Relationship between the creatinine/albumin ratio and the ICU mortality rate of patients with acute respiratory distress syndrome: A retrospective cohort study

Relationship between the creatinine/albumin ratio and the ICU mortality rate of patients with acute respiratory distress syndrome: A retrospective cohort 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 Research Article Relationship between the creatinine/albumin ratio and the ICU mortality rate of patients with acute respiratory distress syndrome: A retrospective cohort study Zixuan Jiang, Yueyue Zhang, Ge Zhang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7144283/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 Background : Creatinine (Cr) and albumin (ALB) are widely recognized as predictive biomarkers of mortality across various disease conditions, including patients with acute respiratory distress syndrome (ARDS). Nevertheless, limited evidence is available regarding the associations between the creatinine-to-albumin ratio and clinical outcomes in ARDS patients. This study aimed to assess the prognostic value of the creatinine-to-albumin ratio in predicting 28-day mortality, all-cause mortality, and the incidence of acute kidney injury (AKI) among patients diagnosed with ARDS. Methods : This retrospective cohort study utilized clinical data extracted from the database of Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts, USA, spanning the period from 2008--2019. The creatinine‒albumin ratio (CAR) was calculated on the basis of measurements taken within 24 hours of patient admission. Kaplan‒Meier (K‒M) analysis was employed to compare 28-day mortality, all-cause mortality, and the incidence of acute kidney injury across the four patient groups. A Cox proportional hazards regression model and RCS were used to assess the relationships between the CAR and the risks of 28-day mortality, long-term all-cause mortality, and acute kidney injury. The predictive performance of the CAR—including its sensitivity, specificity, and AUC—was evaluated via receiver operating characteristic (ROC) curve analysis for the aforementioned outcomes in patients with ARDS. Subgroup analyses were also conducted to further validate the robustness and reliability of our findings. Results : A total of 1,233 patients were enrolled in the study. K‒M analysis revealed statistically significant differences in 28-day mortality, overall all-cause mortality, and the incidence of AKI across CAR quartiles (log-rank P < 0.001). Patients with elevated CAR levels presented increased risks of both 28-day and all-cause mortality, as well as a higher cumulative incidence of AKI. After adjusting for potential confounding factors, the multivariate Cox proportional hazards regression model confirmed a statistically significant association between the CAR and each of the three clinical outcomes. Furthermore, RCS analysis demonstrated a significant U-shaped nonlinear relationship between the CAR and these outcomes. ROC curve analysis revealed that the AUC values for the continuous CAR in predicting 28-day mortality, all-cause mortality, and AKI in patients with ARDS were 0.729, 0.716, and 0.785, respectively. When analyzed by quartiles, the corresponding AUCs were slightly improved at 0.732, 0.719, and 0.794. Subgroup analyses indicated that the associations between the CAR and clinical outcomes were more pronounced among patients aged >65 years, males, and those with a history of myocardial infarction or peripheral vascular disease. Notably, in patients with cerebrovascular disease, the risk of mortality did not increase with increasing CAR (hazard ratio [HR] = 0.88, P = 0.414), and the risk of AKI was attenuated in patients with diabetes and chronic complications (HR = 0.79, P = 0.303). Conclusion: The CAR can serve as an independent predictor of 28-day mortality, all-cause mortality, and acute kidney injury. This study also revealed differences in the CAR among different subgroups, which may provide promising prognostic biomarkers for risk stratification and clinical management of patients with ARDS. Creatinine/albumin ratio acute respiratory distress syndrome 28-day mortality rate all-cause mortality rate MIMIC-IV serum biomarker Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Acute respiratory distress syndrome is a form of acute respiratory failure resulting from noncardiogenic pulmonary edema [ 1 ]. According to numerous studies, the mortality rate associated with ARDS typically ranges between 25% and 40% [ 2 , 3 ]. Patients with ARDS frequently exhibit renal dysfunction and systemic inflammatory responses. Creatinine is widely recognized as a key biomarker for assessing kidney function. However, variations in creatinine levels do not always accurately reflect true alterations in renal function, as its concentration is affected by multiple physiological factors, such as age, sex, muscle mass, and dietary habits [ 4 ]. Albumin, one of the principal plasma proteins, not only serves as an indicator of nutritional status but also plays a crucial role in inflammatory responses, vascular permeability, and fluid homeostasis. Accumulating evidence indicates that the serum ALB concentration is significantly associated with the risk of developing ARDS[ 5 ]. Nevertheless, albumin concentrations can be influenced by a range of factors, including nutritional state, systemic inflammation, hepatic function, and renal function. Therefore, relying solely on the serum ALB concentration as a prognostic marker may fail to fully capture the complexity of a patient’s clinical condition and long-term outcomes. The CAR, which integrates both creatinine and albumin levels, has been proposed as a novel biomarker for predicting disease prognosis. By combining these two clinically relevant parameters, the CAR enables a more comprehensive evaluation of a patient's overall health status and disease severity. Prior studies have demonstrated that the CAR is an independent predictor of all-cause mortality among ICU patients with acute pancreatitis [ 6 ]. However, no published research has yet examined the utility of the CAR in predicting outcomes for patients with ARDS. This study aimed to investigate the prognostic value of the CAR in patients with ARDS by retrospectively analyzing data obtained from the Medical Information Mart for Intensive Care (MIMIC-IV, v2.0) database in the United States. Methods Data source This retrospective cohort study utilized data from the Medical Information Mart for Intensive Care IV (MIMIC-IV v3.0), a publicly accessible critical care database containing structured clinical records of 192,500 unique patients (455,000 hospital admissions) at Beth Israel Deaconess Medical Center in Boston, Massachusetts, USA, between 2008 and 2019. The database offers detailed documentation encompassing demographic characteristics, laboratory test results, medication administration records, continuous longitudinal vital signs, procedural interventions, disease diagnoses (coded according to the International Classification of Diseases [ICD] system), infusion parameters of therapeutic agents, and postdischarge survival follow-up data. Population Enrollment This study utilized data from the latest version of the MIMIC-IV v3.0, which encompasses clinical records collected between 2012 and 2019. Patients diagnosed with ARDS were identified on the basis of standardized International Classification of Diseases coding criteria (ICD-9-CM: 518.51, 518.81; ICD-10-CM: J80, J96.0, J96.00–J96.02, R06.03), resulting in a total of 30,509 identified cases. The inclusion criteria were as follows: age between 18 and 85 years, first hospital admission with an initial intensive care unit (ICU) stay, and ICU length of stay ≥24 hours (n=13,977). Following the exclusion of patients who lacked baseline serum creatinine (Cr) or albumin (Alb) measurements within 24 hours of ICU admission, a final cohort of 1,233 subjects was included in the analysis. These patients were stratified into four groups (Q1–Q4) according to quartile thresholds of the creatinine-to-albumin ratio. A flowchart detailing the selection process of the study population is illustrated in Figure 1. Data Extraction Strategy Data extraction was conducted via Structured Query Language (SQL) through Navicat Premium (Version 16.1.15). The variables included in the analysis were systematically categorized into six domains: demographics, comorbidities, vital signs, laboratory indicators (initial measurements obtained at hospital admission prior to any therapeutic interventions), healthcare utilization and outcomes, and severity scores. Variables with more than 30% missing data were excluded from the analysis. For variables with ≤30% missing values, multiple imputation was carried out via a random forest algorithm. Table 1 presents a comprehensive list of the extracted variables. Calculation of the CAR The CAR was calculated as [(admission creatinine (mg/dl))/(albumin (mg/dl))], and admission creatine and albumin were obtained directly from MIMIC IV. Statistical analysis Continuous variables were first evaluated for normality via the Shapiro‒Wilk test. Nonnormally distributed data were analyzed via the Wilcoxon rank-sum test and summarized as medians (interquartile ranges, IQRs). Categorical variables were compared via either the chi-square test or Fisher’s exact test, as appropriate, and are presented as counts (percentages). To investigate the associations between the CAR index, analyzed as both continuous and quartile-based variables, and 28-day mortality as well as AKI, four progressively adjusted logistic regression models were