Red blood cell distribution width to albumin ratio is correlated with prognosis of patients in coronary care unit

Abstract

BACKGROUND

As red blood cell distribution width (RDW) and albumin have been shown to be independent predictors of mortality from various diseases, this study aimed to investigate the effect of the RDW to albumin ratio (RA) as an independent predictor of the prognosis of patients admitted to the coronary care unit (CCU).

AIM

To use the RDW and albumin level to predict the prognosis of patients in the CCU.

METHODS

Data were obtained from the Medical Information Mart Intensive Care III database. The primary outcome was 365-day all-cause mortality, whereas the secondary outcomes were 30- and 90-day all-cause mortality, hospital length of stay (LOS), and CCU LOS. Cox proportional hazards regression model, propensity score matching, and receiver operating characteristic curve analyses were used.

RESULTS

The hazard ratio (95% confidence interval) of the upper tertile (RA > 4.66) was 1.62 (1.29 to 2.03) when compared with the reference (RA < 3.84) in 365-day all-cause mortality. This trend persisted after adjusting for demographic and clinical variables in the propensity score-matching analysis. Similar trends were observed for the secondary outcomes of hospital and CCU LOS. Receiver operating characteristic curve analysis was performed by combining the RA and sequential organ failure assessment (SOFA) scores, and the C-statistic was higher than that of the SOFA scores (0.733 vs 0.702, P < 0.001).

CONCLUSION

RA is an independent prognostic factor in patients admitted to the CCU. RA combined with the SOFA score can improve the predictive ability of the SOFA score. However, our results should be verified in future prospective studies.

Key Words: Red blood cell distribution width; Albumin; Prognosis; Coronary care unit

Core Tip: This study identifies the red blood cell distribution width to albumin ratio (RA) as an independent predictor of prognosis in coronary care unit patients. Higher RA levels (RA > 4.66) are linked to increased 365-day mortality and longer hospital stays, even after adjusting for clinical factors. Combining RA with the sequential organ failure assessment score significantly improves mortality prediction (C-statistic: 0.733 vs 0.702, P < 0.001). These findings suggest RA can enhance early risk assessment, supporting better clinical decision-making in critically ill patients.

INTRODUCTION

Cardiovascular diseases have emerged as the leading cause of mortality globally, representing approximately one-third of all fatalities. In the United States, these conditions accounted for 17% of the overall national health expenditure, highlighting their substantial impact on the healthcare system[1,2]. The coronary care unit (CCU) was founded in the late 1960s with the objective of decreasing mortality rates following acute myocardial infarction. This was achieved through the identification and proactive management of arrhythmias, in addition to offering a clinical laboratory setting dedicated to the research and treatment of acute coronary syndrome[3]. The CCU was staffed with professional personnel and equipped with critical facilities to treat those with severe cardiovascular disease[4]. Killip and Kimball[5] reported that the mortality rate of acute myocardial infarction in patients treated in conventional wards decreased from 26% to 7% through CCU treatment, owing to the presence of professionally trained personnel in CCU wards who possessed strong execution and correct treatment methods when dealing with patients with cardiac arrest and arrhythmia. Subsequently, CCU was widely used and developed into the modern CCU.

Red blood cell (RBC) distribution width (RDW) represents a critical parameter routinely assessed in clinical blood analyses. Its measurement is straightforward in clinical settings and has been correlated with systemic inflammation. Consequently, RDW has emerged as a valuable biomarker for systemic inflammation, serving as a novel prognostic indicator for various health conditions, particularly in the context of cardiovascular and cerebrovascular diseases[6]. RDW is a hematological indicator reflecting the volume heterogeneity of RBCs[7]. Previous studies found a graded independent relationship between higher RDW levels and the risk of death and cardiovascular events in patients with prior myocardial infarction but no symptomatic heart failure at baseline, even after percutaneous coronary intervention[8,9]. Albumin is a protein of moderate size, possessing a molecular weight ranging from 66 to 69 kDa, and it constitutes more than fifty percent of the total serum composition in the human body. Serum albumin, produced by the liver, serves as an important biochemical indicator of nutritional health and plays a crucial role in maintaining colloidal osmotic pressure. It is classified as a negative acute-phase protein, meaning that its synthesis is inversely related to the level of inflammatory activity[10]. Albumin has multiple functions, such as regulating osmotic pressure, antioxidant, anti-inflammatory[11,12]. Research indicates that a significant decrease in albumin levels contributes to fluid accumulation and edema due to a decline in plasma osmotic pressure among patients, potentially exacerbating heart failure[13]. An investigation found that decreased serum albumin in the human body might be a marker of protein metabolism disorders and low-level catabolic inflammation in patients with chronic heart failure[14]. Numerous contemporary investigations have indicated a correlation between the RDW-to-albumin (RA) ratio and the prognostic outcomes of various medical conditions, such as aortic aneurysms, diabetic retinopathy, and shock[15-17].

