The neutrophil-to-lymphocyte ratio as a predictive marker for acute kidney injury in hospitalized sepsis patients: a prospective cohort study Authors: Pudit Chiamwittayanukul, Bancha Satirapoj, Theerasak Tangwonglert, Ouppatham Supasyndh, Paramat Thimachai, Wisit Kaewput, Narongrit Siriwattanasit Published: 2026 ABSTRACT AbstractBackgroundSepsis-associated acute kidney injury (SA-AKI) is a frequent and severe complication of sepsis that substantially increases morbidity and mortality. The neutrophil-to-lymphocyte ratio (NLR), obtained from routine complete blood counts, has emerged as a potential early inflammatory marker; however, its predictive value for SA-AKI remains insufficiently characterized, particularly in real-world hospital settings in Southeast Asia. This study aimed to evaluate the predictive performance of admission NLR for SA-AKI in hospitalized patients with sepsis.MethodsIn this prospective, single-center cohort study, 112 adult patients meeting Sepsis-3 criteria were consecutively enrolled at a tertiary hospital in Thailand between July 2024 and February 2025. Admission NLR was derived from the first routine complete blood count obtained within 24 h of admission. The primary outcome was the development of AKI within 7 days, defined by Kidney Disease: Improving Global Outcomes (KDIGO) creatinine criteria. Receiver operating characteristic (ROC) analysis evaluated the predictive performance of NLR alone and in combination with routinely available clinical variables. Multivariable logistic regression was performed to adjust for clinically relevant covariates.ResultsThe mean age of the cohort was 77 ± 11 years, and 32 patients (28.6%) developed AKI. Admission NLR was significantly higher in patients who developed AKI than in those who did not (median 20.2 [IQR 11.9–47.3] vs. 9.1 [IQR 5.1–14.3]; P < 0.001). An NLR cut-off of ≥ 12.8 predicted AKI with an area under the curve (AUC) of 0.78 (95% CI 0.69–0.88), sensitivity of 75.0%, and specificity of 72.5%. In multivariable analysis, NLR (adjusted odds ratio [aOR] 5.01, 95% CI 1.75–14.35; P = 0.003) and serum albumin (aOR 0.19, 95% CI 0.06–0.58; P = 0.004) were independently associated with AKI. Combining NLR with serum albumin improved model discrimination, yielding an AUC of 0.83 (95% CI 0.73–0.93). Patients with AKI had a significantly longer hospital stay than those without AKI (median 11 [IQR 6–14] vs. 7 [IQR 4–10.5] days; P = 0.039).ConclusionsAdmission NLR is independently associated with the development of SA-AKI and demonstrates moderate predictive performance for early risk stratification. Integration of NLR with serum albumin further improves model discrimination. These findings support the potential role of routinely available inflammatory and nutritional indices for SA-AKI risk stratification; however, external validation of the NLR cut-off in independent multicenter cohorts is warranted before routine clinical adoption.Keywords: Sepsis-associated acute kidney injury, Neutrophil-to-lymphocyte ratio, Sepsis, Predictive performance, Serum albumin, Risk stratification Background Sepsis-associated acute kidney injury (SA-AKI) is a common and serious complication of sepsis, affecting up to 40–50% of critically ill patients with sepsis and substantially increasing morbidity, mortality, and healthcare costs [1–3]. Early identification of patients at risk for SA-AKI is crucial to enable timely implementation of nephroprotective strategies, yet current diagnostic criteria based on serum creatinine remain limited by their delayed response to injury, often detecting kidney damage only after significant functional decline has already occurred [4]. Timely recognition is therefore of particular clinical importance, as it may enable earlier implementation of nephroprotective measures, including targeted hemodynamic optimization, avoidance of nephrotoxic agents, and intensified renal surveillance. In this context, there remains a critical unmet need for accessible, real-time predictive markers that can be integrated into routine clinical workflows without additional testing or specialized infrastructure. Because the neutrophil-to-lymphocyte ratio is derived from the routinely obtained complete blood count, it may provide a pragmatic approach to early bedside risk stratification, particularly in resource-limited or resource-variable healthcare settings. These persistent gaps between the onset of renal injury and its clinical recognition have driven the search for earlier, more accessible predictive markers. Although several novel biomarkers, including neutrophil gelatinase-associated lipocalin (NGAL) and tissue inhibitor of metalloproteinase-2/insulin-like growth factor binding protein 7 (TIMP-2/IGFBP7), have demonstrated promising predictive performance for early AKI detection, their clinical adoption remains constrained by availability, cost, and infrastructure requirements, particularly in resource-diverse settings where the burden of sepsis is greatest [5, 6]. These limitations underscore the need for simple, universally applicable markers that can be derived from routine laboratory testing without additional cost or specialized equipment. The neutrophil-to-lymphocyte ratio (NLR) reflects