Red cell distribution width (RCDW) has not been fully investigated for its prognostic impact in patients with acute heart failure (AHF) with or without the cardiorenal anemia syndrome (CRAS). A total of 978 patients (age 75 ± 14 years, 70% men, 43% with CRAS) hospitalized for AHF were enrolled. During a median follow-up duration of 31 months, 472 subjects (48%) died. The postdischarge mortality was positively associated with the increasing RCDW. After accounting for age, gender, co-morbidities, hemoglobin, renal function, sodium level, and N-terminal probrain natriuretic peptide, RCDW remained an independent predictor of mortality (hazard ratio [HR] and 95% CI for a 1% increase of RCDW: 1.09, 1.00 to 1.17, p = 0.04). In the subgroups of patients with or without CRAS, RCDW was an independent predictor of total mortality for both subgroups (HR 1.05, 95% CI 1.00 to 1.10 and HR 1.11, 95% CI 1.07 to 1.15, respectively). In conclusion, elevated RCDW was independently associated with mortality in patients hospitalized for AHF, with or without CRAS.
Impaired renal function and anemia are common in patients with heart failure (HF), and they have been associated with higher mortality and morbidity. Red cell distribution width (RCDW) is related to anemia, and Felker et al for the first time have shown that increased RCDW is associated with increased mortality and morbidity in patients with chronic HF (CHF). Higher levels of RCDW also correlate with poorer survival in acute HF (AHF), stroke, and myocardial infarction. Moreover, the prognostic values of RCDW in patients with HF were independent of anemia status. Given the close interplay between the heart, kidneys, and anemia, the prognostic impacts of RCDW in patients with HF and different stages of chronic kidney disease (CKD) have not yet been clarified. We therefore investigated the value of RCDW as a prognostic marker in patients with cardiorenal anemia syndrome (CRAS).
Since September 2003, clinical data of patients who had been hospitalized in a tertiary medical center with a discharge diagnosis of AHF were comprehensively registered in an electronic database (Heart Failure Registry of Taipei Veterans General Hospital, Heart FAilure Registry of Taipei VEteranS General HospiTal [HARVEST] registry). Up to December 2012, a total of 978 AHF with complete hematologic and biochemistry data constituted this study population. The institutional review committee on human research approved the use of the registry data for research purposes. Clinical outcomes and mortality were ascertained using the National Death Registry. Subjects who did not appear in the National Death Registry on censoring dates (December 31, 2012) were considered to be survivors.
Blood samples were obtained before discharge and analyzed using a Beckman Coulter LH 780 analyzer for hematologic data and a Roche cobas 8000 for biochemical data. Anemia was defined according to World Health Organization criteria: hemoglobin <13 g/dl for men and <12 g/dl for women. CRAS is defined as the presence of HF, anemia, and estimated glomerular filtration rate (eGFR) <60 ml/min/1.73 m 2 . Renal function was expressed as eGFR and determined by the abbreviated modification of diet in renal disease formula. Echocardiograms were performed by experienced technicians and independently reviewed by the physicians. The left ventricular ejection fraction (LVEF) was derived from the 2-dimensional–guided M-mode echocardiography.
Continuous variables are presented as the mean and standard deviation, and categorical variables are presented as numbers and percentages. The 2-tailed independent Student t test was used for normally distributed parameters. The Mann–Whitney U test was used for nonnormally distributed parameters, and the chi-square test was used for categorical variables. Spearman’s correlation coefficient was used to evaluate the relation between RCDW and other parameters. Multivariate linear regression included all variables with p <0.1 by means of a stepwise multiple regression model. The variance inflation factor was used to measure collinearity in predictor variables. The variance inflation factor values >10 indicated collinearity in independent variables.
RCDW was categorized according to tertiles of distribution, with the lowest tertile as the reference group. The Kaplan–Meier method was used for survival curve analysis and comparisons were made using the log-rank test. Stepwise Cox proportional hazard models with forward selection were used to calculate the hazard ratio (HR) and 95% CIs for RCDW categories. The Cox model for independent predictors of 5-year mortality was adjusted for age, gender, medical history of old myocardial infarction, diabetes mellitus (DM), medical therapy (including β blockers, angiotensin-converting enzyme inhibitor/angiotensin receptor blockade (ACEI/ARB), and spironolactone), hemoglobin, eGFR, sodium, and N-terminal of the prohormone brain natriuretic peptide (NT-proBNP). NT-proBNP was log transformed before entry because of its nonnormal distribution. Differences were considered statistically significant at the 2-sided p <0.05 level. Statistical analyses were performed using the IBM SPSS software, version 21.0, (SPSS Inc., Chicago, Illinois).
