You are on page 1of 8

Journal of Critical Care (2011) 26, 287294

Albumin and C-reactive protein have prognostic significance in patients with community-acquired pneumonia
Jae Hyuk Lee MD a , Jooyeong Kim MD b,1 , Kyuseok Kim MD, PhD a,, You Hwan Jo MD, PhD a , JoongEui Rhee MD, PhD a , Tae Youn Kim MD a , Sang Hoon Na MD b , Seung Sik Hwang MD, PhD c
a

Department of Emergency Medicine, Seoul National University Bundang Hospital, Gyeonggi-do 463-707, South Korea Department of Emergency Medicine, Seoul National University Hospital, Seoul, South Korea c Department of Social and Preventive Medicine, Inha University College of Medicine, Incheon, South Korea
b

Keywords:
Community-acquired pneumonia; Prognosis; Albumin; C-reactive protein

Abstract Purpose: This study aims to determine the association of commonly used biochemical markers, such as albumin and C-reactive protein (CRP), with mortality and the prognostic performance of these markers combined with the pneumonia severity index (PSI) for mortality and adverse outcomes in patients with community-acquired pneumonia (CAP). Materials and Methods: The data were gathered prospectively for patients hospitalized with CAP via the emergency department. Laboratory values, including CRP and albumin, clinical variables, and the PSI were measured. Primary outcomes were 28-day mortality and survival times. Secondary outcome was admission to the intensive care unit, vasopressor use, or the need for mechanical ventilation during the hospital stay. Results: A total of 424 patients were included. The 28-day mortality was 13.7%. C-reactive protein and albumin were significantly different between survivors and nonsurvivors. In logistic regression analysis, CRP and albumin were independently associated with 28-day mortality (P b .05). Receiver operating characteristic curves showed improved mortality prediction by adding CRP or albumin to the PSI scale. The Cox proportional hazards analysis showed that high serum albumin (3.3 mg/dL) had a hazard ratio of 0.5 (95% confidence interval, 0.3-0.9), and high CRP (14.3 mg/dL) had a hazard ratio of 2.0 (95% confidence interval, 1.1-3.4). For predicting secondary outcome, adding albumin to PSI increased areas under the curve significantly, but CRP did not. Conclusion: Albumin and CRP were associated with 28-day mortality in hospitalized patients with CAP, and these markers increased prognostic performance when combined with the PSI scale. Crown Copyright 2011 Published by Elsevier Inc. All rights reserved.

Corresponding author. Tel.: +82 31 787 7579; fax: +82 31 787 4055. E-mail address: dremkks@snubh.org (K. Kim). 1 Equally contributed as a first author. 0883-9441/$ see front matter. Crown Copyright 2011 Published by Elsevier Inc. All rights reserved. doi:10.1016/j.jcrc.2010.10.007

288

J.H. Lee et al.

1. Introduction
Community-acquired pneumonia (CAP) is an important cause of morbidity and mortality throughout the world [1]. Some severity assessment tools such as the pneumonia severity index (PSI), CURB65 (confusion, urea nitrogen, respiratory rate, blood pressure, 65 years of age), or CRB65 (confusion, respiratory rate, blood pressure, 65 years of age) have been developed and validated [2,3]. However, although these scales have been useful in identifying low-risk patients who can be discharged from the emergency department (ED), they have been appraised less accurate in assessing the prognosis for mortality, the need for intensive care unit (ICU) care, and the need for mechanical ventilation in hospitalized patients with CAP [4,5]. Many biochemical markers are associated with the outcome of CAP, such as the cytokines like tumor necrosis factor and interleukin 6, C-reactive protein (CRP), procalcitonin, and D-dimer [6-9]. Of those, CRP, which is a commonly tested laboratory value in hospitalized patients, is an acute phase reactant considered as a potentially important prognostic variable [10,11]. Serum albumin, which is also a commonly tested laboratory value in hospitalized patients, has been proposed as a reliable predictor of outcome in critically ill patients with infectious diseases [12,13]. Low serum albumin level also has been associated with morbidity and mortality in various diseases [14-18]. The precise mechanism of the protective effects of albumin is not known. However, in animal experiments, human serum albumin has been shown to have some protective effects, such as improvement of arterial hyporeactivity in endotoxemia as an antioxidant, reduced ischemia-reperfusion injury, and antiinflammatory effects [18-21]. The objectives of this study are to identify the association of biochemical markers such as CRP and albumin with mortality and to determine the prognostic significance of these markers combined with PSI for mortality and major adverse outcomes, such as the need for ICU admission, mechanical ventilation, or vasopressor use in hospitalized patients with CAP. We hypothesized that albumin and CRP would be associated with 28-day mortality and improve mortality prediction in hospitalized patients with CAP.

