Abstract

Objective:

This study aimed to investigate the association between the systemic immune-inflammatory index (SII) and mortality in patients with heart failure (HF).

Method:

We conducted a retrospective cohort study of 1,084 HF patients. In this retrospective cohort study, we enrolled patients hospitalized for heart failure between January 2022 and June 2023. Follow-up was conducted via telephone and outpatient visits until death or July 22, 2025, with all-cause and cardiovascular mortality as primary endpoints. Patients were categorized by log-transformed SII (LnSII). Cox models assessed associations between LnSII and mortality, while restricted cubic splines evaluated nonlinearity. Subgroup, mediation (NT-proBNP, LVEF), and sensitivity analyses were performed.

Results:

A higher LnSII was significantly associated with an increased risk of all-cause mortality (fully adjusted HR = 1.59, 95% CI: 1.03–2.46), No statistically significant association was detected with cardiovascular mortality, which may be attributable to the limited number of cardiovascular deaths (n = 60) and consequent reduced statistical power. Subgroup analysis revealed a significant interaction with smoking status (P for interaction=0.023), showing a stronger association between LnSII and all-cause mortality among smokers (HR = 2.41, 95% CI: 1.57–3.68). Mediation analysis indicated that NT-proBNP and LVEF mediated 35.8% and 15.0% of this association, respectively.

Conclusion:

Elevated SII is independently associated with an increased risk of all-cause mortality in HF patients, particularly among smokers, and may serve as a useful prognostic biomarker.

Introduction

Heart failure (HF) represents the terminal stage of various cardiovascular diseases, significantly impairing quality of life. According to the 2024 Summary of China's Cardiovascular Health and Disease Report, the incidence of adult HF remains high. Monitoring data from 2023 indicates that the non-recovery discharge rate (in-hospital death or non-medically advised discharge) among hospitalized HF patients was 10.2%, with an in-hospital mortality rate of 2.6%, a non-medically advised discharge rate of 7.6%, and a 30-day readmission rate of 11.0% (1). Furthermore, the 5-year mortality rate after HF diagnosis reaches 50% (2), and the poor prognosis of HF imposes a significant healthcare burden on patients and the national healthcare system (3, 25). Therefore, strengthening the early detection and treatment of HF is essential to reduce patient mortality. Studies indicate that patients with systemic inflammatory diseases have a significantly higher risk of cardiovascular disease than the general population, suggesting that inflammatory variables can serve as predictive indicators for cardiovascular disease (4).

The Systemic Immune-Inflammation Index (SII) is a newly developed inflammatory biomarker combining lymphocyte, neutrophil, and platelet counts. Studies have confirmed that SII initiates a cascade of cardiac remodeling and promotes pathological changes in HF. However, the relationship between SII and outcome events in HF patients remains unclear (5). Therefore, further investigation into the association between SII and outcome events in HF patients is necessary to provide insights for precision management of HF patients.

Materials and methodsStudy population

This retrospective study consecutively enrolled 2,232 heart failure patients admitted to the Department of Cardiology of a tertiary hospital in China between January 2022 and June 2023. After screening based on inclusion and exclusion criteria, 1,182 cases were identified. With 98 cases lost to follow-up, 1,084 cases were ultimately included. Inclusion criteria: (1) Age 18 years ≤age <90 years; (2) Patients meeting the diagnostic criteria for HEmrEF or HErEF as defined in the *2018 Chinese Guidelines for the Diagnosis and Treatment of Heart Failure* based on symptoms, physical findings, laboratory tests, and echocardiography, or patients with a history of heart failure and a left ventricular ejection fraction (LVEF) ≤ 49% at admission. Exclusion criteria: (1) Incomplete primary data; (2) Malignant tumors, severe infectious diseases, active tuberculosis, autoimmune diseases; (3) Pregnancy; (4) Severe hepatic or renal insufficiency, proteinuria (urine protein ≥2+); (5) In-hospital death; (6) Acute myocardial infarction. The screening process is illustrated in Figure 1.

Researcher screening flowchart.

