Nonlinear association between systemic immune-inflammation index and in-hospital mortality in critically ill patients with chronic obstructive pulmonary disease and atrial fibrillation: a cross-sectional study

Introduction

Chronic obstructive pulmonary disease (COPD) is a heterogeneous lung condition characterized by persistent, often progressive airflow obstruction due to abnormalities in the airways (bronchitis, bronchiolitis) and/or alveoli (emphysema), leading to chronic respiratory symptoms such as dyspnea, cough, sputum production, and exacerbations (1). Globally, COPD is one of the leading causes of death and disability, imposing a significant economic and social burden (2). With an increasingly aging population, the global burden of COPD is projected to escalate in the coming decades (3). Atrial fibrillation (AF), the most common arrhythmic disorder, is closely linked to aging and severely impacts public health and safety (4-6). In clinical practice, COPD with AF is quite common, with prevalence estimates reaching up to 23% in individuals over 65 years old (7). The simultaneous occurrence of these two conditions not only complicates disease management but also potentially exacerbates patient outcomes.

Given that inflammation is a shared underlying mechanism between COPD and AF, and considering their frequent coexistence in critically ill patients, it is important to understand how inflammation may impact clinical outcomes in this population. The systemic immune-inflammation index (SII), calculated from neutrophil count (NEUT), lymphocyte count (LYM), and platelet count (PLT), has emerged as a robust and readily accessible biomarker of systemic inflammation (8,9). Elevated SII has been linked to poor prognoses across various diseases, including those with limited-stage small cell lung cancer (10), primary brain tumor (11), colorectal cancer (12), and COPD (13). However, the prognostic significance of SII in intensive care unit (ICU) patients with concurrent COPD and AF remains largely unexplored. Therefore, this study aimed to investigate the association between SII and in-hospital mortality in ICU patients with both COPD and AF, with the hypothesis that higher SII levels are independently associated with increased risk of in-hospital death. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-266/rc).

Methods

Data sources

Data for this study were obtained from the Medical Information Mart for Intensive Care-IV (MIMIC-IV, version 3.1) database, a freely available clinical resource containing information on over 65,000 patients admitted to the ICU at Beth Israel Deaconess Medical Center (BIDMC), a large tertiary hospital in Boston, MA (Massachusetts), USA (United States of America) (14). The MIMIC-IV database includes data from routine hospital care, such as records from critical care systems, electronic health records, and the Social Security Administration Death Master File. It is widely used in academic research and is well-recognized. One author Wei Guo obtained approval to exploit the database (certification number 754593322). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Study population

Patients with COPD and AF were included in our study. COPD and AF were identified according to the International Classification of Diseases, ninth/ten revision (ICD-9, ICD-10) codes. The ICD codes for COPD include J44, J440, J441, J449, 496, 49120, 49121, and 49122. The ICD codes for AF are 42,731 and I480, I481, I482, I4819, I4820, I4821, I4891. The exclusion criteria were as follows: (I) patients without ICU admission or with repeated ICU admissions; (II) patients aged <18 years; (III) patients without COPD; (IV) patients without AF; (V) length of admission to ICU less than <24 hours, (VI) LYM, NEUT, or PLT missing or zero. The patient selection flowchart is presented in Figure 1.

Figure 1 Flow chart of included patients. AF, atrial fibrillation; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; MIMIC-IV, Medical Information Mart for Intensive Care-IV.

Exposure variable

The highest peripheral blood cell count recorded during the first 24 hours of ICU admission was used to calculate the SII using the following formula: SII = PLT (10³/µL) ×NEUT (10³/µL) / LYM (10³/µL).

