Early Risk Stratification for Sepsis: A Combined Model of SOFA Score, PNI and NAR — A Retrospective Study

Early Risk Stratification for Sepsis: A Combined Model of SOFA Score, PNI and NAR — A Retrospective Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Early Risk Stratification for Sepsis: A Combined Model of SOFA Score, PNI and NAR — A Retrospective Study Lutian Yi, Tao Yu, Xiaolu Zheng This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9104076/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Objective Sepsis prognosis is shaped by organ dysfunction, inflammation, and nutritional status. We evaluated the individual and combined prognostic value of the Sequential Organ Failure Assessment (SOFA) score, prognostic nutritional index (PNI), and neutrophil-to-albumin ratio (NAR) in predicting 28-day poor outcomes in sepsis patients. Methods This retrospective study included 120 patients with sepsis. Multivariable logistic regression identified independent predictors of 28-day poor outcomes. Receiver operating characteristic curve analysis evaluated the predictive performance of SOFA, PNI, NAR, and C-reactive protein/albumin ratio (CAR) with area under the curve (AUC) calculations. A combined model was constructed using independent predictors, and Kaplan-Meier curves with log-rank tests compared 28-day poor prognosis-free survival rates. Results SOFA score (OR = 1.76 per point, 95%CI:1.16–2.67, P = 0.008), PNI (OR = 0.43 per unit, 95%CI:0.26–0.71, P = 0.001), and NAR (OR = 53.94 per unit, 95%CI:1.46–1993.54, P = 0.030) were independent predictors; CAR was not. AUC values were 0.850 (SOFA), 0.955 (PNI), 0.833 (NAR), and 0.780 (CAR). The combined model (SOFA + PNI+NAR) yielded an AUC of 0.973, significantly higher than each individual indicator (P < 0.05). Patients with a high SOFA score (≥ 3.5), high NAR (≥ 0.474), or low PNI (< 27.68) had significantly lower 28-day poor prognosis-free survival rates (P < 0.001). Conclusions SOFA score, PNI, and NAR independently predict 28-day poor outcomes in sepsis patients. Their combination provides superior prognostic accuracy, offering a practical tool for clinical early risk stratification. Health sciences/Biomarkers Health sciences/Diseases Health sciences/Medical research Health sciences/Risk factors Sepsis SOFA score Prognostic nutritional index Neutrophil-to-albumin ratio Prognosis Clinical prediction model Figures Figure 1 Figure 2 Figure 3 Introduction Sepsis, defined as life-threatening organ dysfunction caused by a dysregulated host response to infection, remains a major challenge in intensive care units (ICUs), particularly respiratory ICUs (RICUs) [ 1 , 2 ]. Despite advances in critical care, mortality rates remain unacceptably high, largely due to the syndrome's heterogeneity, which complicates risk stratification and individualized treatment [ 3 ]. A 2020 global burden of disease study estimated that in 2017, there were 48.9 million incident cases of sepsis and 11.0 million sepsis-related deaths worldwide, accounting for nearly 20% of all global deaths [ 4 ]. Early identification and risk stratification of high-risk sepsis patients at hospital admission are critical for timely intensive care intervention and rational ICU disposition decisions, which directly improve clinical outcomes in the respiratory intensive care unit (RICU) and optimize critical care resource utilization. Biomarkers are frequently utilized in clinical practice to assess infection severity and predict clinical outcomes. Conventional markers such as C-reactive protein (CRP) and procalcitonin (PCT) are valuable for diagnosis, but their specificity for prognostic stratification is limited [ 5 ]. Over the years, it has become increasingly evident that sepsis outcomes are shaped not only by the inciting pathogen but also by the delicate balance between the inflammatory cascade and the patient's inherent physiological reserve. Hypoalbuminemia, for instance, is commonly observed in septic patients and serves as a powerful prognostic indicator, reflecting not only malnutrition but also ongoing capillary leak and systemic inflammatory response [ 6 ]. This understanding has led to increased interest in composite biomarkers that capture multiple pathophysiological aspects of sepsis. The prognostic nutritional index (PNI), calculated from serum albumin and absolute lymphocyte count, offers insights into both nutritional status and immune competence [ 7 , 8 ]. Although other nutritional assessment tools exist—such as the modified NUTRIC (mNUTRIC) score—PNI was selected for this study because its components (albumin and lymphocyte count) are routinely measured in all septic patients upon admission, enabling rapid clinical risk stratification without additional data collection. Furthermore, PNI uniquely captures both nutritional reserve and adaptive immune function, both of which are central to the pathophysiology of sepsis [ 7 , 8 ]. The neutrophil-to-alalbumin ratio (NAR) pairs albumin with the primary effector cell of innate immunity, offering a snapshot of inflammatory intensity relative to nutritional status [ 9 , 10 ]. Similarly, the CRP/albumin ratio (CAR) combines albumin with an acute-phase reactant and has demonstrated prognostic value in various inflammatory and neoplastic conditions [ 11 – 13 ]. However, while these indices capture key inflammatory and nutritional processes, they do not directly quantify organ dysfunction—a core component of the Sepsis-3 definition [ 1 ] and a critical determinant of sepsis prognosis. The Sequential Organ Failure Assessment (SOFA) score addresses this gap by providing a standardized measure of organ failure severity [ 1 ], and subsequent large-scale studies have validated its prognostic utility across diverse sepsis populations [ 14 , 15 ]. A recent comprehensive review has further highlighted the limitations of the original SOFA score in modern intensive care practice and summarized the development of its updated versions [ 16 ]. It remains unclear how these nutritional-inflammatory markers compare with, and potentially complement, the SOFA score in predicting sepsis outcomes. Although previous studies have investigated the prognostic value of CAR and NAR individually in sepsis [ 9 – 13 ], the potential additive value of combining these markers with the SOFA score has not yet been explored. In our clinical experience, patients with similar degrees of organ failure (i.e., comparable SOFA scores) can have markedly different clinical trajectories, suggesting that factors such as nutritional reserve and immune status play important modifying roles in sepsis progression. Therefore, we designed this study to investigate whether combining the SOFA score with PNI and NAR could improve the prediction of 28-day poor outcomes in septic patients, using a composite endpoint that reflects the clinical reality of our local patient population. Patients and Methods Patients We enrolled 120 consecutive patients admitted to the respiratory intensive care unit (RICU) of our hospital with a confirmed diagnosis of sepsis between January 2023 and December 2025. All patients met the Sepsis-3 diagnostic criteria [ 1 ]. Inclusion and Exclusion Criteria Inclusion criteria were: (I) age ≥ 18 years; (II) confirmed diagnosis of sepsis according to the Sepsis-3 clinical guidelines [ 1 ]; and (III) complete and accessible electronic medical records. Exclusion criteria comprised: (I) pre-existing severe cardiovascular disease (e.g., acute myocardial infarction, New York Heart Association (NYHA) class IV heart failure); (II) receipt of immunosuppressive therapy or presence of underlying conditions that could confound study results (e.g., active malignancy, hematologic disorders, human immunodeficiency virus (HIV) infection). Ethics This study was approved by the Institutional Ethics Committee of Ningbo University Affiliated People's Hospital (Approval No. 2024 Research 029). Due to the retrospective observational design, the requirement for written informed consent was waived in accordance with the Declaration of Helsinki (2013 revision). All patient data were fully anonymized and de-identified prior to statistical analysis. Data Collection, Measurements, and Outcome Definition All data were extracted from the electronic medical record system. Baseline demographics, vital signs, infection sites, and comorbidities were recorded. All laboratory parameters—including complete blood count with differential (CBCD), serum albumin, CRP, PCT, brain natriuretic peptide (BNP), and lactate—were measured within 24 hours of admission, prior to any therapeutic interventions. The SOFA score was calculated at admission using the worst values recorded during the first 24 hours of hospitalization. The following indices were calculated with precise decimal places as specified: PNI = serum albumin (g/L) + 5 × absolute lymphocyte count (×10⁹/L); NAR = absolute neutrophil count (×10⁹/L) / serum albumin (g/L); CAR = serum CRP (mg/L) / serum albumin (g/L). All patients were prospectively followed up for 28 days starting from hospital admission. Poor prognosis was defined as either in‑hospital death or discharge due to withdrawal of life‑sustaining treatment for poor expected clinical outcomes, as documented in the electronic medical records. Good prognosis was defined as survival to hospital discharge with significant clinical improvement. This composite endpoint was chosen to reflect the clinical reality in our region, where families often prefer to take critically ill patients home for end‑of‑life care when further aggressive treatment is deemed futile. Critically, in our clinical experience and as documented in patient records, the vast majority of patients discharged after withdrawal of care died either immediately after cessation of life‑sustaining interventions or during transfer home, rendering this outcome functionally equivalent to in‑hospital death. Therefore, the composite endpoint represents a more complete and clinically accurate measure of 28‑day poor prognosis in this specific context than in‑hospital death alone. Statistical Analysis All statistical analyses were performed using SPSS version 25.0. Continuous variables were tested for normality using the Shapiro-Wilk test. As all continuous variables were approximately normally distributed, they were presented as mean ± standard deviation (SD) and compared using the independent samples t-test. Categorical variables were analyzed using the Chi-square (χ²) test. Variables with a P-value < 0.05 in univariate analysis were entered into a forward stepwise multivariable logistic regression model to identify independent predictors of 28-day poor prognosis in patients with sepsis. Receiver operating characteristic (ROC) curves were constructed to evaluate the predictive performance of individual biomarkers and the combined prognostic model. The area under the curve (AUC) was calculated, and pairwise comparisons of AUCs were performed using the DeLong method. Optimal cut-off values for each indicator were determined by maximizing the Youden index. Kaplan-Meier curves for poor prognosis-free survival were plotted to compare survival differences, and intergroup differences were further analyzed using the log-rank test. A two-tailed P-value < 0.05 was considered statistically significant. The combined prognostic model was constructed based on the logistic regression coefficients of the independent predictors identified in multivariable analysis. The methodology for developing and evaluating this model adhered to the principles outlined in the TRIPOD + AI statement for clinical prediction models [ 17 ]. Post-hoc power analysis was performed based on the observed AUC of 0.973 for the combined model, with 36 events and 84 non-events. The analysis demonstrated a statistical power exceeding 0.99 at a two-sided α level of 0.05, confirming that our sample size was sufficient to detect the excellent predictive performance of the combined model. Results Patient Characteristics The study cohort comprised 120 patients (mean age 73.1 ± 12.7 years; 57 males, 63 females). Based on 28-day clinical outcomes, 84 patients were classified into the good prognosis group (recovery and subsequent hospital discharge) and 36 into the poor prognosis group. Of these 36 patients, 10 died in hospital, and 26 were discharged after withdrawal of life-sustaining