Abstract
Background
The primary objective was to examine the association between the lactate/albumin ratio (LAR) and the prognosis of patients with acute kidney injury (AKI) undergoing continuous renal replacement therapy (CRRT).
Methods
Utilizing the Medical Information Mart for Intensive Care IV (MIMIC-IV, v2.0) database, we categorized 703 adult AKI patients undergoing CRRT into survival and non-survival groups based on 28-day mortality. Patients were further grouped by LAR tertiles: low (< 0.692), moderate (0.692-1.641), and high (> 1.641). Restricted cubic splines (RCS), Least Absolute Shrinkage and Selection Operator (LASSO) regression, inverse probability treatment weighting (IPTW), and Kaplan-Meier curves were employed.
Results
In our study, the patients had a mortality rate of 50.07% within 28 days and 62.87% within 360 days. RCS analysis revealed a non-linear correlation between LAR and the risk of mortality at both 28 and 360 days. Cox regression analysis, which was adjusted for nine variables identified by LASSO, confirmed that a high LAR (>1.641) served as an independent predictor of mortality at these specific time points (p < 0.05) in AKI patients who were receiving CRRT. These findings remained consistent even after IPTW adjustment, thereby ensuring a reliable and robust outcome. Kaplan-Meier survival curves exhibited a gradual decline in cumulative survival rates at both 28 and 360 days as the LAR values increased (log-rank test, χ2 = 48.630, p < 0.001; χ2 = 33.530, p < 0.001).
Conclusion
A high LAR (>1.641) was found to be an autonomous predictor of mortality at both 28 and 360 days in critically ill patients with AKI undergoing CRRT.
Keywords: Lactate/albumin ratio, acute kidney injury, continuous renal replacement therapy, prognosis, MIMIC-IV database
Introduction
Acute kidney injury (AKI) is a prevalent and consequential clinical syndrome observed across diverse medical settings, notably in the intensive care unit (ICU) [1]. It is distinguished by an abrupt deterioration or cessation of renal function, resulting in multi-organ dysfunction and heightened mortality [2]. Research indicates that AKI manifests in roughly 33%-67% of adult critically ill individuals, with a mortality rate reaching up to 28% [3–6]. In the context of AKI treatment strategies, continuous renal replacement therapy (CRRT) is recognized as an efficacious modality for managing critically ill patients afflicted with AKI [7]. CRRT offers significant advantages for critically ill patients experiencing hemodynamic instability by emulating the kidney’s filtration mechanism, thereby enabling continuous blood purification and fluid balance regulation [8,9]. Regrettably, the mortality rate among patients undergoing CRRT remains alarmingly high, reaching up to 63% as reported [10]. This statistic underscores the immense difficulties encountered in the clinical management of these individuals and underscores the imperative for improved biomarkers to facilitate prognostic evaluation.
The lactate/albumin ratio (LAR) has garnered recent attention as a prospective biomarker and has demonstrated prognostic value in critically ill individuals [11–13]. In the context of AKI, heightened lactate levels frequently signify tissue hypoperfusion and insufficient oxygenation, whereas hypoalbuminemia may indicate malnutrition or chronic inflammation, both of which are linked to unfavorable outcomes in patients with AKI [14,15]. Consequently, the lactate/albumin ratio holds potential as a valuable composite biomarker for evaluating the severity and prognosis of AKI patients.
However, there remains a lack of comprehensive investigation into the correlation between LAR and prognosis in patients with AKI undergoing CRRT. Consequently, this study sought to address this gap by gathering data from the Medical Information Mart for Intensive Care IV (MIMIC-IV, v2.0) database, specifically targeting patients with AKI undergoing CRRT. The primary objective was to examine the relationship between LAR and the prognosis of these patients, thereby introducing novel insights into the management of individuals afflicted with AKI.
Methods
Source of data
The data utilized in this study originates from the MIMIC-IV database [16], a substantial publicly accessible repository. This database was developed by the MIT Computational Physiology Laboratory and encompasses clinical data pertaining to patients admitted to Beth Israel Deaconess Medical Center (BIDMC) between the years 2008 and 2019. To gain access to this data, two of the authors successfully completed the Collaborative Training Initiative in Human Research (CITI) program course and subsequently passed the examination (ID number: 51774135; 53446653). It is important to note that the data employed in this study underwent de-identification and anonymization processes, thereby obviating the need for informed consent. This study reports the following articles according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [17].
Inclusion and exclusion criteria
The study’s eligibility criteria were delineated into three principal cohorts: (1) adult individuals aged 18 or above, identified with AKI as per the Kidney Disease Improving Global Outcomes (KDIGO) standards [18]; (2) subjects who were administered CRRT; and (3) patients receiving their inaugural ICU admission. In contrast, the criteria for exclusion incorporated: (1) patients succumbing to mortality upon the day of ICU admission, (2) patients enduring chronic kidney disease (CKD) at stage 5, and (3) patients with deficient records, such as unrecorded lactate or albumin levels.
