Abstract
Background
The red cell distribution width-to-albumin ratio (RAR) is an emerging prognostic indicator. However, its correlation with clinical endpoints in individuals with sepsis-associated acute respiratory distress syndrome (SA-ARDS) remains unclear. This study examined the relationship between RAR and all-cause mortality (ACM) in the intensive care unit (ICU) patients with SA-ARDS using the MIMIC-IV database. The goal was to identify high-risk individuals early on and improve prognostic assessment.
Methods
Analyses were conducted using data from the MIMIC-IV (V3.1) database. Adult patients who met Sepsis-3 criteria and the Berlin definition of ARDS were included. The outcomes were 30-, 90-, 180-, and 365-day mortality following ICU admission. Kaplan-Meier (KM) curves were utilized to estimate survival at corresponding timepoints. Cox regression was utilized to evaluate the association between RAR and mortality risk, including subgroup analyses. Restricted cubic spline (RCS) regression was adopted to explore the nonlinear correlation of RAR with ACM and to identify a possible cutoff value. Receiver operating characteristic (ROC) curves were applied for quantifying RAR’s predictive ability. Propensity score matching (PSM) was used to validate the findings.
Results
The current study included 6,042 patients. The first, median, and third quartiles of RAR were 4.14, 5.07, and 6.32, respectively. KM analysis revealed a significantly higher mortality risk in patients with high RAR (P < 0.001). Multivariable Cox regression analysis indicated an independent positive correlation of RAR with ACM. RCS curves revealed a nonlinear relationship (inflection point at 5.04). The areas under the curve (AUCs) for RAR predicting 30-, 90-, 180-, and 365-day mortality were 0.634, 0.648, 0.652, and 0.649, respectively. Subgroup analyses supported the association of RAR. PSM confirmed the robustness of these results.
Conclusions
RAR was found to be independently associated with an increased 30-, 90-, 180-, and 365-day ACM in ICU patients with SA-ARDS. RAR may have potential value as an early indicator for identifying high‑risk ICU populations.
Clinical trial number
Not applicable.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12879-026-12827-4.
Keywords: Sepsis, Acute respiratory failure, Red cell distribution width, Albumin, MIMIC-IV
Background
Although the mortality for sepsis patients has decreased to 25%-30% due to an improved understanding of and ability to manage sepsis [1], in-hospital mortality (IHM) remains high for patients with sepsis-associated acute respiratory distress syndrome (SA-ARDS), reaching 43.2% [2]. Early identification and intervention in this population are critical. It has been shown to significantly improve survival [3]. This underscores the importance of recognizing high-risk SA-ARDS patients early in the course of disease.
Sepsis and ARDS are pathophysiologically linked and etiologically complex. Sepsis can directly induce lung injury via severe pulmonary infection or indirectly contribute to tissue damage through systemic inflammation, ultimately progressing to ARDS [4]. Both conditions are considered to be primarily caused by dysregulated inflammation, leading to endothelial damage, oxidative stress, immune dysfunction, and mitochondrial impairment [5–7]. These processes are frequently accompanied by metabolic derangements, including increased catabolism and nitrogen loss, which can worsen organ dysfunction [8–10]. Elevated red cell distribution width (RDW) has been associated with inflammation and oxidative stress and has been linked to patient outcomes in both sepsis and ARDS [11, 12]. RDW is also correlated with malnutrition [13]. Albumin (Alb), a key liver-derived protein, plays a role in nutritional metabolism, immunity, antioxidant activity, and endothelial repair [14–17]. Alb maintains colloidal osmolality balance inside and outside the capillaries. Hypoalbuminemia is related to pulmonary edema and ARDS [18]. Combining RDW and Alb can comprehensively reflect a patient’s inflammatory state, endothelial integrity, and nutritional status. RDW-to-albumin ratio (RAR) has thus become a potential composite marker for evaluating disease severity and predicting outcomes [19].
While RAR has demonstrated utility in predicting outcomes in several critical illnesses, its prognostic value in SA-ARDS patients remains unclear. Given the high inflammatory burden and tissue damage, and prevalent catabolism in this population, we hypothesized that RAR could predict outcomes in this critical subgroup. This study aimed to evaluate the association between RAR and all-cause mortality (ACM) in the intensive care unit (ICU) patients with SA-ARDS. These findings could provide new insights for the early identification of this population and strategies to improve outcomes, potentially leading to better treatment and survival rates.
Methods
Data source
This study employed a retrospective observational approach. Data were extracted from the MIMIC-IV database (V3.1), a comprehensive, publicly available resource that includes data on patients admitted to the ICU of a large, tertiary care hospital in Boston between 2008 and 2022. The first author (Hongzhou Wu) accessed the database by completing the Collaborative Institutional Training Initiative (CITI) program and obtaining certification (certificate number: 65905104). Given the de-identified nature of the dataset and the absence of protected health information, informed consent was waived.
Eligibility criteria
This study included adult patients who met the 2012 Berlin definition of ARDS [20] and the Sepsis-3 diagnostic criteria. SQL queries were used to identify subjects with SA-ARDS in the MIMIC-IV database. First, cases with a PaO₂/FiO₂ ratio ≤ 300 mmHg and a PEEP ≥ 5 cmH₂O within the first 24 h after ICU admission were extracted to identify ARDS, in accordance with the 2012 Berlin Definition. Then, individuals diagnosed with heart failure based on ICD codes were excluded. The final cohort was obtained by intersecting the resulting set with the Sepsis-3 table. Detailed SQL code can be found in Supplementary Material 1. Individuals who met any of the following criteria were excluded: (1) age under 18 years; (2) absence of critical data, including RDW or Alb values; and (3) subjects with heart failure (ICD-9/10 codes beginning with 428 or I50).
Covariates and outcomes
Using SQL queries, patient data within the first 24 h of ICU admission were acquired. The data encompassed: demographics (marital status, sex, race, age, BMI); vital signs (heart rate, blood pressure, respiratory rate [RR]); laboratory values (neutrophil percentage [Neu%], hemoglobin [Hb], red blood cell count [RBC], RDW, Alb, aspartate aminotransferase [AST], indirect bilirubin, alanine aminotransferase [ALT], direct bilirubin, total bilirubin [TBil]); comorbidities (atrial fibrillation [AF], diabetes mellitus [DM], hypertension, cancer, cirrhosis, stroke); severity scores (Sequential Organ Failure Assessment [SOFA] score, Glasgow Coma Scale [GCS] score); medication use (dopamine, dobutamine, epinephrine, norepinephrine, isoproterenol, vasopressin, statins); and supportive therapies (continuous renal replacement therapy [CRRT], mechanical ventilation [MV]). The outcome was ACM at 30, 90, 180, and 365 days.
Handling of missing data and outliers
Variables with ≤ 30% missing data were imputed using multiple imputation. Missing data patterns were visualized using the plot_missing() function. Variables with more than 30% missingness were removed. For variables with ≤ 30% missing values, multiple imputation was performed using chained equations (MICE) coupled with random forest algorithms (method="rf”, m = 5, seed = 123). The results were pooled according to Rubin’s rules. Variables with > 30% missing data were not included in the analysis. Outliers were defined as values beyond the 99th percentile or below the 1st percentile. Any outliers were handled via a capping method.