the balance between innate and adaptive immune responses, integrating information from neutrophil expansion and relative lymphocyte depletion that are characteristic of the host response to sepsis [7, 8]. Elevated NLR has been associated with systemic inflammation, immune dysregulation, and adverse outcomes across a range of clinical contexts, including surgical [9], traumatic [10], and chronic kidney disease populations [11]. In the context of sepsis, prior studies have reported associations between elevated NLR and AKI development, but the reported predictive performance has been highly variable, with AUCs ranging from 0.63 to 0.99, likely reflecting heterogeneity in study designs, populations, and clinical settings [12–14]. Because NLR is derived from routinely available complete blood counts, it can be calculated without additional laboratory procedures, enhancing its feasibility for bedside risk stratification. Importantly, NLR is a non-specific systemic inflammatory marker that reflects the overall balance between innate immune activation and adaptive immune competence, rather than a kidney-specific biomarker. This non-specificity may influence its discriminatory performance for organ-specific outcomes such as AKI compared with kidney-specific biomarkers. Nevertheless, the broad applicability of NLR across diverse clinical contexts, including diabetic complications [15], metabolic syndrome, surgical populations, and critically ill cohorts, underscores its potential utility as a first-line screening tool. A recent meta-analysis confirmed the prognostic value of NLR for predicting mortality in adult sepsis patients [16], further supporting its clinical relevance in this population. Despite a substantial burden of sepsis and diverse clinical challenges, Southeast Asia remains relatively underrepresented in biomarker validation studies [17]. Factors such as tropical co-infections, delayed presentation to care, variability in laboratory infrastructure, and heterogeneous inflammatory phenotypes may influence biomarker performance in this setting [18]. Most prior NLR studies have been conducted in European and North American intensive care cohorts characterized by early presentation and highly protocolized resuscitation, which may limit their generalizability to other healthcare environments [12]. Accordingly, regional validation is essential to determine the performance and applicability of NLR for SA-AKI risk stratification in diverse real-world settings. This prospective study was designed to evaluate the predictive performance of admission NLR for the early prediction of SA-AKI in hospitalized sepsis patients at a tertiary hospital in Southeast Asia. In addition to assessing the discriminatory ability of NLR alone, we examined whether combining NLR with routinely available clinical parameters, including serum albumin and illness severity indices, could enhance predictive performance. We hypothesized that NLR, either alone or in combination with clinical variables, would demonstrate clinically useful discriminatory performance for SA-AKI risk stratification in this population. Methods Study design and setting This study was conducted as a prospective, single-center observational study at Phramongkutklao Hospital, a tertiary-care teaching hospital and referral center in Bangkok, Thailand. The hospital provides comprehensive emergency, medical ward, and intensive care services to both civilian and military populations. The study was carried out between July 2024 and February 2025. All eligible patients were consecutively screened upon presentation to the emergency department and followed throughout their hospital course for the development of acute kidney injury and other clinical outcomes. The study protocol was reviewed and approved by the Institutional Review Board of the Royal Thai Army Medical Department (IRB No. 0431/2567). The study was conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants or their legally authorized representatives prior to enrollment. The manuscript was prepared and reported in accordance with the Standards for Reporting Diagnostic Accuracy Studies (STARD) and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. Study population and eligibility criteria Adult patients who were admitted to the emergency department of Phramongkutklao Hospital with a diagnosis of sepsis were consecutively screened for eligibility between July 2024 and February 2025. Sepsis was defined according to the Sepsis-3 criteria as suspected or documented infection accompanied by an acute increase in the Sequential Organ Failure Assessment (SOFA) score of ≥ 2 points. Inclusion criteria Patients were eligible for enrollment if they met all of the following criteria: Patients were eligible for enrollment if they met all of the following criteria: Age ≥ 20 years. Age ≥ 20 years. Diagnosis of sepsis according to the Sepsis-3 definition. Diagnosis of sepsis according to the Sepsis-3 definition. Admission to the hospital through the emergency department. Admission to the hospital through the emergency department. Availability of complete baseline laboratory data, including complete blood count and serum creatinine, within 24 h of admission. Availability