Data extraction
The MIMIC-IV database was accessed utilizing PostgreSQL software (version V14.5-1) and Navicat Premium 15 software. The data collection encompassed various variables such as age, gender, sequential organ failure assessment (SOFA) score, acute physiology score III (APS III), net ultrafiltration (NUF) rate, laboratory parameters, therapeutic information, coexisting illnesses, and length of ICU stay. Based on previous literature [19], we calculated the NUF rate using the following formula: NUF rate (milliliters per kilogram per hour) = cumulative NUF volume at the end of 48 h (milliliters)/[weight at the initiation of CRRT (kilograms)) × treatment duration in the first 48 h (hours)]. The laboratory parameters encompassed lactate, albumin, LAR, arterial partial pressure of oxygen (PaO2), white blood cell (WBC), red blood cell (RBC), platelet, red blood cell distribution width (RDW), prothrombin time (PT), anion gap, blood urea nitrogen (BUN), creatinine, aspartate aminotransferase (AST), alanine aminotransferase (ALT), bilirubin, glucose, sodium, total calcium, and phosphorus. The therapeutic information entailed the count of individuals undergoing mechanical ventilation (MV), transthoracic echocardiography (TTE), as well as the administration of norepinephrine and dopamine. Furthermore, data regarding comorbidities such as hypertension, diabetes, congestive heart failure, chronic pulmonary disease, malignant tumors, cerebral infarction, cirrhosis, CKD, pulmonary embolism, acute myocardial infarction (AMI), sepsis, and oliguria were extracted. All laboratory variables were based on the first measurements obtained after admission to the ICU, and disease severity scores were determined based on indicators within 24 h of ICU admission.
Groups and clinical outcomes
Patients were stratified into two cohorts based on their 28-day prognosis: the survival group (n = 351) and the non-survival group (n = 352). Furthermore, the entire population was divided into three groups according to tertiles of LAR values: the low LAR group (< 0.692, n = 237), the moderate LAR group (0.692-1.641, n = 232), and the high LAR group (> 1.641, n = 234). The primary outcome measure for this study was 28-day all-cause mortality, while the secondary outcome measure was 360-day all-cause mortality.
Statistical analysis
Continuous data that conformed to a normal distribution were reported as mean ± standard deviation (0x ± s), and between-group comparisons were conducted using the t-test. In cases where the data did not adhere to a normal distribution, the median (interquartile range) [M (QL, QU)] was reported, and between-group comparisons were performed using the Mann-Whitney U test. Discrete data were presented as composition ratios (%), and between-group comparisons were conducted using the chi-square (χ2) test method. Restricted cubic spline (RCS) analysis was employed to visually depict the association between LAR and the risk of mortality. Ten variables were selected through Least Absolute Shrinkage and Selection Operator (LASSO) with 10-fold cross-validated regression analysis, including LAR, age, APS III, NUF rate, creatinine, RDW, PT, phosphate, MV, and norepinephrine. The selected variables were incorporated into univariate and multivariate Cox proportional risk models to explore the correlation between LAR and all-cause mortality. To account for potential confounding factors and account for the weighted nature of the sample, we employed an inverse probability treatment weighting (IPTW) approach utilizing the original population to construct a weighted cohort. This allowed us to compare general information and variables, with those variables exhibiting a P-value less than 0.10 being included in a subsequent multivariate Cox regression analysis. The results were reported as hazard ratios (HR) with corresponding 95% confidence intervals (CI). Additionally, Kaplan-Meier curves were generated both before and after adjustment using IPTW, and cumulative survival rates were compared among the three groups with varying levels of LAR using the log-rank test. Statistical analysis was conducted using Stata14.0 and R3.3.2 software, with a significance level of p < 0.05.
Results
Basic characteristics
In the original cohort, a total of 703 eligible patients were enrolled in the study, as depicted in Figure 1. The patients had a mean age of 60.22 years, with 59.50% being male. Non-survivors exhibited significantly higher age, SOFA score, APS III, lactate, LAR, RDW, PT, bilirubin, total blood calcium, and serum phosphorus (p < 0.05) compared to survivors. Additionally, non-survivors had higher rates of comorbid chronic kidney disease and oliguria, as well as higher rates of receiving MV, norepinephrine, and dopamine (p < 0.05). According to the findings presented in Table 1, the analysis of IPTW yielded a total weighted cohort of 678.82 individuals. Among this cohort, 328.84 individuals belonged to the survival group, while 349.98 individuals belonged to the non-survival group. Notably, several variables exhibited statistically significant differences (p < 0.05) between the survival and non-survival groups within the weighted cohort. These variables included age, SOFA score, APS III, platelets, RDW, PT, bilirubin, anion gap, NUF rate, PaO2, comorbidities (such as hypertension and cirrhosis), proportion of individuals receiving MV and norepinephrine, and length of ICU stay.
Figure 1.
The flowchart showing the systematic process for selecting study participants.
Table 1.
