Abstract
Red blood cell distribution width (RDW) to human serum albumin (ALB) ratio (RDW/ALB Ratio, RAR) is a prognostic factor for adverse outcomes in different disease populations. However, the relationship between RAR and pulmonary embolism outcomes remains unclear. Therefore, this study set out to investigate the association between RAR and the risk of all-cause death in acute pulmonary embolism (APE) patients admitted to the intensive care unit (ICU). This is a retrospective study based on the MIMIC-IV database. The primary outcome was all-cause mortality among patients with APE (in-hospital and 1-year mortality). The relationship between RAR and all-cause mortality was assessed using Cox regression analysis. The survival curve was drawn to evaluate the predictive value of RAR for patient mortality. Correlations and threshold effects between RAR and all-cause mortality were analyzed using the generalized additive model (GAM). The study included 773 patients, and fully adjusted Cox regression models showed that RAR was associated with higher all-cause mortality in the hospital and one year later (all P < .05). In the GAM, the relationship between RAR and all-cause mortality was shown to be nonlinear, with a positive association between RAR and all-cause mortality in APE patients when RAR values were at low to moderate levels. This study revealed a significant association between RAR and the risk of all-cause day death in patients with pulmonary embolism. Higher RAR value was associated with increased in-hospital mortality and 1-year mortality.
Keywords: ALB, all-cause mortality, MIMIC IV, pulmonary embolism, RDW
1. Introduction
Acute pulmonary embolism (APE) is a severe disease with high morbidity and mortality, with an annual incidence of 0.039% to 0.115% and a mortality rate of 11.4% within 14 days. It is the third most lethal cardiovascular disease after acute myocardial infarction and stroke.[1,2] Accurate assessment and risk stratification of patients with APE can help clinicians make optimal treatment decisions. However, there are few studies on the risk stratification of APE patients in the intensive care unit (ICU), and risk management strategies for APE patients in the ICU are rarely mentioned in existing guidelines.[3]
According to relevant studies, the pulmonary embolism severity index (PESI) is the most widely used clinical method to predict 30-day mortality in APE patients. However, the PESI score has many variables with different weights and is complicated to calculate, so it is not well accepted among emergency medical personnel.[4] Therefore, the simplified pulmonary embolism severity index has been developed to predict the risk of death in APE patients.[5] However, the advantages of PESI and simplified pulmonary embolism severity index mainly lie in identifying patients with low to moderate-risk APE, but its ability to identify patients with moderate to high-risk APE is weak.[4] Therefore, there is an urgent need for some simple and reliable methods to assist clinicians in the early assessment of APE patients’ conditions in the ICU and to improve the accuracy of existing prognosis predictions of severe patients.
RAR has been reported as a significant predictor of prognosis in patients with severe acute pancreatitis and type 2 diabetic foot ulcers.[6,7] It has also been shown that RAR is a more accurate predictor of prognosis than RDW alone in severe patients with acute kidney injury.[8] However, the clinical application of RAR in patients with APE has not been evaluated. Therefore, this study aimed to investigate the relationship between RAR and the prognosis of severe APE patients.
2. Methods
2.1. Study design and data source
This cohort study collected data on 773 patients with pulmonary embolism from the MIMIC-IV database. MIMIC-IV is a publicly available clinical database managed by the Massachusetts Institute of Technology. It was officially released in 2020 and collected information from over 70,000 ICU patients at Beth Israel Deaconess Medical Center between 2008 and 2019. It includes demographic characteristics, basic vital signs records, medical intervention records, nursing records, imaging results, discharge summaries, and other medical data.[9] The authors of this article have been granted a license to use the database (certificate number: 55303142).
2.2. Inclusion and exclusion criteria
Firstly, we included information on patients diagnosed with pulmonary embolism during ICU admission according to codes from the International Classification of Diseases, Ninth Revision (ICD-9), and Tenth Revision (ICD-10) (ICD-9: 41511, 41512, 41519, 67380, 67381, 67382, 67383, 67384; ICD-10: 126, 1260, 12601, 12609, 1269, 12690, 12693, 12699). Secondly, the first admission was included only for patients with repeated admissions, and 1376 pulmonary embolism patients’ information was included. Finally, patients who were not hospitalized for more than 24 hours or did not provide RDW or ALB data in medical records were excluded. All included patients were aged ≥ 18. All-cause mortality in the hospital and one year after admission were used as the primary endpoints. The screening process is shown in Figure 1.