Clinical data collectionGeneral information

Recorded patient age, gender, systolic blood pressure, diastolic blood pressure, heart rate, height, weight, smoking history, history of hypertension, history of diabetes, history of stroke, prior percutaneous coronary intervention, prior myocardial infarction, cardiomyopathy, valvular heart disease, and comorbidities (including concomitant atrial fibrillation or atrial flutter, ventricular arrhythmias), medication history (β-blockers, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, angiotensin-neprilysin inhibitors, aldosterone receptor antagonists, sodium-glucose cotransporter 2 inhibitors, antiplatelet agents, anticoagulants), echocardiographic parameters [left ventricular ejection fraction (LVEF), left ventricular end-diastolic volume, left ventricular end-systolic volume], NYHA functional class (I,II,III,IV) and calculate body mass index.

Laboratory tests

Within 24 h of admission, obtain the following parameters using standard laboratory methods: white blood cells, red blood cells, hemoglobin, platelets, lymphocytes, monocytes, neutrophils, total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein, lipoprotein a, N-terminal pro-B-type natriuretic peptide (NT-proBNP), Glomerular Filtration Rate (GFR) and calculate SII. SII (×109/L) = Platelet count (×109/L) × Neutrophil count (×109/L)/Lymphocyte count (×109/L).

Study subject grouping

To approximate a normal distribution for SII values, natural logarithms (LnSII) were taken and subjects grouped by quartiles.

Follow-up and endpoint event definition

All patients were followed up via telephone and outpatient records starting from hospital discharge. Follow-up ended upon death or reached a cutoff date of July 22, 2025. Endpoint events were defined as: All-cause mortality: Total deaths due to any cause after hospital discharge. Cardiovascular mortality: Deaths due to acute myocardial infarction, sudden cardiac death, decompensated heart failure, stroke, or other cardiovascular causes (e.g., aortic dissection).

Statistical methods

Statistical analysis was performed using SPSS 26.0 and R 4.5.1 software. Normally distributed variables were expressed as mean ± standard deviation, with intergroup comparisons using independent samples t-tests. Non-normally distributed variables were presented as median (P25, P75), with intergroup comparisons using the Kruskal–Wallis H-test. Categorical variables were expressed as percentages, with intergroup comparisons using chi-square tests or Fisher's exact tests. Cox proportional hazards models were employed to investigate the association between baseline LnSII and all-cause or cardiac mortality, with results presented as hazard ratios (HR) and 95% confidence intervals (CI). The proportional hazards assumption for the Cox models was assessed by testing the correlation between Schoenfeld residuals and survival time. A global test P-value > 0.05 was considered indicative of no violation of the assumption. To assess the stability of the primary findings against sampling variability, we performed bootstrap resampling with 1,000 iterations. In each iteration, the Cox proportional hazards model (Model 4) was refitted, and the hazard ratio (HR) for the highest versus lowest LnSII quartile (Q4 vs Q1) was estimated. Bias-corrected 95% confidence intervals were derived using the percentile method. Confounders were selected based on prior knowledge. To avoid over-adjustment bias, variables potentially located on causal pathways (LVEF and NT-proBNP) were excluded from the primary analysis model. Restricted cubic splines (RCS) were used to assess potential nonlinear associations. Subgroup analyses compared differences and interactions of variable factors on outcome events. Mediation analysis using the Bootstrap method (Mediation package) explored the potential mediating roles of NT-proBNP and LVEF in the relationship between LnSII and all-cause mortality and cardiac mortality. Sensitivity analyses tested the robustness of results. P < 0.05 was considered statistically significant.

Study resultsComparison of baseline characteristics Among HF patients in different LnSII groups

This study included 1,084 HF patients. With a mean follow-up of 29.3 months (range: 2 days to 39 months), 142 all-cause deaths (13.1%) and 60 cardiac deaths (5.5%) were observed. The cohort comprised 853 men (78.7%) and and 231 women (21.3%). The mean age was 62 years. Compared with patients in the lowest quartile (Q1), those in the highest quartile (Q4) of LnSII were older [median (interquartile range): 66 (57.00, 72.00)], had higher prevalence of stroke and hypertension (52.8%), lower smoking history (36.5%), lower LVEF [39 (34, 45)], and faster heart rate [86 (75, 103)]. Regarding laboratory parameters, patients in the highest LnSII quartile exhibited significant differences in NT-proBNP, white blood cells, red blood cells, hemoglobin, platelets, lymphocytes, monocytes, neutrophils and GFR, all with P < 0.001. Baseline characteristics of participants stratified by LnSII quartile are presented in Tables 1, 2.

