- Email: [email protected]
C-Reactive Protein Levels and Survival in Patients With Moderate to Very Severe COPD
COPD
C-Reactive Protein Levels and Survival in Patients With Moderate to Very Severe COPD* Juan P. de Torres, MD; Victor Pinto-Plata, MD, FCCP; Ciro Casanoua, MD; Hanna Mulleroua, PhD; Elizabeth C6rdoba-Lantk PhD; Mercedes Muros de Fuentes, MD; Armundo Agufrre-Jaime, PhD; and Bartolome R. Celli, MD, FCCP
Background: Serum levels of C-reactive protein (CRP) are increased in patients with COPD and correlate modestly with variables predictive of outcomes. In epidemiologic studies, CRP level is associated with all-cause mortality in patients with mild-to-moderate disease. Objectke: To determine if CRP levels are assodated with survival in patients with moderate to very severe COPD in comparison with other well-known prognostic parameters of the disease. Methods: In 218 stable patients with COPD, we measured baseline serum CRP level, BODE (body mass index, obstruction, dyspnea, and exercise capacity) index and its components, arterial oxygenation (Pao,), inspiratory capacity (IC) to total lung capacity (TLC) ratio, and Charlson comorbidity score. We followed up the patients over time and evaluated the strength of the association between the variables and all-cause mortality. RecluZtcl: During the follow-up time (median,36 months; 25th to 75th percentiles, 24 to 50 months), 54 patients (25%) died. CRP levels were similar between survivors and the deceased (median, 3.8 mgL; 95% confidence interval, 1.9 to 8.1; vs median, 4.5 m@; 95% confidence interval, 2.1 to 11.5; p = 0.22) and was not significantly associated with survival. ConcZusions: In t h i s population of patients with clinically moderate to very severe COPD, the level of CRP level was not associated with survival compared with other prognostic clinical tools such as the BODE index, modified Medical Research Council scale, 6-min walk distance, percentage of predicted FEVI, IC/TLC ratio < 0.25, and Pao,. Other long-term studies of well-characterized patients with COPD could help determine the exact role of CRP levels as a (CHEST 2008; 133:1336-1343) biomarker in patients with clinical COPD. Key words: COPD; C-reactive protein; survival Abbreviations: ATS = American Thoracic Society; BMI = body mass index; BODE = body mass index, obstruction, dyspnea, and exercise capacity; CRP = C-reactive protein; COLD = Global Initiative for Chronic Obstructive Lung Disease; IC = inspiratory capacity; MMRC = modified Medical Research Council; TLC = total lung capacity; 6MWD = Bmin waking distance
OPD, a major cause of morbidity and mortality C worldwide,’ is thought to result from an abnor-
mal pulmonary inflammatory response to noxious particles or gases.2 COPD is associated in some patients with a “low grade systemic inflammation” that may be responsible for the systemic effects of the disease; malnutrition, muscle wasting, osteoporosis, cardiovascular disease, type I1 diabetes mellitus, anemia, and depression.3 Many different inflammatory markers appear to be increased in the serum of stable COPD patient^.^ 1336
The best studied of these markers is the C-reactive protein (CRP), an acute-phase reactant secreted by the liver in response to infection, inflammation, or tissue damage.5 In patients with stable COPD, CRP
For editorial comment see page 1296 levels are directly associated with age, weight, dyspnea, and quality of life,fi.7and inversely related to exercise capacity and the degree of airway obstructi0n.7.~The higher levels of CRP observed in patients Original Research
with COPD increase even more during exacerbations? and appear to decrease in patients receiving inhaled corticosteroids.7J0 The exact role of CRP levels as a predictor of cardiovascular and global mortality in the general population has been postulated but not yet fully accepted." Two epidemiologic studies12.13 have shown that increased CRP levels are independently associated with global and cardiovascular mortality in COPD patients with mild-to-moderate degrees of airway obstruction. To explore whether CRP levels are associated with survival in patients with more severe degree of airway obstruction, we followed up a cohort of well-characterized patients with moderate to very severe COPD6v7and determined if their initial CRP level was associated with their risk of death. We also compared this association with that of other prognostic parameters that can be obtained in clinical practice, the BODE (body mass index, obstruction, dyspnea, and exercise capacity index) and its components,l4 Pao,, and inspiratory capacity (IC)/total lung capacity (TLC),15in predicting survival in the same population.
CRP Measurement
Fasting blood samples were obtained with the patient at rest, before any other test was performed. CRP concentrations were determined using high-sensitivity immunoassay (in Tenerife: CRP assay [latex] HS; Roche Diagnostics; Indianapolis, in Boston: HS-CRP; DPC Immulite; Los Angeles, CA) following manufacturer instructions. The coefficients of variation measured at each site for Roche Latex HS and for DPC Immulite were 6% and 896, respectively. Analytical sensitivity of these high-sensitivity assays is 0.1 m@, and measurement range is 0.1 to 20 mg /L. Clinical Param.eters
Both sites were part of the international BODE study and followed the exact same methodologicguidelines. Body mass index (BMI) was calculated as the weight in kilograms divided by height in metes? Patients performed spimmehy and lung volume measurements following American Thoracic Society (ATS) remmmendations and standad references.16 The 1WL.Cratio was calculated as previously describd with a value of < 0.25 used in the analysis because this threshold had the stmngest association with survival in COPD patients.15 Functional dyspnea was scored using the modified Medical Research Council (MMRC) scale.17The Bmin walking distance test (GMWD) was performed following the ATS recommendation.'* Pao2 was m e a s d in blood fmm the radial artery. The BODE index was ckssified in quartiles as previously descriM.14 The degree of comorbidities was measured using the m d e d Charhn index.19 Survival Determination
MATERIALS AND METHODS Study Population
