Article, Hematology

Red cell distribution width as a prognostic marker in patients with community-acquired pneumonia

Unlabelled imagered cell distribution width as a prognos”>American Journal of Emergency Medicine (2013) 31, 72-79

Original Contribution

Red cell distribution width as a prognostic marker in patients with community-acquired pneumonia?

Jae Hyuk Lee MD, PhD, Hea Jin Chung MD, Kyuseok Kim MD, PhD?, You Hwan Jo MD, PhD, Joong Eui Rhee MD, PhD, Yu Jin Kim MD, PhD, Kyeong Won Kang MD

Department of Emergency Medicine, Seoul National University Bundang Hospital, Gyeonggi-do, Korea

Received 2 February 2012; revised 30 May 2012; accepted 1 June 2012

Abstract

Background: Red Cell Distribution Width is associated with mortality in both the general population and in patients with certain diseases. However, the relationship between RDW and mortality in patients with community-acquired pneumonia is unknown. The objective of this study was to evaluate the association of RDW with mortality in patients with CAP.

Methods: We performed a retrospective analysis of a prospective registry database of patients with CAP. Red cell distribution width was organized into quartiles. The pneumonia severity index and CURB-65 were calculated. The primary outcome was 30-day mortality. Secondary outcomes included the length of hospital stay, admission to the intensive care unit, vasopressor use, and the need for mechanical ventilation.

Results: A total of 744 patients were included. The PSI and CURB-65 were higher in patients with a high RDW. Multivariate logistic regression analysis identified higher categories of RDW, PSI, CURB- 65, and albumin as statistically significant variables. Thirty-day mortality was significantly higher in patients with a higher RDW. Among the secondary outcomes, the length of hospital stay and vasopressor use were significantly different between the groups. In a Cox proportional hazard regression analysis, patients with higher categories of RDW exhibited increased mortality before and after adjustment of the severity scales. receiver operating characteristics curves demonstrated improved Mortality prediction when RDW was added to the PSI or CURB-65.

Conclusion: Red cell distribution width was associated with 30-day mortality, length of hospital stay, and use of vasopressors in hospitalized patients with CAP. The inclusion of RDW improved the prognostic performance of the PSI and CURB-65.

(C) 2013

Introduction

? Source of support: This study was partially supported by grant number 02-2010-025 from SNUBH research fund.

* Corresponding author. Tel.: +82 31 787 7579; fax: + 82 31 787 4055.

E-mail address: [email protected] (K. Kim).

Community-acquired pneumonia is an important cause of morbidity and mortality throughout the world [1] and is the leading infectious cause of death [2,3]. Prognostication is an important part of the management of patients with CAP, and there are several prognostic scales currently in use. These scales include the pneumonia severity

0735-6757/$ – see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2012.06.004

index (PSI), CURB-65 (Confusion, Urea, Respiratory rate, Blood pressure, Age >=65), and CRB-65 (Confusion, Respiratory rate, Blood pressure, Age>=65). However, these scales have some limitations. The PSI is believed to be useful for identifying low-risk patients, whereas the CURB-65 and CRB-65 are believed to be useful for identifying high-risk patients. Therefore, there are wide- spread efforts to improve the prognostic performance of these scales. It has been suggested that certain biomarkers could improve the prognostic capacity of severity scales [4,5]. However, these biomarkers are somewhat expensive to obtain and are not always available immediately. Thus, improving the prognostic performance by adding biomarkers to severity scales has some limitations.

Red cell distribution width is a coefficient of variation of circulating red cells. This measure reftects the heterogeneity of red cell volume and is reported as a component of the complete blood count . Until now, the clinical significance of the RDW has been limited to the differential diagnosis of anemia; however, recent reports have associated an elevated RDW with outcomes in cardiovascular disease, rheumatoid arthritis, colon cancer, and Metabolic syndrome [6-10]. It has also been reported that RDW is a consistent predictor of mortality across different study populations [11]. The exact mechanisms underlying RDW ftuctuations are unknown; however, oxidative stress and inftammation have been suggested as inftuential factors [11]. The CBC is routinely performed in patients admitted from the emergency department (ED) with suspected infections. It is relatively inexpensive and yields valuable information. This study was designed to evaluate the association of RDW with mortality and to determine the prognostic significance of RDW in patients with CAP. We hypothesized that RDW would be associated with 30-day mortality and would exhibit

prognostic significance in patients with CAP.

