally stable analyte (1). There is an extensive literature reporting associations of CRP with clinical parameters in stable COPD including the severity of FEV1 impairment (28), impaired energy metabolism, decreased exercise tolerance and poorer quality of life (29), and the presence of a sputum bacterial pathogen (10). However, evidence to support the use of CRP in confirming or excluding exacerbation in COPD is minimal. One study, examining the use of CRP to identify exacerbations associated with a sputum bacterial pathogen, observed that CRP was not elevated in 16% of patients with hospitalized exacerbations (30). This may have reflected the difficulty arising from raised CRP levels in the stable state, and although the change in CRP concentration may be more useful, few patients will have baseline levels immediately available at the time exacerbation diagnosis is required.
In this study, CRP was the most effective biomarker for differentiating exacerbation from day-to-day symptom variation. CRP is an acute-phase protein produced by the liver in response to IL-6 stimulation (1). CRP is raised in most conditions associated with infection, inflammation, or tissue damage, for which it is a sensitive marker (1). Evidence suggests that CRP may also be implicated in the pathophysiology of COPD: many deaths at exacerbation are due to cardiovascular disease (31) and an elevated CRP concentration is associated with increased cardiovascular morbidity (17). CRP may even have a direct role in the instability of atheromatous plaques, having been identified in such lesions ex vivo (32). Furthermore, by binding to the low-density lipoproteins present in plaques, CRP may activate the complement cascade (33). CRP and IL-6 could therefore provide a link between airway inflammation, systemic inflammation, and cardiovascular disease in COPD that is worthy of further study. Moreover, therapies that reduce CRP in COPD could impact on this excess cardiovascular morbidity.
This study is the first to report on the use of systemic biomarkers, with and without symptoms, in the confirmation of exacerbation in COPD. The strengths of the present analysis are the assessment of 36 candidate biomarkers, in 90 exacerbations from 90 separate patients, with exacerbations meeting both health care use and symptom-based definitions. Moreover, individual symptom changes were also recorded at baseline (allowing us to exclude intercurrent respiratory illnesses) and both the patients and the exacerbations were comprehensively assessed, enabling an analysis of the relationships between plasma markers and clinical indices of exacerbation severity. The completion of daily diary cards enabled us to precisely define the onset and therefore duration of changes in symptoms and PEF. Relationships between plasma inflammatory indices and the clinical severity of exacerbation were absent, consistent with previous studies (30), and this may reflect the heterogeneity of such events. Our patients had a range of disease severity, and we believe these patients to be representative of a wider population with COPD. However, most of the patients were undergoing inhaled corticosteroid therapy, which is known to reduce CRP in COPD (34), and the results may therefore not be applicable to a steroid-naive population. It is important to remark that the methodology employed did not allow us to assess the use of biomarkers in distinguishing exacerbations from other causes of symptom deteriorations in patients with COPD, such as cardiac failure or pneumonia, diagnoses that were excluded in our patients clinically and, where appropriate, with investigations involving such techniques as electrocardiography and chest radiography.
Investigating interrelationships between biomarkers has the potential to yield information about the underlying mechanisms of exacerbation and therefore to aid development of novel therapeutic strategies. We have particularly focused on relationships involving acute-phase proteins given the recognized associations between systemic inflammation and cardiovascular risk (17), and the ready availability of CRP assay in many health care settings. As expected, we demonstrated relationships between the concentrations of plasma IL-6 and the hepatic acute-phase proteins CRP and IL-1Ra. However, we have also demonstrated that IL-6 is associated with markers from both monocytic and, separately, neutrophilic–lymphocytic pathways. The former relationship remained statistically significant when assessing correlations between the changes in biomarker concentration at exacerbation. Such findings suggest that strategies aimed at damping monocyte-macrophage–mediated inflammation may be the most successful in reducing cardiovascular risk in the periexacerbation period.
Regarding interrelationships between inflammatory markers, MPIF-1, a newly identified CC chemokine, appeared to be the most closely associated with plasma IL-6 concentration. So named from a suppressive effect on hematopoiesis, MPIF-1 has chemotactic activity on resting (but not activated) T lymphocytes and monocytes, but is less active on neutrophils (35). To our knowledge, this is the first report of an MPIF-1 assay performed on patients with COPD.