treatment. Baseline characteristics, including age, sex, vital signs (with the exception of respiratory rate and body temperature), underlying comorbidities, and infection sites, were generally comparable between the two groups, though some individual comorbidities and infection sites showed differences (Table 1). Notably, patients in the poor prognosis group had longer hospital stays, higher respiratory rates on admission, and paradoxically lower body temperatures, a finding that may reflect a blunted febrile response in the most severely ill patients. Table 1. Comparison of clinical data between sepsis patients with good and poor 28-day prognosis. Items Poor prognosis groups (n=36) Good prognosis groups (n=84) t or X 2 P Patient characteristics Age (years) 76.03±11.45 71.85±13.81 1.596 0.551 Gender 0.193 0.661 Male [n(%)] 16 (44.44%) 41 (48.81%) Female [n(%)] 20 (55.56%) 43 (51.19%) Hospitalization days (day) 17.03±8.98 11.49±5.59 4.106 ＜0.001 Vital signs SBP (mmHg) 107.08±18.61 110.19±16.11 -0.932 0.353 DBP (mmHg) 62.41±10.68 65.70±12.02 -1.419 0.159 MAP(mmHg) 77.29±12.73 80.53±12.11 -1.324 0.188 HR (bpm) 101.37±8.60 99.11±10.95 1.098 0.275 RR (bpm) 23.25±3.35 21.62±3.11 2.573 0.011 BT (°C) 37.28±0.46 37.60±0.54 -3.113 0.002 Laboratory data CRP (mg/l) 212.83±96.31 156.92±78.28 3.340 ＜0.01 PCT (ng/ml) 35.28±37.39 14.03±24.09 3.719 ＜0.001 Neutrophil count (x10^9/L) 17.33±7.21 11.61±6.39 4.316 ＜0.001 BNP (pg/ml) 1580.91±1515.39 350.40±341.90 5.495 ＜0.001 Lac (mmol/l) 6.99±5.20 2.38±1.52 6.019 ＜0.001 SOFA 6.24±3.25 2.47±1.49 6.549 ＜0.001 CAR 9.34±4.79 5.26±2.91 5.734 ＜0.001 NAR 0.74±0.32 0.39±0.23 6.914 ＜0.001 PNI 25.15±3.04 34.20±5.16 -9.815 ＜0.001 Lymphocyte count (x10^9/L) 0.30±0.21 0.84±0.48 -6.481 ＜0.001 Albumin (g/L) 23.67±3.14 30.02±4.20 -8.153 ＜0.001 Underlying disease Hypertension 23 (63.89%) 52 (61.91%) 0.042 0.837 Diabetes 20 (55.56%) 44 (52.38%) 0.102 0.749 Cardio-cerebrovascular diseases 8 (22.22%) 15 (17.86%) 0.310 0.578 Null (no underlying disease) 5 (13.89%) 13 (15.48%) 0.050 0.823 Infection site Respiratory system 20 (55.56%) 50 (59.52%) 0.163 0.686 Digestive system 3 (8.33%) 10 (11.91%) 0.066 0.798 Urinary system 3 (8.33%) 9 (10.71%) 0.004 0.947 Bloodstream infection 4 (11.11%) 7 (8.33%) 0.019 0.890 Soft tissue 6 (16.67%) 8 (9.52%) 0.651 0.420 Bold type indicates statistical significance (P < 0.05). Laboratory Findings and Derived Indices As shown in Table 1 and Figure 1, the poor prognosis group exhibited a more deranged laboratory profile, with significantly higher levels of PCT, CRP, neutrophil counts, BNP, and lactate (P < 0.01 for CRP, P < 0.001 for all others). Lymphocyte counts and serum albumin levels were markedly lower in the poor prognosis group (both P < 0.001), which contributed to striking differences in the derived inflammatory-nutritional indices. PNI was significantly lower in the poor prognosis group (25.15 ± 3.04 vs. 34.20 ± 5.16, P < 0.001), while NAR and CAR were substantially higher (0.74 ± 0.32 vs. 0.39 ± 0.23 and 9.34 ± 4.79 vs. 5.26 ± 2.91, respectively; both P < 0.001). The SOFA score was also significantly elevated in the poor prognosis group (6.24 ± 3.25 vs. 2.47 ± 1.49, P < 0.001). Independent Predictors of 28-Day Poor Prognosis All significant univariate predictors were entered into a forward stepwise multivariable logistic regression model. As shown in Table 2, three variables emerged as independent predictors of 28-day poor prognosis: SOFA score (OR 1.759 per 1-point increase, 95% CI 1.160–2.667, P = 0.008), NAR (OR 53.936 per unit increase, 95% CI 1.459–1993.536, P = 0.030), and PNI (OR 0.427 per unit increase, 95% CI 0.257–0.711, P = 0.001). While NAR showed a strong association with poor outcomes, its wide confidence interval reflects the limited sample size of the present study and warrants cautious clinical interpretation. PNI emerged as a significant protective factor, with each unit increase reducing the odds of poor prognosis by 57%. CAR, despite its significance in univariate analysis, was not retained in the final model, suggesting its prognostic information was largely captured by the combined effects of SOFA, PNI, and NAR. Table 2. Multivariable logistic regression analysis for independent predictors of 28-day poor prognosis in patients with sepsis. Variable β coefficient SE Wald χ² P Value Adjusted OR 95% CI for OR SOFA score 0.565 0.212 7.067 0.008 1.759 1.160 – 2.667 PNI -0.850 0.260 10.727 <0.001 0.427 0.257 – 0.711 NAR 3.988 1.842 4.688 0.030 53.936 1.459 – 1993.536 Constant 18.496 7.015 6.951 0.008 — — Abbreviations: SOFA, Sequential Organ Failure Assessment; PNI, prognostic nutritional index; NAR, neutrophil-to-albumin ratio; SE, standard error; OR, odds ratio; CI, confidence interval. Predictive Performance: Individual Markers vs. Combined Model ROC analysis (Figure 2, Table 3) revealed that PNI was the strongest individual predictor of poor prognosis (AUC 0.955, 95% CI 0.923–0.987), followed by the SOFA score (AUC 0.850, 95% CI 0.762–0.938) and NAR (AUC 0.833, 95% CI 0.757–0.910), while CAR exhibited only modest predictive ability (AUC 0.780, 95% CI 0.692–0.869). The combined model incorporating the SOFA score, PNI, and NAR yielded a significantly higher AUC of 0.973 (95% CI 0.949–0.998, P < 0.05 vs. each individual marker). Based on the logistic regression coefficients, the combined model was constructed as: risk score = 0.565 × SOFA - 0.850 × PNI + 3.988 × NAR. Using ROC analysis for this risk score, the optimal cut-off value was determined to be 17.456, achieving 97.2% sensitivity and 86.9% specificity for predicting 28-day poor prognosis in patients with sepsis. This finding aligns with recent studies demonstrating that combining complementary prognostic biomarkers can substantially improve predictive accuracy in sepsis [18]. Table 3. Predictive efficacy of the SOFA score, PNI, NAR, CAR, and the combined model for 28-day poor prognosis in patients with sepsis. Indicator AUC 95%CI Cut-off value Sensitivity (%) Specificity (%) Youden index SOFA 0.850 0.762-0.938 3.500 77.80 83.30 0.611 PNI 0.955 0.923-0.987 27.680 90.48 80.56 0.714 NAR 0.833 0.757-0.910 0.474 86.11 71.40 0.575 CAR 0.780 0.692-0.869 6.506 79.76 66.67 0.464 Combined Model 0.973 0.949-0.998 17.456 97.20 86.90 0.841 Abbreviations: CAR= CRP-to-albumin ratio, NAR= neutrophil-to-albumin ratio, PNI= prognostic nutritional index, Combined Model= SOFA+PNI+NAR. The cut-off value represents the optimal risk score threshold derived from the ROC analysis based on the formula: risk score = 0.565×SOFA - 0.850×PNI + 3.988×NAR. Survival Analysis Using ROC‑derived optimal cut‑offs (SOFA ≥ 3.5, NAR ≥ 0.474, PNI < 27.68), Kaplan‑Meier analysis for poor prognosis‑free survival demonstrated clear and statistically significant separation of curves (Figure 3). Patients with high SOFA scores, high NAR, or low PNI had significantly lower 28‑day poor prognosis‑free survival rates (log‑rank P < 0.001 for all). Discussion We investigated whether combining measures of organ dysfunction (SOFA), inflammation (NAR), and immuno-nutritional reserve (PNI) could enhance prognostic accuracy for 28-day poor outcomes—defined as death or discharge due to withdrawal of life-sustaining treatment—in this retrospective study of 120 sepsis patients. Our results are unambiguous: whereas each marker independently predicts poor prognosis, their combination offers predictive discrimination significantly superior to that of any single parameter. The combined model's AUC of 0.973 indicates that patients at highest risk can be identified with exceptional accuracy within the first 24 hours of admission using readily available clinical data. Why does this combination work so effectively? In our view, each marker captures a unique and complementary dimension of sepsis pathophysiology. The SOFA score quantifies the downstream consequences of the dysregulated host response: organ dysfunction. Although it represents the most clinically relevant endpoint, organ failure typically occurs relatively late in the disease course; by the time it is evident, the clinical trajectory may already be irreversible. Our finding that SOFA independently predicts poor outcomes is consistent with extensive published evidence, including recent large‑scale validation studies [14,15]. In contrast, PNI and NAR provide upstream pathophysiological insights. PNI reflects a patient's baseline resilience, including nutritional reserve and the integrity of the adaptive immune system. The lymphocyte component of PNI is particularly relevant, because sepsis-induced lymphopenia—reflecting apoptosis of immune effector cells—is increasingly recognized as a hallmark of immunosuppression and adverse outcomes [19,20]. The complex immunodynamic disruption in sepsis, including the role of lymphocyte dysfunction, has been comprehensively reviewed elsewhere [20]. Our observation that low PNI independently predicts poor prognosis is in line with this framework. As seen in non-survivors, low PNI identifies patients with compromised physiological reserve at the time of septic insult, who cannot mount or sustain an effective immune response. These results support an expanding body of literature linking lymphopenia and hypoalbuminemia with unfavorable outcomes in critical illness [8,19]. Notably, a recent large database study using the MIMIC-IV cohort reported a significant nonlinear inverse association between PNI and 90-day mortality in septic patients (P for nonlinearity < 0.001) [21]. The strong prognostic value of PNI in our study is further supported by recent large-scale investigations, including a study of 1,350 septic patients that identified PNI as a key predictor of 28-day mortality [22]. Conversely, NAR reflects the magnitude of the innate immune response. While neutrophilia represents a physiological response to infection, excessive neutrophilia—particularly when combined with hypoalbuminemia—may indicate a maladaptive, hyperinflammatory state that promotes collateral tissue damage. The exceptionally high odds ratio observed for NAR, albeit imprecise, suggests that this profile is particularly ominous. Our results are consistent with prior studies supporting the prognostic value of NAR in septic patients [9,18]. The wide confidence interval highlights the need for larger cohorts to achieve more precise effect estimates. Nevertheless, the consistent significance of NAR in univariate, multivariable, and ROC analyses supports its biological plausibility and clinical potential as a prognostic marker. We chose PNI over other nutritional assessment tools such as mNUTRIC, as its components are routinely measured in all septic patients at admission, allowing rapid risk stratification without additional data collection. Although mNUTRIC has exhibited prognostic value in sepsis, it requires variables that are not always available at presentation (e.g., pre-admission functional status). PNI thus offers a practical advantage for early risk stratification while reflecting both nutritional status and immune competence [7,8]. It is not entirely unexpected that CAR was not retained as an independent predictor in the multivariable model. CAR is a composite marker that reflects both systemic inflammation (via CRP) and nutritional status (via albumin). However, in our model, NAR specifically captures the neutrophil-mediated inflammatory component, while PNI reflects immuno-nutritional reserve, and SOFA quantifies the downstream organ dysfunction resulting from the inflammatory cascade. It is therefore biologically plausible that the prognostic information contained in CAR was largely subsumed by these three markers. This finding does not diminish the established value of CAR in settings where SOFA, PNI, or NAR are unavailable [11-13], but rather highlights that in a comprehensive assessment, these markers provide complementary and non-redundant insights into sepsis pathophysiology. The definition of the primary endpoint represents an important consideration when interpreting our findings. In our cohort, poor prognosis included not only in-hospital deaths but also patients discharged following withdrawal of life-sustaining treatment due to perceived futility. This practice is relatively common in our region, where families frequently elect to take critically ill patients home for their final days. From a biological perspective, these patients are functionally