Materials and methods

Study design and setting

A retrospective analysis of a prospective registry database of all consecutive patients with CAP was performed in a 950-bed tertiary academic hospital with an annual ED census of 67000. The institutional review board of our institute approved the study and granted a waiver of informed consent.

Selection of participants

We retrospectively analyzed the prospective registry database of patients who visited our ED and were subsequently hospitalized for CAP between April 2008 and March 2011. Community-acquired pneumonia was defined as evidence of a pulmonary infiltrate on 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 a pulmonary infiltrate was absent on the initial chest radiograph, abnormal lung sounds on the initial physical examination and pulmonary infiltrate on a follow-up chest radiograph were accepted as an equivalent. Eligible patients were older than 18 years and had a diagnosis of CAP. Of those, only 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 [12]. We excluded patients based on the following criteria: younger 18 years, had been transferred from another hospital, was discharged from a hospital within the previous 10 days, experienced an episode of pneumonia within the past 30 days, exhibits active pulmonary tuberculosis, has known HIV positivity, or is chronically immunosuppressed (defined as immunosuppression for solid organ transplanta- tion, postsplenectomy, receiving >= 10 mg/d prednisolone or equivalent for b 30 days, treatment with other immunosup- pressive agents, or neutropenia [b 1.0 x 109/L neutrophils]).

Data collection

Baseline clinical information was obtained through structured patient or proxy interviews, bedside assessments, and structured medical reviews. Collected clinical informa- tion included age and sex as well as comorbidities, such as chronic obstructive pulmonary disease; diabetes mellitus; malignancies; and cardiac, hepatic, renal, and central nervous system diseases. Recorded bedside assessments included temperature, respiratory rate, heart rate, and blood pressure. The following initial laboratory data at ED visit were also recorded: leukocyte count, hemoglobin level, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin (MCH), RDW, platelet count, blood glucose, serum creati- nine, Blood urea nitrogen , albumin, total cholesterol, prothrombin time (international normalized ratio), activated partial thromboplastin time, and C-reactive protein (CRP).

Methods of measurement

Complete blood counts were measure by automated hematology analyzer (Sysmex XE-2100; Sysmex Co, Kobe, Japan) during the ED visit, and RDW was reported as a part of the CBC results. Study nurses ascertained deaths by examining medical records during the hospital stay and by telephone after discharge. This study was based on a registry database for CAP. All patients who are diagnosed with CAP in the ED are enrolled in the registry. If patients had been enrolled in our registry, study nurses attempted to determine the disposition from the ED and the status of patients 1 month after the initial ED visit using the medical records for patients who were dead during the hospital stay. If the patients were discharged alive or transferred to other facility, the status of the patients was

ascertained by the presence of the medical records of the outpatient department 1 month after the ED visit. If the medical records did not exist beyond 1 month of the ED visit, telephone calls to patients or surrogates were conducted. In total, 5 attempts within 1 week were made to contact the patients. During the study periods, 744 patients were hospitalized for pneumonia. Medical records in 702 patients identified the status of patients 1 month after the ED visit. Telephone contacts were made in 42 patients. Of those, 10 patients were not reached, and these patients were identified as having been transferred to other facility after acute management for CAP. Telephone contacts to the physician of the transferred hospital were made to ascertain the status of these patients. The severity of the illness was assessed with the PSI and the CURB-65 for each patient during the ED visit.

Outcome measures

The primary end point was 30-day mortality after the ED visit. The survival time was also investigated. The secondary end points were hospital length of stay, use of vasopressors, intensive care unit admission, and mechanical ventilator requirement.