of complete baseline laboratory data, including complete blood count and serum creatinine, within 24 h of admission. Exclusion criteria Patients were excluded if any of the following conditions were present: Pre-existing end-stage kidney disease (ESKD) or requirement for renal replacement therapy at baseline. Patients were excluded if any of the following conditions were present: Pre-existing end-stage kidney disease (ESKD) or requirement for renal replacement therapy at baseline. History of kidney transplantation. History of kidney transplantation. Active malignancy or immunocompromised status, including human immunodeficiency virus (HIV) infection or treatment with systemic immunosuppressive agents. Active malignancy or immunocompromised status, including human immunodeficiency virus (HIV) infection or treatment with systemic immunosuppressive agents. Pregnancy. Pregnancy. Exposure to nephrotoxic agents or intravenous contrast media within 7 days prior to admission. Exposure to nephrotoxic agents or intravenous contrast media within 7 days prior to admission. Requirement for renal replacement therapy at the time of hospital admission. Requirement for renal replacement therapy at the time of hospital admission. Screening and enrollment process A total of 132 consecutive patients with suspected sepsis were initially screened during the study period. After application of the eligibility criteria, 112 patients fulfilled the inclusion criteria and were enrolled in the final analysis. The patient screening and enrollment process is summarized in Fig. 1. Study flow diagram Written informed consent was obtained from all enrolled patients or their legally authorized representatives prior to participation in the study. Data collection and variable definitions Baseline demographic and clinical data were collected at the time of hospital admission using a standardized case record form. Collected variables included age, sex, body mass index, comorbidities (hypertension, diabetes mellitus, chronic kidney disease, cardiovascular disease), medication history, vital signs on admission, and severity of illness scores, including the Acute Physiology and Chronic Health Evaluation II (APACHE II) and SOFA scores. Laboratory data collection Blood and urine samples were obtained as part of routine clinical care at the time of admission (day 0). Laboratory parameters recorded included: complete blood count (CBC), blood urea nitrogen (BUN), serum creatinine, serum electrolytes (sodium, potassium, chloride, bicarbonate), serum albumin, liver function tests and arterial blood gas analysis. All laboratory analyses were performed in the hospital’s central laboratory using standardized automated assays. Serum creatinine levels were serially measured on day 0, day 1, day 2, day 3, and day 7 after admission, or until hospital discharge or death, whichever occurred first. Definition and calculation of NLR The NLR was calculated from the admission CBC as the absolute neutrophil count divided by the absolute lymphocyte count. Only the first available CBC within 24 h of admission was used for NLR calculation to ensure uniform baseline assessment and to minimize time-dependent confounding. Outcome definitions The primary outcome was the AKI within the first 7 days of hospitalization. AKI was diagnosed according to the Kidney Disease: Improving Global Outcomes (KDIGO) 2012 criteria, based solely on changes in serum creatinine: Increase in serum creatinine ≥ 0.3 mg/dL within 48 h, or. Increase in serum creatinine ≥ 0.3 mg/dL within 48 h, or. Increase in serum creatinine to ≥ 1.5 times baseline within 7 days. Increase in serum creatinine to ≥ 1.5 times baseline within 7 days. Urine output criteria were not applied due to incomplete and inconsistent documentation in non–intensive care settings. Secondary outcomes included Requirement for renal replacement therapy (RRT) during hospitalization. Requirement for renal replacement therapy (RRT) during hospitalization. All-cause in-hospital mortality. All-cause in-hospital mortality. Length of hospital stay (LOS), defined as the number of days from admission to hospital discharge or death. Length of hospital stay (LOS), defined as the number of days from admission to hospital discharge or death. Outcome adjudication and data quality control Clinical outcomes, including AKI diagnosis, initiation of RRT, and mortality status, were independently adjudicated by an attending nephrologist who was blinded to NLR values to minimize assessment bias. All collected data were reviewed for completeness and internal consistency prior to analysis. Data completeness exceeded 98% across all variables. No data imputation was performed. Missing values were handled using pairwise deletion during statistical analysis. Sample size calculation Based on prior studies reporting a sensitivity of 90% for NLR in predicting AKI [12] and an expected AKI prevalence of 50% among sepsis patients [1, 13], a minimum sample size of 86 was estimated using n4Studies Plus software, targeting a 95% confidence level and 9% absolute precision [19, 20]. We enrolled 112 participants to account for potential dropouts. Statistical analysis All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY) and