Covariates

The extracted covariates were classified into six main categories: (I) demographics: age and sex; (II) vital signs: heart rate (HR), respiratory rate (RR), systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean blood pressure (MBP); (III) laboratory parameters: white blood cell count (WBC), NEUT, LYM, PLT, calcium, sodium, potassium, creatinine, and glucose; (IV) comorbidities: conditions such as myocardial infarction, congestive heart failure, diabetes, renal disease, and malignant cancer; (V) admission severity scores: the Charlson Comorbidity Index (CCI), the Simplified Acute Physiology Score II (SAPSII), and the Oxford Acute Severity of Illness Score (OASIS); (VI) treatments: β1-blockers, inhaled β2-agonists, and inhaled anticholinergic agents.

The highest recorded values for vital signs, laboratory tests, and initial severity scores within the first 24 hours of ICU admission were used.

Outcomes

The outcome was in-hospital mortality in ICU patients with COPD and AF between 2008 to 2022.

Statistical analysis

The Kolmogorov-Smirnov test was used to assess the normality of variable distributions. Variables following a normal distribution were presented as mean ± standard deviation (SD), while those with a skewed distribution were expressed as median [interquartile range (IQR): 25th–75th percentiles]. Categorical variables were reported as proportions (%). Differences among the three groups were assessed using the Chi-squared test for categorical variables, one-way analysis of variance (ANOVA) for normally distributed continuous variables, and the Kruskal-Wallis test for skewed distributions. Additionally, since SII data exhibited a non-normal distribution, log transformation was applied to achieve a normal distribution (referred to as log-SII).

We used univariate and multivariate binary logistic regression model to test the link between SII and in-hospital mortality with three distinct models. The effect of log-SII on in-hospital mortality was evaluated using binary logistic regression models [odd ratio (OR) and 95% confidence interval (CI)] with adjustment for major covariables. Model 1: adjusted for sex and age. Model 2: further adjusted for covariates that caused at least a 10% change in the matched odds ratio when included in the model, such as HR, RR, calcium, OASIS and β1-blocker. Model 3: additionally adjusted for WBC, sodium, potassium, creatinine, myocardial infract, congestive heart failure, diabetes, renal disease, malignant cancer, inhaled β2-agonists and inhaled anticholinergic agents based on Model 2, as these covariates are clinically relevant and may influence patient prognosis.

For the multivariate analysis, missing data were handled using multiple imputation (MI) based on 5 replications and a chained equation approach in the R MI procedure. The frequency and percentage distribution of missing data are provided in Table S1. Subsequent analyses were conducted using data with missing values.

In addition, restricted cubic spline (RCS) regression was performed with 4 knots at the 5th, 35th, 65th, and 95th percentiles of log-SII to assess linearity and examine the dose-response curve between log-SII and in-hospital mortality after adjusting variables in Model 3. We used a two-piece-wise logistic regression model with smoothing to analyze the association threshold between log-SII and in-hospital mortality after adjusting the variables in Model 3. The likelihood ratio test was used to determine inflection points.

Furthermore, potential modifications of the relationship between log-SII and in-hospital mortality were assessed, including the following variables: sex, age (<65 vs. ≥65 years), myocardial infract (no vs. yes), congestive heart failure (no vs. yes), diabetes (no vs. yes), renal disease (no vs. yes), malignant cancer (no vs. yes). Heterogeneity among subgroups was assessed by multivariate logistic regression, and interactions between log-SII and in-hospital mortality were examined by likelihood ratio testing.

A priori statistical power calculations were not performed, as the sample size was determined entirely by the available data. All analyses were performed using R statistical software (version 4.2.2, http://www.R-project.org, The R Foundation) and the Free Statistics analysis platform (version 2.0, Beijing, China, http://www.clinicalscientists.cn/freestatistics). A two-tailed P value <0.05 was considered statistically significant.

Results

Baseline characteristics

Of the 364,627 patients in the MIMIC-IV 3.1 database, 299,261 were excluded for not being admitted to or readmitted to the ICU. No participants under 18 years were identified in the study population, thus no exclusions based on age criteria were required. Additionally, 58,147 and 4,544 patients were excluded for not having COPD and AF, respectively. A further 374 patients were excluded due to an ICU stay of less than 24 hours. An additional 1,727 patients were excluded because LYM, NEUT, or PLT were missing or zero. Ultimately, 574 patients were included in the analysis. The patient selection process is shown in Figure 1.