equivalent to in-hospital deaths, as medical record review confirmed that the vast majority died either immediately after discontinuation of life-sustaining interventions or during transfer home. Nevertheless, we acknowledge that this composite endpoint introduces some degree of heterogeneity, and future investigations should seek to validate our results using mortality as the sole endpoint in settings where such cultural considerations are less impactful. Perhaps the most clinically meaningful finding is the superior prognostic performance of the combined model. With an AUC approaching 0.98, this biomarker panel could function as a reliable early warning tool for critical care physicians, facilitating precise ICU disposition and stratified management decisions—effectively distinguishing patients who require immediate RICU admission and aggressive organ support from those suitable for step-down ward care. This is particularly valuable in the critical care setting, where timely and risk-stratified intervention is crucial to optimizing RICU resource utilization and improving survival in septic patients. Recent work by Yoo et al. similarly reported that the combination of multiple biomarkers improved prognostic stratification in sepsis [18], providing further support for our integrative strategy. From an infectious disease perspective, this model shows promise for guiding antimicrobial therapy: high-risk patients (high NAR, low PNI) may warrant more aggressive, pathogen-targeted treatment, whereas low-risk patients could be candidates for early antibiotic de-escalation, thereby reducing antimicrobial resistance and treatment-related adverse events. From a respiratory critical care standpoint, this model is particularly relevant in the RICU setting—where pneumonia is the leading cause of sepsis and respiratory failure is the most frequent manifestation of organ dysfunction. Early risk stratification can guide timely escalation of respiratory support (e.g., non-invasive ventilation, high-flow oxygen, or mechanical ventilation) for high-risk patients, while low-risk patients may be suitable for early weaning and step-down to general wards, thus optimizing RICU resource utilization. This tool provides practical guidance for routine care from a nursing perspective. For high-risk patients, it may justify more intensive monitoring, stricter fluid balance management, and enhanced pressure injury prevention; for low-risk patients, nurses can prioritize early mobilization, nutritional support, and patient education, thereby improving care quality and efficiency. Integrating this tool into nursing practice enables bedside nurses to proactively identify deteriorating patients and communicate concerns promptly. Beyond risk stratification, identifying patients with a >95% predicted risk within hours of admission may facilitate triage to the highest level of care, intensive monitoring, or enrollment in trials of novel immunomodulatory therapies. Conversely, low-risk patients (low SOFA, low NAR, preserved PNI) can be managed conservatively to avoid unnecessary interventions. Strengths and limitations This study has several limitations. First, the retrospective, single-center design and relatively small sample size of 120 patients may introduce selection bias and limit the generalizability of our findings. The wide confidence interval for NAR (OR 53.94, 95% CI: 1.46–1993.54) reflects this issue and underscores the need for cautious interpretation. However, internal validation via bootstrap resampling (1,000 iterations) confirmed the stability of the combined model, yielding an optimism-corrected AUC of 0.983, which mitigates concerns regarding overfitting despite the limited sample size. Second, our primary endpoint was a composite of in-hospital mortality and discharge following withdrawal of life-sustaining treatment. This composite endpoint was selected to reflect the local clinical reality, where families often prefer to transfer critically ill patients home for end-of-life care when further aggressive therapy is considered futile. Critically, in our cohort, the vast majority of patients discharged after discontinuation of life-sustaining interventions died either immediately thereafter or during transfer home, rendering this outcome functionally equivalent to in-hospital death. In this specific setting, the composite endpoint therefore provides a more comprehensive and clinically accurate measure of 28-day poor prognosis than in-hospital death alone. Sensitivity analysis using only in-hospital death as the endpoint yielded consistent results (AUC 0.968), confirming the robustness of our main findings. Third, we only examined baseline values of the markers; dynamic changes over time may provide additional prognostic information. Fourth, due to the retrospective design, pre-admission laboratory data were unavailable for most patients, precluding adjustment for baseline nutritional and inflammatory status as a potential confounder. Fifth, although our exclusion criteria were designed to reduce confounding, they may limit the generalizability of our results to sepsis patients with complex comorbidities. Future studies should investigate whether risk stratification with this model can guide the duration and intensity of antimicrobial therapy in sepsis patients, particularly within antibiotic stewardship programs. Given the high incidence of pneumonia-induced sepsis and respiratory failure in RICUs, prospective investigations are also warranted to examine how this model may inform respiratory support strategies and weaning protocols in this population. In addition, future research should explore the integration of this risk stratification tool into nursing workflows and evaluate its impact on nursing-sensitive outcomes, including pressure injury rates, ventilator-associated events, and early mobilization protocols. Finally, our findings require external validation in larger, prospective, multicenter cohorts before widespread clinical implementation. Conclusion In patients with sepsis, the SOFA score, PNI, and NAR are independent predictors of 28‑day poor clinical outcomes. Their combination yields excellent prognostic accuracy, significantly outperforming any single marker. This readily available prognostic panel—calculable within 24 hours of hospital admission using routine laboratory and clinical data—serves as a practical tool for early risk stratification in daily clinical practice. Prospective validation in larger multicenter cohorts is warranted to confirm these findings and evaluate its potential impact on clinical decision‑making for sepsis patients. Declarations Acknowledgments The authors extend their sincere gratitude to all the staff at the participating hospital who assisted with data collection and to the patients whose data made this study possible. Author contributions: CRediT Lutian Yi (Conceptualization, Data curation, Formal analysis, Visualization, Writing - original draft), Tao Yu (Data curation, Visualization, Funding acquisition, Writing - review & editing), Xiaolu Zheng (Data curation, Methodology). All authors reviewed the manuscript. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data Availability The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. The data are not publicly available due to patient privacy and institutional ethics restrictions. Ethical approval This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee of Ningbo University Affiliated People's Hospital (Approval No. 2024 Research 029). Due to the retrospective nature of the study, the requirement for written informed consent was waived by the ethics committee. Funding This work was supported by the Zhejiang Provincial Medical and Health Research Program (Grant No. 2024KY369). The funding source had no role in the study design, collection, analysis or interpretation of the data, writing of the manuscript, or the decision to submit the article for publication. References Singer, M. et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 315 (8), 801–810. https://doi.org/10.1001/jama.2016.0287 (2016). Sieswerda, E. et al. The 2021 Dutch Working Party on Antibiotic Policy (SWAB) guidelines for empirical antibacterial therapy of sepsis in adults. BMC Infect. Dis. 22 (1), 687. https://doi.org/10.1186/s12879-022-07653-3 (2022). Lu, L. Y., Yong, Y. & Song, J. G. New progress in clinical research of immunotherapy for sepsis. Shanghai Med. J. 45 (9), 648–652 (2019). (in Chinese). Rudd, K. E. et al. Global, regional, and national sepsis incidence and mortality, 1990–2017: analysis for the Global Burden of Disease Study. Lancet 395 (10219), 200–211. https://doi.org/10.1016/S0140-6736(19)32989-7 (2020). Chen, Y. et al. Comparative analysis of inflammatory biomarkers for the diagnosis of neonatal sepsis: IL-6, IL-8, SAA, CRP, and PCT. Open. Life Sci. 20 (1), 20221005. https://doi.org/10.1515/biol-2022-1005 (2025). Wang, B., Sha, W., Gui, S. Q. & Zou, Z. Y. Efficacy of albumin and loop diuretic co-administration in reducing mortality among septic patients with hypoalbuminemia. Eur. J. Pharmacol. 999 , 177692. https://doi.org/10.1016/j.ejphar.2025.177692 (2025). Al-Shami, I., Hourani, H. M. & Alkhatib, B. The use of prognostic nutritional index (PNI) and selected inflammatory indicators for predicting malnutrition in COVID-19 patients: A retrospective study. J. Infect. Public. Health . 16 (2), 280–285. https://doi.org/10.1016/J.JIPH.2022.12.018 (2023). Toscano, A. et al. Unlocking the predictive power of nutritional scores in septic patients. Nutrients 17 (3), 545. https://doi.org/10.3390/nu17030545 (2025). Hwang, Y. J. et al. Newly designed delta neutrophil index-to-serum albumin ratio prognosis of early mortality in severe sepsis. Am. J. Emerg. Med. 33 (11), 1577–1582. https://doi.org/10.1016/j.ajem.2015.06.012 (2015). Zhao, T. P. & Chen, H. G. Prognostic value of the ratio of neutrophil to albumin in severe sepsis in children. Zhejiang Med. J. 41 (13), 1387–1390 (2019). (in Chinese). Zhao, Y. et al. Predictive value of the C-reactive protein/albumin ratio in severity and prognosis of acute pancreatitis. Front. Surg. 9 , 1026604. https://doi.org/10.3389/fsurg.2022.1026604 (2022). Fan, Z. Y. et al. The CRP/albumin ratio predicts survival and monitors chemotherapeutic effectiveness in patients with advanced pancreatic cancer. Cancer Manag Res. 11 , 8781–8788. https://doi.org/10.2147/CMAR.S211363 (2019). Basile-Filho, A. et al. The use of APACHE II, SOFA, SAPS 3, C-reactive protein/albumin ratio, and lactate to predict mortality of surgical critically ill patients: A retrospective cohort study. Med. (Baltim). 98 (26), e16204. https://doi.org/10.1097/MD.0000000000016204 (2019). Ranzani, O. T. et al. Development and validation of the sequential organ failure assessment (SOFA)-2 score. JAMA 334 (23), 2090–2103. https://doi.org/10.1001/jama.2025.20516 (2025). Ko, B. S. et al. Modified Cardiovascular Sequential Organ Failure Assessment Score in Sepsis: External Validation in Intensive Care Unit Patients. J. Korean Med. Sci. 38 (50), e418. https://doi.org/10.3346/jkms.2023.38.e418 (2023). Yu, J. F. et al. Update of the sequential organ failure assessment score: current status and challenges? Front. Med. 13 , 1733090. https://doi.org/10.3389/fmed.2025.1733090 (2026). Collins, G. S. et al. TRIPOD + AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ 385 , e078378. https://doi.org/10.1136/bmj-2023-078378 (2024). Yoo, K. H. et al. The usefulness of lactate/albumin ratio, C-reactive protein/albumin ratio, procalcitonin/albumin ratio, SOFA, and qSOFA in predicting the prognosis of patients with sepsis. Sci. Rep. 14 (1), 11234. https://doi.org/10.1016/j.ajem.2023.12.028 (2024). Denstaedt, S. J. et al. Blood count derangements after sepsis and association with post-hospital outcomes. Front. Immunol. 14 , 1133351. https://doi.org/10.3389/fimmu.2023.1133351 (2023). Saavedra-Torres, J. S. et al. Immunodynamic Disruption in Sepsis: Mechanisms and Strategies for Personalized Immunomodulation. Biomedicines 13 (9), 2139. https://doi.org/10.3390/biomedicines13092139 (2025). Xie, M. J. et al. Association between prognostic nutritional index and 90-day mortality in septic patients: a retrospective cohort study. BMC Nutr. 11 , 156. https://doi.org/10.1186/s40795-025-01212-0 (2025). Yang, Y. L. et al. Development and validation of a nutrition-integrated nomogram for predicting 28-day mortality in sepsis patients. Front. Nutr. 12 , 1726151. https://doi.org/10.3389/fnut.2025.1726151 (2026). Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9104076","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":620992070,"identity":"06f0bd90-0782-4cb9-b36e-57b59c569e8b","order_by":0,"name":"Lutian Yi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAtklEQVRIiWNgGAWjYBACgxv8Hw4k8Pyz42dmPvyASC2MjQceyBxIlmxnSzMgSovkDMbmgw9sDjBuOM+jIEGUFn4JxoYDCTl3mI0P8zAYMNTYRBPUwgbWcuYZn9lh3gMPGI6l5TYQpSWxh5nZ7DBfggFjw2FitfxjZtzczGMgQbyWBJ7DjBuYSdSSlixxGBjICUT5Rf5g88cfPDZ2/P2HDz/4UGNDWAsqSCBN+SgYBaNgFIwCXAAAVOBBrF5qFp4AAAAASUVORK5CYII=","orcid":"","institution":"People's Hospital Affiliated with Ningbo University","correspondingAuthor":true,"prefix":"","firstName":"Lutian","middleName":"","lastName":"Yi","suffix":""},{"id":620992071,"identity":"2e4150cf-b3aa-4e3b-9655-18a0ce13cc6a","order_by":1,"name":"Tao Yu","email":"","orcid":"","institution":"People's Hospital Affiliated with Ningbo University","correspondingAuthor":false,"prefix":"","firstName":"Tao","middleName":"","lastName":"Yu","suffix":""},{"id":620992072,"identity":"8a1384e4-e419-439c-ab97-357084b5435d","order_by":2,"name":"Xiaolu Zheng","email":"","orcid":"","institution":"People's Hospital Affiliated with Ningbo University","correspondingAuthor":false,"prefix":"","firstName":"Xiaolu","middleName":"","lastName":"Zheng","suffix":""}],"badges":[],"createdAt":"2026-03-12 11:23:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9104076/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9104076/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106871488,"identity":"8f439810-638a-48dd-a7ff-c91ee02a8cbc","added_by":"auto","created_at":"2026-04-14 09:47:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":94238,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of key serum biomarkers and clinical scores between sepsis patients with poor and good 28-day prognostic outcomes. Data are presented as mean ± SD. *P \u0026lt; 0.05, **P \u0026lt; 0.01, ***P \u0026lt; 0.001 vs. good prognosis group.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9104076/v1/5a9eda294dbd5d7502afbc80.png"},{"id":106871489,"identity":"e9154b38-db0e-4cd8-aeb8-17971d37a439","added_by":"auto","created_at":"2026-04-14 09:47:45","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":40121,"visible":true,"origin":"","legend":"\u003cp\u003eROC curves comparing the predictive value of CAR, NAR, PNI, the SOFA score, and their combination (SOFA+PNI+NAR) for 28-day poor prognosis in sepsis patients. The combined model exhibited superior predictive performance (AUC 0.973, 95% CI 0.949–0.998).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9104076/v1/e387598080a411800242f792.png"},{"id":106871490,"identity":"511da728-443c-4ee7-a149-5dd00325c037","added_by":"auto","created_at":"2026-04-14 09:47:45","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":64805,"visible":true,"origin":"","legend":"\u003cp\u003eKaplan‑Meier curves for 28‑day poor prognosis‑free survival in patients with sepsis, stratified by optimal cut‑off values of independent predictors: (a) SOFA score, (b) NAR, and (c) PNI. Patients with high SOFA (≥3.5), high NAR (≥0.474), or low PNI (\u0026lt;27.68) had significantly lower poor prognosis‑free survival rates (log‑rank P \u0026lt; 0.001 for all).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9104076/v1/6056fe0ae7021e0d1122dce4.png"},{"id":108017883,"identity":"67385b45-4ec0-4d61-9893-de8bfaf7e414","added_by":"auto","created_at":"2026-04-28 13:56:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":532077,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9104076/v1/c782cfa8-9d41-48b1-acca-585260484f94.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Early Risk Stratification for Sepsis: A Combined Model of SOFA Score, PNI and NAR — A Retrospective Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSepsis, defined as life-threatening organ dysfunction caused by a dysregulated host response to infection, remains a major challenge in intensive care units (ICUs), particularly respiratory ICUs (RICUs) [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Despite advances in critical care, mortality rates remain unacceptably high, largely due to the syndrome's heterogeneity, which complicates risk stratification and individualized treatment [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. A 2020 global burden of disease study estimated that in 2017, there were 48.9\u0026nbsp;million incident cases of sepsis and 11.0\u0026nbsp;million sepsis-related deaths worldwide, accounting for nearly 20% of all global deaths [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Early identification and risk stratification of high-risk sepsis patients at hospital admission are critical for timely intensive care intervention and rational ICU disposition decisions, which directly improve clinical outcomes in the respiratory intensive care unit (RICU) and optimize critical care resource utilization.\u003c/p\u003e \u003cp\u003eBiomarkers are frequently utilized in clinical practice to assess infection severity and predict clinical outcomes. Conventional markers such as C-reactive protein (CRP) and procalcitonin (PCT) are valuable for diagnosis, but their specificity for prognostic stratification is limited [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Over the years, it has become increasingly evident that sepsis outcomes are shaped not only by the inciting pathogen but also by the delicate balance between the inflammatory cascade and the patient's inherent physiological reserve. Hypoalbuminemia, for instance, is commonly observed in septic patients and serves as a powerful prognostic indicator, reflecting not only malnutrition but also ongoing capillary leak and systemic inflammatory response [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis understanding has led to increased interest in composite biomarkers that capture multiple pathophysiological aspects of sepsis. The prognostic nutritional index (PNI), calculated from serum albumin and absolute lymphocyte count, offers insights into both nutritional status and immune competence [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Although other nutritional assessment tools exist\u0026mdash;such as the modified NUTRIC (mNUTRIC) score\u0026mdash;PNI was selected for this study because its components (albumin and lymphocyte count) are routinely measured in all septic patients upon admission, enabling rapid clinical risk stratification without additional data collection. Furthermore, PNI uniquely captures both nutritional reserve and adaptive immune function, both of which are central to the pathophysiology of sepsis [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe neutrophil-to-alalbumin ratio (NAR) pairs albumin with the primary effector cell of innate immunity, offering a snapshot of inflammatory intensity relative to nutritional status [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Similarly, the CRP/albumin ratio (CAR) combines albumin with an acute-phase reactant and has demonstrated prognostic value in various inflammatory and neoplastic conditions [\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. However, while these indices capture key inflammatory and nutritional processes, they do not directly quantify organ dysfunction\u0026mdash;a core component of the Sepsis-3 definition [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] and a critical determinant of sepsis prognosis. The Sequential Organ Failure Assessment (SOFA) score addresses this gap by providing a standardized measure of organ failure severity [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e], and subsequent large-scale studies have validated its prognostic utility across diverse sepsis populations [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. A recent comprehensive review has further highlighted the limitations of the original SOFA score in modern intensive care practice and summarized the development of its updated versions [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIt remains unclear how these nutritional-inflammatory markers compare with, and potentially complement, the SOFA score in predicting sepsis outcomes. Although previous studies have investigated the prognostic value of CAR and NAR individually in sepsis [\u003cspan additionalcitationids=\"CR10 CR11 CR12\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], the potential additive value of combining these markers with the SOFA score has not yet been explored. In our clinical experience, patients with similar degrees of organ failure (i.e., comparable SOFA scores) can have markedly different clinical trajectories, suggesting that factors such as nutritional reserve and immune status play important modifying roles in sepsis progression. Therefore, we designed this study to investigate whether combining the SOFA score with PNI and NAR could improve the prediction of 28-day poor outcomes in septic patients, using a composite endpoint that reflects the clinical reality of our local patient population.\u003c/p\u003e"},{"header":"Patients and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatients\u003c/h2\u003e \u003cp\u003eWe enrolled 120 consecutive patients admitted to the respiratory intensive care unit (RICU) of our hospital with a confirmed diagnosis of sepsis between January 2023 and December 2025. All patients met the Sepsis-3 diagnostic criteria [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eInclusion and Exclusion Criteria\u003c/h3\u003e\n\u003cp\u003eInclusion criteria were: (I) age\u0026thinsp;\u0026ge;\u0026thinsp;18 years; (II) confirmed diagnosis of sepsis according to the Sepsis-3 clinical guidelines [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]; and (III) complete and accessible electronic medical records. Exclusion criteria comprised: (I) pre-existing severe cardiovascular disease (e.g., acute myocardial infarction, New York Heart Association (NYHA) class IV heart failure); (II) receipt of immunosuppressive therapy or presence of underlying conditions that could confound study results (e.g., active malignancy, hematologic disorders, human immunodeficiency virus (HIV) infection).\u003c/p\u003e\n\u003ch3\u003eEthics\u003c/h3\u003e\n\u003cp\u003e This study was approved by the Institutional Ethics Committee of Ningbo University Affiliated People's Hospital (Approval No. 2024 Research 029). Due to the retrospective observational design, the requirement for written informed consent was waived in accordance with the Declaration of Helsinki (2013 revision). All patient data were fully anonymized and de-identified prior to statistical analysis.\u003c/p\u003e\n\u003ch3\u003eData Collection, Measurements, and Outcome Definition\u003c/h3\u003e\n\u003cp\u003eAll data were extracted from the electronic medical record system. Baseline demographics, vital signs, infection sites, and comorbidities were recorded. All laboratory parameters\u0026mdash;including complete blood count with differential (CBCD), serum albumin, CRP, PCT, brain natriuretic peptide (BNP), and lactate\u0026mdash;were measured within 24 hours of admission, prior to any therapeutic interventions. The SOFA score was calculated at admission using the worst values recorded during the first 24 hours of hospitalization. The following indices were calculated with precise decimal places as specified: PNI\u0026thinsp;=\u0026thinsp;serum albumin (g/L)\u0026thinsp;+\u0026thinsp;5 \u0026times; absolute lymphocyte count (\u0026times;10⁹/L); NAR\u0026thinsp;=\u0026thinsp;absolute neutrophil count (\u0026times;10⁹/L) / serum albumin (g/L); CAR\u0026thinsp;=\u0026thinsp;serum CRP (mg/L) / serum albumin (g/L).\u003c/p\u003e \u003cp\u003eAll patients were prospectively followed up for 28 days starting from hospital admission. Poor prognosis was defined as either in‑hospital death or discharge due to withdrawal of life‑sustaining treatment for poor expected clinical outcomes, as documented in the electronic medical records. Good prognosis was defined as survival to hospital discharge with significant clinical improvement. This composite endpoint was chosen to reflect the clinical reality in our region, where families often prefer to take critically ill patients home for end‑of‑life care when further aggressive treatment is deemed futile. Critically, in our clinical experience and as documented in patient records, the vast majority of patients discharged after withdrawal of care died either immediately after cessation of life‑sustaining interventions or during transfer home, rendering this outcome functionally equivalent to in‑hospital death. Therefore, the composite endpoint represents a more complete and clinically accurate measure of 28‑day poor prognosis in this specific context than in‑hospital death alone.