R version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were initially assessed for normality using the Shapiro–Wilk test. Normally distributed data were presented as mean ± standard deviation (SD) and compared between groups using Student’s t-test. Non-normally distributed variables were summarized as median and interquartile range (IQR) and compared using the Mann–Whitney U test. Categorical variables were expressed as counts and percentages and compared using the chi-square test or Fisher’s exact test, as appropriate. For the primary analysis, the association between admission NLR and the development of AKI was evaluated using receiver operating characteristic (ROC) curve analysis. The area under the ROC curve (AUC) with 95% confidence intervals (CIs) was calculated to assess discriminatory performance. The optimal NLR cutoff value was determined using the Youden index, and corresponding sensitivity, specificity, and overall predictive accuracy were reported. To identify independent predictors of AKI, univariate and multivariate logistic regression analyses were performed. Variables for the multivariable logistic regression model were selected based on a combination of statistical significance in univariate analysis (P < 0.10) and clinical relevance supported by pathophysiological rationale. To reduce over-parameterization, the final model was limited to a small set of clinically relevant covariates. Multicollinearity among covariates was assessed using variance inflation factors (VIF); all VIF values were below 2.5, suggesting no evidence of problematic multicollinearity. Results of the logistic regression analyses were reported as odds ratios (ORs) with 95% confidence intervals (CIs). Secondary and exploratory analyses included: The association between admission NLR and AKI severity (KDIGO stages 1–3). The association between admission NLR and AKI severity (KDIGO stages 1–3). The predictive performance of combined models incorporating NLR with serum albumin, APACHE II score, and SOFA score, evaluated by comparison of AUC values from ROC analyses. The predictive performance of combined models incorporating NLR with serum albumin, APACHE II score, and SOFA score, evaluated by comparison of AUC values from ROC analyses. All statistical tests were two-tailed, and a P-value < 0.05 was considered statistically significant. Results Participants and baseline characteristics Among the 112 enrolled patients, 32 (28.6%) developed AKI. Patients in the AKI group were significantly older than those without AKI (81.6 ± 10.6 vs. 74.7 ± 13.4 years, P = 0.01). Serum albumin was significantly lower in patients who developed AKI (2.8 ± 0.6 vs. 3.4 ± 0.5 g/dL, P < 0.001). There were no significant differences between the AKI and non-AKI groups in sex, body mass index, baseline serum creatinine, comorbidities, or baseline medication use. Serum lactate levels tended to be higher in the AKI group but did not reach statistical significance (P = 0.07). Detailed baseline characteristics are shown in Table 1. Table 1Baseline characteristicsNote: Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage), as appropriate. Comparisons between AKI and non-AKI groups were performed using the Student’s t test or Mann–Whitney U test for continuous variables and the chi-square test or Fisher’s exact test for categorical variables. Baseline creatinine refers to the first serum creatinine value measured at hospital admission. RAAS = renin–angiotensin–aldosterone system; MAP = mean arterial pressure Baseline characteristics Note: Data are presented as mean ± standard deviation, median (interquartile range), or number (percentage), as appropriate. Comparisons between AKI and non-AKI groups were performed using the Student’s t test or Mann–Whitney U test for continuous variables and the chi-square test or Fisher’s exact test for categorical variables. Baseline creatinine refers to the first serum creatinine value measured at hospital admission. RAAS = renin–angiotensin–aldosterone system; MAP = mean arterial pressure Primary outcome: predictive performance of NLR for AKI Admission NLR was significantly higher in patients who developed AKI compared with those without AKI (median 20.2 [IQR 11.9–47.3] vs. 9.1 [5.1–14.3], P < 0.001) (Fig. 2). In exploratory descriptive analysis, median NLR values appeared to differ across KDIGO stages; however, the small number of patients in each stage precludes meaningful statistical comparison or clinical interpretation of these differences. Distribution of NLR in patients with SA-AKI and those without SA-AKI Using ROC analysis, the optimal NLR cut-off for predicting AKI was ≥ 12.8, yielding a sensitivity of 75% (95% CI 56.6–88.5%), specificity of 72.5% (61.4–81.9%), and overall accuracy of 73.2% (64.0–81.1%) (Table 2). The area under the ROC curve (AUC) for NLR was 0.78 (95% CI 0.686–0.875), indicating moderate discriminatory performance (Fig. 3). At a lower threshold (NLR ≥ 9.9), sensitivity increased to 81.3%, but specificity decreased to 55% (Table 2). Table 2Performance of NLR cut-off values for predicting sepsis-associated acute kidney injuryNote: Sensitivity, specificity, and accuracy were derived from receiver