Red blood cell distribution width to albumin ratio is linked to all-cause mortality in critically ill patients with acute kidney injury: a retrospective cohort study

1. Introduction

Acute kidney injury (AKI) is a serious clinical condition that can cause functional disability and mortality in affected patients, especially patients who are critically ill patients [1]. AKI is a frequent cause of ICU admission and remains the leading cause of hospital mortality, thus placing a heavy burden on both families and society [2, 3]. Therefore, it is critical to be able to identify or predict potential risks in order to direct medical intervention in a timely manner. Over recent years, disease severity scores have shown significant efficacy in predicting prognoses, such as mortality and the length of hospital stay for AKI patients [4, 5]. However, these scores are not perfect predictors of disease severity and prognosis. In addition, these scores require a wide range of clinical data and cannot be used if this clinical information is not available [6]. Therefore, it is imperative to develop efficient and reliable predictive tools that can effectively enhance the management of patients with AKI. Hematological parameters are more advantageous than the more classical disease severity scores with regard to the prognostic assessment of AKI [7]. As an alternative, the red blood cell distribution width (RDW) to albumin ratio (RAR) is an easily available, inexpensive and reliable indicator, making this ratio highly appropriate for use in clinical practice.

The RDW is a common hematological parameter that reflects heterogeneity in the volume of red blood cells (RBCs) in the peripheral circulation. RDW is typically combined with other hematological indicators to facilitate the diagnosis of anemia. The RDW has been proven to be highly effective in predicting the prognosis of various diseases [8, 9], including cardiovascular disease, diabetes and other critical illnesses. More importantly, studies have indicated that RDW is also strongly associated with the mortality of patients with AKI [10, 11]. Inflammation is omnipresent in the development of AKI [12]. Serum albumin is a routine indicator of nutritional status or inflammatory response and levels of serum albumin can be used as a simple prognostic factor for AKI [13].

Currently, the RAR is considered a simple and effective novel biomarker of inflammation. Previous research has also demonstrated the significance of the RAR for determining the prognosis of numerous conditions, such as acute myocardial infarction [14], heart failure [15], cancer [16] and sepsis [17]. Nevertheless, the association between the RAR and the prognosis of AKI has yet to be fully elucidated. The objective of this study was to investigate the correlation between the RAR and mortality from any cause and determine the influence of the RAR on patients with AKI who are critically ill and admitted to the ICU.

2. Methods

2.2 Inclusion and exclusion criteria

Patients diagnosed with AKI were enrolled based on specified criteria provided by the Acute Kidney Injury Network (AKIN), including a rise in serum creatinine (SCr) levels of at least 26.5 µmol/L or 1.5-fold the baseline values within 48 hours, or a urinary output of <0.5 mL/(kg⋅h) for a duration exceeding 6 hours [18]. Due to the majority of data in the MIMIC-III database being documented before the publication of the Kidney Disease Improving Global Outcomes (KDIGO) guidelines [19], we used the AKIN criteria rather than the KDIGO criteria in the present study. It must also be noted that some previous studies have utilized the AKIN criteria for identical rationale [10, 20]. In addition, renal function was evaluated to diagnose AKI within 48 hours of admission, taking into account potential confounding factors, such as the use of antibiotics and hospital-acquired infections during the 7-day follow-up. Baseline values of SCr levels were measured for the first time following admission to the ICU. AKI staging was determined by considering both SCr levels and urinary output within the initial 48 hours of ICU admission. We included patients admitted to the ICU with a diagnosis of AKI within 48 hours of admission and recorded in the MIMIC-III database. Only patients who were above 18-years-of-age at the point of their initial admission were selected. Patients were excluded if they satisfied any of the subsequent conditions: (1) multiple hospital or ICU admissions; (2) discharged or died within 48 hours after ICU admission; (3) missing RDW and albumin records on the first day of admission, and (4) pregnant women and oncology patients. On the first day of ICU admission, the 3826 patients included in the study were categorized into four groups according to RAR quartiles. Fig. 1 shows a flowchart depicting the patient screening process.

Fig. 1.Flowchart of patient selection for analysis. MIMIC-III: Multiparameter Intelligent Monitoring in Intensive Care Database III; AKI: acute kidney injury; ICU: intensive care unit.