Primary data analysis

Impact of RDW on outcomes of patients with CAP patients

To evaluate the association of RDW with 30-day mortality, RDW values were categorized into quartiles using the following cutoffs: less than 13.3%, 13.3% to 14.1%, 14.1% to 15.2%, and greater than 15.2%. Baseline characteristics were compared across the RDW quartiles. An analysis of variance, ?2 test, or Fisher exact test was performed for comparison as appropriate. A multivariate logistic regression analysis was performed to identify independent factors associated with 30-day mortality. A Kaplan-Meier survival analysis was performed to test the equality of survival functions. Cox proportional hazard regression analyses were performed to evaluate the relationship between RDW and outcomes while adjusting for age, sex, severity index scores, and other risk factors that might confound the relationship between RDW and mortality. The proportional hazard assumption was confirmed through the examination of plots of Schoenfeld residuals. In total, 3 models were tested. Model 1 included RDW, age, sex, neoplastic disease, hematocrit, albumin, cholesterol, BUN, and prothrombin time as pre- dictors; model 2 included RDW, PSI, albumin, cholesterol, and prothrombin time; and model 3 included RDW, CURB-65, albumin, cholesterols and prothrombin time.

Test of the added ability of RDW to predict 30-day mortality

To evaluate the added Prognostic ability of the RDW, multivariate logistic regression analyses were performed by

adding RDW to each of the severity scales, and the Area Under the Receiver Operating Characteristic Curve was calculated. The statistical comparison of AUCs was conducted using the method of Hanley and McNeil [13].

All statistical analyses were performed using STATA/IC

10.1 (StataCorp LP, TX, USA). P b .05 was considered statistically significant.

Results

Characteristics of study subjects

During the study period, 744 patients were hospitalized via the ED for CAP. There were no missing data during the study period. The mean age was 70.1 (SD, 15.0) years, and 32% of patients were male. The overall 30-day mortality was 13.4%. The RDW values ranged from 11.1% to 28.4%

(median, 14.1%; interquartile range, 13.3%-15.2%). The baseline characteristics according to the RDW quartiles are presented in Table 1. Patients with a higher RDW were more likely to be older; have a neoplastic disease; and have lower hemoglobin, albumin, and cholesterol levels as well as a higher BUN level and a prolonged prothrombin time. The severity of pneumonia, as assessed by the PSI and CURB-65, was higher in patients with a higher RDW.

Main results

To identify independent risk factors associated with 30- day mortality, multivariate logistic regression analyses were performed, and a higher RDW, albumin, and severity scales, including pneumonia severity score and CURB-65, were identified as statistically significant variables (Table 2). The effect of each quartile of RDW on the clinical outcome of CAP is summarized in Table 3. Thirty-day mortality was significantly higher in patients with a higher RDW. Secondary outcomes, especially the length of the hospital stay and the incidence of vasopressor use, also differed significantly among the RDW quartiles. In Fig. 1, the Kaplan-Meier curves demonstrate the probability of mortal- ity by quartiles of RDW. Patients with a higher RDW had a higher probability of mortality. In the Cox proportional hazard regression analysis, patients with higher categories of RDW had increased mortality before and after adjustment for severity scales (Table 4). In model 1, patients in the fourth quartile of RDW exhibit a 2.31-fold higher mortality rate than those in the first quartile of RDW. The adjustment for severity scales only partially mitigated the effect of RDW on 30-day mortality in patients with CAP.

To evaluate whether RDW improves the prognostic performance of the PSI and CURB-65 in CAP patients, a receiver operating characteristic curve analysis was con- ducted. The AUCs to predict 30-day mortality were 0.74 (95% confidence interval [CI], 0.70-0.79) for the PSI scale

Table 1 Patient characteristics by quartiles of RDW

Variables

Total patients

Quartile (RDW)

P

n = 744

b 13.3 (n = 196)

13.3-14.1 (n = 172)

14.1-15.2 (n = 190)

>= 15.2 (n = 186)

Age (mean +- SD)

70.1 +- 15.0

65.2 +- 17.5

70.6 +- 14.3

73.1 +- 13.8

71.5 +- 12.7

b.001

Male sex (%)

238 (32.0)

62 (33.2)

65 (37.8)

