Sepsis remains a significant cause of adverse outcomes among critically ill individuals, predominantly driven by a dysregulated immune response to infection that results in life-threatening organ dysfunction [1, 2]. Annually, sepsis is estimated to affect approximately 49 million people worldwide, with about 11 million deaths, highlighting its substantial global health burden [3]. Diagnosis of sepsis relies on the Sepsis-3 criteria, which define the condition as life-threatening organ dysfunction caused by a dysregulated host response to infection [4]. Current treatment strategies emphasize the early administration of antibiotics, fluid resuscitation, hemodynamic support, and organ support therapies [5, 6]. Despite these clinical interventions, heterogeneity in disease progression and treatment response complicates clinical management, and mortality rates in septic shock remain unacceptably high.

Numerous clinical scoring systems, including the sequential organ failure assessment (SOFA) and the acute physiology and chronic health evaluation II (APACHE II), are widely used for risk stratification in septic patients [7, 8]. Nonetheless, these models often fail to reflect real-time alterations in immune status and nutritional reserves, both of which play critical roles in the onset and progression of the condition [9, 10, 11]. Recent studies have emphasized the prognostic significance of composite biomarkers that integrate inflammatory and nutritional parameters, demonstrating their potential in enhancing outcomes prediction across various critical illnesses [12, 13, 14].

The systemic immune-inflammation index (SII), derived from platelet (PLT), neutrophil (NEUT), and lymphocyte (LYMPH) counts, serves as a composite measure of both pro-inflammatory activity and immune surveillance [15, 16]. Likewise, the Naples Prognostic Score (NPS) incorporates parameters such as serum albumin, total cholesterol (TC), neutrophil-to-lymphocyte ratio (NLR), and lymphocyte-to-monocyte ratio (LMR), thereby providing a multidimensional assessment of inflammation and nutritional status [17, 18]. While these indices have shown prognostic relevance in oncology and cardiovascular diseases, their role in sepsis remains to be elucidated [19, 20, 21, 22].

Therefore, this study focused on assessing the prognostic implications of SII and NPS in individuals with sepsis, and constructing a nomogram model incorporating these indicators. By integrating routinely available hematological and biochemical parameters, we aimed to establish a practical, cost-effective, and easily applicable prognostic tool for individualized risk stratification in clinical settings. Such an approach may enhance the early recognition of high-risk patients, thereby supporting prompt and targeted therapeutic decision-making. Additionally, the study provides insights into the complex interaction between systemic inflammatory responses, immune dysregulation, and nutritional decline during sepsis progression, providing a novel perspective for prognostic evaluation in critical care.

This study comprehensively evaluated the prognostic application of SII and NPS in predicting 28-day mortality among sepsis patients. Our analysis demonstrated that both elevated SII and NPS $\ge$2 were significantly associated with higher 28-day mortality risk. Notably, combining these two indices with a baseline clinical model significantly improved predictive performance, increasing the AUC from 0.698 to 0.846. These results underscore the clinical potential of SII and NPS as complementary predictors for enhancing mortality risk stratification in septic patients.

The SII, a composite marker derived from NEUT, LYMPH, and PLT counts, reflects both inflammatory activity and immune competence. Initially proposed as a prognostic indicator in oncology, it has since demonstrated significance across a variety of conditions, including cardiac diseases, autoimmune pathologies, and acute critical illnesses. In sepsis, evidence has revealed a correlation between higher SII levels and adverse outcomes [27]. Our study reinforces this association, demonstrating that elevated SII is independently linked to increased risk of early mortality, even after adjusting for established severity scores such as SOFA and APACHE II. Although direct comparisons were beyond the scope of this study, the findings imply that SII offers unique prognostic information complementary to traditional scoring systems. Increased SII may reflect a state of higher systemic inflammation combined with impaired adaptive immunity, potentially contributing to organ dysfunction, reduced pathogen clearance, and worse outcomes. These mechanisms may underlie the observed association between higher SII values and increased mortality, although further studies are needed to confirm this hypothesis.

