Abstract
Background
Red blood cell distribution width (RDW) is a marker that routinely describes circulating erythrocytes variability and may hold potential prognostic value in cardiac surgery. Our objective is to evaluate the association between peak values of RDW and in-hospital mortality among patients receiving post-cardiotomy veno-arterial extracorporeal membrane oxygenation (V-A ECMO) support.
Methods
This single-center, retrospective study included adult patients receiving post-cardiotomy V-A ECMO support between January 2017 and December 2024. The association between peak values of RDW and in-hospital mortality was assessed by multivariable logistic regression model. Spearman correlation analysis was used to determine the correlations between related laboratory factors and peak values of RDW.
Results
A total of 106 patients were included and in-hospital mortality rate was 54.7%. After adjusting for confounding factors, increased peak values of RDW during ECMO support were significantly associated with higher risk of in-hospital mortality (aOR, 2.097; 95% CI, 1.328–3.312; p = 0.001). Peak values of RDW were negatively correlated with lowest values of hemoglobin and positively correlated with peak values of lactate dehydrogenase (LDH), direct bilirubin (DBil), total bilirubin (TBil), neutrophil-to-lymphocyte ratio (NLR) and fibrinogen degradation products (FDP) occurring within ± 12 h of the RDW peak levels.
Conclusions
In patients receiving post-cardiotomy V-A ECMO, the peak values of RDW might be an independent prognostic factor for in-hospital mortality. Potential factors related to increased peak values of RDW included hemolysis, systemic inflammatory responses, and coagulation dysfunction.
Keywords: Red blood cell distribution width, Cardiac surgery, Veno-arterial extracorporeal membrane oxygenation, In-hospital mortality
Introduction
Veno-arterial extracorporeal membrane oxygenation (V-A ECMO) is a widely recognized and effective life-saving procedure for patients with post-cardiotomy cardiogenic shock (CS) [1]. However, patient outcomes continue to be poor, with in-hospital mortality rates ranging from 40 to 80% [2–5]. These patients usually face unique physiological challenges due to dual red blood cells (RBCs) damage resulting from both cardiopulmonary bypass (CPB) of the initial cardiac surgery and subsequent V-A ECMO support, characterized by mechanical shear stress, exposure to artificial surfaces, hemodilution, and systemic inflammatory response. Notably, the hematological consequences of this dual RBCs damage may be reflected in erythrocyte parameters.
Red blood cell distribution width (RDW) is a measure of the variability in the size of circulating RBCs and is traditionally employed in anemia diagnostics. Recently, several studies have shown that an elevated RDW level is an independent predictor of worse outcomes in patients with myocardial infarction [6], heart failure [7] and aortic valve disease [8] and most of these studies have focused on baseline RDW levels in patients with cardiovascular diseases prior to admission or surgery. In patients undergoing post-cardiotomy V-A ECMO, RBCs damage and inflammation may lead to a significant elevation in RDW levels during the V-A ECMO support. However, limited research has explored the potential association between peak values of RDW during V-A ECMO support and patient outcomes. This study aimed to investigate the relationship between peak values of RDW and in-hospital mortality in patients receiving post-cardiotomy V-A ECMO support and analyze the correlation between peak RDW values and other related laboratory parameters.
Materials and methods
Study population
This retrospective study included adult patients (aged ≥ 18 years) who received post-cardiotomy V-A ECMO support between January 2017 and December 2024. The study was approved by the institutional ethics committee/review board of our center (Approval number: 2024–2517). The requirement for informed consent was waived due to the retrospective nature of the study. The exclusion criteria were as follows: (1) patients supported with V-A ECMO for less than 48 h were excluded to ensure a meaningful assessment of RDW dynamics (n = 66); (2) patients with V-A ECMO support following heart transplantation were excluded to maintain a homogeneous study population (n = 41); (3) patients with incomplete or missing data were also excluded (n = 3). Patients were divided into two groups according to their discharge status (Fig. 1).
Fig. 1.
Study flow chart. V-A ECMO, veno-arterial extracorporeal membrane oxygenation
ECMO management
Indications for ECMO support
The indications for post-cardiotomy V-A ECMO support included (1) difficult weaning from CPB despite optimal inotropic agents and vasopressor (with or without intra-aortic balloon pump [IABP]), and (2) the occurrence of refractory CS or cardiac arrest postoperatively.
ECMO circuit
The V-A ECMO circuit consisted of an arterial cannula (Edwards Lifesciences, Irvine, CA, USA), a venous cannula (Edwards Lifesciences), an oxygenator kit (BE- PLS 2050, Maquet, Rastatt, Germany), and a centrifugal pump drive and console (Jostra Medizintechnik AG, Hirrlingen, Germany). Patients undergoing post-cardiotomy V-A ECMO may be managed with either peripheral or central cannulation, depending on their clinical condition. For peripheral V-A ECMO, a venous drainage cannula measuring 21–25 Fr is introduced through the femoral vein, while an arterial return cannula sized 16–18 Fr is inserted into the femoral artery. In cases where clinical assessment indicates inadequate distal perfusion, a distal perfusion cannula is connected to the side port of the arterial cannula, positioned either antegrade into the lower extremity or retrograde into a pedal artery. Central V-A ECMO involves aortic cannulation using an 18–20 Fr cannula, with venous access achieved through a 30–34 Fr cannula placed in the right atrium.
ECMO anticoagulation strategy
The anticoagulation strategy includes continuous infusion of unfractionated heparin to maintain an activated partial thromboplastin time (aPTT) of 60–80 s, while the target for activated clotting time (ACT) is established of 180–220 s.