developed. Model 1 was unadjusted; Model 2 was adjusted for demographic characteristics (age, sex); Model 3 further incorporated metabolic parameters (albumin, anion gap, bicarbonate, blood urea nitrogen, chloride, creatinine, glucose) and erythrocyte indices (mean corpuscular volume [MCV], mean corpuscular hemoglobin [MCH], mean corpuscular hemoglobin concentration [MCHC]); Model 4 additionally controlled for comorbid conditions (myocardial infarction, peripheral vascular disease, cerebrovascular disease, chronic pulmonary disease) and the Charlson Comorbidity Index (CCI), with the lowest CAR quartile (Q1) serving as the reference group. Restricted cubic splines (RCSs, 3 knots) were applied to assess potential dose‒response relationships between continuous CAR levels and clinical outcomes, whereas a recursive partitioning algorithm was used to detect inflection points within nonlinear associations (likelihood ratio test P < 0.05). Stratified analyses were performed according to sex, age, diabetes status (with or without complications), sepsis, peripheral vascular disease, myocardial infarction, chronic pulmonary disease, dementia, and cerebrovascular disease to evaluate heterogeneity via the Breslow‒Day test. All the statistical analyses were conducted via SPSS (v22.0) and R software (v4.3.2), with a two-tailed P value of less than 0.05 considered statistically significant. Results Baseline demographic and clinical characteristics This baseline table presents a comparative analysis of clinical characteristics among 1,233 patients stratified by creatinine‒albumin ratio (CAR) quartiles (Q1–Q4). Overall, the data reveal a statistically significant trend toward worsening clinical profiles with increasing CAR values (Q1: 0.176 to Q4: 0.942). Patients in the highest CAR quartile (Q4) represented the highest proportion of males (63.9%), the longest median hospital length of stay (11.65 days), and the longest ICU duration (5.87 days). This group also presented elevated markers of physiological instability, including increased lactate levels (3.054 mmol/L), INR (1.907), and prothrombin time (PT, 21.21 s), as well as more severe inflammatory responses (albumin: 2.79 g/dL, WBC: 16.76×10³/μL) and greater organ dysfunction (SOFA score: 53.5, serum creatinine: 2.53 mg/dL). Comorbidity analysis indicated that Q4 had the highest Charlson Comorbidity Index (5.27) and the highest incidence of acute kidney injury (50.95%). The 28-day mortality rate in Q4 was 65.4%, significantly exceeding that in Q1 (9.8%) (p < 0.001). Notably, Q4 demonstrated a greater prevalence of sepsis (88.2%) and a greater need for mechanical ventilation (77.6%). Compensatory changes were observed in the platelet count (155×10³/μL) and hemoglobin-related parameters (MCHC: 32.88 g/dL). No significant intergroup differences were found for body temperature, Glasgow Coma Scale (GCS) score, or chloride level (p > 0.05). The detailed results are presented in Table 2 . LOS_hospital, length of hospital stay; LOS_icu, length of ICU stay; total CO2, total carbon dioxide emissions; INR,integer normalized ratio; PT, prothrombin time; PTT, partial thromboplastin time; GCS, Glasgow Coma Scale; HR, heart rate; SAEPS Ⅱ, simplified acute physiology score Ⅱ; OASIS, organ dysfunction assessment system; Temp, temperature; MechVent, mechanical ventilation; SIRS, systemic inflammatory response syndrome; AIDS, acquired immunodeficiency syndrome; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; RBC, red blood cell count; RDW, red blood cell distribution width; WBC, white blood cell count; AKI, acute kidney injury; CAR, creatinine‒albumin ratio; Kaplan‒Meier curve analysis Kaplan‒Meier analysis revealed significant differences in 28-day mortality, overall survival, and AKI incidence across CAR quartiles (overall log-rank P < 0.001). Compared with the lowest CAR quartile (Q1), the highest quartile (Q4) presented significantly lower cumulative survival rates for both 28-day mortality and overall mortality, as well as a greater cumulative incidence of AKI. The results are presented in Figure 2. The predictive efficacy of the CAR for clinical outcomes and the results of multivariate regression analysis We evaluated the predictive performance of both continuous CAR values and CAR quartiles for three key clinical outcomes: 28-day mortality, all-cause mortality, and AKI. As illustrated in Figure 3, the AUC analysis demonstrated that the CAR exhibited robust predictive ability across all evaluated endpoints. For 28-day mortality, the AUC values were 0.729 (95% CI: 0.640–0.824) for the continuous CAR quartile and 0.732 (95% CI: 0.638–0.813) for the CAR quartile. Similarly, for all-cause mortality, the AUC values were 0.716 (95% CI: 0.616–0.805) and 0.719 (95% CI: 0.628–0.800) for the continuous and quartile-based CAR measurements, respectively. Notably, the CAR demonstrated particularly strong predictive accuracy for AKI, with AUC values of 0.785 (95% CI: 0.709–0.853) for the continuous CAR and 0.794 (95% CI: 0.724–0.858) for the CAR quartiles. Collectively, these findings indicate that the CAR has substantial predictive value across multiple critical clinical outcomes. This study utilized multistage-adjusted logistic regression models to investigate the associations between the predictor variable and three clinical outcomes, which were differentially influenced by confounding factors. As presented in Table 3, in the unadjusted model, each one-unit increase in the continuous predictor variable was significantly associated with increased risks of 28-day mortality (OR = 16.43, 95% CI: 10.09–27.85), all-cause mortality (OR = 11.75, 95% CI: 7.536–18.97), and AKI (OR = 3.938, 95% CI: 2.830–5.582). All p values were < 0.001. However, after adjusting for laboratory parameters (e.g., albumin, creatinine) and comorbidities, the direction of the association with mortality outcomes was reversed (Model 4: 28-day mortality OR = 0.117, p = 0.002; all-cause mortality OR = 0.117, p = 0.002), whereas a significant inverse association with AKI remained (Model 4: OR = 0.087, p < 0.001). Quartile-based analysis revealed a clear dose‒response relationship, with the highest quartile (Q4) showing the greatest risk in fully adjusted models: 28-day mortality (OR = 2.506, 95% CI: 1.201–5.245, p = 0.014), all-cause mortality (OR = 2.506, same CI range), and AKI (OR = 13.83, 95% CI: 6.561–30.21, p < 0.001). Notably, AKI maintained statistical significance across the Q2 and Q3 groups in all the models (Model 4: Q2 OR = 2.713, p = 0.002; Q3 OR = 5.215, p < 0.001), whereas the associations between mortality and the Q2–Q3 quartiles became nonsignificant after adjustment. The final model (Model 4) indicated that laboratory markers and the Charlson Comorbidity Index explained 76.3–93.8% of the original risk variance in mortality outcomes, compared with only 39.1% for AKI. This residual association suggests that the predictor variable may exert a direct pathophysiological effect. Detection of nonlinear relationships This series of RCS plots demonstrated a significant U-shaped nonlinear association between the CAR and the risk of 28-day mortality, all-cause mortality, and AKI in critically ill patients (Fig. 4). Specifically, a specific CAR range is associated with the nadir risk for these adverse outcomes. Significant risk elevation is observed when CAR values fall below or rise above this optimal range. This robust U-shaped association was highly statistically significant in both the minimally adjusted model (Model 1) and the model incorporating additional variables (Model 2) for all outcomes (P-nonlinear < 0.0001 for all models/outcomes). Critically, this nonlinear U-shaped relationship remained statistically significant even in the most rigorously adjusted models (Model 3 & Model 4), despite potential alterations in the precise curve morphology or increased estimation uncertainty (wider confidence intervals) with progressive covariate inclusion (Model 3 P-nonlinear = 0.0301, Model 4 P-nonlinear = 0.0130, as applicable across outcomes). Subgroup analyses To further investigate the associations between the CAR and 28-day mortality, long-term all-cause mortality, and acute kidney injury, we performed stratified analyses according to sex, age, diabetes status (with or without complications), sepsis, peripheral vascular disease, myocardial infarction, chronic lung disease, dementia, and cerebrovascular disease. The results are presented in Figure 5. 