Baseline characteristics of the two cohorts.
| Variables | Original cohort |
Weighted cohort |
||||
|---|---|---|---|---|---|---|
| Survival group (n = 351) |
Non-survival group (n = 352) |
P-value | Survival group (n = 328.84) |
Non-survival group (n = 349.98) |
P-value | |
| Age (year) | 58.42 ± 15.59 | 62.01 ± 14.95 | <0.001 | 58.86 ± 15.74 | 62.73 ± 13.70 | 0.012 |
| Sex, male, n (%) | 209 (59.54) | 212 (60.23) | 0.853 | 189.90 (57.70) | 224.50 (64.20) | 0.256 |
| SOFA score | 12.93 ± 3.76 | 14.00 ± 3.82 | <0.001 | 12.80 ± 3.74 | 13.72 ± 3.62 | 0.019 |
| APS III | 93.77 ± 26.42 | 104.51 ± 28.07 | <0.001 | 93.19 ± 26.20 | 101.44 ± 26.51 | 0.006 |
| NUF rate (mL/kg/h) | 3.35 (2.71, 4.49) | 3.21 (2.30, 4.27) | 0.005 | 3.61 (2.80, 4.72) | 3.04 (2.03, 4.07) | <0.001 |
| Laboratory parameters | ||||||
| Lactate (mmol/L) | 2.30 (1.50, 4.20) | 3.70 (1.95, 7.40) | <0.001 | 2.50 (1.50, 4.98) | 3.50 (1.90, 6.10) | 0.011 |
| Albumin (g/dL) | 2.92 ± 0.73 | 2.91 ± 0.78 | 0.860 | 2.90 ± 0.74 | 2.85 ± 0.75 | 0.538 |
| LAR | 0.83 (0.52, 1.56) | 1.37 (0.67, 2.85) | <0.001 | 0.90 (0.58, 1.72) | 1.36 (0.67, 2.54) | 0.013 |
| PaO2 (mmHg) | 87.00 (52.00, 171.00) | 73.00 (47.00, 130.50) | 0.005 | 101.00 (59.00, 173.26) | 67.81 (46.00, 116.90) | <0.001 |
| WBC (×109/L) | 12.90 (7.90, 18.60) | 13.35 (9.05, 19.70) | 0.164 | 13.06 (7.57, 17.50) | 12.10 (7.80, 17.60) | 0.737 |
| RBC (×1012/L) | 3.50 ± 0.91 | 3.43 ± 0.96 | 0.324 | 3.61 ± 0.98 | 3.48 ± 0.92 | 0.257 |
| Platelet (×109/L) | 153.50 (97.00, 238.00) | 153.50 (83.00, 216.50) | 0.126 | 156.00 (99.00, 247.00) | 135.66 (83.29, 203.84) | 0.041 |
| RDW (%) | 15.90 ± 2.71 | 16.86 ± 3.42 | <0.001 | 15.77 ± 2.69 | 16.50 ± 3.34 | 0.08 |
| PT (s) | 16.10 (13.30, 21.90) | 18.60 (14.20,27.55) | <0.001 | 15.71 (13.30, 21.70) | 18.70 (14.50, 29.23) | 0.011 |
| Anion gap (mmol/L) | 20.82 ± 7.01 | 21.51 ± 7.97 | 0.223 | 21.19 ± 6.50 | 20.03 ± 6.62 | 0.099 |
| BUN (mg/dL) | 14.60 (8.19, 24.21) | 13.35 (7.65, 21.89) | 0.122 | 13.17 (6.76, 23.14) | 12.82 (7.83, 20.65) | 0.834 |
| Creatinine (umol/L) | 236.68 (141.44, 397.80) | 194.48 (123.76, 313.82) | 0.002 | 203.32 (123.76, 388.96) | 194.48 (123.76, 318.33) | 0.258 |
| AST (U/L) | 95.00 (42.00, 440.00) | 127.00 (53.00, 408.50) | 0.142 | 114.73 (47.19, 749.07) | 135.00 (60.00, 456.92) | 0.582 |
| ALT (U/L) | 55.00 (23.00, 254.00) | 55.00 (26.00, 205.00) | 0.700 | 55.00 (25.05, 268.42) | 59.93 (27.00, 257.15) | 0.832 |
| Bilirubin (umol/L) | 18.81 (8.55, 59.85) | 29.93 (10.26, 98.33) | <0.001 | 15.39 (8.55, 51.30) | 30.78 (11.97, 74.62) | <0.001 |
| Glucose (mmol/L) | 7.72 (5.83, 11.22) | 7.69 (5.67, 11.30) | 0.576 | 7.67 (5.80, 10.83) | 8.14 (5.98, 12.05) | 0.473 |
| Sodium (mmol/L) | 136.62 ± 6.67 | 137.28 ± 7.21 | 0.210 | 136.43 ± 6.71 | 137.54 ± 7.13 | 0.134 |
| Total calcium (mmol/L) | 1.97 ± 0.37 | 2.03 ± 0.42 | 0.037 | 1.97 ± 0.38 | 2.40 ± 1.44 | 0.243 |
| Phosphate (mmol/L) | 1.75 ± 0.86 | 1.90 ± 0.87 | 0.025 | 1.71 ± 0.80 | 2.07 ± 1.29 | 0.199 |
| Therapies, n (%) | ||||||
| MV | 323 (92.02) | 347 (98.58) | <0.001 | 300.10 (91.30) | 342.10 (97.80) | 0.014 |
| TTE | 157 (44.73) | 153 (43.47) | 0.736 | 167.20 (50.80) | 158.40 (45.30) | 0.374 |