Baseline clinical characteristics of the included patients are summarized in Table 1. Patients were categorized into three groups based on log-SII levels: Q1 (<2.9), Q2 (2.9–3.4), and Q3 (>3.4). Of the patients, 345 (60.1%) were male, with an average age of 74.1±9.4 years. The in-hospital mortality rate was 20.4% (117 patients). Significant differences were observed between the groups in variables such as age, HR, RR, DBP, MBP, WBC, NEUT, LYM, PLT, creatinine, glucose, congestive heart failure, renal disease, SAPSII, OASIS, use of β1-blockers, inhaled β2-agonists, and inhaled anticholinergic agents (all P<0.05).

Association between SII and in-hospital mortality in patients with COPD and AF

In the univariate logistic regression analysis (Table S2), we found that log-SII (hazard ratio =3.22, 95% CI: 1.99–5.20, P<0.001) and high log-SII (hazard ratio =5.09, 95% CI: 2.81–9.24, P<0.001, Q3) were significantly associated with in-hospital mortality in ICU patients with COPD and AF. Other factors related to in-hospital mortality are listed in Table S2.

In the multivariate logistic regression analysis, three models were adjusted. When log-SII was treated as a continuous variable, each unit increase was associated with a hazard ratio of 1.72 (95% CI: 1.00–2.94, P=0.048; Table 2, Model 3). When analyzed as a categorical variable, patients in the high log-SII group had a 2.78-fold higher risk of in-hospital mortality compared to those in the low log-SII group (hazard ratio =2.78, 95% CI: 1.37–5.62, P=0.005), independent of potential confounders (Table 2, Model 3). Furthermore, multivariate logistic regression analyses using unimputed missing data yielded consistent results (Table S3).

Nonlinear association and threshold effect analyses between SII and in-hospital mortality

RCS was used to better characterize and visually represent the relationship between log-SII and in-hospital mortality. As shown in Figure 2, a nonlinear association was observed between log-SII and in-hospital mortality (P for non-linearity =0.01). In the two-piecewise regression models, the adjusted OR for in-hospital mortality was 21.8 (95% CI: 3.3–143.7, P=0.001) in participants with log-SII between 2.9 and 3.6. However, no significant association was found between log-SII and in-hospital mortality in participants with log-SII <2.9 or >3.6 (Table 3).

Figure 2 Relationship between the log-SII and In-hospital mortality. Restricted cubic spline analysis depicting the nonlinear association between log-SII and in-hospital mortality in patients with COPD and AF. Adjusted for all covariates as per Model 3. The red line represents the estimated values, while the green area indicates the corresponding 95% confidence intervals. Only 95% of the data was displayed. Model 1: adjusted for sex, age; Model 2: adjusted for covariates included in Model 1 that caused at least a 10% change in the matched odds ratio, such as HR, RR, calcium, OASIS, and β1-blocker; Model 3: further adjusted for WBC, sodium, potassium, creatinine, myocardial infract, congestive heart failure, diabetes, renal disease, malignant cancer, inhaled β2-agonists and inhaled anticholinergic agents, based on Model 2, as these covariates are clinically relevant and may influence patient prognosis. AF, atrial fibrillation; COPD, chronic obstructive pulmonary disease; HR, heart rate; OASIS, Oxford Acute Severity of Illness Score; RR, respiratory rate; SII, systemic immune-inflammation index; WBC, white blood cell count.

Subgroup analysis

To evaluate the stability of the association between SII and in-hospital mortality, we performed curve fitting analyses stratified by sex, age, myocardial infarction, congestive heart failure, diabetes, renal disease, and malignant cancer. The OR in all subgroups were greater than 1, and no significant interactions were observed (all P for interaction >0.05) (Figure 3).