58 (30.5)

50 (26.9)

.159

Comorbidities, n (%)

Heart failure

18 (2.4)

2 (1.0)

2 (1.2)

8 (4.2)

6 (3.2)

.120

Renal failure

83 (11.2)

16 (8.2)

15 (8.7)

24 (12.6)

28 (15.1)

.113

Liver disease

44 (5.9)

7 (3.6)

11 (6.4)

15 (7.9)

11 (5.9)

.322

COPD

148 (19.9)

38 (19.4)

39 (22.7)

38 (20.0)

33 (17.7)

.704

Neoplasm

195 (26.2)

35 (17.9)

26 (15.1)

55 (29.0)

79 (42.5)

b.001

neurologic disease

187 (25.1)

42 (21.4)

44 (25.6)

56 (29.5)

45 (24.2)

.331

Diabetes mellitus

222 (29.8)

48 (24.5)

50 (29.1)

63 (33.2)

61 (32.8)

.207

Laboratory findings

WBC

12.7 +- 6.8

13.0 +- 5.8

13.2 +- 7.1

12.8 +- 5.9

11.7 +- 8.1

.138

Hemoglobin level

12.1 +- 2.2

13.0 +- 1.8

12.5 +- 2.0

12.3 +- 2.0

10.8 +- 2.3

b.001

Hematocrit (%)

36.1 +- 6.4

38.0 +- 5.2

36.8 +- 6.1

36.8 +- 6.0

32.8 +- 6.7

b.001

MCV

93.4 +- 6.7

93.8 +- 4.7

93.0 +- 5.3

94.1 +- 6.4

92.5 +- 9.3

.061

MCH

31.3 +- 2.6

32.0 +- 1.9

31.4 +- 2.1

31.5 +- 2.3

30.5 +- 3.6

b.001

Platelet (x 103/mm3)

252.8 +- 127.0

252.1 +- 110.5

257.4 +- 110.4

256.4 +- 153.2

245.7 +- 128.8

.811

Glucose (mg/dL)

164.6 +- 111.6

160.0 +- 80.7

178.6 +- 144.3

164.7 +- 123.0

156.3 +- 90.3

.252

Albumin (g/dL)

3.4 +- 0.6

3.7 +- 0.5

3.5 +- 0.6

3.4 +- 0.5

3.2 +- 0.6

b.001

Cholesterol (g/dL)

143.0 +- 40.3

146.5 +- 34.9

144.6 +- 40.4

148.4 +- 42.6

132.3 +- 41.4

b.001

BUN (mg/dL)

24.4 +- 16.9

19.5 +- 11.6

22.8 +- 16.2

26.3 +- 18.2

29.0 +- 19.3

b.001

Creatinine (mg/dL)

1.4 +- 1.3

1.2 +- 1.0

1.3 +- 1.2

1.5 +- 1.5

1.6 +- 1.5

.052

arterial pH

7.43 +- 0.07

7.44 +- 0.07

7.42 +- 0.07

7.42 +- 0.07

7.43 +- 0.09

.392

PaO2

68.6 +- 29.7

69.3 +- 25.1

67.2 +- 25.3

68.7 +- 35.8

69.1 +- 31.1

.921

PaCO2

35.2 +- 10.9

35.1 +- 9.3

36.2 +- 12.0

35.7 +- 10.8

34.0 +- 11.2

.320

HCO3

22.6 +- 5.1

23.0 +- 3.7

22.8 +- 5.4

22.8 +- 5.6

21.8 +- 5.3

.104

Na

134.7 +- 7.6

134.7 +- 9.8

134.1 +- 6.1

135.0 +- 6.6

135.1 +- 7.0

.627

K

4.2 +- 0.7

4.1 +- 0.5

4.2 +- 0.7

4.2 +- 0.7

4.3 +- 0.7

.059

Cl

99.6 +- 6.7

99.9 +- 5.3

98.5 +- 7.0

99.6 +- 7.0

100.3 +- 7.4

.082

PT (INR)

1.2 +- 0.4

1.14 +- 0.30

1.19 +- 0.47

1.21 +- 0.28

1.27 +- 0.37

.004

apt (s)