In summary, in patients with COPD presenting with increased symptoms, CRP was the most effective of 36 biomarkers studied to confirm exacerbation. The CRP cutoff may be chosen to maximize sensitivity or specificity, as required. CRP alone was neither sufficiently sensitive nor specific to confirm exacerbation, and the magnitude or rise in CRP concentration did not reflect the clinical severity of the event. Finally, interrelationships between biomarkers at exacerbation suggest that monocytic pathways may be an important target for future research, with the aim of reducing the considerable cardiovascular morbidity associated with this prevalent disease.
| 1. | Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest 2003;111:1805–1812. |
| 2. | Gan WQ, Man SFP, Senthilselvan A, Sin DD. Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and a meta-analysis. Thorax 2004;59:574–580. |
| 3. | Donaldson GC, Seemungal TAR, Bhowmik A, Wedzicha JA. Relationship between exacerbation frequency and lung function decline in chronic obstructive pulmonary disease. Thorax 2002;57:847–852. |
| 4. | Seemungal TAR, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;157:1418–1422. |
| 5. | Ashton CM, Ferguson JA, Goldacre MJ. In-patient workload in medical specialities. 2. Profiles of individual diagnoses from linked statistics. Q J Med 1995;88:661–672. |
| 6. | Soler-Cataluña J, Martínez-García MA, Román Sánchez P, Salcedo E, Navarro M, Ochando R. Severe acute exacerbations and mortality in patients with chronic obstructive pulmonary disease. Thorax 2005;60: 925–931. |
| 7. | Rodriguez-Roisin R. Toward a consensus definition for COPD exacerbations. Chest 2000;117:398S–401S. |
| 8. | Hurst JR, Donaldson GC, Perera WR, Wilkinson TM, Bilello JA, Hagan GW, Vessey RS, Wedzicha JA. Changes in plasma biomarkers at exacerbation of COPD: a proteomic approach [abstract]. Proc Am Thorac Soc 2006;3:A839. |
| 9. | Hurst JR, Donaldson GC, Perera WR, Wilkinson TM, Bilello JA, Hagan GW, Vessey RS, Wedzicha JA. Utility of plasma biomarkers to confirm exacerbation in COPD [abstract]. Proc Am Thorac Soc 2006;3:A271. |
| 10. | Hurst JR, Wilkinson TMA, Perera WR, Donaldson GC, Wedzicha JA. Relationships among bacteria, upper airway, lower airway, and systemic inflammation in COPD. Chest 2005;127:1219–1226. |
| 11. | Donaldson GC, Seemungal TA, Patel IS, Bhowmik A, Wilkinson TM, Hurst JR, MacCallum PK, Wedzicha JA. Airway and systemic inflammation and decline in lung function in patients with COPD. Chest 2005;128:1995–2004. |
| 12. | Hurst JR, Perera WR, Wilkinson TMA, Donaldson GC, Wedzicha JA. Systemic and upper and lower airway inflammation at exacerbation of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2006;173:71–78. |
| 13. | Wilkinson TMA, Hurst JR, Perera WR, Wilks M, Donaldson GC, Wedzicha JA. Effect of interactions between lower airway bacterial and rhinoviral infection in exacerbations of COPD. Chest 2006;129: 317–324. |
| 14. | Anthonisen NR, Manfreda J, Warren CP, Hershfield ES, Harding GK, Nelson NA. Antibiotic therapy in exacerbations of chronic obstructive pulmonary disease. Ann Intern Med 1987;106:196–204. |
| 15. | Bhowmik A, Seemungal TAR, Sapsford RJ, Wedzicha JA. Relation of sputum inflammatory markers to symptoms and lung function changes in COPD exacerbations. Thorax 2000;55:114–120. |
| 16. | Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978;8:283–298. |
| 17. | Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, Lowe GD, Pepys MB, Gudnason V. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 2004;350:1387–1397. |
| 18. | Gompertz S, O'Brien C, Bayley DL, Hill SL, Stockley RA. Changes in bronchial inflammation during acute exacerbations of chronic bronchitis. Eur Respir J 2001;17:1112–1119. |
