Predictive value of hematological parameters in prognosis of congestive heart failure patients in a tertiary hospital, Addis Ababa, Ethiopia

Abstract

Background

Hematological parameters are increasingly recognized as cost-effective prognostic indicators in chronic heart failure (CHF), offering insight into disease severity and progression.

Objective

To evaluate selected hematological parameters and their prognostic utility in patients with CHF, using the New York Heart Association (NYHA) as a clinical severity reference.

Methods

A total of 206 adult CHF patients were recruited from the cardiology department of St. Paul’s Hospital Millennium Medical College (SPHMMC), Addis Ababa. Venous blood samples (5 mL) were collected in EDTA tubes and analyzed using the DxH 800 Beckman Coulter Hematology Analyzer. Differences in hematological parameters across NYHA classes were evaluated, and prognostic performance was assessed using receiver operating characteristic (ROC) curve analysis on selected parameters. A p-value < 0.05 was considered statistically significant.

Results

All three parameters—Hct, RDW, and ALC—differed significantly between NYHA class II and class IV patients (p = 0.035, p = 0.002, and p = 0.035, respectively). ROC analysis demonstrated that RDW and ALC provided greater predictive value than Hct. For disease progression from NYHA class I to III, RDW yielded an AUC of 0.663 (95% CI, 0.549–0.777; p = 0.009) at a cut-off of 13.5%, while ALC achieved an AUC of 0.657 (95% CI, 0.543–0.771; p = 0.012) at a cut-off of 1.72 × 10⁹/L. For progression from class I to IV, RDW showed an AUC of 0.708 (95% CI, 0.608–0.808; p < 0.001) at a cut-off of 15.9%, and ALC an AUC of 0.622 (95% CI, 0.514–0.729; p = 0.035) at a cut-off of 1.95 × 10⁹/L.

Conclusion

Hematocrit (Hct), red cell distribution width (RDW), and absolute lymphocyte count (ALC) varied significantly across NYHA functional classes, indicating prognostic relevance in chronic heart failure. Among these, RDW and ALC demonstrated stronger discriminatory performance. Together, these readily available parameters may serve as practical adjuncts for risk stratification in CHF, particularly in resource-limited settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-025-05434-9.

Introduction

Heart failure (HF) is a complex clinical syndrome and a major global health challenge, affecting an estimated 26 million individuals worldwide [1]. It accounts for over one million hospital admissions annually in the United States and Europe alone [1]. The European Society of Cardiology (ESC) and the American Heart Association (AHA)/American College of Cardiology (ACC) define HF as a syndrome characterized by typical symptoms—such as dyspnea, fatigue, and ankle swelling—and signs, including elevated jugular venous pressure, pulmonary crackles, and peripheral edema. These manifestations arise from structural and/or functional cardiac abnormalities that impair cardiac output and/or elevate intracardiac pressures, either at rest or during exertion [2, 3].

The New York Heart Association (NYHA) functional classification remains a cornerstone in the clinical assessment of HF, offering a practical framework for evaluating symptom severity and predicting prognosis [4]. It is integrated into most major risk stratification models and endorsed by recent ESC guidelines for the diagnosis and management of both acute and chronic HF [3].

Hematological parameters are invaluable for HF prognostication owing to their accessibility, affordability, and established clinical relevance [5–7]. They enable clinicians to monitor disease progression, facilitate early interventions, and tailor treatments to improve patient outcomes. Incorporating these markers into routine practice provides a cost-effective strategy for risk stratification and prognosis prediction [7–9]. Key hematologic indices and ratios—including red cell distribution width (RDW), lymphocyte, neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), platelet-to-lymphocyte ratio (PLR), hematocrit (Hct), serum ferritin, and transferrin saturation—yield insights into HF etiology, progression, and overall prognosis [2–7, 10, 11].

A growing body of evidence positions hematologic indices as critical predictors of HF outcomes, rather than mere adjuncts [3, 12–17]. Notably, RDW is consistently elevated in patients with advanced HF or those who succumb to the disease, compared with survivors [18]. Recent studies have reinforced the prognostic value of various hematological and inflammatory markers in chronic CHF, with RDW emerging as a prominent indicator [9, 12, 13, 19–21]. Investigations by Q. Zhang et al. and Liang et al. have demonstrated a strong association between elevated RDW and heightened mortality risk in CHF patients, supporting its role as a cost-effective biomarker for clinical risk stratification [12, 13].

Beyond RDW, other blood parameters offer prognostic utility. Majmundar et al. have underscored the significance of absolute lymphocyte count (ALC) as a predictor of adverse outcomes, including mortality, in CHF patients [5]. Similarly, various studies have assessed the prognostic impact of hematocrit (Hct) in heart failure. Hct is another important hematological parameter that correlates with poorer HF outcomes [14–16, 22].