Statistical analysis
Patients were stratified into quartiles based on RAR values: Q1 (n = 1,508, RAR ≤ 4.14), Q2 (n = 1,511, 4.14 < RAR ≤ 5.07), Q3 (n = 1,512, 5.07 < RAR ≤ 6.32), and Q4 (n = 1,511, RAR > 6.32). Categorical variables were described utilizing frequencies and percentages. Continuous variables were assessed for normality via the Shapiro-Wilk test. Mean (SD) was applied for normally distributed variables, while median (quartile) was adopted for non-normally distributed variables. Kaplan-Meier (KM) curves and Cox regression were utilized to evaluate RAR’s association with mortality risk. Three nested Cox regression models were constructed to refine the analysis. Model 1 (base model) included only the RAR and made no adjustments for any covariates. Model 2 (demographic model) incorporated adjustments for demographics (marital status, sex, race, age). Model 3 (fully adjusted model) was built upon Model 2 by incorporating demographics, comorbidities, pathophysiological variables, severity scores, and interventions. Model 3 made further adjustments for several indicators, including inflammatory/anemia indicators (Neu%, Hb, anion gap, glucose, platelet count [PLT]), comorbidities (AF, DM, hypertension, cancer, stroke, cirrhosis), hepatic/renal function indicators (TBil, creatinine [Cre], blood urea nitrogen [BUN]), medication use (dopamine, dobutamine, epinephrine, norepinephrine, isoproterenol, vasopressin, statins, aspirin, clopidogrel), supportive therapies (MV, CRRT), and severity scores (SOFA, GCS). The proportional hazards assumption in the Cox model was verified by computing Schoenfeld residuals. Restricted cubic spline (RCS) regression was employed to examine the dose-response association of RAR with ACM, assess the linearity, and evaluate RAR’s predictive ability for SA-ARDS. Interaction tests and subgroup analyses were performed to evaluate the consistency of RAR’s predictive ability across different subgroups defined by demographics, medication use, and therapeutic interventions. Propensity score matching (PSM) was utilized to ensure the robustness of the findings. To address potential bias from very brief ICU stays, a sensitivity analysis was performed. This analysis excluded patients with an ICU stay under 48 h. Statistical analyses were conducted in R software (V4.4.1). All tests adopted the same significance level (P < 0.05).
Results
Included population
Using SQL queries, 11,438 patients who met the 2012 Berlin definition for ARDS were identified from a pool of 364,627 participants in the MIMIC-IV 3.1 database. Of those patients, 7,628 satisfied the Sepsis-3 criteria as well. After excluding patients under 18 years of age, those with missing RDW or Alb data, and those with ICU stays of less than 48 h, 6,042 patients were included in this study (Fig. 1).
Fig. 1.
Flowchart of the included population
Comparisons of between-group characteristics
The current study comprised 6,042 patients. Their baseline characteristics are presented in Table 1. The participants were stratified by RAR quartile (Q1, Q2, Q3, Q4). Among these patients, 3,654 were male (60.5%), and 2,388 were female (39.5%). Of the patients, 2,936 were younger than 60 years old (48.6%), while 3,106 were 60 years old or older (51.4%). During the 365-day follow-up period, 2,160 (35.7%) patients died. Compared to the other RAR quartiles, participants in the highest RAR quartile (Q4) exhibited elevated heart rate, RR, AST, anion gap, BUN, and SOFA scores. They also had a higher prevalence of cancer and cirrhosis, as well as a greater utilization rate of epinephrine, norepinephrine, vasopressin, and CRRT. Conversely, participants in Q4 exhibited lower diastolic (DBP) and systolic blood pressure (SBP), Hb, hematocrit, PLT, and a lower proportion of individuals with prior stroke, as well as decreased usage of aspirin and clopidogrel. There was no significant difference in the prevalence of AR between Q4 and the other quartiles (Q1, Q2, Q3). The number of deaths at 30, 90, 180, and 365 days was significantly higher in the Q4 group (p < 0.001). To clarify the differences between each quartile group, intergroup pairwise p-value comparisons were conducted (Supplementary Material 2). The differences observed in these comparisons generally aligned with the overall trend.
Table 1.
Baseline characteristics of patients grouped per RAR quartiles
| ALL | Q1 | Q2 | Q3 | Q4 | P for overall | |
|---|---|---|---|---|---|---|
| N = 6,042 | N = 1,508 | N = 1,511 | N = 1,512 | N = 1,511 | ||
| Demographics, n% | ||||||
| Sex | < 0.001 | |||||
| Male | 3,654 (60.5%) | 1,006 (66.7%) | 934 (61.8%) | 884 (58.5%) | 830 (54.9%) | |
| Female | 2,388 (39.5%) | 502 (33.3%) | 577 (38.2%) | 628 (41.5%) | 681 (45.1%) | |
| Age group | 0.018 | |||||
| < 60 | 2,936 (48.6%) | 774 (51.3%) | 697 (46.1%) | 714 (47.2%) | 751 (49.7%) | |
| ≥ 60 | 3,106 (51.4%) | 734 (48.7%) | 814 (53.9%) | 798 (52.8%) | 760 (50.3%) | |
| Race | 0.479 | |||||
| White | 3,579 (59.2%) | 877 (58.2%) | 897 (59.4%) | 912 (60.3%) | 893 (59.1%) | |
| Non-White | 1,230 (20.4%) | 308 (20.4%) | 297 (19.7%) | 295 (19.5%) | 330 (21.8%) | |
| Unknown | 1,233 (20.4%) | 323 (21.4%) | 317 (21.0%) | 305 (20.2%) | 288 (19.1%) | |