28-day mortality This study performed a subgroup analysis of 28-day mortality. The overall results revealed a significantly elevated mortality risk across the entire population (HR = 1.87, 95% CI: 1.66–2.10, P < 0.001). With respect to demographic characteristics, patients older than 65 years (HR = 2.06) and males (HR = 2.10) presented greater mortality risks than did those aged ≤65 years (HR = 1.75) and females (HR = 1.67), although the interaction effects for age and sex were not statistically significant (P = 0.088 and 0.072, respectively). In the diabetes subgroup, diabetic patients without chronic complications presented a lower risk than nondiabetic patients did (HR = 1.50 vs. 2.00; interaction P = 0.061), whereas the subgroup with chronic complications (HR = 1.21, P = 0.586) did not present a significant difference, likely due to the small sample size (n = 69). Notably, patients with cerebrovascular disease exhibited a significant inverse interaction effect (P < 0.001), with no increased mortality risk (HR = 0.88, P = 0.414), in contrast to those without cerebrovascular disease, who had an HR of 2.13 (P < 0.001). Among the other disease subgroups, patients with a history of myocardial infarction presented the highest risk (HR = 2.25), and both peripheral vascular disease (HR = 2.15) and chronic pulmonary disease (HR = 1.61) were significantly associated with increased mortality. All-cause mortality This study conducted subgroup analyses of all-cause mortality, which revealed a significantly elevated overall risk (HR = 1.82, 95% CI: 1.62–2.04, P < 0.001). Subgroup analyses revealed statistically significant interaction effects for age (Pinteraction = 0.032), sex (Pinteraction = 0.05), and cerebrovascular disease (Pinteraction < 0.001). Specifically, patients older than 65 years (HR = 2.06 vs. ≤65 years, HR = 1.68) and males (HR = 2.07 vs. females, HR = 1.62) presented increased mortality risks, whereas patients with cerebrovascular disease presented no increased mortality risk (HR = 0.93, P = 0.628), suggesting potential protective associations. The key high-risk subgroups included those with a history of myocardial infarction (HR = 2.25, P < 0.001) and peripheral vascular disease (HR = 2.15, P 1.5, P 0.05). Acute kidney injury This subgroup analysis of patients with AKI revealed a significantly elevated overall risk (HR = 1.94, 95% CI: 1.69–2.22, P 65 years: HR = 2.05) and sexes (male: HR = 2.13; female: HR = 1.72) (all P < 0.001), although the interaction effects of age and sex were not statistically significant (P = 0.577 and 0.153, respectively). Among the comorbidities, diabetic patients with chronic complications presented a significantly lower AKI risk (HR = 0.79, P = 0.303) and a notable interaction effect (P < 0.001), whereas those without complications presented a greater risk (HR = 1.51, P = 0.006). Nonseptic patients demonstrated substantially greater AKI risk than did septic patients (HR = 2.80 vs. 1.75, Pinteraction = 0.002), with myocardial infarction (HR = 2.28) and chronic pulmonary disease (HR = 2.27) identified as additional high-risk factors. The dementia subgroup (n = 11) yielded unreliable results (HR = 40192.97, P = 1.000) because of an insufficient sample size, whereas peripheral vascular disease and cerebrovascular disease did not show significant between-group differences. Discussion This study investigated the connection between the CAR and clinical outcomes in patients diagnosed with acute ARDS. The results revealed a notable U-shaped nonlinear correlation between CAR levels and the likelihood of 28-day mortality, overall mortality, and the occurrence of acute kidney injury. Within a certain range of CAR values, the risk of these negative outcomes was lowest; however, when CAR values fell outside this optimal range, the risk of experiencing these three adverse events increased significantly. The AUC analysis demonstrated strong predictive accuracy for all studied outcomes, especially for acute kidney injury. These observations imply that the CAR could serve as a valuable biomarker for evaluating the prognosis of ARDS patients. By integrating measurements of both creatinine and albumin, the CAR offers a more holistic evaluation of a patient's general health condition and illness severity, thereby enhancing the accuracy of outcome predictions for ARDS patients. In our research, initial patient data displayed a distinct dose‒response pattern relative to CAR levels. As the CAR quartiles increased, patients tended to present with increasingly severe inflammation, coagulation abnormalities, tissue hypoxia, and multiorgan dysfunction. These findings support the notion that the CAR captures both renal impairment and systemic inflammatory or nutritional disturbances, making it a well-suited marker for assessing the overall burden of critical illness. In the Kaplan‒Meier analysis, notable differences were observed in both 28-day mortality and overall all-cause mortality across patient groups categorized by different CAR quartiles (log-rank test, P < 0.001). To further explore this relationship, a multivariable logistic regression analysis was conducted, revealing that confounding variables significantly influenced the association between CAR levels and mortality outcomes. Initially, in the unadjusted model, each unit increase in CAR was linked to a 16-fold higher risk of 28-day mortality (OR = 16.43). However, after adjusting for key clinical factors—including laboratory markers such as albumin and creatinine, as well as the Charlson Comorbidity Index—the direction of this association changed dramatically (OR = 0.117 in Model 4). This shift suggests that comorbidities and laboratory parameters substantially modify the impact of the CAR on mortality risk, potentially by interacting with the CAR to influence patient survival. For example, the level of albumin, which serves as a marker for both nutritional status and systemic inflammation, may reflect worsened health conditions when it is decreased, thereby contributing to increased mortality risk [ 7 ]. Moreover, concurrent conditions such as cardiovascular diseases and diabetes can further deteriorate a patient's clinical condition and increase the likelihood of mortality [ 8 , 9 ]. In contrast to mortality outcomes, the association between the CAR and acute kidney injury remained robust even after full adjustment for confounding variables. The area under the curve (AUC) values were 0.785 (95% confidence interval: 0.709–0.853) for continuous CAR measurements and 0.794 (95% confidence interval: 0.724–0.858) for quartile-based analysis. These results indicate that the CAR is a highly effective biomarker for identifying patients at risk of developing acute kidney injury, with superior predictive performance compared with several traditional indicators. For instance, the level of serum creatinine, a widely used marker of renal function, is associated with a delayed response. Prior studies have shown that creatinine levels typically do not increase until 48 to 72 hours after renal injury has occurred, at which time significant kidney function may be lost [ 10 ]. Changes in the CAR could act as early signals of the development of renal impairment. In a stepwise multivariable logistic regression analysis, the initial unadjusted continuous CAR assessment revealed a marked increase in the likelihood of AKI with each unit increase (OR = 3.938, 95% CI: 2.830–5.582, p < 0.001). Following complete adjustment for confounding factors, the CAR still exhibited a significant inverse relationship with AKI (Model 4 OR = 0.087, p < 0.001), indicating its considerable independent predictive capacity for the occurrence of AKI. Importantly, in Model 4, laboratory variables and the Charlson Comorbidity Index together accounted for 76.3–93.8% of the initial variance in mortality risk, whereas AKI alone contributed merely 39.1%. This residual association suggests that the CAR may have a direct pathophysiological impact on kidney damage, separate from the patient's overall clinical status. Additional subgroup analyses revealed higher mortality risks among older individuals (over 65 years of age) and male patients, which might be linked to age-associated reductions in organ functional reserves and variations in hormonal control mechanisms [ 11 ]. Patients with cerebrovascular diseases exhibited a significant inverse interaction (P < 0.001). The mortality risk was not elevated in relation to the CAR values (HR = 0.88, P = 0.414). This observation may be attributed to the long-term and regular use of antiplatelet agents or statins among patients with cerebrovascular conditions. Antiplatelet medications can suppress the interaction between platelets and endothelial cells, enhance endothelial function, and reduce microvascular permeability, thereby mitigating pulmonary edema and inflammatory responses associated with ARDS. The evidence suggests that antiplatelet therapy is linked to a lower incidence and mortality rates of ARDS, as well as a reduced requirement for mechanical ventilation [ 12 ]. Notably, patients with diabetes and chronic complications demonstrated a markedly reduced risk of AKI, accompanied by a significant interaction effect. This potential protective effect may be associated with long-term glycemic control, the use of medications that reduce albumin loss, compensatory mechanisms preserving renal function, or the specific renal benefits of drugs such as sodium‒glucose cotransporter 2 (SGLT2) inhibitors [ 13 , 14 ]. However, given the limited sample size, further research is necessary to validate these observations. Our research represents the first effort to investigate the connection between the CAR and both mortality risk and the development of acute kidney injury among patients diagnosed with acute respiratory distress syndrome, utilizing publicly accessible datasets. The findings from this study could offer healthcare professionals additional insights for the early diagnosis of acute respiratory distress syndrome, evaluation of disease severity, and targeted interventions for patients with unfavorable prognoses. Nevertheless, certain limitations should be acknowledged. First, as a retrospective analysis, the data were sourced from a single-center database. Despite the MIMIC-IV database encompassing a substantial sample size, potential biases—such as selection bias or disease spectrum bias—may still exist, warranting multicenter studies to confirm the generalizability of the observed protective effects of cerebrovascular diseases. Second, the analysis did not account for several possible confounding variables, which might influence the association between CAR values and clinical outcomes. Finally, some subgroup analyses—such as those involving patients with diabetes and chronic