Figure 3 Subgroup analyses of the association between log-SII and in-hospital mortality. Subgroup analyses for the association between log-SII and in-hospital mortality. Each stratified group was adjusted for the all covariates in Model 3. Model 1: adjusted for sex, age; Model 2: adjusted for covariates included in Model 1 that caused at least a 10% change in the matched odds ratio, such as HR, RR, calcium, OASIS, and β1-blocker; Model 3: further adjusted for WBC, sodium, potassium, creatinine, myocardial infract, congestive heart failure, diabetes, renal disease, malignant cancer, inhaled β2-agonists and inhaled anticholinergic agents, based on Model 2, as these covariates are clinically relevant and may influence patient prognosis. CI, confidence interval; HR, heart rate; OASIS, Oxford Acute Severity of Illness Score; OR, odds ratio; RR, respiratory rate; SII, systemic immune-inflammation index; WBC, white blood cell count.

Discussion

In this study, we observed a nonlinear relationship between SII and in-hospital mortality in ICU patients with COPD and AF, with higher SII levels closely associated with lower in-hospital survival rates. This association remained robust even after adjusting for potential confounders.

The use of SII has garnered significant attention due to its accessibility and established potential as a biomarker in various clinical contexts. Numerous studies have explored SII as a prognostic marker, highlighting its ability to predict diverse survival outcomes. For example, research has demonstrated associations between SII and survival outcomes before and after treatment for colorectal cancer (15), progression-free survival in cervical cancer patients undergoing immunotherapy (16), and all-cause mortality in both cancer patients and those without cancer (17,18). Furthermore, SII has been linked to the incidence of various diseases, including stroke (19), chronic respiratory disease (20) and cataracts (8). Recent studies have also underscored the potential of SII in evaluating the efficacy of immunotherapy (21). Our findings align with previous studies, reinforcing the strong association between higher SII levels and increased in-hospital mortality in patients with COPD and AF. Additionally, we observed a nonlinear relationship between SII and mortality. However, due to limitations in sample size, we were unable to definitively determine whether this relationship follows an S-shaped, or another specific curve type. This uncertainty is reflected in the wide 95% CI in the threshold effect analysis, particularly when log-SII is less than 2.9 (Table 3).

The strong association between SII and COPD with AF can also be explained from a pathophysiological perspective. The pathological changes in COPD involve both inflammatory and structural alterations, which progressively worsen with the severity of airflow obstruction and persist even after smoking cessation (2). The inflammatory response is characterized by an increased number of macrophages, T lymphocytes, and B lymphocytes, along with an elevated presence of neutrophils in the airway lumen (22,23). Cigarette smoke activates alveolar macrophages and epithelial cells, triggering the release of chemotactic factors that attract circulating leukocytes to the lungs, thereby amplifying the inflammatory response (24). Beyond localized pulmonary inflammation, systemic inflammation also plays a crucial role. A study by Gan et al. found that COPD patients exhibited significantly higher levels of inflammatory markers compared to controls, including C-reactive protein, fibrinogen, leukocytes, and tumor necrosis factor-alpha (25). Moreover, systemic inflammation exacerbates comorbid conditions such as cardiovascular disease, diabetes, and osteoporosis (23). As discussed in the introduction, AF is closely linked to COPD, and accumulating evidence suggests that inflammation plays a pivotal role in the initiation and perpetuation of AF, as well as in AF-related thrombosis and electrical and structural atrial remodeling (26-28). A study by Chen et al. further demonstrated that inflammatory cell infiltration is increased in the atrial myocardium of patients with AF (29). Moreover, a recent study revealed that the border region between epicardial adipose tissue (EAT) and atrial tissue is a hotspot for intense inflammatory and fibrotic activity. EAT has been shown to transport tissue-resident memory T cells into atrial cardiomyocytes, where they play a role in modulating calcium flux and activating inflammatory and apoptotic signaling pathways in patients with AF (30). Taken together, these findings highlight the significant role of inflammation in the pathogenesis and progression of both COPD and AF.