43.2 +- 11.9

44.5 +- 13.6

42.1 +- 13.2

41.9 +- 8.0

44.3 +- 12.2

.060

CRP (mg/dL)

13.5 +- 9.4

12.9 +- 8.4

14.3 +- 10.2

13.2 +- 10.1

13.7 +- 8.9

.524

PSI class

b.001

I, II

132 (17.7)

66 (50.0)

35 (26.5)

15 (11.4)

16 (12.1)

III

136 (18.3)

50 (36.8)

40 (29.4)

31 (22.8)

15 (11.0)

IV

300 (40.3)

61 (20.3)

68 (22.7)

82 (27.3)

89 (29.7)

V

176 (23.7)

19 (10.8)

29 (16.5)

62 (35.2)

66 (37.5)

CURB-65

b.001

0

96 (12.9)

47 (49.0)

22 (22.9)

13 (13.5)

14 (14.6)

1

214 (28.8)

71 (33.2)

58 (27.1)

52 (24.3)

33 (15.4)

2

253 (34.0)

48 (19.0)

55 (21.7)

71 (28.1)

79 (31.2)

3

133 (17.9)

25 (18.8)

30 (22.6)

42 (31.6)

36 (27.1)

4

38 (5.1)

4 (10.5)

5 (13.2)

10 (26.3)

19 (50.0)

5

10 (1.3)

1 (10.0)

2 (20.0)

2 (20.0)

5 (50.0)

Abbreviations: COPD, chronic obstructive pulmonary disease; MCV, mean corpuscular volume; INR, international normalized ratio.

and 0.74 (95% CI, 0.69-0.79) for the CURB-65 scale. The AUCs were significantly increased when RDW was added to either scale (P b .05; Fig. 2).

Limitations

There are several limitations of this study. First, it was conducted at a single institution and only included patients

who were hospitalized via the ED. Thus, it cannot be generalized to all patients with CAP. Second, we did not gather and analyze data about anemia, transfusion status, or nutritional deficiencies, such as iron, vitamin B12, or folate deficiencies. Red cell distribution width is associated with anemia and red cell transfusion. However, a multivariate analysis demonstrated a significant association between the RDW and mortality, even after adjusting for hematocrit.

Variables Odds ratio (95% CI) P

RDW b 13.3

Reference

RDW 13.3-14.1

0.73 (0.28-1.91)

.518

RDW 14.1-15.2

1.11 (0.47-2.62)

.815

RDW >= 15.2

2.37 (1.04-5.42)

.040

PSI class I, II

Reference

PSI class III

1.47 (0.24-8.93)

.678

PSI class IV

4.76 (1.01-22.53)

.049

PSI class V

7.10 (1.42-35.42)

.017

CURB-65

0

Reference

1

1.34 (0.25-7.20)

.730

2

2.26 (0.45-11.30)

.323

3

2.44 (0.46-12.82)

.292

4

3.42 (0.58-20.06)

.174

5

37.02 (2.49-550.32)

.009

Hematocrit

1.03 (0.99-1.07)

.191

MCH

1.01 (0.92-1.10)

.869

Albumin

0.19 (0.11-0.33)

b.001

Cholesterol

1.01 (1.00-1.01)

.073

Prothrombin time

1.14 (0.57-2.26)

.716

Moreover, the RDW values gathered in this study were from initial CBCs before any definitive medical interventions. Third, in this study, patients were enrolled after the initial ED

Table 2 Multivariate logistic regression analysis

Table 3 Community-acquired pneumonia outcomes stratified by RDW

evaluation and laboratory testing. In addition, the ED physician who was responsible for enrolment of patients knew the results of the laboratory values, including RDW, before determining the disposition of patients. Thus, this might be a potential source of bias. However, in our ED, the disposition of patients with CAP was largely determined according to the PSI after the initial ED evaluation [12]. In the PSI, RDW was not considered during the calculation. Thus, during this study, disposition was not inftuenced by initial RDW values, and this study was performed only in hospitalized patients according to the PSI and the general status of the patients. Fourth, this analysis does not compare RDW with physician judgement and may not be beneficial in clinical practice. However, the current severity calculation index for CAP has its limitations. Our findings may be useful if the severity calculation index will be revised.