Inflammatory indices have also gained traction as prognostic tools in CHF assessment. The systemic immune-inflammation index (SII), which integrates neutrophil, lymphocyte, and platelet counts, provides a holistic measure of systemic inflammation. Elevated SII levels are linked to higher mortality and adverse events in CHF patients, highlighting its potential for risk stratification [20, 21]. Likewise, the monocyte-to-lymphocyte ratio (MLR) serves as a simple yet potent biomarker, with higher values indicating intensified inflammation and worse outcomes [4, 11]. These indices illuminate inflammatory pathways driving HF progression and aid in identifying high-risk patients amenable to targeted therapies.

Integrating hematological and inflammatory biomarkers into routine evaluations enhances predictive accuracy for CHF outcomes, fostering personalized management and timely interventions to optimize prognosis. Nonetheless, further research is essential to validate these biomarkers and elucidate their therapeutic implications. Advanced CHF is associated with impaired bone marrow function, leading to broad hematopoietic dysfunction. CD34 + progenitor cells from CHF patients show reduced burst-forming unit–erythroid (BFU-E) colony formation compared with healthy controls, signaling compromised hematopoietic capacity with implications for disease progression and management [17]. This dysfunction may underpin anemia and other hematological anomalies common in HF, thereby affecting clinical outcomes and treatment strategies.

Hematological markers are increasingly valued for their diagnostic and prognostic roles, particularly in resource-limited settings such as Ethiopia, due to their accessibility and low cost. Parameters like Hct, RDW, and ALC are especially useful when combined with traditional clinical assessments and imaging, improving diagnostic precision and prognostic evaluations.

This study evaluated the diagnostic and prognostic significance of these hematologic parameters in patients with congestive HF, offering novel insights tailored to the Ethiopian population, where such data are scarce. By stratifying patients according to NYHA functional class and examining class-specific variations in hematological profiles within a tertiary care setting, the research aims to guide clinical decision-making and enhance patient outcomes through identification of accessible biomarkers reflective of disease severity and progression.

Materials and methods

Study area

The study was conducted at St. Paul’s Hospital Millennium Medical College (SPHMMC), located in Addis Ababa, Ethiopia. SPHMMC is one of the country’s few tertiary-level referral hospitals and serves as a major teaching and research institution. On average, the hospital accommodates more than 1,200 outpatient visits daily. In general, the hospital admits more than eight (8) CHF cases per week and more than 400 annually.

Study design and duration

A cross-sectional study design was used to enroll participants who have been admitted with CHF cases at the inpatient and outpatient departments of SPHMMC, over the period of six months from February 2022 to June 2022.

Study population

The study includes all CHF patients who have either been in follow-up or are newly admitted patients at St. Paul’s Hospital Millennium Medical College (SPHMMC) during the study period.

Sample size and sampling method

CHF patients who had been referred to SPHMMC and complied with the inclusion criteria during the study period were generally considered for sampling, of which 206 CHF participants were selected for the study. The 206 patients were calculated using a single population proportion formula based on a previous study at SPHMMC [22]. A convenience sampling was used to recruit the study participants. The final analytical cohort included 206 patients of which 176 were non-anemic. The median duration between CHF diagnosis and blood sampling was 16 months (interquartile range [IQR]: 5–36 months). Among non-anemic patients, the median duration was 15 months (IQR: 6–34 months).

Inclusion criteria

Eligible participants were adults aged 18 years or older who had a confirmed diagnosis of chronic heart failure (CHF) and were classified as at least New York Heart Association (NYHA) functional class I at the time of evaluation.

Exclusion criteria

Patients were excluded if they had undergone a major surgical procedure within the three months preceding enrollment or if they presented with comorbid conditions known to significantly affect hematological parameters. These included active infections, end-stage renal disease requiring dialysis, advanced hepatic disease, malignancies, or primary hematologic disorders. Additionally, individuals with a history of blood transfusion within six months prior to enrollment were not considered for inclusion. Therefore, Of 234 patients initially screened, 28 were excluded: 10 due to missing hematologic data, 8 due to concurrent infection, 5 with chronic kidney disease on dialysis, and 5 due to incomplete NYHA documentation.

Data collection procedure

Experienced laboratory technologists and phlebotomists collected 5 mL venous blood samples from adult participants with strict adherence to standard operating procedures and guidelines using EDTA vacutainer tubes. In addition to the main sample collection unit of the laboratory, samples were collected in the emergency and cardiology departments of the hospital. Moreover, socio-demographic and clinical data were collected using a data extraction format specifically adapted for this study from the patient’s medical record.

Hematological analyses

After the sample passed all the pre-analytical quality checks like labeling and sample integrity, it was carefully delivered to the hematological department of the hospital, where it was analyzed with the 5-part DxH 800 Beckman Coulter Hematologic Analyzer. This instrument operates based on the Coulter principle. According to this method, as a particle passes through the sensing zone when the liquid is drawn from the container, a volume of the electrolyte equivalent to the immersed volume of the particle is displaced from the sensing zone. This causes a short-term change in the resistance across the aperture. This resistance change can be measured as either a voltage or current pulse. By measuring the number of pulses and their amplitudes, it is possible to have information about the number of particles and the volume of each particle.