Data collection
Patient data were collected from the electronic medical record system at our center. The information collected included demographic data, preoperative comorbidities, laboratory test results, V-A ECMO settings, transfusion data, patient outcomes, and complications. We collected RDW values within the first 6 h after V-A ECMO initiation, the daily RDW values and the peak values of RDW during V-A ECMO support. In our ICU, RDW was routinely measured two times per day. Additional measurements were performed for patients with unstable conditions. Daily RDW values were defined as the highest RDW recorded during each 24-hour period after ECMO initiation. The peak RDW was defined as the highest RDW recorded during the entire ECMO support period.
Outcomes
The primary outcome was in-hospital mortality after V-A ECMO initiation. Secondary outcomes included intensive care unit (ICU) and hospital length of stay, bleeding complications, blood product transfusion, pulmonary infection, limb complications, cerebral vascular accident, IABP implantation, the requirement of continuous renal replacement therapy (CRRT). Bleeding complications during ECMO support were defined as any of the following: (1) re-thoracotomy for bleeding; (2) cannulation-site bleeding; or (3) diffuse or uncontrolled non-surgical-related bleeding (e.g., gastrointestinal, respiratory, or intracranial).
Statistical analysis
The baseline characteristics of study patients were compared after grouping by discharge status. Normally distributed data were presented as mean ± standard deviation (SD), and non-normally distributed variables were presented as the median with interquartile range (25th − 75th percentile). Categorical variables were reported as frequencies and percentages (%). Statistical comparisons were conducted using the Chi-square test or Fisher’s exact test for categorical data, while the Student’s t-test or Mann-Whitney U test was employed for continuous data, based on the distribution of the data.
The association between peak value of RDW during post-cardiotomy V-A ECMO support and in-hospital mortality was evaluated using the multivariable logistic regression model with adjusted odds ratios (ORs) and 95% confidence intervals (CI). The model included sex, lactate, lactate dehydrogenase (LDH), hemoglobin, and fibrinogen within 6 h after V-A ECMO initiation, as well as peak values of RDW during V-A ECMO as covariates. A stepwise selection based on univariable analyses and previous literature was used to identify confounding variables for inclusion in the models. The variance inflation factor (VIF) was calculated to evaluate the collinearity of potential variables. All variables had a VIF less than 5 indicating absence of collinearity. The Spearman linear correlation coefficients (rs, range − 1 to 1) were computed to quantify the relationship between peak values of RDW during V-A ECMO and other related laboratory factors. To ensure the analysis captured the most clinically relevant pathological changes associated with RDW elevation, these laboratory factors were collected as their maximum values occurring within ± 12 h of the RDW peak levels. The coefficient describes the strength of a correlation (0.00–0.10.00.10, negligible correlation; 0.10–0.39, weak correlation; 0.40–0.69, moderate correlation; 0.70–0.89, strong correlation; 0.90–1.00.90.00, very strong correlation) [9].
A two-sided p-value ≤ 0.05 was regarded as statistically significant. Statistical analyses were performed using GraphPad Prism (version 8; GraphPad Software; La Jolla, CA) and SPSS statistical software (version 26.0, Armonk, NY, USA: IBM Corp).
Results
Patient characteristics
A total of 106 patients were included in final analysis after exclusion, with mean age of 56.1 ± 13.2 years and comprising 58.5% males (Table 1). The in-hospital mortality for patients receiving post-cardiotomy V-A ECMO support was 54.7%. The proportion of male patients was significantly lower in non-survivors (48.3% vs. 70.8%; P = 0.019). There were no significant differences in age, common comorbidities, laboratory tests before cardiac surgery, operation type, aortic cross-clamp time, CPB time and operative time between the two groups.
Table 1.
Demographic and operative data of the study population
| Total population (N = 106) | Survivors (N = 48) |
Non-Survivors (N = 58) |
P | |
|---|---|---|---|---|
| Age, years | 56.1 ± 13.2 | 54.1 ± 15.1 | 57.8 ± 11.4 | 0.148 |
| Male sex (%) | 62 (58.5) | 34 (70.8) | 28 (48.3) | 0.019 |
| BMI, kg/m2 | 24.8 ± 3.8 | 24.8 ± 3.7 | 24.7 ± 3.9 | 0.905 |
| Comorbidities (%) | ||||
| Diabetes mellitus | 13 (12.3) | 7 (14.6) | 6 (10.3) | 0.508 |
| Hypertension | 45 (42.5) | 20 (41.7) | 25 (43.1) | 0.882 |
| Hyperlipidemia | 44 (41.5) | 18 (37.5) | 26 (44.8) | 0.446 |