| Marital status | 0.008 | |||||
| Single | 1,863 (30.8%) | 458 (30.4%) | 464 (30.7%) | 450 (29.8%) | 491 (32.5%) | |
| Divorced or widowed | 787 (13.0%) | 160 (10.6%) | 222 (14.7%) | 220 (14.6%) | 185 (12.2%) | |
| Married | 2397 (39.7%) | 628 (41.6%) | 564 (37.3%) | 598 (39.6%) | 607 (40.2%) | |
| Unknown | 995 (16.5%) | 262 (17.4%) | 261 (17.3%) | 244 (16.1%) | 228 (15.1%) | |
| Vital signs | ||||||
| SBP, mmHg | 117 [102;138] | 120 [105;141] | 118 [102;138] | 118 [102;138] | 114 [100;133] | < 0.001 |
| DBP, mmHg | 60.0 [52.0;70.0] | 63.0 [54.0;72.0] | 61.0 [52.0;70.0] | 60.0 [51.0;69.0] | 58.0 [49.5;68.0] | < 0.001 |
| Heart rate, beats/minute | 91.0 [79.0;107] | 86.0 [75.0;99.0] | 89.0 [77.0;104] | 94.0 [79.0;109] | 97.0 [84.0;113] | < 0.001 |
| RR, breaths/minute | 19.0 [16.0;24.0] | 18.0 [15.0;22.0] | 19.0 [16.0;23.0] | 19.0 [16.0;24.0] | 20.0 [16.0;25.0] | < 0.001 |
| Laboratory tests | ||||||
| ALT, U/L | 31.0 [17.0;67.0] | 29.0 [18.0;58.0] | 31.0 [17.0;65.0] | 33.0 [17.0;73.0] | 31.0 [17.0;72.0] | 0.149 |
| AST, U/L | 45.0 [25.0;104] | 37.0 [23.0;73.0] | 44.0 [25.0;100] | 48.0 [27.0;122] | 54.0 [27.0;132] | < 0.001 |
| Anion gap, mEq/L | 14.0 [12.0;17.0] | 14.0 [12.0;16.0] | 14.0 [12.0;17.0] | 14.0 [12.0;17.0] | 14.0 [11.0;18.0] | 0.135 |
| Total bilirubin, mg/dL | 0.60 [0.40;1.50] | 0.50 [0.40;0.90] | 0.60 [0.40;1.10] | 0.70 [0.40;1.70] | 1.00 [0.50;3.10] | < 0.001 |
| BUN, mg/dL | 19.0 [13.0;31.0] | 16.0 [12.0;22.0] | 18.0 [13.0;29.0] | 21.0 [14.0;35.0] | 24.0 [15.0;41.0] | <0.001 |
| Creatinine, mg/dL | 1.00 [0.70;1.50] | 0.90 [0.70;1.20] | 1.00 [0.70;1.50] | 1.00 [0.70;1.60] | 1.10 [0.70;1.90] | < 0.001 |
| Chloride, mEq/L | 105 [101;108] | 104 [101;108] | 104 [100;108] | 105 [101;109] | 105 [100;110] | 0.001 |
| Glucose, mg/dL | 136 [109;177] | 136 [111;176] | 140 [111;182] | 136 [110;177] | 131 [105;174] | <0.001 |
| Hb, g/dL | 10.6 [9.00;12.4] | 12.0 [10.3;13.5] | 11.1 [9.50;12.7] | 10.3 [8.90;11.8] | 9.40 [8.10;10.9] | < 0.001 |
| Neu% | 80.0 [70.9;86.4] | 78.6 [69.8;85.1] | 79.9 [70.4;86.6] | 80.8 [72.0;87.0] | 81.0 [71.0;87.0] | <0.001 |
| PLT, K/uL | 182 [119;258] | 196 [147;253] | 184 [126;255] | 174 [112;259] | 164 [90.0;265] | <0.001 |
| Comorbidities, n% | ||||||
| AF | 1,537 (25.4%) | 367 (24.3%) | 392 (25.9%) | 413 (27.3%) | 365 (24.2%) | 0.152 |
| Cancer | 1,788 (29.6%) | 383 (25.4%) | 398 (26.3%) | 486 (32.1%) | 521 (34.5%) | < 0.001 |
| Cirrhosis | 1,049 (17.4%) | 105 (6.96%) | 207 (13.7%) | 297 (19.6%) | 440 (29.1%) | < 0.001 |
| Diabetes | 1,778 (29.4%) | 424 (28.1%) | 483 (32.0%) | 453 (30.0%) | 418 (27.7%) | 0.038 |
| Hypertension | 3,330 (55.1%) | 904 (59.9%) | 878 (58.1%) | 819 (54.2%) | 729 (48.2%) | < 0.001 |
| Stroke | 731 (12.1%) | 227 (15.1%) | 200 (13.2%) | 174 (11.5%) | 130 (8.60%) | < 0.001 |
| Treatment, n% | ||||||
| Aspirin | 1,234 (20.4%) | 423 (28.1%) | 324 (21.4%) | 273 (18.1%) | 214 (14.2%) | < 0.001 |
| Clopidogrel | 147 (2.43%) | 52 (3.45%) | 48 (3.18%) | 21 (1.39%) | 26 (1.72%) | < 0.001 |
| Dopamine | 173 (2.86%) | 46 (3.05%) | 47 (3.11%) | 36 (2.38%) | 44 (2.91%) | 0.615 |
| Epinephrine | 3,191 (52.8%) | 584 (38.7%) | 757 (50.1%) | 826 (54.6%) | 1024 (67.8%) | < 0.001 |
| Norepinephrine | 2,666 (44.1%) | 479 (31.8%) | 608 (40.2%) | 681 (45.0%) | 898 (59.4%) | < 0.001 |
| Phenylephrine | 517 (8.56%) | 137 (9.08%) | 127 (8.41%) | 133 (8.80%) | 120 (7.94%) | 0.7 |
| Statins | 1,287 (21.3%) | 464 (30.8%) | 353 (23.4%) | 268 (17.7%) | 202 (13.4%) | < 0.001 |
| Vasopressin | 1,263 (20.9%) | 165 (10.9%) | 258 (17.1%) | 348 (23.0%) | 492 (32.6%) | < 0.001 |
| CRRT | 247 (4.09%) | 23 (1.53%) | 45 (2.98%) | 70 (4.63%) | 109 (7.21%) | < 0.001 |
| MV | 4,744 (78.5%) | 1,169 (77.5%) | 1,192 (78.9%) | 1,184 (78.3%) | 1,199 (79.4%) | 0.642 |
| Scores | ||||||
| GCS | 15.0 [15.0;15.0] | 15.0 [15.0;15.0] | 15.0 [15.0;15.0] | 15.0 [15.0;15.0] | 15.0 [15.0;15.0] | 0.446 |
| SOFA | 7.00 [4.00;10.0] | 5.00 [3.00;8.00] | 6.00 [4.00;9.00] | 7.00 [5.00;10.0] | 8.00 [5.50;12.0] | < 0.001 |
| Survival, n% | ||||||
| 30 days | 4,748 (78.6%) | 1,335 (88.5%) | 1,230 (81.4%) | 1,169 (77.3%) | 1,014 (67.1%) | < 0.001 |
| 90 days | 4,326 (71.6%) | 1,276 (84.6%) | 1,143 (75.6%) | 1,050 (69.4%) | 857 (56.7%) | < 0.001 |
| 180 days | 4,116 (68.1%) | 1,246 (82.6%) | 1,084 (71.7%) | 1,000 (66.1%) | 786 (52.0%) | < 0.001 |
| 365 days | 3,882 (64.3%) | 1,202 (79.7%) | 1,029 (68.1%) | 929 (61.4%) | 722 (47.8%) | < 0.001 |
SBP: Systolic blood pressure; DBP: Diastolic blood pressure; RR: Respiratory rate; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; BUN: Blood urea nitrogen; Hb: Hemoglobin; Neu%: Neutrophils%; PLT: Platelet; AF: Atrial fibrillation; CRRT: Continuous renal replacement therapy; MV: Mechanical ventilation.
KM survival curves
A KM survival analysis was performed to evaluate the relationship between RAR and mortality risk in patients with SA-ARDS (Fig. 2). The KM curves for 30-day mortality (Fig. 2A) demonstrated a significant increase in mortality with increasing RAR quartiles. Compared to patients in Q1, those in Q4 exhibited a substantially elevated mortality risk. This trend was consistently observed in the 90-, 180-, and 365-day survival curves (Fig. 2B and C, and 2D).