Discussion

This study demonstrated that a higher RDW was associated with increased 30-day mortality in patients with CAP, and this effect was especially pronounced in patients with an RDW greater than 15.2%. The secondary outcomes of length of hospital stay and Vasopressor requirement were also affected by an increasing RDW. In addition, RDW

30-d mortality, n (%) PSI

I, II III IV V

CURB-65 0

1

2

3

4

5

Secondary outcomes HLOS, median (IQR)

ICU admission, n (%)

Vasopressor use, n (%)

MV use, n (%) VDD median (IQR)

100 (13.4)

11 (5.6)

12 (7.0)

24 (12.6)

53 (28.5)

b.001

2/132 (1.5%)

4/136 (2.9%)

41/300 (13.7)

53/176 (30.1)

0/66 (0.0%)

1/50 (2.0%)

6/61 (9.8%)

4/19 (21.1%)

0/35 (0.0%)

1/40 (2.5%)

4/68 (5.9%)

7/29 (24.1%)

0/15 (0.0%)

1/31 (3.2%)

10/82 (12.2%)

13/62 (21.0%)

2/16 (12.5%)

1/15 (6.7%)

21/89 (23.6%)

29/66 (43.9%)

.026

.730

.011

.026

2/96 (2.1%)

11/214 (5.1%)

36/253 (14.2%)

27/133 (20.3%)

16/38 (42.1%)

8/10 (80%)

1/47 (2.1%)

0/71 (0.0%)

4/48 (8.3%)

4/25 (16.0%)

1/4 (25.0%)

1/1 (100.0%)

0/22 (0.0%)

3/58 (5.2%)

2/55 (3.6%)

3/30 (10.0%)

2/5 (40.0%)

2/2 (100.0%)

1/13 (7.7%)

3/52 (5.8%)

7/71 (9.9%)

10/42 (23.8%)

2/10 (20.0%)

1/2 (50.0%)

0/14 (0.0%)

5/33 (15.2%)

23/79 (29.1%)

10/36 (27.8%)

11/19 (57.9%)

4/5 (80.0%)

.392

.006

b.001

.282

.237

1.000

11 (7-18)

10 (6-15)

11 (8-17.5)

11.5 (8-21)

12 (8-20)

.004

107 (14.4)

18 (9.2)

28 (16.3)

29 (15.3)

32 (17.2)

.075

103 (13.8)

10 (5.1)

23 (13.4)

32 (16.8)

38 (20.4)

b.001

102 (13.7)

9 (3-17)

17 (8.7)

8 (5-15)

25 (14.5)

10 (3-28)

28 (14.7)

11 (5-17)

32 (17.2)

7 (3-13.5)

.080

.100

Abbreviations: HLOS, hospital length of stay; MV, mechanical ventilator; VDD, ventilator-dependent days.

Variables Total patients

Quartile (RDW)

P

(n = 744)

Quartile 1 (RDW b 13.3,

Quartile 2

(13.3 <= RDW b 14.1,

Quartile 3

(14.1 <= RDW b 15.2,

Quartile 4

(RDW >= 15.2,

n = 196)

n = 172)

n = 190)

n = 186)

Fig. 1 Kaplan-Meier survival curves.

combined with standard PSIs improved mortality prediction capability compared with each severity index alone.

Red cell distribution width has been reported to be a strong predictor of mortality in the general population of adults 45 years or older [14]; among outpatients with cardiovascular disease, cancer, or chronic lower respiratory tract disease [15]; and in heart failure populations [6]. In critically ill patients, RDW was also reported to be significantly associated with the risk of death and of blood stream infection [16]. Although a dose-response relationship between RDW quartiles and 30- day mortality was not obvious, the mortality of patients with a high-range RDW (b 15.2%) was significantly greater than that of patients with an RDW less than 13.3%. In addition, the length of hospital stay and the use of vasopressors were significantly increased in patients in the highest range of RDW. Thus, initial RDW measurements could be a prognostic marker in patients with CAP.