operating characteristic (ROC) curve analysis. Cut-off values were selected based on optimal Youden index. AKI was defined using KDIGO serum creatinine criteria within 7 days of admission Performance of NLR cut-off values for predicting sepsis-associated acute kidney injury Note: Sensitivity, specificity, and accuracy were derived from receiver operating characteristic (ROC) curve analysis. Cut-off values were selected based on optimal Youden index. AKI was defined using KDIGO serum creatinine criteria within 7 days of admission Receiver operating characteristic (ROC) curve analysis for prediction of sepsis-associated acute kidney injury (SA-AKI). Abbreviations: AKI, acute kidney injury; AUC, area under the curve; CI, confidence interval; NLR, neutrophil-to-lymphocyte ratio; SOFA, Sequential Organ Failure Assessment; APACHE II, Acute Physiology and Chronic Health Evaluation II Secondary outcomes The discriminatory performance of admission NLR for predicting RRT was poor, with an AUC of 0.435 (95% CI 0.056–0.814). For in-hospital mortality, NLR demonstrated modest predictive ability, with an AUC of 0.603 (95% CI 0.376–0.829). In contrast, NLR demonstrated moderate predictive performance for the composite outcome of AKI or mortality, with an AUC of 0.754 (95% CI, 0.610–0.857). (Fig. 4). Predictive performance of admission NLR for secondary outcomes. Note: ROC curves were generated using predicted probabilities derived from logistic regression models with admission NLR as the predictor. Abbreviations: NLR, neutrophil-to-lymphocyte ratio; AKI, acute kidney injury; RRT, renal replacement therapy; ROC, receiver operating characteristic; AUC, area under the curve Patients who developed AKI had significantly greater illness severity at presentation, as reflected by higher APACHE II scores (17.7 ± 5.1 vs. 13.4 ± 4.9, P < 0.001) and SOFA scores (4.1 ± 2.6 vs. 2.1 ± 2.5, P = 0.001) compared with those without AKI. The use of vasopressors was more frequent in the AKI group (25.0% vs. 8.8%, P = 0.023), whereas rates of ICU admission and mechanical ventilation did not differ significantly between groups. Length of hospital stay was significantly longer in patients with AKI (median 11 [IQR 6–14] vs. 7 [4–10.5] days, P = 0.039). Although in-hospital mortality was numerically higher in the AKI group (15.6% vs. 6.3%), this difference did not reach statistical significance (P = 0.116) (Table 3). Table 3Clinical parameters and outcomes stratified by acute kidney injury statusNote: Values are presented as odds ratios (OR) with 95% confidence intervals. Multivariable logistic regression analysis was performed to identify independent predictors of acute kidney injury. The adjusted model included neutrophil-to-lymphocyte ratio (NLR), serum albumin, age, vasopressor use, serum lactate, and use of renin–angiotensin–aldosterone system (RAAS) inhibitorsAbbreviations: NLR, neutrophil-to-lymphocyte ratio; AKI, acute kidney injury; RAAS = renin–angiotensin–aldosterone system Clinical parameters and outcomes stratified by acute kidney injury status Note: Values are presented as odds ratios (OR) with 95% confidence intervals. Multivariable logistic regression analysis was performed to identify independent predictors of acute kidney injury. The adjusted model included neutrophil-to-lymphocyte ratio (NLR), serum albumin, age, vasopressor use, serum lactate, and use of renin–angiotensin–aldosterone system (RAAS) inhibitors Abbreviations: NLR, neutrophil-to-lymphocyte ratio; AKI, acute kidney injury; RAAS = renin–angiotensin–aldosterone system Multivariate analysis: independent predictors of AKI In univariate analysis, admission NLR, serum albumin, age, and vasopressor use were significantly associated with the development of AKI. In the multivariate logistic regression model, only NLR and serum albumin remained independently associated with AKI. Admission NLR was independently associated with higher odds of developing AKI (adjusted OR 5.01, 95% CI 1.75–14.35, P = 0.003), whereas higher serum albumin was independently associated with lower odds of AKI (adjusted OR 0.19, 95% CI 0.06–0.58, P = 0.004). Age, vasopressor use, serum lactate, and use of renin–angiotensin–aldosterone system inhibitors (RAAS inhibitors) were not independently associated with AKI in the multivariable model (Table 4). Table 4Multivariable logistic regression analysis of independent predictors of acute kidney injuryNote: Values are presented as odds ratios (OR) with 95% confidence intervals. Multivariable logistic regression analysis was performed to identify independent predictors of acute kidney injury. The adjusted model included neutrophil-to-lymphocyte ratio (NLR), serum albumin, age, vasopressor use, serum lactate, and use of renin–angiotensin–aldosterone system (RAAS) inhibitorsAbbreviations: NLR, neutrophil-to-lymphocyte ratio; AKI, acute kidney injury; RAAS = renin–angiotensin–aldosterone system Multivariable logistic regression analysis of independent predictors of acute kidney injury Note: Values are presented as odds ratios (OR) with 95% confidence intervals. Multivariable logistic regression analysis was performed to identify independent predictors of acute kidney injury. The adjusted model included