Baseline dataTotal count (n = 1,084)Q1 (2.36–5.84) (n = 271)Q2 (5.84–6.22) (n = 271)Q3 (6.22–6.66) (n = 271)Q4 (6.66–8.78) (n = 271)x2/t/Z-valueP-valueAge62 (54.3,70.0)62 (54.0,69.0)60 (53.0,67.0)61 (53.8,70.0)66 (57.0,72.0)16.591<0.001Gender (n,%)8.3440.039 Male853 (78.7)220 (81.2)224 (82.4)211 (78.1)198 (73.1) Female231 (21.3)51 (18.8)48 (17.6)59 (21.9)73 (26.9)Pre-existing Conditions and Comorbidities (n,%)Smoking History484 (44.6)129 (47.6)129 (47.4)127 (47.0)99 (36.5)9.6540.022Hypertension486 (44.8)108 (39.9)105 (38.6)130 (28.1)143 (52.8)15.0840.002Type 2 Diabetes256 (23.6)66 (24.4)58 (21.3)59 (21.9)73 (26.9)2.9970.392Cardiomyopathy228 (21.0)59 (21.8)64 (23.5)59 (21.9)46 (17.0)3.9060.272Valvular Heart Disease51 (4.7)11 (4.1)9 (3.3)13 (4.8)18 (6.6)3.7100.295Myocardial Infarction355 (32.7)96 (35.4)96 (35.3)84 (31.9)77 (28.4)4.0920.252PCI398 (36.7)107 (39.5)102 (37.5)104 (38.5)85 (31.4)4.6820.197Cerebrovascular accident64 (5.9)12 (4.4)14 (5.1)9 (3.3)29 (10.7)15.7800.001Atrial fibrillation/flutter237 (21.9)70 (25.8)56 (20.6)48 (17.8)63 (23.2)5.6970.127Ventricular arrhythmia289 (26.7)66 (24.4)74 (27.2)76 (28.1)73 (26.9)1.0950.778Prior medication history (n,%)β-blockers978 (90.2)239 (88.2)254 (93.4)238 (88.1)247 (91.1)5.9230.115ACEI/ARB/ARNI952 (87.8)235 (86.7)243 (89.3)239 (88.5)235 (86.7)1.3270.723MRA766 (70.7)178 (65.7)189 (69.5)191 (70.7)208 (76.8)8.2730.041SGLT2 inhibitors715 (66.0)169 (62.4)181 (66.5)175 (64.8)190 (70.1)3.8410.279Antiplatelet agents680 (62.7)171 (63.1)178 (65.4)170 (63.0)161 (59.4)2.1550.541Anticoagulants259 (23.9)72 (26.6)58 (21.3)59 (21.9)70 (25.8)3.2320.257Heart ratea78.5 (69,92)77 (68,92)79 (69,90)78 (70,91.5)86 (75,103)24.347<0.001Systolic blood pressure (mmHg)a123 (109,138)120 (107,139)124 (110,137)122 (110,139)123 (106,138)2.6100.456Diastolic blood pressure (mmHg)a78 (68,89)76 (67,90)80 (67.5,90)79 (69.5,91)77 (68,88)2.2070.531LVEDV (mL)a176 (141,221)172 (145,215)177 (136,221)178 (129,228)171 (136,211)4.6520.199LVESV (mL)a104 (77,138)103 (82,134)102 (72,143)103 (71.5,140)104 (75,134)1.4280.699LVEF(%)a41 (35,46)41 (35,46)42 (33.5,47)41 (36,45)39 (34,45)7.4480.059HFmrEF547 (50.5)141 (52.0)143 (52.8)139 (51.3)124 (45.8)HFrEFa537 (49.5)130 (48.0)128 (47.2)132 (48.7)147 (54.2)NYHA classa32.150<0.001I58 (5.4)18 (6.6)21 (7.7)13 (4.8)6 (2.2)II345 (31.8)89 (32.8)97 (35.8)81 (29.9)78 (28.8)III435 (40.1)125 (46.1)96 (35.4)110 (40.6)104 (38.4)IV246(22.7)39(14.4)57(21.0)67(24.7)83(30.6)BMI[(kg/m2)]a24.4(22.0,26.8)24.2(21.9,25.9)24.2(22.0,25.8)24.5(22.1,26.5)24.5(21.5,27.0)1.2780.734