complications or individuals with dementia—were based on small sample sizes, thereby limiting the reliability of the results. Further investigations with larger cohorts are necessary to validate these findings. Conclusion Our findings indicate that the CAR functions as an independent prognostic marker for 28-day mortality, overall mortality, and the development of AKI. The study also highlights variations in CAR across different patient subgroups, offering valuable insights for clinical decision-making and guiding future research directions. Further investigations are warranted to evaluate the practical utility of the CAR in predicting outcomes for patients with ARDS. Declarations Acknowledgments We express our sincere appreciation to the staff and patients involved in the development of the MIMIC-IV database, from which the research data for this study were sourced. Author contributions Zixuan Jiang contributed to the study design, data analysis, and manuscript writing. Yueyue Zhang contributed to the data extraction, data analysis, and manuscript revision. Ge Zhang: Writing – review & editing, supervision, formal analysis. Funding No specific funding was received for this article from public, commercial, or nonprofit organizations. Availability of data and materials The data are available from the corresponding author upon reasonable request. This research received approval from the appropriate review bodies at the Massachusetts Institute of Technology and Beth Israel Deaconess Medical Center and was conducted in full compliance with applicable laws, regulations, and institutional guidelines. As the database is publicly available and all patient information has been deidentified, the ethics review committee or institutional review board has exempted the need for written informed consent from participants, their legal representatives, or immediate family members, thereby requiring no further ethical clearance. Consent for publication Not applicable. Competing interests We affirm that all designated authors have reviewed and endorsed the manuscript and that there are no individuals who met the authorship criteria but were omitted from the author list. References Huppert, LA, Matthay, MA, Ware, LB. Pathogenesis of Acute Respiratory Distress Syndrome. SEMIN RESP CRIT CARE. 2019; 40 (1): 31–39. Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med 2005;353(16):1685–1693 Sadana, D, Kaur, S, Sankaramangalam, K, et al. Mortality associated with acute respiratory distress syndrome, 2009–2019: a systematic review and meta-analysis. CRIT CARE RESUSC. 2022; 24 (4): 341–351. Ávila, M, Mora Sánchez, MG, Bernal Amador, AS, et al. The Metabolism of Creatinine and Its Usefulness to Evaluate Kidney Function and Body Composition in Clinical Practice. Biomolecules. 2025; 15 (1): Wang, HX, Huang, XH, Ma, LQ, et al. Association between lactate-to-albumin ratio and short-time mortality in patients with acute respiratory distress syndrome. J CLIN ANESTH. 2024; 99 111632. Wang, J, Li, H, Luo, H, et al. Association between serum creatinine to albumin ratio and short- and long-term all-cause mortality in patients with acute pancreatitis admitted to the intensive care unit: a retrospective analysis based on the MIMIC-IV database. Front Immunol. 2024; 15 1373371. You, T, Zhou, YR, Liu, XC, et al. Risk Factors and Clinical Characteristics of Neonatal Acute Respiratory Distress Syndrome Caused by Early Onset Sepsis. Front Pediatr. 2022; 10 847827. Azoulay, E, Lemiale, V, Mourvillier, B, et al. Management and outcomes of acute respiratory distress syndrome patients with and without comorbid conditions. INTENS CARE MED. 2018; 44 (7): 1050–1060. You, T, Zhou, YR, Liu, XC, et al. Risk Factors and Clinical Characteristics of Neonatal Acute Respiratory Distress Syndrome Caused by Early Onset Sepsis. Front Pediatr. 2022; 10 847827. Ronco, C, Bellomo, R, Kellum, JA. Acute kidney injury. LANCET. 2019; 394 (10212): 1949–1964. Nilsson, BO. Modulation of the inflammatory response by estrogens with focus on the endothelium and its interactions with leukocytes. INFLAMM RES. 2007; 56 (7): 269–73. Chen, CM, Lu, HC, Tung, YT, et al. Antiplatelet Therapy for Acute Respiratory Distress Syndrome. Biomedicines. 2020; 8 (7): Fadini, GP, Longato, E, Morieri, ML, et al. Comparative renal outcomes of matched cohorts of patients with type 2 diabetes receiving SGLT2 inhibitors or GLP-1 receptor agonists under routine care. DIABETOLOGIA. 2024; 67 (11): 2585–2597. Bae, JH. SGLT2 Inhibitors and GLP-1 Receptor Agonists in Diabetic Kidney Disease: Evolving Evidence and Clinical Application. Diabetes Metab J. 2025; 49 (3): 386–402. Tables Tables 1 to 3 are available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table1.pdf Table2.pdf Table3.pdf 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-7144283","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":495560851,"identity":"8d6c440f-7b48-4ef3-92c0-e0f33de2a1d7","order_by":0,"name":"Zixuan Jiang","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Zixuan","middleName":"","lastName":"Jiang","suffix":""},{"id":495560852,"identity":"7fa0a4b0-7c1b-457c-9b49-5fdd76c0615c","order_by":1,"name":"Yueyue Zhang","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Yueyue","middleName":"","lastName":"Zhang","suffix":""},{"id":495560853,"identity":"a905bed7-1298-48dc-b142-5e2f56951363","order_by":2,"name":"Ge Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5ElEQVRIiWNgGAWjYBACefbGBoOECgk5BoYDQC4bEVoMew4fKPhwxsKYeC0MN9ISPs5sq0hsAPOI0cLYkGO4mbdNIn1+4xkDhg9lhxn4Zzfg18LOcMbYmOecRG5jwxkDxhnnDjNI3DlAwJbGHjNjnjKJ3GaGMwbMvG2HGQwkEgi47DCP+W8eNol0NpCWv0RpOcaWYDijTSKBB6SFkRgthj3MBww+nJEwnMFwrOBgz7l0HokbBLTIyz8ERWWdvPyMwxsf/CizluOfQchhcCBxAByZPMSqBwL+BhIUj4JRMApGwYgCAHMfRb9fd0joAAAAAElFTkSuQmCC","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Ge","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-07-17 02:53:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7144283/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7144283/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88472677,"identity":"25fcb707-99b1-41b5-8fec-2d0c65af82fe","added_by":"auto","created_at":"2025-08-06 19:40:26","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":183670,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart for participants from the MIMIC-IV (v 3.0).\u003c/p\u003e","description":"","filename":"11.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/67726e5152b4826ab6fd8f00.jpg"},{"id":88472678,"identity":"21b50d99-6178-45de-b68f-fd2d04f23136","added_by":"auto","created_at":"2025-08-06 19:40:26","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":295644,"visible":true,"origin":"","legend":"\u003cp\u003eKaplan‒Meier survival analysis curves for 28-day mortality (A), all-cause mortality (B), and acute kidney injury (C)\u003c/p\u003e","description":"","filename":"12.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/9a38b849b6d55a042cc54aaa.jpg"},{"id":88473234,"identity":"389ad291-db52-47d2-8d91-b634ab10efda","added_by":"auto","created_at":"2025-08-06 19:56:26","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":295119,"visible":true,"origin":"","legend":"\u003cp\u003eROC curves for the prediction of 28-day mortality (A), all-cause mortality (B), and acute kidney injury (C) inpatients with ARDS.\u003c/p\u003e","description":"","filename":"14.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/b12aa40f177521e38825cea8.jpg"},{"id":88472682,"identity":"d18018b3-9b6c-4e61-931d-c228d8c54dfc","added_by":"auto","created_at":"2025-08-06 19:40:26","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":81114,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/4382e7923ffe664e78f6fcf6.jpg"},{"id":88473318,"identity":"bfa1cda0-38db-4fb0-8aef-9c2ba8a21245","added_by":"auto","created_at":"2025-08-06 20:04:26","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":559500,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot of the stratified analysis of the CAR and 28-day all-cause mortality (A), all-cause mortality (B), and acute kidney injury (C) rates.\u003c/p\u003e","description":"","filename":"17.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/53b296b06473d2231c85b8a4.jpg"},{"id":93713915,"identity":"982eb6a5-a8e3-4fbd-ad7e-c2eb3b1c8bc5","added_by":"auto","created_at":"2025-10-16 18:53:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2005434,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/afd45108-6af2-410d-985f-6d7495c37a39.pdf"},{"id":88473231,"identity":"3e3b0075-a3af-4bf3-8d9d-d38d32ea4331","added_by":"auto","created_at":"2025-08-06 19:56:26","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":91483,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/2464bb00c8c006de910817f3.pdf"},{"id":88472759,"identity":"ca151460-f270-4e19-8239-0a263bc21379","added_by":"auto","created_at":"2025-08-06 19:48:26","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":96844,"visible":true,"origin":"","legend":"","description":"","filename":"Table2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/6725b297edb49f79dc8e01a8.pdf"},{"id":88472762,"identity":"61e619eb-cf34-4281-89a5-2f0d28d85722","added_by":"auto","created_at":"2025-08-06 19:48:26","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":84164,"visible":true,"origin":"","legend":"","description":"","filename":"Table3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7144283/v1/0ca47caaae18c1accf143a4a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Relationship between the creatinine/albumin ratio and the ICU mortality rate of patients with acute respiratory distress syndrome: A retrospective cohort