Data quality assurance

To ensure the quality of test results, quality control tools were used based on the guideline of the manufacturer of the auto-analyzer. Overall, to verify the analytical performance of the machine, three types of quality control materials (normal, low, and high samples) were run prior to daily operation. In line with this, SOPs have been strictly followed from specimen collection through labeling, transportation, and processing to test result issuing.

Data analysis and interpretation

All data, including hematological, demographic, and clinical variables, were initially recorded in Microsoft Excel 2016 and subsequently imported into IBM^® SPSS^® Statistics for Windows, Version 26.0 for analysis. Descriptive statistics were generated to summarize study variables, and results were presented in tables and figures where appropriate.

The Shapiro–Wilk test was used to assess the normality of continuous variables (p > 0.05 indicating approximate normality). For normally distributed variables, one-way ANOVA was employed to compare mean values across the four New York Heart Association (NYHA) functional classes I through IV. MANOVA was additionally performed to examine differences across multiple dependent variables simultaneously. Prior to conducting ANOVA, MANOVA, and supplementary regression analyses, statistical assumptions were assessed. Homogeneity of variances and covariances was evaluated using Levene’s test and Box’s M test, respectively. Multicollinearity was examined through inter-variable correlations and variance inflation factors (VIF), with all values < 2, indicating no problematic collinearity. For regression analyses, residuals were checked for independence, normality, and homoscedasticity. All assumptions were satisfactorily met. To evaluate the diagnostic performance and prognostic value of selected hematological parameters in relation to CHF severity, receiver operating characteristic (ROC) curve analysis was performed. A p-value < 0.05 was considered statistically significant in all analyses.

Results

Sociodemographic characteristics

The study included 206 patients with chronic heart failure (CHF), with a 100% response rate. Of these, 109 (53%) were males and 97 (47%) were females. The mean age of the study participants was 48.3 ± 17.3 (Table 1).

Table 1.

Sociodemographic characteristics of CHF patients (n = 206)

Parameters N = 206	Age (years) (SD)	Gender
		Male (%)	Female (%)
Mean	48.32 (17.34)
Minimum	18
Maximum	85
Frequency		109 (53%)	97 (47%)

Summary of patienst age and sex distribution. Values are presented as mean (standard deviation,SD), minimum and maximum age, and absolute frequencies with percentages for sex categories

Clinical characteristics

The distribution of New York Heart Association (NYHA) functional classes was 40 patients (19.4%) in class I, 49 (23.8%) in class II, 48 (23.3%) in class III, and 69 (33.5%) in class IV (Table 2). Systemic hypertension was present in 128 patients (62.1%) and diabetes mellitus in 53 patients (25.7%). Chronic kidney disease affected 45 patients (21.8%), and anemia affected 30 patients (14.6%). Other cardiovascular disease was noted in 68 patients (33.0%). Regarding lifestyle factors, 43 patients (20.9%) were current smokers, and 62 (30.1%) reported heavy alcohol use (Table 2).

Table 2.

Clinical characteristics and baseline comorbidities of CHF patients

Variables	Frequency	Percent (%)
NYHA Class
NYHA Class Ⅰ	40	19.4
NYHA Class Ⅱ	49	23.8
NYHA Class Ⅲ	48	23.3
NYHA Class Ⅳ	69	33.5
Diabetic status
Diabetic	53	25.7
Non-Diabetic	152	73.8
Hypertension
Hypertensive	128	62.1
Non-hypertensive	78	37.9
Other CVDs
Have CVDs	68	33
Non-CVDs	138	67
Anemia
Anemic	30	14.6
Non-anemic	176	85.4
Renal failures or dysfunction
Renal failure	45	21.8
Without renal dysfunction	161	78.2
Smoking habit
Smoker	43	20.9
Non-smoker	163	79.1
Alcohol consumption
Drinker	62	30.1
Non-drinker	144	69.9

Distribution of NYHA functional classes and frequencies of major comorbidities (diabetes, hypertension, anemia, chronic kidney disease, and other cardiovascular diseases), along with lifestyle factors (smoking and alcohol use). Data are shown as frequencies and percentages

Hematological parameters by NYHA class

A one-way analysis of variance (ANOVA) was performed to compare hematologic parameters across the four NYHA classes. Significant between-group differences were observed for hemoglobin (F = 3.536, P = 0.016), hematocrit (F = 3.349, P = 0.020), and absolute lymphocyte count (F = 4.009, P = 0.008) (Table 3). Other parameters (including red blood cell count, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, platelet count, neutrophil count, monocyte count, and eosinophil count) did not differ significantly by NYHA class (all P > 0.05).