| Chronic kidney disease | 5 (4.7) | 2 (4.2) | 3 (5.2) | 0.808 |
| Cerebrovascular disease | 12 (11.3) | 5 (10.4) | 7 (12.1) | 0.789 |
| COPD | 7 (6.6) | 3 (6.3) | 4 (6.9) | 0.894 |
| LEVDD before surgery, mm | 52 (45, 60) | 52 (45, 59) | 55 (45, 60) | 0.345 |
| LVEF before surgery, % | 58 (50, 64) | 59 (52, 63.5) | 55 (47.3, 64.8) | 0.534 |
| Laboratory tests before cardiac surgery | ||||
| Hemoglobin, g/L | 135 (120.8, 143) | 137 (125, 143) | 135 (118, 143.3) | 0.534 |
| RDW, % | 13.2 (12.4, 14.4) | 12.9 (12.1, 13.8) | 13.2 (12.5, 14.6) | 0.275 |
| Creatinine, umol/L | 76.5 (64.8, 95.8) | 72.1 (60.7, 94.8) | 81.3 (69.8, 96.7) | 0.093 |
| BUN, mmol/L | 6.5 (5.5, 8.3) | 6.5 (5.5, 8.6) | 6.5 (5.4, 8.3) | 0.883 |
| LDH, IU/L | 195.5 (163.7, 235) | 192 (161.8, 250.5) | 200.5 (166.8, 231.3) | 0.965 |
| AST, IU/L | 24 (18, 31.3) | 25.5 (20.3, 33.5) | 24 (16.8, 31) | 0.209 |
| ALT, IU/L | 19 (14, 28.3) | 18.5 (13.3, 31.5) | 19 (14, 26.5) | 0.861 |
| TBil, umol/L | 13.9 (9, 19.9) | 11.6 (8.6, 18.4) | 14.4 (9.4, 22.4) | 0.230 |
| DBil, umol/L | 4.2 (2.3, 7.4) | 3.9 (2.4, 7) | 4.4 (2.3, 8) | 0.770 |
| NLR | 2.4 (1.8, 3.9) | 2.1 (1.8, 3.6) | 2.7 (1.9, 4.5) | 0.143 |
| D-dimer, ug/ml | 0.4 (0.2, 1.8) | 0.4 (0.3, 1.6) | 0.4 (0.2, 1.9) | 0.920 |
| aPTT, s | 37.5 (34.8, 41) | 38.2 (35.2, 41.2) | 35.9 (34.4, 40.6) | 0.208 |
| Fibrinogen, g/L | 3.3 (2.9, 3.9) | 3.6 (3, 3.9) | 3.5 (2.9, 4.2) | 0.526 |
| FDP, ug/ml | 2.5 (2.4, 5.1) | 2.5 (2.1, 4.1) | 2.5 (2.3, 6.6) | 0.254 |
| INR, R | 1.1 (0.9, 1.2) | 1.1 (0.9, 1.2) | 1.1 (0.9, 1.1) | 0.455 |
| Emergency surgery (%) | 21 (19.8) | 7 (14.6) | 14 (24.1) | 0.219 |
| Operation type (%) | 0.295 | |||
| CABG only | 25 (23.6) | 9 (18.8) | 16 (27.6) | |
| Valve operation only | 21 (19.8) | 10 (20.8) | 11 (19) | |
| Major aortic operation | 13 (12.3) | 4 (8.3) | 9 (15.5) | |
| CABG and valve operation | 13 (12.3) | 7 (14.6) | 6 (10.3) | |
| CABG and Major aortic operation | 16 (15.1) | 6 (12.5) | 10 (17.2) | |
| Other | 18 (17) | 12 (25) | 6 (10.3) | |
| CPB time, min | 247 (164, 391) | 237 (151.5, 362) | 273 (173, 421) | 0.182 |
| Aortic cross-clamp time, min | 116 (91, 183.2) | 117.5 (77.3, 178.8) | 116 (94, 186.5) | 0.307 |
| Operative time, min | 456.5 (306.7, 691.8) | 430 (273.5, 673) | 476 (358.8, 739.3) | 0.217 |
Bold P indicates a statistically significant difference (P < 0.05)
BMI, body mass index; COPD, chronic obstructive pulmonary disease; LEVDD, left ventricular end-diastolic dimension; LVEF, left ventricular ejection fraction; RDW, red blood cell distribution width; BUN, blood urea nitrogen; LDH, lactate dehydrogenase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; TBil, total bilirubin; DBil, direct bilirubin; NLR, neutrophil-to-lymphocyte ratio; aPTT, activated partial thromboplastin time; FDP, fibrin degradation products; INR, international normalized ratio; CABG, coronary artery bypass grafting; CPB, cardiopulmonary bypass
Among all involved patients, fifty-four patients (50.9%) received post-cardiotomy V-A ECMO support due to difficult weaning from CPB and 52 patients (49.1%) due to post-cardiotomy cardiogenic shock or cardiac arrest (Table 2). Seventeen patients (16%) had a central arterial cannulation site. Within 6 h after ECMO initiation, the levels of lactate (12.7 vs. 6.5 mmol/L; p < 0.001) and LDH (991 vs. 656 IU/L; p = 0.001) were significantly higher in non-survivors, whereas the levels of hemoglobin (84.5 vs. 92 g/L; p < 0.001), RBCs (2.8 vs. 3.2 × 1012/L; p = 0.020) and fibrinogen (1.6 vs. 2.3 g/L; p < 0.001) were significantly lower in non-survivors. No significant difference was observed in RDW levels within 6 h after ECMO initiation. The peak values of RDW during ECMO support were significantly higher in non-survivors (17.2 vs. 16.2%; p < 0.001). The maximum values of lactate (16.1 vs. 12 mmol/L; p = 0.002), LDH (1785 vs. 741.5 IU/L; p < 0.001), aspartate aminotransferase (AST, 696.5 vs. 319 IU/L; p = 0.011), alanine aminotransferase (ALT, 304.5 vs. 96.5; p = 0.008), neutrophil-to-lymphocyte ratio (NLR, 32.7 vs. 26.7; p < 0.001) occurring within ± 12 hours of the RDW peak levels were significantly higher in non-survivors. Peak levels of total and direct bilirubin also tended to be higher in non-survivors, although these differences did not reach statistical significance (p = 0.082 and p = 0.055, respectively). The lowest values of hemoglobin (64 vs. 71 g/L; p < 0.001) and RBCs (2.1 vs. 2.4 × 1012/L; p < 0.001) were significantly lower in non-survivors. RDW levels increased over the course of ECMO support in both groups. From day four after ECMO initiation, survivors exhibited significantly lower RDW levels than non-survivors, with this difference persisting throughout the remainder duration of ECMO support (Fig. 2).
