Abstract
Introduction
The prognostic value of the ratio of haemoglobin to red cell distribution width (HRR) in different types of heart failure (HF) is not well known.
Method and Results
We analysed the long-term prognostic value of HRR in patients with HF using the Cox proportional risk model and Kaplan-Meier method. We reviewed consecutive 972 HF patients. The overall mortality rate was 45.68%. Mortality was 52.22% in the HFrEF group and 40.99% in the HFpEF + HFmrEF group. Cox regression showed that when HRR increased by 1 unit, the risk of all-cause death in all HF patients decreased by 22.8% (HR: 0.772, 95% CI: 0.724, 0.823, p < 0.001), in the HFpEF + HFmrEF group it decreased by 15.5% (HR: 0.845, 95% CI: 0.774, 0.923, p < 0.001), and in the HFrEF group it decreased by 36.1% (HR: 0.639, 95% CI: 0.576, 0.709, p < 0.0001). Subgroup analysis showed that there were interactions between the EF and HRR groups. The group in which HRR best predicted all-cause death from HF was group 1 (EF <40%, HRR <9.45), followed by group 2 (EF <40%, HRR ≥9.45), and group 3 (EF ≥40%, HRR <9.45). HRR had no predictive value in group 4 (EF ≥40%, HRR ≥9.45).
Conclusion
HRR is an important predictor of all-cause mortality in patients with HF, especially HFrEF. There is an interaction between HRR group and LVEF group.
Keywords: Heart failure, Ratio of haemoglobin to red cell distribution width, Ejection fraction, Prognostic value, Interaction
Introduction
Heart failure (HF), a clinical syndrome of impaired ventricular filling or ejection function due to structural or functional abnormalities of the heart, is a highly prevalent disease with substantial risks of rehospitalization, mortality, and morbidity worldwide [1]. European preventive cardiology studies show that HF patients increased from 33.5 million in 1990 to 64.3 million in 2017 globally [2]. According to a 2022 Chinese study, the prevalence of HF in China is 1.1%, with 12.05 million patients [3]. The 5-year mortality rate of HF is 55.4% [4]. HF has emerged as a pressing public health issue, necessitating proactive identification of new risk factors even in cases where drug interventions are already being administered.
HF has become a growing public health problem that requires active identification of risk factors when medicine has a treatment to offer. The 2016 European Guidelines classify HF into three categories by left ventricular ejection fraction (LVEF): HF with reduced ejection fraction (HFrEF): LVEF <40%; HF with middle-range ejection fraction (HFmrEF): LVEF 40–49%; and HF with preserved ejection fraction (HFpEF): LVEF ≥50% [5]. Different types of HF not only have different diagnostic criteria but also have different pathophysiologies, clinical characteristics, and treatment methods [6].
Haemoglobin (Hb) and red cell distribution width (RDW) are common blood routine indicators. These two indicators have been found to predict the prognosis of patients with HF [7, 8]. The VICTORIA study showed that lower Hb was associated with a higher risk of cardiovascular death, hospitalization for HF, and all-cause mortality [9]. On the one hand, anaemia can accelerate the progression of HF, increasing its risk by 66% compared to non-anaemia [10]. On the other hand, HF itself can lead to the development of anaemia, and the two conditions interact, forming a vicious cycle that increases the risk of death. RDW reflects the variability of red blood cell volume size. High RDW levels are associated with high levels of oxidative stress and chronic inflammation [11]. Recent studies have found that high RDW is associated with an increased risk of all-cause death from CHD [12], atrial fibrillation [13], stroke [14], and HF [15].
The ratio of Hb to RDW (HRR) for various diseases have appeared in some articles. Baseline RDW and Hb levels in patients with chronic HF (CHF) after 2 years of follow-up are important predictors of mortality in this population, for example, as proposed by Łukasz Wołowiec et al. [16]. Jikai Song et al. [17] proposed that there was a negative correlation between Hb and RDW-SD ratio and 3-month readmission in the Chinese elderly patients with HF, and Hb/RDW-SD <1.78 and Hb/RDW-SD> 2.17 both showed a strong negative correlation with 3-month readmission in HF patients. Eldad Rahamim et al. [18] proposed the lower Hb/RDW ratio as a significant independent predictor of the combined endpoint of death or cardiovascular rehospitalization in patients with HF. Wen-Juan Xiu et al. [19] proposed that HRR predicts postoperative mortality after percutaneous coronary intervention in patients with CAHD. Zuoan Qin et al. [20] proposed a non-linear relationship between all-cause mortality and HRR values in ischemic stroke patients with AF. When the HRR value was less than or equal to 9.74, all-cause mortality was inversely associated with the HRR value. In conclusion, HRR demonstrated a good predictive effect in various fields as a new indicator, probably because it standardized RDW and Hb levels in each patient, as a ratio of oxidative stress and chronic inflammation, malnutrition, iron deficiency, bleeding, and chronic kidney disease. Thus, we explored the role of the HRR in predicting HF and quantified the differences in the predictive value of HRR in different types of HF, as well as the interaction between LVEF group and HRR group.
