Effectiveness of the Relative Lymphocyte Count to Predict One-Year Mortality in Patients With Acute Heart Failure

Effectiveness of the Relative Lymphocyte Count to Predict One-Year Mortality in Patients With Acute Heart Failure Julio Núñez, MDa,*, Eduardo Núñez, MD, MPHa, Gema Miñana, MDa, Juan Sanchis, MDa, Vicent Bodí, MDa, Eva Rumiz, MDa, Patricia Palau, MDa, Myriam Olivares, MDa, Pilar Merlos, MDa, Clara Bonanad, MDa, Luis Mainar, MDb, and Angel Llàcer, MDa Several works have endorsed a significant role of the immune system and inflammation in the pathogenesis of heart failure. As indirect evidence, an association between a low relative lymphocyte count (RLC%) and worse outcomes found in this population has been suggested. Nevertheless, the role of RLC% for risk stratification in a large and nonselected population of patients with acute heart failure (AHF) has not yet been determined. Thus, the aim of this study was to determine the association between low RLC% and 1-year mortality in patients with AHF and consequently to define whether it has any role for early risk stratification. A total of 1,192 consecutive patients admitted for AHF were analyzed. Total white blood cell and differential counts were measured on admission. RLC% (calculated as absolute lymphocyte count/total white blood cell count) was categorized in quintiles and its association with all-cause mortality at 1 year assessed using Cox regression. At 1 year, 286 deaths (24%) were identified. A negative trend was observed between 1-year mortality rates and quintiles of RLC%: 31.5%, 27.2%, 23.1%, 23%, and 15.5% in quintiles 1 to 5, respectively (p for trend <0.001). After thorough covariate adjustment, only patients in the lowest quintile (<9.7%) showed an increased risk for mortality (hazard ratio 1.76, 95% confidence interval 1.17 to 2.65, p ⴝ 0.006). When RLC% was modeled with restricted cubic splines, a stepped increase in risk was observed patients in quintile 1: those with RLC% values <7.5% and <5% showed 1.95- and 2.66-fold increased risk for death compared to those in the top quintile. In conclusion, in patients with AHF, RLC% is a simple, widely available, and inexpensive biomarker, with potential for identifying patients at increased risk for 1-year mortality. © 2011 Elsevier Inc. All rights reserved. (Am J Cardiol 2011;107:1034 –1039) Current evidence supports a pathogenic role of the immune system and inflammation in the pathogenesis of heart failure (HF).1– 4 A low relative lymphocyte count (RLC%) has been related to worse outcomes in selected patients with HF.5–9 In the present study, we sought to determine the association between low RLC% assessed on admission and 1-year all-cause mortality in patients consecutively admitted with diagnoses of acute HF (AHF). Methods We prospectively studied a cohort of 1,268 patients, consecutively admitted to the cardiology department at Hospital Clínico Universitario de Valencia from January 1, 2004, to July 1, 2009, with diagnoses of AHF. AHF was defined as the rapid onset of symptoms and signs secondary to abnormal cardiac function and the presence of objective evidence of

a Servicio de Cardiología, Hospital Clínico Universitario, INCLIVA, Universitat de Valencia; and bServicio de Cardiología, Hospital de Manises, Valencia, Spain. Manuscript received August 31, 2010; revised manuscript received and accepted November 12, 2010. This study was supported by a grant from Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III, RED HERACLES RD06/0009/ 1001, Madrid, Spain, Beca Fundación Española del Corazón Basic Research and PI080128. *Corresponding author: Tel: 34-652856689; fax: 34-963862658. E-mail address: [email protected] (J. Núñez).

0002-9149/11/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2010.11.029

structural or functional abnormalities of the heart at rest (such as cardiomegaly, third heart sound, cardiac murmur, abnormality on echocardiography, or increased natriuretic peptides).10,11 Patients with clinical evidence of cancer (n ⫽ 18), hematologic disturbances (n ⫽ 12), chronic inflammatory disease (n ⫽ 11), pneumonia (n ⫽ 16), sepsis (n ⫽ 8), and treatment with corticosteroids (n ⫽ 11) were excluded, leaving a final study sample of 1,192 patients. Total white blood cell (WBC), neutrophil, lymphocyte, and monocyte counts were obtained on admission, using an automated blood cell counter. RLC% was calculated as the ratio between lymphocyte count and total WBC count. Follow-up was limited to 1 year, and patients were censored if they died, underwent cardiac valve replacement, or underwent cardiac transplantation within this period. All-cause mortality and cardiovascular (CV) mortality were chosen as the main and secondary end points, respectively. Survival status was ascertained either during the patient’s hospitalization, by phone contact with the patient or family members, or by routine clinic visits. Information about the cause of death was extracted from the patient’s clinical chart, adjudicated by an investigator who was blinded to lymphocyte values, and further categorized according to the American Heart Association classification.12 Deaths were considered not CV in origin if a specific non-CV cause was identified. Otherwise, CV origin was considered and included sudden death, progressive HF death, deaths attributable to other CV causes (such as myocardial infarction, stroke, etc.), and unknown causes of death. For the present study, www.ajconline.org