| 19. | Spruit MA, Gosselink R, Troosters T, Kasran A, Gayan-Ramirez G, Bogaerts P, Bouillon R, Decramer M. Muscle force during an acute exacerbation in hospitalised patients with COPD and its relationship with CXCL8 and IGF-I. Thorax 2003;58:752–756. |
| 20. | Crooks SW, Bayley DL, Hill SL, Stockley RA. Bronchial inflammation in acute bacterial exacerbations of chronic bronchitis: the role of leukotriene B4. Eur Respir J 2000;15:274–280. |
| 21. | Hill AT, Bayley D, Stockley RA. The interrelationship of sputum inflammatory markers in patients with chronic bronchitis. Am J Respir Crit Care Med 1999;160:893–898. |
| 22. | Calikoglu M, Sahin G, Unlu A, Ozturk C, Tamer L, Ercan B, Kanik A, Atik U. Leptin and TNF-α levels in patients with chronic obstructive pulmonary disease and their relationship to nutritional parameters. Respiration (Herrlisheim) 2004;71:45–50. |
| 23. | Roland M, Bhowmik A, Sapsford RJ, Seemungal TA, Jeffries DJ, Warner TD, Wedzicha JA. Sputum and plasma endothelin-1 levels in exacerbations of chronic obstructive pulmonary disease. Thorax 2001;56:30–35. |
| 24. | Fiorini G, Crespi S, Rinaldi M, Oberti E, Vigorelli R, Palmieri G. Serum ECP and MPO are increased during exacerbations of chronic bronchitis with airway obstruction. Biomed Pharmacother 2000;54:274–278. |
| 25. | Wedzicha JA, Seemungal TA, MacCallum PK, Paul EA, Donaldson GC, Bhowmik A, Jeffries DJ, Meade TW. Acute exacerbations of chronic obstructive pulmonary disease are accompanied by elevations of plasma fibrinogen and serum IL-6 levels. Thromb Haemost 2000;84: 210–215. |
| 26. | Stockley RA, Burnett D. α-Antitrypsin and leukocyte elastase in infected and noninfected sputum. Am Rev Respir Dis 1979;120:1081–1086. |
| 27. | Shindo K, Hirai Y, Fukumura M, Koide K. Plasma levels of leukotriene E4 during clinical course of chronic obstructive pulmonary disease. Prostaglandins Leukot Essent Fatty Acids 1997;56:213–217. |
| 28. | Mannino DM, Ford ES, Redd SC. Obstructive and restrictive lung disease and markers of inflammation: data from the Third National Health and Nutrition Examination. Am J Med 2003;114:758–762. |
| 29. | Broekhuizen R, Wouters EFM, Creutzberg EX, Schols AMWJ. Raised CRP levels mark metabolic and functional impairment in advanced COPD. Thorax 2006;61:17–22. |
| 30. | Dev D, Wallace E, Sankaran R, Cunniffe J, Govan JR, Wathen CG, Emmanuel FX. Value of C-reactive protein measurements in exacerbations of chronic obstructive pulmonary disease. Respir Med 1998;92: 664–667. |
| 31. | Hansell AL, Walk JA, Soriano JB. What do chronic obstructive pulmonary disease patients die from? A multiple cause coding analysis. Eur Respir J 2003;22:809–814. |
| 32. | Torzewski J, Torzewski M, Bowyer DE, Frohlich M, Koenig W, Waltenberger J, Fitzsimmons C, Hombach V. C-reactive protein frequently colocalizes with the terminal complement complex in the intima of early atherosclerotic lesions of human coronary arteries. Arterioscler Thromb Vasc Biol 1998;18:1386–1392. |
| 33. | Bhakdi S, Torzewski M, Klouche M, Hemmes M. Complement and atherogenesis: binding of CRP to degraded, nonoxidised LDL enhances complement activation. Arterioscler Thromb Vasc Biol 1999;19: 2348–2354. |
| 34. | Man SFP, Sin DD. Effects of corticosteroids on systemic inflammation in chronic obstructive pulmonary disease. Proc Am Thorac Soc 2005;2: 78–82. |
| 35. | Patel VP, Kreider BL, Li Y, Li H, Leung K, Salcedo T, Nardelli B, Pippalla V, Gentz S, Thotakura R, Parmelee D, Gentz R, Garotta G. Molecular and functional characterization of two novel human C–C chemokines as inhibitors of two distinct classes of myeloid progenitors. J Exp Med 1997;185:1163–1172. |