Limitation

The findings of this study should be interpreted in light of several limitations. First, as a retrospective cohort study, it inherently cannot establish causal relationship. Instead, our findings highlight a strong association, which may serve as a foundation for future mechanistic or prospective studies. Second, the study is not a randomized controlled trial, leaving the potential for residual confounding. However, multivariable regression analyses and subgroup analyses were performed to minimize this concern. Third, the data were sourced from the MIMIC-IV database, which represents ICU patients from a single institution. This limitation restricts the generalizability of our findings to other populations and healthcare settings. Fourth, the threshold effect analysis should be interpreted with caution given the wider CIs observed at lower Log-SII values, reflecting reduced precision due to sample size limitations. Therefore, further large-scale, multicenter studies are required to validate these results.

Conclusions

We identified a nonlinear association between log-SII and in-hospital mortality in ICU patients with COPD and AF, with higher SII levels correlating with an increased risk of mortality.

Acknowledgments

We extend our gratitude to the MIMIC database for providing the platform and to all contributors for sharing valuable datasets. Special thanks to Qilin Yang from the Department of Critical Care, The Second Affiliated Hospital of Guangzhou Medical University, Guangzhou, China, for assistance in data extraction from the MIMIC database.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-266/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-266/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-266/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Celli B, Fabbri L, Criner G, et al. Definition and Nomenclature of Chronic Obstructive Pulmonary Disease: Time for Its Revision. Am J Respir Crit Care Med 2022;206:1317-25. [Crossref] [PubMed]
2. GOLD report - global initiative for chronic obstructive lung disease - GOLD. Accessed November 28, 2024. Available online: https://goldcopd.org/2025-gold-report/
3. Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med 2006;3:e442. [Crossref] [PubMed]
4. Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the Management of Atrial Fibrillation Developed in Collaboration With EACTS. Rev Esp Cardiol (Engl Ed) 2017;70:50. [Crossref] [PubMed]
5. Schnabel RB, Yin X, Gona P, et al. 50 year trends in atrial fibrillation prevalence, incidence, risk factors, and mortality in the Framingham Heart Study: a cohort study. Lancet 2015;386:154-62. [Crossref] [PubMed]
6. Zhang J, Johnsen SP, Guo Y, et al. Epidemiology of Atrial Fibrillation: Geographic/Ecological Risk Factors, Age, Sex, Genetics. Card Electrophysiol Clin 2021;13:1-23. [Crossref] [PubMed]
7. Simons SO, Elliott A, Sastry M, et al. Chronic obstructive pulmonary disease and atrial fibrillation: an interdisciplinary perspective. Eur Heart J 2021;42:532-40. [Crossref] [PubMed]
8. Mahemuti N, Jing X, Zhang N, et al. Association between Systemic Immunity-Inflammation Index and Hyperlipidemia: A Population-Based Study from the NHANES (2015-2020). Nutrients 2023;15:1177. [Crossref] [PubMed]
9. Zhao Y, Shao W, Zhu Q, et al. Association between systemic immune-inflammation index and metabolic syndrome and its components: results from the National Health and Nutrition Examination Survey 2011-2016. J Transl Med 2023;21:691. [Crossref] [PubMed]
10. Zhang Z, Zhao Y, Wen J, et al. Impact of systemic immune-inflammation index and its evaluation of optimal threshold in patients with limited-stage small cell lung cancer: a retrospective study based on 572 cases. Transl Cancer Res 2025;14:371-82. [Crossref] [PubMed]