Heart Failure/RLC% and Long-Term Mortality in AHF

1035

Table 1 Baseline characteristics stratified by relative lymphocyte count quintiles Variable Age (years) Women Previous admission for AHF Hypertension Dyslipidemia Current smoker Previous smoker Ischemic heart disease Valvular heart disease Acute decompensated HF Acute pulmonary edema Chronic pulmonary obstructive disease Stroke Peripheral artery disease Radiologic pleural effusion Peripheral edema Previous use of diuretics Previous use of ␤ blockers Previous use of angiotensin converting enzyme inhibitors/ angiotensin II receptor blockers Previous use of statins Heart rate (beats/min) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Atrial fibrillation QRS interval ⬎120 ms Hemoglobin (g/dl) Serum creatinine (mg/dl) Uric acid (mg/dl) Sodium (mEq/L) Troponin I* (ng/ml) Troponin I ⬎0.2 ng/ml Brain natriuretic peptide* (pg/ml) Antigen carbohydrate 125* (U/ml) Echocardiography Left ventricular ejection fraction ⱕ50% Left atrial diameter (mm) Left ventricular diastolic diameter (mm) Septum (mm) Posterior wall (mm) Medications ␤ blockers Diuretics Spironolactone Angiotensin converting enzyme inhibitors Angiotensin II receptor blockers Statins Oral anticoagulants Nitrates Digoxin

Q1 (1.5%–9.7%) (n ⫽ 239)

Q2 (9.7%–14%) (n ⫽ 238)

Q3 (14.1%–18.4%) (n ⫽ 239)

Q4 (18.4%–24.6%) (n ⫽ 238)

Q5 (24.7%–62.7%) (n ⫽ 238)

p Value for Trend

75 ⫾ 9 125 (52.5%) 107 (45%) 191 (80.2%) 99 (41.6%) 24 (10.1%) 42 (17.6%) 99 (41.6%) 70 (29.4%) 163 (68.5%) 59 (24.8%) 55 (23.1%)

74 ⫾ 11 119 (49.8%) 88 (36.8%) 184 (77%) 104 (43.5%) 21 (8.8%) 58 (24.3%) 91 (38.1%) 64 (26.8%) 173 (72.4%) 46 (19.2%) 58 (24.3%)

73 ⫾ 12 125 (52.5%) 96 (40.3%) 182 (76.5%) 100 (42%) 24 (10.1%) 45 (18.9%) 70 (29.4%) 77 (32.3%) 176 (73.9%) 35 (14.7%) 50 (21%)

72 ⫾ 12 119 (49.8%) 84 (35.1%) 176 (73.6%) 99 (41.4%) 27 (11.3%) 38 (15.9%) 87 (36.4%) 73 (30.5%) 181 (75.7%) 38 (15.9%) 44 (18.4%)

71 ⫾ 12 118 (49.6%) 70 (29.4%) 188 (79%) 112 (47.1%) 28 (11.8%) 40 (16.8) 108 (45.4%) 56 (23.5%) 127 (53.4%) 77 (32.3%) 43 (18.1%)

0.001 0.566 0.001 0.494 0.385 0.348 0.208 0.555 0.389 0.005 0.162 0.056

28 (11.8%) 16 (6.7%) 110 (46.2%) 139 (58.4%) 165 (68.3%) 45 (18.9%) 109 (45.8%)

31 (13%) 16 (6.7%) 122 (51%) 139 (58.2%) 149 (62.3%) 57 (23.8%) 115 (48.1%)

21 (8.8%) 16 (6.7%) 115 (48.3%) 149 (62.6%) 156 (65.5%) 63 (26.5%) 112 (47.1%)

20 (8.4%) 10 (4.2%) 103 (43.1%) 134 (56.1%) 140 (58.6%) 67 (28%) 105 (43.9%)