| Norepinephrine | 259 (73.79) | 334 (94.89) | <0.001 | 245.00 (74.50) | 325.80 (93.10) | <0.001 |
| Dopamine | 35 (9.97) | 57 (16.19) | 0.014 | 34.60 (10.50) | 47.70 (13.60) | 0.353 |
| Coexisting illness, n (%) | ||||||
| Hypertension | 108 (30.77) | 110 (31.25) | 0.890 | 95.30 (29.00) | 143.70 (41.10) | 0.051 |
| Diabetes | 112 (31.91) | 97 (27.56) | 0.207 | 94.20 (28.70) | 84.90 (24.30) | 0.366 |
| Congestive heart failure | 102 (29.06) | 110 (31.25) | 0.527 | 88.10 (26.80) | 109.40 (31.30) | 0.401 |
| Chronic pulmonary disease | 79 (22.51) | 85 (24.15) | 0.607 | 73.40 (22.30) | 79.90 (22.80) | 0.918 |
| Malignant tumors | 56 (15.95) | 48 (13.64) | 0.387 | 49.40 (15.00) | 58.70 (16.80) | 0.652 |
| Cerebral infarction | 28 (7.98) | 26 (7.39) | 0.769 | 26.70 (8.10) | 19.30 (5.50) | 0.252 |
| Cirrhosis | 92 (26.21) | 107 (30.40) | 0.218 | 78.10 (23.70) | 124.00 (35.40) | 0.050 |
| CKD | 99 (28.21) | 74 (21.02) | 0.027 | 80.20 (24.40) | 64.00 (18.30) | 0.157 |
| Pulmonary embolism | 12 (3.42) | 11 (3.13) | 0.827 | 9.30 (2.80) | 13.30 (3.80) | 0.564 |
| AMI | 47 (13.39) | 57 (16.19) | 0.295 | 50.90 (15.50) | 52.70 (15.10) | 0.935 |
| Sepsis | 335 (95.44) | 332 (94.32) | 0.499 | 317.30 (96.50) | 329.40 (94.10) | 0.220 |
| Oliguria | 171 (48.72) | 200 (56.82) | 0.031 | 147.20 (44.70) | 177.10 (50.60) | 0.357 |
| Length of ICU stay (day) | 12.97 (7.10, 20.93) | 5.45 (2.41, 11.25) | <0.001 | 14.07 (7.19, 24.11) | 5.22 (2.36, 11.63) | <0.001 |
Abbreviations: SOFA: sequential organ failure assessment, APS III: acute physiology score II, NUF: net ultrafiltration, LAR: lactate/albumin ratio, PaO2: arterial partial pressure of oxygen, WBC: white blood cell, RBC: red blood cell, RDW: red blood cell distribution width, PT: prothrombin time, BUN: blood urea nitrogen, AST: aspartate aminotransferase, ALT: alanine aminotransferase, MV: mechanical ventilation, TTE: transthoracic echocardiography, CKD: chronic kidney disease, AMI: acute myocardial infarction, ICU intensive care unit.
LAR and the risk of all-cause mortality
According to the data presented in Table 2, the overall mortality rate in the study population was 50.07% at 28 days and 62.87% at 360 days. Statistical analysis revealed significant disparities in both 28-day and 360-day mortality rates across the three groups (χ2 = 33.895, p < 0.001; χ2 = 14.901, p = 0.001). Notably, patients with elevated LAR values (>1.641) exhibited a significantly higher mortality rate.
Table 2.
All-cause mortality at 28 and 360 days across three groups of patients with AKI undergoing CRRT.
| Groups | 28-day |
360-day |
||||||
|---|---|---|---|---|---|---|---|---|
| Survival (n = 351) |
Non-survival (n = 352) |
χ2 | P-value | Survival (n = 261) |
Non-survival (n = 442) |
χ2 | P- value | |
| LAR <0.692 (n = 237) | 145 (41.31) | 92 (26.14) | 106 (40.61) | 131 (29.64) | ||||
| LAR 0.692-1.641 (n = 232) | 124 (35.33) | 108 (30.68) | 90 (34.48) | 142 (32.13) | ||||
| LAR >1.641 (n = 234) | 82 (23.36) | 152 (43.18) | 33.895 | <0.001 | 65 (24.90) | 169 (38.24) | 14.901 | 0.001 |
Abbreviations: AKI: acute kidney injury, CRRT: continuous renal replacement therapy, LAR: lactate/albumin ratio.