Figure 1.
The flowchart of patients’ selection.
2.3. Data extraction
Patients were evenly divided into 3 groups and then further categorized into the Low RAR group, Middle RAR group, and High RAR group according to their RAR levels. The mean and standard deviation of continuous variables were expressed, while the frequency and percentage of categorical variables were expressed. The means and percentages of the groups were compared using one-way ANOVA, Fisher exact test, Kruskal-Wallis H test, and chi-square test. We evaluated the differences in all-cause mortality rates between patient groups using Kaplan–Meier (KM) methodology. Four different models were used, including the unadjusted model and three other models with increments of adjusted factors. Using a generalized additive model (GAM), the nonlinear relationships were also identified. In the case of nonlinear relationships, a two-segment linear regression model was employed to calculate the threshold effect of RAR on APE mortality. When the ratio between RAR and APE mortality changed significantly in the GAM, the effect values of the inflection points and values before and after these points were calculated. Additionally, interaction and stratified analyses were performed for age, gender, hypertension, diabetes, chronic lung disease, chronic renal disease (CKD), liver disease, malignancy, and obesity. In all analyses, the P value of 0.05 was considered significant, and the analysis was performed using package R (version 4.3.2, http://www.Rproject.org) and EmpowerStats software (version 4.0, http://www.empowerstats.com).
3. Results
3.1. Patient and hospital characteristics
There were 773 APE patients included in the study. Table 1 shows the baseline characteristics. They were categorized into 3 groups based on RAR levels, including 258 patients in the Low RAR group (2.68–4.71), 258 in the Middle RAR group (4.71–6.09), and 258 in the High RAR group (6.09–15.45). The overall mean age of patients was 61.43 ± 16.91, and approximately 52.59% were male. A total of 137 patients died in the hospital, and a total of 287 patients died one year after treatment. Middle and High RAR groups had significantly higher heart rhythm, respiratory rate, blood urea nitrogen, Cr, white blood cells, RDW, alanine aminotransferase, and significantly lower systolic pressure, ALB, hemoglobin, and platelet than the Low RAR group. Meanwhile, the Middle and High RAR groups also had a higher prevalence rate of liver disease and malignancy. There was no significant difference in the prevalence of hypertension, diabetes, chronic lung disease, coronary atherosclerotic disease (CAD), congestive heart failure, CKD, obesity, age, gender, race, blood oxygen saturation, body temperature, Na + (Sodium), K + (Kalium), fasting blood glucose and aspartate transaminase among different RAR groups.
Table 1.
Baseline characteristics of participants.
| RDW/Alb ratio | All | Low | Middle | High | P value | P value* |
|---|---|---|---|---|---|---|
| Number | 773 | 257 | 258 | 258 | ||
| RAR, mean (SD) | 5.70 (1.82) | 4.01 (0.46) | 5.38 (0.38) | 7.71 (1.62) | <.001 | <.001 |
| Age, yr | 61.43 (16.91) | 61.24 (16.31) | 61.72 (17.52) | 61.33 (16.93) | .942 | .769 |