Methods
Study Population
This is a retrospective observational study. Our research data were obtained from the database of the medical system of the First Affiliated Hospital of Kunming Medical University. We continuously collected 1,200 patients with CHF who were admitted from December 2016 to December 2020 and then completed telephone follow-up in March 2022. Eligible patients were those who were admitted with CHF graded as New York Heart Association (NYHA) functional class III or IV, along with a brain natriuretic peptide (BNP) level of ≥500 pg/mL. We excluded patients who died during hospitalization, died within 30 days of follow-up, or lacked blood routine or cardiac ultrasound data. We also excluded patients with malignancies, infectious diseases, blood disorders, or severe renal or liver dysfunction, patients lost to follow-up, and patients unwilling to cooperate with telephone follow-up. Finally, 972 patients with CHF were included in the analysis.
Data Collection
Demographic and clinical information, electrocardiograms, cardiac ultrasound data, and blood samples were collected. Patients’ age and sex, body mass index (BMI), blood pressure (BP), NYHA cardiac classification, past history, drug condition, heart rate (HR), QRS width, QT intervals, left atrium diameter (LAd), left ventricular end-diastolic diameter (LVDd), right atrium diameter (RAd), right ventricle diameter (RVd), LVEF, white blood cells (WBC), red blood cells (RBCs), platelets (PLT), RDW, Hb, D-dimer, sodium, potassium, chloride, albumin (Alb), globulin (Glb), alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), uric acid (UA), and estimated glomerular filtration rate (eGFR) were collected. Blood samples were drawn from all patients after an overnight fast (8–12 h) and sent to the laboratory of the First Affiliated Hospital of Kunming Medical University. The primary endpoint of this study was all-cause mortality. Investigators collected survival data by conducting telephone interviews with patients or their families. Authors have ensured that all data were fully anonymized during or after data collection.
Statistical Methods
The analysis was performed in SPSS 25.0. To assess the prognostic value of HRR, a multivariate Cox proportional hazards model was built to calculate the hazard ratio (HR), 95% confidence interval (CI), and p values, illustrated by the forest plot. The cumulative incidence of long-term follow-up events was analysed by Kaplan-Meier analysis. The RCS (restricted cubic spline) analysis was performed using R 4.2.2.We drew log-rank curves to analyse the differences between survival curves. The interactions between HRR and LVEF were assessed by introducing a cross-product term in the Cox regression analysis models. All reported p values are two tailed. ANOVA was used to make interaction plots. Variables with p values below 0.05 were considered statistically significant.
Results
Baseline Patient Characteristics
Data from 972 HF patients were analysed, including 406 for HFrEF and 566 for HFpEF + HFmrEF. Compared with HFpEF + HFmrEF patients, the HFrEF group had a higher percentage of males, a younger mean age, lower systolic BP, higher BNP and higher UA, wider QRS wave, greater LAd, LVDd, and RVd (p < 0.001), lower WBC, lower RBC, lower Hb, lower Hb/RDW ratio, lower PLT, higher Alb, lower Glb, higher ALT, higher AST, higher D-dimer, and greater RAd (p < 0.05) (Table 1).
Table 1.
Baseline characteristics according to LVEF
| Variables | Total (n = 972) | LVEF | p value | |
|---|---|---|---|---|
| HFrEF | HFpEF+HFmrEF | |||
| n = 406 | n = 566 | |||
| Basic characteristics | ||||
| Male, n (%) | 597 (61.42) | 285 (70.20) | 312 (55.12) | <0.001 |
| Age, years | 67.00±12.29 | 64.12±12.08 | 69.07±12.04 | <0.001 |
| HR, beats/min | 84.18±20.94 | 85.20±19.29 | 83.45±22.04 | 0.201 |
| Systolic BP, mm Hg | 122.31±22.51 | 117.45±19.47 | 125.81±23.86 | <0.001 |
| BMI, kg/m2 | 22.98±3.73 | 22.75±3.74 | 23.15±3.72 | 0.100 |
| NYHA grade, n (%) | 328 (33.74) | 155 (38.18) | 173 (30.57) | 0.013 |
| Medical history | ||||
| Coronary disease, n (%) | 500 (51.44) | 189 (46.55) | 311 (54.95) | 0.01 |
| Hypertension, n (%) | 543 (55.86) | 183 (45.07) | 360 (63.60) | <0.001 |
| Diabetes, n (%) | 266 (27.37) | 96 (23.65) | 170 (30.04) | 0.028 |
| Atrial fibrillation, n (%) | 341 (35.08) | 112 (27.59) | 229 (40.46) | <0.001 |
| Stroke, n (%) | 133 (13.68) | 46 (11.33) | 87 (15.37) | 0.071 |
| Laboratory indicators | ||||