Table 2.
Details on V-A ECMO settings and laboratory tests
| Total population (N = 106) |
Survivors (N = 48) |
Non-Survivors (N = 58) |
P | |
|---|---|---|---|---|
| ECMO indications (%) | 0.338 | |||
| Difficult weaning from CPB | 54 (50.9) | 22 (45.8) | 32 (55.2) | |
| Post-cardiotomy shock or cardiac arrest | 52 (49.1) | 26 (54.2) | 26 (44.8) | |
| Central arterial V-A ECMO | 17 (16) | 7 (14.6) | 10 (17.2) | 0.710 |
| ECMO settings | ||||
| Initial flow rate, mL/min | 3100 (2800, 3225) | 3100 (3000, 3337) | 3000 (2675, 3200) | 0.262 |
| Initial speed, rpm | 3000 (2812, 3200) | 3035 (2900,3200) | 2955 (2800, 3103) | 0.068 |
| Maximal flow rate, mL/min | 3400 (3200, 3617.5) | 3400 (3300, 3577) | 3425 (3200, 3763) | 0.967 |
| Maximal speed, rpm | 3200 (3100, 3300) | 3200 (3120, 3300) | 3200 (3098, 3302) | 0.626 |
| Laboratory tests within 6 h after ECMO initiation | ||||
| Hemoglobin, g/L | 88 (76.8, 106) | 92 (84.3, 115) | 84.5 (72, 95.5) | < 0.001 |
| RBCs, ×1012/L | 2.9 (2.4, 3.5) | 3.2 (2.6, 3.7) | 2.8 (2.3, 3.2) | 0.020 |
| RDW, % | 14.6 (14.1, 15.6) | 14.6 (13.8, 15.6) | 14.8 (14.2, 15.7) | 0.256 |
| Lactate, mmol/L | 9.9 (6.3, 14) | 6.5 (3.9, 9.9) | 12.7 (9.1, 17) | < 0.001 |
| Creatinine, umol/L | 123 (90.8, 153.8) | 111.8 (76.2, 153.8) | 128.6 (97.5, 157.4) | 0.183 |
| BUN, mmol/L | 10.5 (7.9, 14.9) | 10.2 (6.9, 14) | 12 (8.2, 17.5) | 0.120 |
| LDH, IU/L | 823 (568.5, 1484) | 656 (390.5, 1075.8) | 991 (624.7, 2072.8) | 0.001 |
| AST, IU/L | 160 (89.7, 474.3) | 145.5 (73.5, 332.7) | 194 (96.7, 518) | 0.202 |
| ALT, IU/L | 61.5 (20, 198.7) | 45.5 (19.3, 162.8) | 71.5 (21.5, 268.8) | 0.219 |
| TBil, umol/L | 33.9 (23.1, 51.6) | 36.9 (22.7, 54.1) | 33.2 (23.1, 48.8) | 0.621 |
| DBil, umol/L | 13.7 (8.3, 24.9) | 13.8 (8.1, 26.2) | 13.4 (8.4, 24.2) | 0.889 |
| NLR | 15.2 (11.2, 19.6) | 14.7 (8.9, 20.4) | 15.3 (12.5, 19.4) | 0.489 |
| D-dimer, ug/ml | 2.72 (0.9, 7.4) | 1.92 (0.8, 6.6) | 2.9 (1.1, 7.8) | 0.149 |
| aPTT, s | 86 (60.6, 159.1) | 82 (57.4, 119) | 91.8 (63.1, 180) | 0.169 |
| Fibrinogen, g/L | 1.8 (1.4, 2.5) | 2.3 (1.6, 3.2) | 1.6 (1.3, 1.9) | < 0.001 |
| FDP, ug/ml | 8.1 (2.9, 23.6) | 6.5 (2.7, 18.2) | 8.9 (3.1, 29.1) | 0.220 |
| INR, R | 1.7 (1.5, 2.3) | 1.7 (1.5, 2.2) | 1.9 (1.5, 2.7) | 0.493 |
| Peak values of RDW during ECMO support, % | 16.6 (15.9, 18.1) | 16.2 (15.6, 17.2) | 17.2 (16.1, 18.8) | < 0.001 |
| Peak values of laboratory tests within ± 12 h of the RDW peak levels | ||||
| Lactate, mmol/L | 13.4 (8.5, 19.8 | 12 (7.2, 15.3) | 16.1 (10.5, 22.4) | 0.002 |
| Creatinine, umol/L | 169.7 (134.6, 210.9) | 157.8 (127.6, 195.8) | 172.8 (143, 233) | 0.108 |
| BUN, mmol/L | 20 (12.3, 28.5) | 18.8 (12.4, 30.8) | 20.6 (12.2, 28.4) | 0.990 |
| LDH, IU/L | 1155.5 (652.8, 2584.5) | 741.5 (553.3, 1281) | 1785 (851, 3754) | < 0.001 |
| AST, IU/L | 380 (170.5, 1593) | 319 (119, 856.5) | 696.5 (204.5, 3138) | 0.011 |
| ALT, IU/L | 158 (49, 927) | 96.5 (44, 352.3) | 304.5 (70, 1547) | 0.008 |
| TBil, umol/L | 87.7 (48.7, 149.8) | 82.8 (37.9, 135.7) | 110 (57.3, 196) | 0.082 |
| DBil, umol/L | 53.6 (25.5, 97.5) | 44.8 (18.6, 83.6) | 71 (32.8, 127.4) | 0.055 |
| NLR | 28.9 (24.4, 36.6) | 26.7 (22.8, 30.7) | 32.7 (26.6, 38.9) | < 0.001 |
| D-dimer, ug/ml | 12.3 (6.3, 20) | 11.7 (6.1, 20) | 12.8 (6.3, 20) | 0.863 |
| Fibrinogen, g/L | 4.8 (3.9, 6.1) | 4.9 (3.9, 5.8) | 4.7 (3.8, 6.3) | 0.829 |
| FDP, ug/ml | 46.5 (23.1, 100.7) | 47.8 (21.6, 101.5) | 38.7 (23.6, 94.1) | 0.992 |
| Lowest values of laboratory tests within ± 12 h of the RDW peak levels | ||||