| Gender, n(%) | .148 | - | ||||
| Male | 395 (51.10%) | 144 (56.03%) | 124 (48.06%) | 127 (49.22%) | ||
| Female | 378 (48.90%) | 113 (43.97%) | 134 (51.94%) | 131 (50.78%) | ||
| Race, n(%) | .898 | - | ||||
| White | 481 (62.23%) | 154 (59.92%) | 165 (63.95%) | 162 (62.79%) | ||
| Black | 93 (12.03%) | 33 (12.84%) | 28 (10.85%) | 32 (12.40%) | ||
| Asian | 23 (2.98%) | 6 (2.33%) | 8 (3.10%) | 9 (3.49%) | ||
| Other | 176 (22.77%) | 64 (24.90%) | 57 (22.09%) | 55 (21.32%) | ||
| Heart rate (beats/min), mean(SD) | 98.16 (22.05) | 91.69 (20.19) | 98.87 (21.68) | 103.90 (22.57) | <.001 | <.001 |
| Respiration rate (beats/min), mean(SD) | 21.84 (6.89) | 20.47 (6.31) | 22.10 (6.57) | 22.93 (7.53) | <.001 | <.001 |
| SBP (mm Hg), mean (SD) | 124.31 (23.48) | 129.55 (22.99) | 123.03 (22.60) | 120.36 (23.95) | <.001 | <.001 |
| SpO2(%), mean (SD) | 96.06 (4.30) | 96.30 (3.76) | 95.93 (4.37) | 95.95 (4.73) | .552 | .869 |
| Tempeature (°C), mean (SD) | 36.85 (0.80) | 36.84 (0.61) | 36.87 (0.85) | 36.84 (0.91) | .933 | .361 |
| Na + (mmol/L), mean (SD) | 138.11 (5.74) | 138.22 (4.51) | 138.34 (6.11) | 137.76 (6.42) | .482 | .075 |
| K + (mmol/L), mean (SD) | 4.10 (0.74) | 4.15 (0.70) | 4.12 (0.78) | 4.03 (0.74) | .191 | .207 |
| Glucose (mmol/L), mean (SD) | 8.02 (4.17) | 8.02 (3.67) | 8.17 (4.29) | 7.88 (4.52) | .725 | .436 |
| ALB(g/dL), mean (SD) | 2.86 (0.63) | 3.46 (0.42) | 2.85 (0.33) | 2.27 (0.44) | <.001 | <.001 |
| BUN (mg/dL), median (Q1,Q3) | 18.00 (12.00–28.00) | 16.00 (12.00-22.00) | 18.00 (12.00-26.00) | 22.50 (14.00-37.00) | <.001 | <.001 |
| Cr (mg/dL), median (Q1,Q3) | 0.90 (0.70-1.20) | 0.90 (0.70-1.10) | 0.90 (0.60-1.20) | 0.95 (0.70-1.40) | .114 | .032 |
| HGB (g/dL), mean(SD) | 10.86 (2.28) | 12.06 (1.99) | 10.82 (2.17) | 9.71 (2.04) | <.001 | <.001 |
| WBC (103/μL), median (Q1,Q3) | 11.40 (8.20-16.20) | 10.50 (8.17-14.60) | 11.70 (8.53-15.47) | 12.50 (8.15-17.88) | .259 | .044 |
| PLT (103/μL), median (Q1,Q3) | 204.00 (141.00–284.00) | 221.00 (169.25–281.00) | 195.00 (125.00–270.00) | 196.00 (124.75–297.00) | .408 | .01 |
| RDW(%), mean (SD) | 15.35 (2.31) | 13.76 (1.10) | 15.26 (1.70) | 17.01 (2.58) | <.001 | <.001 |
| ALT(IU/L), median (Q1,Q3) | 26.00 (15.00-53.00) | 27.00 (15.50-61.50) | 28.00 (16.00-48.00) | 21.00 (12.00-52.75) | .155 | .033 |
| AST(IU/L), median (Q1,Q3) | 34.00 (20.00-64.00) | 30.00 (21.00-61.25) | 37.00 (21.00-58.50) | 35.50 (19.00-71.00) | .325 | .576 |
| 3.85 (0.39) | 4.88 (0.27) | 5.93 (0.36) | 8.17 (1.56) | <.001 | <.001 | |
| Hypertension, n(%) | .41 | - | ||||
| No | 450 (58.21%) | 141 (54.86%) | 155 (60.08%) | 154 (59.69%) | ||
| Yes | 323 (41.79%) | 116 (45.14%) | 103 (39.92%) | 104 (40.31%) | ||
| Diabetes, n(%) | .299 | - | ||||
| No | 594 (76.84%) | 206 (80.16%) | 195 (75.58%) | 193 (74.81%) | ||
| Yes | 179 (23.16%) | 51 (19.84%) | 63 (24.42%) | 65 (25.19%) | ||
| Chronic lung disease, n(%) | .783 | - | ||||
| No | 715 (92.50%) | 237 (92.22%) | 237 (91.86%) | 241 (93.41%) | ||
| Yes | 58 (7.50%) | 20 (7.78%) | 21 (8.14%) | 17 (6.59%) | ||
| CAD, n(%) | .935 | - | ||||
| No | 722 (93.40%) | 239 (93.00%) | 241 (93.41%) | 242 (93.80%) | ||