| LgBNP | 3.16±0.27 | 3.28±0.26 | 3.08±0.26 | <0.001 |
| WBC (109/L) | 6.76 (5.44–8.60) | 6.63 (5.33–8.14) | 6.88 (5.50–9.10) | 0.022 |
| RBC (1012/L) | 4.42±0.84 | 4.35±0.91 | 4.47±0.78 | 0.033 |
| Hb, g/L | 133.56±24.70 | 131.37±25.58 | 135.13±23.92 | 0.02 |
| RDW, % | 14.80±1.89 | 14.87±1.80 | 14.75±1.94 | 0.344 |
| Hb/RDW ratio | 9.20±2.12 | 9.01±2.19 | 9.33±2.06 | 0.019 |
| PLT (109/L) | 192.00 (147.00–241.75) | 184.50 (142.00–138.00) | 198.00 (149.75–245.00) | 0.042 |
| Potassium, mmol/L | 3.92±0.60 | 3.94±0.57 | 3.91±0.62 | 0.444 |
| Sodium, mmol/L | 141.26±4.24 | 141.17±4.17 | 141.32±4.30 | 0.584 |
| Chloride, mmol/L | 103.23±4.52 | 102.97±4.33 | 103.41±4.64 | 0.132 |
| Alb, g/L | 37.05±4.79 | 37.49±4.33 | 36.74±5.07 | 0.016 |
| Glb, g/L | 31.30±5.88 | 30.86±5.81 | 31.62±5.90 | 0.046 |
| ALT, IU/L | 24.30 (16.30–40.00) | 25.80 (17.10–45.30) | 23.50 (15.23–37.00) | 0.003 |
| AST, IU/L | 27.00 (20.00–41.00) | 28.30 (21.00–43.28) | 26.20 (19.60–40.00) | 0.037 |
| UA, μmol/L | 477.10 (369.00–588.40) | 514.7 (416.98–632.88) | 446.10 (353.85–557.55) | <0.001 |
| TC, mmol/L | 3.67±0.99 | 3.66±0.97 | 3.68±1.01 | 0.753 |
| eGFR, mL/min | 45.80±19.10 | 44.30±14.68 | 44.48±20.72 | 0.905 |
| LgD-dimer | 2.77±0.52 | 2.82±0.50 | 2.73±0.53 | 0.008 |
| ECG parameters and cardiac ultrasound index | ||||
| QRS wave, ms | 106 (95–130) | 116 (100–150) | 102 (92–119) | <0.001 |
| LAd, mm | 42.64±9.41 | 44.14±8.42 | 41.57±9.93 | <0.001 |
| LVDd, mm | 56.41±12.13 | 63.59±10.27 | 51.30±10.68 | <0.001 |
| RAd, mm | 52.42±12.34 | 53.45±11.91 | 51.68±12.60 | 0.029 |
| RVd, mm | 70 (63–78) | 74 (66–81) | 68 (62–75) | <0.001 |
| LVEF, % | 43 (33–58) | 31 (26–35) | 55.5 (46–66) | <0.001 |
| Drug conditions | ||||
| SGLT-2I, n (%) | 188 (19.34) | 79 (19.46) | 109 (19.26) | 0.938 |
| β-Receptor blockers, n (%) | 742 (76.34) | 300 (73.89) | 442 (78.09) | 0.129 |
| Diuretics, n (%) | 796 (81.89) | 323 (79.56) | 473 (83.57) | 0.109 |
| ACEI/ARB/ARNI, n (%) | 526 (54.12) | 198 (48.77) | 328 (57.95) | 0.092 |
Normally distributed continuous variables were compared between groups by the independent-sample t test, and non-normally distributed data were compared using the Mann-Whitney U rank-sum test. The χ2 test was used to compare categorical variables between groups.
NYHA grade, NYHA IV grade; TC, total cholesterol; SGLT-2I, sodium-glucose cotransporter 2 inhibitor; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; ARNI, angiotensin receptor-enkephalinase inhibitor.
p values compare the HFpEF + HFmrEF group with the HFrEF group. p < 0.05 indicated statistical significance.
Survival Differences in Patients with Different Types of HF
The median survival time in all patients was 1,236 days, and the all-cause mortality rate was 45.68%. In the HFrEF group, the median survival time was 901.00 ± 97.26 days, and the all-cause mortality rate was 52.22%. In the HFpEF + HFmrEF group, the all-cause mortality rate was 40.99%. There was a significant difference in all-cause mortality between the HFrEF group and the HFpEF + HFmrEF group (p < 0.001). All-cause mortality was higher in the HFrEF group than in the HFpEF + HFmrEF group among male patients only (52.28 vs. 39.10%, p = 0.001), while among female patients, there was no significant difference in all-cause mortality between the HFrEF and HFpEF + HFmrEF groups (p = 0.001, p < 0.05) (Fig. 1 and Table 2).
Fig. 1.
Cumulative probability function of death in different groups. Log-rank test, p = 0.001. It can be considered that the difference in the HFrEF group and the HFpEF + HFmrEF group was statistically significant.
Table 2.
Comparison of all-cause deaths between HFrEF and HFpEF + HFmrEF group
| Comparison/group | HFrEF (n = 406) | HFpEF+HFmrEF (n = 566) | Total (n = 972) | p value |
|---|---|---|---|---|
| Median survival time, days | 901.00±97.26 | – | 1,236 | – |
| Age, years | 64.12±12.08 | 69.07±12.04 | 67.00±12.29 | <0.001 |
| Total all-cause mortality, n (%) | 212 (52.22) | 232 (40.99) | 444 (45.68) | 0.001 |
| All-cause mortality, n (%) | ||||
| Male | 149 (52.28) | 122 (39.10) | 271 (45.40) | 0.001 |
| Female | 63 (52.07) | 110 (43.31) | 173 (46.13) | 0.112 |
Quantitative data with a normal distribution are described as X ± S; comparisons were performed using the t test. The categorical data are described as n (%); comparisons were performed using the χ2 test.
p value <0.05 indicated statistical significance. p values compare the HFpEF + HFmrEF group with the HFrEF group.