Baseline characteristics of HF patients by LnSII quartiles.

PCI, Percutaneous coronary intervention; ACEI, Angiotensin-Converting Enzyme Inhibitor; ARB, Angiotensin Receptor Blocker; ARNI, Angiotensin-N-Endopeptidase Inhibitor; MRA, Mineralocorticoid Receptor Antagonist; SGLT2i, Sodium-Glucose Transporter 2 Inhibitor; LVEDV, Left ventricular end-diastolic volume; LVESV, Left ventricular end-systolic volume; LVEF, Left ventricular ejection fraction; HErEF, Heart Failure with reduced Ejection Fraction; HEmrEF, Heart Failure with mildly reduced Ejection Fraction; BMI, Body Mass Index.

Laboratory resultsTotal count (n = 1,084)Q1 (2.36–5.84) (n = 271)Q2 (5.84–6.22) (n = 271)Q3 (6.22–6.66) (n = 271)Q4 (6.66–8.78) (n = 271)x2/t/Z-valueP-valueNT-proBNP ((pg/mL)a1,705 (582.5, 4,490)1,510 (556, 4,000)1,630 (547.5, 3,490)1,280 (483, 5,705)3,670 (1,310, 10,000)58.323<0.001WBC (×109/L)a6.25 (5.15,7.51)5.50 (4.19,6.75)5.74 (4.55,6.88)6.41 (5.43,7.86)6.90 (5.94,8.33)146.494<0.001RBC (×1012/L)a4.81 (4.37, 5.21)4.75 (4.31, 5.15)4.78 (4.42, 5.16)4.81 (4.31, 5.12)4.64 (4.14, 5.17)14.1570.003HGB (g/L)a151 (137, 163)149 (136, 163)153 (139.5, 163.5)151 (133.5, 160.5)143 (126, 157)33.718<0.001PLT (×109/L)a181 (143, 221)139 (107, 176)160 (135, 193)196 (157.5, 237.5)221 (178, 278)299.301<0.001LYM (×109/L)a1.44 (1.05, 1.86)1.68 (1.27, 2.31)1.43 (1.11,,1.92)1.43 (1.08,,1.83)0.95 (0.73,,1.21)195.964<0.001MON (×109/L)a0.39 (0.31,,0.50)0.37 (0.28, 0.45)0.38 (0.28, 0.47)0.39 (0.31, 0.47)0.42 (0.32, 0.57)44.260<0.001NE (×109/L)a4.11 (3.29, 5.20)3.24 (2.44, 3.93)3.68 (3.01, 4.48)4.59 (3.71, 5.32)5.38 (4.49, 6.75)381.352<0.001TC (mmol/L)a3.50 (2.93, 4.24)3.52 (2.98, 4.29)3.32 (2.90, 4.06)3.35 (2.78, 4.36)3.47 (2.94, 4.12)1.5110.690TG (mmol/L)a1.31 (0.94, 1.87)1.44 (0.94, 2.13)1.40 (0.95, 1.99)1.27 (0.89, 1.77)1.19 (0.92, 1.68)6.2500.100HDL-C (mmol/L)a0.97 (0.83, 1.13)0.94 (0.82, 1.09)0.96 (0.83, 1.10)0.96 (0.82, 1.14)0.95 (0.79, 1.22)1.0470.790LDL-C (mmol/L)a2.19 (1.78, 2.75)2.25 (1.81, 2.84)2.04 (1.72, 2.66)2.10 (1.66, 2.79)2.21 (1.81, 2.71)2.4760.480Lp (a)(g/L)a14.24 (6.6, 31.38)8.8 (5.04, 26.92)11.89 (6.55, 26.02)13.83 (6.49, 28.96)16.61 (7.74, 36.63)4.3180.229GFR88.1 (76.4, 102.0)89.9 (77.6, 102.0)90.5(81.0, 103.0)88.5(77.0, 103.0)83.7(67.4, 98.6)19.814<0.001

Laboratory characteristics of HF patients by LnSII quartiles.