11. Pramaswari AAAA, Sofyan HR, Savitri I, et al. AB037. Association between systemic immune-inflammation index with hypercoagulable state in primary brain tumor: a retrospective study in Dr. Cipto Mangunkusumo Hospital, Jakarta, Indonesia. Chin Clin Oncol 2024;13:AB037.
12. Sun J, Yang R, Wu H, et al. Prognostic value of preoperative combined with postoperative systemic immune-inflammation index for disease-free survival after radical rectal cancer surgery: a retrospective cohort study. Transl Cancer Res 2024;13:371-80. [Crossref] [PubMed]
13. Zhang Y, Tan X, Hu S, et al. Relationship Between Systemic Immune-Inflammation Index and Risk of Respiratory Failure and Death in COPD: A Retrospective Cohort Study Based on the MIMIC-IV Database. Int J Chron Obstruct Pulmon Dis 2024;19:459-73. [Crossref] [PubMed]
14. Johnson AEW, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data 2023;10:1. [Crossref] [PubMed]
15. Tan Y, Hu B, Li Q, et al. Prognostic value and clinicopathological significance of pre-and post-treatment systemic immune-inflammation index in colorectal cancer patients: a meta-analysis. World J Surg Oncol 2025;23:11. [Crossref] [PubMed]
16. Chen Q, Zhai B, Li J, et al. Systemic immune-inflammatory index predict short-term outcome in recurrent/metastatic and locally advanced cervical cancer patients treated with PD-1 inhibitor. Sci Rep 2024;14:31528. [Crossref] [PubMed]
17. Wu S, Liu Z, Li X, et al. Association between systemic immune-inflammation index and the risk of all-cause, cancer and non-cancer mortality in the general population: results from national health and nutrition examination survey 2005-2018. BMC Public Health 2025;25:227. [Crossref] [PubMed]
18. Zeng D, Zeng Q, Li S, et al. Evaluating body roundness index and systemic immune inflammation index for mortality prediction in MAFLD patients. Sci Rep 2025;15:330. [Crossref] [PubMed]
19. Shi R, Tian Y, Tian J, et al. Association between the systemic immunity-inflammation index and stroke: a population-based study from NHANES (2015-2020). Sci Rep 2025;15:381. [Crossref] [PubMed]
20. Zhan Y, Yang R, Feng J, et al. Joint association of systemic immune-inflammation index and phenotypic age acceleration with chronic respiratory disease: a cross-sectional study. BMC Public Health 2025;25:186. [Crossref] [PubMed]
21. He T, Xu B, Wang LN, et al. The prognostic value of systemic immune-inflammation index in patients with unresectable hepatocellular carcinoma treated with immune-based therapy. Biomark Res 2025;13:10. [Crossref] [PubMed]
22. Barnes PJ. Cellular and molecular mechanisms of chronic obstructive pulmonary disease. Clin Chest Med 2014;35:71-86. [Crossref] [PubMed]
23. Barnes PJ. Inflammatory mechanisms in patients with chronic obstructive pulmonary disease. J Allergy Clin Immunol 2016;138:16-27. [Crossref] [PubMed]
24. Donnelly LE, Barnes PJ. Chemokine receptors as therapeutic targets in chronic obstructive pulmonary disease. Trends Pharmacol Sci 2006;27:546-53. [Crossref] [PubMed]
25. Gan WQ, Man SF, Senthilselvan A, et al. Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and a meta-analysis. Thorax 2004;59:574-80. [Crossref] [PubMed]
26. Guo Y, Lip GY, Apostolakis S. Inflammation in atrial fibrillation. J Am Coll Cardiol 2012;60:2263-70. [Crossref] [PubMed]
27. Hu YF, Chen YJ, Lin YJ, et al. Inflammation and the pathogenesis of atrial fibrillation. Nat Rev Cardiol 2015;12:230-43. [Crossref] [PubMed]
28. Korantzopoulos P, Letsas KP, Tse G, et al. Inflammation and atrial fibrillation: A comprehensive review. J Arrhythm 2018;34:394-401. [Crossref] [PubMed]
29. Chen MC, Chang JP, Liu WH, et al. Increased inflammatory cell infiltration in the atrial myocardium of patients with atrial fibrillation. Am J Cardiol 2008;102:861-5. [Crossref] [PubMed]
30. Vyas V, Sandhar B, Keane JM, et al. Tissue-resident memory T cells in epicardial adipose tissue comprise transcriptionally distinct subsets that are modulated in atrial fibrillation. Nat Cardiovasc Res 2024;3:1067-82. [Crossref] [PubMed]