18 (7.6%) 23 (9.7%) 78 (32.8%) 106 (44.5%) 131 (55%) 56 (23.5%) 101 (42.4%)

0.034 0.515 0.001 0.003 0.001 0.126 0.285

56 (23.5%) 99 ⫾ 27 144 ⫾ 34 78 ⫾ 18 106 (44.5%) 65 (27.3%) 12.4 ⫾ 1.8 1.43 ⫾ 0.68 7.9 ⫾ 2.6 137 ⫾ 5 0 (0.43) 81 (34%) 254 (328) 58.3 (109.4)

61 (25.5%) 100 ⫾ 27 148 ⫾ 34 81 ⫾ 18 111 (46.4%) 66 (27.6%) 12.5 ⫾ 1.8 1.34 ⫾ 0.68 8.1 ⫾ 2.6 139 ⫾ 5 0 (0.32) 68 (28.4%) 266 (328) 81.8 (128.2)

65 (27.3%) 101 ⫾ 31 149 ⫾ 36 83 ⫾ 22 115 (48.3%) 60 (25.2%) 12.7 ⫾ 1.8 1.25 ⫾ 0.45 7.8 ⫾ 2.4 139 ⫾ 5 0 (0.22) 65 (27.3%) 285 (339) 65.3 (110.5)

69 (28.9%) 100 ⫾ 30 145 ⫾ 34 81 ⫾ 20 108 (45.2%) 63 (26.4%) 12.6 ⫾ 1.7 1.26 ⫾ 0.49 8 ⫾ 2.4 140 ⫾ 4 0 (0.22) 62 (25.9%) 270 (370) 75.8 (115)

72 (30.2%) 107 ⫾ 31 162 ⫾ 38 89 ⫾ 21 90 (37.8%) 77 (32.3%) 13.2 ⫾ 2 1.26 ⫾ 0.48 7.7 ⫾ 2.1 139 ⫾ 4 0 (0.40) 77 (32.3%) 257 (370) 42.6 (94.6)

0.065 0.072 ⬍0.001 ⬍0.001 0.149 0.337 ⬍0.001 0.053 0.503 ⬍0.001 0.275 0.530 0.931 0.122

107 (45%)

101 (42.3%)

115 (48.3%)

128 (53.6%)

124 (52.1%)

0.012

43 ⫾ 8 54 ⫾ 9

44 ⫾ 8 54 ⫾ 9

45 ⫾ 9 56 ⫾ 9

44 ⫾ 7 57 ⫾ 10

43 ⫾ 7 57 ⫾ 9

11.6 ⫾ 2.2 11.3 ⫾ 1.8

11.5 ⫾ 2.4 11.2 ⫾ 2.2

11.6 ⫾ 2.8 11.4 ⫾ 2.2

11.2 ⫾ 2.2 11.2 ⫾ 1.9

11.3 ⫾ 2.4 11.3 ⫾ 2.1

0.175 0.768

110 (46.2%) 235 (98.7%) 43 (18.1%) 108 (45.4%)

113 (47.3%) 233 (97.5%) 41 (17.1%) 97 (40.6%)

128 (53.8%) 234 (98.3%) 39 (16.4%) 105 (44.1%)

140 (58.6%) 235 (98.3%) 50 (20.9%) 91 (38.1%)

103 (54.6%) 233 (97.9%) 49 (20.6%) 113 (47.5%)

0.006 0.761 0.269 0.869

50 (21%) 83 (34.9%) 98 (41.2%) 58 (24.4%) 68 (28.6%)

68 (28.4%) 76 (31.8%) 108 (45.2%) 53 (22.2%) 59 (24.7%)

77 (32.3%) 72 (30.2%) 96 (40.3%) 33 (13.9%) 66 (27.7%)

74 (31%) 90 (37.7%) 99 (41.4%) 45 (18.8%) 61 (25.5%)

66 (27.7%) 100 (42%) 79 (33.2%) 49 (20.6%) 48 (20.2%)

0.083 0.040 0.050 0.183 0.074

Data are expressed as mean ⫾ SD or as number (percentage) unless otherwise specified. * Median (interquartile range). Q ⫽ quintile.