Fig. 2.
KM curves for different levels of RAR on ACM in SA-ARDS. (A) 30-day KM curve; (B) 90-day KM curve; (C) 180-day KM curve; (D) 365-day KM curve
Relationship between RAR and mortality risk in SA-ARDS patients
A Cox regression analysis was performed to examine the independent effect of RAR on mortality risk in SA-ARDS patients (Table 2). RAR was found to be significantly linked to mortality. In the unadjusted model (Model 1), RAR was positively related to 30-, 90-, 180-, and 365-day mortality (P < 0.001). Following adjustments for demographics (marital status, sex, race, age), inflammatory and anemia indicators (Neu%, Hb, anion gap, glucose, PLT), comorbidities (AF, DM, hypertension, cancer, stroke, cirrhosis), hepatic and renal function indicators (TBil, Cre, BUN), medication use (dopamine, dobutamine, epinephrine, norepinephrine, isoproterenol, vasopressin, statins, aspirin, clopidogrel), supportive therapies (MV, CRRT), and severity scores (SOFA, GCS) (Model 3), the association remained significant (P < 0.001).
Table 2.
Cox proportional hazard models (PHMs) for ACM
| Variables | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | |
| 30-day ACM | ||||||
| RAR | 1.23(1.20–1.26) | < 0.001 | 1.23(1.20–1.27) | < 0.001 | 1.15(1.11–1.19) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.68(1.39–2.03) | < 0.001 | 1.66(1.37–2.01) | < 0.001 | 1.50(1.24–1.82) | < 0.001 |
| Q3 | 2.09(1.74–2.51) | < 0.001 | 2.08(1.73–2.50) | < 0.001 | 1.66(1.37–2.02) | < 0.001 |
| Q4 | 3.19(2.68–3.79) | < 0.001 | 3.24(2.73–3.86) | < 0.001 | 2.26(1.86–2.75) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 90-day ACM | ||||||
| RAR | 1.25(1.22–1.28) | < 0.001 | 1.25(1.23–1.28) | < 0.001 | 1.17(1.14–1.20) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.67(1.42–1.97) | < 0.001 | 1.65(1.40–1.95) | < 0.001 | 1.48(1.25–1.75) | < 0.001 |
| Q3 | 2.16(1.85–2.53) | < 0.001 | 2.17(1.85–2.54) | < 0.001 | 1.71(1.44–2.02) | < 0.001 |
| Q4 | 3.32(2.86–3.86) | < 0.001 | 3.42(2.94–3.98) | < 0.001 | 2.34(1.97–2.77) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 180-day ACM | ||||||
| RAR | 1.27(1.25–1.30) | < 0.001 | 1.26(1.23–1.29) | < 0.001 | 1.18(1.14–1.21) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.73(1.48–2.02) | < 0.001 | 1.71(1.46–1.99) | < 0.001 | 1.53(1.30–1.79) | < 0.001 |
| Q3 | 2.15(1.86–2.50) | < 0.001 | 2.16(1.86–2.51) | < 0.001 | 1.71(1.46-2.00) | < 0.001 |
| Q4 | 3.36(2.92–3.87) | < 0.001 | 3.46(3.00-3.99) | < 0.001 | 2.38(2.03–2.79) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 365-day ACM | ||||||
| RAR | 1.24(1.21–1.27) | < 0.001 | 1.25(1.22–1.28) | < 0.001 | 1.17(1.14–1.20) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.69(1.46–1.95) | < 0.001 | 1.67(1.44–1.92) | < 0.001 | 1.50(1.30–1.74) | < 0.001 |
| Q3 | 2.13(1.86–2.45) | < 0.001 | 2.14(1.86–2.46) | < 0.001 | 1.72(1.48–1.99) | < 0.001 |
| Q4 | 3.23(2.83–3.69) | < 0.001 | 3.33(2.92–3.80) | < 0.001 | 2.35(2.02–2.73) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
Model 1: No adjustment
Model 2: Adjusted for marriage, sex, race, and age
Model 3: Adjusted for marriage, sex, race, age, DBP, RR, SBP, HR, anion gap, AST, BUN, chloride, Cre, glucose, Hb, Neu%, PLT, AF, cancer, cirrhosis, hypertension, DM, stroke, aspirin, clopidogrel, dobutamine, dopamine, epinephrine, norepinephrine, phenylephrine, statins, vasopressin, MV, CRRT, GCS, and SOFA
For categorical RAR (quartiles), the hazard ratio (HR) for 30-, 90-, 180-, and 365-day mortality increased with increasing RAR quartiles (Q1, Q2, Q3, Q4) in the unadjusted model (Model 1). The HR for Q4 was 3.19 (2.68–3.79), 3.32 (2.86–3.86), 3.36 (2.92–3.87), and 3.23 (2.83–3.69), respectively (p < 0.001). After adjusting for demographics, inflammatory and anemia indicators, comorbidities, hepatic and renal function indicators, medication use, supportive therapies, and severity scores (Model 3), the association remained significant (p < 0.001).
Exploration of nonlinearity
RCS regression revealed a nonlinear relationship between RAR and ACM at 30, 90, 180, and 365 days in SA-ARDS patients (p for nonlinearity: 0.014, 0.004, 0.002, and < 0.001; Fig. 3). The estimated ideal RAR cutoff for ACM was 5.04. Above this threshold, the mortality risk increased substantially.
Fig. 3.
RCS curves for RAR on mortality risk in SA-ARDS at different observation times. (A) 30-day RCS; (B) 90-day RCS; (C) 180-day RCS; (D) 365-day RCS
Predictive value of RAR in SA-ARDS
Time-dependent receiver operating characteristic (ROC) curve analysis indicated that RAR predicted ACM in SA-ARDS patients (Fig. 4) with an area under the curve (AUC) of 0.634 (95% confidence interval [CI]: 0.618–0.651) at 30 days, 0.648 (0.633–0.664) at 90 days, 0.652 (0.637–0.666) at 180 days, and 0.649 (0.635–0.664) at 365 days. At all timepoints, the predictive performance of RAR was significantly superior to that of RDW or Alb alone. The predictive ability of RAR was inferior to the SOFA score at 30 and 90 days, comparable to the SOFA score at 180 days, and superior to the SOFA score at 365 days.
Fig. 4.
AUC curves for RAR on ACM in SA-ARDS at different observation times. (A) 30-day ROC; (B) 90-day ROC; (C) 180-day ROC; (D) 365-day ROC. RDW: Red cell distribution width; RAR: RDW-to-albumin ratio; ALB: Albumin
Subgroup analyses
Using stratified Cox PHMs, this study evaluated the effect modification by incorporating interaction terms (RAR level×variable) for 17 clinically relevant subgroup variables, including age, sex, AF, hypertension, cancer, DM, cirrhosis, stroke, and the use of dopamine, epinephrine, norepinephrine, isoproterenol, vasopressin, aspirin, clopidogrel, CRRT, and MV. Separate models were constructed for each endpoint (30-, 90-, 180-, and 365-day mortality) (Fig. 5). The results indicated no significant interaction between RAR and the majority of variables (p for interaction > 0.05). This suggested robust generalizability of the risk association across a broad population. However, potential interaction effects were observed with age (30-day p for interaction = 0.039), cancer (30-day p = 0.002 for interaction; 90-day p = 0.019 for interaction), and vasopressor use (epinephrine: 90-, 180-, and 365-day p < 0.05 for interaction; norepinephrine: 90- and 180-day p < 0.05 for interaction; vasopressin: 90-, 180-, and 365-day p < 0.05 for interaction).