study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAcute respiratory distress syndrome is a form of acute respiratory failure resulting from noncardiogenic pulmonary edema [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. According to numerous studies, the mortality rate associated with ARDS typically ranges between 25% and 40% [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Patients with ARDS frequently exhibit renal dysfunction and systemic inflammatory responses. Creatinine is widely recognized as a key biomarker for assessing kidney function. However, variations in creatinine levels do not always accurately reflect true alterations in renal function, as its concentration is affected by multiple physiological factors, such as age, sex, muscle mass, and dietary habits [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Albumin, one of the principal plasma proteins, not only serves as an indicator of nutritional status but also plays a crucial role in inflammatory responses, vascular permeability, and fluid homeostasis. Accumulating evidence indicates that the serum ALB concentration is significantly associated with the risk of developing ARDS[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Nevertheless, albumin concentrations can be influenced by a range of factors, including nutritional state, systemic inflammation, hepatic function, and renal function. Therefore, relying solely on the serum ALB concentration as a prognostic marker may fail to fully capture the complexity of a patient\u0026rsquo;s clinical condition and long-term outcomes. The CAR, which integrates both creatinine and albumin levels, has been proposed as a novel biomarker for predicting disease prognosis. By combining these two clinically relevant parameters, the CAR enables a more comprehensive evaluation of a patient's overall health status and disease severity. Prior studies have demonstrated that the CAR is an independent predictor of all-cause mortality among ICU patients with acute pancreatitis [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. However, no published research has yet examined the utility of the CAR in predicting outcomes for patients with ARDS. This study aimed to investigate the prognostic value of the CAR in patients with ARDS by retrospectively analyzing data obtained from the Medical Information Mart for Intensive Care (MIMIC-IV, v2.0) database in the United States.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eData source\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective cohort study utilized data from the Medical Information Mart for Intensive Care IV (MIMIC-IV v3.0), a publicly accessible critical care database containing structured clinical records of 192,500 unique patients (455,000 hospital admissions) at Beth Israel Deaconess Medical Center in Boston, Massachusetts, USA, between 2008 and 2019. The database offers detailed documentation encompassing demographic characteristics, laboratory test results, medication administration records, continuous longitudinal vital signs, procedural interventions, disease diagnoses (coded according to the International Classification of Diseases [ICD] system), infusion parameters of therapeutic agents, and postdischarge survival follow-up data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePopulation Enrollment \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study utilized data from the latest version of the MIMIC-IV v3.0, which encompasses clinical records collected between 2012 and 2019. Patients diagnosed with ARDS were identified on the basis of standardized International Classification of Diseases coding criteria (ICD-9-CM: 518.51, 518.81; ICD-10-CM: J80, J96.0, J96.00\u0026ndash;J96.02, R06.03), resulting in a total of 30,509 identified cases. The inclusion criteria were as follows: age between 18 and 85 years, first hospital admission with an initial intensive care unit (ICU) stay, and ICU length of stay \u0026ge;24 hours (n=13,977). Following the exclusion of patients who lacked baseline serum creatinine (Cr) or albumin (Alb) measurements within 24 hours of ICU admission, a final cohort of 1,233 subjects was included in the analysis. These patients were stratified into four groups (Q1\u0026ndash;Q4) according to quartile thresholds of the creatinine-to-albumin ratio. A flowchart detailing the selection process of the study population is illustrated in Figure 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Extraction Strategy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData extraction was conducted via Structured Query Language (SQL) through Navicat Premium (Version 16.1.15). The variables included in the analysis were systematically categorized into six domains: demographics, comorbidities, vital signs, laboratory indicators (initial measurements obtained at hospital admission prior to any therapeutic interventions), healthcare utilization and outcomes, and severity scores. Variables with more than 30% missing data were excluded from the analysis. For variables with \u0026le;30% missing values, multiple imputation was carried out via a random forest algorithm. Table 1 presents a comprehensive list of the extracted variables.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCalculation of\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ethe CAR\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThe\u0026nbsp;\u003c/strong\u003eCAR was calculated as [(admission\u0026nbsp;creatinine (mg/dl))/(albumin (mg/dl))], and admission creatine and albumin were obtained directly from MIMIC IV.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eContinuous variables were first evaluated for normality via the Shapiro‒Wilk test. Nonnormally distributed data were analyzed via the Wilcoxon rank-sum test and summarized as medians (interquartile ranges, IQRs). Categorical variables were compared via either the chi-square test or Fisher\u0026rsquo;s exact test, as appropriate, and are presented as counts (percentages). To investigate the associations between the CAR index, analyzed as both continuous and quartile-based variables, and 28-day mortality as well as AKI, four progressively adjusted logistic regression models were developed. Model 1 was unadjusted; Model 2 was adjusted for demographic characteristics (age, sex); Model 3 further incorporated metabolic parameters (albumin, anion gap, bicarbonate, blood urea nitrogen, chloride, creatinine, glucose) and erythrocyte indices (mean corpuscular volume [MCV], mean corpuscular hemoglobin [MCH], mean corpuscular hemoglobin concentration [MCHC]); Model 4 additionally controlled for comorbid conditions (myocardial infarction, peripheral vascular disease, cerebrovascular disease, chronic pulmonary disease) and the Charlson Comorbidity Index (CCI), with the lowest CAR quartile (Q1) serving as the reference group. Restricted cubic splines (RCSs, 3 knots) were applied to assess potential dose‒response relationships between continuous CAR levels and clinical outcomes, whereas a recursive partitioning algorithm was used to detect inflection points within nonlinear associations (likelihood ratio test P \u0026lt; 0.05). Stratified analyses were performed according to sex, age, diabetes status (with or without complications), sepsis, peripheral vascular disease, myocardial infarction, chronic pulmonary disease, dementia, and cerebrovascular disease to evaluate heterogeneity via the Breslow‒Day test. All the statistical analyses were conducted via SPSS (v22.0) and R software (v4.3.2), with a two-tailed P value of less than 0.05 considered statistically significant.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eBaseline demographic and clinical characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis baseline table presents a comparative analysis of clinical characteristics among 1,233 patients stratified by\u0026nbsp;creatinine‒albumin ratio (CAR) quartiles (Q1\u0026ndash;Q4). Overall, the data reveal a statistically significant trend toward worsening clinical profiles with increasing CAR values (Q1: 0.176 to Q4: 0.942). Patients in the highest CAR quartile (Q4)\u0026nbsp;represented the highest proportion of males (63.9%), the longest median hospital length of stay (11.65 days), and the longest ICU duration (5.87 days). This group also presented elevated markers of physiological instability, including increased lactate levels (3.054 mmol/L), INR (1.907), and prothrombin time (PT, 21.21 s), as well as more severe inflammatory responses (albumin: 2.79 g/dL, WBC: 16.76\u0026times;10\u0026sup3;/\u0026mu;L) and greater organ dysfunction (SOFA score: 53.5, serum creatinine: 2.53 mg/dL). Comorbidity analysis indicated that Q4 had the highest Charlson Comorbidity Index (5.27) and the highest incidence of acute kidney injury (50.95%). The 28-day mortality rate in Q4 was 65.4%, significantly exceeding that\u0026nbsp;in Q1 (9.8%) (p \u0026lt; 0.001). Notably, Q4 demonstrated a greater prevalence of sepsis (88.2%) and a greater need for mechanical ventilation (77.6%). Compensatory changes were observed in the platelet count (155\u0026times;10\u0026sup3;/\u0026mu;L) and\u0026nbsp;hemoglobin-related parameters (MCHC: 32.88 g/dL). No significant intergroup differences were found for body temperature, Glasgow Coma Scale (GCS) score, or chloride level (p \u0026gt; 0.05). The detailed results are presented in\u003cstrong\u003e\u0026nbsp;Table 2\u003c/strong\u003e.