Of a total of 978 subjects, 680 subjects (70%) were men and the median age was 78.5 years (interquartile range 69.2 to 83.6 years). The median follow-up was 31 months (interquartile range 10.7 to 54.8 months). Overall, 38% of patients had DM, 61% of patients had anemia, and 63% of patients were in CKD stage III to V. Reduced LV systolic function defined by an LVEF <50% was found in 40% of the study population. ACEI/ARB and beta-blocker usage were 82% and 63%, respectively. The baseline characteristics of patients divided according to tertiles of RCDW is detailed in Table 1 . Across tertiles of RCDW distribution, a high RCDW was associated with lower LVEF, an increased right ventricular systolic pressure, lower hemoglobin levels, and higher serum creatinine levels.
|Variable||Red Cell Distribution Width||P value|
(n = 325)
|>14.3 ≤ 15.9 |
(n = 327)
|> 15.9 |
(n = 326)
|Coronary artery disease||119(35%)||104(32%)||87(28%)||0.114|
|White blood cell count (x10 9 /L)||7.8±3.1||7.3±2.9||7.0±3.0||0.002|
|Platelet (x10 9 /L)||215±79||195±84||196±98||0.002|
|Blood urea nitrogen (mg/dl)||30±17||33±19||39±24||<0.001|
|Estimated glomeruler filtration rate (ml/min/1.73 m 2 )||56±28||53±29||51±30||0.109|
|High-density lipoprotein cholesterol (mg/dl)||44±14||44±14||39±15||<0.001|
|Low-density lipoprotein cholesterol (mg/dl)||100±35||96±33||85±31||<0.001|
|Uric acid (mg/dl)||8±3||9±3||9±3||0.004|
|C-reactive protein (mg/dl)||3.4±4.6||3.4±4.7||3.4±4.7||0.991|
|Log(N-terminal of the prohormone brain natriuretic peptide (pg/ml))||3.6±0.6||3.8±0.6||3.7±0.6||0.034|
|Left ventricular ejection fraction (%)||58±18||53±23||52±20||<0.001|
|Right ventricular systolic pressure (mm Hg)||40±15||45±16||48±17||<0.001|
The correlations between hematologic and biochemical variables and RCDW are provided in Table 2 . Multiple linear regression showed that hypertension, right ventricular systolic pressure, and hemoglobin were independently correlated to RCDW ( Table 2 ).
|Univariate analysis||Multivariate analysis ∗|
|Beta coefficient||P value||Beta coefficient||P value|
|Left ventricular ejection fraction||-0.061||0.057|
|Right ventricular systolic pressure||0.190||<0.001||0.181||0.025|
|White blood cell count||-0.06||0.061|
|Blood urea nitrogen||0.167||<0.001|
|Estimated glomeruler filtration rate||-0.049||0.126|
|Log(N-terminal of the prohormone brain natriuretic peptide)||0.088||0.228|
During the median follow-up duration of 31 months, 472 patients (48%) died. As detailed in Table 3 , several univariate predictors of mortality are provided. In multivariate Cox proportional hazard analysis, age, the use of ACEI/ARB, RCDW, and the serum sodium level were independently associated with a 5-year survival rate in patients with AHF. Of these predictor factors, RCDW showed a strong predictor power (HR 1.09 per 1% increase in RCDW, 95% CI 1.00 to 1.17; p = 0.04; Table 3 ). A Kaplan–Meier analysis showed a graded increased probability of death across the tertiles of RCDW (log-rank p <0.001; Figure 1 ).
|HR ∗||P value||HR †||P value||HR ‡||P value|
|Age||1.02 (1.01-1.03)||<0.01||1.02 (1.02-1.03)||<0.01||1.04 (1.01-1.07)||<0.01|
|Male||1.13 (0.93-1.38)||0.23||1.23 (0.99-1.51)||0.06||1.58 (0.82-3.01)||0.17|
|Diabetes mellitus||1.24 (1.03-1.49)||0.02||1.34 (1.10-1.62)||<0.01|
|Beta-blocker||0.70 (0.58-0.84)||<0.01||0.78 (0.64-0.96)||0.02|
|Angiotensin converting- enzyme inhibitor/Angiotensin receptor blockade||0.65 (0.52-0.82)||<0.01||0.72 (0.56-0.92)||0.01||0.28 (0.14-0.59)||<0.01|
|Hemoglobin||0.90 (0.87-0.94)||<0.01||0.97 (0.93-1.02)||0.26||0.93 (0.82-1.06)||0.30|
|Red cell distribution width||1.09 (1.06-1.12)||<0.01||1.10 (1.06-1.13)||<0.01||1.09 (1.00-1.17)||0.04|
|Estimated glomeruler filtration rate||0.99 (0.99-0.99)||<0.01||0.99 (0.99-1.00)||<0.01|
|Sodium||0.98 (0.96-1.00)||0.025||0.98 (0.96-1.00)||0.03||0.91 (0.86-0.96)||<0.01|
|Log(N-terminal of the prohormone brain natriuretic peptide)||1.76 (1.14-2.73)||0.01||1.37 (0.82-2.29)||0.23|