Investigating the Relationship between HRR and Survival in Individuals with HF Using Restricted Cubic Splines
In this study, we collected data on the outcome survival, the continuous predictor variable HRR. Possible non-linear relationships between the change in HRR and survival were examined by a Cox regression model with RCS. We conducted RCS with 4 knots to flexibly model the association. The RCS analysis suggested a non-linear association of HRR with survival. The inflection point of the RCS curve was identified at HRR = 9.45, representing a turning point in the relationship between the HRR and the survival (Fig. 2). It can be seen that HRR is less than 9.45, the risk of death decreases rapidly with the increase of HRR, HRR is greater than 9.45, and the risk of death does not decrease with the increase of HRR.
Fig. 2.
Restrictive cubic spline (RCS) of the HRR. Association between HRR and survival with the RCS function. Model with 4 knots located at 5th, 35th, 65th, and 90th percentiles. Y-axis represents the HR to present survival for any value of HRR compared to individuals with 9.45 of HRR.
Predictive Role of HRR in Different EF Groups in the Multivariable Cox Regression Model
We included 34 risk factors, such as age, sex, past medical history, ALT, AST, eGFR, potassium, sodium, chloride, WBC, RBC, PLT, RDW, Hb, HRR, BNP, and D-dimer, in the multivariable Cox regression model. In the HFrEF group, each 1-unit increase in HRR, as a continuous variable, was associated with a 36.2% lower risk of all-cause death (HR: 0.638, 95% CI: 0.575–0.708, p < 0.0001). When HRR was dichotomized around the median, patients with HRR ≥9.45 had a 45.4% lower risk of death than patients with HRR <9.45 (HR: 0.546, 95% CI: 0.308–0.782, p = 0.001). In the HFpEF + HFmrEF group, each 1-unit increase in HRR was associated with a 15.3% lower risk of all-cause death (HR: 0.847, 95% CI: 0.775–0.924, p < 0.0001). When HRR was analysed as a categorical variable, patients with HRR ≥9.45 had a 30.4% lower risk of death than patients with HRR <9.45 (HR: 0.696, 95% CI: 0.507–0.956, p = 0.025) (Table 3).
Table 3.
Predictive value of HRR among different EF groups
| HFrEF group (HR, 95% CI, p value) | HFpEF+HFmrEF group (HR, 95% CI, p value) | Total (972 people) (HR, 95% CI, p value) | |
|---|---|---|---|
| HRRa | 0.638 (0.575, 0.708, 0.000) | 0.847 (0.775, 0.924, 0.000) | 0.772 (0.724, 0.822, 0.000) |
| HRR <9.45 | Reference | Reference | Reference |
| HRR ≥9.45 | 0.546 (0.381, 0.782, 0.001) | 0.696 (0.507, 0.956, 0.025) | 0.629 (0.500, 0.791, 0.000) |
The multivariate Cox regression included 34 correction factors: sex, age, grade NYHA, previous history (coronary heart disease, hypertension, diabetes, peripheral vascular disease, atrial fibrillation), heart rate, QRS width, OT interval, LAd, LVDd, RAd, RVd, WBC, RBC, PLT, potassium, sodium, chloride, albumin, globulin, ALT, AST, TBIL, UA, eGFR, systolic BP, diastolic BP, BMI, lgBNP, lgD-dimer, and HRR.
aRepresents HRR as a continuous variable, and the above HRR value is the reference.
Interaction Analysis of Risk Factors Affecting All-Cause Death from HF
EF Interacts with HRR
The interaction map generated by ANOVA is illustrated in Figure 3 (interaction plot). The two lines intersect clearly, meaning that there was a strong interaction, i.e., that the accuracy of using HRR to predict all-cause death from HF was influenced by LVEF. To investigate whether other factors related to HF outcomes had a joint effect on HRR, we conducted multivariate Cox regression analyses and calculated the interaction p values between different variable subgroups (age, sex, LVEF, NYHA grade, and lgBNP [levels of BNP were normalized by log10 transformation]) and HRR. The predictive value of HRR for all-cause death from HF within each subgroup was then presented via a forest map, as shown in Figure 4 (forest map). Our results illustrate that HRR interacted with EF and age. In HFrEF, the risk of all-cause death in patients with HRR ≥9.45 was 0.546 times that in patients with HRR <9.45. In HFpEF + HFmrEF, the risk of all-cause death in patients with HRR ≥9.45 was 0.696 times that in patients with HRR <9.45. Combining these findings with the interaction plot, we can conclude that there is indeed an interaction between HRR grouping and LVEF grouping.
Fig. 3.
LVEF and HRR interaction plot. Interaction between HRR and LVEF groups, p value <0.001. The interaction plot suggests an interaction between HRR and LVEF.
Fig. 4.
Interaction forest plot. p value <0.05 indicated statistical significance. The multivariate Cox regression included 34 correction factors: sex, age, grade NYHA, previous history (coronary heart disease, hypertension, diabetes, peripheral vascular disease, atrial fibrillation), heart rate, QRS width, OT interval, LAd, LVDd, RAd, RVd, WBC, RBC, PLT, potassium, sodium, chloride, albumin, globulin, ALT, AST, TBIL, UA, eGFR, systolic BP, diastolic BP, BMI, lgBNP, lgD-dimer, and HRR. HR of calculated HRR values with 95% CI and p value.