The analysis of RCS in Figure 2 revealed a non-linear association between LAR and the risk of all-cause mortality at 28 and 360 days in patients with AKI undergoing CRRT (χ2 = 18.560, p < 0.001; χ2 = 10.340, p = 0.016). As LAR levels rose, the likelihood of patient mortality also increased. However, this increase gradually decelerated until reaching a plateau.
Figure 2.
RCS Analyses of the correlation between lactate/albumin ratio (LAR) and the risk of 28-day (a) and 360-day (b) all-cause mortality in patients with acute kidney injury (AKI) undergoing continuous renal replacement therapy (CRRT).
COX regression analysis
A LASSO regression analysis was conducted with 10-fold cross-validation to examine the relationship between the patient’s 28-day survival and the independent variables from the baseline characteristics listed in Table 1. Figure 3 displays the ten variables that were assessed for non-zero parameters at the optimal λ value of 0.039, which include LAR, age, APS III, NUF rate, creatinine, RDW, PT, phosphate, MV, and norepinephrine. Subsequently, the Cox regression analysis was performed, as presented in Table 3. In the unadjusted model, the HR (95% CI) for 28-day mortality in the high LAR group was 2.304 (1.777-2.987), indicating that a high LAR (> 1.641) was associated with an increased risk of 28-day mortality in patients (p < 0.05). Model II, which adjusted for the other nine variables selected through LASSO analysis, confirmed that high LAR (> 1.641) was an independent risk factor for 28-day all-cause mortality in patients with AKI undergoing CRRT (HR 1.597, CI 1.203-2.120, p = 0.001). Cox regression analysis of 360-day all-cause mortality in patients with AKI undergoing CRRT yielded similar findings.
Figure 3.
Clinical features were selected using LASSO regression with cross-validation. (a) The selection results for the optimal parameter λ (cross-validation). Two dotted lines represent two specific λ values: lambda.min (left) and lambda.lse (right). (b) The dynamic process chart of feature selection through LASSO regression. Each curve represents the trajectory of a single independent variable coefficient, with the vertical axis showing the value of the coefficient for the independent variable, the lower horizontal axis indicating log (λ), and the upper horizontal axis indicating the number of variables with non-zero coefficients present in the model at that time.
Table 3.
Cox proportional hazard model assessing all-cause mortality in patients with AKI undergoing CRRT across three groups.
| Variables | Unadjusted |
Model I |
Model II |
|||||||
|---|---|---|---|---|---|---|---|---|---|---|
| HR | 95%CI | P-value | HR | 95%CI | P-value | HR | 95%CI | P-value | ||
| Original cohort | 28-day mortality | |||||||||
| LAR <0.692 | 1 | 1 | 1 | |||||||
| LAR 0.692-1.641 | 1.302 | 0.986- 1.719 | 0.063 | 1.189 | 0.894-1.582 | 0.234 | 1.038 | 0.779-1.385 | 0.797 | |
| LAR >1.641 | 2.304 | 1.777-2.987 | <0.001 | 1.838 | 1.387- 2.435 | <0.001 | 1.597 | 1.203-2.120 | 0.001 | |
| 360-day mortality | ||||||||||
| LAR <0.692 | 1 | 1 | 1 | |||||||
| LAR 0.692-1.641 | 1.222 | 0.963-1.549 | 0.098 | 1.142 | 0.895-1.458 | 0.285 | 1.016 | 0.794-1.300 | 0.900 | |
| LAR >1.641 | 1.881 | 1.496-2.364 | <0.001 | 1.599 | 1.247-2.051 | <0.001 | 1.427 | 1.111-1.834 | 0.005 | |
| weighted cohort | 28-day mortality | |||||||||
| LAR <0.692 | 1 | 1 | 1 | |||||||
| LAR 0.692-1.641 | 1.263 | 0.867-1.838 | 0.223 | 1.297 | 0.907-1.855 | 0.154 | 1.200 | 0.826-1.743 | 0.337 | |
| LAR >1.641 | 1.855 | 1.144-3.006 | 0.012 | 1.939 | 1.261-2.983 | 0.003 | 1.939 | 1.325-2.836 | <0.001 | |
| 360-day mortality | ||||||||||
| LAR <0.692 | 1 | 1 | 1 | |||||||
| LAR 0.692-1.641 | 1.149 | 0.824-1.602 | 0.412 | 1.164 | 0.851-1.591 | 0.342 | 1.120 | 0.827-1.519 | 0.464 | |
| LAR >1.641 | 1.563 | 1.004-2.432 | 0.048 | 1.614 | 1.084-2.404 | 0.019 | 1.698 | 1.220-2.362 | 0.002 | |
Original cohort:
Unadjusted model adjusted for no variables.
Model I adjusted for age, APS III, NUF rate, creatinine, RDW, PT and phosphate.
Model II adjusted for Model I plus MV and norepinephrine.
Matched cohort:
Unadjusted model adjusted for no variables.
Model I adjusted for age, SOFA score, APS III, platelet, RDW, PT, bilirubin, anion gap, NUF rate and PaO2.
