- Email: [email protected]
The association between red cell distribution width and poor outcomes in hospitalized patients with influenza
Journal of Critical Care 41 (2017) 166–169
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
The association between red cell distribution width and poor outcomes in hospitalized patients with influenza Guy Topaz ⁎, Yona Kitay-Cohen, Lee Peled, Wesal Gharra, Keren Kaminer, Mayan Eitan, Lamis Mahamid, Lotan Shilo Department of Internal Medicine “C”, Meir Hospital, Kfar-Saba and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
a r t i c l e
i n f o
Available online xxxx Keywords: Influenza Red cell distribution width Hospital complications Mortality
a b s t r a c t Purpose: To examine an association between red blood cell distribution width (RDW) and the prognosis of influenza patients. Methods: We conducted a retrospective analysis of patients hospitalized with influenza during 2012–2015 in the internal medicine wards of one medical center. RDW measurements during hospitalization were analyzed. Primary outcome was complicated hospitalization (defined as at least one of: length of stay ≥7 days, need for mechanical ventilation, septic shock, transfer to intensive-care, or 30-day mortality). Secondary outcome was 30day mortality. Results: 153 patients were included, mean age: 62.5 ± 1, 82 (54%) male; 84 (55%) had a high RDW value (N14.5%) during hospitalization. Patients with high and low RDW (≤14.5%) had similar age and comorbidity profiles, but those with high RDW had lower hemoglobin and higher creatinine levels. Patients with high RDW had a higher rate of complicated hospitalization (32.5% vs. 10.3%, p b 0.01) and a trend for increased 30-day mortality. In a multivariate regression model, high RDW was a predictor of complicated hospitalization (OR 5.03, 95% CI 1.81–13.93, p b 0.01). Each 1-point increase in RDW was associated with a 29% increase in the risk for the primary outcome. Conclusion: RDW N 14.5% was a predictor of severe hospital complications in patients with influenza. © 2017 Elsevier Inc. All rights reserved.
1. Introduction
2. Materials and methods
Influenza infection is usually self-limited, but occasionally may lead to substantial morbidity and even to mortality, especially in an elderly and immunocompromised population [1]. Seasonal influenza occurs mainly during winter and is responsible for an estimated 3–5 million cases of severe illness and up to 500,000 deaths annually world-wide [2]. Red blood cell distribution width (RDW) is a measure of the degree of heterogeneity of erythrocyte volume. Prior studies have shown that elevated RDW is associated with increased mortality among patients with various disease states, such as cardiovascular disease [3-5], stroke [6], renal disease [7], chronic obstructive pulmonary disease [8], septic shock [9] and community acquired pneumonia [10].To the best of our knowledge, RDW levels among influenza patients have not been evaluated previously. This study examined the association between RDW levels and short-term outcomes among influenza patients.
2.1. Study population and design
⁎ Corresponding author at: Department of Internal Medicine C, Meir Medical Center, Tchernichovsky St. 59, Kfar Sava 4428164, Israel. E-mail address: [email protected] (G. Topaz).
http://dx.doi.org/10.1016/j.jcrc.2017.05.014 0883-9441/© 2017 Elsevier Inc. All rights reserved.
We analyzed the data of all adult patients (age N 18) hospitalized with the diagnosis of influenza infection in the5 internal medicine wards of Meir Medical Center from January 1, 2012 through December 31, 2015. Meir Medical Center is a 760-bed, tertiary care university hospital. Influenza was diagnosed using polymerase chain reaction kit (3 M integrated cycler, Focus Diagnostics, Cypress, CA, USA). During the study period, RDW and other laboratory tests were analyzed in a single central laboratory using standardized automated kits (Advia 2120, Siemens, Erlangen, Germany). Data were collected from electronic medical records. Mortality data were confirmed using the Israel Central Bureau of Statistics. Data collection was approved by the hospital ethics committee. 2.2. Definitions For analysis, patients were categorized according to RDW values during hospitalization: high RDW (at least one RDW N 14.5%) and low RDW (no RDW value N 14.5%). The cutoff of 14.5% was determined as the upper limit of normal values provided by our laboratory. Primary
G. Topaz et al. / Journal of Critical Care 41 (2017) 166–169
167
outcome was complicated hospitalization, defined as at least one of the following: hospital stay N 7 days, need for mechanical ventilation, septic shock, transfer to the Intensive Care Unit (ICU) and 30-day mortality. Secondary outcome was 30-day, all-cause mortality. Patients were followed until December 31, 2015 for all-cause mortality. Patients were further subdivided according to the change in RDW levels during hospitalization (discharge RDW minus baseline RDW). A significant change was defined as more than ± 0.6%.
analysis (mean age 62.5 ± 1 years, 53.6%men). Eighty-four (55%) patients were categorized in the high RDW group and 69 (45%) in the low RDW group. Patients in the two groups were of similar age and had similar comorbidity profiles. The mean hemoglobin level at admission was lower and the mean creatinine level was higher for patients in the high RDW than in the low RDW group (11.2 ± 1.8 vs. 12.9 ± 1.3, p = 0.00 and 1.9 ± 2.0 vs. 1.2 ± 0.6, p b 0.01, respectively) (Table 1).