This is a prospective observational study of patients from the two sites of the BODE cohort that had measured CRP levels at baseline.6.7 Patients from the pulmonary clinics at the Hospital Universitario Ntra Sra de Candelaria, in Tenerife, Spain, and from Caritas St Elizabeth's Medical Center in Boston, MA, signed the informed consent approved by the Human Institutional Review Boards. The patients were consecutively included ifthey had smoked B 20 pack-years and had a postbronchodilator FEV,/FVC ratio < 0.7 after 400 kg of inhaled albuterol. Patients were excluded if they had a history of asthma, bronchiectasis, tuberculosis, malignancy, arthritis, connective tissue disorders, or inflammatory bowel disease. The patients were clinically stable (no exacerbation for 2 months) at the time of the evaluation. *From the Respiratory Research Unit (Dn. de Torres, C6rdobaLank, and Aguirre-Jaime),Pulmonary (Dr. Casanova), and Biochemical Analysis (Dr. Mums de Fuentes) Departments, Hospital Nh-a Sra de Candelaria, Tenerife, Spain; Pulmonary and Critical Care Department (Drs. Pinto-Plata and Celli),Caritas St Elizabeth's Medical Center, Boston, MA; and GlaxoSmithKline (Dr. Mullerova), Worldwide Epidemiol Department, London, UK. The author have no conflicts3interest to disclose. Funding was provided partially by the Canarian Research and Health Foundation. Manuscript received October 1,2007; revision accepted January 18, 2008. Repduction of this article is rohibited without written permission from the American College o!Chest Physicians (w.chestjoumal. org/misc/re rintsshtml). C o m s p o d n c e to: Juan P. de Torres, MD, Unidud de Investigaci6n Hospital Ntra Sra de Candelaria, Ctra del Rosa& s/n, 38010 Santa Cruz de Tenerife, Canary Islands, Spain; e-mail: [email protected] DOI: 10.1378/chest.07-2433 www.chestjoumal.org
Patient survival status and cause of death were determined through medical record review and telephone interviews with family
Table l-Clinical and Physiologic Characteristics of the 218 Patients Enrolled in the Study* Characteristics
Data
Follow-up, mo Age*yr Gender Male Female Pack-yr history CRP level, mg/L BMI,kg/me FEV,, % predicted FVC,96 predicted GMWD, m IC/rLC < 0.25 MMRC scale 2 2 Pao,, mm Hg Charlson scale Cardiovascular risk factors or disease Use of inhaled corticosteroids GOLD stage Stage I Stage I1 Stage I11 Stage IV BODE index Quartiles First and semnd Third and fourth
36 (24-50) 6529 138 (64) 80 (36) 55 (40-79) 4 (1.9-9) 26 2 5 46 2 19 77 2 22 426 2 107 70 (32) 117 (53) 70? 12 2 (1-4) 73 (33) 110 (50)
157 (72) 61 (28)
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%). CHEST / 133 I 6 I JUNE, 2008
1337
members. In Boston, the status was canfirmed using the national social security registry. All-cause mortality during the follow-up period was used as the outcome measure. We also established the causes of death and classified them as respiratory, including respiratory failure and respiratory infection; cardiovascular, including sudden death, myocardial infarction, and cerebrovascular events; cancer, including all types of cancers; and others. Statistical Analysis
All continuous variables are expressed as mean (SD), median (25th to 75th percentiles), or as percentage, depending on distribution. CRP values were logarithmic transformed because of nonnomal distribution. Pearson x2 test, Mann-Whitney U test, or Student t test were used for two-group comparisons. To evaluate the strength of the association of death and the other variables, we performed a Cox regression analysis adjusting for age, gender, pack-year history, presence of cardiovascular risk factors or disease, and treatment with inhaled corticosteroids (none of the patients were receiving oral corticosteroids). Results are reported as hazard ratios with 95% confidence intervals; p I0.05 was considered statistically significant. Statistical analysis was performed using a commercial statistical package (SPSS version 13.0; SPSS; Chicago, IL).
RESULTS The clinical and physiologic charwteristics of all patients are summarized in Table 1. This predominantly male COPD population included patients with Global Initiative for Chronic Obstructive Lung Disease (GOLD)/ATS/European Respiratory Society stages I1 to IV.Most had a normal BMI, with 32% of them
having an I W L C ratio < 0.25. Half had an MMRC score > 2 and some degree of comorbidity. At the end of the follow-up time, 25% (54patients) had died and 7% (15 patients) were not available for follow-up. The characteristics of the 15 patients not available for follow-up were similar to those of the 149 patients who were alive at the end of the study. Table 2 shows the clinical and physiologic parameters in patients from Tenerife and Boston. No W e r ences were found in age, gender, BMI, comorbidity scores, Pao,, and CRP levels, but the patients from Boston had a greater smoking history, degree of airway obstruction, dyspnea, and higher BODE scores. Table 3 compares the clinical and physiologic parameters in survivors and nonsurvivors. Independent of study site, there were no differences in CRP, age, BMI, and comorbidity Charlson index. Differences were found in the intensity of smoking, degree of airflow obstruction, BODE index and its components (except the BMI), Pao,, and in ICRLC. The cause of death could be obtained only in the 23 patients from the Tenerife cohort. The living status and date of death could be retrieved in all of the Boston patients, but the actual cause of death was not available for most of them. In the available data, the most frequent cause of death was respiratory (43%),then cardiovascular diseases (26%),and finally all types of cancers (13%).
Table 2-Clinical and Physiologic Churacterietics of the 218 Patients Enrolled in the Study According to Study Site* Characteristics Follow-up, mo
4%yr Gender Male Female Pack-yr history CRP level, m e BMI, kg/m2 FEV,, % predicted WC, % predicted GMWD, m
ICRLC ratio < 0.25 MMRC scale 2 2 Pao,, mm Hg Charlson scale GOLD stage Stage I Stage I1 Stage I11 Stage N BODE index Quartiles First and second Third and fourth
Tenerife (n = 130)
Boston (n = 88)
30 (24-38) 63 2 10
57 (43-67) 65 2 8
83 (64) 47 (36) 50 (40-70) 4.1 (1.8-7.7) 27 5 5 53 2 20 83 2 23 460 2 102 29 (22) 40 (44) 71 2 11 2 (24)
55 (63) 33 (37) 65 (42-88) 3.9 (2-9.7) 27 2 5 35 2 10 67? 19 374 2 94 41 (46) 77 (88) 68 2 13 3 (27-5)
18 (14) 51 (39) 41 (32) 20 (15) 2 (1-3
0 (0 5 (6) 55 (62) 28 (32) 5 (35)
115 (89) 15 (11)
p Value 0.001 0.24 0.35
0.028 0.55 0.77 < 0.001 < 0.001 < 0.001 0.003 0.002 0.08
0.56 0.005
< 0.001
42 (48) 46 (52)
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%). 1338
Original Research
hPatfsnta Alive or Deceased at the End of the Study
Table 3-Diflerencee in the Studied Parametm
pQJJod* Parameters
Abve In = 149)
Deceased (n = 54)
Age. yr Pack-yr history CRP level, in@ BMI, k g h 2 FEV,, % predicted FVC, % predicted GMWD, m ICmLC ratio < 0.25 MMRC scale 2 2 Pao,, mm Hg Charlson scale Cardiovascular risk factors or disease Use of inhaled corticosteroids GOLD stage Stages I and I1 Stages 111 and IV BODE index Quartiles First and second Third and fourth
6428 56 (41-79) 3.8 (1.9-8.1) 27 t 5 4 8 k 19 79 2 22 446 2 97 33 (22) 76 (51) 72 2 10 2 (13) 54 (36) 80 (53)
66 2 9 71 (5p96) 4.5 (2.1-11.5) 26 2 5 35 2 14 66222 353 t 110 37 (65) 40 (75) 64 2 14 2(14 19 (35) 30 (55)
58 (38) 91 (62) 3( 1 4
5 (10) 49 (90) 5 (3-6)
118 (79) 31 (21)
D
Value
0.07 0.002 0.23 0.15 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.26 0.27 0.35 < 0.001
< 0.001 < 0.001
24 (45) 30 (55)
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%).