2.2. Study design and population


We prospectively collected data of patients who visited our ED and were subsequently hospitalized for CAP between April 2008 and March 2010. Community-acquired pneumonia was defined as a pulmonary infiltrate on the chest radiograph and symptoms consistent with pneumonia, including cough, dyspnea, fever, and/or pleuritic chest pain, which were not acquired in a hospital or a nursing home. If pulmonary infiltrate was not seen in the initial chest radiograph, abnormal lung sounds on the initial physical examination and pulmonary infiltrate on a followup chest radiograph were considered as an equivalent. Eligible patients were older than 18 years and had a diagnosis of CAP. Of those, patients who were hospitalized were included. In our ED, hospital admission is decided by the attending physician of the ED based on the PSI scale and other medical conditions [22]. We excluded patients based on the following criteria: younger than 18 years, transferred from another hospital, discharged from a hospital within the previous 10 days, an episode of pneumonia within the past 30 days, active pulmonary tuberculosis, known to be positive for HIV, or chronically immunosuppressed (defined as immunosuppression for solid organ transplantation, postsplenectomy, receiving 10 mg/d prednisolone or equivalent for more than 30 days, treatment with other immunosuppressive agents or neutropenia [b1.0 109/L neutrophils]).

2.3. Data collection


The data for this study were part of a prospective quality improvement study to implement PSI as an admission decision tool in our institution [22]. Baseline clinical information obtained by structured patient or proxy interviews, bedside assessment, and structured medical reviews were gathered. Collected clinical information included age and sex as well as comorbid diseases such as chronic obstructive pulmonary disease; cardiac, liver, renal, and malignant diseases; diabetes mellitus; and central nervous system diseases. Recorded bedside assessment included temperature, respiratory and heart rates, and blood pressure. The following initial laboratory data also were recorded: leukocyte count, hemoglobin level, platelet count, blood glucose, serum creatinine, blood urea nitrogen, albumin, total cholesterol, prothrombin time (international normalized ratio), activated partial thromboplastin time, and CRP. Patients were scored by PSI using the prediction rule of Fine et al [2] and were classified accordingly into one of 4 risk groups: groups I-II, less than 70 points; group III, 70 to 90 points; group IV, 91 to 130 points; and group V, more than 131 points. The primary end point was 28-day mortality after the ED visit. The survival time was also investigated. The secondary end point was vasopressor use, the need for

2. Materials and methods


2.1. Study setting
A prospective observational study was performed in a 950-bed tertiary academic hospital with an annual ED census of 67 000. The institutional review board of our institute approved the study.

Prognostic significance of albumin and C-reactive protein mechanical ventilation, or admission to an ICU during the hospital stay.

289 hazard effect of biochemical variables on survival, we conducted Cox proportional hazards regression analysis with identified variables and adjusted by the PSI scale. Statistical analysis was done using STATA IC/10.1 (Stata Corp. LP, TX, USA).