2.3. Statistical analysis
3.1. RDW and in-hospital complications
Data are presented as numbers and percentages for nominal data and as means and standard deviations for continuous parameters. Differences between the high RDW and low RDW groups were analyzed with chi-square or Fisher's exact test, as appropriate. Continuous variables were examined for normality (Shapiro-Wilk test) and data were analyzed accordingly. The t-test was used for normally distributed variables and the Mann-Whitney for non-parametric variables. Receiver operating characteristic (ROC) curves were used to evaluate the ability of maximum RDW and the change in RDW to predict unfavorable outcomes. A logistic regression model was applied to estimate odds ratios for complicated hospitalization. In a multivariate logistic regression model, the following variables were accounted forage, gender, high RDW, change in RDW N ±0.6% and hemoglobin levels at admission. A Cox proportional hazard model was applied to estimate hazard ratios for all-cause mortality until the end of the follow-up period. The variables accounted for in this model were diabetes mellitus, creatinine at admission, hemoglobin at admission, high RDW and complicated hospitalization. p b 0.05 was considered statistically significant. Data were analyzed with SPSS Version 21 (IBM Corporation, Armonk, NY, USA).
Complicated hospitalization rates were significantly higher in the high than in the low RDW group (32.5% vs. 10.3%, p b 0.01). In-hospital characteristics and complications are depicted in Table 2. ROC curve analysis revealed that the maximum RDW of 14.5% was the optimal cut-off value associated with complicated hospitalization. Risk assessment using RDW level as a continuous variable demonstrated that each 1-point increase in RDW was associated with a 29% increase in the risk for the primary endpoint of complicated hospitalization. A multivariate logistic regression model revealed that high RDW was the only factor examined that was an independent predictor of complicated hospitalization (OR 5.03, 95% CI 1.81 to 13.93, p b 0.01). A sub-analysis of the 35 patients with complicated hospitalization revealed a high RDW value in 27 (77%). The mean maximum RDW of patients with complicated hospitalization was significantly higher than that of patients with a non-complicated hospitalization (15.9 ± 1.9 vs. 14.9 ± 1.8, p b 0.01). The same trend was noted when comparing patients with and without complicated hospitalization, with mean RDW levels at baseline and at discharge higher in the complicated hospitalization group (14.8 ± 1.6 vs. 14.4 ± 1.6, p = 0.2 and 15.3 ± 2.0 vs. 14.8 ± 1.8, p = 0.13, respectively).
3. Results Among 162 patients hospitalized with a diagnosis of influenza during the study period, 9 were excluded from the analysis due to incomplete data. A total of 153 influenza patients were included in this
Table 1 Characteristics of patients hospitalized with influenza virus infection. Variables
RDW N 14.5% (n = 84)
RDW ≤ 14.5% (n = 69)
Total (n = 153)
p-Value
Age Male gender Comorbidities Asthma COPD Diabetes mellitus Hypertension Hyperlipidemia CRF Tobacco use Baseline laboratory values Hemoglobin (g/dl) WBC (K/microL) PLT (K/microL)
64 ± 12.7 51.2
60.9 ± 15.7 56.5
62.5 ± 14.2 53.6
0.19 0.52
4.8 7.1 27.4 28.6 28.6 13.1 13.1
7.2 2.9 20.3 30.4 37.7 5.8 14.5
5.9 5.2 24.2 29.4 32.7 9.8 13.7
0.73 0.29 0.31 0.8 0.23 0.17 0.8
Creatinine (g/dL) Bilirubin (mg/dL) CRP (mg/dL) Baseline RDW (%)
11.2 ± 1.8
12.9 ± 1.3
11.9 ± 1.8
0.00
6.8 ± 4.8 186 ± 88
6.8 ± 3.1 193 ± 62
0.89 0.56
1.9 ± 2.0 0.61 ± 0.35 11.2 ± 7.9 15.3 ± 1.6
1.2 ± 0.6 0.68 ± 0.68 9.8 ± 6.1 13.4 ± 0.6
6.8 ± 4.1 189.4 ± 77.3 1.36 ± 1.4 0.64 ± 0.5 9.9 ± 6.1 14.5 ± 1.6
0.00 0.43 0.31 0.00
WBC, white blood cells; PLT, platelets; RDW, red blood cell distribution width; COPD, chronic obstructive pulmonary disease; CRF, chronic renal failure. Data are presented as means ± SD or percentages of presented cases. Normal ranges of values of laboratory data: Hemoglobin (12–16), WBC (4.8–10.8), PLT (150–400), creatinine (0.5–1.2), bilirubin (0.2–1.5), CRP (0.0–0.5).
3.2. RDW and mortality Thirty-day mortality rates were higher for patients with a high RDW value than for those without (2.4% vs. 0%, p = 0.5). Two patients died due to complications related to influenza infection. Four patients (2.7%) died within 90 days due to influenza complications, 3 had a high RDW value (3.6% vs. 1.5%, p = 0.63). During a mean follow-up of 21 ±10 months, 24 patients (15.9%) died. A high RDW was associated with higher long-term all-cause mortality (21.7% vs. 8.8%, p = 0.04). However, a multivariate Cox regression model analysis demonstrated that only complicated hospitalization, and not high RDW, was
Table 2 In-hospital characteristics and outcomes of patients hospitalized with influenza virus infection according to RDW levels. Variables
RDW N 14.5% (n = 84)
RDW ≤ 14.5% (n = 69)
Total (n = 153)
p-Value
Acute renal failure Septic shock Syncope COPD exacerbation Max fever, °C ICU transfer Mechanical ventilation Oseltamivir treatment Hospitalization length, days 30-day mortality rate
10.7 3.6 2.4 4.8 38.1 ± 0.88 4.9 2.4
8.8 0 5.9 0 38.0 ± 0.9 2.9 0
9.9 2.0 3.9 2.6 38.0 ± 0.8 4.0 1.3
0.7 0.25 0.41 0.13 0.03 0.69 0.5
76.2
77.9
77.0
0.08
6.4 ± 5.9
4.6 ± 2.9
5.6 ± 4.9
0.02
2.4
0
1.3
0.5
COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; RDW, red blood cell distribution width; Max, maximum. Data are presented as means ± SD or percentages of presented cases.