Fig. 5.
30-, 90-, 180-, and 365-day subgroup analysis. (A) 30-day forest plot; (B) 90-day forest plot; (C) 180-day forest plot; (D) 365-day forest plot
Sensitivity analysis
PSM was employed to balance the covariates. Patients were matched at a 1:1 ratio based on a propensity score with an RAR cutoff of 5.04. A total of 2,954 patients (1,477 matched pairs) were successfully matched. Then, a Cox regression analysis was performed on the matched group to determine if RAR independently affected the mortality risk of SA-ARDS patients (Table 3). The results indicated a positive association between RAR and ACM at 30, 90, 180, and 365 days in the unadjusted model (Model 1) (p < 0.05). After adjusting for demographic, comorbidity, pathophysiological, and severity score variables (Model 3), this association remained significant (p < 0.05). For categorical variables, the HR for ACM at 30, 90, 180, and 365 days increased with increasing RAR in the unadjusted model (Model 1) compared to the Q1 model (p < 0.05). This pattern persisted after full adjustment for confounders (Model 3) (p < 0.05). Following PSM, an elevated RAR remained linked to an increased HR for mortality risk at 30, 90, 180, and 365 days. To illustrate the covariate balance after PSM, Supplementary Material 3 presents the standardized mean differences for each variable before and after matching, facilitating an assessment of the matching effectiveness. To minimize potential bias from excessively short ICU stays, patients with ICU stays of less than 48 h were excluded. The final cohort comprised 5,450 individuals. Cox regression results were consistent with the primary findings (Table 4): the association between elevated RAR and increased mortality risk at 30, 90, 180, and 365 days remained statistically significant and materially unchanged.
Table 3.
Cox PHMs for ACM after PSM
| Variables | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR(95% CI) | P | HR(95% CI) | P | HR(95% CI) | P | |
| 30-day ACM | ||||||
| RAR | 1.15(1.09–1.21) | < 0.001 | 1.15(1.09–1.21) | < 0.001 | 1.09(1.03–1.15) | 0.002 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.56(1.18–2.06) | 0.002 | 1.52(1.15–2.01) | 0.003 | 1.45(1.09–1.92) | 0.010 |
| Q3 | 1.69(1.28–2.22) | < 0.001 | 1.65(1.25–2.17) | < 0.001 | 1.66(1.26–2.20) | < 0.001 |
| Q4 | 2.22(1.66–2.96) | < 0.001 | 2.19(1.64–2.92) | < 0.001 | 1.91(1.42–2.58) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 90-day ACM | ||||||
| RAR | 1.17(1.12–1.22) | < 0.001 | 1.24(1.20–1.29) | < 0.001 | 1.11(1.06–1.17) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.50(1.18–1.90) | < 0.001 | 1.47(1.16–1.86) | 0.001 | 1.38(1.08–1.75) | 0.009 |
| Q3 | 1.63(1.29–2.06) | < 0.001 | 1.61(1.27–2.03) | < 0.001 | 1.58(1.25-2.00) | < 0.001 |
| Q4 | 2.26(1.77–2.89) | < 0.001 | 2.27(1.78–2.90) | < 0.001 | 1.93(1.49–2.48) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 180-day ACM | ||||||
| RAR | 1.18(1.13–1.23) | < 0.001 | 1.18(1.13–1.23) | < 0.001 | 1.12(1.07–1.17) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.64(1.31–2.05) | < 0.001 | 1.59(1.28–1.99) | < 0.001 | 1.47(1.17–1.84) | < 0.001 |
| Q3 | 1.65(1.32–2.06) | < 0.001 | 1.63(1.30–2.04) | < 0.001 | 1.58(1.26–1.98) | < 0.001 |
| Q4 | 2.41(1.91–3.04) | < 0.001 | 2.43(1.93–3.07) | < 0.001 | 2.01(1.58–2.55) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 365-day ACM | ||||||
| RAR | 1.17(1.13–1.22) | < 0.001 | 1.18(1.13–1.23) | < 0.001 | 1.11(1.07–1.16) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.66(1.35–2.05) | < 0.001 | 1.62(1.31–1.99) | < 0.001 | 1.50(1.21–1.85) | < 0.001 |
| Q3 | 1.68(1.36–2.06) | < 0.001 | 1.66(1.35–2.05) | < 0.001 | 1.61(1.31–1.99) | < 0.001 |
| Q4 | 2.44(1.96–3.03) | < 0.001 | 2.46(1.98–3.06) | < 0.001 | 2.04(1.63–2.55) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
Model 1: No adjustment
Model 2: Adjusted for marriage, sex, race, and age
Model 3: Adjusted for marriage, sex, race, age, DBP, RR, SBP, HR, anion gap, AST, BUN, chloride, Cre, glucose, Hb, Neu%, PLT, AF, cancer, cirrhosis, hypertension, DM, stroke, aspirin, clopidogrel, dobutamine, dopamine, epinephrine, norepinephrine, phenylephrine, statins, vasopressin, MV, CRRT, GCS, and SOFA
Table 4.