\u003c/p\u003e\n\u003cdiv\u003e\u003cbr\u003e\u003c/div\u003e\n\u003cp\u003eLOS_hospital,\u0026nbsp;length of hospital stay; LOS_icu, length of ICU stay; total CO2, total carbon dioxide emissions; INR,integer normalized ratio; PT, prothrombin time; PTT, partial thromboplastin time; GCS, Glasgow Coma Scale; HR, heart rate; SAEPS Ⅱ, simplified acute physiology score Ⅱ; OASIS, organ dysfunction assessment system; Temp, temperature; MechVent, mechanical ventilation; SIRS, systemic inflammatory response syndrome; AIDS, acquired immunodeficiency syndrome; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; RBC, red blood cell count; RDW, red blood cell distribution width; WBC, white blood cell count; AKI, acute kidney injury; CAR, creatinine‒albumin ratio;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKaplan‒Meier\u0026nbsp;curve analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eKaplan‒Meier analysis revealed significant differences in 28-day mortality, overall survival, and AKI incidence across CAR quartiles (overall log-rank P \u0026lt; 0.001). Compared with the lowest CAR quartile (Q1), the highest quartile (Q4) presented significantly lower cumulative survival rates for both 28-day mortality and overall mortality, as well as a greater cumulative incidence of AKI. The results are presented in Figure\u0026ensp;2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThe predictive efficacy of\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ethe CAR for clinical outcomes and the results of multivariate regression analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe evaluated the predictive performance of both continuous CAR values and CAR quartiles for three key clinical outcomes: 28-day mortality, all-cause mortality, and AKI. As illustrated in Figure 3, the AUC analysis demonstrated that the CAR exhibited robust predictive ability across all evaluated endpoints. For 28-day mortality, the AUC values were 0.729 (95% CI: 0.640\u0026ndash;0.824) for the continuous CAR quartile and 0.732 (95% CI: 0.638\u0026ndash;0.813) for the CAR quartile. Similarly, for all-cause mortality, the AUC values were 0.716 (95% CI: 0.616\u0026ndash;0.805) and 0.719 (95% CI: 0.628\u0026ndash;0.800) for the continuous and quartile-based CAR measurements, respectively. Notably, the CAR demonstrated particularly strong predictive accuracy for AKI, with AUC values of 0.785 (95% CI: 0.709\u0026ndash;0.853) for the continuous CAR and 0.794 (95% CI: 0.724\u0026ndash;0.858) for the CAR quartiles. Collectively, these findings indicate that the CAR has substantial predictive value across multiple critical clinical outcomes.\u003c/p\u003e\n\u003cp\u003eThis study utilized multistage-adjusted logistic regression models to investigate the associations between the predictor variable and three clinical outcomes, which were differentially influenced by confounding factors. As presented in Table 3, in the unadjusted model, each one-unit increase in the continuous predictor variable was significantly associated with increased risks of 28-day mortality (OR = 16.43, 95% CI: 10.09\u0026ndash;27.85), all-cause mortality (OR = 11.75, 95% CI: 7.536\u0026ndash;18.97), and AKI (OR = 3.938, 95% CI: 2.830\u0026ndash;5.582). All p values were \u0026lt; 0.001. However, after adjusting for laboratory parameters (e.g., albumin, creatinine) and comorbidities, the direction of the association with mortality outcomes was reversed (Model 4: 28-day mortality OR = 0.117, p = 0.002; all-cause mortality OR = 0.117, p = 0.002), whereas a significant inverse association with AKI remained (Model 4: OR = 0.087, p \u0026lt; 0.001). Quartile-based analysis revealed a clear dose‒response relationship, with the highest quartile (Q4) showing the greatest risk in fully adjusted models: 28-day mortality (OR = 2.506, 95% CI: 1.201\u0026ndash;5.245, p = 0.014), all-cause mortality (OR = 2.506, same CI range), and AKI (OR = 13.83, 95% CI: 6.561\u0026ndash;30.21, p \u0026lt; 0.001). Notably, AKI maintained statistical significance across the Q2 and Q3 groups in all the models (Model 4: Q2 OR = 2.713, p = 0.002; Q3 OR = 5.215, p \u0026lt; 0.001), whereas the associations between mortality and the Q2\u0026ndash;Q3 quartiles became nonsignificant after adjustment. The final model (Model 4) indicated that laboratory markers and the Charlson Comorbidity Index explained 76.3\u0026ndash;93.8% of the original risk variance in mortality outcomes, compared with only 39.1% for AKI. This residual association suggests that the predictor variable may exert a direct pathophysiological effect.\u003c/p\u003e\n\u003ch3\u003eDetection\u0026nbsp;of nonlinear relationships\u003c/h3\u003e\n\u003cp\u003eThis series of RCS plots demonstrated a significant U-shaped nonlinear association between the CAR and the risk of 28-day mortality, all-cause mortality, and AKI in critically ill patients (Fig. 4). Specifically, a specific CAR range is associated with the nadir risk for these adverse outcomes. Significant risk elevation is observed when CAR values fall below or rise above this optimal range. This robust U-shaped association was highly statistically significant in both the minimally adjusted model (Model 1) and the model incorporating additional variables (Model 2) for all outcomes (P-nonlinear \u0026lt; 0.0001 for all models/outcomes). Critically, this nonlinear U-shaped relationship remained statistically significant even in the most rigorously adjusted models (Model 3 \u0026amp; Model 4), despite potential alterations in the precise curve morphology or increased estimation uncertainty (wider confidence intervals) with progressive covariate inclusion (Model 3 P-nonlinear = 0.0301, Model 4 P-nonlinear = 0.0130, as applicable across outcomes).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSubgroup analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further investigate the\u0026nbsp;associations between the CAR and 28-day mortality, long-term all-cause mortality, and acute kidney injury, we performed stratified analyses according to sex, age, diabetes status (with or without complications), sepsis, peripheral vascular disease, myocardial infarction, chronic lung disease, dementia, and cerebrovascular disease. The results are presented in Figure 5.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e28-day mortality\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study performed a subgroup analysis of 28-day mortality. The overall results revealed a significantly elevated mortality risk across the entire population (HR = 1.87, 95% CI: 1.66\u0026ndash;2.10, P \u0026lt; 0.001). With respect to demographic characteristics, patients older than 65 years (HR = 2.06) and males (HR = 2.10)\u0026nbsp;presented greater mortality risks than did those aged \u0026le;65 years (HR = 1.75) and females (HR = 1.67), although the interaction effects for age and\u0026nbsp;sex were not statistically significant (P = 0.088 and 0.072, respectively). In the diabetes subgroup, diabetic patients without chronic complications presented a lower risk than nondiabetic patients did (HR = 1.50 vs. 2.00; interaction P = 0.061), whereas the subgroup with chronic complications (HR = 1.21, P = 0.586) did not present a significant difference, likely due to the small sample size (n = 69). Notably, patients with cerebrovascular disease exhibited a significant inverse interaction effect (P \u0026lt; 0.001), with no increased mortality risk (HR = 0.88, P = 0.414), in contrast to those without cerebrovascular disease, who had an HR of 2.13 (P \u0026lt; 0.001). Among the other disease subgroups, patients with a history of myocardial infarction presented the highest risk (HR = 2.25), and both peripheral vascular disease (HR = 2.15) and chronic pulmonary disease (HR = 1.61) were significantly associated with increased mortality.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAll-cause mortality\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study conducted subgroup analyses of all-cause mortality, which revealed a significantly elevated overall risk (HR = 1.82, 95% CI: 1.62\u0026ndash;2.04, P \u0026lt; 0.001). Subgroup analyses\u0026nbsp;revealed statistically significant interaction effects for age (Pinteraction = 0.032), sex (Pinteraction = 0.05), and cerebrovascular disease (Pinteraction \u0026lt; 0.001). Specifically, patients older than 65 years (HR = 2.06 vs. \u0026le;65 years, HR = 1.68) and males (HR = 2.07 vs. females, HR = 1.62) presented increased mortality risks, whereas patients with cerebrovascular disease presented no increased mortality risk (HR = 0.93, P = 0.628), suggesting potential protective associations. The key high-risk subgroups included those with a history of myocardial infarction (HR = 2.25, P \u0026lt; 0.001) and peripheral vascular disease (HR = 2.15, P \u0026lt; 0.001). Notably, although subgroups of diabetes (with or without chronic comorbidities), sepsis, and chronic pulmonary disease all presented elevated risks (HR \u0026gt; 1.5, P \u0026lt; 0.01), their interaction effects were not statistically significant (Pinteraction \u0026gt; 0.05).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcute kidney injury\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis subgroup analysis of patients with AKI revealed a significantly elevated overall risk (HR = 1.94, 95% CI: 1.69\u0026ndash;2.22, P \u0026lt; 0.001). Demographic analysis revealed significantly greater AKI risks in both age groups (\u0026le;65 years: HR = 1.93; \u0026gt;65 years: HR = 2.05) and sexes (male: HR = 2.13; female: HR = 1.72) (all P \u0026lt; 0.001), although the interaction effects of age and sex were not statistically significant (P = 0.577 and 0.153, respectively). Among the comorbidities, diabetic patients with chronic complications presented a significantly lower AKI risk (HR = 0.79, P = 0.303) and a notable interaction effect (P \u0026lt; 0.001), whereas those without complications presented a greater risk (HR = 1.51, P = 0.006). Nonseptic patients demonstrated substantially greater AKI risk than did septic patients (HR = 2.80 vs. 1.75, Pinteraction = 0.002), with myocardial infarction (HR = 2.28) and chronic pulmonary disease (HR = 2.27) identified as additional high-risk factors. The dementia subgroup (n = 11) yielded unreliable results (HR = 40192.97, P = 1.000) because of an insufficient sample size, whereas peripheral vascular disease and cerebrovascular disease did not show significant between-group differences.