2.4. Statistical analysis


Categorical variables were analyzed using the 2 test or Fisher exact test, and continuous variables were analyzed using Student t test or the Mann-Whitney U test. To identify independent risk factors for 28-day mortality among the variables that showed statistically significant associations in univariate analysis (P b .05), multivariate logistic regression analysis was performed. To evaluate the prognostic value of independent risk factors and PSI class, we repeated the multivariate logistic regression analysis using identified risk factors, and the area under the receiver operating characteristic (ROC) curve (AUC) was calculated. Statistical comparison of the AUC between PSI alone and PSI combined with each independent risk factor was conducted using the method of Hanley and McNeil [23]. To identify the
Table 1 Variables Baseline characteristics and initial severity of patients Total patients n = 424 Age (y) Male sex (%) Comorbidities, n (%) Heart failure Renal failure Liver disease COPD Neoplasm Neurologic disease Diabetes mellitus Initial vital signs Mean arterial pressure (mm Hg) Respiratory rate (/min) Body temperature (C) Laboratory findings White blood cell count (103/mm3) Hemoglobin level (g/dL) Platelet count (103/mm3) Glucose (mg/dL) Albumin (g/dL) Cholesterol (g/dL) Blood urea nitrogen (mg/dL) Creatinine (mg/dL) Prothrombin time (INR) Activated partial thromboplastin time (s) CRP (mg/dL) PSI, n (%) I, II III IV V 70.4 15.6 148 (35.0) 12 (2.8) 44 (10.4) 26 (6.1) 84 (19.8) 104 (24.5) 107 (25.2) 124 (29.3) 109.6 21.9 23.4 5.8 37.4 1.0 12.8 7.1 12.1 2.3 281.1 139.3 169.3 129.9 3.4 0.6 145.4 41.6 25.6 18.4 1.4 1.1 1.2 0.3 42.9 11.6 13.6 9.8 72 (17.0) 73 (17.2) 176 (41.5) 103 (24.3)

3. Results
During the study period, 424 patients were hospitalized via the ED for CAP. The inclusion criterion was hospitalized patients for CAP via ED, and there were no missing data during the study period. Thus, all patients were included for analysis. Mean (SD) age was 70.4 (15.6) years, and 148 patients (35%) were male. The 28-day mortality was 13.7% (58 patients). The demographic characteristics and initial severity are shown in Table 1. Patients who died were older; had a lower rate of chronic obstructive pulmonary diseases and a higher rate of neoplastic diseases; had lower mean arterial pressure and more rapid respiratory rate; and had

Nonsurvivor n = 58 (13.7%) 78.0 12.3 19 (32.8) 1 (1.7) 3 (5.2) 3 (5.2) 5 (8.6) 25 (43.1) 18 (31.0) 13 (22.4) 101.0 28.8 26.8 8.2 37.2 1.0 14.1 8.6 11.3 2.6 280.0 154.9 163.0 118.1 3.0 0.6 136.8 56.2 34.9 23.0 1.7 0.9 1.3 0.3 43.9 14.4 18.0 10.3 0 1 (1.7) 25 (43.1) 32 (55.2)

Survivor n = 366 (86.3%) 69.2 15.8 129 (35.3) 11 (3.0) 41 (11.2) 23 (6.3) 79 (21.6) 79 (21.6) 89 (24.3) 111 (30.3) 110.9 20.3 22.9 5.1 37.4 1.0 12.5 6.8 12.3 2.2 281.3 137.0 209.7 185.1 3.5 0.6 146.8 38.8 24.1 17.2 1.4 1.2 1.2 0.3 42.7 11.0 13.0 9.5 72 (19.7) 72 (19.7) 151 (41.3) 71 (19.4)

P b.001 .712 .585 .162 .743 .021 b.001 .274 .218 .001 b.001 .204 .119 .003 .949 .012 b.001 .094 b.001 .093 b.001 .458 b.001 b.001

Data are expressed as mean SD or number (%). COPD indicates chronic obstructive pulmonary disease; INR, international normalized ratio.

290
Table 2 Variables Mean arterial pressure Hemoglobin level Albumin PT, INR CRP PSI class I-III PSI class IV PSI class V
OR indicates odds ratio.

J.H. Lee et al.