Table 3.

Comparison of hematological parameters across NYHA classes by one-way ANOVA

Parameters	Groups	Sum of Squares	df	Mean Square	F	Sig.
Hemoglobin	Between Groups	136.774	3	45.591	3.536	0.016
	Within Groups	2604.799	202	12.895
	Total	2741.573	205
Red Blood Cell	Between Groups	11.29	3	3.763	2.22	0.087
	Within Groups	342.391	202	1.695
	Total	353.681	205
Hematocrit	Between Groups	1221.481	3	407.16	3.349	0.02
	Within Groups	24560.482	202	121.587
	Total	25781.963	205
MCV	Between Groups	155.389	3	51.796	1.277	0.284
	Within Groups	8196.145	202	40.575
	Total	8351.535	205
MCH	Between Groups	44.891	3	14.964	2.079	0.104
	Within Groups	1453.627	202	7.196
	Total	1498.518	205
MCHC	Between Groups	17.017	3	5.672	2.008	0.114
	Within Groups	570.503	202	2.824
	Total	587.52	205
MPV	Between Groups	5.753	3	1.918	1.455	0.228
	Within Groups	266.267	202	1.318
	Total	272.02	205
Neutrophil	Between Groups	78.14	3	26.047	1.314	0.271
	Within Groups	4004.704	202	19.825
	Total	4082.845	205
Monocytes	Between Groups	1.29	3	0.43	0.341	0.796
	Within Groups	254.741	202	1.261
	Total	256.03	205
Platelets	Between Groups	6002.741	3	2000.914	0.21	0.89
	Within Groups	1929164.274	202	9550.318
	Total	1935167.015	205
Lymphocytes	Between Groups	12.97	3	4.323	4.009	0.008
	Within Groups	217.827	202	1.078
	Total	230.797	205
Eosinophil	Between Groups	1.111	3	0.37	0.724	0.539
	Within Groups	103.29	202	0.511
	Total	104.401	205

Analysis of variance (ANOVA) results showing between-group and within-group variance, F-values, and significance levels for hematologic indices (including hemoglobin, hematocrit, red cell distribution width [RDW], and leukocyte subsets)

Post hoc pairwise comparisons were conducted to identify which classes differed. For hematocrit, the only significant contrast was between NYHA class II and class IV (mean difference − 5.620; 95% confidence interval [CI] − 10.957 to − 0.284; P = 0.035) (Table 4).

Table 4.

Post hoc analysis of hematocrit differences between NYHA classes

Dependent Variable	NYHA Classification (I)	NYHA Classification (J)	Mean Difference (I-J)	Sig.	95% Confidence Interval
					Lower Bound	Upper Bound
Hematocrit	NYHA Class Ⅰ	NYHA Class Ⅱ	5.299	0.112	−0.787	11.386
		NYHA Class Ⅲ	3.805	0.374	−2.309	9.921
		NYHA Class Ⅳ	−0.321	0.999	−5.998	5.355
	NYHA Class Ⅱ	NYHA Class Ⅰ	−5.299	0.112	−11.386	0.787
		NYHA Class Ⅲ	−1.493	0.909	−7.294	4.307
		NYHA Class Ⅳ	−5.620	0.035*	−10.957	−0.284
	NYHA Class Ⅲ	NYHA Class Ⅰ	−3.805	0.374	−9.921	2.309
		NYHA Class Ⅱ	1.493	0.909	−4.307	7.294
		NYHA Class Ⅳ	−4.127	0.195	−9.496	1.241
	NYHA Class Ⅳ	NYHA Class Ⅰ	0.321	0.999	−5.355	5.998
		NYHA Class Ⅱ	5.620	0.035	0.284	10.957
		NYHA Class Ⅲ	4.127	0.195	−1.241	9.496

Pairwise comparisons of hematocrit values between NYHA functional classes using Tukey’s HSD test. Results include mean differences, p-values, and 95% confidence intervals (CI). Statistically significant differences are denoted with an asterisk (*)

For absolute lymphocyte count, patients in class I had higher counts than those in class II (mean difference + 0.620; 95% CI 0.047–1.194; P = 0.028), class III (mean difference + 0.703; 95% CI 0.127–1.279; P = 0.010), and class IV (mean difference + 0.562; 95% CI 0.027–1.096; P = 0.035) (Table 5).

Table 5.