Table 4 shows the comparison of c h c a l and physiologic parameters from patients with initial CRP value 2 3 m@ or < 3 m@. We chose this cut-off value because it was previously proved to determine patient survival in the cardiovascular literature20.21 and in the COPD cohort reported by DahI and coworkers.13 No differences were found in all of the parameters except for a higher BMI and a lower Pao, in patients with higher CRP levels. Figure 1 shows no Werences in the Kaplan-Meier survival curves of patients with lower or higher CRP levels.
Table 5 shows the results of the Cox proportional hazard regression analysis. The analysis showed that most of the evaluated variables were associated with all-cause mortality, whereas the CRP did not.
DISCUSSION The most important finding of this study is that a baseline serum CRP level was not significantly associated with survival status in patients with clinically
Table I-Comparison of Clinical and Phyeiologic Parameters Between Patlents With Znitial CRP Levels s 3 mg/L or > 3 m a *
Pardmeters
CRP=3mg/L(n=87)
Deceased Age, yr Pack-yr history BMI, kg/m2 FEV,, % predicted FVC. % predicted GMWD, m I W L C ratio < 0.25 MMRC scale 2 2 Pao,, m m Hg Charlson scale GOLD stage Stages I and 11 Stages I11 and IV BODE index Quartiles First and second Third and fourth
20 (23) 63 t 9 50 (40-70) 2525 43 2 21 80 2 22 438 -C 103 30 (34) 47 (54) 73 2 11 2 (13)
CRP > 3 mg/L (n = 131)
34 (26) 6429 60 (40-80) 28 2 5
44 2 17 75 2 22 417 2 109
48 (36) 70 (53) 68 2 12 3 (1-4)
32 (27) 55 (63) 3 (1-5)
42 (32) 89 (68) 3 (2-5)
61 (71) 26 (29)
96 (73)
p Value 0.84 0.82 0.23 0.001 0.22 0.09 0.12 0.63 0.93 0.002 0.06 0.10
0.06
35 (27)
0.70
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%). www.chestjoumal.org
CHEST / 133 / 6 / JUNE, 2008
1339
FIGURE1. Kaplan-Meier suMval curves showing the accumulative mortality of patients participating in the study depending on a baseline CRP level > 3 mg/L or < 3 mg/L. The up er anel represents the entire cohort, whereas the lower panels provide similar information at eacg o f t h e study sites; p > 0.05, log-rank comparison between both groups.
stable moderate to very severe COPD. Numerous prospective studies have shown that the serum level of CRP is independently associated with future mortality. Levels of CRP are associated with cardiovascular risk of death in the general population21as well as in patients with other conditions, such as
diabetes22 and renal disease.= Althoug+ the important body of data built around CRP levels as a cardiovascular mortality marker suggests a role in its routine determination, cardiologists have not fully accepted yet its use in the risk assessment of individual patients.11
Table 5-Regression Analysis With Global Mortality aa Main Outcome and Each Studied Variable as a Predictor This Outcome* Regression Analysis Variables CRP
BMI FEV, % predicted MRC scale 2 2 BMWD ICRLC ratio < 0.25
'Hazard Ratio
95% Confidence Interval
1.00 0.94 0.96
0.98-1.02 0.89-1.00 0.94-O.9H
2.00 0.99
1.063.70
PdO,
4.66 0.95
0.98-0.99 2.53458 0.93-0.97
BODE index
1.41
1.22-1.61
p Valuel 0.56 0.07 0.001 0.03
< 0.001 < 0.001
< 0.001 < 0.001
*All analyses adjusted for age, gender, pack-year history, cardiovascular risk factors or disease, and treatment with inhaled corticosteroids. 1340
Original Research
To our knowledge, only two previous studies have examined the possible association of CRP levels and COPD outcomes. The first, by Man et al,I2 measured CRP levels (mean, 1.32 mg/L; interquartile range, 0.62 to 3.06 m@) in individuals from the Lung Health Study with mild-to-moderate ainvay obstruction (mean FEV,, 78 2 9% of predicted). They found that patients in the lowest CRP quartile had the highest risk of all-cause mortality, cancer deaths, and cardiovascular events, and suggested that CRP could be a marker of outcome in COPD. Recently, Dahl et al,13 using data from the Copenhagen City Heart Epidemiologic Study, found that the CRP level (mean, 3.32 m@; 25th to 75th percentiles, 1.6 to 6.9) was a strong and independent predictor of COPD outcomes represented by hospitalization and death. Both studies were adjusted by other confounding factors such as age, gender, race, smoking intensity, airway obstruction, and cardiovascular state. Our study population had much smaller size compared with those epidemiologic studies, and our patients had higher CRP levels (median, 4.1 m e ; 25th to 75th percentiles, 1.9 to 9 m@) than those from the previously reported studies. This is likely due to the greater degree of airway obstruction and disease severity in patients attending clinics compared with those enrolled in epidemiologic studies. CRP levels of the patients who died tended to be higher than those who survived, but the difference failed to reach statistical significance. When we categorized patients by the initial CRP level of 1 3 in@ or < 3 mg/L, we again found no differences in clinical or physiologic characteristics except for a higher BMI and lower Pao, in those patients with CRP levels 2 3 m@. This information could have different interpretations in our population of patients with moderate to very severe disease: first, that this threshold value is only valid to evaluate cardiovascular mortality in mild-to-moderate disease; second, that a merent threshold value (probably higher considering most COPD studies median CRP values) should be chosen to evaluate the importance of this biomarker in patients with severe to very severe disease; and third, that the level of CRP is not associated with increased risk of mortality in this particular degree of COPD severity. Indeed, in our cohort, CRP levels were not associated with survival, whereas the BODE index, a multidimensional prognostic tool,'* and most of its components were significant predictors of outcome. That CRP levels in patients iyith wellcharacterized COPD may not provide independent information about outcome in patients with clinically diagnosed COPD is not unique to our study. The same observation has also been found in the Framingham study,% in which elevated CRP levels did not www.chestjoumal.org
'
provide furthers prognostic information beyond traditional office risk factor assessment. Our study has three important differences compared with those of Man et all2 and Dahl and coworkers.13 They relate primarily to the populations included and the measured variables. With regard to study size, the previous studies were epidemiologic in design with large number of patients, whereas ours is a clinical study with 218 well-characterized COPD patients. In relation to the populations recruited, The Lung Health Study12 included patients with mild COPD likely with little or no systemic consequences of the disease, and in whom the main cause of mortality was cardiovascular disease, a finding that has been discussed by others.