Multivariate logistic regression analysis OR (95% CI) 1.00 (0.98-1.01) 1.01 (0.87-1.17) 0.37 (0.19-0.73) 1.02 (0.39-2.70) 1.04 (1.00-1.07) Reference 16.96 (2.24-128.42) 36.41 (4.71-281.39) P .506 .889 .004 .962 .025 .006 .001

To identify the effect of initial serum albumin and CRP levels on survival time, Cox proportional hazards regression analysis was performed. As shown in Table 4, the group with albumin 3.3 mg/dL or more had a hazard ratio of 0.5 (95% CI, 0.3-0.9; P = .03), and the group with CRP 14.3 mg/dL or more had a hazard ratio of 2.0 (95% CI, 1.1-3.4; P = .02). Fig. 2 shows the survival curves according to serum albumin and CRP levels. Patients with low serum albumin or high CRP had higher mortality rates (P b .05 for both).

3.2. Secondary outcome


To identify additive effects of albumin and CRP with PSI in predicting ICU admission, vasopressor use, or the need for mechanical ventilation, ROC curve analysis was done, and AUCs were calculated. For predicting secondary end points, albumin had an additive role with PSI, but CRP did not. However, the combination of albumin, CRP, and PSI increased AUC significantly compared with PSI alone (Fig. 3).

lower hemoglobin values, lower blood glucose, lower albumin, higher blood urea nitrogen, more prolonged prothrombin time, higher CRP values, and higher initial severity scores (P b .05).

3.1. Primary outcome


To identify the factors associated with 28-day mortality in patients with CAP, we performed multivariate logistic regression analysis with significantly different variables between survivors and nonsurvivors. In multivariate logistic regression analysis after adjustment for PSI class, albumin and initial CRP were identified as statistically significant variables (Table 2). The cutoff values for albumin and CRP, using the highest values of sensitivity and specificity after constructing an ROC curve, were 3.3 mg/dL and 14.3 mg/ dL, respectively. In patients who were categorized into the same PSI class, especially PSI classes IV and V, mortality was higher in patients who had low serum albumin (b3.3 mg/ dL) or high CRP (14.3 mg/dL) than in patients who had high serum albumin (3.3 mg/dL) or low CRP (b14.3 mg/ dL). In patients who had albumin less than 3.3 mg/dL, mortality was significantly higher than in patients with albumin 3.3 mg/dL or more (22.1% vs 6.8%, P b .05). About CRP value, mortality was higher in patients with CRP 14.3 mg/dL or more than in patients with CRP less than 14.3 mg/ dL (20.2% vs 9.2%, P b .05) (Table 3). The AUCs to predict 28-day mortality was 0.66 (95% confidence interval [CI], 0.60-0.72) for albumin, 0.61 (95% CI, 0.55-0.68) for CRP, and 0.76 (95% CI, 0.71-0.81) for PSI. The AUCs significantly increased when albumin or CRP was added to the PSI (Fig. 1).

4. Discussion
In hospitalized patients with CAP, the initial value of serum albumin and CRP was strongly and independently associated with 28-day mortality, and the addition of serum albumin or CRP concentrations to PSI improved mortality prediction compared with the PSI scale alone. Low serum albumin concentration and high CRP value were also associated with reduced survival of patients with CAP. For predicting ICU admission, vasopressor use, or the need for mechanical ventilation, only albumin had an additive role when combined with PSI. The PSI was derived from hospitalized patients with CAP to identify low-risk patients who could be safely managed in outpatient departments. In addition, the PSI can stratify the risk of short-term mortality in patients with CAP. In EDs, patients with CAP can be correctly stratified using PSI score, and this reduces admission rates and medical costs significantly [24-27]. Accurate prognostication is especially important in critically ill patients, and many biochemical markers could reflect the severity of diseases. However, currently available severity assessment tools have relatively

Table 3 PSI class I, II III IV V Total

Mortality according to initial serum albumin and CRP concentration and PSI class Albumin b3.3 mg/dL 0/15 0/23 16/88 26/64 42/190 (0.0%) (0.0%) (18.2%) (40.6%) (22.1%) Albumin 3.3 mg/dL 0/57 1/50 9/88 6/39 16/234 (0.0%) (2.0%) (10.2%) (15.4%) (6.8%) P .495 .131 .007 b.001 CRP b14.3 mg/dL 0/49 1/43 9/100 13/59 23/251 (0.0%) (2.3%) (9.0%) (22.0%) (9.2%) CRP 14.3 mg/dL 0/23 (0.0%) 0/30 (0.0%) 16/76 (21.1%) 19/44 (43.2%) 35/173 (20.2%) P .400 .023 .022 .001

Data were presented as numbers of death/number of patients (% of death).