168
G. Topaz et al. / Journal of Critical Care 41 (2017) 166–169
Table 3 Cox proportional hazard model for long-term mortality among patients with influenza, adjusted for all significant variables in univariate analysis. Variables
OR
Diabetes mellitus Creatinine admission Hemoglobin Max RDW N 14.5 Complicated hospitalization
2.1 1.09 0.79 1.16 3.26
95% CI
p-Value
Lower
Upper
0.87 0.90 0.60 0.36 1.32
5.12 1.31 1.04 3.72 8.07
0.10 0.37 0.09 0.80 0.01
RDW, red blood cell distribution width.
associated with long-term mortality (HR 3.26, 95% CI 1.32–8.07, p = 0.01) (Table 3). Change in RDW levels during the index hospitalization. We examined an association between the change in RDW levels during the index hospitalization and the primary outcome of complicated hospitalization. 1) Baseline to maximum values. An increase in RDW of N 1.5% between the baseline RDW and a consecutive measure was significantly associated with complicated hospitalization (OR 3.5, 95% CI 1.57–7.79, p b 0.01).This increase was not a significant independent predictor of complicated hospitalization in multivariate analysis (Fig. 1). 2) Baseline to discharge values. An absolute difference in RDW (N± 0.6%) was significantly associated with complicated hospitalization (OR 3.4, 95% CI 1.26–8.97, p = 0.01); however, this change was not a significant independent predictor of complicated hospitalization in multivariate analysis. Interestingly, the mean change in RDW during hospitalization was greater for patients with complicated hospitalization (0.86 ± 0.9 vs. 0.5 ± 0.6, p = 0.04) (Fig. 2). 4. Discussion In this study of Influenza patients hospitalized in internal medicine wards, our main findings were: 1) A RDW N 14.5% at any time during hospitalization was independently associated with a complicated hospital course and showed a trend for increased short-term mortality. 2) A significant absolute change of N0.6%in RDW during hospitalization was associated with a complicated hospitalization; however, in contrast to an elevated RDW level, this was not statistically significant in multivariate analysis. 3) Among patients with a complicated hospitalization, the mean RDW level was higher and the mean change in RDW during hospitalization was greater, compared to patients without a complicated hospitalization. 4) An elevated RDW level during hospitalization was
not associated with long-term mortality, with a mean follow-up period of 21 ± 10 months. Seasonal influenza infection may cause severe illness and can be fatal. The clinical course of the disease is mainly dependent on host and viral characteristics [1,11]. Early administration of neuraminidase inhibitors, such as oseltamivir (75 mg twice daily), has been shown to improve outcomes among influenza patients [12]. According to the World Health Organization, severe illness may require the use of higher doses (150 mg twice daily) of oseltamivir [12]. Therefore, early recognition of the development of complications is important. Though not a specific marker, elevated RDW has been shown to predict worse prognosis in patients with chronic medical conditions [3-6] and acute infectious diseases [9,10,13,14]. In acute infectious states, RDW levels have been shown to predict mortality when measured at different intervals from hospital admission: baseline [10,15], day 3 [16] and day 8 [9]. In our study, RDW level N 14.5% measured at any point during the index hospitalization was strongly associated with severe complications in patients with influenza. This finding, as far as we know, has not been previously described. Our findings support the predictive value of a significant change in RDW for adverse outcomes. Interestingly, a fall or rise in RDW of N 0.6%, regardless of baseline RDW values, was associated with a complicated course among influenza patients. This finding supports a study that showed a rise in RDW of 0.4% throughout hospitalization was associated with in-hospital mortality among heterogeneous patients hospitalized in an internal medicine ward [17]. Gorelik et al. recently showed that among 980 communityacquired pneumonia patients, even a smaller absolute RDW change of ± 0.4% was associated with worse short-term outcomes [18]. We believe that serial RDW measurements may be informative, in addition to baseline RDW levels. We demonstrated that high RDW is associated with a positive trend for 30-day mortality; yet is not significantly associated with longer term all-cause mortality. In contrast, prior studies showed baseline RDW to be associated with worse outcomes in the settings of sepsis [9,14], pneumonia [10,15] and endocarditis [13]. Moreover, Patel et al. found that RDW is a strong predictor of mortality in the general population of adults 45 years and older [19]. The exact mechanism by which RDW is associated with short-term poor outcomes in this scenario is not yet understood. It is possible that elevated RDW, which reflects deregulation of erythrocyte homeostasis with impaired red blood cell production [20], can serve as a marker of inflammation and oxidative state [21,22]. Elevated RDW could be related to interactions among various factors, such as erythropoietin, proinflammatory cytokines and the hematopoietic response to erythropoietin [23]. RDW seems to reflect general health status.