K-M Curves for Different HRR Groups under Different EF Groups
To investigate the predictive value of different HRR values in different LVEF groups, patients were divided into four groups: group 1: EF <40% and HRR <9.45; group 2: EF <40% and HRR ≥9.45; group 3: EF ≥40% and HRR <9.45; group 4: EF ≥40% and HRR ≥9.45. We analysed survival probability between groups using the K-M method. There was no significant difference in patient survival between group 2 and group 4. Group 1 had the lowest probability of survival, group 3 had the second lowest probability of survival, and group 2 and group 4 both had the third lowest probability of survival (Fig. 5: K-M survival analysis plot).
Fig. 5.
K-M method survival analysis curves for HRR group and LVEF group. p value <0.05 indicated statistical significance. The data were divided into four groups based on the values of different LVEF (1: HFrEF; 2: HFpEF + HFmrEF) and different HRR (1: HRR <9.45; 2: HRR ≥9.45) (group 1: EF <40% and HRR <9.45; group 2: EF <40% and HRR ≥9.45; group 3: EF ≥40% and HRR <9.45; group 4: EF ≥40% and HRR ≥9.45).
The HRR Values in the Four Groups of Patients with Different LVEF Predict Poor Prognosis
There was a difference in the value of HRR in predicting the risk of all-cause mortality between the four groups (Fig. 6). The results showed that the death risk was reduced by 49.2% (HR: 0.508, 95% CI: 0.443, 0.584, p < 0.001) when HRR increased by 1 unit. For each unit increase in HRR in group 2, death risk decreased by 33.0% (HR: 0.670, 95% CI: 0.470, 0.953, p = 0.026). For each 1-unit increase in HRR group 3, the risk of death decreased by 18.8% (HR: 0.812, 95% CI: 0.716, 0.921, p = 0.001). In group 4, the HRR had no value in predicting all-cause death in HF patients (p = 0.312).
Fig. 6.
Prognostic difference of HRR values between the four groups. The multivariate Cox regression included 34 correction factors: sex, age, grade NYHA, previous history (coronary heart disease, hypertension, diabetes, peripheral vascular disease, atrial fibrillation), heart rate, QRS width, OT interval, LAd, LVDd, RAd, RVd, WBC, RBC, PLT, potassium, sodium, chloride, albumin, globulin, ALT, AST, TBIL, UA, eGFR, systolic BP, diastolic BP, BMI, lgBNP, lgD-dimer, and HRR. HR of calculated HRR values with 95% CI and p value.
Discussion
Our results indicate that patients with HFrEF have a higher all-cause mortality rate than patients with HFpEF+HFmrEF. This difference was observed among male patients but not among female patients. Additionally, lower HRR was an independent predictor of death, irrespective of CHF subtype (HFrEF or HFpEF + HFmrEF). Our study presents several noteworthy findings that emphasize the prognostic implications of HRR in CHF patients.
First, there were differences in the predictive value of HRR between the various HF subtypes. At the same HRR value, the HFrEF group had a 20.9% higher risk of all-cause death than the HFpEF + HFmrEF group. In this study, HFmrEF was not considered a separate phenotype of HF since it has a similar aetiology and treatment as HFpEF and a comparable mortality rate [21]; mid-range LVEF impairment is suggestive of a HFpEF population with progressive dysfunction. Thus, HFpEF and HFmrEF were jointly analysed. Importantly, the epidemiology of HFpEF + HFmrEF is different from that of HFrEF. Specifically, HFpEF patients are mostly females, older, and more often have hypertension and atrial fibrillation than HFrEF patients, who are younger and had a higher history of ischaemic heart disease, cardiomyopathy, or diabetes and higher BNP [22]. Consistent with previous studies [23], our results showed that the adjusted mortality for HFpEF and HFmrEF was lower than that for HFrEF (52.22 vs. 40.99%, p = 0.001). LVEF is a crucial factor that can influence all-cause death in patients with HF. As HF progresses, activation of the cardiac sympathetic nervous system and the RAAS leads to cardiac remodelling, larger reductions in LVEF, insufficient blood supply to tissues and organs, aggravated chronic inflammatory reactions, liver and kidney dysfunction, malignant arrhythmia, and eventually death [24]. Hence, at the same Hb and RDW values, the risk of death in patients with HFrEF is greater than that in patients with HF PEF + HFmrEF. Even if the HRR value increases by the same unit, the risk of HFrEF-related death is still higher due to the physiopathological differences between HFrEF and HFpEF + HFmrEF.