Model II adjusted for Model I plus hypertension, cirrhosis, MV and norepinephrine.
Abbreviations: AKI: acute kidney injury, CRRT: continuous renal replacement therapy, LAR: lactate/albumin ratio, APS III: acute physiology score III, NUF: net ultrafiltration, RDW: red blood cell distribution width, PT: prothrombin time, MV: mechanical ventilation, SOFA: sequential organ failure assessment, PaO2: arterial partial pressure of oxygen, HR: hazard ratio, CI: confidence interval.
To ascertain the consistency of the findings, variables with a P-value less than 0.1 in the weighted cohort, were incorporated into a separate Cox regression analysis, as depicted in Table 3. The outcomes remained consistent, suggesting that a high LAR (> 1.641) was an autonomous risk factor for mortality due to any cause within 28 and 360 days among patients with AKI undergoing CRRT (p < 0.05).
Kaplan-Meier survival curves
The Kaplan-Meier survival curves, depicted in Figure 4, exhibited a gradual decline in cumulative survival rates at both 28 and 360 days as the LAR values increased (log-rank test, χ2 = 48.630, p < 0.001; χ2 = 33.530, p < 0.001). Notably, patients in the high LAR group had the lowest survival rates. These findings remained consistent even after conducting IPTW, as illustrated in Figure 5.
Figure 4.
Kaplan-Meier Survival curves for the cumulative survival rates at 28-day (a) and 360-day (b) across different levels of lactate/albumin ratio (LAR).
Figure 5.
Kaplan-Meier Survival curves for the cumulative survival rates at 28-day (a) and 360-day (b) after IPTW across different levels of lactate/albumin ratio (LAR).
Discussion
This retrospective study demonstrated a strong association between LAR values and all-cause mortality in AKI patients undergoing CRRT. Patients with high LAR values had a significantly worse prognosis. After adjusting for the other nine variables identified through LASSO analysis, Cox regression analysis confirmed that high LAR (>1.641) was an independent risk factor for all-cause mortality at 28 and 360 days in AKI patients undergoing CRRT (p < 0.05). These findings remained consistent even after IPTW adjustment, thereby ensuring a reliable and robust outcome. Furthermore, the Kaplan-Meier survival curves, both prior to and subsequent to the implementation of IPTW analysis, provided additional confirmation that there exist substantial disparities in survival rates among patients exhibiting varying levels of LAR. These outcomes implied that LAR possesses the potential to serve as an uncomplicated and dependable biomarker, aiding in the identification of individuals undergoing CRRT who are at a heightened susceptibility for AKI.
Lactate, a byproduct of cellular hypoxia and anaerobic metabolism, is frequently utilized as a measure of insufficient tissue perfusion and metabolic acidosis [20,21]. Within this framework, heightened blood lactate concentrations have demonstrated a strong correlation with unfavorable outcomes in critically ill individuals [22,23]. Nevertheless, the reliance solely on lactate levels may be subject to the influence of diverse physiological and pathological factors [24], thereby constraining its efficacy as an autonomous prognostic indicator. In addition to its role in maintaining colloid osmolality, albumin, a prominent plasma protein, serves as an indicator of nutritional status and systemic inflammatory response [25]. In critically ill patients, a decrease in albumin levels is often observed, potentially linked to heightened protein catabolism, diminished synthesis, or increased loss. Consequently, hypoalbuminemia is believed to be correlated with disease severity and a less favorable prognosis [15,26]. A recent systematic review has further demonstrated that hypoproteinemia is associated with a heightened susceptibility to developing AKI [27]. The LAR serves as an encompassing assessment tool that considers metabolic status, nutritional status, and inflammatory response. The fluctuation of this ratio may offer a more precise indication of the physiological and pathological condition of critically ill patients, surpassing the predictive capability of any individual indicator. Consequently, an escalating number of clinicians and researchers are devoting attention to this composite indicator.
A study conducted in Korea with a sample size of 3,499 individuals revealed that the LAR exhibited the highest AUROC for predicting 28-day mortality in sepsis, suggesting its utility as a valuable prognostic factor [12]. Prior investigations have also highlighted the significance of elevated LAR levels upon admission as a strong predictor of in-hospital mortality in heart failure following myocardial infarction, outperforming lactate levels, SOFA score, and other relevant prognostic markers [28]. One previous study, derived from the MIMIC-IV database, has also confirmed that high LAR (≥0.659) on ICU admission is an independent risk factor for 30-day and 360-day mortality in patients with AKI [29]. Nevertheless, there is currently a dearth of studies examining the correlation between LAR and AKI patients undergoing CRRT. From a pathophysiological perspective, individuals with AKI undergoing CRRT may exhibit heightened levels of tissue ischemia, cellular injury, and inflammatory reactions. These patients frequently endure more pronounced renal dysfunction and may additionally encounter multiple organ dysfunction syndrome (MODS). From a clinical perspective, patients with severe AKI who require CRRT represent a significant subgroup of AKI, often with more severe conditions and higher mortality rates, with short-term mortality exceeding 50% [30]. This group has received relatively less attention. To our knowledge, there have been no reports on the correlation between LAR levels and the prognosis of AKI patients requiring CRRT. Therefore, we conducted a retrospective study utilizing data from the MIMIC-IV database to investigate the association between LAR levels and the prognosis of 703 AKI patients undergoing CRRT. Cox regression analyses revealed that a high LAR (>1.641) independently contributed to the risk of mortality at 28 and 360 days in critically ill AKI patients undergoing CRRT, a finding robustly supported by IPTW adjustment. This discovery has the potential to assist clinicians in identifying patients at a higher risk, enabling them to intervene earlier and closely monitor their condition.