| Yes | 51 (6.60%) | 18 (7.00%) | 17 (6.59%) | 16 (6.20%) | ||
| CHF, n(%) | .19 | - | ||||
| No | 687 (88.87%) | 228 (88.72%) | 223 (86.43%) | 236 (91.47%) | ||
| Yes | 86 (11.13%) | 29 (11.28%) | 35 (13.57%) | 22 (8.53%) | ||
| CKD, n(%) | .595 | - | ||||
| No | 681 (88.10%) | 229 (89.11%) | 223 (86.43%) | 229 (88.76%) | ||
| Yes | 92 (11.90%) | 28 (10.89%) | 35 (13.57%) | 29 (11.24%) | ||
| Liver disease, n(%) | <.001 | - | ||||
| No | 528 (68.31%) | 200 (77.82%) | 180 (69.77%) | 148 (57.36%) | ||
| Yes | 245 (31.69%) | 57 (22.18%) | 78 (30.23%) | 110 (42.64%) | ||
| Malignancy, n(%) | <.001 | - | ||||
| No | 515 (66.62%) | 188 (73.15%) | 182 (70.54%) | 145 (56.20%) | ||
| Yes | 258 (33.38%) | 69 (26.85%) | 76 (29.46%) | 113 (43.80%) | ||
| Obesity, n(%) | .499 | - | ||||
| No | 664 (85.90%) | 218 (84.82%) | 219 (84.88%) | 227 (87.98%) | ||
| Yes | 109 (14.10%) | 39 (15.18%) | 39 (15.12%) | 31 (12.02%) | ||
| In-hospital mortality | <.001 | - | ||||
| No | 636 (82.28%) | 232 (90.27%) | 205 (79.46%) | 199 (77.13%) | ||
| Yes | 137 (17.72%) | 25 (9.73%) | 53 (20.54%) | 59 (22.87%) | ||
| 1-yr all-cause mortality | <.001 | - | ||||
| No | 486 (62.87%) | 200 (77.82%) | 160 (62.02%) | 126 (48.84%) | ||
| Yes | 287 (37.13%) | 57 (22.18%) | 98 (37.98%) | 132 (51.16%) |
ALB = albumin, ALT = alanine aminotransferase, AST = aspartate transaminase, BUN = blood urea nitrogen, CAD = coronary atherosclerotic disease, CHF = congestive heart failure, CKD = completely knocked down, Cr = creatinine, HGB = hemoglobin, K+ = kalium, Lac = lactic acid, Na+ = Sodium, PLT = platelet, RAR = RDW/ALB ratio, RDW = red blood cell distribution width, SBP = systolic pressure, SpO2 = blood oxygen saturation, WBC = white blood cells.
3.2. Kaplan–Meier curves
We plotted Kaplan–Meier survival curves for patients (Fig. 2). Compared to the Low RAR group, the mortality rate of the middle and high RAR groups is significantly higher. However, there seems to be little difference in the KM curve for in-hospital mortality between the middle and high RAR level groups.
Figure 2.
Kaplan–Meier curves of the RDW/ALB ratio for mortality with PE. A high RDW/ALB ratio was significantly associated with higher mortality than a low RDW/ALB ratio (P < .001). Kaplan–Meier (KM) survival curves of In-hospital (A), 1-yr (B).
3.3. All-cause mortality of pulmonary embolism is associated with PDW/ALB ratio
We have established three adjustment models (Table 2). In Model 1 (adjusted for age, gender, and race) and Model 2 (adjusted for age, gender, race, hypertension, diabetes, chronic lung disease, CAD, congestive heart failure, CKD, liver disease, malignancy, and obesity), there was a statistically significant increase in the risk of all-cause in-hospital mortality and all-cause mortality within one year after admission in patients with middle to high RAR levels. In the fully adjusted model (Model 3), which adjusts for all other variables as well as those in Models 1 and 2 mentioned above, there was also a significant correlation between RAR level and increased risk of all-cause mortality of pulmonary embolism in the middle to high RAR groups during hospitalization or within one year after admission.
Table 2.