Second, both LVEF and HRR strongly impact the prognosis of patients with HF. During the interaction test, a significant HRR group × LVEF group interaction was found (interaction p < 0.0001). There is no study on whether there is an interactive effect on HF outcome between HRR and LVEF. Therefore, we further studied the relationship between different HRR groups and different EF groups. After adjusting for 34 influencing factors, HRR levels were found to be independently correlated with mortality in some CHF subgroups. The highest predictive value of HRR was observed in group 1 (EF <40% and HRR <9.45), and the second highest predictive value of HRR was observed in group 2 (EF <40% and HRR ≥9.45). In group 3 (EF ≥40% and HRR <9.45), the third highest predictive value of HRR was observed. However, in group 4 (EF ≥40% and HRR ≥9.45), HRR had no value in predicting all-cause death in HF patients. This could be because the ejection fraction of patients in group 4 is ≥ 40%, and the HRR level is above the 50th percentile, causing offsetting self-compensation of changes in Hb and RDW. In Group 1, HRR had the highest predicted all-cause death risk value because patients had both low LVEF and low HRR. HRR was always below the 50th percentile of HRR in the HF population, and the systemic effect caused by changing by one unit was more pronounced, making patients more likely to die. Therefore, the best predictive HRR value was found in group 1.
Third, regardless of the subgroup, HRR below the 50th percentile consistently predicts all-cause mortality independently in patients with HF. A patient with an HRR of less than 9.45 has lower Hb levels and a higher RDW than more than 50% of HF patients. Changing the HRR by one unit increases the risk of death more than for patients with HRR ≥9.45. Both Hb levels and RDW have been shown to be associated with the prognosis of HF. Anaemia can cause a series of haemodynamic, neurohormonal, and kidney changes that increase myocardial load. Studies have found that patients with lower Hb levels have lower cardiac output and more severe myocardial damage, worse kidney function, and adverse left ventricular remodelling and hypertrophy, making it a risk factor for patients with HF [25]. Furthermore, increasing RDW is a predictor of multiple cardiovascular diseases’ morbidity and mortality [26]. Several studies have shown that increased RDW is associated with increased all-cause mortality from HF [27], coronary heart disease [28], atrial fibrillation [29], stroke [30], and other diseases. Felker et al. [31] (2007) first identified the potential predictive function of RDW in CHF patients, and since then, increasing evidence showed that RDW is a key factor in the incidence of HF and the risk of poor prognosis [32]. A meta-analysis reported the advantage of including RDW and Hb into the prognosis ratio, showing for the first time that it is a powerful tool for predicting cardiovascular disease outcome [33]. On the one hand, both parameters based on Hb and RDW have a significant prognostic value for the prognosis of HF, and combining these parameters has a significant prognostic utility as a ratio for HF; on the other hand, the pathophysiology of HF is complex and highly variable, with many overlapping pathogenic mechanisms of different importance; minor changes (such as atrial fibrillation or ischaemia) may lead to the development of HF. Regardless of baseline cardiac status, various “amplification mechanisms” persist and lead to cardiac decompensation. These mechanisms include neurohormones and inflammatory activation, among others [34].
Our study described the Hb and RDW situation of the patient through the HRR value; the lower the HRR value, the heavier the anaemia, the wider the red blood cell volume distribution, and the higher the mortality rate. Multiple other studies have found that decreased HRR suggests a poor prognosis of atrial fibrillation [20], cerebral infarction [35], cancer [36], or other diseases. Therefore, HRR can be a simple yet practical new indicator for predicting all-cause death risk in HF patients.
Conclusion
HRR is an important prognostic tool for predicting mortality outcome in patients with HF. HRR had better predictive value for all-cause death in HFrEF patients than in patients with HFpEF + HFmrEF. There was an interaction between LVEF group and HRR group. HRR was most effective in predicting all-cause death from HF in group 1 (EF <40% and HRR <9.45), followed by group 2 (EF <40% and HRR ≥9.45) and group 3 (EF ≥40% and HRR <9.45). In group 4 (EF ≥40% and HRR ≥9.45), HRR had no predictive value for all-cause death from HF. HRR below the 50th percentile has value in predicting the risk of all-cause death in different types of HF.
Acknowledgments
All authors have reviewed the final version of the manuscript and approved it for publication. We certify that this manuscript has not been published in whole or in part nor is it being considered for publication elsewhere.
Statement of Ethics
The study was conducted in accordance with the Declaration of Helsinki and approved by Medical Ethics Committee of the First Affiliated Hospital of Kunming Medical University. The ethics number of the study was (2022) Ethics L No.173. Written informed consent was obtained from participators in this study. All participators signed informed consent at follow-up.
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
The research was funded by the Yunnan Provincial Health Commission Clinical Medical Center (ZX2019-03-01) and by the Applied Basic Research Program of the Science and Technology Hall of Yunnan Province and Kunming Medical University (Project No. 202301AY070001-130).
Author Contributions
Jing Zhou, Wenfang Ma, Yu Wan, Yanji Zhou, Wen Wan, and Lixing Chen conceptualized and designed the survey, conducted the statistical analyses, drafted the first manuscript, and approved the final manuscript as submitted. Lixing Chen, Chenggong Xu, Hongxia Li, Jing Zhou, and Wenyi Gu conducted the data collection and statistical analyses. All authors agreed to the submission of the final manuscript.
Funding Statement
The research was funded by the Yunnan Provincial Health Commission Clinical Medical Center (ZX2019-03-01) and by the Applied Basic Research Program of the Science and Technology Hall of Yunnan Province and Kunming Medical University (Project No. 202301AY070001-130).
Data Availability Statement
The date that support the findings of this study are not publicly available due to ethical issues but are available from corresponding author upon reasonable request.