0.986 ⬍0.001

1036

The American Journal of Cardiology (www.ajconline.org)

Table 2 Hazard ratios for 1-year all-cause mortality and cardiovascular mortality Cox models All-cause mortality* RLC% splines† RLC% quintiles Q5 Q4 Q3 Q2 Q1 CV mortality§ RLC% splines† RLC% quintiles Q5 Q4 Q3 Q2 Q1

Unadjusted HR (95% CI)

p Value

Adjusted HR (95% CI)

p Value

NA

⬍0.001‡

NA

0.007‡

1 1.59 (1.05–2.42) 1.58 (1.04–2.40) 1.90 (1.27–2.84) 2.22 (1.49–3.29)

0.002‡ 0.028 0.031 0.002 ⬍0.001

1 1.23 (0.80–1.89) 1.24 (0.81–1.91) 1.30 (0.86–1.98) 1.76 (1.17–2.65)

0.065‡ 0.338 0.315 0.219 0.006

NA

⬍0.001‡

NA

0.012‡

1 1.28 (0.83–1.98) 1.34 (0.85–2.11) 1.13 (0.69–1.82) 1.81 (1.19–2.75)

0.055‡ 0.264 0.213 0.631 0.006

1 1.64 (1.06–2.55) 1.56 (1.00–2.44) 1.63 (1.05–2.54) 2.07 (1.35–3.16)

0.023‡ 0.028 0.049 0.030 0.001

* Covariates for the adjusted model: age, HF cause (ischemic, valvular, and others), categories of left ventricular ejection fraction (⬍35%, 35% to 49%, and ⱖ50%), systolic blood pressure, the interaction between atrial fibrillation and heart rate, serum creatinine ⬎1.3 mg/dl, log serum brain natriuretic peptide, serum antigen carbohydrate 125, Charlson index categories (1 and 0), pleural effusion, treatment with inotropic medication during hospitalization, and statin prescription at discharge. † RLC% per change from 25th to 75th percentile. ‡ Omnibus p value. § Covariates for the adjusted competing risk model: age, HF cause (ischemic, valvular, and others), categories of left ventricular ejection fraction (⬍35%, 35% to 49%, and ⱖ50%), systolic blood pressure ⬍100 mm Hg, heart rate, serum creatinine ⬎1.3 mg/dl, log serum brain natriuretic peptide, log serum antigen carbohydrate 125, treatment with inotropic medication during hospitalization, and Charlson index ⬎0. In addition, this model accounted for non-CV mortality as a competing event. CI ⫽ confidence interval; HR ⫽ hazard ratio; NA ⫽ not applicable; Q ⫽ quintile.

those patients who died out of hospital (n ⫽ 39) were assumed to have died of CV causes. This study was approved by an institutional review committee, and all patients provided written informed consent. Baseline characteristics were compared among quintiles of RLC% distribution. For continuous, normally distributed variables, comparisons were calculated using analysis of variance; for highly skewed variables, the Kruskal-Wallis rank test was used. The chi-square test was used for the comparison of discrete variables. Cumulative mortality curves and their differences were estimated with appropriate methods: Kaplan-Meier and Peto-Peto Prentice tests for total mortality and cumulative incidence function and Gray tests for CV mortality. The independent effects of RLC% quintiles on long-term mortality and CV mortality were assessed by using Cox regression and Cox regression suited for competing risk events.13 All variables listed in Table 1 were tested with prognostic purposes. Candidate covariates for multivariate analyses were chosen on the basis of previous medical knowledge, independent of their p values. Backward stepwise selection that kept the nominal type I error at 0.05 was used.14 During selection, the functional form of continuous variables was modeled by restricted cubic splines. The covariates included in the final models are listed at the bottom of Table 2. For the 2 end points, the proportionality assumption was tested. Model discrimination was assessed using Harrell’s C-statistic. Cox model calibration was tested using the Gronnesby and Borgan test.15 The increment in the prognostic utility of adding lymphocyte count to the clinical model was determined using the integrated discrimination improvement index.16 Two-sided p values