Hazard ratios (HRs) for all-cause mortality based on RDW/ALB ratios in pulmonary embolism patients.
| RDW/ALB ratio | Crude model | Model I | Model II | Model III | ||||
|---|---|---|---|---|---|---|---|---|
| HR(95%CI) | P | HR(95%CI) | P | HR(95%CI) | P | HR(95%CI) | P | |
| In-hospital mortality | ||||||||
| Low | Ref | Ref | Ref | Ref | ||||
| Middle | 2.25 (1.40, 3.62) | .0008 | 3.25 (1.70, 6.20) | .0003 | 2.16 (1.34, 3.48) | .0017 | 2.22 (1.38, 3.57) | .0011 |
| High | 2.54 (1.59, 4.05) | <.0001 | 2.17 (1.11, 4.24) | .0228 | 2.31 (1.43, 3.72) | .0006 | 2.58 (1.61, 4.12) | <.0001 |
| 1-yr all-cause mortality | ||||||||
| Low | Ref | Ref | Ref | Ref | ||||
| Middle | 1.91 (1.38, 2.65) | <.0001 | 1.90 (1.37, 2.64) | <.0001 | 1.88 (1.35, 2.62) | .0002 | 2.03 (1.32, 3.12) | .0012 |
| High | 2.78 (2.03, 3.79) | <.0001 | 2.90 (2.13, 3.97) | <.0001 | 2.50 (1.82, 3.44) | <.0001 | 2.15 (1.39, 3.33) | .0006 |
Crude model: we did not adjust other covariants.
Model I: we adjusted age gender and race.
Model II: we adjusted age, gender, race, hypertension, diabetes, chronic lung disease, CAD, CHF, CKD, liver disease, malignancy and obesity.
Model III: we adjusted age, gender, race, heart rate, respiration rate, SBP, SpO2, Tempeature, hypertension, diabetes, chronic lung disease, CAD, CHF, CKD, liver disease, malignancy, obesity, Na+, K+, glucose, BUN, Cr, HGB, WBC, PLT, ALT and AST.
ALB = albumin, ALT = alanine aminotransferase, AST = aspartate transaminase, BUN = blood urea nitrogen, CAD = coronary atherosclerotic disease, CHF = congestive heart failure, CI = confidence interval, CKD = completely knocked down, Cr = creatinine, HGB = hemoglobin, K+ = kalium, Lac = lactic acid, Na+ = Sodium, PLT = platelet, RAR = RDW/ALB ratio, RDW = red blood cell distribution width, Ref = reference, SBP = systolic pressure, SpO2 = blood oxygen saturation, WBC = white blood cells.
3.4. A nonlinear relationship analysis
Since RAR is a continuous variable, it is necessary to analyze it for its linear relationship with pulmonary embolism mortality (Fig. 3). In this study, we found that the relationship between RAR and all-cause mortality in pulmonary embolism was nonlinear (same adjustment as in Model 3). Based on a two-segment linear regression model, we determined that in-hospital and one year mortality inflection points were 5.19 and 5.46, respectively. The effect sizes (95% CI, P value) on the left side of the inflection point were 2.11 (1.34–3.33, 0.0013) and 1.81 (1.36–2.40, 0.001) respectively. On the right side of the inflection point, we observed no definite relationship between RAR and all-cause mortality due to pulmonary embolism with values of 0.93 (0.83–1.03, 0.1633) and 1.01 (0.91–1.11, 0.889), respectively; the results are shown in Table 3.
Figure 3.
The relationship between RAR and all-cause mortality of In-hospital (A), 1-yr (B). A nonlinear relationship between them was detected after adjusting for age, gender, race, heart rate, respiration rate, SBP, SpO2, temperature, hypertension, diabetes, chronic lung disease, CAD, CHF, CKD, liver disease, malignancy, obesity, Na+, K+, glucose, BUN, Cr, HGB, WBC, PLT, ALT, and AST. ALT = alanine aminotransferase, AST = aspartate transaminase, BUN = blood urea nitrogen, CAD = coronary atherosclerotic disease, CHF = congestive heart failure, CKD = completely knocked down, Cr = creatinine, HGB = hemoglobin, K+ = kalium, Na+ = Sodium, PLT = platelet, RAR = RDW/ALB ratio, SBP = systolic pressure, WBC = white blood cells.
Table 3.
The results of the two-piece wise linear regression model.
| RDW/ALB ratio | Inflection point | HR(95%CI) | P value |
|---|---|---|---|
| In-hospital mortality | <5.19 | 2.11 (1.34,3.33) | .0013 |
| ≥5.19 | 0.93 (0.83,1.03) | .1633 | |
| Likelihood ratio test | .003 | ||
| 1-yr all-cause mortality | |||
| <5.46 | 1.81 (1.36, 2.40) | <.0001 | |
| ≥5.46 | 1.01 (0.91, 1.11) | .889 | |
| Likelihood ratio test | <.001 |
Adjusted: age, gender, race, heart rate, respiration rate, SBP, SpO2, Tempeature, hypertension, diabetes, chronic lung disease, CAD, CHF, CKD, liver disease, malignancy, obesity, Na+, K+, glucose, BUN, Cr, HGB, WBC, PLT, ALT and AST.
