Low lymphocyte count has been considered a predictive marker of unfavorable outcomes for patients with heart failure (HF). Baseline blood samples for complete blood counts, differential counts, renal function tests. and lipid profile were prospectively obtained to assess the association between lymphocyte count and clinical outcomes in 305 patients with HF (average New York Heart Association [NYHA] class 2.8). The mean follow-up duration was 4.7 years (range 8 months to 8.4 years), and 111 patients (36%) died during the follow-up period. The mean lymphocyte count for the group was 1,803.64 ± 740.3, and the mean left ventricular ejection fraction (LVEF) was 37%. Patients with low lymphocyte counts (<1,600 median count) after 8 years had significantly lower survival rates than those with lymphocyte counts ≥1,600 (58% vs 72%, p = 0.012). The prediction of poorest survival was for patients in NYHA class III or IV and with lymphocyte counts <1,600. Regression analysis showed that lymphocyte level, the LVEF, and NYHA class were predictors of mortality. Of these, NYHA class was the most prominent predictor, followed by lymphocyte count, which was even more significant than the LVEF (hazard ratio 0.76, p = 0.037). In conclusion, the findings of this study demonstrate that total lymphocyte count is an important prognostic factor, inversely associated with predicted mortality. Although the total low lymphocyte count was correlated with a lower NYHA class and a lower LVEF, it emerged as an independent death risk factor in patients with chronic HF.
The aim of the present study was to determine different lymphocyte count patterns and their prognostic association with outcomes in patients with heart failure (HF).
Methods
All patients in this prospective study were recruited successively from the special HF unit at Tel Aviv Sourasky Medical Center. Baseline blood samples for complete blood count, including automatic differential counts, renal function tests, and lipid profile levels were obtained from January 2000 and July 2001. Systolic HF was defined as a left ventricular ejection fraction (LVEF) <40% as determined by echocardiography. On the first visit, all participants were examined by a physician, who obtained their medical histories; listed medications used; performed physical examinations; measured blood pressure at rest, heart rate, and weight; determined New York Heart Association (NYHA) class; and performed echocardiography. Ischemic heart disease was defined according to a history of confirmed myocardial infarction, coronary artery bypass grafting, or coronary angiography. Patients were followed up by an HF specialist at least once every 3 months and often much more frequently.
Excluded were patients with malignant diseases with diffuse metastases; those who received chemotherapy and/or radiotherapy or medications known to affect the complete blood count; patients with acute and severe renal failure, active hepatic disease, severe pulmonary disease (e.g., chronic obstructive pulmonary disease), and acute infectious disease; and those <3 months since acute myocardial infarction or heart surgery. The end point of the study was all-cause mortality or hospitalization. The study was approved by the institutional ethics committee, and each subject provided informed consent to participate in the study.
The laboratory parameters were grouped as follows: lymphocyte count <1,600 versus ≥1,600, NYHA class I or II versus III or IV, and LVEF <40% versus ≥40%. The Kaplan-Meier method was used to compute the product-limit estimate survivor function of morbidity and mortality for each lymphocyte group.
A comparison of survival curves between the 2 groups was carried out using a log-rank test and Wilcoxon’s test. Regression analysis of survival data based on a Cox proportional-hazards model was used to explain the effect of laboratory values on morbidity and mortality hazard rates. The variables thought to be related to survival were white blood cell count × 1,000, C-reactive protein, the LVEF, lymphocyte count, and NYHA class. A backward elimination process was used to identify important prognostic factors from these 9 explanatory variables. Lymphocyte count, the LVEF, and NYHA were identified as the most important prognostic factors for mortality.
The Cox proportional-hazards model was used to predict the survival function for 2 covariate sets. The first set consisted of 4 combinations of the lymphocyte group and the NYHA group. The second set consisted of 4 combinations of the lymphocyte group and the LVEF group. Pearson’s correlations were calculated to find the linear association between the lymphocyte group and laboratory variables.