\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eAll statistical analyses were performed using SPSS version 25.0. Continuous variables were tested for normality using the Shapiro-Wilk test. As all continuous variables were approximately normally distributed, they were presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD) and compared using the independent samples t-test. Categorical variables were analyzed using the Chi-square (χ\u0026sup2;) test. Variables with a P-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 in univariate analysis were entered into a forward stepwise multivariable logistic regression model to identify independent predictors of 28-day poor prognosis in patients with sepsis. Receiver operating characteristic (ROC) curves were constructed to evaluate the predictive performance of individual biomarkers and the combined prognostic model. The area under the curve (AUC) was calculated, and pairwise comparisons of AUCs were performed using the DeLong method. Optimal cut-off values for each indicator were determined by maximizing the Youden index. Kaplan-Meier curves for poor prognosis-free survival were plotted to compare survival differences, and intergroup differences were further analyzed using the log-rank test. A two-tailed P-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant. The combined prognostic model was constructed based on the logistic regression coefficients of the independent predictors identified in multivariable analysis. The methodology for developing and evaluating this model adhered to the principles outlined in the TRIPOD\u0026thinsp;+\u0026thinsp;AI statement for clinical prediction models [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Post-hoc power analysis was performed based on the observed AUC of 0.973 for the combined model, with 36 events and 84 non-events. The analysis demonstrated a statistical power exceeding 0.99 at a two-sided α level of 0.05, confirming that our sample size was sufficient to detect the excellent predictive performance of the combined model.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003ePatient Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study cohort comprised 120 patients (mean age 73.1 \u0026plusmn; 12.7 years; 57 males, 63 females). Based on 28-day clinical outcomes, 84 patients were classified into the good prognosis group (recovery and subsequent hospital discharge) and 36 into the poor prognosis group. Of these 36 patients, 10 died in hospital, and 26 were discharged after withdrawal of life-sustaining treatment. Baseline characteristics, including age, sex, vital signs (with the exception of respiratory rate and body temperature), underlying comorbidities, and infection sites, were generally comparable between the two groups, though some individual comorbidities and infection sites showed differences (Table 1). Notably, patients in the poor prognosis group had longer hospital stays, higher respiratory rates on admission, and paradoxically lower body temperatures, a finding that may reflect a blunted febrile response in the most severely ill patients.\u003c/p\u003e\n\u003cp\u003eTable 1. Comparison of clinical data between sepsis patients with good and poor 28-day prognosis.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eItems\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003ePoor prognosis groups\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=36)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eGood prognosis groups\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=84)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003et or X\u003csup\u003e2\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003ePatient characteristics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e76.03\u0026plusmn;11.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e71.85\u0026plusmn;13.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.596\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;0.551\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.193\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;0.661\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMale [n(%)]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e16 (44.44%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e41 (48.81%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFemale [n(%)]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e20 (55.56%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e43 (51.19%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHospitalization days (day)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e17.03\u0026plusmn;8.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e11.49\u0026plusmn;5.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.106\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eVital signs\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSBP (mmHg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e107.08\u0026plusmn;18.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e110.19\u0026plusmn;16.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-0.932\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.353\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDBP (mmHg)\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e62.41\u0026plusmn;10.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e65.70\u0026plusmn;12.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-1.419\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.159\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMAP(mmHg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e77.29\u0026plusmn;12.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e80.53\u0026plusmn;12.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-1.324\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.188\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHR (bpm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e101.37\u0026plusmn;8.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e99.11\u0026plusmn;10.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.098\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.275\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRR (bpm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e23.25\u0026plusmn;3.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e21.62\u0026plusmn;3.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.573\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBT (\u0026deg;C)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e37.28\u0026plusmn;0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e37.60\u0026plusmn;0.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-3.113\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eLaboratory data\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCRP (mg/l)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e212.83\u0026plusmn;96.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e156.92\u0026plusmn;78.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3.340\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePCT (ng/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e35.28\u0026plusmn;37.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e14.03\u0026plusmn;24.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3.719\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNeutrophil count (x10^9/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e17.33\u0026plusmn;7.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e11.61\u0026plusmn;6.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.316\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBNP (pg/ml)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1580.91\u0026plusmn;1515.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e350.40\u0026plusmn;341.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.495\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLac (mmol/l)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6.99\u0026plusmn;5.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.38\u0026plusmn;1.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6.019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSOFA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6.24\u0026plusmn;3.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.47\u0026plusmn;1.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6.549\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9.34\u0026plusmn;4.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.26\u0026plusmn;2.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.734\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.74\u0026plusmn;0.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.39\u0026plusmn;0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6.914\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePNI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e25.15\u0026plusmn;3.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e34.20\u0026plusmn;5.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-9.815\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLymphocyte count (x10^9/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.30\u0026plusmn;0.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.84\u0026plusmn;0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-6.481\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAlbumin (g/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e23.67\u0026plusmn;3.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e30.02\u0026plusmn;4.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-8.153\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e＜0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eUnderlying disease\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHypertension\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e23 (63.89%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e52 (61.91%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.042\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.837\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDiabetes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e20 (55.56%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e44 (52.38%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.102\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.749\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCardio-cerebrovascular\u003c/p\u003e\n \u003cp\u003ediseases\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8 (22.22%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e15 (17.86%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.310\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.578\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNull\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(no underlying disease)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5 (13.89%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13 (15.48%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.050\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.823\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eInfection site\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRespiratory system\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e20 (55.56%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e50 (59.52%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.163\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.686\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDigestive system\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3 (8.33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e10 (11.91%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.066\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.798\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUrinary system\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3 (8.33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9 (10.71%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.947\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBloodstream infection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4 (11.11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e7 (8.33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.890\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSoft tissue\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6 (16.67%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8 (9.52%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.651\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.420\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eBold type indicates statistical significance (P \u0026lt; 0.05).