Prognostic significance of albumin and C-reactive protein

291

Fig. 1 Receiver operating characteristic curves for 28-day mortality. Albumin or CRP increased the AUC significantly when combined with PSI.

low sensitivity and positive predictive value in severe CAP [28]. Thus, these tools need to be modified to predict outcomes more accurately in patients with severe CAP. Recently, some biochemical markers, such as procalcitonin, D-dimer, B-type natriuretic peptide, or N-terminal pro-B-type natriuretic peptide, have been reported to have prognostic power in patients with CAP; incorporating these markers into severity scales could improve mortality prediction [6-8,2931]. Serum albumin and CRP concentrations are commonly measured as an initial evaluation of critically ill patients with infectious diseases, and these values had prognostic significance in patients with CAP in our study. C-reactive protein is an acute phase reactant associated with the severity of inflammation. There has been some controversy about the prognostic significance of CRP in CAP. Some studies reported that CRP was associated with
Table 4 Variables PSI class I-III PSI class IV PSI class V Albumin b3.3 mg/dL Albumin 3.3 mg/dL CRP b14.3 mg/dL CRP 14.3 mg/dL
HR indicates hazard ratio.

Cox proportional hazard regression analysis HR (95% CI) Reference 17.2 (2.3-127.8) 43.1 (5.8-318.7) Reference 0.5 (0.3-0.9) Reference 2.0 (1.1-3.4) P .005 .000 .027 .015

mortality in hospitalized patients with CAP [6,10] and that adding CRP to PSI scales improved mortality prediction [6]. However, other studies found no association of admission CRP with mortality in patients with CAP [32,33]. These controversies might have originated from discrepancies in the baseline characteristics of the included patients. The study reporting that CRP had a prognostic significance was conducted in hospitalized patients. However, other studies showing that CRP was not significantly associated with prognosis included all patients with CAP, hospitalized or not, or hospitalized elderly patients. Thus, the baseline characteristics of those patients might have been different. This inference is supported by our finding that the mortality rate was not different in PSI I-III, regardless of CRP level (Table 3). In our study, only hospitalized patients who were older than 18 years were included for analysis, and CRP had an additive effect in predicting 28-day mortality when combined with PSI. Thus, CRP level might be useful as a prognostic factor, at least in hospitalized patients with CAP. However, CRP did not have a role in predicting secondary outcome (ICU admission, vasopressor use, or the need for mechanical ventilation). This finding is similar to that of Charles et al [33]. Their study showed that albumin was an independent risk factor to predict intensive respiratory and vasopressor support, but CRP was not. Initial serum albumin level was independently associated with short-term mortality and had an additive effect in predicting 28-day mortality, ICU admission, vasopressor

292

J.H. Lee et al.

Fig. 2 Kaplan-Meier survival curves according to serum albumin and CRP levels. Kaplan-Meier survival estimates according to serum albumin level (A) and initial CRP level (B).

use, and the need for mechanical ventilation. This finding is similar to that of a study that identified the relationship between serum albumin concentration and mortality in patients with CAP [3,33]. Artero et al [13] also reported that serum albumin concentration was an independent risk factor for mortality in patients with community-acquired severe sepsis or septic shock. Community-acquired pneumonia might be one cause of severe sepsis or septic shock. Thus, serum albumin concentration might be correlated with the outcomes of infectious diseases. The initial value of serum albumin might be a result of malnutrition, underlying

disease, or infectious processes. Thus, these underlying conditions might be the direct cause of mortality. The precise mechanism, however, needs to be determined. In this study, the mortality prediction by PSI was improved by adding initial serum albumin or CRP values. Pneumonia severity index consists of age, comorbidity, clinical features of the patients, blood parameters, and radiological findings. However, even with identical PSI scores, patients' ability to overcome infectious processes could be different. For example, although 2 patients have the same malignant disease, their stage might not be same, and