= Thus, it is possible that the elevated CRP was probably expressing the cardiovascular mortality risk. The third difference between our cohort and those reported previously reside in the variables measured. In the epidemiologic cohorts, spirometry was the only clinical variable used to evaluate COPD severity, and the outcome was determined using the ICD codes with all of its shortcomings. Our cohort was smaller, however, and had a higher percentage of deaths (25% vs 6%) over a shorter period of time, included patients with COPD as seen by clinicians, and the patients were more comprehensively evaluated using tools that have been shown to be associated with mortality better than the FEV,. It is possible that the level of CRP may be useful in epidemiologic studies but that it loses some of its association with survival once the patient has more clinically evident COPD. We believe that our findings should raise a word of caution to the excitement that this inflammatory marker has generated in the COPD field. CRP is an acute-phase reactant that increases in a very sensitive but nonspecific way in most forms of tissue damage, inflammation, and infection,5 all very dynamic processes in patients with COPDq9Indeed, it seems that there are genetic influences in the way patients with COPD react to inflarnmation,Z6 a factor that could have an important role in their serum CRP levels.27 Further CRP levels are influenced by many different factors such as cardiovascular disease, degree of physical activity, diabetes, renal disease, hypertension, metabolic syndrome, obstructive sleep apnea, smoking status, diet, and medications (corticosteroids, statins, angiotensin-converting enzyme inhibitors), to name a few. To our knowledge, only one study7 has determined the stability of CRP over time in a single cohort of a rather small number of patients. As suggested by Donaldson,28longitudinal studies are needed to confirm the stability of the CRP signal and its relationship with other clinical factors. Until then, the measurement of CRP level at CHEST f 133 f 6 f JUNE, 2008
1341
one single point may not be useful. Probably, serial measurements indicating its trend in relationship to interventions (long-term oxygen therapy, rehabilitation, inhaled corticosteroid, or statin therapy) could be one of the possible ways to evaluate the use of this inflammatory marker in COPD. Our study had several limitations. First, the relatively small population compared with the numbers reported in epidemiologic studies may be underpowered to adequately study the association between survival and CRP levels as and end point. Second, the findings are restricted to the population included in the study: patients from pulmonary clinics with moderate to very severe COPD. Third, the exact proportion of the causes of death may not be accurate because it was obtained from review of the medical records and the exact causes of death in many patients from the Boston cohort were not determined. This is a known limitation of many COPD survival studies.25 However, we used allcause mortality as the primary outcome, and survival status was known for 203 of the 218 patients. Fourth, even though we selected patients with no known illness that could cause increased CRP, many of the known factors likely to influence CRP could have occurred over time. However, this occurrence would have made our results even more relevant because they would have favored the presence of an association. An important additional strength of our study is that we observed the same findings in both study sites. This suggests that the lack of relationship between CRP and survival over time is consistent and reproducible in clinically and geographically different populations. In summary, we have found that in this population of patients with moderate to very severe COPD from a pulmonary clinic, CRP levels are not associated with survival status. CRP does not add information to the risk assessment provided by other tools such as the multidimensional BODE index and its components or the IC/TLC ratio. Other long-term studies in different settings are needed to evaluate the role of CRP and other markers2Qin patients with clinical COPD. REFERENCES 1 Pauwels RA, Rabe KF. Burden and clinical features of chronic obstructive pulmonary disease (COPD). Lancet 2004; 364613-620 2 Pauwels R, Buist SA, Calverley P, et al. Global strategy for the diagnosis, management and prevention of Chronic obstructive pulmonary disease: NHLBVWHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) workshop summary. Am J Respir Crit Care Med 2001;163:1256-1276 3 Agusti A. Systemic effects of chronic obstructive pulmoniuy disease. Proc Am Thorac SOC2005; 2:367-370 1342
10
11 12 13 14
15 16 17 18 19 20 21
22
23 24
Can WQ, Man SFP, Senthilselvan A, et al. Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and meta-analysis. Thorax 2004; 59:574-580 Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest 2003; 111:1805-1812 de Torres JP, Cordoba-Lanus E, Lopez-Aguilar C, et al. C-reactive protein levels and clinically important predictive outcomes in stable COPD patients. Eur Respir J 2006; 27:902-907 Pinto-Plata VM, Mullerova H, Toso JF, et al. C-reactive protein in patients with COPD, control smokers, and nonsmokers. Thorax 2006; 61:23-28 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 Hurst JR, Donaldson GC, Perera WR, et al. Use of plasma hiomarkers at exacerbation of chronic obstructive pulmonary disease. Am J Respir Crit Care Med; 2006; 1742367-874 Sin DD, La& P,-York E, et al. Effects of fluticasone on systemic markers of inflammation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2004; 170: 760-765 Lowe GD, Pepys MB. C-reactive protein and cardiovascular disease: weighing the evidence. Curr Atheroscler Rep 2006; 8~421-428 Man SFP, Connett JE, Anthonisen NR, et al. C reactive protein and mortality in mild to moderate chronic obstructive pulmonary disease. Thorax 2006; 61849-853 Dahl M, Vestbo J, Lange P, et al. C-reactive protein as a predictor of prognosis in COPD. Am J Respir Crit Care Med 2007; 175250-255 C e k BR, Cote CG, Marin JM, et al. The body mass index, airflow obstruction, dyspnea and exercise capacity index in chronic obstructive pulmonary disease. N Engl J Med 2004; 350:1005-1012 Casanova C, Cote C, de Torres JP, et al. The inspiratory to total lung capacity ratio predicts mortality in patients with COPD. Am J Respir Crit Care Med 2005; 171:591577 Quanjer PH. Standardized lung function testing: report of the Working P q for the European Community for Steel and Coal. Bull Eur Pbysiopath Respir 1983; 19(suppl5):22-27 Brooks SM. Survehlkce for respiratory hazards. ATS News 1982; 8:12-16 ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med 2002; 166:lll-117 Charlson M, Szatrowsky T, Peterson J, et al. Validation of a combined comorbidity index. J Clin Epidemiol 1994; 47: 1245-1251 Myers GL, Rifai N, Tracy RP, et al. CDC/AHA Workshop on Markers of Inflammation and Cardiovascular Disease. Circulation 2004; 11Oe545-e549 Ridker P. Clinical application of C-reactive protein for cardiovascular disease detection and prevention. Circulation 2003; 107:363-369 Linnemann B, Voigt W, Novel W, et al. C-reactive protein is a strong independent predictor of death in type 2 diabetes: association with multiple facets of the metabolic syndrome. Exp Clin Endocrinol Diabet 2006: 114:127-134 Racki S, Zaputovic L, Mavric Z, et al. C-reactive protein is a strong predictor of mortality in hemodialysis patients. Ren Fail 2006; 427-433 Wilson PW, Nam BH, Pencina M, et al. C-reactive protein and risk of cardiovascular disease in men and women from the Framingham Heart Study. Arch Intern Med 2005; 165:24732478 Original Research
25 Hansel1 A, Walk JA, Soriano JB. What do chronic obstruotive pulmonary disease patients die from? A multiple cause coding analysis. Eur Respir J 2003; 22:809-814 26 Yanbaeva DG, Dentener MA, Creutzberg EC, et al. Systemic inflammation in COPD: is genetic susceptibilitya key factor? COPD 2006; 3:51-61 27 Hersh CP, Miller DT, Kwiatkowski DJ, et al. Genetic deter-