| Hemoglobin, g/L | 65.5 (60.7, 75) | 71 (64, 80) | 64 (57.7, 68) | < 0.001 |
| RBCs, ×1012/L | 2.1 (2, 2.5) | 2.4 (2.1, 2.7) | 2.1 (1.9, 68) | < 0.001 |
Bold P indicates a statistically significant difference (P < 0.05)
V-A ECMO, veno-arterial extracorporeal membrane oxygenation; CPB, cardiopulmonary bypass; RBCs, red blood cells; RDW, red blood cell distribution width; BUN, blood urea nitrogen; LDH, lactate dehydrogenase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; TBil, total bilirubin; DBil, direct bilirubin; NLR, neutrophil-to-lymphocyte ratio; aPTT, activated partial thromboplastin time; FDP, fibrin degradation products; INR, international normalized ratio
Fig. 2.
RDW levels in survivors and non-survivors during ECMO support. ECMO, extracorporeal membrane oxygenation; RDW, red blood cell distribution width * p < 0.05, ** p < 0.01
Outcomes and multivariable logistic regression analysis
Significantly more fresh frozen plasma (FFP) was transfused in non-survivors (2000 vs. 1600 mL; p = 0.033). Non-survivors showed an increased use of CRRT (84.5% vs. 60.4%; p = 0.005) and longer postoperative ventilation time (226 vs. 278.5 h; p = 0.023). No significant difference was found in duration of ECMO between the two groups (149.8 vs. 140.4 h; p = 0.343). Non-survivors exhibited significantly shorter ICU length of stay (9.5 vs. 20.5 days; p < 0.001) and hospital length of stay (17.5 vs. 32 days; p < 0.001, Table 3).
Table 3.
Outcomes and complications of the study population
| Total population (N = 106) |
Survivors (N = 48) |
Non-Survivors (N = 58) |
P | |
|---|---|---|---|---|
| Blood product transfusion during ECMO support | ||||
| RBCs, units | 18 (12, 27.1) | 18 (10, 24) | 19 (13.5, 28.4) | 0.138 |
| Platelet, units | 3.5 (1, 5) | 3 (2.5, 4.8) | 4 (2, 5.5) | 0.064 |
| FFP, mL | 1800 (987.5, 2450) | 1600 (600, 2400) | 2000 (1200, 2800) | 0.033 |
| Bleeding complications during ECMO support, % | 51 (48.1) | 21 (43.8) | 30 (51.7) | 0.413 |
| Complications during hospitalization | ||||
| IABP, % | 69 (65.1) | 32 (66.7) | 37 (63.8) | 0.757 |
| CRRT, % | 78 (73.6) | 29 (60.4) | 49 (84.5) | 0.005 |
| Pulmonary infection, % | 34 (32.1) | 14 (29.2) | 20 (34.5) | 0.559 |
| Limb complications, % | 13 (12.3) | 7 (14.6) | 6 (10.3) | 0.508 |
| Cerebral vascular accident, % | 8 (7.5) | 4 (8.3) | 4 (6.9) | 1.000 |
| Reoperation, % | 63 (59.4) | 24 (50) | 39 (67.2) | 0.072 |
| Duration of ECMO, h | 141.8 (107.8, 198.8) | 140.4 (111, 169.5) | 149.8 (94.5, 241.4) | 0.343 |
| ECMO weaning, % | 68 (64.2) | 48 (100) | 20 (34.5) | < 0.001 |
| Ventilation time, h | 249.5 (188, 410.5) | 278.5 (206.3, 496) | 226 (133.5, 365.5) | 0.023 |
| ICU length of stay, days | 13 (8, 21.2) | 20.5 (12, 30.8) | 9.5 (6, 15) | < 0.001 |
| Hospital length of stay, days | 21.3 (15, 35) | 32 (21.3, 42) | 17.5 (11, 23) | < 0.001 |
Bold P indicates a statistically significant difference (P < 0.05)
ECMO, extracorporeal membrane oxygenation; RBCs, red blood cells; FFP, fresh frozen plasma; IABP, intra-aortic balloon pump; CRRT, continuous renal replacement therapy; ICU, intensive care unit
Based on multivariable logistic regression analysis, increased peak values of RDW during ECMO support were independently associated with higher risk of in-hospital mortality (aOR, 2.097; 95% CI, 1.328–3.312; p = 0.001, Table 4). Confounding variables included sex, lactate, LDH, hemoglobin, and fibrinogen within 6 h after V-A ECMO initiation.