neutrophil-to-lymphocyte ratio (NLR), serum albumin, age, vasopressor use, serum lactate, and use of renin–angiotensin–aldosterone system (RAAS) inhibitors Abbreviations: NLR, neutrophil-to-lymphocyte ratio; AKI, acute kidney injury; RAAS = renin–angiotensin–aldosterone system Combined predictive models of NLR with clinical variables The addition of serum albumin to NLR significantly improved the predictive performance for AKI, yielding an AUC of 0.83 (95% CI 0.73–0.93). Further integration of APACHE II or SOFA scores resulted in incremental improvements in model discrimination, with an AUC of 0.86 (95% CI 0.77–0.95) for NLR + albumin + APACHE II, 0.84 (95% CI 0.73–0.94) for NLR + albumin + SOFA, and 0.85 (95% CI 0.76–0.95) for the fully combined model (NLR + albumin + APACHE II + SOFA). Across all combined models, specificity remained consistently high (approximately 89–91%), while sensitivity ranged from 42.9% to 62.5% (Table 5). Table 5Predictive performance of combined NLR-based models for predicting acute kidney injuryNote: Area under the curve (AUC), sensitivity, specificity, and accuracy were derived from receiver operating characteristic (ROC) analyses using predicted probabilities generated from multivariable logistic regression models. Models included combinations of neutrophil-to-lymphocyte ratio (NLR), serum albumin, Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and Sequential Organ Failure Assessment (SOFA) score as indicatedAbbreviations: NLR, neutrophil-to-lymphocyte ratio; AUC, area under the curve; APACHE II, Acute Physiology and Chronic Health Evaluation II; SOFA, Sequential Organ Failure Assessment Predictive performance of combined NLR-based models for predicting acute kidney injury Note: Area under the curve (AUC), sensitivity, specificity, and accuracy were derived from receiver operating characteristic (ROC) analyses using predicted probabilities generated from multivariable logistic regression models. Models included combinations of neutrophil-to-lymphocyte ratio (NLR), serum albumin, Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and Sequential Organ Failure Assessment (SOFA) score as indicated Abbreviations: NLR, neutrophil-to-lymphocyte ratio; AUC, area under the curve; APACHE II, Acute Physiology and Chronic Health Evaluation II; SOFA, Sequential Organ Failure Assessment Discussion This prospective study evaluated the NLR as a predictor of SA-AKI in a Southeast Asian hospitalized cohort. We found that admission NLR was significantly higher in patients who developed AKI and remained independently associated with AKI after multivariable adjustment. With an AUC of 0.78 and a sensitivity of 75% at the optimal cut-off, NLR demonstrated moderate discriminatory performance for early AKI risk stratification rather than definitive diagnosis. The advanced mean age of our cohort (77 ± 11 years) reflects the demographic profile of sepsis admissions at our institution. As a result, these findings may not be directly generalizable to younger sepsis populations or to healthcare settings with different demographic characteristics. External validation in more diverse cohorts is therefore warranted. The inclusion of both intensive care and general ward patients supports the applicability of our findings across different levels of care, particularly in settings where access to novel biomarkers remains limited [4, 21]. At the optimal cut-off value of 12.8, NLR demonstrated a sensitivity of 75% and a specificity of 72%, with an AUC of 0.78, indicating moderate discriminatory performance for predicting SA-AKI. Although this performance is lower than that reported in highly selected ICU cohorts with AUCs approaching 0.90–0.95 [12, 13], the difference likely reflects the heterogeneity of our real-world population, which included both ICU and general ward patients with a predominance of KDIGO stage 1 cases. The observed AKI incidence of 29% is consistent with that reported in non-ICU sepsis populations. Compared with novel kidney-specific biomarkers, NLR offers the advantage of immediate availability from routine complete blood counts without additional cost or specialized infrastructure. However, as a non-specific systemic inflammatory marker rather than a kidney-specific biomarker, the moderate predictive performance of NLR is expected and should be interpreted within this context [5, 22]. Accordingly, our findings support the role of NLR as a widely applicable first-line risk stratification marker, particularly in settings where access to emerging biomarkers is limited, while acknowledging that its moderate sensitivity precludes its use as a standalone diagnostic test. The combination of NLR and serum albumin improved the discriminatory performance for SA-AKI prediction, achieving an AUC of 0.83 with high specificity (~ 89%). Serum albumin, which reflects both nutritional reserve and systemic capillary leak in sepsis, provides complementary information to the inflammatory signal represented by NLR [23, 24]. The incremental improvement observed with this combination suggests that integrating multiple routinely available clinical parameters may enhance early risk stratification beyond the use of a single biomarker alone. This approach