⬍0.05 were considered statistically significant for all analyses. All analyses were performed using Stata version 11.1 (StataCorp LP, College Station, Texas). Results The mean age in our sample was 73 ⫾ 11 years; 50.8% were women, and 54% exhibited left ventricular ejection fractions ⬎50%. The medians for total WBC, neutrophil count, lymphocyte count, and monocyte count were 9.4 ⫻ 103 cells/ml (interquartile range 7.5 to 12.10), 6.90 ⫻ 103 cells/ml (5.10 to 9.10 interquartile range), 1.40 ⫻ 103 cells/ml (1.01 to 20 interquartile range), and 0.46 ⫻ 103 cells/ml (0.35 to 0.60 interquartile range), respectively, and the median RLC% was 16% (interquartile range 10.8% to 22.7%). Overall, there was an inverse and monotonic association between quintiles of RLC% and most clinical parameters indicative of chronicity and congestion (Table 1). Indeed, higher proportions of previous admission for AHF, admission as acute decompensated HF, previous stroke, pleural effusion, peripheral edema, previous use of diuretics, and preserved systolic function were found from quintiles 1 to 5 of RLC%. Likewise, a higher mean age and lower systolic and diastolic blood pressures, serum hemoglobin, and serum sodium were observed paralleling the decrease in RLC% quintiles (Table 1). At 1-year follow-up, 286 deaths (24.0%) were ascertained, with 67 (5.6%) occurring during the index hospitalization. There was an inverse trend between 1-year mortality and RLC% quintiles: 31.5%, 27.2%, 23.1%, 23%, and 15.5% for quintiles 1 to 5, respectively (p for trend ⬍0.001). The cumulative mortality risk was significantly higher for

Heart Failure/RLC% and Long-Term Mortality in AHF

1037

Figure 1. (a) Cumulative rates of 1-year mortality by quintiles of RLC% (Kaplan-Meier). (b) Functional form of the adjusted association between RLC% and 1-year mortality. RLC% was modeled with restrictive cubic splines (4 degrees of freedom). Under a multivariate setting, RLC% was related in a nonlinear fashion with 1-year mortality; in patients in the lowest quintile, a pronounced gradient of risk was observed compared to the rest of the continuum. Q ⫽ quintile.

Figure 2. (a) Cumulative rates of CV mortality by quintiles of RLC% (cumulative incidence function method). (b) Functional adjusted form of the relation between RLC% and CV mortality. RLC% was modeled with restrictive cubic splines (4 degrees of freedom). Under a multivariate setting, RLC% was related in a nonlinear fashion with 1-year CV mortality. Q ⫽ quintile.

patients in quintile 1, intermediate for those in quintiles 2 to 4, and lower for those in quintile 5, differences that were notable from the beginning of follow-up (Figure 1). Furthermore, when tested in the multivariate Cox model, RLC% emerged as an independent predictor of 1-year mor-

tality (Table 2). Patients in quintile 1 of RLC% (⬍9.7%) displayed an adjusted increased risk for all-cause mortality compared to those in quintile 5 (hazard ratio 1.76, 95% confidence interval 1.17 to 2.65, p ⫽ 0.006). Intermediate quintiles (quintiles 2 to 4) were not associated with 1-year

1038

The American Journal of Cardiology (www.ajconline.org)

mortality (Table 2). Harrell’s C-statistic of the final Cox model was 0.754. The Gronnesby and Borgan test of goodness of fit showed good model calibration (p ⫽ 0.913). Furthermore, the added utility of adding RLC% to the base model was also supported by an integrated discrimination improvement index of 0.68% (p ⫽ 0.007). When all WBC count variables (RLC%, total WBC, and neutrophil, lymphocyte, and monocyte counts) were forced into the Cox model (adjusted by one another), only RLC% was retained in the final multivariate model. No significant interactions were observed between RLC% and age ⬎75 years (p ⫽ 0.469), gender (p ⫽ 0.296), left ventricular ejection fraction ⬍50% (p ⫽ 0.575), ischemic cause of HF (p ⫽ 0.147), previous admission for AHF (p ⫽ 0.507), and acute decompensated HF clinical presentation (p ⫽ 0.455), indicating an uniform prognostic effect of RLC% among these representative subgroups of patients. Because during model selection, RLC% was retained and modeled with restricted cubic splines with 4 degrees of freedom (because of a lack of linearity assumption), our next step was to determine its functional form of relating with the log hazard for 1-year mortality. Overall, we can characterize the association as inverse, and definitively nonlinear, with a pronounced exponential decrease in mortality up to 12%; then, the curve becomes flat up to 25%, and finally, a less stepped decrease in risk for the rest of the continuum was observed (Figure 1). By changing the reference value, patients with RLC% ⬍7.5% (n ⫽ 133) and ⬍5% (n ⫽ 52) displayed 1.95-fold (p ⬍0.001) and 2.66fold (p ⬍0.001) adjusted increased risk for mortality compared to those in the upper quintile. For the CV end point, the results paralleled those observed for 1-year total mortality (Table 2). Two hundred thirty-eight deaths (20.0%) were ascertained as CV, and 63 of these (5.3%) occurred during the index admission. There was an inverse trend between mortality and RLC% quintiles. From quintile 1 to quintile 5, 1-year mortality rates were 25.5%, 20.6%, 19.7%, 20.6%, and 13.5%, respectively (p ⫽ 0.026). The cumulative mortality risk was significantly higher for patients in quintile 1, intermediate for those in quintiles 2 to 4, and lower for those in quintile 5, differences that were notable from the beginning of the follow-up period (Figure 2). Furthermore, when tested in a multivariate setting, RLC% remained independently associated with 1-year CV mortality (Table 2). The model’s area under the curve was 0.794; the integrated discrimination improvement index resulting from the comparison of the multivariate models, with and without RLC%, was estimated at 0.5% (p ⫽ 0.026). When RLC% was modeled with restricted cubic splines with 4 degrees of freedom, the association between RLC% and CV mortality resembles the curve for all-cause mortality (Figure 2). Discussion In the present study, using a nonselected hospitalized population of patients with AHF, we have shown that RLC% was inversely and, in a nonlinear fashion, associated with all-cause and CV mortality at 1 year. Importantly, our results can be viewed as a confirmation of