Prognostic significance of albumin and C-reactive protein

293

Fig. 3 Receiver operating characteristic curves for secondary outcome. Secondary outcome was admission to the ICU, vasopressor use, or the need for mechanical ventilation. Albumin increased the AUCs significantly when combined with the PSI, but CRP did not. However, the combination of albumin, CRP, and PSI increased these AUCs compared with PSI alone to predict secondary outcome.

serum albumin could partly cover this difference in PSI by augmenting their nutritional status, and this could explain the additive effects of albumin with PSI in predicting mortality in CAP. The additive effect of CRP could be inferred from the fact that the severity of inflammation is an important factor in prognosis as is the patient's underlying condition. In this study, we compared the performance of CURB65 by adding albumin or CRP to predict mortality. However, the additive effects of albumin and CRP to predict mortality were not observed in CURB65 (data not shown). C-reactive protein and albumin might not be routinely checked in all patients with CAP in other institutions, and some physicians may argue that it is not cost-effective. This viewpoint is supported by our results indicating no role for these markers in PSI classes I to III. On the other hand, these markers had a significant role in severe CAP (Tables 3 and 4). Therefore, CRP and albumin should be on the laboratory panel in patients with CAP requiring admission. This study has some limitations. First, it was conducted in a single institution and only included patients who were hospitalized via the ED. Thus, it cannot be extrapolated to patients who were discharged from the ED, nor can it be generalized to all patients with CAP. However, it may be that most patients with severe CAP who require hospitalization are admitted via the ED. Second, in this study, we did not include other markers that have been associated with outcomes of infectious diseases, such as lactate, D-dimer, and procalcitonin. These markers may be associated with

outcomes, and adding these markers might improve mortality prediction. In this study, we concluded that initial serum albumin and CRP levels were associated with 28-day mortality in hospitalized patients with CAP and that these 2 values have prognostic significance when combined with the PSI. Moreover, albumin, but not CRP, had an additive effect in predicting ICU admission, vasopressor use, or the need for mechanical ventilation.

References
[1] Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis 2007;44(Suppl 2):S27-S72. [2] Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify lowrisk patients with community-acquired pneumonia. N Engl J Med 1997;336:243-50. [3] Lim WS, van der Eerden MM, Laing R, et al. Defining community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax 2003;58:377-82. [4] Angus DC, Marrie TJ, Obrosky DS, et al. Severe community-acquired pneumonia: use of intensive care services and evaluation of American and British Thoracic Society Diagnostic criteria. Am J Respir Crit Care Med 2002;166:717-23. [5] Ewig S, de Roux A, Bauer T, et al. Validation of predictive rules and indices of severity for community acquired pneumonia. Thorax 2004;59:421-7.

294
[6] Menendez R, Martinez R, Reyes S, et al. Biomarkers improve mortality prediction by prognostic scales in community-acquired pneumonia. Thorax 2009;64:587-91. [7] Chalmers JD, Singanayagam A, Scally C, et al. Admission D-dimer can identify low-risk patients with community-acquired pneumonia. Ann Emerg Med 2009;53:633-8. [8] Huang DT, Weissfeld LA, Kellum JA, et al. Risk prediction with procalcitonin and clinical rules in community-acquired pneumonia. Ann Emerg Med 52 2008;e42:48-58. [9] Mira JP, Max A, Burgel PR. The role of biomarkers in communityacquired pneumonia: predicting mortality and response to adjunctive therapy. Crit Care 12 Suppl 2008;6:S5. [10] Chalmers JD, Singanayagam A, Hill AT. C-reactive protein is an independent predictor of severity in community-acquired pneumonia. Am J Med 2008;121:219-25. [11] Castell JV, Gomez-Lechon MJ, David M, et al. Acute-phase response of human hepatocytes: regulation of acute-phase protein synthesis by interleukin-6. Hepatology 1990;12:1179-86. [12] Dominguez de Villota E, Mosquera JM, Rubio JJ, et al. Association of a low serum albumin with infection and increased mortality in critically ill patients. Intensive Care Med 1980;7:19-22. [13] Artero A, Zaragoza R, Camarena JJ, et al. Prognostic factors of mortality in patients with community-acquired bloodstream infection with severe sepsis and septic shock. J Crit Care 2010. [14] Hedlund JU, Hansson LO, Ortqvist AB. Hypoalbuminemia in hospitalized patients with community-acquired pneumonia. Arch Intern Med 1995;155:1438-42. [15] Kaysar M, Augustine T, Jim L, et al. Predictors of length of stay between the young and aged in hospitalized community-acquired pneumonia patients. Geriatr Gerontol Int 2008;8:227-33. [16] Apolinario Hidalgo R, Suarez Cabrera M, Geijo Martinez MP, et al. [Prognostic factors in community acquired pneumonia. Prospective multicenter study in internal medical departments]. An Med Interna 2007;24:467-72. [17] Yap FH, Joynt GM, Buckley TA, et al. Association of serum albumin concentration and mortality risk in critically ill patients. Anaesth Intensive Care 2002;30:202-7. [18] de Mutsert R, Grootendorst DC, Indemans F, et al. Association between serum albumin and mortality in dialysis patients is partly explained by inflammation, and not by malnutrition. J Ren Nutr 2009;19:127-35. [19] Meziani F, Kremer H, Tesse A, et al. Human serum albumin improves arterial dysfunction during early resuscitation in mouse endotoxic model via reduced oxidative and nitrosative stresses. Am J Pathol 2007;171:1753-61.