Table 4.
Univariable and multivariable logistic regression analysis for in-hospital mortality
| Covariates | Univariable | Multivariable | ||
|---|---|---|---|---|
| OR (95% CI) | P | aOR (95% CI) | P | |
| Male sex | 0.384 (0.171–0.862) | 0.020 | ||
| Lactate within 6 h after ECMO initiation | 1.276 (1.149–1.416) | < 0.001 | ||
| LDH within 6 h after ECMO initiation | 1.000 (1.000–1.001.000.001) | 0.020 | 1.000 (1.000–1.001.000.001) | 0.043 |
| Hemoglobin within 6 h after ECMO initiation | 0.959 (0.937–0.982) | < 0.001 | ||
| Fibrinogen within 6 h after ECMO initiation | 0.375 (0.216–0.648) | < 0.001 | 0.217 (0.083–0.566) | 0.002 |
| Peak values of RDW during ECMO support | 1.694 (1.252–2.293) | < 0.001 | 2.097 (1.328–3.312) | 0.001 |
ECMO, extracorporeal membrane oxygenation; LDH, lactate dehydrogenase; RDW, red blood cell distribution width; aOR, adjusted odds ratio; OR, odds ratio; CI, confidential interval
Correlation between peak values of RDW and related factors
Peak values of RDW were negatively correlated with lowest values of hemoglobin occurring within ± 12 h of the RDW peak levels (Spearman rs = − 0.248, p = 0.010; weak correlation). It was positively correlated with peak values of LDH (Spearman rs = 0.369, p < 0.001; weak correlation), direct bilirubin (DBil, Spearman rs = 0.389, p < 0.001; weak correlation), total bilirubin (TBil, Spearman rs = 0.355, p < 0.001; weak correlation), NLR (Spearman rs = 0.556, p < 0.001; moderate correlation) and fibrinogen degradation products (FDP, Spearman rs = 0.264, p = 0.007; weak correlation). The peak values of ALT (Spearman rs = 0.167, p = 0.087), AST (Spearman rs = 0.129, p = 0.187), lactate (Spearman rs = 0.077, p = 0.447) and D-dimer (Spearman rs = 0.117, p = 0.348) had no correlation with peak values of RDW (Fig. 3).
Fig. 3.
Correlation between peak values of RDW and laboratory parameters: A lowest values of hemoglobin; B peak values of LDH; C peak values of ALT; D peak values of AST; E peak values of DBil; F peak values of TBil; G peak values of lactate; H peak values of NLR; I peak values of FDP; J peak values of D-dimer.RDW, red blood cell distribution width; LDH, lactate dehydrogenase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; TBil, total bilirubin; DBil, direct bilirubin; NLR, neutrophil-to-lymphocyte ratio; FDP, fibrin degradation products
Discussion
In this single-center, retrospective study, we reviewed 106 adult patients receiving post-cardiotomy V-A ECMO support and investigated the relationship between the peak values of RDW during V-A ECMO support and in-hospital mortality. The increased peak values of RDW were associated with an increased risk of in-hospital mortality. Elevated peak RDW levels may be indicative of underlying hemolysis, systemic inflammatory responses, and coagulation dysfunction. As an easily obtainable laboratory indicator, the peak RDW levels during ECMO support could provide clinicians with valuable prognostic information, aiding in the optimization of treatment strategies and management.
Previous studies have shown that RDW at admission serves as an independent predictor of major adverse complications and long-term mortality in various clinical settings, including patients with heart failure, those undergoing cardiac rehabilitation, noncardiac surgery, and individuals receiving V-A ECMO support [10–16]. In this study, we evaluated the association between peak values of RDW during ECMO support and short-term prognosis, specifically focusing on patients who received post-cardiotomy V-A ECMO. These patients frequently encounter unique physiological challenges, including recent major surgery, related tissue trauma as well as dilution coagulopathy, shear stress-related blood disruption, and inflammation-coagulation-fibrinolysis system activation caused by prior CPB [17, 18]. Subsequent ECMO-induced hemolysis, repeated blood transfusions and inflammation may directly contribute to erythrocyte heterogeneity and affect their overall recovery and survival rates. Our findings extend prior researches by demonstrating that dynamic RDW monitoring during ECMO—rather than a single baseline measurement—provides critical prognostic value.
The exact biological mechanisms underlying the association between higher peak value of RDW and mortality are currently unclear. Several potential mechanisms including hemolysis, RBCs transfusion, inflammation, oxidative stress and ischemia-perfusion injury can be considered in cardiac surgery patients [19–21]. Previous studies suggested oxidative stress contributes to increased anisocytosis by disrupting erythropoiesis and impairing RBCs membrane deformability, thereby shortening the lifespan of circulating erythrocytes and ultimately leading to elevated RDW levels [22, 23]. In our study, we explored the correlations between peak values of RDW and several clinical parameters, and found that peak RDW levels were positively associated with the peak values of indirect hemolysis biomarkers, including LDH, TBil and DBil, as well as the systemic inflammation biomarker NLR. These findings suggested that hemolysis and inflammation may contribute to RDW elevation during ECMO support and may partially explain its association with poor outcomes. Notably, markers of oxidative stress and direct hemolysis, such as cell-free hemoglobin, were not available in our dataset. Incorporating these biomarkers in future studies may provide a more comprehensive understanding of the biological pathways linking RDW to adverse outcomes.