Post hoc analysis of absolute lymphocyte count differences between NYHA classes

Dependent Variable	NYHA Classification (I)	NYHA Classification (J)	Mean Difference (I-J)	Sig.	95% Confidence Interval
					Lower Bound	Upper Bound
Lymphocytes	NYHA Class Ⅰ	NYHA Class Ⅱ	0.620	0.028*	0.047	1.194
		NYHA Class Ⅲ	0.703	0.01*	0.127	1.279
		NYHA Class Ⅳ	0.562	0.035*	0.027	1.096
	NYHA Class Ⅱ	NYHA Class Ⅰ	−0.620	0.028	−1.194	−0.047
		NYHA Class Ⅲ	0.082	0.98	−0.463	0.628
		NYHA Class Ⅳ	−0.058	0.99	−0.561	0.443
	NYHA Class Ⅲ	NYHA Class Ⅰ	−0.703	0.01	−1.279	−0.127
		NYHA Class Ⅱ	−0.082	0.98	−0.628	0.463
		NYHA Class Ⅳ	−0.141	0.887	−0.646	0.364
	NYHA Class Ⅳ	NYHA Class Ⅰ	−0.562	0.035	−1.096	−0.027
		NYHA Class Ⅱ	0.058	0.99	−0.443	0.561
		NYHA Class Ⅲ	0.141	0.887	−0.364	0.646

Pairwise comparisons of lymphocyte count between NYHA classes using Tukey’s HSD test. Results are expressed as mean differences with p-values and 95% CI. Significant differences are marked with an asterisk (*)

Because RDW did not satisfy the equal-variance assumption, we used the Games–Howell test for RDW comparisons. RDW was significantly higher in class III vs. I (mean difference − 1.406; 95% CI − 2.796 to − 0.016; P = 0.046) and in class IV vs. I (mean difference − 1.707; 95% CI − 2.926 to − 0.489; P = 0.002) (Table 6). Overall, RDW, hematocrit, and lymphocyte count showed statistically significant variation across NYHA classes.

Table 6.

Post hoc analysis of red cell distribution width (RDW) between NYHA classes (Games–Howell test)

Dependent Variable	NYHA Classification (I)	NYHA Classification (J)	Mean Difference (I-J)	Sig.	95% Confidence Interval
					Lower Bound	Upper Bound
RDW	NYHA Class Ⅰ	NYHA Class Ⅱ	−1.821	0.071	−3.752	0.109
		NYHA Class Ⅲ	−1.406	0.046*	−2.796	−0.016
		NYHA Class Ⅳ	−1.707	0.002*	−2.926	−0.489
	NYHA Class Ⅱ	NYHA Class Ⅰ	1.821	0.071	−0.109	3.752
		NYHA Class Ⅲ	0.415	0.952	−1.639	2.469
		NYHA Class Ⅳ	0.113	0.999	−1.836	2.063
	NYHA Class Ⅲ	NYHA Class Ⅰ	1.406	0.046	0.016	2.796
		NYHA Class Ⅱ	−0.415	0.952	−2.469	1.639
		NYHA Class Ⅳ	−0.301	0.944	−1.717	1.114
	NYHA Class Ⅳ	NYHA Class Ⅰ	1.707	0.002	0.489	2.926
		NYHA Class Ⅱ	−0.113	0.999	−2.063	1.836
		NYHA Class Ⅲ	0.301	0.944	−1.114	1.717

Because RDW violated the equal-variance assumption, the Games–Howell test was applied. Pairwise mean differences, p-values, and 95% CI are presented. Statistically significant findings are indicated with an asterisk (*)

Multivariate analysis

A multivariate analysis of variance (MANOVA) was performed to assess the joint effects of clinical predictors on the hematologic outcomes (hematocrit, RDW, and lymphocyte count). Predictors included age, sex, NYHA class, and comorbidities (diabetes, hypertension, other cardiovascular disease, anemia, and renal disease) (Table 7). The MANOVA revealed significant multivariate effects for age (Wilks’ λ = 0.90; F(3,202) = 6.67; P < 0.001), sex (Wilks’ λ = 0.95; F(3,202) = 2.80; P = 0.040), NYHA class (Wilks’ λ = 0.80; F(9,618) = 6.25; P < 0.001), and anemia (Wilks’ λ = 0.84; F(3,202) = 10.48; P < 0.001). Other cardiovascular diseases showed a marginal effect (Wilks’ λ = 0.96; F(3,202) = 2.80; P = 0.040), whereas diabetes mellitus (P = 0.123), hypertension (P = 0.321), and renal disease (P = 0.123) were not significant multivariate predictors (Table 7). To see the effect of individual predictors (comorbidities) on the three haematological parameters separately please see supplementary Tables 1–3.

Table 7.