ALB = albumin, ALT = alanine aminotransferase, AST = aspartate transaminase, BUN = blood urea nitrogen, CAD = coronary atherosclerotic disease, CHF = congestive heart failure, CI = confidence interval, CKD = completely knocked down, Cr = creatinine, HGB = hemoglobin, K+ = kalium, Lac = lactic acid, Na+ = Sodium, PLT = platelet, RAR = RDW/ALB ratio, RDW = red blood cell distribution width, Ref = reference, SBP = systolic pressure, SpO2 = blood oxygen saturation, WBC = white blood cells.
4. Discussion
The goal of this study was to clarify the relationship between RAR and all-cause mortality among patients with APE in the intensive care unit. According to our analysis, mortality was significantly higher in APE patients with middle to high RAR levels than in patients with low RAR levels. Based on the fully adjusted model, middle and high levels of RAR were definitively associated with all-cause mortality during hospitalization or within one year after admission. However, we also found a nonlinear relationship between RAR and all-cause mortality inAPE, with different correlations located on the left and right sides of the inflection points: a positive correlation on the left and no statistically significant correlation on the right. A continued increase in RAR when RAR was ≥ 5.19 (in-hospital mortality), 5.24 (1 year mortality) was not reflected in the trend in APE all-cause mortality. This is similar to what we found in the KM curve, where there seems to be little difference between the middle and high RAR level groups within one month of patient admission. Even so, it is still reasonable to assume that RAR levels can be considered a prognostic marker for patients with APE.
There is an association between inflammatory response and oxidative stress in patients with APE and their prognosis.[10] On the one hand, after a pulmonary embolism, ischemia, hypoxia, or ischemia-reperfusion lead to a significant increase in the inflammatory response and oxidative stress, which can be seen in the laboratory indicators such as increased IL-6, IL-8, ROS, and decreased SOD in the patients’ blood. Halici et al[11] found that total oxidant status, oxidative stress index (OSI), and total antioxidant status remained high at the end of the first month of treatment. On the other hand, many cohort studies have confirmed that the degree of the inflammatory response and oxidative stress strongly correlate with patient mortality, such as elevated white blood cells, Neutrophil-to-lymphocyte ratio, platelet-to-hemoglobin ratio, IL-6, and ROS, suggesting poor prognosis.[12,13] In addition to these indicators, RDW, lymphocyte ratio, IL-6, and ROS are all indicators of poor prognosis. Besides these indicators, RDW and ALB can also indicate inflammation and oxidative stress in patients with APE, as well as their prognosis.
The RDW is a measure of red blood cell size variability that is a standard component of routine blood tests in clinical practice. It is primarily used in the evaluation of hematological disorders. RDW increases with age, physical activity, and pregnancy in physiological states, but it is also strongly associated with inflammation, oxidative stress, ineffective erythropoiesis, and impaired erythrocyte membrane.[14] RDW is associated with many cardiovascular, cerebrovascular, and respiratory diseases, including their development, progression, and death risk.[15] In cohort studies of APE, patients with elevated RDW are more likely to be diagnosed with massive pulmonary embolism, and this group of patients has a worse prognosis.[16] It has also been shown that RDW is an independent predictor of 30-day mortality in patients with APE. A cohort study including 309 patients with APE found that patients with an initial RDW ≥ 14.6% after admission had significantly higher 30-day mortality than those with an RDW < 14.6%, and therefore RDW was considered an independent predictor of 30-day mortality in APE patients.[17] This study also found a sensitivity of 52.2% and specificity of 87.8% in predicting 30-day mortality in patients with APE when RDW ≥ 16%.Yazici et al[18] investigated the predictive value of dynamic changes in RDW for APE mortality, which measured RDW values of APE patients at admission and 24 hours after admission, and patients were divided into three groups based on the two measurements: normal group (both measurements were normal), decreased group (RDW values > 14.5% at admission and decreased RDW values at 24 hours after admission) and increased group (RDW values > 14.5% at 24h postadmission and increased compared to RDW values at admission). Patients in the increased group were found to have a higher 30-day mortality rate compared to the normal and decreased groups (19% vs 0% vs 3.4%, P = .001). Although the cutoff values for RDW prediction varied across studies, all suggested that RDW may be an independent predictor of 30-day mortality in patients with APE.