w.chestjoumaI.org
minants of C-reactive protein in COPD. Eur Respir J 2006; 28:1156-1162 28 Donaldson G. C-reactive protein: does it predict mortality? Am J Respir Crit Care Med 2006; 175:209-210 29 Jones PW, Agusti AGN. Outcomes and markers in the assessment of chronic obstructive pulmonary disease. Eur Respir J 2006; 27:822-832
CHEST I 133 I 6 I JUNE, 2008
1343
C-Reactive Protein Levels and Survival in Patients With Moderate to Very Severe COPD* Juan P. de Torres, MD; Victor Pinto-Plata, MD, FCCP; Ciro Casanoua, MD; Hanna Mulleroua, PhD; Elizabeth C6rdoba-Lantk PhD; Mercedes Muros de Fuentes, MD; Armundo Agufrre-Jaime, PhD; and Bartolome R. Celli, MD, FCCP
Background: Serum levels of C-reactive protein (CRP) are increased in patients with COPD and correlate modestly with variables predictive of outcomes. In epidemiologic studies, CRP level is associated with all-cause mortality in patients with mild-to-moderate disease. Objectke: To determine if CRP levels are assodated with survival in patients with moderate to very severe COPD in comparison with other well-known prognostic parameters of the disease. Methods: In 218 stable patients with COPD, we measured baseline serum CRP level, BODE (body mass index, obstruction, dyspnea, and exercise capacity) index and its components, arterial oxygenation (Pao,), inspiratory capacity (IC) to total lung capacity (TLC) ratio, and Charlson comorbidity score. We followed up the patients over time and evaluated the strength of the association between the variables and all-cause mortality. RecluZtcl: During the follow-up time (median,36 months; 25th to 75th percentiles, 24 to 50 months), 54 patients (25%) died. CRP levels were similar between survivors and the deceased (median, 3.8 mgL; 95% confidence interval, 1.9 to 8.1; vs median, 4.5 m@; 95% confidence interval, 2.1 to 11.5; p = 0.22) and was not significantly associated with survival. ConcZusions: In t h i s population of patients with clinically moderate to very severe COPD, the level of CRP level was not associated with survival compared with other prognostic clinical tools such as the BODE index, modified Medical Research Council scale, 6-min walk distance, percentage of predicted FEVI, IC/TLC ratio < 0.25, and Pao,. Other long-term studies of well-characterized patients with COPD could help determine the exact role of CRP levels as a (CHEST 2008; 133:1336-1343) biomarker in patients with clinical COPD. Key words: COPD; C-reactive protein; survival Abbreviations: ATS = American Thoracic Society; BMI = body mass index; BODE = body mass index, obstruction, dyspnea, and exercise capacity; CRP = C-reactive protein; COLD = Global Initiative for Chronic Obstructive Lung Disease; IC = inspiratory capacity; MMRC = modified Medical Research Council; TLC = total lung capacity; 6MWD = Bmin waking distance
OPD, a major cause of morbidity and mortality C worldwide,’ is thought to result from an abnor-
mal pulmonary inflammatory response to noxious particles or gases.2 COPD is associated in some patients with a “low grade systemic inflammation” that may be responsible for the systemic effects of the disease; malnutrition, muscle wasting, osteoporosis, cardiovascular disease, type I1 diabetes mellitus, anemia, and depression.3 Many different inflammatory markers appear to be increased in the serum of stable COPD patient^.^ 1336
The best studied of these markers is the C-reactive protein (CRP), an acute-phase reactant secreted by the liver in response to infection, inflammation, or tissue damage.5 In patients with stable COPD, CRP
For editorial comment see page 1296 levels are directly associated with age, weight, dyspnea, and quality of life,fi.7and inversely related to exercise capacity and the degree of airway obstructi0n.7.~The higher levels of CRP observed in patients Original Research
with COPD increase even more during exacerbations? and appear to decrease in patients receiving inhaled corticosteroids.7J0 The exact role of CRP levels as a predictor of cardiovascular and global mortality in the general population has been postulated but not yet fully accepted." Two epidemiologic studies12.13 have shown that increased CRP levels are independently associated with global and cardiovascular mortality in COPD patients with mild-to-moderate degrees of airway obstruction. To explore whether CRP levels are associated with survival in patients with more severe degree of airway obstruction, we followed up a cohort of well-characterized patients with moderate to very severe COPD6v7and determined if their initial CRP level was associated with their risk of death. We also compared this association with that of other prognostic parameters that can be obtained in clinical practice, the BODE (body mass index, obstruction, dyspnea, and exercise capacity index) and its components,l4 Pao,, and inspiratory capacity (IC)/total lung capacity (TLC),15in predicting survival in the same population.
CRP Measurement
Fasting blood samples were obtained with the patient at rest, before any other test was performed. CRP concentrations were determined using high-sensitivity immunoassay (in Tenerife: CRP assay [latex] HS; Roche Diagnostics; Indianapolis, in Boston: HS-CRP; DPC Immulite; Los Angeles, CA) following manufacturer instructions. The coefficients of variation measured at each site for Roche Latex HS and for DPC Immulite were 6% and 896, respectively. Analytical sensitivity of these high-sensitivity assays is 0.1 m@, and measurement range is 0.1 to 20 mg /L. Clinical Param.eters
Both sites were part of the international BODE study and followed the exact same methodologicguidelines. Body mass index (BMI) was calculated as the weight in kilograms divided by height in metes? Patients performed spimmehy and lung volume measurements following American Thoracic Society (ATS) remmmendations and standad references.16 The 1WL.Cratio was calculated as previously describd with a value of < 0.25 used in the analysis because this threshold had the stmngest association with survival in COPD patients.15 Functional dyspnea was scored using the modified Medical Research Council (MMRC) scale.17The Bmin walking distance test (GMWD) was performed following the ATS recommendation.'* Pao2 was m e a s d in blood fmm the radial artery. The BODE index was ckssified in quartiles as previously descriM.14 The degree of comorbidities was measured using the m d e d Charhn index.19 Survival Determination
MATERIALS AND METHODS Study Population