Multivariate analysis of variance (MANOVA) testing for joint effect of predictors on RDW, ALC and Hct

Predictor	Wilks’ Lambda	F-value	df	p-value
Age	0.90	6.67	3	< 0.001*
Gender	0.95	2.80	3	0.040*
NYHA	0.80	6.25	9	< 0.001*
Diabetic	0.97	1.94	3	0.123
Hypertensive	0.98	1.17	3	0.321
CVD	0.96	2.80	3	0.040
Anemic	0.84	10.48	3	< 0.001*
Renal	0.97	1.94	3	0.123

Joint effects of demographic and clinical predictors (age, sex, NYHA class, diabetes, hypertension, cardiovascular disease, anemia, and renal disease) on hematocrit, RDW, and lymphocyte count. Reported values include Wilks’ Lambda, F-values, degrees of freedom (df), and p-values. Significant results (*p< 0.05) are note

Receiver operating characteristic (ROC) analysis

Receiver operating characteristic (ROC) curve analysis was conducted to evaluate the ability of haematological parameters in predicting the CHF across the different stages. The top three hematologic markers were (absolute lymphocyte count, RDW, and hematocrit) to distinguish early-stage (NYHA I) from advanced (NYHA III or IV) CHF, based on the calculated area under the curve (AUC). For each marker we determined optimal cut-off values by maximizing the Youden index (sensitivity + specificity).

Comparing NYHA class I versus III, absolute lymphocyte count had an AUC of 0.657 (95% CI 0.543–0.771; P = 0.012); at the optimal cut-off of 1.718×10⁹/L, sensitivity was 50.0% and specificity 79.2%. RDW had an AUC of 0.663 (95% CI 0.549–0.777; P = 0.009); at a cut-off of 13.5%, sensitivity was 87.5% and specificity 40.0%. Hematocrit did not significantly discriminate class I from III (AUC = 0.412; P = 0.106) (Table 8).

Table 8.

Receiver operating characteristic (ROC) analysis comparing NYHA class I versus class III

parameters	AUC (95% CI)	p-Value	Cut-off points	Sensitivity (%) 95% CI	Specificity (%) 95% CI	LR+	LR-
RDW (%)	0.663(0.549, 0.777)	0.009*	13.5	87.5 (76.4, 93.8)	40.0 (30.1, 50.8)	1.46	0.31
Hct (%)	0.412(0.306, 0.518)	0.106	39.15	42.0 (29.9, 55.0)	97.1 (90.7, 99.2)	14.5	0.60
Lymph (109/L)	0.657(0.543, 0.771)	0.012*	1.718	50.0 (37.5, 62.5)	79.2 (69.5, 86.4)	2.40	0.63

Diagnostic performance of RDW, hematocrit, and lymphocyte count in distinguishing early-stage (NYHA I) from more advanced CHF (NYHA III). Results are presented as area under the curve (AUC) with 95% CI, optimal cut-off values, sensitivity, and specificity, LR-Likelihood ratio, Significant results are marked with an asterisk (*)

In the comparison of NYHA class I versus IV, RDW demonstrated stronger discrimination (AUC = 0.708; 95% CI 0.608–0.808; P < 0.001). Using a cut-off of 15.9%, sensitivity was 53.6% and specificity 82.5%. Absolute lymphocyte count had AUC = 0.622 (95% CI 0.514–0.729; P = 0.035); at a cut-off of 1.95×10⁹/L, sensitivity was 42.5% and specificity 78.0%. Again, hematocrit provided no predictive value (AUC = 0.520; P = 0.725) (Table 9). Collectively, these results indicate that RDW and lymphocyte count offer moderate prognostic discrimination between early and advanced CHF, with RDW performing particularly well for distinguishing NYHA class I from IV. The AUC and the 95% CI for the remaining parameters are presented in supplementary Tables 4 and 5 for all cohort and non-anemic group respectively. RDW, Lymph, and PLR are consistently the top 3 predictors in both the full cohort and the non-anemic subgroup. Performance is slightly higher in the non-anemic subgroup, especially for RDW (0.635 vs. 0.612).

Table 9.

Receiver operating characteristic (ROC) analysis comparing NYHA class I versus class IV

parameters	AUC (95% CI)	p-Value	Cut-off points	Sensitivity (%) 95% CI	Specificity (%) 95% CI	LR+	LR-
RDW (%)	0.708(0.608, 0.808)	0.001*	15.9	53.6 (41.9, 64.9)	82.5 (73.9, 88.8)	3.06	0.56
Hct (%)	0.520(0.411, 0.630)	0.725	40.35	60.9 (49.2, 71.5)	56.3 (46.1, 66.0)	1.39	0.69
Lymph (109/L)	0.622(0.514, 0.729)	0.035*	1.95	42.5 (31.3, 54.5)	78.0 (68.7, 85.2)	1.93	0.74

Diagnostic performance of RDW, hematocrit, and lymphocyte count in differentiating NYHAclass I from NYHA class IV. AUC values with 95% CI, cut-off thresholds, sensitivity, and specificity are shown, LR- Likelihood ratio, Significant results are marked with an asterisk (*)

We have further used the DeLong test to account for potential correlation between the haematological parameters in their prediction during the ROC curve analysis. RDW and absolute lymphocyte count remained the top predictors for the given cut-off points in Table 8 (RDW vs. Hct: p = 0.002, Lymph vs. Hct: p = 0.001) and for Table 9 (RDW vs. Hct: p = 0.005, RDW vs. Lymph: p = 0.048). The detail data for the remaining parameters in anemic and non anemic group is presented in supplementary table 6a, b and 7a, b respectively.