3. Results

3.2 The incidence rate for allcause mortality between different groups

Fig. 2 shows the Kaplan-Meier survival analysis curves; these demonstrate the occurrence of all-cause mortality in critically ill patients with AKI, categorized into quartile groups based on the RAR. The survival graphs indicated that individuals in the high RAR category experienced higher rate of mortality within 28 days when compared to those in the remaining categories (p < 0.001). This suggests that patients in the high-RAR quartile generally exhibited lower rates of short-term survival. There was a notable variance in mortality rate between the four groups (Q1: 11.4% vs. Q2: 17.3% vs. Q3: 20.9% vs. Q4: 30.5%, log-rank p < 0.001, Fig. 2A). Similar survival curves were obtained for in-hospital and ICU mortality rates during follow-up.

Fig. 2.Kaplan-Meier survival analysis curves for mortality. (A) represents 28-day mortality; (B) represents in-hospital mortality; (C) represents in-ICU mortality. Q1 (RAR ≤4.11), Q2 (4.11 < RAR ≤ 5.00), Q3 (5.00 < RAR ≤ 6.33), Q4 (RAR >6.33).

4. Discussion

In this study, we discovered an association between elevated RAR levels and higher mortality rates within 28 days, during hospitalization, and in the ICU for patients with AKI. Our data also suggested that elevated RAR levels could potentially be used as a predictive factor for increased mortality rates in critically ill patients with AKI. However, it is important to consider confounding factors including multiple clinical and laboratory variables, as these factors may weaken the observed correlation between RAR levels and mortality rates.

AKI originates from a complex interaction between renal insufficiency and subsequent inflammation and coagulation [21]. Studies have indicated that RDW is closely linked to unfavorable prognosis in critically ill patients with AKI and represents an indicator of chronic inflammation that is considered a valid predictor for the prognosis of AKI [10, 11, 22]. Although we identified the predictive value of RDW for all-cause mortality risk has been revealed, the mechanisms underlying the relationship between these factors remain unclear. Both inflammation and oxidative stress may play a key role. Inflammatory cytokines inhibit iron metabolism and bone marrow function, thereby suppressing the production of erythropoietin; this subsequently affects the maturation of erythrocytes, thus leading to impaired erythrocyte synthesis and increased RDW values [23]. AKI patients in the ICU often exhibit symptoms associated with the activation of oxidative stress, including metabolic disorders, sepsis, and hemodynamic dysregulation [24]. The rate of erythrocyte destruction and the release of immature erythrocytes into the circulation are also increased by oxidative stress, consequently leading to an increase in RDW values. Serum albumin, a vital protein with anti-inflammatory, nutritional, and hematological characteristics, effectively suppresses platelet activation and aggregation [25]. Serum albumin is traditionally evaluated to assess the nutritional status of patients who are critically ill. Hypoproteinemia is regarded as a dependable indicator of prognosis in numerous critically ill patients with chronic conditions [26]. Research has shown that low levels of albumin levels may be associated with hemodilution due to blood volume overload, chronic inflammation, malnutrition, and cachexia; these factors may lead to reduced albumin synthesis [15]. Malnutrition is common in critically ill patients, and both nutritional status and chronic inflammation are known to reduce the levels of albumin in patients with AKI. Junhua Lv et al. [27] recently reported a strong correlation between mortality rates and serum albumin levels in critically ill patients with AKI. Furthermore, these authors suggested that there is an inverse relationship between albumin levels and mortality rates at both day 28 and day 90 in critically ill patients with AKI. Our study expands on previous discoveries by additionally identifying the amalgamation of these indicators as favorable indicators for the prognosis of critically ill individuals suffering from AKI.