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study investigated the connection between the CAR and clinical outcomes in patients diagnosed with acute ARDS. The results revealed a notable U-shaped nonlinear correlation between CAR levels and the likelihood of 28-day mortality, overall mortality, and the occurrence of acute kidney injury. Within a certain range of CAR values, the risk of these negative outcomes was lowest; however, when CAR values fell outside this optimal range, the risk of experiencing these three adverse events increased significantly. The AUC analysis demonstrated strong predictive accuracy for all studied outcomes, especially for acute kidney injury. These observations imply that the CAR could serve as a valuable biomarker for evaluating the prognosis of ARDS patients. By integrating measurements of both creatinine and albumin, the CAR offers a more holistic evaluation of a patient's general health condition and illness severity, thereby enhancing the accuracy of outcome predictions for ARDS patients. In our research, initial patient data displayed a distinct dose‒response pattern relative to CAR levels. As the CAR quartiles increased, patients tended to present with increasingly severe inflammation, coagulation abnormalities, tissue hypoxia, and multiorgan dysfunction. These findings support the notion that the CAR captures both renal impairment and systemic inflammatory or nutritional disturbances, making it a well-suited marker for assessing the overall burden of critical illness.\u003c/p\u003e\u003cp\u003eIn the Kaplan‒Meier analysis, notable differences were observed in both 28-day mortality and overall all-cause mortality across patient groups categorized by different CAR quartiles (log-rank test, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). To further explore this relationship, a multivariable logistic regression analysis was conducted, revealing that confounding variables significantly influenced the association between CAR levels and mortality outcomes. Initially, in the unadjusted model, each unit increase in CAR was linked to a 16-fold higher risk of 28-day mortality (OR\u0026thinsp;=\u0026thinsp;16.43). However, after adjusting for key clinical factors\u0026mdash;including laboratory markers such as albumin and creatinine, as well as the Charlson Comorbidity Index\u0026mdash;the direction of this association changed dramatically (OR\u0026thinsp;=\u0026thinsp;0.117 in Model 4). This shift suggests that comorbidities and laboratory parameters substantially modify the impact of the CAR on mortality risk, potentially by interacting with the CAR to influence patient survival.\u003c/p\u003e\u003cp\u003eFor example, the level of albumin, which serves as a marker for both nutritional status and systemic inflammation, may reflect worsened health conditions when it is decreased, thereby contributing to increased mortality risk [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Moreover, concurrent conditions such as cardiovascular diseases and diabetes can further deteriorate a patient's clinical condition and increase the likelihood of mortality [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. In contrast to mortality outcomes, the association between the CAR and acute kidney injury remained robust even after full adjustment for confounding variables. The area under the curve (AUC) values were 0.785 (95% confidence interval: 0.709\u0026ndash;0.853) for continuous CAR measurements and 0.794 (95% confidence interval: 0.724\u0026ndash;0.858) for quartile-based analysis. These results indicate that the CAR is a highly effective biomarker for identifying patients at risk of developing acute kidney injury, with superior predictive performance compared with several traditional indicators. For instance, the level of serum creatinine, a widely used marker of renal function, is associated with a delayed response. Prior studies have shown that creatinine levels typically do not increase until 48 to 72 hours after renal injury has occurred, at which time significant kidney function may be lost [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Changes in the CAR could act as early signals of the development of renal impairment. In a stepwise multivariable logistic regression analysis, the initial unadjusted continuous CAR assessment revealed a marked increase in the likelihood of AKI with each unit increase (OR\u0026thinsp;=\u0026thinsp;3.938, 95% CI: 2.830\u0026ndash;5.582, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Following complete adjustment for confounding factors, the CAR still exhibited a significant inverse relationship with AKI (Model 4 OR\u0026thinsp;=\u0026thinsp;0.087, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), indicating its considerable independent predictive capacity for the occurrence of AKI. Importantly, in Model 4, laboratory variables and the Charlson Comorbidity Index together accounted for 76.3\u0026ndash;93.8% of the initial variance in mortality risk, whereas AKI alone contributed merely 39.1%. This residual association suggests that the CAR may have a direct pathophysiological impact on kidney damage, separate from the patient's overall clinical status.\u003c/p\u003e\u003cp\u003eAdditional subgroup analyses revealed higher mortality risks among older individuals (over 65 years of age) and male patients, which might be linked to age-associated reductions in organ functional reserves and variations in hormonal control mechanisms [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Patients with cerebrovascular diseases exhibited a significant inverse interaction (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The mortality risk was not elevated in relation to the CAR values (HR\u0026thinsp;=\u0026thinsp;0.88, P\u0026thinsp;=\u0026thinsp;0.414). This observation may be attributed to the long-term and regular use of antiplatelet agents or statins among patients with cerebrovascular conditions. Antiplatelet medications can suppress the interaction between platelets and endothelial cells, enhance endothelial function, and reduce microvascular permeability, thereby mitigating pulmonary edema and inflammatory responses associated with ARDS. The evidence suggests that antiplatelet therapy is linked to a lower incidence and mortality rates of ARDS, as well as a reduced requirement for mechanical ventilation [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Notably, patients with diabetes and chronic complications demonstrated a markedly reduced risk of AKI, accompanied by a significant interaction effect. This potential protective effect may be associated with long-term glycemic control, the use of medications that reduce albumin loss, compensatory mechanisms preserving renal function, or the specific renal benefits of drugs such as sodium‒glucose cotransporter 2 (SGLT2) inhibitors [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. However, given the limited sample size, further research is necessary to validate these observations.\u003c/p\u003e\u003cp\u003eOur research represents the first effort to investigate the connection between the CAR and both mortality risk and the development of acute kidney injury among patients diagnosed with acute respiratory distress syndrome, utilizing publicly accessible datasets. The findings from this study could offer healthcare professionals additional insights for the early diagnosis of acute respiratory distress syndrome, evaluation of disease severity, and targeted interventions for patients with unfavorable prognoses. Nevertheless, certain limitations should be acknowledged. First, as a retrospective analysis, the data were sourced from a single-center database. Despite the MIMIC-IV database encompassing a substantial sample size, potential biases\u0026mdash;such as selection bias or disease spectrum bias\u0026mdash;may still exist, warranting multicenter studies to confirm the generalizability of the observed protective effects of cerebrovascular diseases. Second, the analysis did not account for several possible confounding variables, which might influence the association between CAR values and clinical outcomes. Finally, some subgroup analyses\u0026mdash;such as those involving patients with diabetes and chronic complications or individuals with dementia\u0026mdash;were based on small sample sizes, thereby limiting the reliability of the results. Further investigations with larger cohorts are necessary to validate these findings.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOur findings indicate that the CAR functions as an independent prognostic marker for 28-day mortality, overall mortality, and the development of AKI. The study also highlights variations in CAR across different patient subgroups, offering valuable insights for clinical decision-making and guiding future research directions. Further investigations are warranted to evaluate the practical utility of the CAR in predicting outcomes for patients with ARDS.