Currently, it is essential to promptly recognize critically ill patients who exhibit a poor prognosis. This early identification is crucial for enhancing the treatment efficacy for patients in the CCU and for discerning the prognostic factors associated with these severely ill individuals[18,19]. At present, numerous risk assessment instruments have been employed to assess the prognosis of patients in the CCU, potentially offering valuable predictive insights. Nevertheless, these tools exhibit certain drawbacks, including a lack of comprehensive clinical data necessary for the formulation of risk scores, limited predictive accuracy, and the emergence of new scoring systems that may not meet acceptable standards, thereby impeding their broader implementation. Therefore, determining an acceptable and effective scoring method to predict the prognosis of CCU patients is crucial. As far as we are aware, there has been no prior research assessing the prognostic significance of RA in patients admitted to the CCU. The objective of this study was to examine the correlation between RA levels and the outcomes of CCU patients. In this study, we used real-world data obtained from the Medical Information Mart for Intensive Care III database (MIMIC III) database to evaluate the relationship between RA and prognosis in CCU patients.

MATERIALS AND METHODS

Data source

In this study, the MIMIC III v1.4, was utilized. This publicly accessible, single-center database comprises information on patients admitted to the intensive care unit (ICU) of a prominent tertiary care facility. MIMIC III v1.4 encompasses data related to 53423 admissions of adult patients, specifically those aged 16 years and older, who were admitted to the ICU from 2001 to 2012. Approval for the use of MIMIC III was granted by the Institutional Review Boards of both the Beth Israel Deaconess Medical Center and the Massachusetts Institute of Technology. The database includes a comprehensive array of information, such as demographic details, laboratory test results, vital signs, patient vital status, and various chart events. Additionally, it documents the International Classification of Diseases, Ninth Revision codes, as well as hourly physiological data, which were meticulously recorded and verified by ICU nursing staff. Furthermore, written evaluations of radiological images conducted by specialists during the corresponding timeframe were also incorporated into this database[20]. We engaged in an online course provided by the National Institutes of Health, successfully completed the examination on Protecting Human Research Participants, and subsequently applied for the necessary access permissions. Following this, we proceeded to extract pertinent data from the database for our study, ensuring the anonymity of patient identity information throughout the entire process in order to safeguard their privacy. It is important to note that our institutional ethics review committee did not require approval for the research protocols or informed consent, as our investigation was centered on a retrospective analysis of publicly available databases.

Population selection criteria

Among the more than 50000 patients in the database, we included patients admitted to the CCU in the MIMIC III database, the included subjects met the following criteria: (1) Age ≥ 16 years; (2) Had multiple admissions only input the first admission; and (3) The RDW and albumin data were not missing. The exclusion criteria were as follows: (1) A secondary diagnosis of cancer or liver disease on admission; and (2) The patient’s individual data is missing by more than 10%. Liver disease encompasses a range of pathological alterations affecting the liver, which include viral hepatitis, cirrhosis, fatty liver disease, alcoholic liver disease, portal hypertension, hepatic encephalopathy, hepatorenal syndrome, liver necrosis, among several other hepatic disorders and their associated complications, while specifically excluding liver cancer. Ultimately, our study comprised a total of 2218 patients.