This is a prospective observational study of patients from the two sites of the BODE cohort that had measured CRP levels at baseline.6.7 Patients from the pulmonary clinics at the Hospital Universitario Ntra Sra de Candelaria, in Tenerife, Spain, and from Caritas St Elizabeth's Medical Center in Boston, MA, signed the informed consent approved by the Human Institutional Review Boards. The patients were consecutively included ifthey had smoked B 20 pack-years and had a postbronchodilator FEV,/FVC ratio < 0.7 after 400 kg of inhaled albuterol. Patients were excluded if they had a history of asthma, bronchiectasis, tuberculosis, malignancy, arthritis, connective tissue disorders, or inflammatory bowel disease. The patients were clinically stable (no exacerbation for 2 months) at the time of the evaluation. *From the Respiratory Research Unit (Dn. de Torres, C6rdobaLank, and Aguirre-Jaime),Pulmonary (Dr. Casanova), and Biochemical Analysis (Dr. Mums de Fuentes) Departments, Hospital Nh-a Sra de Candelaria, Tenerife, Spain; Pulmonary and Critical Care Department (Drs. Pinto-Plata and Celli),Caritas St Elizabeth's Medical Center, Boston, MA; and GlaxoSmithKline (Dr. Mullerova), Worldwide Epidemiol Department, London, UK. The author have no conflicts3interest to disclose. Funding was provided partially by the Canarian Research and Health Foundation. Manuscript received October 1,2007; revision accepted January 18, 2008. Repduction of this article is rohibited without written permission from the American College o!Chest Physicians (w.chestjoumal. org/misc/re rintsshtml). C o m s p o d n c e to: Juan P. de Torres, MD, Unidud de Investigaci6n Hospital Ntra Sra de Candelaria, Ctra del Rosa& s/n, 38010 Santa Cruz de Tenerife, Canary Islands, Spain; e-mail: [email protected] DOI: 10.1378/chest.07-2433 www.chestjoumal.org
Patient survival status and cause of death were determined through medical record review and telephone interviews with family
Table l-Clinical and Physiologic Characteristics of the 218 Patients Enrolled in the Study* Characteristics
Data
Follow-up, mo Age*yr Gender Male Female Pack-yr history CRP level, mg/L BMI,kg/me FEV,, % predicted FVC,96 predicted GMWD, m IC/rLC < 0.25 MMRC scale 2 2 Pao,, mm Hg Charlson scale Cardiovascular risk factors or disease Use of inhaled corticosteroids GOLD stage Stage I Stage I1 Stage I11 Stage IV BODE index Quartiles First and semnd Third and fourth
36 (24-50) 6529 138 (64) 80 (36) 55 (40-79) 4 (1.9-9) 26 2 5 46 2 19 77 2 22 426 2 107 70 (32) 117 (53) 70? 12 2 (1-4) 73 (33) 110 (50)
157 (72) 61 (28)
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%). CHEST / 133 I 6 I JUNE, 2008
1337
members. In Boston, the status was canfirmed using the national social security registry. All-cause mortality during the follow-up period was used as the outcome measure. We also established the causes of death and classified them as respiratory, including respiratory failure and respiratory infection; cardiovascular, including sudden death, myocardial infarction, and cerebrovascular events; cancer, including all types of cancers; and others. Statistical Analysis
All continuous variables are expressed as mean (SD), median (25th to 75th percentiles), or as percentage, depending on distribution. CRP values were logarithmic transformed because of nonnomal distribution. Pearson x2 test, Mann-Whitney U test, or Student t test were used for two-group comparisons. To evaluate the strength of the association of death and the other variables, we performed a Cox regression analysis adjusting for age, gender, pack-year history, presence of cardiovascular risk factors or disease, and treatment with inhaled corticosteroids (none of the patients were receiving oral corticosteroids). Results are reported as hazard ratios with 95% confidence intervals; p I0.05 was considered statistically significant. Statistical analysis was performed using a commercial statistical package (SPSS version 13.0; SPSS; Chicago, IL).
RESULTS The clinical and physiologic charwteristics of all patients are summarized in Table 1. This predominantly male COPD population included patients with Global Initiative for Chronic Obstructive Lung Disease (GOLD)/ATS/European Respiratory Society stages I1 to IV.Most had a normal BMI, with 32% of them
having an I W L C ratio < 0.25. Half had an MMRC score > 2 and some degree of comorbidity. At the end of the follow-up time, 25% (54patients) had died and 7% (15 patients) were not available for follow-up. The characteristics of the 15 patients not available for follow-up were similar to those of the 149 patients who were alive at the end of the study. Table 2 shows the clinical and physiologic parameters in patients from Tenerife and Boston. No W e r ences were found in age, gender, BMI, comorbidity scores, Pao,, and CRP levels, but the patients from Boston had a greater smoking history, degree of airway obstruction, dyspnea, and higher BODE scores. Table 3 compares the clinical and physiologic parameters in survivors and nonsurvivors. Independent of study site, there were no differences in CRP, age, BMI, and comorbidity Charlson index. Differences were found in the intensity of smoking, degree of airflow obstruction, BODE index and its components (except the BMI), Pao,, and in ICRLC. The cause of death could be obtained only in the 23 patients from the Tenerife cohort. The living status and date of death could be retrieved in all of the Boston patients, but the actual cause of death was not available for most of them. In the available data, the most frequent cause of death was respiratory (43%),then cardiovascular diseases (26%),and finally all types of cancers (13%).
Table 2-Clinical and Physiologic Churacterietics of the 218 Patients Enrolled in the Study According to Study Site* Characteristics Follow-up, mo
4%yr Gender Male Female Pack-yr history CRP level, m e BMI, kg/m2 FEV,, % predicted WC, % predicted GMWD, m
ICRLC ratio < 0.25 MMRC scale 2 2 Pao,, mm Hg Charlson scale GOLD stage Stage I Stage I1 Stage I11 Stage N BODE index Quartiles First and second Third and fourth
Tenerife (n = 130)
Boston (n = 88)
30 (24-38) 63 2 10
57 (43-67) 65 2 8
83 (64) 47 (36) 50 (40-70) 4.1 (1.8-7.7) 27 5 5 53 2 20 83 2 23 460 2 102 29 (22) 40 (44) 71 2 11 2 (24)
55 (63) 33 (37) 65 (42-88) 3.9 (2-9.7) 27 2 5 35 2 10 67? 19 374 2 94 41 (46) 77 (88) 68 2 13 3 (27-5)
18 (14) 51 (39) 41 (32) 20 (15) 2 (1-3
0 (0 5 (6) 55 (62) 28 (32) 5 (35)
115 (89) 15 (11)
p Value 0.001 0.24 0.35
0.028 0.55 0.77 < 0.001 < 0.001 < 0.001 0.003 0.002 0.08
0.56 0.005
< 0.001
42 (48) 46 (52)
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%). 1338
Original Research
hPatfsnta Alive or Deceased at the End of the Study
Table 3-Diflerencee in the Studied Parametm
pQJJod* Parameters
Abve In = 149)
Deceased (n = 54)
Age. yr Pack-yr history CRP level, in@ BMI, k g h 2 FEV,, % predicted FVC, % predicted GMWD, m ICmLC ratio < 0.25 MMRC scale 2 2 Pao,, mm Hg Charlson scale Cardiovascular risk factors or disease Use of inhaled corticosteroids GOLD stage Stages I and I1 Stages 111 and IV BODE index Quartiles First and second Third and fourth
6428 56 (41-79) 3.8 (1.9-8.1) 27 t 5 4 8 k 19 79 2 22 446 2 97 33 (22) 76 (51) 72 2 10 2 (13) 54 (36) 80 (53)
66 2 9 71 (5p96) 4.5 (2.1-11.5) 26 2 5 35 2 14 66222 353 t 110 37 (65) 40 (75) 64 2 14 2(14 19 (35) 30 (55)
58 (38) 91 (62) 3( 1 4
5 (10) 49 (90) 5 (3-6)
118 (79) 31 (21)
D
Value
0.07 0.002 0.23 0.15 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.26 0.27 0.35 < 0.001
< 0.001 < 0.001
24 (45) 30 (55)
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%).
Table 4 shows the comparison of c h c a l and physiologic parameters from patients with initial CRP value 2 3 m@ or < 3 m@. We chose this cut-off value because it was previously proved to determine patient survival in the cardiovascular literature20.21 and in the COPD cohort reported by DahI and coworkers.13 No differences were found in all of the parameters except for a higher BMI and a lower Pao, in patients with higher CRP levels. Figure 1 shows no Werences in the Kaplan-Meier survival curves of patients with lower or higher CRP levels.