Cox PH models for ACM after exclusion of ICU LOS < 2 days
| Variables | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | |
| 30-day ACM | ||||||
| RAR | 1.21(1.18–1.24) | < 0.001 | 1.22(1.18–1.25) | < 0.001 | 1.13(1.10–1.27) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.58(1.31–1.91) | < 0.001 | 1.59(1.32–1.92) | < 0.001 | 1.43(1.18–1.73) | < 0.001 |
| Q3 | 1.90(1.58–2.28) | < 0.001 | 1.91(1.59–2.29) | < 0.001 | 1.52(1.25–1.85) | < 0.001 |
| Q4 | 2.80(2.36–3.33) | < 0.001 | 2.89(2.43–3.44) | < 0.001 | 2.00(1.64–2.43) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 90-day ACM | ||||||
| RAR | 1.23(1.20–1.26) | < 0.001 | 1.24(1.21–1.27) | < 0.001 | 1.16(1.12–1.19) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.60(1.36–1.89) | < 0.001 | 1.61(1.36–1.90) | < 0.001 | 1.44(1.22–1.71) | < 0.001 |
| Q3 | 2.03(1.73–2.38) | < 0.001 | 2.04(1.74–2.40) | < 0.001 | 1.61(1.36–1.91) | < 0.001 |
| Q4 | 3.03(2.61–3.53) | < 0.001 | 3.16(2.71–3.68) | < 0.001 | 2.16(1.81–2.56) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 180-day ACM | ||||||
| RAR | 1.23(1.20–1.26) | < 0.001 | 1.24(1.21–1.27) | < 0.001 | 1.16(1.13–1.19) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.73(1.48–2.02) | < 0.001 | 1.69(1.44–1.97) | < 0.001 | 1.51(1.28–1.77) | < 0.001 |
| Q3 | 2.15(1.86–2.50) | < 0.001 | 2.02(1.74–2.36) | < 0.001 | 1.60(1.36–1.88) | < 0.001 |
| Q4 | 3.36(2.92–3.87) | < 0.001 | 3.19(2.76–3.69) | < 0.001 | 2.18(1.85–2.57) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
| 365-day ACM | ||||||
| RAR | 1.22(1.20–1.25) | < 0.001 | 1.23(1.21–1.26) | < 0.001 | 1.15(1.12–1.19) | < 0.001 |
| RAR quartiles | ||||||
| Q1 | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) | |||
| Q2 | 1.67(1.44–1.94) | < 0.001 | 1.67(1.44–1.93) | < 0.001 | 1.50(1.29–1.75) | < 0.001 |
| Q3 | 2.07(1.79–2.38) | < 0.001 | 2.08(1.80–2.40) | < 0.001 | 1.66(1.43–1.94) | < 0.001 |
| Q4 | 3.01(2.62–3.45) | < 0.001 | 3.12(2.72–3.58) | < 0.001 | 2.18(1.87–2.54) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
Model 1: No adjustment
Model 2: Adjusted for marriage, sex, race, and age
Model 3: Adjusted for marriage, sex, race, age, DBP, RR, SBP, HR, anion gap, AST, BUN, chloride, Cre, glucose, Hb, Neu%, PLT, AF, cancer, cirrhosis, hypertension, DM, stroke, aspirin, clopidogrel, dobutamine, dopamine, epinephrine, norepinephrine, phenylephrine, statins, vasopressin, MV, CRRT, GCS, and SOFA
Discussion
This appears to be the first study examining the association between RAR and ACM risk in SA-ARDS patients. KM analysis and Cox regression indicated that elevated RAR was significantly associated with increased 30-, 90-, 180-, and 365-day ACM after ICU admission. Furthermore, RCS regression further revealed a nonlinear relationship (P for nonlinearity < 0.05), with a notably increased mortality risk when RAR exceeded 5.04. ROC curve analysis showed that RAR had modest discriminatory ability for long-term (> 180 days) outcomes in SA-ARDS patients. The predictive performance of RAR for long-term mortality was comparable to that of the SOFA score and slightly exceeded it at 365 days. Subgroup analyses revealed that this association was robust across a broad population. PSM revealed that elevated RAR was an independent predictor of mortality risk in SA-ARDS. Based on these findings and the readily available nature of RAR, it holds promise as a prognostic biomarker in SA-ARDS patients.
RDW, a commonly available clinical indicator, reflects erythrocyte volume heterogeneity and can be used to assess anemia severity in patients [21]. It is now widely recognized that inflammatory mediators (e.g., interleukin-1, tumor necrosis factor-α) can induce erythrocyte apoptosis through activation of related pathways, thereby increasing RDW [22]. Furthermore, inflammation may affect erythrocyte maturation, causing myelosuppression, reduced erythropoietin production, and decreased iron bioavailability. This results in the release of immature erythrocytes into circulation, thereby increasing RDW [23]. Erythrocytes and hematopoietic stem cells are highly sensitive to elevated reactive oxygen species (ROS), which are caused by oxidative stress. Excessive ROS can also trigger hemolysis, leading to stress erythropoiesis and increased RDW [24]. Moreover, studies by Kato et al. [25] suggest that elevated RDW may be linked to the activation of the renin-angiotensin system. Critically ill patients often exhibit elevated RDW levels. Yu et al. [26] demonstrated that higher RDW demonstrated a stronger correlation with increased 30-day mortality in ARDS patients. RDW has been shown to have significant predictive value in various diseases, including infectious diseases [27], hypoxic diseases [28], cardiovascular diseases [29], cancer [30], and respiratory diseases [26].
Alb, a major colloid in the human body, is recognized for its role in critical physiological processes such as inflammation resolution and immune modulation [17]. It maintains intravascular oncotic pressure and exerts anti-inflammatory and antioxidant effects. Previous studies have linked hypoalbuminemia to nutritional depletion [31, 32]. Furthermore, excessive inflammation can impair hepatic Alb synthesis, leading to hypoalbuminemia. Hypoalbuminemia is closely related to pulmonary edema formation in ARDS [32]. Alb also has prognostic implications. For example, Kumar et al. found that hypoalbuminemia was strongly associated with an elevated 28-day mortality risk in individuals with post-traumatic ARDS [33].
RAR combines the predictive values of RDW and Alb in critically ill patients, reflecting the effects of inflammation, nutritional depletion, and edema combined. From a pathophysiological perspective, therefore, RAR is a relatively comprehensive biomarker for capturing these changes in critically ill individuals. As far as we are aware, this is the first study to apply RAR to the prognostic assessment of SA-ARDS patients. Higher RAR was linked to elevated 30-, 90-, 180-, and 365-day mortality risk in SA-ARDS patients. This finding is consistent with our initial hypothesis. A steep increase in mortality risk was observed when RAR was greater than approximately 5.04. This threshold should be considered exploratory and cohort‑specific and requires validation in independent populations before it can be applied in clinical practice. The temporal complementarity of RAR and the SOFA score is particularly noteworthy. RAR demonstrated modest yet notable discriminatory performance and was slightly superior to the SOFA score in predicting 365‑day mortality. This suggests the potential value of RAR in chronic risk stratification.
Subgroup analyses confirmed the stability of RAR’s predictive ability. Interaction analyses indicated that the increased risk associated with RAR in patients aged 60 years and older within the first 30 days of ICU admission attenuated with longer observation periods (p for interaction = 0.039). This may indicate that, within the initial 30 days, age is a significant determinant of prognosis, but that, beyond this timeframe, other factors, such as comorbidities, organ function status, and impaired immune function, become more influential on prognosis, thereby diminishing the effect of advanced age on mortality risk. This aligns with the review by Vallet et al. [34]. Significantly elevated HR values at 30 and 90 days in patients with malignancy indicate that cancer significantly increases the risk of short-term mortality following ICU admission. The less pronounced differences in HR values at 180 and 365 days may indicate that the acute risk associated with malignancy diminishes over time with treatment. These findings are consistent with those of Le Borgne et al. in 2022 [35]. Lower HR values observed in patients receiving vasopressors (epinephrine, norepinephrine, vasopressin) may be attributable to higher perfusion pressures maintained by these drugs and their potential endothelial protective mechanisms [36, 37].
In summary, RAR shows promise as a prognostic indicator for ICU patients with SA-ARDS. It surpasses the SOFA score in long-term (365 days) prediction and fills the void of long-term prognostic indicators for this population. This is due to its rapid and easily accessible nature. RAR could potentially contribute to the future risk stratification and management of SA‑ARDS patients.