23
21
19
17
15
13
11 Baseline RDW
Maximal RDW
Baseline RDW
Maximal RDW
Fig. 1. Baseline and maximum mean RDW in patients with complicated (n = 35) and non-complicated (n = 118) hospitalizations.
G. Topaz et al. / Journal of Critical Care 41 (2017) 166–169
169
23
RDWLevel(%)
21
19
17
15
13
11 Baseline RDW
RDW at discharge
Baseline RDW RDW at discharge
Fig. 2. Baseline and discharge mean RDW in patients with complicated (n = 35) and non-complicated (n = 118) hospitalizations.
Therefore, influenza patients with elevated RDW were presumably in poor general condition. Patients with high RDW levels had lower hemoglobin levels, which could influence RDW levels. However, a multivariate analysis model, including both factors, demonstrated that high RDW, and not hemoglobin levels, was associated with complicated hospitalization. This study had a few limitations. The retrospective design precluded determining whether a causal relationship exists between RDW and complicated hospitalization. Furthermore, measuring RDW levels during hospitalization and after medical intervention may influence the levels of RDW. Finally, the cohort was small, resulting in a lack of statistical power. Thus, a larger prospective study should be conducted. 5. Conclusion RDW N 14.5% is associated with a complicated hospitalization in patients with influenza. Combined with other known risk factors, RDW may serve as a simple tool to improve risk stratification among patients with influenza. Conflict of interests None. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] Nichol KL. Influenza vaccination in the elderly: impact on hospitalization and mortality. Drugs Aging 2005;22:495–515. [2] Simpson CR, Lone N, Kavanagh K, et al. Seasonal influenza vaccine effectiveness community-based national-linked database to determine the effectiveness of the seasonal trivalent influenza vaccine. Southampton (UK): NIHR Journals Library. Health Services and Delivery Research; 2013 Nov. [3] Khaki S, Mortazavi SH, Bozorgi A, et al. Relationship between red blood cell distribution width and mortality of patients with acute myocardial infarction referring to Tehran Heart Center. Crit Pathw Cardiol 2015;14:112–5. [4] Sahin O, Akpek M, Sarli B, et al. Association of red blood cell distribution width levels with severity of coronary artery disease in patients with non-ST elevation myocardial infarction. Med Princ Pract 2015;24:178–83.
[5] Topaz G, Haim M, Kusniec J, et al. Association between red cell distribution width and mortality after cardiac resynchronization therapy. Isr Med Assoc J 2015;17: 505–9. [6] Ani C, Ovbiagele B. Elevated red blood cell distribution width predicts mortality in persons with known stroke. J Neurol Sci 2009;277:103–8. [7] Lippi G, Targher G, Montagnana M, et al. Relationship between red blood cell distribution width and kidney function tests in a large cohort of unselected outpatients. Scand J Clin Lab Invest 2008;68:745–8. [8] Seyhan EC, Özgül MA, Tutar N, et al. Red blood cell distribution and survival in patients with chronic obstructive pulmonary disease. COPD 2013;10:416–24. [9] Leonardo L, María MM, Martín, et al. Red blood cell distribution width during the first week is associated with severity and mortality in septic patients. PLoS One 2014;9:e105436. [10] Lee JH, Chung HJ, Kim K, et al. Red cell distribution width as a prognostic marker in patients with community-acquired pneumonia. Am J Emerg Med 2013;31:72–9. [11] Uraki R, Kiso M, Shinya K, et al. Virulence determinants of pandemic A(H1N1)2009 influenza virus in a mouse model. J Virol 2013 Feb;87(4):2226–33. [12] World Health Organization. WHO guidelines for pharmacological management of pandemic influenza A(H1N1) 2009 and other influenza viruses. Geneva: World Health Organization; 2010 Feb. Available from http://www.who.int/csr/resources/ publications/swineflu/h1n1_guidelines_pharmaceutical_mngt.pdf?ua=1. [13] Guray Y, Ipek EG, Guray U, et al. Red cell distribution width predicts mortality in infective endocarditis. Arch Cardiovasc Dis 2014;107:299–307. [14] Mahmood NA, Mathew J, Kang B, et al. Broadening of the red blood cell distribution width is associated with increased severity of illness in patients with sepsis. Int J Crit Illn Inj Sci 2014;4:278–82. [15] Braun E, Domany E, Kenig Y, et al. Elevated red cell distribution width predicts poor outcome in young patients with community acquired pneumonia. Crit Care 2011; 15:R194. [16] Ku NS, Kim HW, Oh HJ, et al. Red blood cell distribution width is an independent predictor of mortality in patients with gram-negative bacteremia. Shock 2012;38: 123–7. [17] Shteinshnaider M, Barchel D, Almoznino-Sarafian D, et al. Prognostic significance of changes in red cell distribution width in an internal medicine ward. Eur J Intern Med 2015;26:616–22. [18] Gorelik O, Izhakian S, Barchel D, et al. Changes in red cell distribution width during hospitalization for community-acquired pneumonia: clinical characteristics and prognostic significance. Lung 2016;194:985–95. [19] 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;169:515–23. [20] Salvagno GL, Sanchis-Gomar F, Picanza A, et al. Red blood cell distribution width: a simple parameter with multiple clinical applications. Crit Rev Clin Lab Sci 2015;52: 86–105. [21] 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;133:628–32. [22] Pierce CN, Larson DF. Inflammatory cytokine inhibition of erythropoiesis in patients implanted with a mechanical circulatory assist device. Perfusion 2005;20:183. [23] Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med 2005;352: 1011–23.