Results
A total of 334 consecutive outpatients with congestive HF–related symptoms were found suitable to participate in this study. Twenty-nine were excluded for technical reasons, noncompliance, or lack of follow-up information. The remaining 305 patients constituted the study cohort. The mean follow-up duration was 4.7 years (range 8 months to 8.4 years), and 163 patients (53%) were admitted to the hospital during the course of their follow-up. All admissions were due to HF exacerbation. The general clinical and laboratory data of all study patients are listed in Table 1 . There were no significant differences between the 2 groups in mean NYHA class (2.8), mean lymphocyte count (1,803.64 ± 740.3), and mean LVEF (37%). Because of the small number of patients in NYHA class I and the unequal numbers of patients in the various NYHA groups, it was decided to divide all the patients into groups in NYHA class I or II and in NYHA class III or IV.
| Variable | Value | Laboratory | Value |
|---|---|---|---|
| Mean age (years) ⁎ | 72 | Hemoglobin (g%) | 12.8 ± 6.5 |
| Men | 69 (69.0%) | Platelets (per ml) | 214,000.56 ± 51,234.4 |
| Body mass index (kg/m 2 ) ⁎ | 26.1 | WBCs (per ml) | 7,213.39 ± 2,840.3 |
| Hyperlipidemia | 192 (63.6%) | PMN leukocytes (per ml) | 5,386.34 ± 1,994.4 |
| Smokers | 70 (23.0%) | Lymphocytes (per ml) | 1,803.64 ± 740.3 |
| Hypertension | 205 (67.0%) | Monocytes | 719.73 ± 616.6 |
| Diabetes mellitus | 116 (38.0%) | SGPT (mg/dl) | 24.32 ± 12.2 |
| Ischemic heart disease | 245 (80%) | Albumin (mg/dl) | 3.91 ± 1.3 |
| Valvular heart disease | 60 (20%) | Creatine clearance (ml/min) | 42.32 ± 13.8 |
| Systolic dysfunction (LVEF <40%) | 220 (72%) | Total cholesterol (mg/dl) | 177.45 ± 45.3 |
| HF with preserved LVEF (>40%) | 86 (28%) | LDL cholesterol (mg/dl) | 43.12 ± 16.2 |
| Mean NYHA class ⁎ | 2.8 | HDL cholesterol (mg/dl) | 137.56 ± 67.9 |
| NYHA classes I and II | 89 (29%) | Triglycerides (mg/dl) | 139.43 ± 67.8 |
| NYHA classes III and IV | 226 (74%) | Hs-CRP (mg/dl) | 3.43 ± 1.6 |
| Follow-up (years) ⁎ | 4.7 | NT-proBNP (pg/ml) | 1,991 ± 1,020.2 |
| Admissions | 104 (34%) | Glucose | 134.21 ± 53.1 |
| Mortality | 111 (36%) | Alkaline phosphatase | 64.31 ± 19.2 |
To determine the predictive value of lymphocyte count on mortality, we divided our cohort according to median absolute lymphocyte count, 1,600 (which is also the midrange value of absolute lymphocyte count in our laboratory).
Those with lymphocyte counts <1,600 were assigned to group 1, and those with lymphocyte counts ≥1,600 were assigned to group 2. Figure 1 shows the Kaplan-Meier survival-mortality curves by lymphocyte count, and Figure 2 illustrates the Kaplan-Meier morbidity curve, which examines the time to first hospitalization in the 2 groups of patients. The survival rate of group 1 patients was significantly lower than that of group 2 patients (p <0.02). The Cox regression curves showed the survival of patients with different lymphocyte counts according to NYHA class I or II and class III or IV and according to LVEFs <40% and ≥40%. Figure 3 shows that patients who were in NYHA class I or II with lymphocyte counts <1,600 had a poorer survival rate than those in group 2 with lymphocyte counts ≥1,600. No significant association was found between the LVEF, NYHA class, and morbidity in any of the lymphocyte groups. Figure 4 displays the mortality rate according to lymphocyte count and the LVEF. The best survival rate was in the subgroup with lymphocyte counts ≥1,600 and LVEFs ≥40%.
Table 2 lists Pearson’s correlations for mortality between lymphocyte count and other laboratory and clinical variables. Only N-terminal pro–B-type natriuretic peptide level was correlated with the lymphocyte groups. Pearson’s correlation analysis of morbidity revealed no correlations between the groups with different lymphocyte counts and other laboratory and clinical variables. Regression analysis yielded the mortality hazard ratio of the different clinical and laboratory variables ( Table 3 ). We chose variables associated with inflammation despite the relatively scant evidence of predictive value, because these are relevant to understanding the importance of lymphocytes. Only the lymphocyte group, the LVEF group, and NYHA class emerged as significant. NYHA class had the highest HR, with lymphocyte count being the second most important value (hazard ratio 0.76, p = 0.037). The only significant morbidity hazard ratio was that of NYHA class. The 1-, 3-, and 5-year mortality was higher in group 1 patients: 75 group 1 patients (42%) died during follow-up (2,500 days) compared with 36 group 2 patients (28%).