Recently, Braun et al [17] reported that RDW was associated with mortality and morbidity in young patients hospitalized with CAP. Their finding is similar to our results.

However, they only included patients who were younger than 60 years, which they cited as a limitation of their study. Age is a significant prognostic factor in various diseases, including CAP. In our study, we included all patients who were hospitalized for CAP. The mean age of our study population was 70.1 years old, and most patients (N 80% of study population) were older than 60 years. Thus, our findings support RDW as a significant prognostic factor in patients with CAP across all ages.

Red cell distribution width reftects the heterogeneity of the red cell population and has been used in diagnosing certain anemias, especially those that resulted from iron, vitamin B12, or folate deficiencies. An increased RDW can be caused by conditions, such as the release of a greater number of immature cells into the circulation (blood loss), abnormal hemoglobin level (Sickle cell anemia), and hemolysis. Thus, anemia was a possible cause of the increased RDW values in our data. Hemoglobin level and hematocrit values varied significantly across the quartiles of RDW. However, in multivariate logistic regression analyses, RDW was

Table 4

Hazard ratios by categories of RDW

Model

Hazard ratio (95% CI)

Test for trend

RDW

quartile 1

RDW

quartile 2

RDW

quartile 3

RDW

quartile 4

P value

Model 1

1.0 (reference)

0.86 (0.36-2.05)

1.28 (0.59-2.74)

2.31 (1.12-4.79)

b.001

Model 2

1.0 (reference)

0.76 (0.32-1.81)

1.06 (0.49-2.28)

2.06 (1.01-4.20)

b.001

Model 3

1.0 (reference)

0.88 (0.37-2.08)

1.37 (0.64-2.95)

2.45 (1.20-4.98)

b.001

Model 1 is adjusted for age, sex, neoplastic disease, hematocrit, albumin, cholesterol, BUN, and prothrombin time. Model 2 is adjusted for PSI, albumin, cholesterol, and prothrombin time. Model 3 is adjusted for CURB-65, albumin, cholesterol, and prothrombin time.

Fig. 2 Receiver operating characteristic curves and AUC of the different models to predict 30-day mortality.

independently associated with the 30-day mortality of patients with CAP after adjusting for hematocrit values. The exact mechanisms of an association of RDW with the mortality of CAP need to be determined. It has been suggested that inftammation and oxidative stress affect red cell homeostasis. A previous study demonstrated that RDW displayed a strong, graded association with inftammatory biomarkers in general outpatient populations [18], and another study indicated that serum antioxidant levels, including selenium and carotenoids, were associated with RDW in older women [19]. Community-acquired pneumonia is an infectious disease that results in inftammatory and oxidative stress to the host. If these stresses are severe, mortality will be increased. In our data, patients with a higher RDW had a tendency toward higher severity index scores,

and the overall 30-day mortality was also higher in patients with a higher RDW. However, the levels of CRP, an inftammatory biomarker, were not different across the quartile ranges of RDW. Nonetheless, CRP-insensitive inftammatory mechanisms may affect RDW in these patients. In our study, age, neoplastic disease, hemoglobin level, albumin, cholesterol, BUN, and prothrombin time were found to vary significantly across the quartile ranges of RDW in univariate analyses. Among these variables, albumin was significantly correlated with RDW and was identified as an independent predictor of mortality in multivariate analyses. In our previous report, albumin was found to be an independent predictor of mortality in patients with CAP [5], and this finding was confirmed in the current study. Thus, low serum albumin levels in patients with high RDW

values may reftect the severity of CAP. The exact relationship remains to be determined.