Table 5 shows the results of the Cox proportional hazard regression analysis. The analysis showed that most of the evaluated variables were associated with all-cause mortality, whereas the CRP did not.
DISCUSSION The most important finding of this study is that a baseline serum CRP level was not significantly associated with survival status in patients with clinically
Table I-Comparison of Clinical and Phyeiologic Parameters Between Patlents With Znitial CRP Levels s 3 mg/L or > 3 m a *
Pardmeters
CRP=3mg/L(n=87)
Deceased Age, yr Pack-yr history BMI, kg/m2 FEV,, % predicted FVC. % predicted GMWD, m I W L C ratio < 0.25 MMRC scale 2 2 Pao,, m m Hg Charlson scale GOLD stage Stages I and 11 Stages I11 and IV BODE index Quartiles First and second Third and fourth
20 (23) 63 t 9 50 (40-70) 2525 43 2 21 80 2 22 438 -C 103 30 (34) 47 (54) 73 2 11 2 (13)
CRP > 3 mg/L (n = 131)
34 (26) 6429 60 (40-80) 28 2 5
44 2 17 75 2 22 417 2 109
48 (36) 70 (53) 68 2 12 3 (1-4)
32 (27) 55 (63) 3 (1-5)
42 (32) 89 (68) 3 (2-5)
61 (71) 26 (29)
96 (73)
p Value 0.84 0.82 0.23 0.001 0.22 0.09 0.12 0.63 0.93 0.002 0.06 0.10
0.06
35 (27)
0.70
*Data are presented as median (25th to 75th percentiles), mean 2 SD, or No. (%). www.chestjoumal.org
CHEST / 133 / 6 / JUNE, 2008
1339
FIGURE1. Kaplan-Meier suMval curves showing the accumulative mortality of patients participating in the study depending on a baseline CRP level > 3 mg/L or < 3 mg/L. The up er anel represents the entire cohort, whereas the lower panels provide similar information at eacg o f t h e study sites; p > 0.05, log-rank comparison between both groups.
stable moderate to very severe COPD. Numerous prospective studies have shown that the serum level of CRP is independently associated with future mortality. Levels of CRP are associated with cardiovascular risk of death in the general population21as well as in patients with other conditions, such as
diabetes22 and renal disease.= Althoug+ the important body of data built around CRP levels as a cardiovascular mortality marker suggests a role in its routine determination, cardiologists have not fully accepted yet its use in the risk assessment of individual patients.11
Table 5-Regression Analysis With Global Mortality aa Main Outcome and Each Studied Variable as a Predictor This Outcome* Regression Analysis Variables CRP
BMI FEV, % predicted MRC scale 2 2 BMWD ICRLC ratio < 0.25
'Hazard Ratio
95% Confidence Interval
1.00 0.94 0.96
0.98-1.02 0.89-1.00 0.94-O.9H
2.00 0.99
1.063.70
PdO,
4.66 0.95
0.98-0.99 2.53458 0.93-0.97
BODE index
1.41
1.22-1.61
p Valuel 0.56 0.07 0.001 0.03
< 0.001 < 0.001
< 0.001 < 0.001
*All analyses adjusted for age, gender, pack-year history, cardiovascular risk factors or disease, and treatment with inhaled corticosteroids. 1340
Original Research
To our knowledge, only two previous studies have examined the possible association of CRP levels and COPD outcomes. The first, by Man et al,I2 measured CRP levels (mean, 1.32 mg/L; interquartile range, 0.62 to 3.06 m@) in individuals from the Lung Health Study with mild-to-moderate ainvay obstruction (mean FEV,, 78 2 9% of predicted). They found that patients in the lowest CRP quartile had the highest risk of all-cause mortality, cancer deaths, and cardiovascular events, and suggested that CRP could be a marker of outcome in COPD. Recently, Dahl et al,13 using data from the Copenhagen City Heart Epidemiologic Study, found that the CRP level (mean, 3.32 m@; 25th to 75th percentiles, 1.6 to 6.9) was a strong and independent predictor of COPD outcomes represented by hospitalization and death. Both studies were adjusted by other confounding factors such as age, gender, race, smoking intensity, airway obstruction, and cardiovascular state. Our study population had much smaller size compared with those epidemiologic studies, and our patients had higher CRP levels (median, 4.1 m e ; 25th to 75th percentiles, 1.9 to 9 m@) than those from the previously reported studies. This is likely due to the greater degree of airway obstruction and disease severity in patients attending clinics compared with those enrolled in epidemiologic studies. CRP levels of the patients who died tended to be higher than those who survived, but the difference failed to reach statistical significance. When we categorized patients by the initial CRP level of 1 3 in@ or < 3 mg/L, we again found no differences in clinical or physiologic characteristics except for a higher BMI and lower Pao, in those patients with CRP levels 2 3 m@. This information could have different interpretations in our population of patients with moderate to very severe disease: first, that this threshold value is only valid to evaluate cardiovascular mortality in mild-to-moderate disease; second, that a merent threshold value (probably higher considering most COPD studies median CRP values) should be chosen to evaluate the importance of this biomarker in patients with severe to very severe disease; and third, that the level of CRP is not associated with increased risk of mortality in this particular degree of COPD severity. Indeed, in our cohort, CRP levels were not associated with survival, whereas the BODE index, a multidimensional prognostic tool,'* and most of its components were significant predictors of outcome. That CRP levels in patients iyith wellcharacterized COPD may not provide independent information about outcome in patients with clinically diagnosed COPD is not unique to our study. The same observation has also been found in the Framingham study,% in which elevated CRP levels did not www.chestjoumal.org
'
provide furthers prognostic information beyond traditional office risk factor assessment. Our study has three important differences compared with those of Man et all2 and Dahl and coworkers.13 They relate primarily to the populations included and the measured variables. With regard to study size, the previous studies were epidemiologic in design with large number of patients, whereas ours is a clinical study with 218 well-characterized COPD patients. In relation to the populations recruited, The Lung Health Study12 included patients with mild COPD likely with little or no systemic consequences of the disease, and in whom the main cause of mortality was cardiovascular disease, a finding that has been discussed by others.= Thus, it is possible that the elevated CRP was probably expressing the cardiovascular mortality risk. The third difference between our cohort and those reported previously reside in the variables measured. In the epidemiologic cohorts, spirometry was the only clinical variable used to evaluate COPD severity, and the outcome was determined using the ICD codes with all of its shortcomings. Our cohort was smaller, however, and had a higher percentage of deaths (25% vs 6%) over a shorter period of time, included patients with COPD as seen by clinicians, and the patients were more comprehensively evaluated using tools that have been shown to be associated with mortality better than the FEV,. It is possible that the level of CRP may be useful in epidemiologic studies but that it loses some of its association with survival once the patient has more clinically evident COPD. We believe that our findings should raise a word of caution to the excitement that this inflammatory marker has generated in the COPD field. CRP is an acute-phase reactant that increases in a very sensitive but nonspecific way in most forms of tissue damage, inflammation, and infection,5 all very dynamic processes in patients with COPDq9Indeed, it seems that there are genetic influences in the way patients with COPD react to inflarnmation,Z6 a factor that could have an important role in their serum CRP levels.27 Further CRP levels are influenced by many different factors such as cardiovascular disease, degree of physical activity, diabetes, renal disease, hypertension, metabolic syndrome, obstructive sleep apnea, smoking status, diet, and medications (corticosteroids, statins, angiotensin-converting enzyme inhibitors), to name a few. To our knowledge, only one study7 has determined the stability of CRP over time in a single cohort of a rather small number of patients. As suggested by Donaldson,28longitudinal studies are needed to confirm the stability of the CRP signal and its relationship with other clinical factors. Until then, the measurement of CRP level at CHEST f 133 f 6 f JUNE, 2008