previous and similar findings in a large and nonselected population with AHF. In recent observational studies, RLC% has emerged as a prognostic marker in patients with HF.5–9 However, most of these studies were small6,8,9 and selectively included patients with left ventricular systolic dysfunction5–9 or patients in an ambulatory setting.6 – 8 Remarkably, RLC% was incorporated into the Seattle Heart Failure Model scoring system because of the observed inverse association with survival at 1, 2, and 3 years.7 Although the exact pathophysiologic mechanisms behind this association are not fully clarified, several mechanisms have been proposed.17 A great amount of evidence supports the notion that leukocytosis, neutrophilia, and lymphopenia are a frequent response to systemic stress, an effect most likely mediated by elevated levels of cortisol and catecholamines.18 –21 Nevertheless, in 129 outpatients with HF, Huehnergarth et al6 did not find any significant association between RLC% and nonsalivary morning cortisol and only a weak correlation with evening salivary cortisol (r ⫽ ⫺0.13, p ⫽ 0.051). Moreover, and contrary to what has been observed in other acute CV diseases, leukocytosis and neutrophilia have not been consistently associated with prognosis in HF, which suggests that neurohormonal activation may not be the only pathophysiologic mechanism behind these findings. In recent years, HF has been characterized as a syndrome in which immunologic activation and inflammation prevailed.1– 4 Indeed, functional alterations of peripheral blood mononuclear cells have been observed in advanced cases of HF, indicating chronic activation of lymphocytes and monocytes by high levels of cytokines.1– 4,22 One of the most plausible mediators for this immune-inflammatory activation, especially during acute episodes, is bacterial endotoxin translocation due to intestinal bacterial proliferation. Anker et al23–25 has postulated that during repetitive episodes of decompensation, particularly when associated with severe systemic congestion, bacterial endotoxin translocation from the gut into the circulation may occur, with subsequent activation of immune response and cytokine release. It is known that administration of Escherichia coli lipopolysaccharide to humans causes a biphasic response in neutrophils, with initial decrease and posterior increases, but a more uniform response in lymphocytes, evidenced by a nadir of lymphopenia at 1.5 and 4 hours.26 Also, the changes in circulating neutrophil, lymphocyte, and monocyte counts during intravenous infusion of recombinant tumor necrosis factor–␣ resemble the changes seen after a bolus injection of E. coli lipopolysaccharide,26,27 suggesting that its effect in experimental endotoxemia is mediated, at least in part, through this cytokine. Moreover, a dose-response effect of E. coli lipopolysaccharide was observed, augmenting the robustness of this hypothesis.26,27 In this regard, a significant correlation between plasma levels of soluble tumor necrosis receptor–1 and the relative number of lymphocyte subsets has been recently reported in HF.28 Although various mechanisms have been postulated, enhanced lymphocyte apoptosis appears to be 1 of the key mechanisms implicated in the pathogenesis of lymphopenia in HF,17,29,30 and it seems to happen regardless of the origin.29 In agreement with Anker et al’s23–25 postulate, we

