Acute pancreatitis (AP) can trigger systemic inflammation and organ dysfunction and carries substantial short‑term mortality risk, especially in severe cases. Early, accurate prognostic tools are important for triage and management. Red cell distribution width (RDW) reflects erythrocyte size heterogeneity and associates with inflammation, oxidative stress, and nutritional state; serum albumin is a nutritional and inflammatory marker. The red cell distribution width‑to‑albumin ratio (RAR) combines these measures and has emerged as a potential prognostic biomarker in acute conditions. The Sequential Organ Failure Assessment (SOFA) score quantifies multiorgan dysfunction and is commonly used to predict mortality in critically ill patients, including those with AP. This study tested the hypothesis that combining RAR with the SOFA score improves prediction of 28‑day all‑cause mortality in ICU patients with AP.
Data were obtained from the publicly available MIMIC‑IV (v3.1) database, which includes detailed ICU admissions at Beth Israel Deaconess Medical Center from 2008 to 2022. From 546,028 total admissions and 49,30 patients identified with AP by ICD‑9/10 codes, 1,280 AP patients admitted to the ICU were reviewed. After excluding patients younger than 18 years, repeat ICU admissions (only first admission retained), ICU stays under 24 hours, and those lacking RDW, albumin, or SOFA data, the final analytic cohort comprised 702 adult ICU patients with AP.
The study derived the RAR by dividing RDW by serum albumin (details of unit handling were reported in the source). The SOFA score was calculated according to its standard components to quantify organ dysfunction. The authors combined RAR with SOFA to create a single RAR‑modified SOFA score; specifics of the combination method (for example additive or weighted) were described in the source material. The combined indicator was then evaluated as a predictor of 28‑day all‑cause mortality.
The relationship between the RAR‑modified SOFA score and 28‑day mortality was assessed using univariate and multivariate Cox proportional hazards regression models to identify independent associations after adjustment for covariates. The optimal cutoff value for the combined score was established using the Youden index from Receiver Operating Characteristic (ROC) curve analysis. Discriminatory performance was compared among the RAR‑modified SOFA score, RAR alone, and SOFA alone using ROC curves. Kaplan‑Meier (K‑M) survival analysis compared mortality between groups defined by the Youden cutoff. Subgroup analyses examined the robustness of associations across different patient strata.
In both univariate and multivariate Cox regression analyses, the RAR‑modified SOFA score was identified as an independent risk factor for 28‑day all‑cause mortality among ICU patients with AP. The multivariate models adjusted for relevant clinical covariates as described in the source. The source reports that the combined indicator retained statistical significance after adjustment, indicating an independent association with short‑term mortality.
ROC curve analysis demonstrated that the RAR‑modified SOFA score had superior predictive ability for 28‑day mortality compared with either RAR or SOFA alone. Using the Youden index, the optimal cutoff value for the combined indicator was identified as 12.69, which was then used to stratify patients into higher‑ and lower‑risk groups for survival analyses.
Kaplan‑Meier survival curve analysis showed a significantly higher 28‑day mortality in the high‑value group (RAR‑modified SOFA score ≥ 12.69) than in the low‑value group (< 12.69). This stratification by the Youden‑derived threshold visibly separated survival curves, supporting the combined score’s clinical discrimination for short‑term outcome.
The authors performed subgroup analyses to test whether the association between the RAR‑modified SOFA score and 28‑day mortality persisted across different patient subsets. The study reports that the RAR‑modified SOFA score remained an independent predictor in these analyses, indicating robustness of the finding across examined subpopulations. Specific subgroup definitions and results are presented in the source article’s tables and figures.
Combining an inflammation/nutrition marker (RAR) with an organ dysfunction score (SOFA) produced a single, readily available prognostic indicator with improved accuracy for predicting 28‑day mortality in ICU patients with AP compared with either component alone. Because RDW, albumin, and SOFA components are commonly measured in ICU practice, the RAR‑modified SOFA score could be implemented without additional testing and may aid early risk stratification and clinical decision making.
This analysis is retrospective and used a single public ICU database (MIMIC‑IV v3.1). The source provides data availability details: the dataset is accessible via MIMIC‑IV (DOI provided in the article). The study acknowledges its retrospective design and the limits of registry data; any further external validation in other cohorts or prospective settings would be required to confirm generalizability. The source contains additional details on exclusions, covariates, and subgroup definitions.
In this MIMIC‑IV‑based retrospective cohort of 702 ICU patients with acute pancreatitis, the RAR‑modified SOFA score was independently associated with 28‑day all‑cause mortality and demonstrated better predictive performance than RAR or SOFA alone. A combined‑score cutoff of 12.69 identified a high‑risk group with significantly higher short‑term mortality. The authors propose that the RAR‑modified SOFA score is a simple, accessible tool that may enhance early prognosis assessment in AP, though prospective validation is warranted.