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
The association between red cell distribution width and poor outcomes in hospitalized patients with influenza Guy Topaz ⁎, Yona Kitay-Cohen, Lee Peled, Wesal Gharra, Keren Kaminer, Mayan Eitan, Lamis Mahamid, Lotan Shilo Department of Internal Medicine “C”, Meir Hospital, Kfar-Saba and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
a r t i c l e
i n f o
Available online xxxx Keywords: Influenza Red cell distribution width Hospital complications Mortality
a b s t r a c t Purpose: To examine an association between red blood cell distribution width (RDW) and the prognosis of influenza patients. Methods: We conducted a retrospective analysis of patients hospitalized with influenza during 2012–2015 in the internal medicine wards of one medical center. RDW measurements during hospitalization were analyzed. Primary outcome was complicated hospitalization (defined as at least one of: length of stay ≥7 days, need for mechanical ventilation, septic shock, transfer to intensive-care, or 30-day mortality). Secondary outcome was 30day mortality. Results: 153 patients were included, mean age: 62.5 ± 1, 82 (54%) male; 84 (55%) had a high RDW value (N14.5%) during hospitalization. Patients with high and low RDW (≤14.5%) had similar age and comorbidity profiles, but those with high RDW had lower hemoglobin and higher creatinine levels. Patients with high RDW had a higher rate of complicated hospitalization (32.5% vs. 10.3%, p b 0.01) and a trend for increased 30-day mortality. In a multivariate regression model, high RDW was a predictor of complicated hospitalization (OR 5.03, 95% CI 1.81–13.93, p b 0.01). Each 1-point increase in RDW was associated with a 29% increase in the risk for the primary outcome. Conclusion: RDW N 14.5% was a predictor of severe hospital complications in patients with influenza. © 2017 Elsevier Inc. All rights reserved.
1. Introduction
2. Materials and methods
Influenza infection is usually self-limited, but occasionally may lead to substantial morbidity and even to mortality, especially in an elderly and immunocompromised population [1]. Seasonal influenza occurs mainly during winter and is responsible for an estimated 3–5 million cases of severe illness and up to 500,000 deaths annually world-wide [2]. Red blood cell distribution width (RDW) is a measure of the degree of heterogeneity of erythrocyte volume. Prior studies have shown that elevated RDW is associated with increased mortality among patients with various disease states, such as cardiovascular disease [3-5], stroke [6], renal disease [7], chronic obstructive pulmonary disease [8], septic shock [9] and community acquired pneumonia [10].To the best of our knowledge, RDW levels among influenza patients have not been evaluated previously. This study examined the association between RDW levels and short-term outcomes among influenza patients.
2.1. Study population and design
⁎ Corresponding author at: Department of Internal Medicine C, Meir Medical Center, Tchernichovsky St. 59, Kfar Sava 4428164, Israel. E-mail address: [email protected] (G. Topaz).
http://dx.doi.org/10.1016/j.jcrc.2017.05.014 0883-9441/© 2017 Elsevier Inc. All rights reserved.
We analyzed the data of all adult patients (age N 18) hospitalized with the diagnosis of influenza infection in the5 internal medicine wards of Meir Medical Center from January 1, 2012 through December 31, 2015. Meir Medical Center is a 760-bed, tertiary care university hospital. Influenza was diagnosed using polymerase chain reaction kit (3 M integrated cycler, Focus Diagnostics, Cypress, CA, USA). During the study period, RDW and other laboratory tests were analyzed in a single central laboratory using standardized automated kits (Advia 2120, Siemens, Erlangen, Germany). Data were collected from electronic medical records. Mortality data were confirmed using the Israel Central Bureau of Statistics. Data collection was approved by the hospital ethics committee. 2.2. Definitions For analysis, patients were categorized according to RDW values during hospitalization: high RDW (at least one RDW N 14.5%) and low RDW (no RDW value N 14.5%). The cutoff of 14.5% was determined as the upper limit of normal values provided by our laboratory. Primary
G. Topaz et al. / Journal of Critical Care 41 (2017) 166–169
167
outcome was complicated hospitalization, defined as at least one of the following: hospital stay N 7 days, need for mechanical ventilation, septic shock, transfer to the Intensive Care Unit (ICU) and 30-day mortality. Secondary outcome was 30-day, all-cause mortality. Patients were followed until December 31, 2015 for all-cause mortality. Patients were further subdivided according to the change in RDW levels during hospitalization (discharge RDW minus baseline RDW). A significant change was defined as more than ± 0.6%.
analysis (mean age 62.5 ± 1 years, 53.6%men). Eighty-four (55%) patients were categorized in the high RDW group and 69 (45%) in the low RDW group. Patients in the two groups were of similar age and had similar comorbidity profiles. The mean hemoglobin level at admission was lower and the mean creatinine level was higher for patients in the high RDW than in the low RDW group (11.2 ± 1.8 vs. 12.9 ± 1.3, p = 0.00 and 1.9 ± 2.0 vs. 1.2 ± 0.6, p b 0.01, respectively) (Table 1).