1341
one single point may not be useful. Probably, serial measurements indicating its trend in relationship to interventions (long-term oxygen therapy, rehabilitation, inhaled corticosteroid, or statin therapy) could be one of the possible ways to evaluate the use of this inflammatory marker in COPD. Our study had several limitations. First, the relatively small population compared with the numbers reported in epidemiologic studies may be underpowered to adequately study the association between survival and CRP levels as and end point. Second, the findings are restricted to the population included in the study: patients from pulmonary clinics with moderate to very severe COPD. Third, the exact proportion of the causes of death may not be accurate because it was obtained from review of the medical records and the exact causes of death in many patients from the Boston cohort were not determined. This is a known limitation of many COPD survival studies.25 However, we used allcause mortality as the primary outcome, and survival status was known for 203 of the 218 patients. Fourth, even though we selected patients with no known illness that could cause increased CRP, many of the known factors likely to influence CRP could have occurred over time. However, this occurrence would have made our results even more relevant because they would have favored the presence of an association. An important additional strength of our study is that we observed the same findings in both study sites. This suggests that the lack of relationship between CRP and survival over time is consistent and reproducible in clinically and geographically different populations. In summary, we have found that in this population of patients with moderate to very severe COPD from a pulmonary clinic, CRP levels are not associated with survival status. CRP does not add information to the risk assessment provided by other tools such as the multidimensional BODE index and its components or the IC/TLC ratio. Other long-term studies in different settings are needed to evaluate the role of CRP and other markers2Qin patients with clinical COPD. REFERENCES 1 Pauwels RA, Rabe KF. Burden and clinical features of chronic obstructive pulmonary disease (COPD). Lancet 2004; 364613-620 2 Pauwels R, Buist SA, Calverley P, et al. Global strategy for the diagnosis, management and prevention of Chronic obstructive pulmonary disease: NHLBVWHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) workshop summary. Am J Respir Crit Care Med 2001;163:1256-1276 3 Agusti A. Systemic effects of chronic obstructive pulmoniuy disease. Proc Am Thorac SOC2005; 2:367-370 1342
10
11 12 13 14
15 16 17 18 19 20 21
22
23 24
Can WQ, Man SFP, Senthilselvan A, et al. Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and meta-analysis. Thorax 2004; 59:574-580 Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest 2003; 111:1805-1812 de Torres JP, Cordoba-Lanus E, Lopez-Aguilar C, et al. C-reactive protein levels and clinically important predictive outcomes in stable COPD patients. Eur Respir J 2006; 27:902-907 Pinto-Plata VM, Mullerova H, Toso JF, et al. C-reactive protein in patients with COPD, control smokers, and nonsmokers. Thorax 2006; 61:23-28 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 Hurst JR, Donaldson GC, Perera WR, et al. Use of plasma hiomarkers at exacerbation of chronic obstructive pulmonary disease. Am J Respir Crit Care Med; 2006; 1742367-874 Sin DD, La& P,-York E, et al. Effects of fluticasone on systemic markers of inflammation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2004; 170: 760-765 Lowe GD, Pepys MB. C-reactive protein and cardiovascular disease: weighing the evidence. Curr Atheroscler Rep 2006; 8~421-428 Man SFP, Connett JE, Anthonisen NR, et al. C reactive protein and mortality in mild to moderate chronic obstructive pulmonary disease. Thorax 2006; 61849-853 Dahl M, Vestbo J, Lange P, et al. C-reactive protein as a predictor of prognosis in COPD. Am J Respir Crit Care Med 2007; 175250-255 C e k BR, Cote CG, Marin JM, et al. The body mass index, airflow obstruction, dyspnea and exercise capacity index in chronic obstructive pulmonary disease. N Engl J Med 2004; 350:1005-1012 Casanova C, Cote C, de Torres JP, et al. The inspiratory to total lung capacity ratio predicts mortality in patients with COPD. Am J Respir Crit Care Med 2005; 171:591577 Quanjer PH. Standardized lung function testing: report of the Working P q for the European Community for Steel and Coal. Bull Eur Pbysiopath Respir 1983; 19(suppl5):22-27 Brooks SM. Survehlkce for respiratory hazards. ATS News 1982; 8:12-16 ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med 2002; 166:lll-117 Charlson M, Szatrowsky T, Peterson J, et al. Validation of a combined comorbidity index. J Clin Epidemiol 1994; 47: 1245-1251 Myers GL, Rifai N, Tracy RP, et al. CDC/AHA Workshop on Markers of Inflammation and Cardiovascular Disease. Circulation 2004; 11Oe545-e549 Ridker P. Clinical application of C-reactive protein for cardiovascular disease detection and prevention. Circulation 2003; 107:363-369 Linnemann B, Voigt W, Novel W, et al. C-reactive protein is a strong independent predictor of death in type 2 diabetes: association with multiple facets of the metabolic syndrome. Exp Clin Endocrinol Diabet 2006: 114:127-134 Racki S, Zaputovic L, Mavric Z, et al. C-reactive protein is a strong predictor of mortality in hemodialysis patients. Ren Fail 2006; 427-433 Wilson PW, Nam BH, Pencina M, et al. C-reactive protein and risk of cardiovascular disease in men and women from the Framingham Heart Study. Arch Intern Med 2005; 165:24732478 Original Research
25 Hansel1 A, Walk JA, Soriano JB. What do chronic obstruotive pulmonary disease patients die from? A multiple cause coding analysis. Eur Respir J 2003; 22:809-814 26 Yanbaeva DG, Dentener MA, Creutzberg EC, et al. Systemic inflammation in COPD: is genetic susceptibilitya key factor? COPD 2006; 3:51-61 27 Hersh CP, Miller DT, Kwiatkowski DJ, et al. Genetic deter-
w.chestjoumaI.org
minants of C-reactive protein in COPD. Eur Respir J 2006; 28:1156-1162 28 Donaldson G. C-reactive protein: does it predict mortality? Am J Respir Crit Care Med 2006; 175:209-210 29 Jones PW, Agusti AGN. Outcomes and markers in the assessment of chronic obstructive pulmonary disease. Eur Respir J 2006; 27:822-832
CHEST I 133 I 6 I JUNE, 2008
1343