The NPS, which incorporates inflammatory markers (NLR and LMR) along with nutritional indicators (serum albumin and TC), was originally formulated for oncology patients to evaluate both systemic inflammation and nutritional deficits. Unlike with other commonly used nutritional risk scores in critical care, such as NRS-2002 or modified Nutrition Risk in the Critically Ill (mNUTRIC) score, NPS integrates inflammatory and nutritional parameters into a single, easily obtainable score, potentially enhancing early risk stratification in the ICU setting [28, 29]. In our cohort, an NPS score of 2 or higher was independently associated with higher 28-day mortality. This association may be due to the combined effects of malnutrition and immune dysregulation on sepsis outcomes: hypoalbuminemia and hypocholesterolemia often reflect catabolic stress, impaired hepatic synthesis, and compromised cellular immunity, while elevated NLR and reduced LMR indicate systemic inflammation and LYMPH depletion, both of which can exacerbate organ dysfunction and increase mortality risk [30, 31]. Hence, integrating inflammatory ratios, such as NLR and LMR, and nutritional markers into the NPS provides a more comprehensive evaluation of patient vulnerability and improves prognostic accuracy in critically ill septic patients.

Importantly, combining both SII and NPS into a predictive model yielded a synergistic effect, significantly enhancing the model’s discriminative capability. This result suggests that incorporating an SII with an NPS more effectively captures the multifactorial pathophysiology of sepsis, where immune dysregulation and metabolic impairment coexist. The construction of a nomogram offers a convenient, bedside-friendly tool for individualized risk prediction. Its accuracy was supported by calibration analysis, which revealed strong concordance between predicted and observed mortality rates, and by DCA, which indicated consistent clinical benefit across various thresholds.

Compared with previously reported predictive models for sepsis patients, our combined SII + NPS model demonstrated comparable or superior performance. For instance, a nomogram integrating N-terminal pro–B-type natriuretic peptide (NT-proBNP), lactate, partial pressure of arterial oxygen/fraction of inspired oxygen (PaO₂/FiO₂), MAP, and hematocrit achieved an AUC of 0.820 (95% CI: 0.780–0.860) in the training cohort, which was lower than the AUC of 0.846 observed in our study [32]. Another model, which combined age, SOFA score, CRP, mechanical ventilation, and vasopressor use, reported AUCs of 0.849 in the training set and 0.837 in the external validation set, values comparable to our results [33]. These findings suggest that incorporating SII and NPS enhances prognostic value and may improve early risk stratification in sepsis.

The principal novelty of this study lie in integrating two readily available, non-invasive, and cost-effective biomarkers—SII and NPS—into a unified prognostic model for sepsis. Unlike previous studies that often focus on single parameters or rely heavily on advanced and expensive biomarkers, our approach depends on routine laboratory tests, enhancing their feasibility and generalizability in real-world clinical settings. However, we acknowledge certain limitations in our study. First, the retrospective design and single-center dataset may have introduced selection bias and limited the broader applicability. Second, although internal validation confirmed the model’s reliability, external and prospective validation are warranted to confirm its predictive robustness. Finally, potentially informative biomarkers, including PCT, interleukin-6, and lactate, were excluded due to data unavailability, which may have reduced the overall comprehensiveness of the model.

In summary, both SII and NPS are independent predictors of early mortality risk in sepsis. Their combined use significantly improves the predictive accuracy of traditional severity models. The nomogram developed from these parameters shows good calibration and holds promise as a clinically valuable tool for personalized mortality risk estimation. However, prospective multicenter investigations are required to validate the model’s robustness and extend its applicability to broader patient populations.

$\bullet$ This study evaluated the prognostic significance of the SII and NPS in sepsis.

$\bullet$ Univariate screening followed by multivariate logistic regression identified SII and NPS ($\ge$2) as independent predictors of 28-day mortality.

$\bullet$ ROC analysis revealed that combining SII and NPS has higher discriminative capability than either index alone.

$\bullet$ A nomogram developed from SII and NPS shows good calibration and strong clinical applicability, supporting its use in individualized risk prediction for sepsis.

All data included in this study are available from the corresponding author upon reasonable request.

XYC and YL designed the study. XYC, KL, and SDX performed the experiments and data analyses. XYC drafted the manuscript. YL supervised the project. All authors contributed to the important editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

This study was conducted in accordance with the principles of the Declaration of Helsinki and its later amendments. The research protocol was reviewed and approved by the Ethics Committee of The People’s Hospital of Cangnan (Approval No. 2024021). Given the retrospective and non-interventional nature of the study, which involved the use of anonymized data extracted from electronic medical records, the requirement for written informed consent was formally waived by the Ethics Committee.

Not applicable.

This work was supported by the Basic Research Project of Wenzhou City (No. Y20240181).

The authors declare no conflict of interest.