NT-proBNP, N-terminal pro-B-type natriuretic peptide precursor; WBC, white blood cell count; RBC, red blood cell count; HGB, hemoglobin count; PLT, platelet count; LYM, lymphocyte count; MON, monocyte count; NE: neutrophil count; TC, total cholesterol count; TG, Triglyceride count; HDL-C, High-density lipoprotein cholesterol; LDL-C, Low-density lipoprotein cholesterol; Lp(a), Lipoprotein(a); GFR, Glomerular Filtration Rate.

Association of LnSII with all-cause and cardiovascular mortality

During the follow-up period, a total of 142 all-cause deaths (13.1%) and 60 cardiovascular deaths (5.5%) were observed. A Cox proportional hazards model was used to assess the association between LnSII quartiles (Q1: 2.36–5.84; Q2: 5.84–6.22; Q3: 6.22–6.66; Q4: 6.66–8.78) and all-cause mortality (ACM). In models adjusted for various combinations of variables, the risk of all-cause mortality was significantly increased in the fourth quartile (Q4) (HR = 1.59, 95% CI: 1.03–2.46 in the fully adjusted Model 4), whereas the second and third quartiles (Q2, Q3) showed no statistically significant difference compared with the reference group (Q1). To further address potential residual confounders, we constructed an extended Model 4, which additionally included atrial fibrillation/flutter, prior myocardial infarction, and hemoglobin as covariates. The results indicate that elevated SII levels remain an independent predictor of all-cause mortality in patients with heart failure. To further explore the mechanisms underlying SII, this study incorporated markers reflecting cardiac function and stress (LVEF and NT-proBNP) into the adjustment models (Model 5). The HRs (95% CI) for ACM were 1.00 (reference), 0.64 (0.38–1.10), 0.92 (0.56–1.51), and 1.34 (0.86–2.09), respectively. The association between LnSII and mortality was significantly weakened, suggesting that the effect of SII on mortality risk may be partially mediated through exacerbation of cardiac dysfunction and myocardial stress, as shown in Table 3.

HR (95%Cl), P-valueModel1Model2Model3Model4Model5All-cause MortalityLnSII1.64 (1.31-2.06),<0.0011.48 (1.18, 1.85),<0.0011.49 (1.19–1.87),<0.0011.50 (1.20–1.88),<0.0011.29 (1.04–1.61), 0.024LnSII (quartile)Q1Ref.Ref.Ref.Ref.Ref.Q20.64 (0.37–1.10), 0.1060.66 (0.39–1.13), 0.1300.65 (0.38–1.11), 0.1140.69 (0.41–1.19), 0.1820.64 (0.38–1.10), 0.109Q30.89 (0.55–1.46), 0.6590.90 (0.55–1.46), 0.6500.91 (0.55–1.48), 0.6970.97 (0.59–1.59), 0.8890.92 (0.56–1.51), 0.733Q41.78 (1.16–2.73), 0.0081.60 (1.04–2.45), 0.0321.58 (1.03–2.44), 0.0371.59 (1.03–2.46), 0.0361.34 (0.86–2.09), 0.196P for trend<0.0010.0100.0110.0130.073Cardiovascular MortalityLnSII1.20 (0.83–1.74), 0.3371.13 (0.79–1.63), 0.4991.15 (0.80–1.66), 0.4451.09 (0.77–1.53), 0.6380.99 (0.70–1.38), 0.928LnSII (quartile)Q1Ref.Ref.Ref.Ref.Ref.Q20.63 (0.30–1.29), 0.2020.64 (0.31–1.32), 0.2230.64 (0.31–1.32), 0.2290