J.H. Lee et al.


[20] Yao X, Miao W, Li M, et al. Protective effect of albumin on VEGF and brain edema in acute ischemia in rats. Neurosci Lett 2010;472: 179-83. [21] Lang Jr JD, Figueroa M, Chumley P, et al. Albumin and hydroxyethyl starch modulate oxidative inflammatory injury to vascular endothelium. Anesthesiology 2004;100:51-8. [22] Jo S, Kim K, Jung K, et al.: The effects of incorporating a pneumonia severity index in the admission protocol for community-acquired pneumonia. J Emerg Med In Press, 2010 [23] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839-43. [24] Jo S, Kim K, Jung K, et al.: The Effects of Incorporating a Pneumonia Severity Index in the Admission Protocol for Community-Acquired Pneumonia. J Emerg Med, In Press, 2010. [25] Carratala J, Fernandez-Sabe N, Ortega L, et al. Outpatient care compared with hospitalization for community-acquired pneumonia: a randomized trial in low-risk patients. Ann Intern Med 2005;142: 165-72. [26] Marrie TJ, Lau CY, Wheeler SL, et al. A controlled trial of a critical pathway for treatment of community-acquired pneumonia. CAPITAL Study Investigators. Community-Acquired Pneumonia Intervention Trial Assessing Levofloxacin. JAMA 2000;283:749-55. [27] Marras TK, Gutierrez C, Chan CK. Applying a prediction rule to identify low-risk patients with community-acquired pneumonia. Chest 2000;118:1339-43. [28] Buising KL, Thursky KA, Black JF, et al. A prospective comparison of severity scores for identifying patients with severe community acquired pneumonia: reconsidering what is meant by severe pneumonia. Thorax 2006;61:419-24. [29] Jeong KY, Kim K, Kim TY, et al.: Prognostic value of N-terminal probrain natriuretic peptide in hospitalised patients with communityacquired pneumonia. Emerg Med J, In Press [30] Mueller C, Laule-Kilian K, Scholer A, et al. B-type natriuretic peptide for risk stratification in community-acquired pneumonia. J Intern Med 2005;258:391-3. [31] Christ-Crain M, Breidthardt T, Stolz D, et al. Use of B-type natriuretic peptide in the risk stratification of community-acquired pneumonia. J Intern Med 2008;264:166-76. [32] Thiem U, Niklaus D, Sehlhoff B, et al. C-reactive protein, severity of pneumonia and mortality in elderly, hospitalised patients with community-acquired pneumonia. Age Ageing 2009;38:693-7. [33] Charles PG, Wolfe R, Whitby M, et al. SMART-COP: a tool for predicting the need for intensive respiratory or vasopressor support in community-acquired pneumonia. Clin Infect Dis 2008;47:375-84.