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe express our sincere appreciation to the staff and patients involved in the development of the MIMIC-IV database, from which the research data for this study\u0026nbsp;were sourced.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZixuan Jiang contributed to\u0026nbsp;the study design, data analysis, and manuscript writing. Yueyue Zhang contributed to the data extraction, data analysis, and manuscript revision. Ge Zhang: Writing \u0026ndash; review \u0026amp; editing, supervision, formal analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo specific funding was received for this article from public, commercial, or nonprofit organizations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003eThis research received approval from the appropriate review bodies at the Massachusetts Institute of Technology and Beth Israel Deaconess Medical Center and was conducted in full compliance with applicable laws, regulations, and institutional guidelines. As the database is publicly available and all patient information has been deidentified, the ethics review committee or institutional review board has exempted the need for written informed consent from participants, their legal representatives, or immediate family members, thereby requiring no further ethical clearance.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe affirm that all designated authors have reviewed and endorsed the manuscript and\u0026nbsp;that there are no individuals who met the authorship criteria\u003c/p\u003e\n\u003cp\u003ebut were omitted from the author list.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eHuppert, LA, Matthay, MA, Ware, LB. Pathogenesis of Acute Respiratory Distress Syndrome. SEMIN RESP CRIT CARE. 2019; 40 (1): 31\u0026ndash;39.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med 2005;353(16):1685\u0026ndash;1693\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSadana, D, Kaur, S, Sankaramangalam, K, et al. Mortality associated with acute respiratory distress syndrome, 2009\u0026ndash;2019: a systematic review and meta-analysis. CRIT CARE RESUSC. 2022; 24 (4): 341\u0026ndash;351.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e\u0026Aacute;vila, M, Mora S\u0026aacute;nchez, MG, Bernal Amador, AS, et al. The Metabolism of Creatinine and Its Usefulness to Evaluate Kidney Function and Body Composition in Clinical Practice. Biomolecules. 2025; 15 (1):\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang, HX, Huang, XH, Ma, LQ, et al. Association between lactate-to-albumin ratio and short-time mortality in patients with acute respiratory distress syndrome. J CLIN ANESTH. 2024; 99 111632.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang, J, Li, H, Luo, H, et al. Association between serum creatinine to albumin ratio and short- and long-term all-cause mortality in patients with acute pancreatitis admitted to the intensive care unit: a retrospective analysis based on the MIMIC-IV database. Front Immunol. 2024; 15 1373371.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYou, T, Zhou, YR, Liu, XC, et al. Risk Factors and Clinical Characteristics of Neonatal Acute Respiratory Distress Syndrome Caused by Early Onset Sepsis. Front Pediatr. 2022; 10 847827.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAzoulay, E, Lemiale, V, Mourvillier, B, et al. Management and outcomes of acute respiratory distress syndrome patients with and without comorbid conditions. INTENS CARE MED. 2018; 44 (7): 1050\u0026ndash;1060.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYou, T, Zhou, YR, Liu, XC, et al. Risk Factors and Clinical Characteristics of Neonatal Acute Respiratory Distress Syndrome Caused by Early Onset Sepsis. Front Pediatr. 2022; 10 847827.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRonco, C, Bellomo, R, Kellum, JA. Acute kidney injury. LANCET. 2019; 394 (10212): 1949\u0026ndash;1964.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNilsson, BO. Modulation of the inflammatory response by estrogens with focus on the endothelium and its interactions with leukocytes. INFLAMM RES. 2007; 56 (7): 269\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen, CM, Lu, HC, Tung, YT, et al. Antiplatelet Therapy for Acute Respiratory Distress Syndrome. Biomedicines. 2020; 8 (7):\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFadini, GP, Longato, E, Morieri, ML, et al. Comparative renal outcomes of matched cohorts of patients with type 2 diabetes receiving SGLT2 inhibitors or GLP-1 receptor agonists under routine care. DIABETOLOGIA. 2024; 67 (11): 2585\u0026ndash;2597.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBae, JH. SGLT2 Inhibitors and GLP-1 Receptor Agonists in Diabetic Kidney Disease: Evolving Evidence and Clinical Application. Diabetes Metab J. 2025; 49 (3): 386\u0026ndash;402.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 to 3 are available in the Supplementary Files section.\u003c/p\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":"Creatinine/albumin ratio, acute respiratory distress syndrome, 28-day mortality rate, all-cause mortality rate, MIMIC-IV, serum biomarker","lastPublishedDoi":"10.21203/rs.3.rs-7144283/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7144283/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e: \u0026nbsp;Creatinine (Cr) and albumin (ALB) are widely recognized as predictive biomarkers of mortality across various disease conditions, including patients with acute respiratory distress syndrome (ARDS). Nevertheless, limited evidence is available regarding the associations between the creatinine-to-albumin ratio and clinical outcomes in ARDS patients. This study aimed to assess the prognostic value of the creatinine-to-albumin ratio in predicting 28-day mortality, all-cause mortality, and the incidence of acute kidney injury (AKI) among patients diagnosed with ARDS.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e: This retrospective cohort study utilized clinical data extracted from the database of Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts, USA, spanning the period from 2008--2019. The creatinine‒albumin ratio (CAR) was calculated on the basis of measurements taken within 24 hours of patient admission. Kaplan‒Meier (K‒M) analysis was employed to compare 28-day mortality, all-cause mortality, and the incidence of acute kidney injury across the four patient groups. A Cox proportional hazards regression model and RCS were used to assess the relationships between the CAR and the risks of 28-day mortality, long-term all-cause mortality, and acute kidney injury. The predictive performance of the CAR—including its sensitivity, specificity, and AUC—was evaluated via receiver operating characteristic (ROC) curve analysis for the aforementioned outcomes in patients with ARDS. Subgroup analyses were also conducted to further validate the robustness and reliability of our findings.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e: A total of 1,233 patients were enrolled in the study. K‒M analysis revealed statistically significant differences in 28-day mortality, overall all-cause mortality, and the incidence of AKI across CAR quartiles (log-rank P \u0026lt; 0.001). Patients with elevated CAR levels presented increased risks of both 28-day and all-cause mortality, as well as a higher cumulative incidence of AKI. After adjusting for potential confounding factors, the multivariate Cox proportional hazards regression model confirmed a statistically significant association between the CAR and each of the three clinical outcomes. Furthermore, RCS analysis demonstrated a significant U-shaped nonlinear relationship between the CAR and these outcomes. ROC curve analysis revealed that the AUC values for the continuous CAR in predicting 28-day mortality, all-cause mortality, and AKI in patients with ARDS were 0.729, 0.716, and 0.785, respectively. When analyzed by quartiles, the corresponding AUCs were slightly improved at 0.732, 0.719, and 0.794. Subgroup analyses indicated that the associations between the CAR and clinical outcomes were more pronounced among patients aged \u0026gt;65 years, males, and those with a history of myocardial infarction or peripheral vascular disease. Notably, in patients with cerebrovascular disease, the risk of mortality did not increase with increasing CAR (hazard ratio [HR] = 0.88, P = 0.414), and the risk of AKI was attenuated in patients with diabetes and chronic complications (HR = 0.79, P = 0.303).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion: \u003c/strong\u003eThe CAR can serve as an independent predictor of 28-day mortality, all-cause mortality, and acute kidney injury. This study also revealed differences in the CAR among different subgroups, which may provide promising prognostic biomarkers for risk stratification and clinical management of patients with ARDS.\u003c/p\u003e","manuscriptTitle":"Relationship between the creatinine/albumin ratio and the ICU mortality rate of patients with acute respiratory distress syndrome: A retrospective cohort study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-06 19:40:21","doi":"10.21203/rs.3.rs-7144283/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":"6cd95da2-78cb-417a-b71d-44724ef5181f","owner":[],"postedDate":"August 6th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-11-19T17:38:23+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-06 19:40:21","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7144283","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7144283","identity":"rs-7144283","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}