may be particularly relevant in clinical settings where access to emerging biomarkers is limited, although its performance remains constrained by the moderate sensitivity observed in the present study. Although dedicated mechanistic investigations were not performed, our observations are biologically consistent with several pathophysiological pathways implicated in sepsis-associated AKI. Elevated NLR reflects both excessive innate immune activation, including neutrophil extracellular trap formation and reactive oxygen species generation, and relative suppression of adaptive immunity through sepsis-induced lymphocyte apoptosis and immunoparalysis [8, 25, 26]. These processes have been implicated in renal microcirculatory impairment and tubular injury in experimental models [27]. The improved performance of the combined NLR-albumin model is consistent with the concept that endothelial dysfunction and capillary leak, reflected by hypoalbuminemia, represent complementary pathways to renal injury in sepsis, integrating inflammatory and vascular signals within a single predictive framework [28, 29]. From a clinical perspective, our findings suggest that admission NLR may serve as a simple first-line risk stratification marker for sepsis-associated AKI, derived immediately from routine complete blood counts. In particular, the cut-off value of NLR ≥ 12.8 identifies a subgroup of patients at substantially higher odds for developing AKI. However, these observations should be considered hypothesis-generating rather than prescriptive, as formal prospective validation of multi-tier risk categories was not performed in this study. The optimal NLR cut-off of ≥ 12.8 identified in this study was derived from internal ROC analysis using the Youden index and should be regarded as a data-driven estimate rather than a validated clinical threshold. External validation in independent, multicenter cohorts with diverse demographics and clinical settings is required before this threshold can be recommended for routine clinical use. Importantly, our results do not support the use of NLR as a sole determinant of clinical decision-making; rather, NLR should be interpreted in conjunction with established clinical, hemodynamic, and biochemical risk factors. The potential role of NLR-based risk stratification in guiding intensified renal surveillance and nephroprotective strategies warrants prospective validation in multicenter interventional trials. This study has several important strengths. First, the prospective design and standardized data collection minimize information bias and enhance internal validity. Second, the inclusion of both intensive care and general ward patients reflects real-world clinical practice and improves the generalizability of our findings within hospitalized sepsis populations. Third, all laboratory variables, including NLR and serum creatinine, were derived from routine clinical measurements, supporting the feasibility of applying our results in everyday settings. Several limitations should be acknowledged. First, this was a single-center study conducted at a tertiary referral hospital in Thailand with a predominantly elderly cohort, which may limit the generalizability of our findings to younger sepsis populations and to other healthcare settings. External validation across more diverse populations is needed before broader clinical application. Second, the relatively modest sample size (n = 112, with 32 AKI events) may have limited the statistical power for subgroup analyses and the precision of our predictive performance estimates. The limited number of AKI events (n = 32) relative to the number of covariates in the multivariable model resulted in an events-per-variable ratio of approximately 5.3, which is below the conventional rule-of-thumb threshold of 10 events per variable. This raises the possibility of model overfitting and may have reduced the stability of the regression estimates. Accordingly, the results of the multivariable analysis should be interpreted with caution and considered hypothesis-generating pending validation in larger cohorts. The small sample size within individual KDIGO stages precluded meaningful stage-specific analyses; therefore, these observations should be regarded as exploratory and hypothesis-generating only. Third, the sample size calculation was based on an assumed AKI prevalence of 50% from prior ICU-based sepsis studies; however, the observed AKI prevalence in our cohort was 28.6%. This lower-than-expected event rate may have reduced the precision of the diagnostic performance estimates, widened confidence intervals, and limited the interpretability of secondary and stage-specific exploratory analyses. Validation in larger cohorts with adequate event rates is warranted. Fourth, baseline serum creatinine was defined as the first measured value at hospital admission and may therefore not have reflected the true pre-illness baseline kidney function. In patients with pre-existing chronic kidney disease or kidney injury already present before hospital arrival, this approach may have resulted in misclassification or underestimation of AKI incidence. Ideally, a stable pre-admission outpatient creatinine value would