Date extraction

Structured Query Language was used to extract the data on the first day of CCU admission in the MIMIC III including basic vital signs, demographic data, basic laboratory parameters, comorbidities and scoring system before treatment. Age, gender and ethnicity were demographic information. Mean arterial pressure, heart rate, respiratory rate (RR), temperature and pulse oximetry-derived oxygen saturation (SpO₂) belonged to vital signs. Comorbidities included congestive heart failure (CHF), atrial fibrillation, chronic obstructive pulmonary disease, coronary atherosclerotic heart disease (CAD), pulmonary circulation and renal failure. Laboratory findings included RDW, albumin, RA, serum creatinine, glucose, serum urea nitrogen, hematocrit, hemoglobin, white blood cell count, platelet, sodium, potassium, international normalized ratio (INR) and prothrombin time (PT). In addition, sequential organ failure assessment (SOFA) was also included. The RA was calculated as the RDW as the numerator divided by albumin using the same blood samples according to the formula: RDW (fl)/Albumin (g/dL). Used the date of patient’s admission as the starting date for follow-up. Obtained the patient’s date of death from the United States government’s Social Security Death Index records. The 365-day all-cause mortality was the primary outcome of our study. We selected 30-day, 90-day all-cause mortality, hospital length of stay (LOS) and CCU LOS as secondary outcomes.

Statistical analysis

Study population were divided into three groups based on RA in the baseline characteristics, with categorical data represented by frequency (percentage) and continuous data represented by mean ± SD or IQR. We did comparisons between groups by the χ²-test or Fisher’s exact test for categorical variables and the variance analysis or the Kruskal-Wallis test for continuous ones. In order to examine the association between RA and the results, the Cox proportional hazards model was used. The analysis of the results was conducted in accordance with the tertiles of RA levels, with the initial tertile serving as the reference category. Outcomes were reported as hazard ratios (HR) accompanied by 95% confidence intervals (CI) or as β coefficients with corresponding 95%CI. Two distinct adjusted models were employed for the multivariate analysis. The selection of confounders was predicated on their association with the outcome, the presence of effect estimate modifications exceeding 10%, or their clinical significance as determined by healthcare professionals within our analytical framework. In the first adjustment model (adjust I), we accounted for covariates such as age, gender, and ethnicity. Meanwhile, in the second adjustment model (adjust II), additional covariates were incorporated, including age, gender, ethnicity, glucose levels, hematocrit, potassium, INR, sodium, serum urea nitrogen, white blood cell count, heart rate, RR, temperature, SpO₂, CAD, CHF, atrial fibrillation, chronic obstructive pulmonary disease, and the SOFA score. To further confirm the association of RA and the 365-day all-cause mortality, we conducted a sensitivity analysis using propensity score matching (PSM) because the 365-day all-cause mortality groups differed significantly across baseline characteristics. The details of PSM were shown in Supplementary Table 1 and Supplementary Figure 1. Covariate balance was assessed in these matched participants by checking P value between the two groups (survival group and non-survival group). We considered a covariate to be well balanced if the P value was more than 0.05. Finally, conditional logistic regression was employed to determine the association of RA and the 365-day all-cause mortality among the populations after PSM. To enhance the assessment of the predictive capability of RA, we conducted a receiver operating characteristic (ROC) curve analysis focused on the 365-day all-cause mortality, utilizing both the SOFA score and the combination of the SOFA score with RA. The analysis was executed using R software (version 3.6.1, available at http://www.r-project.org). All statistical tests were two-tailed, with a significance threshold set at P < 0.05.

RESULTS

Subject characteristics

After excluding patients who did not meet the inclusion criteria, 2218 patients were included in the study (Figure 1). The patients were stratified into three distinct groups based on the tertiles of RA. The baseline characteristics categorized by RA tertile are detailed in Table 1. A total of 739 (33.32%) patients were in the low RA group (RA < 3.84), 736 (33.18%) were in the medium RA group (3.84-4.66), and 743 (33.50%) were in the high RA group (RA > 4.66). Participants with higher calibrated RA (> 4.66) were more likely to be older, black, and report a history of CHF, atrial fibrillation, pulmonary circulation, and renal failure, as well as higher heart rate, RR, serum creatinine, serum urea nitrogen, RDW, potassium, INR, and PT than those with lower calibrated RA (< 3.84). Compared to the low RA group, the high RA group had lower levels of mean arterial pressure, temperature, SpO₂, CAD, albumin, glucose, hematocrit, hemoglobin, and sodium. The scoring system (SOFA) was also higher than that of those with lower RA (< 3.84).

Figure 1 Flow chart of patient selection. CCU: Coronary care unit; MIMIC III: Medical Information Mart for Intensive Care III; RDW: Red blood cell distribution width.

Table 1 Baseline characteristics of the study population, n (%).