References
- 1. Bozkurt B, Coats AJ, Tsutsui H, Abdelhamid M, Adamopoulos S, Albert N, et al. Universal definition and classification of heart failure: a report of the Heart Failure Society of America, Heart Failure Association of the European Society of Cardiology, Japanese Heart Failure Society and Writing Committee of the Universal Definition of Heart Failure. Eur J Heart Fail. 2021;23(3):352–80. [DOI] [PubMed] [Google Scholar]
- 2. Bragazzi NL, Zhong W, Shu J, Abu Much A, Lotan D, Grupper A, et al. Burden of heart failure and underlying causes in 195 countries and territories from 1990 to 2017. Eur J Prev Cardiol. 2021;28(15):1682–90. [DOI] [PubMed] [Google Scholar]
- 3. Yang Y, Jiang C, Fu G. Interpretation of Chinese expert consensus 2022 on the comprehensive management of patients with chronic heart failure exacerbation (J). Cardio-Cerebrovascular Dis Prev Treat. 2022;22(02):1–3. [Google Scholar]
- 4. Summary of China cardiovascular health and disease report 2021 (J). Chin Circ J. 2022;37(06):553–78. [Google Scholar]
- 5. Ponikowski P, Voors AA, Anker SD, Bueno H, Cleland JGF, Coats AJS, et al. , ESC Scientific Document Group . 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC) developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J. 2016;37(27):2129–200. [DOI] [PubMed] [Google Scholar]
- 6. Liang M, Bian B, Yang Q. Characteristics and long-term prognosis of patients with reduced, mid-range, and preserved ejection fraction: a systemic review and meta-analysis. Clin Cardiol. 2022;45(1):5–17 Epub 2022 Jan 18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yamauchi T, Sakata Y, Takada T, Nochioka K, Miura M, Tadaki S, et al. , CHART-2 investigators . Prognostic impact of anemia in patients with chronic heart failure with special reference to clinical background: report from the CHART-2 study. Circ J. 2015;79(9):1984–93. [DOI] [PubMed] [Google Scholar]
- 8. Pilling LC, Atkins JL, Kuchel GA, Ferrucci L, Melzer D. Red cell distribution width and common disease onsets in 240,477 healthy volunteers followed for up to 9 years. PLoS One. 2018;13(9):e0203504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Ezekowitz JA, Zheng Y, Cohen-Solal A, Melenovský V, Escobedo J, Butler J, et al. Hemoglobin and clinical outcomes in the vericiguat global study in patients with heart failure and reduced ejection fraction (VICTORIA). Circulation. 2021;144(18):1489–99. [DOI] [PubMed] [Google Scholar]
- 10. He SW, Wang LX. The impact of anemia on the prognosis of chronic heart failure: a meta-analysis and systemic review. Congest Heart Fail. 2009;15(3):123–30. [DOI] [PubMed] [Google Scholar]
- 11. Förhécz Z, Gombos T, Borgulya G, Pozsonyi Z, Prohászka Z, Jánoskuti L. Red cell distribution width in heart failure: prediction of clinical events and relationship with markers of ineffective erythropoiesis, inflammation, renal function, and nutritional state. Am Heart J. 2009;158(4):659–66. [DOI] [PubMed] [Google Scholar]
- 12. Y S, Chandregowda, Cp DG. Relationship between red cell distribution width and clinical outcome in patients with acute coronary syndrome. J Assoc Physicians India. 2022;70(4):11–2. [Google Scholar]
- 13. Adamsson Eryd S, Borné Y, Melander O, Persson M, Smith JG, Hedblad B, et al. Red blood cell distribution width is associated with incidence of atrial fibrillation. J Intern Med. 2014;275(1):84–92. [DOI] [PubMed] [Google Scholar]
- 14. Kim J, Kim YD, Song TJ, Park JH, Lee HS, Nam CM, et al. Red blood cell distribution width is associated with poor clinical outcome in acute cerebral infarction. Thromb Haemost. 2012;108(2):349–56. [DOI] [PubMed] [Google Scholar]
- 15. Uemura Y, Shibata R, Takemoto K, Uchikawa T, Koyasu M, Watanabe H, et al. Elevation of red blood cell distribution width during hospitalization predicts mortality in patients with acute decompensated heart failure. J Cardiol. 2016;67(3):268–73. [DOI] [PubMed] [Google Scholar]
- 16. Wołowiec Ł, Rogowicz D, Banach J, Buszko K, Surowiec A, Błażejewski J, et al. Prognostic significance of red cell distribution width and other red cell parameters in patients with chronic heart failure during two years of follow-up. Kardiol Pol. 2016;74(7):657–64. [DOI] [PubMed] [Google Scholar]