Nevertheless, this study has several limitations. First, it was unable to capture all relevant patient variables. Unmeasured factors, such as socioeconomic status and lifestyle habits, may influence these patients’ prognoses. Second, the study only considered the initial laboratory results after ICU admission and did not account for their dynamic changes. Third, the impact of pre-ICU Alb infusions on RAR values was not considered. The lack of dynamic measurements for Alb and RDW also limited trend analysis and long-term risk assessment. Fourth, as a retrospective study using the MIMIC-IV database, the potential for misclassification of ARDS cases cannot be ruled out. Fifth, some of the adjusted variables used in this research, such as SOFA score and vasopressor use, may act as mediators of illness severity rather than true confounders. This could potentially attenuate the observed associations. Sixth, excluding patients with heart failure may affect the generalizability of the findings. Seventh, the predictive performance of RAR was moderate (AUC: approximately 0.63–0.65), warranting a cautious interpretation. Eighth, the RAR threshold of 5.04, identified through RCS analysis, is an exploratory finding specific to our cohort that requires external validation. Therefore, future prospective studies are necessary to validate the association between RAR and prognosis.
Conclusions
This study found that RAR is independently associated with an increased ACM at 30, 90, 180, and 365 days in ICU patients with SA-ARDS. RAR may serve as a convenient and accessible indicator for identifying high‑risk patients.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
Not applicable.
Abbreviations
- ACM
All-cause mortality
- AF
Atrial fibrillation
- Alb
Albumin
- ALT
Alanine aminotransferase
- AST
Aspartate aminotransferase
- AUC
Area under the curve
- BUN
Blood urea nitrogen
- CITT
Collaborative Institutional Training Initiative
- Cre
Creatinine
- CRRT
Continuous renal replacement therapy
- DM
DataObjectiabetes mellitus
- Hb
Hemoglobin
- HR
Hazard ratio
- ICU
Intensive care unit
- KM
Kaplan-Meier
- MV
Mechanical ventilation
- Neu%
Neutrophil percentage
- PSM
Propensity score matching
- RAR
Red cell distribution width-to-albumin ratio
- RBC
Rred blood cell count
- RCS
Restricted cubic spline
- ROC
Receiver operating characteristic
- RR
Respiratory rate
- SA-ARDS
Sepsis-associated acute respiratory distress syndrome
- SBP
Systolic blood pressure
- TBil
Total bilirubin
Author contributions
Hongzhou Wu: Conceptualization, methodology, software, formal analysis, investigation, resources, data curation, writing - original draft, visualization; Min He: Formal analysis, investigation, resources; Yujia Gui: Validation, investigation; Jinyi Zhou: Validation, software; Liting Yue: Validation; Xiaoqin Zhang: Validation, writing - review & editing, supervision, project administration. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding
The authors declare that they did not receive any funding from any source.
Data availability
A retrospective observational approach was used in this study. It employed data extracted from the MIMIC-IV database (V3.1).
Declarations
Ethics approval and consent to participate
The MIMIC-IV database was approved by the Massachusetts Institute of Technology (Cambridge, MA) and the Beth Israel Deaconess Medical Center (Boston, MA). Consent was obtained for the collection of original data. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committees, as well as the 1964 Helsinki Declaration and its subsequent amendments, or other ethical standards. Therefore, ethical approval and informed consent were waived for this study.
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.
Hongzhou Wu and Min He contributed equally to this work and should be considered as co-first authors
References
- 1.Lever A, Mackenzie I, Sepsis. Definition, epidemiology, and diagnosis. BMJ. 2007;335(7625):879–83. 10.1136/bmj.39346.495880.AE. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Weng L, Zeng XY, Yin P, Wang LJ, Wang CY, Jiang W, et al. Sepsis-related mortality in china: A descriptive analysis. Intensive Care Med. 2018;44(7):1071–80. 10.1007/s00134-018-5203-z. [DOI] [PubMed] [Google Scholar]
- 3.Dellinger RP, Rhodes A, Evans L, Alhazzani W, Beale R, Jaeschke R, et al. Surviving sepsis campaign. Crit Care Med. 2023;51(4):431–44. 10.1097/ccm.0000000000005804. [DOI] [PubMed] [Google Scholar]
- 4.Gavelli F, Castello LM, Avanzi GC. Management of sepsis and septic shock in the emergency department. Intern Emerg Med. 2021;16(6):1649–61. 10.1007/s11739-021-02735-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zhang H, Wang Y, Qu M, Li W, Wu D, Cata JP, et al. Neutrophil, neutrophil extracellular traps and endothelial cell dysfunction in sepsis. Clin Transl Med. 2023;13(1):e1170. 10.1002/ctm2.1170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Azoulay E, Zuber J, Bousfiha AA, Long Y, Tan Y, Luo S, et al. Complement system activation: bridging physiology, pathophysiology, and therapy. Intensive Care Med. 2024;50(11):1791–803. 10.1007/s00134-024-07611-4. [DOI] [PubMed] [Google Scholar]
- 7.Pokharel MD, Fu P, Garcia-Flores A, Yegambaram M, Lu Q, Sun X, et al. Inflammatory lung injury is associated with endothelial cell mitochondrial fission and requires the nitration of Rhoa and cytoskeletal remodeling. Free Radic Biol Med. 2024;221:125–35. 10.1016/j.freeradbiomed.2024.05.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Thompson BT, Chambers RC, Liu KD. Acute respiratory distress syndrome. N Engl J Med. 2017;377(6):562–72. 10.1056/NEJMra1608077. [DOI] [PubMed] [Google Scholar]
- 9.Ware LB, Matthay MA. The acute respiratory distress syndrome. N Engl J Med. 2000;342(18):1334–49. 10.1056/nejm200005043421806. [DOI] [PubMed] [Google Scholar]
- 10.Matthay MA, Zemans RL. The acute respiratory distress syndrome: pathogenesis and treatment. Annu Rev Pathol. 2011;6:147–63. 10.1146/annurev-pathol-011110-130158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Huang Y, Jiang S, Li W, Fan Y, Leng Y, Gao C. Establishment and effectiveness evaluation of a scoring system-raas (rdw, age, Apache ii, sofa) for sepsis by a retrospective analysis. J Inflamm Res. 2022;15:465–74. 10.2147/jir.S348490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hu ZD, Lippi G, Montagnana M. Diagnostic and prognostic value of red blood cell distribution width in sepsis: A narrative review. Clin Biochem. 2020; 77:1-6. 10.1016/j.clinbiochem.2020.01.001IF: 2.1 Q2 B3. [DOI] [PubMed]