ALB is a medium-sized protein synthesized by the liver, which is mainly responsible for regulating colloid osmotic pressure in the body and, similar to RDW, also plays an essential role in the inflammatory response and oxidative stress.[19] Numerous studies have suggested that low ALB is associated with a poor prognosis in patients with acute thrombotic disease. González-Pacheco et al[20] found that patients with CAD with ALB levels < 3.50 g/dL had an increased risk of secondary heart failure and in-hospital mortality. In a prospective study that included 75 acute stroke cases, the mean ALB was significantly higher in patients with a good prognosis (3.03 g/dL) than in those with a poor prognosis (2.08 g/dL).[21] Pulmonary embolism, a type of vascular infarction and local ischemic disease, has a similar profile. In a cohort study including 1032 APE patients, patients with hypoalbuminemia had higher heart rates, lower systolic blood pressure, lower SpO2, and higher mortality (30 days, 90 days after diagnosis).[22]
In this study, RAR, as a combination of RAW and ALB, was associated with all-cause mortality of APE in the full model, even after adjusting for all the variables.
Although this study provides strong evidence that RAR predicts all-cause mortality in patients with APE based on an extensive database, it still has some limitations. Firstly, we chose to use ICD-9-CM and ICD-10-CM to identify APE or other comorbidities, which may be subject to partial error. Secondly, our study was limited by the retrospective study design and may have defects in selection bias and missing data, and information on the causes of morbidity and death in APE patients was not available either. Thirdly, the database did not provide echocardiography, troponin, and N-terminal pro-BNP, which are important indicators for evaluating APE. Further randomized controlled trials are needed to validate the findings in the future.
5. Conclusion
In conclusion, RAR has partial predictive value for predicting in-hospital mortality and one year mortality of APE patients in the ICU, especially for patients with low RAR levels, which indicates a better prognosis. RAR detection is simple and easy. When evaluating the prognosis of APE patients in the ICU, adding this indicator can more accurately identify patients with poor prognosis, so as to develop personalized programs, strengthen treatment and improve the prognosis of APE patients.
Author contributions
Conceptualization: Ding Chaowei.
Data curation: Ding Chaowei, Ziping Zhang, Jiayong Qiu, Dan Du.
Formal analysis: Ding Chaowei, Dan Du.
Methodology: Ding Chaowei.
Writing – original draft: Ding Chaowei.
Writing – review & editing: Zexin Liu.
Abbreviations:
- ALB
- albumin
- APE
- acute pulmonary embolism
- CAD
- coronary atherosclerotic disease
- GAM
- generalized additive model
- ICU
- intensive care unit
- MIMIC-IV
- medical information mart for intensive care IV
- Na+
- Sodium
- RAR
- RDW/ALB ratio
- RDW
- red blood cell distribution width
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Ethical approval to report this case series was obtained from Laboratory for Computational Physiology at the Massachusetts Institute of Technology.
The authors have no funding and conflicts of interest to disclose.
How to cite this article: Ding C, Zhang Z, Qiu J, Du D, Liu Z. Association of red blood cell distribution width to albumin ratio with the prognosis of acute severe pulmonary embolism: A cohort study. Medicine 2023;102:47(e36141).
Contributor Information
Chaowei Ding, Email: 398992987@qq.com.
Ziping Zhang, Email: sangerzzp@icloud.com.
Jiayong Qiu, Email: Jiayong5201@163.com.
Dan Du, Email: 947642356@qq.com.