The results of this study underscore the significance of the LAR as a prognostic indicator for both short-term (28-day) and long-term (360-day) mortality. Despite the multitude of variables affecting survival over a year, the LAR has demonstrated its predictive utility as a biomarker, reflecting both acute physiological stress and chronic physiological depletion. Elevated lactate levels are indicative of tissue hypoperfusion and metabolic acidosis, with sustained elevation potentially indicating ongoing perfusion insufficiency. Furthermore, the LAR is linked to tumor metabolism and metabolic disorders, which may have implications for long-term prognoses [31–33]. Simultaneously, hypoalbuminemia is recognized as a prognostic indicator for long-term survival in medical inpatients, impacting one- and five-year mortality rates [34]. Consequently, the LAR offers a comprehensive assessment of a patient’s immediate and long-term health condition, providing valuable insights into prognostic implications. This finding is in alignment with existing research, including the use of serum lactate levels as an indicator to assess the one-year mortality risk in patients with acute pancreatitis [35], as well as an independent predictor of mortality within six months for patients with acute and chronic liver failure [36]. Additionally, LAR has shown a significant correlation with overall mortality within 90 days following an acute myocardial infarction [37]. These findings underscore the utility of LAR as a comprehensive risk assessment tool for acute and chronic conditions.
This study possesses multiple notable strengths. Firstly, it stands as the pioneering investigation into the association between the LAR and critically ill patients with AKI undergoing CRRT, thereby offering novel insights and data to enhance the scientific comprehension within this domain. Moreover, the study’s utilization of the IPTW technique to account for selection bias and confounding variables, coupled with the subsequent application of Cox regression and survival analysis, exemplified the meticulousness of the statistical methods employed and the durability of the findings obtained. As a retrospective study, it is important to acknowledge its inherent limitation in establishing causality, as it can only reveal correlation. Furthermore, this research utilized indicators from the initial day of patient admission to the ICU and did not incorporate parameters preceding the initiation of CRRT or implement ongoing monitoring throughout the hospital stay. While this approach facilitates early prediction to some extent, it also introduces a degree of bias. Lastly, given its single-center nature, caution should be exercised in generalizing the results, and their applicability to other regions or groups should be further investigated. Consequently, future research endeavors should consider employing prospective or randomized controlled trials to establish causality and broaden the scope of the study, thereby validating and reinforcing the obtained findings.
Conclusion
A high LAR (>1.641) was found to be an autonomous predictor of mortality at both 28 and 360 days in critically ill patients with AKI undergoing CRRT. This observation has the potential to assist clinicians in the early identification of high-risk patients, although it necessitates validation through additional research.
Acknowledgements
The authors appreciate the researchers at the MIT Laboratory for Computational Physiology for publicly sharing of the MIMIC-IV clinical database.
Funding Statement
The author(s) reported there is no funding associated with the work featured in this article.
Author contributions
Literature search: HL, JHL and JYZ; Study design: HL, JHL, and LZ; Data collection: JM and LZ; Data analysis: JM and LZ; Manuscript writing: HL, JM and JYZ. All authors read and approved the final manuscript.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
Publicly available datasets were analyzed in this study. The data is accessible at https://physionet.org/content/mimiciv/2.0/. The datasets utilized and examined during this study can be obtained from the corresponding author upon reasonable request.