The RAR has been reported as a simple and novel indicator of inflammation that can be widely used to determine the prognosis of various inflammatory diseases. The RAR combines RDW and albumin indicators to reflect the levels of these two factors and the pathological states of both hematopoietic dysfunction and hypoproteinemia. The RAR has been demonstrated to be a better predictor of adverse disease outcomes than RDW and albumin levels alone [17, 28]. Qingwei Ni et al. [15] reported that the RAR is a promising predictor of mortality rates in patients with heart failure. Elevated RAR levels have also been reported to be significantly associated with increased all-cause mortality in patients with sepsis [17] and chronic obstructive pulmonary disease (COPD) [29]. AKI is a common cause of death in the ICU along with the above-mentioned diseases. Our present results show that the RAR is closely related to mortality rates in critically ill patients with AKI, although the influence of confounding factors, such as comorbidities, cannot be completely excluded. As the RAR increases, the clinical prognosis of critically ill patients with AKI is likely to be exacerbated. It is interesting to note that several severity scoring systems, including the SOFA and SAPS II, have been widely used to predict the prognosis of critically ill patients. However, Shigehiko Uchino et al. [30] concluded that these evaluation methods do not exhibit a significant level of differentiation or accuracy in terms of forecasting death rates among individuals suffering from acute kidney failure. In addition, the scoring systems of SOFA and SAPS II scoring systems are complex and may be time-consuming. In contrast, RDW and albumin results are readily available in the laboratory and are simpler and more efficient. However, RAR can be affected by the patient’s course of treatment resulting in less stable results; consequently, we do not have sufficient evidence to conclude that RAR is superior to existing severity scoring systems. Notably, in this study, we discovered a positive correlation between RAR levels and both SOFA and SAPS II scores in critically ill patients with AKI; these findings concur with those reported previously [31]. Further studies are now needed to evaluate the prognostic efficacy of combining the RAR indicator and severity scoring systems to predict the outcomes of patients with AKI. This combined approach has the potential to provide more accurate predictive results for critically ill patients with AKI.

In addition, we used subgroup analysis to investigate the correlation between RAR levels and 28-day mortality rates among critically ill patients with AKI and found that individuals with AKI and diabetes had elevated 28-day mortality rates when compared to those without diabetes. Previous studies have shown that individuals with diabetes have a higher likelihood of developing AKI compared to those without diabetes [32]. Recently, Guozhi Jiang et al. [33] showed that patients with diabetes and AKI had a higher risk of all-cause mortality than patients without AKI; we obtained similar results in the present study. In addition, we identified a relatively higher mortality rate in AKI patients with CHD than patients without CHD. Therefore, the RAR could serve as a significant indicator for forecasting the outlook of individuals with AKI who have diabetes or heart failure. We are concerned about the findings of our subgroup analysis due to the potential impact of variations among diverse populations. Thus, additional research is required to validate our discoveries.

The present study also has some limitations that need to be considered. First, the study was conducted as a retrospective cohort investigation, and although we made significant attempts to account for potential confounding factors, it proved challenging to completely mitigate selection and confounding biases which represent a limitation for all retrospective studies. Second, only patients diagnosed with AKI within 48 hours of ICU admission and who had RDW and albumin levels tested on the first day were included; this may be an important source of selection bias and potential confounding; therefore, our conclusions require a well-designed multicenter prospective study for further validation. Third, the generalizability of our findings is inherently limited by the fact that the investigated study period was from 2001 to 2012; during this period, the diagnostic criteria and treatment choices of the investigated patient population have changed. Consequently, the results of this study are only used as a reference to support subsequent studies. Fourth, the MIMIC-III database only included patients with AKI admitted to the ICU; thus, it remains unclear whether our conclusions apply to other AKI populations.

5. Conclusions

In critically ill AKI patients, we identified a positive correlation between the RAR and the rates of all-cause mortality within 28 days, during hospitalization, and in the ICU. The monitoring of RAR levels is cost-effective, efficient, simple, and has the potential to be an important and practical indicator for identifying a poor prognosis in critically ill patients with AKI. However, further research is required to determine its value in clinical practice.

Availability of data and materials

The data presented in this study are available on reasonable request from the corresponding author.

Author contributions

FL and LL—were responsible for the study concept and study design; critical revision of the manuscript for important intellectual content. LL—performed data extraction; XTL and MC—were responsible for data analysis; XTL—drafting of the manuscript. All authors have read and approved the final manuscript.

Ethics approval and consent to participate

Approval for the project was granted by the Institutional Review Boards (IRB) at both the Massachusetts Institute of Technology (MIT, Cambridge, MA, USA) and Beth Israel Deaconess Medical Center (BIDMC). Private patient information was removed from the database to protect patient privacy; thus, informed consent was waived.

Acknowledgment

We would like to thank the Massachusetts Institute of Technology and the Beth Israel Deaconess Medical Center for the MIMIC-III database.

Funding

This research received no external funding.

Conflict of interest

The authors declare no conflict of interest.