References
- [1].Fleitas Sosa D, Lehr AL, Zhao H, et al. Impact of pulmonary embolism response teams on acute pulmonary embolism: a systematic review and meta-analysis. Eur Respirat Rev. 2022;31:220023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Wendelboe AM, Raskob GE. Global burden of thrombosis: epidemiologic aspects. Circ Res. 2016;118:1340–7. [DOI] [PubMed] [Google Scholar]
- [3].Konstantinides SV, Meyer G, Becattini C, et al. 2019 ESC Guidelines for the diagnosis and management of acute pulmonary embolism developed in collaboration with the European Respiratory Society (ERS): the task force for the diagnosis and management of acute pulmonary embolism of the European Society of Cardiology (ESC). Eur Respir J. 2019;54:1901647. [DOI] [PubMed] [Google Scholar]
- [4].Elias A, Mallett S, Daoud-Elias M, et al. Prognostic models in acute pulmonary embolism: a systematic review and meta-analysis. BMJ Open. 2016;6:e010324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Venetz C, Jiménez D, Mean M, et al. A comparison of the original and simplified pulmonary embolism severity index. Thromb Haemost. 2011;106:423–8. [DOI] [PubMed] [Google Scholar]
- [6].Wang X, Zhu L, Tao K, et al. Red cell distribution width to serum albumin ratio as an early prognostic marker for severe acute pancreatitis: a retrospective study. Arab J Gastroenterol. 2022;23:206–9. [DOI] [PubMed] [Google Scholar]
- [7].Hong J, Hu X, Liu W, et al. Impact of red cell distribution width and red cell distribution width/albumin ratio on all-cause mortality in patients with type 2 diabetes and foot ulcers: a retrospective cohort study. Cardiovasc Diabetol. 2022;21:91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Gao C, Peng L. Association and prediction of red blood cell distribution width to albumin ratio in all-cause mortality of acute kidney injury in critically ill patients. Front Med. 2023;10:1047933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Johnson AEW, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data. 2023;10:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Jo JY, Lee MY, Lee JW, et al. Leukocytes and systemic inflammatory response syndrome as prognostic factors in pulmonary embolism patients. BMC Pulm Med. 2013;13:74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Halici B, Sarinc Ulasli S, Günay E, et al. Assessment of inflammatory biomarkers and oxidative stress in pulmonary thromboembolism: follow-up results. Inflammation. 2014;37:1186–90. [DOI] [PubMed] [Google Scholar]
- [12].Köse N, Yildirim T, Akin F, et al. Prognostic role of NLR, PLR, and LMR in patients with pulmonary embolism. Bosn J Basic Med Sci. 2020;20:248–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Ozbeyaz NB, Gokalp G, Gezer AE, et al. Novel marker for predicting the severity and prognosis of acute pulmonary embolism: platelet-to-hemoglobin ratio. Biomarkers Med. 2022;16:915–24. [DOI] [PubMed] [Google Scholar]
- [14].Salvagno GL, Sanchis-Gomar F, Picanza A, et al. Red blood cell distribution width: a simple parameter with multiple clinical applications. Crit Rev Clin Lab Sci. 2015;52:86–105. [DOI] [PubMed] [Google Scholar]
- [15].Haybar H, Pezeshki SMS, Saki N. Evaluation of complete blood count parameters in cardiovascular diseases: an early indicator of prognosis? Exp Mol Pathol. 2019;110:104267. [DOI] [PubMed] [Google Scholar]
- [16].Omar HR, Mirsaeidi M, Rashad R, et al. Association of serum albumin and severity of pulmonary embolism. Medicina (Kaunas, Lithuania). 2020;56:26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Zhou XY, Chen HL, Ni SS. Red cell distribution width in predicting 30-day mortality in patients with pulmonary embolism. J Crit Care. 2017;37:197–201. [DOI] [PubMed] [Google Scholar]
- [18].Yazici S, Kiriş T, Sadik Ceylan U, et al. Relation between dynamic change of red cell distribution width and 30-day mortality in patients with acute pulmonary embolism. Clin Respir J. 2018;12:953–60. [DOI] [PubMed] [Google Scholar]
- [19].Belinskaia DA, Voronina PA, Shmurak VI, et al. serum albumin in health and disease: esterase, antioxidant, transporting and signaling properties. Int J Mol Sci . 2021;22:10318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].González-Pacheco H, Amezcua-Guerra LM, Sandoval J, et al. Prognostic implications of serum albumin levels in patients with acute coronary syndromes. Am J Cardiol. 2017;119:951–8. [DOI] [PubMed] [Google Scholar]
- [21].Abubakar S, Sabir A, Ndakotsu M, et al. Low admission serum albumin as prognostic determinant of 30-day case fatality and adverse functional outcome following acute ischemic stroke. Pan Afr Med J. 2013;14:53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Hoskin S, Chow V, Kritharides L, et al. Incidence and impact of hypoalbuminaemia on outcomes following acute pulmonary embolism. Heart Lung Circ. 2020;29:280–7. [DOI] [PubMed] [Google Scholar]