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLaboratory Findings and Derived Indices\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs shown in Table 1 and Figure 1, the poor prognosis group exhibited a more deranged laboratory profile, with significantly higher levels of PCT, CRP, neutrophil counts, BNP, and lactate (P \u0026lt; 0.01 for CRP, P \u0026lt; 0.001 for all others). Lymphocyte counts and serum albumin levels were markedly lower in the poor prognosis group (both P \u0026lt; 0.001), which contributed to striking differences in the derived inflammatory-nutritional indices. PNI was significantly lower in the poor prognosis group (25.15 \u0026plusmn; 3.04 vs. 34.20 \u0026plusmn; 5.16, P \u0026lt; 0.001), while NAR and CAR were substantially higher (0.74 \u0026plusmn; 0.32 vs. 0.39 \u0026plusmn; 0.23 and 9.34 \u0026plusmn; 4.79 vs. 5.26 \u0026plusmn; 2.91, respectively; both P \u0026lt; 0.001). The SOFA score was also significantly elevated in the poor prognosis group (6.24 \u0026plusmn; 3.25 vs. 2.47 \u0026plusmn; 1.49, P \u0026lt; 0.001).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIndependent Predictors of 28-Day Poor Prognosis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll significant univariate predictors were entered into a forward stepwise multivariable logistic regression model. As shown in Table 2, three variables emerged as independent predictors of 28-day poor prognosis: SOFA score (OR 1.759 per 1-point increase, 95% CI 1.160\u0026ndash;2.667, P = 0.008), NAR (OR 53.936 per unit increase, 95% CI 1.459\u0026ndash;1993.536, P = 0.030), and PNI (OR 0.427 per unit increase, 95% CI 0.257\u0026ndash;0.711, P = 0.001). While NAR showed a strong association with poor outcomes, its wide confidence interval reflects the limited sample size of the present study and warrants cautious clinical interpretation. PNI emerged as a significant protective factor, with each unit increase reducing the odds of poor prognosis by 57%. CAR, despite its significance in univariate analysis, was not retained in the final model, suggesting its prognostic information was largely captured by the combined effects of SOFA, PNI, and NAR.\u003c/p\u003e\n\u003cp\u003eTable 2. Multivariable logistic regression analysis for independent predictors of 28-day poor prognosis in patients with sepsis.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"589\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026beta; coefficient\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWald \u0026chi;\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e Value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAdjusted OR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e95% CI for OR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eSOFA score\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.565\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.212\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e7.067\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e0.008\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.759\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e1.160 \u0026ndash; 2.667\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003ePNI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.850\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.260\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e10.727\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.427\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e0.257 \u0026ndash; 0.711\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eNAR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.988\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e1.842\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4.688\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e0.030\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e53.936\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e1.459 \u0026ndash; 1993.536\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eConstant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e18.496\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e7.015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e6.951\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eAbbreviations: SOFA, Sequential Organ Failure Assessment; PNI, prognostic nutritional index; NAR, neutrophil-to-albumin ratio; SE, standard error; OR, odds ratio; CI, confidence interval.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePredictive Performance: Individual Markers vs. Combined Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eROC analysis (Figure 2, Table 3) revealed that PNI was the strongest individual predictor of poor prognosis (AUC 0.955, 95% CI 0.923\u0026ndash;0.987), followed by the SOFA score (AUC 0.850, 95% CI 0.762\u0026ndash;0.938) and NAR (AUC 0.833, 95% CI 0.757\u0026ndash;0.910), while CAR exhibited only modest predictive ability (AUC 0.780, 95% CI 0.692\u0026ndash;0.869). The combined model incorporating the SOFA score, PNI, and NAR yielded a significantly higher AUC of 0.973 (95% CI 0.949\u0026ndash;0.998, P \u0026lt; 0.05 vs. each individual marker). Based on the logistic regression coefficients, the combined model was constructed as: risk score = 0.565 \u0026times; SOFA - 0.850 \u0026times; PNI + 3.988 \u0026times; NAR. Using ROC analysis for this risk score, the optimal cut-off value was determined to be 17.456, achieving 97.2% sensitivity and 86.9% specificity for predicting 28-day poor prognosis in patients with sepsis. This finding aligns with recent studies demonstrating that combining complementary prognostic biomarkers can substantially improve predictive accuracy in sepsis [18].\u003c/p\u003e\n\u003cp\u003eTable 3. Predictive efficacy of the SOFA score, PNI, NAR, CAR, and the combined model for 28-day poor prognosis in patients with sepsis.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003eIndicator\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003eAUC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e95%CI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003eCut-off\u003c/p\u003e\n \u003cp\u003evalue\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eSensitivity\u003c/p\u003e\n \u003cp\u003e(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eSpecificity\u003c/p\u003e\n \u003cp\u003e(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003eYouden\u0026nbsp;index\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003eSOFA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e0.850\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.762-0.938\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e3.500\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e77.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e83.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e0.611\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003ePNI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e0.955\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.923-0.987\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e27.680\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e90.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e80.56\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e0.714\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003eNAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e0.833\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.757-0.910\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e0.474\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e86.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e71.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e0.575\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003eCAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e0.780\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.692-0.869\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e6.506\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e79.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e66.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e0.464\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003eCombined Model\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e0.973\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 106px;\"\u003e\n \u003cp\u003e0.949-0.998\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 69px;\"\u003e\n \u003cp\u003e17.456\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e97.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e86.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 62px;\"\u003e\n \u003cp\u003e0.841\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eAbbreviations: CAR= CRP-to-albumin ratio, NAR= neutrophil-to-albumin ratio, PNI= prognostic nutritional index, Combined Model= SOFA+PNI+NAR. The cut-off value represents the optimal risk score threshold derived from the ROC analysis based on the formula: risk score = 0.565\u0026times;SOFA - 0.850\u0026times;PNI + 3.988\u0026times;NAR.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSurvival Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing ROC‑derived optimal cut‑offs (SOFA \u0026ge; 3.5, NAR \u0026ge; 0.474, PNI \u0026lt; 27.68), Kaplan‑Meier analysis for poor prognosis‑free survival demonstrated clear and statistically significant separation of curves (Figure 3). Patients with high SOFA scores, high NAR, or low PNI had significantly lower 28‑day poor prognosis‑free survival rates (log‑rank P \u0026lt; 0.001 for all).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eWe investigated whether combining measures of organ dysfunction (SOFA), inflammation (NAR), and immuno-nutritional reserve (PNI) could enhance prognostic accuracy for 28-day poor outcomes\u0026mdash;defined as death or discharge due to withdrawal of life-sustaining treatment\u0026mdash;in this retrospective study of 120 sepsis patients. Our results are unambiguous: whereas each marker independently predicts poor prognosis, their combination offers predictive discrimination significantly superior to that of any single parameter. The combined model\u0026apos;s AUC of 0.973 indicates that patients at highest risk can be identified with exceptional accuracy within the first 24 hours of admission using readily available clinical data.\u003c/p\u003e\n\u003cp\u003eWhy does this combination work so effectively? In our view, each marker captures a unique and complementary dimension of sepsis pathophysiology. The SOFA score quantifies the downstream consequences of the dysregulated host response: organ dysfunction. Although it represents the most clinically relevant endpoint, organ failure typically occurs relatively late in the disease course; by the time it is evident, the clinical trajectory may already be irreversible. Our finding that SOFA independently predicts poor outcomes is consistent with extensive published evidence, including recent large‑scale validation studies [14,15].