The ROC curves are illustrated in Fig. 1 (NYHA class I vs. III) and Fig. 2 (NYHA class I vs. IV).

Fig. 1.

Receiver operating characteristic (ROC) curves for distinguishing NYHA class I from NYHA class III CHF patients. The graphs show ROC curves for (A) absolute lymphocyte count and (B) red cell distribution width (RDW). The y-axis represents sensitivity (true-positive rate), and the x-axis represents 1 – specificity (false-positive rate). The diagonal reference line (AUC = 0.5) indicates no discriminative power. Curves above the diagonal demonstrate predictive value. The optimal cut-off point for each parameter was determined using the Youden index. AUC values, 95% confidence intervals, and significance levels are reported in Table 8

Fig. 2.

Receiver operating characteristic (ROC) curves for distinguishing NYHA class I from NYHA class IV CHF patients. The graphs show ROC curves for (A) absolute lymphocyte count and (B) red cell distribution width (RDW). Axes are as described in Fig. 1. The ROC analysis illustrates moderate discriminative capacity of lymphocyte count and stronger performance of RDW in differentiating patients with early-stage (class I) from advanced (class IV) heart failure. AUC values, 95% confidence intervals, and significance levels are reported in Table 9

Discussion

This study identified three hematological parameters—red cell distribution width (RDW), hematocrit (Hct), and absolute lymphocyte count (ALC)—as significant markers associated with the severity of chronic heart failure (CHF). The principal findings were threefold. First, red cell distribution width (RDW) and absolute lymphocyte count (ALC) demonstrated the most consistent association with disease severity: both markers differed significantly across NYHA classes in pairwise comparisons and achieved statistically significant discriminatory performance in ROC analysis. Second, although hematocrit (Hct) showed a statistically significant difference across NYHA classes by ANOVA (F = 3.349, P = 0.020), its discriminatory performance in ROC analyses was negligible and non-significant. Third, multivariate modeling confirmed that age, sex, NYHA class, and anemia have joint, statistically significant effects on the combined hematologic outcomes highlighting the multifactorial determinants of hematologic change in CHF.

In ROC analyses comparing NYHA class I with class III, RDW had significant discriminatory potential with an AUC of 0.663 and ALC had an AUC of 0.657 whereas Hct had an AUC of 0.412 (P = 0.106). Similarly, in comparisons of class I versus class IV, RDW performed even better with an AUC of 0.708 and ALC remained modestly discriminative; AUC = 0.622, while Hct had an AUC = 0.520; P = 0.725. These ROC results are supported by the pattern of pairwise post hoc contrasts: RDW differed significantly between class I and classes III and IV. Moreover, ALC was consistently higher in class I compared with classes II–IV. Taken together, these findings justify RDW and ALC as robust hematological parameters for discriminating early versus advanced NYHA classes in our cohort.

Lymphocyte count and disease progression

Absolute lymphocyte count showed a progressive decline with advancing NYHA class, with significant differences between class I and classes II–IV. This trend is consistent with prior studies linking lymphopenia to adverse outcomes in CHF [14, 15]. Lymphocyte depletion is biologically plausible, given the chronic systemic inflammation, oxidative stress, and neurohormonal activation characteristic of CHF. Elevated circulating cortisol and pro-inflammatory cytokines such as IL-6 and TNF-α are known to impair lymphopoiesis and accelerate lymphocyte apoptosis, contributing to immune dysfunction in decompensated HF [9, 16–19]. The association between lymphopenia and higher NYHA class underscores its potential role as a simple immunological marker of disease progression [5, 20, 21].

Red cell distribution width as a prognostic marker

RDW was significantly elevated in advanced CHF, demonstrating the strongest discriminative performance in ROC analysis. These findings align with prior reports showing RDW as a robust predictor of morbidity and mortality in HF [6, 13, 23, 24]. Increased RDW reflects anisocytosis, often driven by iron-restricted erythropoiesis, chronic inflammation, or impaired bone marrow function [25–27]. Previous studies have shown correlations between RDW and inflammatory mediators, including IL-6, TNF-α, soluble cytokine receptors, and high-sensitivity C-reactive protein [27, 28]. The progressive rise in RDW observed in our cohort reinforces the role of systemic inflammation and ineffective hematopoiesis in CHF pathophysiology and supports RDW as a surrogate marker of disease severity [27, 29–31].