Pneumonia severity scales including the PSI and CURB-65 have proven useful for the identification of a low and high risk of death among patients with CAP. However, the PSI was derived from hospitalized patients with CAP to identify low- risk patients who could be safely managed as outpatients, whereas the CURB-65 was designed to identify patients at a high risk of death. Thus, the prognostic capabilities of the PSI and CURB-65 alone may be limited. Many studies have reported that added biomarkers could improve the performance of these severity scales. In our study, the mortality prediction of both the PSI and CURB-65 was improved by the addition of RDW as a severity criterion. In patients who present to an ED for CAP, a CBC is almost uniformly performed and is relatively inexpensive. Red cell distribution width is com- monly reported as a component of the CBC. Thus, RDW can be used as a biomarker in predicting outcomes in patients with CAP without incurring additional cost.

In this study, we concluded that the initial RDW was associated with 30-day mortality, length of hospital stay, and the use of vasopressors in hospitalized patients with CAP and that RDW improved the prognostic performance of severity scales, such as the PSI or CURB-65.

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):S27-72.
  2. Almirall J, Bolibar I, Vidal J, et al. Epidemiology of community- acquired pneumonia in adults: a population-based study. Eur Respir J 2000;15(4):757-63.
  3. Armstrong GL, Conn LA, Pinner RW. Trends in infectious disease mortality in the United States during the 20th century. JAMA 1999;281(1):61-6.
  4. Menendez R, Martinez R, Reyes S, et al. Biomarkers improve mortality prediction by prognostic scales in community-acquired pneumonia. Thorax 2009;64(7):587-91.
  5. Lee JH, Kim J, Kim K, et al. Albumin and C-reactive protein have prognostic significance in patients with community-acquired pneu- monia. J Crit Care 2011;26(3):287-94.
  6. Felker GM, Allen LA, Pocock SJ, et al. Red cell distribution width as a novel prognostic marker in heart failure: data from the CHARM Program and the Duke Databank. J Am Coll Cardiol 2007;50(1):40-7.
  7. Ani C, Ovbiagele B. Elevated red blood cell distribution width predicts mortality in persons with known stroke. J Neurol Sci 2009;277(1-2): 103-8.
  8. Spell DW, Jones Jr DV, Harper WF, et al. The value of a complete blood count in predicting cancer of the colon. Cancer Detect Prev 2004;28(1):37-42.
  9. Lee WS, Kim TY. Relation between red blood cell distribution width and inflammatory biomarkers in rheumatoid arthritis. Arch Pathol Lab Med 2010;134(4):505-6.
  10. Sanchez-Chaparro MA, Calvo-Bonacho E, Gonzalez-Quintela A, et al. Higher red blood cell distribution width is associated with the metabolic syndrome: results of the Ibermutuamur CArdiovascular RIsk assessment study. Diabetes Care 2010;33(3):e40.
  11. Patel KV, Semba RD, Ferrucci L, et al. Red cell distribution width and mortality in older adults: a meta-analysis. J Gerontol A Biol Sci Med Sci 2010;65(3):258-65.
  12. Jo S, Kim K, Jung K, et al. The effects of incorporating a pneumonia severity index into the admission protocol for community-acquired pneumonia. J Emerg Med 2010.
  13. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148(3):839-43.
  14. Patel KV, Ferrucci L, Ershler WB, et al. Red blood cell distribution width and the risk of death in middle-aged and older adults. Arch Intern Med 2009;169(5):515-23.
  15. Perlstein TS, Weuve J, Pfeffer MA, et al. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch Intern Med 2009;169(6):588-94.
  16. Bazick HS, Chang D, Mahadevappa K, et al. Red cell distribution width and all-cause mortality in critically ill patients. Crit Care Med 2011;39(8):1913-21.
  17. Braun E, Domany E, Kenig Y, Mazor Y, Makhoul BF, Azzam ZS. Elevated red cell distribution width predicts poor outcome in young patients with community acquired pneumonia. Crit Care 2011;15(4): R194.
  18. Lippi G, Targher G, Montagnana M, et al. Relation between red blood cell distribution width and inflammatory biomarkers in a large cohort of unselected outpatients. Arch Pathol Lab Med 2009;133(4):628-32.
  19. Semba RD, Patel KV, Ferrucci L, et al. Serum antioxidants and inflammation predict red cell distribution width in older women: the Women’s Health and Aging Study I. Clin Nutr 2010;29(5):600-4.