- 17. Song J, Yu T, Yan Q, Zhang Q, Wang L. Association of hemoglobin to red blood cell distribution width-standard deviation (RDW-SD) ratio and 3-month readmission in elderly Chinese patients with heart failure: a retrospective cohort study. Int J Gen Med. 2023;16:303–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Rahamim E, Zwas DR, Keren A, Elbaz-Greener G, Ibrahimli M, Amir O, et al. The ratio of hemoglobin to red cell distribution width: a strong predictor of clinical outcome in patients with heart failure. J Clin Med. 2022;11(3):886. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Xiu WJ, Zheng YY, Wu TT, Hou XG, Yang Y, Ma YT, et al. Hemoglobin-to-red-cell distribution width ratio is a novel predictor of long-term patient outcomes after percutaneous coronary intervention: a retrospective cohort study. Front Cardiovasc Med. 2022;9:726025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Qin Z, Liao N, Lu X, Duan X, Zhou Q, Ge L. Relationship between the hemoglobin-to-red cell distribution width ratio and all-cause mortality in ischemic stroke patients with atrial fibrillation: an analysis from the MIMIC-IV database. Neuropsychiatr Dis Treat. 2022;18:341–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Cheng RK, Cox M, Neely ML, Heidenreich PA, Bhatt DL, Eapen ZJ, et al. Outcomes in patients with heart failure with preserved, borderline, and reduced ejection fraction in the Medicare population. Am Heart J. 2014;168(5):721–30. [DOI] [PubMed] [Google Scholar]
- 22. Schwinger RHG. Pathophysiology of heart failure. Cardiovasc Diagn Ther. 2021;11(1):263–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Vergaro G, Ghionzoli N, Innocenti L, Taddei C, Giannoni A, Valleggi A, et al. Noncardiac versus cardiac mortality in heart failure with preserved, midrange, and reduced ejection fraction. J Am Heart Assoc. 2019;8(20):e013441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Polyakova EA, Mikhaylov EN, Sonin DL, Cheburkin YV, Galagudza MM. Neurohumoral, cardiac and inflammatory markers in the evaluation of heart failure severity and progression. J Geriatr Cardiol. 2021;18(1):47–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Groenveld HF, Januzzi JL, Damman K, van Wijngaarden J, Hillege HL, van Veldhuisen DJ, et al. Anemia and mortality in heart failure patients: a systematic review and meta-analysis. J Am Coll Cardiol. 2008;52(10):818–27. [DOI] [PubMed] [Google Scholar]
- 26. Wołowiec Ł, Rogowicz D, Banach J, Buszko K, Surowiec A, Błażejewski J, et al. Prognostic significance of red cell distribution width and other red cell parameters in patients with chronic heart failure during two years of follow-up. Kardiol Pol. 2016;74(7):657–64. [DOI] [PubMed] [Google Scholar]
- 27. Borné Y, Smith J, Melander O, Hedblad B, Engström G. Red cell distribution width and risk for first hospitalization due to heart failure: a population-based cohort study. Eur J Heart Fail. 2011;13(12):1355–61. [DOI] [PubMed] [Google Scholar]
- 28. Tonelli M, Sacks F, Arnold M, Moye L, Davis B, Pfeffer M, et al. Relation between red blood cell distribution width and cardiovascular event rate in people with coronary disease. Circulation. 2008;117(2):163–8. [DOI] [PubMed] [Google Scholar]
- 29. Ertaş G, Aydin C, Sönmez O, Erdoğan E, Turfan M, Tasal A, et al. Red cell distribution width predicts new-onset atrial fibrillation after coronary artery bypass grafting. Scand Cardiovasc J. 2013;47(3):132–5. [DOI] [PubMed] [Google Scholar]
- 30. Ramírez-Moreno J, Gonzalez-Gomez M, Ollero-Ortiz A, RoaMontero A, Gómez-Baquero M, Constantino-Silva A. Relation between red blood cell distribution width and ischemic stroke: a case-control study. Int J Stroke. 2013;8(6):E36. [DOI] [PubMed] [Google Scholar]
- 31. Felker GM, Allen LA, Pocock SJ, Shaw LK, McMurray JJV, Pfeffer MA, et al. Red cell distribution width as a novel prognostic marker in heart failure: data from the CHARM program and the Duke databank. J Am Coll Cardiol. 2007;50(1):40–7. [DOI] [PubMed] [Google Scholar]
- 32. Huang Y-L, Hu Z-D, Liu S-J, Sun Y, Qin Q, Qin BD, et al. Prognostic value of red blood cell distribution width for patients with heart failure: a systematic review and meta-analysis of cohort studies. PLoS One. 2014;9(8):e104861–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Hou H, Sun T, Li C, Li Y, Guo Z, Wang W, et al. An overall and dose-response meta-analysis of red blood cell distribution width and CVD outcomes. Sci Rep. 2017;7:43420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Song J, Yu T, Yan Q, Zhang Q, Wang L. Association of hemoglobin to red blood cell distribution width-standard deviation (RDW-SD) ratio and 3-month readmission in elderly Chinese patients with heart failure: a retrospective cohort study. Int J Gen Med. 2023;16:303–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Wang J, Chen Z, Yang H, Li H, Chen R, Yu J. Relationship between the hemoglobin-to-red cell distribution width ratio and all-cause mortality in septic patients with atrial fibrillation: based on propensity score matching method. J Cardiovasc Dev Dis. 2022;9(11):400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Petrella F, Casiraghi M, Radice D, Cara A, Maffeis G, Prisciandaro E, et al. Prognostic value of the hemoglobin/red cell distribution width ratio in resected lung adenocarcinoma. Cancers. 2021;13(4):710. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The date that support the findings of this study are not publicly available due to ethical issues but are available from corresponding author upon reasonable request.