This study found that there was no significant difference in RDW levels between survivors and non-survivors during the early stage of post-cardiotomy V-A ECMO support, especially at the first 3 days. A statistically significant difference in RDW levels was observed beginning on day four. This delayed difference can be explained by the distinct pathophysiological phases of ECMO support. In the immediate post-operative and early stage of ECMO, all patients, regardless of ultimate outcome, experience profound acute physiological disturbances—including systemic inflammation, mechanical hemolysis, and stress responses—which can cause a universal, non-specific elevation in RDW levels. With the duration of ECMO support increasing, survivors may begin to exhibit signs of physiological recovery and stabilization, allowing their RDW levels to plateau or slightly increase. In contrast, non-survivors are likely subjected to persistent or cumulative injuries, including ongoing tissue hypoperfusion, bone marrow suppression, or dysregulated inflammation, which drive a continued rise in RDW levels. Therefore, RDW, as a dynamic biomarker integrating cumulative physiological stress over time, may hold potential prognostic value in the later stages of ECMO support and warrants close monitoring and clinical attention.
RDW represents a valuable and cost-effective laboratory parameter that provides valuable prognostic information. Specifically, serial RDW measurements could help clinicians identify a subgroup of high-risk patients who, despite surviving the initial critical phase, exhibit a trajectory of increasing physiological stress and poor recovery potential. A persistent rise in RDW beyond day 3–4 could trigger a more intensive diagnostic workup to address potential sources of ongoing hemolysis, inflammation, or organ dysfunction. Future research should explore the incorporation of serial RDW measurements into multimodal prognostic algorithms. The combination of RDW with other dynamic biomarkers may improve risk stratification and potentially guide individualized therapy.
The authors acknowledge some limitations. First, this study is a single-center retrospective analysis, and may not fully represent the broader population; therefore, future validation in multicenter studies is needed. Second, due to the retrospective observational nature of the study, the possibility of unmeasured confounding persists despite our efforts to adjust for known variables. Third, as VA-ECMO initiation is often performed as an emergency procedure, potential data bias cannot be completely excluded despite the use of multivariable analysis. Fourth, the absence of oxidative stress markers and free hemoglobin measurements limits our ability to further elucidate the biological mechanisms underlying the association between RDW and mortality.
Conclusion
Our findings indicate that an increase in the peak values of RDW may be associated with higher in-hospital mortality in patients with post-cardiotomy V-A ECMO. Potential factors related to the increased peak values of RDW during V-A ECMO support include hemolysis and systemic inflammatory responses. The prognostic value of RDW for mortality and potential mechanisms in these high-risk patients should be further examined in future researches.
Acknowledgements
Not applicable.
Abbreviations
- aPTT
activated partial thromboplastin time
- aOR
adjusted odds ratio
- AST
Aspartate aminotransferase
- ALT
Alanine aminotransferase
- BMI
Body mass index
- BUN
Blood urea nitrogen
- CS
Cardiogenic shock
- CPB
Cardiopulmonary bypass
- CRRT
Continuous renal replacement therapy
- CI
Confidence intervals
- COPD
Chronic obstructive pulmonary disease
- DBil
Direct bilirubin
- FDP
Fibrinogen degradation products
- FFP
Fresh frozen plasma
- IABP
Intra-aortic balloon pump
- ICU
Intensive care unit
- INR
International normalized ratio
- LDH
Lactate dehydrogenase
- LEVDD
Left ventricular end-diastolic dimension
- LVEF
Left ventricular ejection fraction
- NLR
Neutrophil-to-lymphocyte ratio
- ORs
Odds ratios
- RDW
Red blood cell distribution width
- RBCs
Red blood cells
- SD
Standard deviation
- TBil
Total bilirubin
- V-A ECMO
Veno-arterial extracorporeal membrane oxygenation
- VIF
Variance inflation factor
Author contributions
HZ, JW and BYJ designed the study, extracted the data and drafted the manuscript; GL, SJY and LYB collected data; TLW, YT, JW and JRZ performed the statistical analysis. JRZ and ZZL prepared Figs. 1, 2 and 3. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.
Funding
The authors have no funding sources to disclose.
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval and consent to participate
This study was approved by the ethical committee of Fuwai Hospital (No. 2024–2517). Given the observational nature of the study, individual patient consent was waived by the Ethics Committee.
Consent for publication
All the authors consent to the publication of the manuscript and support material.
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.
Han Zhang and Jing Wang contributed equally.