Characteristics	RA < 3.84 (n = 739)	RA 3.84-4.66 (n = 736)	RA > 4.66 (n = 743)	P value
RA	3.45 ± 0.28	4.20 ± 0.23	5.89 ± 1.37	< 0.001
Clinical parameters
Age, years	63.86 ± 15.31	70.18 ± 14.98	70.36 ± 14.75	< 0.001
Gender				< 0.001
Male	510 (69.01)	393 (53.40)	404 (54.37)
Female	229 (30.99)	343 (46.60)	339 (45.63)
Ethnicity				0.038
White	497 (67.25)	523 (71.06)	510 (68.64)
Black	46 (6.22)	54 (7.34)	69 (9.29)
Other	196 (26.52)	159 (21.60)	164 (22.07)
Vital signs
MAP, mmHg	80.33 ± 10.91	77.51 ± 11.50	74.03 ± 11.40	< 0.001
Heart rate, times/minute	79.21 ± 15.91	80.94 ± 15.62	85.06 ± 17.20	< 0.001
RR, times/minute	18.52 ± 3.44	19.30 ± 3.88	19.67 ± 4.23	< 0.001
Temperature, °C	36.79 ± 0.60	36.79 ± 0.75	36.68 ± 0.82	0.001
SpO2, %	97.19 ± 1.79	96.89 ± 2.58	96.87 ± 3.65	0.047
Comorbidities
CHF	163 (22.06)	256 (34.78)	241 (32.44)	< 0.001
Atrial fibrillation	190 (25.71)	271 (36.82)	303 (40.78)	< 0.001
COPD	15 (2.03)	19 (2.58)	12 (1.62)	0.425
CAD	477 (64.55)	435 (59.10)	348 (46.84)	< 0.001
Pulmonary circulation	122 (16.51)	192 (26.09)	264 (35.53)	< 0.001
Renal failure	73 (9.88)	155 (21.06)	199 (26.78)	< 0.001
Laboratory parameters
Albumin, g/dL	3.93 ± 0.33	3.44 ± 0.29	2.87 ± 0.51	< 0.001
Serum creatinine, mg/dL	1.17 ± 1.17	1.51 ± 1.34	2.04 ± 1.82	< 0.001
Glucose, mg/dL	121.72 ± 41.36	119.21 ± 39.28	113.50 ± 45.35	< 0.001
Serum urea nitrogen, mg/dL	21.15 ± 16.61	29.52 ± 20.64	38.54 ± 27.38	< 0.001
Hematocrit, %	34.03 ± 5.84	31.13 ± 5.19	29.03 ± 5.35	< 0.001
Hemoglobin, g/dL	11.75 ± 2.05	10.56 ± 1.79	9.64 ± 1.75	< 0.001
WBC, 109/L	9.97 ± 3.80	10.27 ± 4.35	11.17 ± 6.79	0.158
Platelet, 109/L	210.02 ± 68.56	213.47 ± 88.36	216.30 ± 105.17	0.722
RDW, fl	13.52 ± 0.88	14.43 ± 1.18	16.41 ± 2.24	< 0.001
Sodium, mmol/L	136.35 ± 4.25	136.19 ± 4.96	135.86 ± 5.63	0.030
Potassium, mmol/L	3.73 ± 0.48	3.79 ± 0.54	3.84 ± 0.64	0.002
INR	1.27 ± 0.48	1.42 ± 0.69	1.60 ± 0.85	< 0.001
PT, second	14.09 ± 3.93	15.33 ± 5.56	16.54 ± 6.09	< 0.001
Scoring system
SOFA	3.21 ± 2.58	4.24 ± 2.96	5.67 ± 3.41	< 0.001
Primary outcome
365-day mortality	96 (12.99)	172 (23.37)	307 (41.32)	< 0.001
Secondary outcomes
30-day mortality	54 (7.31)	93 (12.64)	167 (22.48)	< 0.001
90-day mortality	70 (9.47)	120 (16.30)	218 (29.34)	< 0.001
Hospital LOS, day	7.12 ± 6.53	9.42 ± 9.55	11.84 ± 11.05	< 0.001
CCU LOS, day	3.69 ± 4.29	4.76 ± 6.34	5.88 ± 7.41	< 0.001

Hazard ratios (95% confidence interval) were derived from Cox proportional hazards regression models. The data came from patients in Medical Information Mart for Intensive Care III database. RA: Red blood cell distribution width to albumin; MAP: Mean arterial pressure; RR: Respiratory rate; SpO₂: Pulse oximetry-derived oxygen saturation; CHF: Congestive heart failure; COPD: Chronic obstructive pulmonary disease; CAD: Coronary atherosclerotic heart disease; WBC: White blood cell; RDW: Red blood cell distribution width; INR: International normalized ratio; PT: Prothrombin time; SOFA: Sequential organ failure assessment; LOS: Length of stay; CCU: Coronary care unit.