have provided a more accurate reference, but such data were not consistently available in this cohort. In addition, urine output criteria were not incorporated into the AKI definition because of incomplete and inconsistent documentation in non-ICU settings. As a result, oliguria-predominant AKI may have been underrecognized, and the reported AKI incidence and predictive performance estimates may represent conservative estimates. Fifth, early in-hospital mortality before sufficient serial creatinine assessment may have introduced survivorship bias or a competing-risk effect. Patients who died early in the hospital course may have had insufficient time for AKI to be detected using creatinine-based KDIGO criteria, which could have led to underascertainment of AKI and underestimation of the association between elevated NLR and adverse renal outcomes. Sixth, several potential confounders were not fully accounted for, including infection source, timing of antimicrobial therapy, and post-admission exposure to nephrotoxic agents. These factors may have influenced inflammatory burden, illness severity, and subsequent AKI risk. Seventh, only a single admission NLR measurement was used; serial NLR measurements were not performed, precluding assessment of dynamic trajectories that may further enhance predictive performance. Eighth, novel biomarkers such as NGAL, KIM-1, or TIMP-2/IGFBP7 were not measured for direct head-to-head comparison with NLR. Ninth, the optimal NLR cut-off of ≥ 12.8 identified in this study was derived from internal ROC analysis using the Youden index and should be regarded as a data-driven estimate rather than a validated clinical threshold. External validation in independent, multicenter cohorts with diverse demographics and clinical settings is essential before this threshold can be recommended for routine clinical use. Finally, as an observational study, causal inferences cannot be established, and residual confounding from unmeasured variables cannot be entirely excluded despite multivariable adjustment. Future studies are warranted to extend these findings in several important directions. First, multicenter validation across diverse patient populations and healthcare settings is needed to confirm the generalizability of NLR-based risk stratification for SA-AKI. Second, prospective evaluation of dynamic changes in NLR and the optimal timing of measurement may further refine its predictive utility. Third, the development and validation of integrated risk models combining NLR with other readily available clinical and biochemical parameters may enhance predictive performance. Such approaches may incorporate advanced analytical methods, provided that model transparency and clinical interpretability are preserved [3, 14, 30]. Conclusions In this prospective cohort study, admission NLR was independently associated with the development of SA-AKI and demonstrated moderate discriminatory performance for early risk stratification. The combination of NLR with serum albumin further improved model performance, supporting the value of integrating multiple readily available clinical parameters. Although NLR showed good specificity, its moderate sensitivity indicates that it should not be used as a standalone predictive tool, but rather as a component of a multimodal risk assessment approach. Our findings highlight the potential of routinely available laboratory indices to contribute to early AKI risk stratification in sepsis. The identified NLR cut-off value requires external validation in independent multicenter cohorts before routine clinical use. Larger multicenter studies with prospective validation are required to confirm these results and to define the optimal clinical role of NLR in diverse healthcare settings. Acknowledgements All nephrology staffs from division of nephrology, Department of Medicine, Phramongkutklao Hospital and College of medicine. Author contributions PC designed the study, collected data, performed statistical analysis, and drafted the manuscript. BS provided supervision and critical revision of the manuscript. TT contributed to data collection and interpretation. OS contributed to study design and critical revision. NS conceived the study, supervised all aspects of the work, and critically revised the manuscript. PT and WK contributed to data collection and interpretation. All authors read and approved the final manuscript. Funding A part of the present study was supported by department of medicine, Phramongkutklao Hospital and College of medicine. Data availability The datasets generated and/or analyzed during the current study are available from the corresponding author, Narongrit Siriwattanasit, upon reasonable request (Email: nsiriwattanasit@gmail.com). Declarations Ethics approval and consent to participate The study was approved by the Ethics Committee of the Institute Review Board at the Royal Thai Army Medical Department and was conducted according to the Declaration of Helsinki. Informed consent was obtained from all participants. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are available from the corresponding author, Narongrit Siriwattanasit, upon reasonable request (Email: nsiriwattanasit@gmail.com).