- 13.Galindo-Martín CA, Chong-Aviña PA, Godinez-Breacher V, Aportela-Vázquez VA, Bueno-Hernández G, Gante-García MF, et al. Malnutrition: muscle wasting, inflammation, rdw, and their relation with adverse outcomes. Cir Cir. 2024;92(2):150–8. 10.24875/ciru.22000634. [DOI] [PubMed]
- 14.Colombo G, Clerici M, Giustarini D, Rossi R, Milzani A, Dalle-Donne I. Redox albuminomics: oxidized albumin in human diseases. Antioxid Redox Signal. 2012;17(11):1515–27. 10.1089/ars.2012.4702. [DOI] [PubMed] [Google Scholar]
- 15.Alcaraz-Quiles J, Casulleras M, Oettl K, Titos E, Flores-Costa R, Duran-Güell M, et al. Oxidized albumin triggers a cytokine storm in leukocytes through p38 mitogen-activated protein kinase: role in systemic inflammation in decompensated cirrhosis. Hepatology. 2018;68(5):1937–52. 10.1002/hep.30135. [DOI] [PubMed] [Google Scholar]
- 16.Taverna M, Marie AL, Mira JP, Guidet B. Specific antioxidant properties of human serum albumin. Ann Intensive Care. 2013;3(1):4. 10.1186/2110-5820-3-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Torres LN, Chung KK, Salgado CL, Dubick MA, Torres Filho IP. Low-volume resuscitation with normal saline is associated with microvascular endothelial dysfunction after hemorrhage in rats, compared to colloids and balanced crystalloids. Crit Care. 2017;21(1):160. 10.1186/s13054-017-1745-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Matejovic M, Krouzecky A, Rokyta R Jr., Novak I. Fluid challenge in patients at risk for fluid loading-induced pulmonary edema. Acta Anaesthesiol Scand. 2004;48(1):69–73. 10.1111/j.1399-6576.2004.00270.x. [DOI] [PubMed] [Google Scholar]
- 19.Hao M, Jiang S, Tang J, Li X, Wang S, Li Y, et al. Ratio of red blood cell distribution width to albumin level and risk of mortality. JAMA Netw Open. 2024;7(5):e2413213. 10.1001/jamanetworkopen.2024.13213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, et al. Acute respiratory distress syndrome: the Berlin definition. JAMA. 2012;307(23):2526–33. 10.1001/jama.2012.5669. [DOI] [PubMed] [Google Scholar]
- 21.Salvagno GL, Sanchis-Gomar F, Picanza A, Lippi G. Red blood cell distribution width: A simple parameter with multiple clinical applications. Crit Rev Clin Lab Sci. 2015;52(2):86–105. 10.3109/10408363.2014.992064. [DOI] [PubMed] [Google Scholar]
- 22.Mahnashi MH, Jabbar Z, Alamgeer, Irfan HM, Asim MH, Akram M, et al. Venlafaxine demonstrated anti-arthritic activity possibly through down regulation of tnf-α, il-6, il-1β, and cox-2. Inflammopharmacology. 2021;29(5):1413–25. 10.1007/s10787-021-00849-0. [DOI] [PubMed] [Google Scholar]
- 23.Purtle SW, Moromizato T, McKane CK, Gibbons FK, Christopher KB. The association of red cell distribution width at hospital discharge and out-of-hospital mortality following critical illness*. Crit Care Med. 2014;42(4):918–29. 10.1097/ccm.0000000000000118. [DOI] [PubMed] [Google Scholar]
- 24.Ghaffari S. Oxidative stress in the regulation of normal and neoplastic hematopoiesis. Antioxid Redox Signal. 2008;10(11):1923–40. 10.1089/ars.2008.2142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kato H, Ishida J, Imagawa S, Saito T, Suzuki N, Matsuoka T, et al. Enhanced erythropoiesis mediated by activation of the renin-angiotensin system via angiotensin Ii type 1a receptor. Faseb J. 2005;19(14):2023–5. 10.1096/fj.05-3820fje. [DOI] [PubMed] [Google Scholar]
- 26.Yu XS, Chen ZQ, Hu YF, Chen JX, Xu WW, Shu J, et al. Red blood cell distribution width is associated with mortality risk in patients with acute respiratory distress syndrome based on the Berlin definition: A propensity score matched cohort study. Heart Lung. 2020;49(5):641–5. 10.1016/j.hrtlng.2020.04.008. [DOI] [PubMed] [Google Scholar]
- 27.Wu H, Liao B, Cao T, Ji T, Huang J, Ma K. Diagnostic value of Rdw for the prediction of mortality in adult sepsis patients: A systematic review and meta-analysis. Front Immunol. 2022;13:997853. 10.3389/fimmu.2022.997853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Badrick T, Richardson AM, Arnott A, Lidbury BA. The kinetics of haemoglobin and ferritin in longitudinal community patients with iron deficiency or hypoxia. Diagnosis (Berl). 2017;4(1):35–41. 10.1515/dx-2016-0031. [DOI] [PubMed] [Google Scholar]
- 29.García-Escobar A, Lázaro-García R, Goicolea-Ruigómez J, González-Casal D, Fontenla-Cerezuela A, Soto N, et al. Red blood cell distribution width is a biomarker of red cell dysfunction associated with high systemic inflammation and a prognostic marker in heart failure and cardiovascular disease: A potential predictor of atrial fibrillation recurrence. High Blood Press Cardiovasc Prev. 2024;31(5):437–49. 10.1007/s40292-024-00662-0. [DOI] [PubMed] [Google Scholar]
- 30.Hu L, Li M, Ding Y, Pu L, Liu J, Xie J, et al. Prognostic value of Rdw in cancers: A systematic review and meta-analysis. Oncotarget. 2017;8(9):16027–35. 10.18632/oncotarget.13784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Gatta A, Verardo A, Bolognesi M, Hypoalbuminemia. Intern Emerg Med. 2012;7(Suppl 3):S193–9. 10.1007/s11739-012-0802-0. [DOI] [PubMed] [Google Scholar]
- 32.Franch-Arcas G. The meaning of hypoalbuminaemia in clinical practice. Clin Nutr. 2001;20(3):265–9. 10.1054/clnu.2001.0438. [DOI] [PubMed] [Google Scholar]
- 33.Kumar M, Jain K, Chauhan R, Meena SC, Luthra A, Thakur H, et al. Hypoalbuminemia: incidence and its impact on acute respiratory distress syndrome and 28-day outcome in trauma patients. Eur J Trauma Emerg Surg. 2023;49(5):2305–14. 10.1007/s00068-023-02318-5. [DOI] [PubMed] [Google Scholar]
- 34.Vallet H, Schwarz GL, Flaatten H, de Lange DW, Guidet B, Dechartres A. Mortality of older patients admitted to an icu: A systematic review. Crit Care Med. 2021;49(2):324–34. 10.1097/ccm.0000000000004772. [DOI] [PubMed] [Google Scholar]
- 35.Le Borgne P, Feuillassier L, Schenck M, Herbrecht JE, Janssen-Langenstein R, Simand C, et al. Comparison of short- and long-term mortality in patients with or without cancer admitted to the Icu for septic shock: A retrospective observational study. Cancers (Basel). 2022;14(13). 10.3390/cancers14133196. [DOI] [PMC free article] [PubMed]
- 36.Farand P, Hamel M, Lauzier F, Plante GE, Lesur O. Review article: organ perfusion/permeability-related effects of norepinephrine and vasopressin in sepsis. Can J Anaesth. 2006;53(9):934–46. 10.1007/bf03022837. [DOI] [PubMed] [Google Scholar]
- 37.Joffre J, Lloyd E, Wong E, Chung-Yeh C, Nguyen N, Xu F, et al. Catecholaminergic vasopressors reduce toll-like receptor agonist-induced microvascular endothelial cell permeability but not cytokine production. Crit Care Med. 2021;49(3):e315–26. 10.1097/ccm.0000000000004854. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
A retrospective observational approach was used in this study. It employed data extracted from the MIMIC-IV database (V3.1).