Association between RA and results

During the 365-day, 30-day, and 90-day follow-up periods, 575, 314, and 408 deaths were recorded, respectively. The relationships between RA and all-cause mortality, hospital LOS, and CCU LOS in patients admitted to the CCU in the MIMIC III database are shown in Table 2. In relation to the primary endpoint of all-cause mortality at 365 days, our analysis indicated that elevated levels of RA were linked to a heightened risk of mortality. Specifically, the HR with a 95%CI for individuals in the highest tertile (RA > 4.66) was identified as 1.62 (ranging from 1.29 to 2.03) in comparison to the reference group (RA < 3.84). Upon controlling for variables such as age, sex, and ethnicity in the first adjusted model, a discernible increasing trend was noted within the upper tertile (RA > 4.66) with an HR of 1.65 (1.31 to 2.08). Further adjustments for additional confounding factors in the second adjusted model retained the statistical significance of this upward trend in the upper tertile, yielding an HR of 1.30 (1.01 to 1.69). Corresponding trends were also evident for secondary outcomes related to LOS in both the hospital and the CCU. Among the secondary outcomes concerning 30-day and 90-day all-cause mortality, an increasing trend in the upper tertile (RA > 4.66) was only observed in the non-adjusted and first adjusted models. For 30-day all-cause mortality, the HR was 1.40 (1.03 to 1.90) and 1.45 (1.06 to 1.97), respectively. For 90-day all-cause mortality, it was 1.45 (1.11 to 1.89) and 1.49 (1.14 to 1.95), respectively.

Table 2 The β value or hazard ratios or odds ratio for outcomes across groups of red blood cell distribution width-to-albumin ratios (95% confidence interval).

Mortality	Non-adjusted, HR	Non-adjusted, P value	Adjust I, HR	Adjust I, P value	Adjust II, HR	Adjust II, P value	PSM, OR	PSM, P value
365-day all-cause mortality
RA	1.18 (1.12, 1.25)	< 0.001	1.20 (1.13, 1.26)	< 0.001	1.08 (1.01, 1.15)	0.026	1.24 (1.10, 1.41)	< 0.001
Tertiles < 3.84	1.00		1.00		1.00		1.00
Tertiles 3.84-4.66	1.09 (0.85, 1.41)	0.478	1.09 (0.85, 1.40)	0.499	1.02 (0.79, 1.34)	0.859	1.28 (0.87, 1.87)	0.209
Tertiles > 4.66	1.62 (1.29, 2.03)	< 0.001	1.65 (1.31, 2.08)	< 0.001	1.30 (1.01, 1.69)	0.045	1.92 (1.28, 2.89)	0.002
30-day all-cause mortality
RA	1.14 (1.06, 1.22)	< 0.001	1.16 (1.08, 1.24)	< 0.001	1.01 (0.93, 1.10)	0.839
Tertiles < 3.84	1.00		1.00		1.00
Tertiles 3.84-4.66	1.03 (0.74, 1.44)	0.863	1.05 (0.75, 1.48)	0.765	0.99 (0.68, 1.43)	0.944
Tertiles > 4.66	1.40 (1.03, 1.90)	0.032	1.45 (1.06, 1.97)	0.019	1.06 (0.73, 1.52)	0.770
90-day all-cause mortality
RA	1.17 (1.10, 1.25)	< 0.001	1.19 (1.12, 1.27)	< 0.001	1.05 (0.97, 1.13)	0.198
Tertiles < 3.84	1.00		1.00		1.00
Tertiles 3.84-4.66	1.02 (0.76, 1.37)	0.879	1.03 (0.77, 1.39)	0.837	0.96 (0.70, 1.32)	0.795
Tertiles > 4.66	1.45 (1.11, 1.89)	0.007	1.49 (1.14, 1.95)	0.004	1.10 (0.81, 1.50)	0.545
Hospital LOS, β, days
RA	1.70 (1.41, 1.99)	< 0.001	1.76 (1.46, 2.06)	< 0.001	1.20 (0.85, 1.55)	< 0.001
Tertiles < 3.84	0.00		0.00		0.00
Tertiles 3.84-4.66	2.30 (1.36, 3.24)	< 0.001	2.54 (1.58, 3.51)	< 0.001	1.54 (0.52, 2.56)	0.003
Tertiles > 4.66	4.72 (3.77, 5.66)	< 0.001	4.98 (4.02, 5.94)	< 0.001	2.73 (1.61, 3.85)	< 0.001
CCU LOS, β, days
RA	0.66 (0.47, 0.86)	< 0.001	0.72 (0.52, 0.92)	< 0.001	0.23 (0.00, 0.46)	0.047
Tertiles < 3.84	0.00		0.00		0.00
Tertiles 3.84-4.66	1.07 (0.44, 1.70)	< 0.001	1.23 (0.59, 1.87)	< 0.001	0.67 (0.00, 1.34)	0.049
Tertiles > 4.66	2.20 (1.57, 2.82)	< 0.001	2.37 (1.73, 3.01)	< 0.001	0.80 (0.06, 1.53)	0.034