2.3. Statistical analysis
3.1. RDW and in-hospital complications
Data are presented as numbers and percentages for nominal data and as means and standard deviations for continuous parameters. Differences between the high RDW and low RDW groups were analyzed with chi-square or Fisher's exact test, as appropriate. Continuous variables were examined for normality (Shapiro-Wilk test) and data were analyzed accordingly. The t-test was used for normally distributed variables and the Mann-Whitney for non-parametric variables. Receiver operating characteristic (ROC) curves were used to evaluate the ability of maximum RDW and the change in RDW to predict unfavorable outcomes. A logistic regression model was applied to estimate odds ratios for complicated hospitalization. In a multivariate logistic regression model, the following variables were accounted forage, gender, high RDW, change in RDW N ±0.6% and hemoglobin levels at admission. A Cox proportional hazard model was applied to estimate hazard ratios for all-cause mortality until the end of the follow-up period. The variables accounted for in this model were diabetes mellitus, creatinine at admission, hemoglobin at admission, high RDW and complicated hospitalization. p b 0.05 was considered statistically significant. Data were analyzed with SPSS Version 21 (IBM Corporation, Armonk, NY, USA).
Complicated hospitalization rates were significantly higher in the high than in the low RDW group (32.5% vs. 10.3%, p b 0.01). In-hospital characteristics and complications are depicted in Table 2. ROC curve analysis revealed that the maximum RDW of 14.5% was the optimal cut-off value associated with complicated hospitalization. Risk assessment using RDW level as a continuous variable demonstrated that each 1-point increase in RDW was associated with a 29% increase in the risk for the primary endpoint of complicated hospitalization. A multivariate logistic regression model revealed that high RDW was the only factor examined that was an independent predictor of complicated hospitalization (OR 5.03, 95% CI 1.81 to 13.93, p b 0.01). A sub-analysis of the 35 patients with complicated hospitalization revealed a high RDW value in 27 (77%). The mean maximum RDW of patients with complicated hospitalization was significantly higher than that of patients with a non-complicated hospitalization (15.9 ± 1.9 vs. 14.9 ± 1.8, p b 0.01). The same trend was noted when comparing patients with and without complicated hospitalization, with mean RDW levels at baseline and at discharge higher in the complicated hospitalization group (14.8 ± 1.6 vs. 14.4 ± 1.6, p = 0.2 and 15.3 ± 2.0 vs. 14.8 ± 1.8, p = 0.13, respectively).
3. Results Among 162 patients hospitalized with a diagnosis of influenza during the study period, 9 were excluded from the analysis due to incomplete data. A total of 153 influenza patients were included in this
Table 1 Characteristics of patients hospitalized with influenza virus infection. Variables
RDW N 14.5% (n = 84)
RDW ≤ 14.5% (n = 69)
Total (n = 153)
p-Value
Age Male gender Comorbidities Asthma COPD Diabetes mellitus Hypertension Hyperlipidemia CRF Tobacco use Baseline laboratory values Hemoglobin (g/dl) WBC (K/microL) PLT (K/microL)
64 ± 12.7 51.2
60.9 ± 15.7 56.5
62.5 ± 14.2 53.6
0.19 0.52
4.8 7.1 27.4 28.6 28.6 13.1 13.1
7.2 2.9 20.3 30.4 37.7 5.8 14.5
5.9 5.2 24.2 29.4 32.7 9.8 13.7
0.73 0.29 0.31 0.8 0.23 0.17 0.8
Creatinine (g/dL) Bilirubin (mg/dL) CRP (mg/dL) Baseline RDW (%)
11.2 ± 1.8
12.9 ± 1.3
11.9 ± 1.8
0.00
6.8 ± 4.8 186 ± 88
6.8 ± 3.1 193 ± 62
0.89 0.56
1.9 ± 2.0 0.61 ± 0.35 11.2 ± 7.9 15.3 ± 1.6
1.2 ± 0.6 0.68 ± 0.68 9.8 ± 6.1 13.4 ± 0.6
6.8 ± 4.1 189.4 ± 77.3 1.36 ± 1.4 0.64 ± 0.5 9.9 ± 6.1 14.5 ± 1.6
0.00 0.43 0.31 0.00
WBC, white blood cells; PLT, platelets; RDW, red blood cell distribution width; COPD, chronic obstructive pulmonary disease; CRF, chronic renal failure. Data are presented as means ± SD or percentages of presented cases. Normal ranges of values of laboratory data: Hemoglobin (12–16), WBC (4.8–10.8), PLT (150–400), creatinine (0.5–1.2), bilirubin (0.2–1.5), CRP (0.0–0.5).
3.2. RDW and mortality Thirty-day mortality rates were higher for patients with a high RDW value than for those without (2.4% vs. 0%, p = 0.5). Two patients died due to complications related to influenza infection. Four patients (2.7%) died within 90 days due to influenza complications, 3 had a high RDW value (3.6% vs. 1.5%, p = 0.63). During a mean follow-up of 21 ±10 months, 24 patients (15.9%) died. A high RDW was associated with higher long-term all-cause mortality (21.7% vs. 8.8%, p = 0.04). However, a multivariate Cox regression model analysis demonstrated that only complicated hospitalization, and not high RDW, was
Table 2 In-hospital characteristics and outcomes of patients hospitalized with influenza virus infection according to RDW levels. Variables
RDW N 14.5% (n = 84)
RDW ≤ 14.5% (n = 69)
Total (n = 153)
p-Value
Acute renal failure Septic shock Syncope COPD exacerbation Max fever, °C ICU transfer Mechanical ventilation Oseltamivir treatment Hospitalization length, days 30-day mortality rate
10.7 3.6 2.4 4.8 38.1 ± 0.88 4.9 2.4
8.8 0 5.9 0 38.0 ± 0.9 2.9 0
9.9 2.0 3.9 2.6 38.0 ± 0.8 4.0 1.3
0.7 0.25 0.41 0.13 0.03 0.69 0.5
76.2
77.9
77.0
0.08
6.4 ± 5.9
4.6 ± 2.9
5.6 ± 4.9
0.02
2.4
0
1.3
0.5
COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; RDW, red blood cell distribution width; Max, maximum. Data are presented as means ± SD or percentages of presented cases.