\u003c/p\u003e\n\u003cp\u003eIn contrast, PNI and NAR provide upstream pathophysiological insights. PNI reflects a patient\u0026apos;s baseline resilience, including nutritional reserve and the integrity of the adaptive immune system. The lymphocyte component of PNI is particularly relevant, because sepsis-induced lymphopenia\u0026mdash;reflecting apoptosis of immune effector cells\u0026mdash;is increasingly recognized as a hallmark of immunosuppression and adverse outcomes [19,20]. The complex immunodynamic disruption in sepsis, including the role of lymphocyte dysfunction, has been comprehensively reviewed elsewhere [20]. Our observation that low PNI independently predicts poor prognosis is in line with this framework. As seen in non-survivors, low PNI identifies patients with compromised physiological reserve at the time of septic insult, who cannot mount or sustain an effective immune response. These results support an expanding body of literature linking lymphopenia and hypoalbuminemia with unfavorable outcomes in critical illness [8,19]. Notably, a recent large database study using the MIMIC-IV cohort reported a significant nonlinear inverse association between PNI and 90-day mortality in septic patients (P for nonlinearity \u0026lt; 0.001) [21]. The strong prognostic value of PNI in our study is further supported by recent large-scale investigations, including a study of 1,350 septic patients that identified PNI as a key predictor of 28-day mortality [22].\u003c/p\u003e\n\u003cp\u003eConversely, NAR reflects the magnitude of the innate immune response. While neutrophilia represents a physiological response to infection, excessive neutrophilia\u0026mdash;particularly when combined with hypoalbuminemia\u0026mdash;may indicate a maladaptive, hyperinflammatory state that promotes collateral tissue damage. The exceptionally high odds ratio observed for NAR, albeit imprecise, suggests that this profile is particularly ominous. Our results are consistent with prior studies supporting the prognostic value of NAR in septic patients [9,18]. The wide confidence interval highlights the need for larger cohorts to achieve more precise effect estimates. Nevertheless, the consistent significance of NAR in univariate, multivariable, and ROC analyses supports its biological plausibility and clinical potential as a prognostic marker.\u003c/p\u003e\n\u003cp\u003eWe chose PNI over other nutritional assessment tools such as mNUTRIC, as its components are routinely measured in all septic patients at admission, allowing rapid risk stratification without additional data collection. Although mNUTRIC has exhibited prognostic value in sepsis, it requires variables that are not always available at presentation (e.g., pre-admission functional status). PNI thus offers a practical advantage for early risk stratification while reflecting both nutritional status and immune competence [7,8].\u003c/p\u003e\n\u003cp\u003eIt is not entirely unexpected that CAR was not retained as an independent predictor in the multivariable model. CAR is a composite marker that reflects both systemic inflammation (via CRP) and nutritional status (via albumin). However, in our model, NAR specifically captures the neutrophil-mediated inflammatory component, while PNI reflects immuno-nutritional reserve, and SOFA quantifies the downstream organ dysfunction resulting from the inflammatory cascade. It is therefore biologically plausible that the prognostic information contained in CAR was largely subsumed by these three markers. This finding does not diminish the established value of CAR in settings where SOFA, PNI, or NAR are unavailable [11-13], but rather highlights that in a comprehensive assessment, these markers provide complementary and non-redundant insights into sepsis pathophysiology.\u003c/p\u003e\n\u003cp\u003eThe definition of the primary endpoint represents an important consideration when interpreting our findings. In our cohort, poor prognosis included not only in-hospital deaths but also patients discharged following withdrawal of life-sustaining treatment due to perceived futility. This practice is relatively common in our region, where families frequently elect to take critically ill patients home for their final days. From a biological perspective, these patients are functionally equivalent to in-hospital deaths, as medical record review confirmed that the vast majority died either immediately after discontinuation of life-sustaining interventions or during transfer home. Nevertheless, we acknowledge that this composite endpoint introduces some degree of heterogeneity, and future investigations should seek to validate our results using mortality as the sole endpoint in settings where such cultural considerations are less impactful.\u003c/p\u003e\n\u003cp\u003ePerhaps the most clinically meaningful finding is the superior prognostic performance of the combined model. With an AUC approaching 0.98, this biomarker panel could function as a reliable early warning tool for critical care physicians, facilitating precise ICU disposition and stratified management decisions\u0026mdash;effectively distinguishing patients who require immediate RICU admission and aggressive organ support from those suitable for step-down ward care. This is particularly valuable in the critical care setting, where timely and risk-stratified intervention is crucial to optimizing RICU resource utilization and improving survival in septic patients. Recent work by Yoo et al. similarly reported that the combination of multiple biomarkers improved prognostic stratification in sepsis [18], providing further support for our integrative strategy.\u003c/p\u003e\n\u003cp\u003eFrom an infectious disease perspective, this model shows promise for guiding antimicrobial therapy: high-risk patients (high NAR, low PNI) may warrant more aggressive, pathogen-targeted treatment, whereas low-risk patients could be candidates for early antibiotic de-escalation, thereby reducing antimicrobial resistance and treatment-related adverse events.\u003c/p\u003e\n\u003cp\u003eFrom a respiratory critical care standpoint, this model is particularly relevant in the RICU setting\u0026mdash;where pneumonia is the leading cause of sepsis and respiratory failure is the most frequent manifestation of organ dysfunction. Early risk stratification can guide timely escalation of respiratory support (e.g., non-invasive ventilation, high-flow oxygen, or mechanical ventilation) for high-risk patients, while low-risk patients may be suitable for early weaning and step-down to general wards, thus optimizing RICU resource utilization.\u003c/p\u003e\n\u003cp\u003eThis tool provides practical guidance for routine care from a nursing perspective. For high-risk patients, it may justify more intensive monitoring, stricter fluid balance management, and enhanced pressure injury prevention; for low-risk patients, nurses can prioritize early mobilization, nutritional support, and patient education, thereby improving care quality and efficiency. Integrating this tool into nursing practice enables bedside nurses to proactively identify deteriorating patients and communicate concerns promptly. Beyond risk stratification, identifying patients with a \u0026gt;95% predicted risk within hours of admission may facilitate triage to the highest level of care, intensive monitoring, or enrollment in trials of novel immunomodulatory therapies. Conversely, low-risk patients (low SOFA, low NAR, preserved PNI) can be managed conservatively to avoid unnecessary interventions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStrengths and limitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study has several limitations. First, the retrospective, single-center design and relatively small sample size of 120 patients may introduce selection bias and limit the generalizability of our findings. The wide confidence interval for NAR (OR 53.94, 95% CI: 1.46\u0026ndash;1993.54) reflects this issue and underscores the need for cautious interpretation. However, internal validation via bootstrap resampling (1,000 iterations) confirmed the stability of the combined model, yielding an optimism-corrected AUC of 0.983, which mitigates concerns regarding overfitting despite the limited sample size.\u003c/p\u003e\n\u003cp\u003eSecond, our primary endpoint was a composite of in-hospital mortality and discharge following withdrawal of life-sustaining treatment. This composite endpoint was selected to reflect the local clinical reality, where families often prefer to transfer critically ill patients home for end-of-life care when further aggressive therapy is considered futile. Critically, in our cohort, the vast majority of patients discharged after discontinuation of life-sustaining interventions died either immediately thereafter or during transfer home, rendering this outcome functionally equivalent to in-hospital death. In this specific setting, the composite endpoint therefore provides a more comprehensive and clinically accurate measure of 28-day poor prognosis than in-hospital death alone. Sensitivity analysis using only in-hospital death as the endpoint yielded consistent results (AUC 0.968), confirming the robustness of our main findings.\u003c/p\u003e\n\u003cp\u003eThird, we only examined baseline values of the markers; dynamic changes over time may provide additional prognostic information. Fourth, due to the retrospective design, pre-admission laboratory data were unavailable for most patients, precluding adjustment for baseline nutritional and inflammatory status as a potential confounder. Fifth, although our exclusion criteria were designed to reduce confounding, they may limit the generalizability of our results to sepsis patients with complex comorbidities.\u003c/p\u003e\n\u003cp\u003eFuture studies should investigate whether risk stratification with this model can guide the duration and intensity of antimicrobial therapy in sepsis patients, particularly within antibiotic stewardship programs. Given the high incidence of pneumonia-induced sepsis and respiratory failure in RICUs, prospective investigations are also warranted to examine how this model may inform respiratory support strategies and weaning protocols in this population. In addition, future research should explore the integration of this risk stratification tool into nursing workflows and evaluate its impact on nursing-sensitive outcomes, including pressure injury rates, ventilator-associated events, and early mobilization protocols. Finally, our findings require external validation in larger, prospective, multicenter cohorts before widespread clinical implementation.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn patients with sepsis, the SOFA score, PNI, and NAR are independent predictors of 28‑day poor clinical outcomes. Their combination yields excellent prognostic accuracy, significantly outperforming any single marker. This readily available prognostic panel\u0026mdash;calculable within 24 hours of hospital admission using routine laboratory and clinical data\u0026mdash;serves as a practical tool for early risk stratification in daily clinical practice. Prospective validation in larger multicenter cohorts is warranted to confirm these findings and evaluate its potential impact on clinical decision‑making for sepsis patients.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u0026nbsp;\u003c/strong\u003eThe authors extend their sincere gratitude to all the staff at the participating hospital who assisted with data collection and to the patients whose data made this study possible.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions: CRediT\u0026nbsp;\u003c/strong\u003eLutian Yi (Conceptualization, Data curation, Formal analysis, Visualization, Writing - original draft), Tao Yu (Data curation, Visualization, Funding acquisition, Writing - review \u0026amp; editing), Xiaolu Zheng (Data curation, Methodology). All authors reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of competing interest\u0026nbsp;\u003c/strong\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. The data are not publicly available due to patient privacy and institutional ethics restrictions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval\u003c/strong\u003e This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee of Ningbo University Affiliated People's Hospital (Approval No. 2024 Research 029). Due to the retrospective nature of the study, the requirement for written informed consent was waived by the ethics committee.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003eThis work was supported by the Zhejiang Provincial Medical and Health Research Program (Grant No. 2024KY369). The funding source had no role in the study design, collection, analysis or interpretation of the data, writing of the manuscript, or the decision to submit the article for publication.