Non-adjusted and adjust I and II and propensity score matching were derived from Cox proportional hazards regression models. Adjust I covariates were adjusted for age, gender and ethnicity. Adjust II covariates were adjusted for age, gender, ethnicity, glucose, hematocrit, potassium, international normalized ratio, sodium, serum urea nitrogen, white blood cell, heart rate, respiratory rate, temperature, pulse oximetry-derived oxygen saturation_, coronary atherosclerotic heart disease, congestive heart failure, atrial fibrillation, chronic obstructive pulmonary disease, sequential organ failure assessment. propensity score matching further adjusted for age, ethnicity, respiratory rate, congestive heart failure, serum creatinine, serum urea nitrogen, hematocrit, hemoglobin, potassium, prothrombin time based on Supplementary Table 2. HR: Hazard ratio; PSM: Propensity score matching; OR: Odds ratio; RA: Red blood cell distribution width to albumin; LOS: Length of stay; CCU: Coronary care unit.

PSM analyses

A PSM analysis was performed to evaluate the association between RA and all-cause mortality within a 365-day period among patients in the CCU. The initial characteristics of patients categorized into various RA groups are presented in Supplementary Table 2. However, according to the 365-day all-cause mortality, age, ethnicity, RR, CHF, serum creatinine, serum urea nitrogen, hematocrit, hemoglobin, potassium, and PT were not well matched with the baseline characteristics of the patients. After adjusting for these variables in the conditional logistic regression analysis of PSM, we found that a higher RA was related to an increased risk of the primary outcome of 365-day all-cause mortality. The odds ratio (95%CI) of the upper tertile (RA > 4.66) was 1.92 (1.28 to 2.89) when compared with the reference (RA < 3.84).

ROC curve analysis

To assess the potential prognostic significance of RA in forecasting the survival outcomes of patients in the CCU, an ROC curve analysis was conducted. The results, illustrated in Figure 2, indicate that the C-statistic obtained from the combination of RA and the SOFA score (0.733) surpassed that derived from the SOFA score alone (0.702), with a statistically significant difference (P < 0.001).

Figure 2 Receiver operating characteristic curve for 365-day mortality combining sequential organ failure assessment and red blood cell distribution width to albumin. Model 1: Sequential organ failure assessment + red blood cell distribution width to albumin; Model 2: Sequential organ failure assessment; AUC: Area under the curve.

DISCUSSION

Our findings can be summarized as follows: First, RA was closely related to the study’s main outcome (365-day all-cause mortality rate), and the upper tertile of RA was related to increased mortality risk. Second, the upper tertile of RA was associated with increased 30-day and 90-day all-cause mortality, hospital LOS, and CCU LOS. Third, after adjusting for potential confounding factors and PSM, RA was determined to be an independent predictor of the clinical outcomes of patients in the CCU, although there was no statistically significant difference in 30-day and 90-day all-cause mortality after adjusting for additional variables. Fourth, RA combined with the SOFA score proved to be a better predictor of outcomes than the SOFA score alone. This may indicate a close correlation between RA and disease prognosis in patients admitted to the CCU.