References
- 1.Djordjevic I, Eghbalzadeh K, Sabashnikov A, et al. Central vs peripheral venoarterial ECMO in postcardiotomy cardiogenic shock. J Card Surg. 2020;35:1037–42. 10.1111/jocs.14526. [DOI] [PubMed] [Google Scholar]
- 2.Mariani S, Heuts S, van Bussel BCT, et al. Patient and management variables associated with survival after postcardiotomy extracorporeal membrane oxygenation in adults: the PELS-1 multicenter cohort study. J Am Heart Assoc. 2023;12:e029609. 10.1161/jaha.123.029609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chen SW, Tsai FC, Lin YS, et al. Long-term outcomes of extracorporeal membrane oxygenation support for postcardiotomy shock. J Thorac Cardiovasc Surg. 2017;154:469–77. 10.1016/j.jtcvs.2017.02.055. [DOI] [PubMed] [Google Scholar]
- 4.Rastan AJ, Dege A, Mohr M, et al. Early and late outcomes of 517 consecutive adult patients treated with extracorporeal membrane oxygenation for refractory postcardiotomy cardiogenic shock. J Thorac Cardiovasc Surg. 2010;139:302–11. 10.1016/j.jtcvs.2009.10.043. [DOI] [PubMed] [Google Scholar]
- 5.Biancari F, Dalén M, Fiore A, et al. Multicenter study on postcardiotomy venoarterial extracorporeal membrane oxygenation. J Thorac Cardiovasc Surg. 2020;159:1844–54. 10.1016/j.jtcvs.2019.06.039. [DOI] [PubMed] [Google Scholar]
- 6.Wan J, Luo P, Du X, Yan H. Preoperative red cell distribution width predicts postoperative cognitive dysfunction after coronary artery bypass grafting. Biosci Rep. 2020. 10.1042/bsr20194448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Giamouzis G, Kourek C, Magouliotis DE, et al. The prognostic role of RDW in hospitalized heart failure patients with and without chronic kidney disease. J Clin Med. 2024. 10.3390/jcm13237395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ruan L, Zhu L, Su L, et al. Better prognosis in surgical aortic valve replacement patients with lower red cell distribution width: A MIMIC-IV database study. PLoS ONE. 2024;19:e0306258. 10.1371/journal.pone.0306258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang J, Gao S, Wang T, et al. First-day platelet count is associated with in-hospital mortality in adult postcardiotomy extracorporeal membrane oxygenation. ASAIO J. 2025. 10.1097/mat.0000000000002393. [DOI] [PubMed] [Google Scholar]
- 10.Isik T, Kurt M, Tanboga IH, et al. The impact of admission red cell distribution width on long-term cardiovascular events after primary percutaneous intervention: a four-year prospective study. Cardiol J. 2016;23:281–8. 10.5603/CJ.a2015.0080. [DOI] [PubMed] [Google Scholar]
- 11.Liu C, Zhang K, Zhang T, et al. Higher preoperative red blood cell distribution width increases the risk of myocardial injury after noncardiac surgery in advanced-age patients: a retrospective cohort study. Clin Interv Aging. 2023;18:169–79. 10.2147/cia.S392778. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zhang Q, Zhou B, Li X, Cong H. In-hospital changes in the red blood cell distribution width and mortality in critically ill patients with heart failure. ESC Heart Fail. 2023;10:3287–98. 10.1002/ehf2.14513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lu SY, Ortoleva J, Colon K, et al. Red blood cell distribution width predicts mortality of adult patients receiving veno-arterial extracorporeal membrane oxygenation. Perfusion. 2024;39:935–42. 10.1177/02676591231169850. [DOI] [PubMed] [Google Scholar]
- 14.Lazzeroni D, Moderato L, Marazzi PL, et al. Red blood cell distribution width as a novel prognostic marker after myocardial revascularization or cardiac valve surgery. Sci Rep. 2021;11:7889. 10.1038/s41598-021-87075-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Xie KH, Liu LL, Liang YR, et al. Red cell distribution width: a novel predictive biomarker for stroke risk after transient ischaemic attack. Ann Med. 2022;54:1167–77. 10.1080/07853890.2022.2059558. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Skjelbakken T, Lappegård J, Ellingsen TS, et al. Red cell distribution width is associated with incident myocardial infarction in a general population: the Tromsø study. J Am Heart Assoc. 2014. 10.1161/jaha.114.001109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schaefer AK, Latus M, Riebandt J, et al. Bleeding and thrombotic events in post-cardiotomy extracorporeal life support. Eur J Cardiothorac Surg. 2023. 10.1093/ejcts/ezad072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jin Y, Cui Y, Zhang Y, et al. Hemostatic complications and systemic heparinization in pediatric post-cardiotomy veno-arterial extracorporeal membrane oxygenation failed to wean from cardiopulmonary bypass. Transl Pediatr. 2022;11:1458–69. 10.21037/tp-22-104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wang XD, Zhao ZZ, Yang XY, et al. Association between red cell distribution width and liver injury after cardiac and aortic aneurysm surgery with cardiopulmonary bypass. J Cardiothorac Vasc Anesth. 2024;38:3065–75. 10.1053/j.jvca.2024.09.004. [DOI] [PubMed] [Google Scholar]
- 20.Chen W, Yu P, Chen C, et al. Association between the red blood cell distribution width and 30-day mortality in intensive care patients undergoing cardiac surgery: a retrospective observational study based on the medical information mart for intensive care-IV database. Ann Lab Med. 2024;44:401–9. 10.3343/alm.2023.0345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.von Meijenfeldt GCI, van der Laan MJ, Zeebregts C, Christopher KB. Red cell distribution width at hospital discharge and out-of hospital outcomes in critically ill non-cardiac vascular surgery patients. PLoS One. 2018;13:e0199654. 10.1371/journal.pone.0199654. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zou Z, Zhuang Y, Liu L, et al. Role of elevated red cell distribution width on acute kidney injury patients after cardiac surgery. BMC Cardiovasc Disord. 2018;18:166. 10.1186/s12872-018-0903-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hunziker S, Celi LA, Lee J, Howell MD. Red cell distribution width improves the simplified acute physiology score for risk prediction in unselected critically ill patients. Crit Care. 2012;16:R89. 10.1186/cc11351. [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.
Data Availability Statement
No datasets were generated or analysed during the current study.