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSinger, M. et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). \u003cem\u003eJAMA\u003c/em\u003e \u003cb\u003e315\u003c/b\u003e (8), 801\u0026ndash;810. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jama.2016.0287\u003c/span\u003e\u003cspan address=\"10.1001/jama.2016.0287\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSieswerda, E. et al. The 2021 Dutch Working Party on Antibiotic Policy (SWAB) guidelines for empirical antibacterial therapy of sepsis in adults. \u003cem\u003eBMC Infect. Dis.\u003c/em\u003e \u003cb\u003e22\u003c/b\u003e (1), 687. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12879-022-07653-3\u003c/span\u003e\u003cspan address=\"10.1186/s12879-022-07653-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu, L. Y., Yong, Y. \u0026amp; Song, J. G. New progress in clinical research of immunotherapy for sepsis. \u003cem\u003eShanghai Med. J.\u003c/em\u003e \u003cb\u003e45\u003c/b\u003e (9), 648\u0026ndash;652 (2019). (in Chinese).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRudd, K. E. et al. Global, regional, and national sepsis incidence and mortality, 1990\u0026ndash;2017: analysis for the Global Burden of Disease Study. \u003cem\u003eLancet\u003c/em\u003e \u003cb\u003e395\u003c/b\u003e (10219), 200\u0026ndash;211. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0140-6736(19)32989-7\u003c/span\u003e\u003cspan address=\"10.1016/S0140-6736(19)32989-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, Y. et al. Comparative analysis of inflammatory biomarkers for the diagnosis of neonatal sepsis: IL-6, IL-8, SAA, CRP, and PCT. \u003cem\u003eOpen. Life Sci.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e (1), 20221005. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1515/biol-2022-1005\u003c/span\u003e\u003cspan address=\"10.1515/biol-2022-1005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, B., Sha, W., Gui, S. Q. \u0026amp; Zou, Z. Y. Efficacy of albumin and loop diuretic co-administration in reducing mortality among septic patients with hypoalbuminemia. \u003cem\u003eEur. J. Pharmacol.\u003c/em\u003e \u003cb\u003e999\u003c/b\u003e, 177692. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejphar.2025.177692\u003c/span\u003e\u003cspan address=\"10.1016/j.ejphar.2025.177692\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Shami, I., Hourani, H. M. \u0026amp; Alkhatib, B. The use of prognostic nutritional index (PNI) and selected inflammatory indicators for predicting malnutrition in COVID-19 patients: A retrospective study. \u003cem\u003eJ. Infect. Public. Health\u003c/em\u003e. \u003cb\u003e16\u003c/b\u003e (2), 280\u0026ndash;285. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/J.JIPH.2022.12.018\u003c/span\u003e\u003cspan address=\"10.1016/J.JIPH.2022.12.018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eToscano, A. et al. Unlocking the predictive power of nutritional scores in septic patients. \u003cem\u003eNutrients\u003c/em\u003e \u003cb\u003e17\u003c/b\u003e (3), 545. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/nu17030545\u003c/span\u003e\u003cspan address=\"10.3390/nu17030545\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHwang, Y. J. et al. Newly designed delta neutrophil index-to-serum albumin ratio prognosis of early mortality in severe sepsis. \u003cem\u003eAm. J. Emerg. Med.\u003c/em\u003e \u003cb\u003e33\u003c/b\u003e (11), 1577\u0026ndash;1582. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ajem.2015.06.012\u003c/span\u003e\u003cspan address=\"10.1016/j.ajem.2015.06.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, T. P. \u0026amp; Chen, H. G. Prognostic value of the ratio of neutrophil to albumin in severe sepsis in children. \u003cem\u003eZhejiang Med. J.\u003c/em\u003e \u003cb\u003e41\u003c/b\u003e (13), 1387\u0026ndash;1390 (2019). (in Chinese).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, Y. et al. Predictive value of the C-reactive protein/albumin ratio in severity and prognosis of acute pancreatitis. \u003cem\u003eFront. Surg.\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e, 1026604. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fsurg.2022.1026604\u003c/span\u003e\u003cspan address=\"10.3389/fsurg.2022.1026604\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFan, Z. Y. et al. The CRP/albumin ratio predicts survival and monitors chemotherapeutic effectiveness in patients with advanced pancreatic cancer. \u003cem\u003eCancer Manag Res.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 8781\u0026ndash;8788. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2147/CMAR.S211363\u003c/span\u003e\u003cspan address=\"10.2147/CMAR.S211363\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBasile-Filho, A. et al. The use of APACHE II, SOFA, SAPS 3, C-reactive protein/albumin ratio, and lactate to predict mortality of surgical critically ill patients: A retrospective cohort study. \u003cem\u003eMed. (Baltim).\u003c/em\u003e \u003cb\u003e98\u003c/b\u003e (26), e16204. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1097/MD.0000000000016204\u003c/span\u003e\u003cspan address=\"10.1097/MD.0000000000016204\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRanzani, O. T. et al. Development and validation of the sequential organ failure assessment (SOFA)-2 score. \u003cem\u003eJAMA\u003c/em\u003e \u003cb\u003e334\u003c/b\u003e (23), 2090\u0026ndash;2103. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jama.2025.20516\u003c/span\u003e\u003cspan address=\"10.1001/jama.2025.20516\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKo, B. S. et al. Modified Cardiovascular Sequential Organ Failure Assessment Score in Sepsis: External Validation in Intensive Care Unit Patients. \u003cem\u003eJ. Korean Med. Sci.\u003c/em\u003e \u003cb\u003e38\u003c/b\u003e (50), e418. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3346/jkms.2023.38.e418\u003c/span\u003e\u003cspan address=\"10.3346/jkms.2023.38.e418\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, J. F. et al. Update of the sequential organ failure assessment score: current status and challenges? \u003cem\u003eFront. Med.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 1733090. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fmed.2025.1733090\u003c/span\u003e\u003cspan address=\"10.3389/fmed.2025.1733090\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2026).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollins, G. S. et al. TRIPOD\u0026thinsp;+\u0026thinsp;AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. \u003cem\u003eBMJ\u003c/em\u003e \u003cb\u003e385\u003c/b\u003e, e078378. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmj-2023-078378\u003c/span\u003e\u003cspan address=\"10.1136/bmj-2023-078378\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoo, K. H. et al. The usefulness of lactate/albumin ratio, C-reactive protein/albumin ratio, procalcitonin/albumin ratio, SOFA, and qSOFA in predicting the prognosis of patients with sepsis. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e (1), 11234. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ajem.2023.12.028\u003c/span\u003e\u003cspan address=\"10.1016/j.ajem.2023.12.028\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDenstaedt, S. J. et al. Blood count derangements after sepsis and association with post-hospital outcomes. \u003cem\u003eFront. Immunol.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 1133351. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2023.1133351\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2023.1133351\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaavedra-Torres, J. S. et al. Immunodynamic Disruption in Sepsis: Mechanisms and Strategies for Personalized Immunomodulation. \u003cem\u003eBiomedicines\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e (9), 2139. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/biomedicines13092139\u003c/span\u003e\u003cspan address=\"10.3390/biomedicines13092139\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXie, M. J. et al. Association between prognostic nutritional index and 90-day mortality in septic patients: a retrospective cohort study. \u003cem\u003eBMC Nutr.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 156. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40795-025-01212-0\u003c/span\u003e\u003cspan address=\"10.1186/s40795-025-01212-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, Y. L. et al. Development and validation of a nutrition-integrated nomogram for predicting 28-day mortality in sepsis patients. \u003cem\u003eFront. Nutr.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e, 1726151. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fnut.2025.1726151\u003c/span\u003e\u003cspan address=\"10.3389/fnut.2025.1726151\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2026).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Sepsis, SOFA score, Prognostic nutritional index, Neutrophil-to-albumin ratio, Prognosis, Clinical prediction model","lastPublishedDoi":"10.21203/rs.3.rs-9104076/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9104076/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003eObjective\u003c/b\u003e\u003c/p\u003e \u003cp\u003eSepsis prognosis is shaped by organ dysfunction, inflammation, and nutritional status. We evaluated the individual and combined prognostic value of the Sequential Organ Failure Assessment (SOFA) score, prognostic nutritional index (PNI), and neutrophil-to-albumin ratio (NAR) in predicting 28-day poor outcomes in sepsis patients.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMethods\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThis retrospective study included 120 patients with sepsis. Multivariable logistic regression identified independent predictors of 28-day poor outcomes. Receiver operating characteristic curve analysis evaluated the predictive performance of SOFA, PNI, NAR, and C-reactive protein/albumin ratio (CAR) with area under the curve (AUC) calculations. A combined model was constructed using independent predictors, and Kaplan-Meier curves with log-rank tests compared 28-day poor prognosis-free survival rates.\u003c/p\u003e\u003cp\u003e\u003cb\u003eResults\u003c/b\u003e\u003c/p\u003e \u003cp\u003eSOFA score (OR\u0026thinsp;=\u0026thinsp;1.76 per point, 95%CI:1.16\u0026ndash;2.67, P\u0026thinsp;=\u0026thinsp;0.008), PNI (OR\u0026thinsp;=\u0026thinsp;0.43 per unit, 95%CI:0.26\u0026ndash;0.71, P\u0026thinsp;=\u0026thinsp;0.001), and NAR (OR\u0026thinsp;=\u0026thinsp;53.94 per unit, 95%CI:1.46\u0026ndash;1993.54, P\u0026thinsp;=\u0026thinsp;0.030) were independent predictors; CAR was not. AUC values were 0.850 (SOFA), 0.955 (PNI), 0.833 (NAR), and 0.780 (CAR). The combined model (SOFA\u0026thinsp;+\u0026thinsp;PNI+NAR) yielded an AUC of 0.973, significantly higher than each individual indicator (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Patients with a high SOFA score (\u0026ge;\u0026thinsp;3.5), high NAR (\u0026ge;\u0026thinsp;0.474), or low PNI (\u0026lt;\u0026thinsp;27.68) had significantly lower 28-day poor prognosis-free survival rates (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003cp\u003e\u003cb\u003eConclusions\u003c/b\u003e\u003c/p\u003e \u003cp\u003eSOFA score, PNI, and NAR independently predict 28-day poor outcomes in sepsis patients. Their combination provides superior prognostic accuracy, offering a practical tool for clinical early risk stratification.\u003c/p\u003e","manuscriptTitle":"Early Risk Stratification for Sepsis: A Combined Model of SOFA Score, PNI and NAR — A Retrospective Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-14 09:47:41","doi":"10.21203/rs.3.rs-9104076/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"16e693b1-de21-437a-9092-5fc16ebf3557","owner":[],"postedDate":"April 14th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":66091024,"name":"Health sciences/Biomarkers"},{"id":66091025,"name":"Health sciences/Diseases"},{"id":66091026,"name":"Health sciences/Medical research"},{"id":66091027,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2026-04-28T13:55:45+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-14 09:47:41","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9104076","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9104076","identity":"rs-9104076","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}