Cardiovascular disease imposes a heavy burden on the medical system worldwide, particularly because of its high mortality and incidence rates[21,22]. Over the past 50 years, improvements in medical equipment and professional skills in the CCU have improved the prognosis of a large proportion of critically ill patients[23]. Therefore, to improve the prognosis of CCU patients further, it is important to study useful biomarkers for CCU prognosis early. Chronic inflammation plays an important role in the development of adverse cardiac remodeling and heart failure. Research has found that in patients with severe heart disease, levels of pro-inflammatory circulating factors such as interleukin-6, growth differentiation factor 15, tumor necrosis factor α and galectin-4[24]. The inflammatory process can lead to myocardial damage, and an increase in the levels of inflammatory factors is associated with the deterioration and progression of heart diseases. Ischemic cardiomyopathy and idiopathic dilated cardiomyopathy are associated with inflammation[25]. Inflammation has a detrimental impact on the hematopoietic capabilities of the bone marrow, resulting in the suppression of RBC maturation. This impairment is characterized by reticulocytosis and an elevated RDW. Furthermore, oxidative stress contributes to an increased RDW by reducing the lifespan of RBCs, causing them to prematurely enter the peripheral circulation. This mechanism may elucidate the aforementioned observation[23]. RDW, an inflammatory marker in the human body, has long been associated with adverse cardiovascular events and death in many cardiovascular disease prognostic studies[26-28].

Human serum albumin is synthesized by the liver, and its plasma concentration is influenced by its synthesis ratio, dilution level, and loss of exogenous albumin[11,29-31]. When the human body is in an inflammatory state, not only during infection, trauma, and surgery, but also in patients with heart failure, plasma albumin levels decrease, which is also related to the survival rate of patients with advanced heart failure[32]. There are various causes of decreased plasma albumin levels in patients with heart failure, including malnutrition, hemodilution, reduced protein synthesis caused by liver congestion, inflammatory consumption, increased metabolic activity, and proteinuria loss[33]. Evidence suggests that hypoalbuminemia is common in patients with heart failure, which could cause pulmonary edema and elevated cardiac pressure. Severe hypoalbuminemia promotes fluid retention and edema by reducing plasma osmotic pressure, which may further exacerbate heart and kidney failure[13]. A study found that human serum albumin may be a biomarker of protein metabolism disorders and low-level catabolic inflammation in patients with chronic heart failure[14]. Several studies have shown that the serum albumin level in the human body correlates to varying degrees with the prognosis of various diseases, including cardiovascular and cerebrovascular diseases, which are primarily caused by malnutrition and inflammation[11,17,34-36]. A study involving over 1000 patients with CAD reported that albumin reduction after percutaneous coronary intervention for coronary artery disease can predict higher all-cause mortality in patients[20].

Our findings are consistent with those of previous studies evaluating the prognostic value of RA in other clinical settings[15-17]. Several studies on the prognosis of patients with cardiovascular diseases have combined multiple indicators[37-40]. Our study found that RA was an independent predictor of prognosis in patients admitted to the CCU. RA, which combines RDW and albumin level, is advantageous for evaluating inflammatory processes. Moreover, ROC analysis showed that the combination of RA and SOFA enhances the predictive ability of SOFA. As a promising novel biomarker, RA can be swiftly and easily accessed through laboratory tests conducted upon admission, exhibiting notable advantages in terms of its straightforwardness. Due to its affordability, accessibility, and capacity to forecast mortality rates, RA holds significant clinical relevance for patients admitted to the CCU, thereby assisting clinicians in making prompt clinical decisions.

Our investigation presents several notable strengths. We meticulously adjusted our analysis to account for potential confounding variables that could impact the association between RA and patient outcomes in those admitted to the CCU. Additionally, we conducted multiple validations of our findings through various models. Nevertheless, our study is not without limitations. Firstly, the retrospective, single-center design of our research may introduce bias, highlighting the need for additional prospective studies to mitigate this concern. Secondly, while RA data is generally accessible in clinical settings, we observed a loss of RA information within the database, which could have resulted in selection bias. Thirdly, the initial findings imply that RA may function as a risk-adjusted instrument with prognostic relevance for patients in the CCU. In order to solidify RA as a prognostic biomarker, it is essential to confirm its clinical significance. Further investigations are necessary to substantiate our findings.

CONCLUSION

Through the validation of multiple models, RA was identified as an independent prognostic factor in patients admitted to the CCU. ROC analysis indicated that the incorporation of RA alongside the SOFA score enhances the predictive capacity of the SOFA score. Nevertheless, it is imperative that our findings be validated through future prospective research.