Hematocrit and prognostic utility

The analysis of hematocrit levels in patients with heart failure reveals significant variations across different NYHA functional classes. Specifically, patients classified as NYHA class IV exhibit markedly lower hematocrit values compared to those in class II. Despite this, hematocrit demonstrates a relatively limited capacity to predict clinical outcomes when compared to other hematological markers such as red cell distribution width (RDW) and absolute lymphocyte count (ALC). Earlier studies have shown a U-shaped link between hematocrit values and mortality, suggesting that both low and high hematocrit levels are tied to poorer outcomes [31]. This suggests that deviations from normal hematocrit levels, whether low or high, can contribute to increased morbidity and mortality in heart failure patients. While anemia, characterized by low hematocrit, is a common comorbidity in chronic heart failure (CHF), its impact on prognosis appears to be largely indirect [32, 33]. Our findings support the notion that hematocrit can serve as an auxiliary biomarker in the assessment of heart failure severity. However, its utility as a standalone prognostic indicator is limited. When compared to more robust markers such as RDW and ALC, Hct has less discriminative potential (AUC = 0.520; P = 0.725). Consequently, clinicians should consider hematocrit as part of a broader panel of hematological parameters rather than relying on it exclusively for prognostic evaluation in heart failure management.

New York Heart Association (NYHA) classification and anemia were identified as the most significant predictors of hematological abnormalities in patients with congestive heart failure (CHF). These factors demonstrated a strong association with alterations in blood cell counts and other hematological parameters. Conversely, variables such as diabetes mellitus, hypertension, and renal disease did not show direct significant correlations with hematological changes. This suggests that their influence on hematological profiles may be predominantly indirect, mediated through the severity of heart failure and the presence of anemia. The findings underscore the intricate relationship between various comorbid conditions and hematopoietic dysregulation, emphasizing the need for comprehensive management strategies that address both cardiac function and hematological health in CHF patients.

Clinical implications

Our results demonstrate that RDW and ALC can provide clinically meaningful information in routine CHF assessment. In low-resource settings such as Ethiopia, where access to advanced biomarkers and imaging may be limited, these inexpensive and widely available hematological parameters offer practical tools for risk stratification. Incorporating RDW and lymphocyte count into standard CHF monitoring could enable earlier identification of patients at risk of functional decline, facilitating timely intervention. RDW remains the most significant predictor in all groups for NYHA class discrimination. Its predictive potential is even much higher in non-anemic subgroup. This reinforces that RDW is the single most powerful routine hematologic parameter for NYHA class discrimination, especially when anemia is excluded.

Limitations

The study was conducted at a single tertiary hospital, which may limit the generalizability of the findings to broader CHF populations. In addition, the absence of a control group without heart failure restricts the ability to distinguish disease-specific hematologic alterations from background variability. Finally, the sample size, although sufficient for the analyses performed, was modest for multivariable modeling, and the risk of overfitting cannot be fully excluded. Despite these considerations, the study provides novel and clinically relevant insights into the prognostic value of RDW, ALC, and Hct in an under-represented population.

Conclusion

This study demonstrates that specific hematological parameters—particularly red cell distribution width (RDW), absolute lymphocyte count (ALC), and hematocrit (Hct)—are associated with disease severity in chronic heart failure (CHF). All three markers showed measurable variation across NYHA functional classes; however, RDW and ALC provided more consistent discriminatory performance in ROC analyses, while Hct demonstrated prognostic associations that were comparatively less robust. These findings suggest that routine hematological indices, especially RDW and ALC, may serve as practical adjuncts for risk stratification. Given their accessibility, affordability, and ease of integration into standard laboratory testing, these parameters hold promise as adjunctive tools for early identification of patients at risk of progression, particularly in resource-constrained settings. Larger, multicenter studies are warranted to validate these results and clarify their role in global CHF management.

Abbreviations

ACC

American College of Cardiology

ACCF/AHA

American College of Cardiology Foundation/American Heart Association

ADHF

Acute decompensated heart failure AHA American heart association

ALC

Absolute lymphocyte count CAD Coronary artery disease

CBC

Complete blood count CHF Congestive heart failure

ECG

Electrocardiogram ESC European Society of Cardiology

Hct

Hematocrit HF Heart failure

HFpEF

Heart failure with preserved ejection fraction

Hgb

Hemoglobin MCHC Mean cell hemoglobin concentration

MCH

Mean cell hemoglobin MCV Mean cell volume

NLR

Neutrophil to Lymphocyte Ratio MLR Monocyte to lymphocyte ratio

NYHA

New York Heart Association functional classification

HCT

Packed cell volume PLR Platelet to Lymphocyte Ratio

RBC

Red blood cell RDW Red blood cell distribution width

SPHMMC

St. Paul’s Hospital Millennium Medical College

SII

Systemic immune-inflammation index

Authors’ contributions

BY: has designed the study, conducted data collection and analysis, and written the first draft. MN has designed the study, conducted the analysis, and edited the manuscript. Both authors read and approved the final version.

Funding

No funding was available for this study.

Data availability

All relevant data are provided in the manuscript, and the raw data will be shared upon reasonable request of the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.