Heart Failure/RLC% and Long-Term Mortality in AHF

have observed in the present study that low RLC% was more prevalent in patients with previous admission for AHF, with clinical presentation of acute decompensated HF and severe systemic congestion. Whether low RLC% is directly involved in the pathogenesis of AHF or whether it is just a marker of disease severity is an issue this study was not intended to address, and thus, it will require further investigations. The fact that lymphopenia may predispose patients to infections, which in turn are a well-known cause of death and a precipitating factor for HF decompensation, may be seen as an indirect evidence of a direct effect.10,11 Some limitations must be addressed. First, this was a singlecenter observational study in which, by design, different types of bias and residual confounding may have been operating. Second, the adjudication of specific cause of death was done mainly using chart review, which may introduce some error into the competing risk estimates. Third, with the present data, we cannot reveal the exact pathophysiologic mechanisms behind this association. Last, by modeling RLC% baseline values, we were not able to infer the effect of repeated measures on the association found. 1. Fildes JE, Shaw SM, Yonan N, Williams SG. The immune system and chronic heart failure: is the heart in control? J Am Coll Cardiol 2009;53:1013–1020. 2. Torre-Amione G. Immune activation in chronic heart failure. Am J Cardiol 2005;95(suppl):3C– 8C. 3. von Haehling S, Schefold JC, Lainscak M, Doehner W, Anker SD. Inflammatory biomarkers in heart failure revisited: much more than innocent bystanders. Heart Fail Clin 2009;5:549 –560. 4. Yndestad A, Damas JK, Oie E, Ueland T, Gullestad L, Aukrust P. Systemic inflammation in heart failure—the whys and wherefores. Heart Fail Rev 2006;11:83–92. 5. Acanfora D, Gheorghiade M, Trojano L, Furgi G, Pasini E, Picone C, Papa A, Iannuzzi GL, Bonow RO, Rengo F. Relative lymphocyte count: a prognostic indicator of mortality in elderly patients with congestive heart failure. Am Heart J 2001;142:167–173. 6. Huehnergarth KV, Mozaffarian D, Sullivan MD, Crane BA, Wilkinson CW, Lawler RL, McDonald GB, Fishbein DP, Levy WC. Usefulness of relative lymphocyte count as an independent predictor of death/ urgent transplant in heart failure. Am J Cardiol 2005;95:1492–1495. 7. Levy WC, Mozaffarian D, Linker DT, Sutradhar SC, Anker SD, Cropp AB, Anand I, Maggioni A, Burton P, Sullivan MD, Pitt B, Poole-Wilson PA, Mann DL, Packer M. The seattle heart failure model: Prediction of survival in heart failure. Circulation 2006;113:1424 –1433. 8. Ommen SR, Hodge DO, Rodeheffer RJ, McGregor CG, Thomson SP, Gibbons RJ. Predictive power of the relative lymphocyte concentration in patients with advanced heart failure. Circulation 1998;97:19 –22. 9. Rudiger A, Burckhardt OA, Harpes P, Muller SA, Follath F. The relative lymphocyte count on hospital admission is a risk factor for long-term mortality in patients with acute heart failure. Am J Emerg Med 2006;24:451– 454. 10. Dickstein K, Cohen-Solal A, Filippatos G, McMurray JJ, Ponikowski P, Poole-Wilson PA, Stromberg A, van Veldhuisen DJ, Atar D, Hoes AW, Keren A, Mebazaa A, Nieminen M, Priori SG, Swedberg K, Vahanian A, Camm J, De Caterina R, Dean V, Funck-Brentano C, Hellemans I, Kristensen SD, McGregor K, Sechtem U, Silber S, Tendera M, Widimsky P, Zamorano JL. Esc guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur Heart J 2008;29:2388 –2442. 11. Nieminen MS, Bohm M, Cowie MR, Drexler H, Filippatos GS, Jondeau G, Hasin Y, Lopez-Sendon J, Mebazaa A, Metra M, Rhodes A, Swedberg K, Priori SG, Garcia MA, Blanc JJ, Budaj A, Dean V, Deckers J, Burgos EF, Lekakis J, Lindahl B, Mazzotta G, Morais J,

12.

13. 14. 15.

16.

17. 18.

19. 20. 21. 22. 23. 24. 25.

26. 27.

28.

29. 30.