168
G. Topaz et al. / Journal of Critical Care 41 (2017) 166–169
Table 3 Cox proportional hazard model for long-term mortality among patients with influenza, adjusted for all significant variables in univariate analysis. Variables
OR
Diabetes mellitus Creatinine admission Hemoglobin Max RDW N 14.5 Complicated hospitalization
2.1 1.09 0.79 1.16 3.26
95% CI
p-Value
Lower
Upper
0.87 0.90 0.60 0.36 1.32
5.12 1.31 1.04 3.72 8.07
0.10 0.37 0.09 0.80 0.01
RDW, red blood cell distribution width.
associated with long-term mortality (HR 3.26, 95% CI 1.32–8.07, p = 0.01) (Table 3). Change in RDW levels during the index hospitalization. We examined an association between the change in RDW levels during the index hospitalization and the primary outcome of complicated hospitalization. 1) Baseline to maximum values. An increase in RDW of N 1.5% between the baseline RDW and a consecutive measure was significantly associated with complicated hospitalization (OR 3.5, 95% CI 1.57–7.79, p b 0.01).This increase was not a significant independent predictor of complicated hospitalization in multivariate analysis (Fig. 1). 2) Baseline to discharge values. An absolute difference in RDW (N± 0.6%) was significantly associated with complicated hospitalization (OR 3.4, 95% CI 1.26–8.97, p = 0.01); however, this change was not a significant independent predictor of complicated hospitalization in multivariate analysis. Interestingly, the mean change in RDW during hospitalization was greater for patients with complicated hospitalization (0.86 ± 0.9 vs. 0.5 ± 0.6, p = 0.04) (Fig. 2). 4. Discussion In this study of Influenza patients hospitalized in internal medicine wards, our main findings were: 1) A RDW N 14.5% at any time during hospitalization was independently associated with a complicated hospital course and showed a trend for increased short-term mortality. 2) A significant absolute change of N0.6%in RDW during hospitalization was associated with a complicated hospitalization; however, in contrast to an elevated RDW level, this was not statistically significant in multivariate analysis. 3) Among patients with a complicated hospitalization, the mean RDW level was higher and the mean change in RDW during hospitalization was greater, compared to patients without a complicated hospitalization. 4) An elevated RDW level during hospitalization was
not associated with long-term mortality, with a mean follow-up period of 21 ± 10 months. Seasonal influenza infection may cause severe illness and can be fatal. The clinical course of the disease is mainly dependent on host and viral characteristics [1,11]. Early administration of neuraminidase inhibitors, such as oseltamivir (75 mg twice daily), has been shown to improve outcomes among influenza patients [12]. According to the World Health Organization, severe illness may require the use of higher doses (150 mg twice daily) of oseltamivir [12]. Therefore, early recognition of the development of complications is important. Though not a specific marker, elevated RDW has been shown to predict worse prognosis in patients with chronic medical conditions [3-6] and acute infectious diseases [9,10,13,14]. In acute infectious states, RDW levels have been shown to predict mortality when measured at different intervals from hospital admission: baseline [10,15], day 3 [16] and day 8 [9]. In our study, RDW level N 14.5% measured at any point during the index hospitalization was strongly associated with severe complications in patients with influenza. This finding, as far as we know, has not been previously described. Our findings support the predictive value of a significant change in RDW for adverse outcomes. Interestingly, a fall or rise in RDW of N 0.6%, regardless of baseline RDW values, was associated with a complicated course among influenza patients. This finding supports a study that showed a rise in RDW of 0.4% throughout hospitalization was associated with in-hospital mortality among heterogeneous patients hospitalized in an internal medicine ward [17]. Gorelik et al. recently showed that among 980 communityacquired pneumonia patients, even a smaller absolute RDW change of ± 0.4% was associated with worse short-term outcomes [18]. We believe that serial RDW measurements may be informative, in addition to baseline RDW levels. We demonstrated that high RDW is associated with a positive trend for 30-day mortality; yet is not significantly associated with longer term all-cause mortality. In contrast, prior studies showed baseline RDW to be associated with worse outcomes in the settings of sepsis [9,14], pneumonia [10,15] and endocarditis [13]. Moreover, Patel et al. found that RDW is a strong predictor of mortality in the general population of adults 45 years and older [19]. The exact mechanism by which RDW is associated with short-term poor outcomes in this scenario is not yet understood. It is possible that elevated RDW, which reflects deregulation of erythrocyte homeostasis with impaired red blood cell production [20], can serve as a marker of inflammation and oxidative state [21,22]. Elevated RDW could be related to interactions among various factors, such as erythropoietin, proinflammatory cytokines and the hematopoietic response to erythropoietin [23]. RDW seems to reflect general health status.
23
21
19
17
15
13
11 Baseline RDW
Maximal RDW
Baseline RDW
Maximal RDW
Fig. 1. Baseline and maximum mean RDW in patients with complicated (n = 35) and non-complicated (n = 118) hospitalizations.