1039

Oto A, Smiseth OA, Dickstein K, Albuquerque A, Conthe P, Crespo-Leiro M, Ferrari R, Follath F, Gavazzi A, Janssens U, Komajda M, Moreno R, Singer M, Singh S, Tendera M, Thygesen K. Executive summary of the guidelines on the diagnosis and treatment of acute heart failure: the Task Force on Acute Heart Failure of the European Society of Cardiology. Eur Heart J 2005;26:384 – 416. Lloyd-Jones D, Adams RJ, Brown TM, Carnethon M, Dai S, De Simone G, Ferguson TB, Ford E, Furie K, Gillespie C, Go A, Greenlund K, Haase N, Hailpern S, Ho PM, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A, McDermott MM, Meigs J, Mozaffarian D, Mussolino M, Nichol G, Roger VL, Rosamond W, Sacco R, Sorlie P, Stafford R, Thom T, Wasserthiel-Smoller S, Wong ND, Wylie-Rosett J. Executive summary: heart disease and stroke statistics—2010 update: A report from the American Heart Association. Circulation 2010;121:948 –954. Fine J, Gray R. A proportional hazard model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94:496 –509. Sauerbrei W, Royston P, Binder H. Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Stat Med 2007;26:5512–5528. May S, Hosmer DW. Type goodness-of-fit statistics for the Cox proportional hazards model. In: Balakrishnana N, Rao CR, eds. Advances in Survival Analysis: Handbook of Statistics. Amsterdam, The Netherlands: Elsevier, 2004;383–394. Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: From area under the roc curve to reclassification and beyond. Stat Med 2008;27:157– 172. Núñez J, Miñana G, Bodí V, Núñez E, Sanchis J, Husser O, Llàcer A. Low lymphocyte count and cardiovascular diseases. Curr Med Chem. In press. Maisel AS, Knowlton KU, Fowler P, Rearden A, Ziegler MG, Motulsky HJ, Insel PA, Michel MC. Adrenergic control of circulating lymphocyte subpopulations. Effects of congestive heart failure, dynamic exercise, and terbutaline treatment. J Clin Invest 1990;85:462– 467. Maisel AS, Michel MC. Beta-adrenoceptor control of immune function in congestive heart failure. Br J Clin Pharmacol 1990;30(suppl): 49S–53S. Mignini F, Traini E, Tomassoni D, Vitali M, Streccioni V. Leucocyte subset redistribution in a human model of physical stress. Clin Exp Hypertens 2008;30:720 –731. Thomson SP, McMahon LJ, Nugent CA. Endogenous cortisol: a regulator of the number of lymphocytes in peripheral blood. Clin Immunol Immunopathol 1980;17:506 –514. Tamariz L, Hare JM. Inflammatory cytokines in heart failure: Roles in aetiology and utility as biomarkers. Eur Heart J 2010;31:768 –770. Anker SD, Egerer KR, Volk HD, Kox WJ, Poole-Wilson PA, Coats AJ. Elevated soluble cd14 receptors and altered cytokines in chronic heart failure. Am J Cardiol 1997;79:1426 –1430. Krack A, Sharma R, Figulla HR, Anker SD. The importance of the gastrointestinal system in the pathogenesis of heart failure. Eur Heart J 2005;26:2368 –2374. Niebauer J, Volk HD, Kemp M, Dominguez M, Schumann RR, Rauchhaus M, Poole-Wilson PA, Coats AJ, Anker SD. Endotoxin and immune activation in chronic heart failure: A prospective cohort study. Lancet 1999;353:1838 –1842. Andreasen AS, Krabbe KS, Krogh-Madsen R, Taudorf S, Pedersen BK, Moller K. Human endotoxemia as a model of systemic inflammation. Curr Med Chem 2008;15:1697–1705. Michie HR, Spriggs DR, Manogue KR, Sherman ML, Revhaug A, O’Dwyer ST, Arthur K, Dinarello CA, Cerami A, Wolff SM, et al. Tumor necrosis factor and endotoxin induce similar metabolic responses in human beings. Surgery 1988;104:280 –286. von Haehling S, Schefold JC, Jankowska E, Doehner W, Springer J, Strohschein K, Genth-Zotz S, Volk HD, Poole-Wilson P, Anker SD. Leukocyte redistribution: Effects of beta blockers in patients with chronic heart failure. PLoS One 2009;4:e6411. Agnoletti L, Curello S, Malacarne F, Airà P, Cargnonia A, Valgimiglid M, Ferrari R. Immune activation in severe heart failure. Does etiology play a role? Eur Heart J 2004;6(suppl):F22–F29. Gupta S, Agrawal A, Agrawal S, Su H, Gollapudi S. A paradox of immunodeficiency and inflammation in human aging: lessons learned from apoptosis. Immun Ageing 2006;3:5.