G. Topaz et al. / Journal of Critical Care 41 (2017) 166–169
169
23
RDWLevel(%)
21
19
17
15
13
11 Baseline RDW
RDW at discharge
Baseline RDW RDW at discharge
Fig. 2. Baseline and discharge mean RDW in patients with complicated (n = 35) and non-complicated (n = 118) hospitalizations.
Therefore, influenza patients with elevated RDW were presumably in poor general condition. Patients with high RDW levels had lower hemoglobin levels, which could influence RDW levels. However, a multivariate analysis model, including both factors, demonstrated that high RDW, and not hemoglobin levels, was associated with complicated hospitalization. This study had a few limitations. The retrospective design precluded determining whether a causal relationship exists between RDW and complicated hospitalization. Furthermore, measuring RDW levels during hospitalization and after medical intervention may influence the levels of RDW. Finally, the cohort was small, resulting in a lack of statistical power. Thus, a larger prospective study should be conducted. 5. Conclusion RDW N 14.5% is associated with a complicated hospitalization in patients with influenza. Combined with other known risk factors, RDW may serve as a simple tool to improve risk stratification among patients with influenza. Conflict of interests None. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] Nichol KL. Influenza vaccination in the elderly: impact on hospitalization and mortality. Drugs Aging 2005;22:495–515. [2] Simpson CR, Lone N, Kavanagh K, et al. Seasonal influenza vaccine effectiveness community-based national-linked database to determine the effectiveness of the seasonal trivalent influenza vaccine. Southampton (UK): NIHR Journals Library. Health Services and Delivery Research; 2013 Nov. [3] Khaki S, Mortazavi SH, Bozorgi A, et al. Relationship between red blood cell distribution width and mortality of patients with acute myocardial infarction referring to Tehran Heart Center. Crit Pathw Cardiol 2015;14:112–5. [4] Sahin O, Akpek M, Sarli B, et al. Association of red blood cell distribution width levels with severity of coronary artery disease in patients with non-ST elevation myocardial infarction. Med Princ Pract 2015;24:178–83.
[5] Topaz G, Haim M, Kusniec J, et al. Association between red cell distribution width and mortality after cardiac resynchronization therapy. Isr Med Assoc J 2015;17: 505–9. [6] Ani C, Ovbiagele B. Elevated red blood cell distribution width predicts mortality in persons with known stroke. J Neurol Sci 2009;277:103–8. [7] Lippi G, Targher G, Montagnana M, et al. Relationship between red blood cell distribution width and kidney function tests in a large cohort of unselected outpatients. Scand J Clin Lab Invest 2008;68:745–8. [8] Seyhan EC, Özgül MA, Tutar N, et al. Red blood cell distribution and survival in patients with chronic obstructive pulmonary disease. COPD 2013;10:416–24. [9] Leonardo L, María MM, Martín, et al. Red blood cell distribution width during the first week is associated with severity and mortality in septic patients. PLoS One 2014;9:e105436. [10] Lee JH, Chung HJ, Kim K, et al. Red cell distribution width as a prognostic marker in patients with community-acquired pneumonia. Am J Emerg Med 2013;31:72–9. [11] Uraki R, Kiso M, Shinya K, et al. Virulence determinants of pandemic A(H1N1)2009 influenza virus in a mouse model. J Virol 2013 Feb;87(4):2226–33. [12] World Health Organization. WHO guidelines for pharmacological management of pandemic influenza A(H1N1) 2009 and other influenza viruses. Geneva: World Health Organization; 2010 Feb. Available from http://www.who.int/csr/resources/ publications/swineflu/h1n1_guidelines_pharmaceutical_mngt.pdf?ua=1. [13] Guray Y, Ipek EG, Guray U, et al. Red cell distribution width predicts mortality in infective endocarditis. Arch Cardiovasc Dis 2014;107:299–307. [14] Mahmood NA, Mathew J, Kang B, et al. Broadening of the red blood cell distribution width is associated with increased severity of illness in patients with sepsis. Int J Crit Illn Inj Sci 2014;4:278–82. [15] Braun E, Domany E, Kenig Y, et al. Elevated red cell distribution width predicts poor outcome in young patients with community acquired pneumonia. Crit Care 2011; 15:R194. [16] Ku NS, Kim HW, Oh HJ, et al. Red blood cell distribution width is an independent predictor of mortality in patients with gram-negative bacteremia. Shock 2012;38: 123–7. [17] Shteinshnaider M, Barchel D, Almoznino-Sarafian D, et al. Prognostic significance of changes in red cell distribution width in an internal medicine ward. Eur J Intern Med 2015;26:616–22. [18] Gorelik O, Izhakian S, Barchel D, et al. Changes in red cell distribution width during hospitalization for community-acquired pneumonia: clinical characteristics and prognostic significance. Lung 2016;194:985–95. [19] 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;169:515–23. [20] Salvagno GL, Sanchis-Gomar F, Picanza A, et al. Red blood cell distribution width: a simple parameter with multiple clinical applications. Crit Rev Clin Lab Sci 2015;52: 86–105. [21] 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;133:628–32. [22] Pierce CN, Larson DF. Inflammatory cytokine inhibition of erythropoiesis in patients implanted with a mechanical circulatory assist device. Perfusion 2005;20:183. [23] Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med 2005;352: 1011–23.