References
- 1.Yang SY, Chiou TT, Shiao CC, et al. Nomenclature and diagnostic criteria for acute kidney injury - 2020 consensus of the Taiwan AKI-task force. J Formos Med Assoc. 2022;121(4):1–10. [DOI] [PubMed] [Google Scholar]
- 2.Turgut F, Awad AS, Abdel-Rahman EM.. Acute kidney injury: medical causes and pathogenesis. J Clin Med. 2023;12(1):375. doi: 10.3390/jcm12010375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hoste E, Kellum JA, Selby NM, et al. Global epidemiology and outcomes of acute kidney injury. Nat Rev Nephrol. 2018;14(10):607–625. doi: 10.1038/s41581-018-0052-0. [DOI] [PubMed] [Google Scholar]
- 4.Fujarski M, Porschen C, Plagwitz L, et al. Prediction of acute kidney injury in the intensive care unit: preliminary findings in a european open access database. Stud Health Technol Inform. 2022;294:139–140. [DOI] [PubMed] [Google Scholar]
- 5.Schmidt-Ott KM, Swolinsky J.. Prevention of acute kidney injury. Dtsch Med Wochenschr. 2022;147(5):236–245. doi: 10.1055/a-1609-0722. [DOI] [PubMed] [Google Scholar]
- 6.Komaru Y, Inokuchi R, Iwagami M, et al. Inverse correlation between incidence and mortality of acute kidney injury in critically ill patients: a systematic review. Shock. 2020;54(3):280–284. doi: 10.1097/SHK.0000000000001511. [DOI] [PubMed] [Google Scholar]
- 7.Baeg SI, Jeon J, Kang D, et al. Impact of protocolized fluid management on electrolyte stability in patients undergoing continuous renal replacement therapy. Front Med (Lausanne). 2022;9:915072. doi: 10.3389/fmed.2022.915072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Teixeira JP, Neyra JA, Tolwani A.. Continuous KRT: a contemporary review. Clin J Am Soc Nephrol. 2023;18(2):256–269. doi: 10.2215/CJN.04350422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Claure-Del Granado R, Clark WR.. Continuous renal replacement therapy principles. Semin Dial. 2021;34(6):398–405. doi: 10.1111/sdi.12967. [DOI] [PubMed] [Google Scholar]
- 10.Prasad B, Urbanski M, Ferguson TW, et al. Early mortality on continuous renal replacement therapy (CRRT): the prairie CRRT study. Can J Kidney Health Dis. 2016;3:36. doi: 10.1186/s40697-016-0124-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Guo W, Zhao L, Zhao H, et al. The value of lactate/albumin ratio for predicting the clinical outcomes of critically ill patients with heart failure. Ann Transl Med. 2021;9(2):118–118. doi: 10.21037/atm-20-4519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yoo KH, Choi SH, Suh GJ, et al. The usefulness of lactate/albumin ratio, C-reactive protein/albumin ratio, procalcitonin/albumin ratio, SOFA, and qSOFA in predicting the prognosis of patients with sepsis who presented to emergency departments. Am J Emerg Med. 2023;78:1–7. doi: 10.1016/j.ajem.2023.12.028. [DOI] [PubMed] [Google Scholar]
- 13.Chen S, Guan S, Yan Z, et al. The lactate to albumin ratio linked to all-cause mortality in critically ill patients with septic myocardial injury. Front Cardiovasc Med. 2023;10:1233147. doi: 10.3389/fcvm.2023.1233147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kim SG, Lee J, Yun D, et al. Hyperlactatemia is a predictor of mortality in patients undergoing continuous renal replacement therapy for acute kidney injury. BMC Nephrol. 2023;24(1):11. doi: 10.1186/s12882-023-03063-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Yang K, Yang N, Sun W, et al. The association between albumin and mortality in patients with acute kidney injury: a retrospective observational study. BMC Nephrol. 2023;24(1):332. doi: 10.1186/s12882-023-03323-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Goldberger AL, Amaral LA, Glass L, et al. PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals. Circulation. 2000;101(23):E215–20. doi: 10.1161/01.cir.101.23.e215. [DOI] [PubMed] [Google Scholar]
- 17.von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61(4):344–349. doi: 10.1016/j.jclinepi.2007.11.008. [DOI] [PubMed] [Google Scholar]
- 18.Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120(4):c179–84. doi: 10.1159/000339789. [DOI] [PubMed] [Google Scholar]
- 19.Murugan R, Kerti SJ, Chang CH, et al. Association of net ultrafiltration rate with mortality among critically ill adults with acute kidney injury receiving continuous venovenous hemodiafiltration: a secondary analysis of the randomized evaluation of normal vs augmented level (RENAL) of renal replacement therapy trial. JAMA Netw Open. 2019;2(6):e195418. doi: 10.1001/jamanetworkopen.2019.5418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Garcia-Alvarez M, Marik P, Bellomo R.. Sepsis-associated hyperlactatemia. Crit Care. 2014;18(5):503. doi: 10.1186/s13054-014-0503-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Emhoff CW, Messonnier LA.. Concepts of lactate metabolic clearance rate and lactate clamp for metabolic inquiry: a mini-review. Nutrients. 2023;15(14):3213. doi: 10.3390/nu15143213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Haas SA, Lange T, Saugel B, et al. Severe hyperlactatemia, lactate clearance and mortality in unselected critically ill patients. Intensive Care Med. 2016;42(2):202–210. doi: 10.1007/s00134-015-4127-0. [DOI] [PubMed] [Google Scholar]
- 23.Bruno RR, Wernly B, Flaatten H, et al. Lactate is associated with mortality in very old intensive care patients suffering from COVID-19: results from an international observational study of 2860 patients. Ann Intensive Care. 2021;11(1):128. doi: 10.1186/s13613-021-00911-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ramzan I, Khan MH, Sharma S, et al. Effects of circadian rhythm on anaerobic performance and blood lactate level: a systematic review. Sleep Breath. 2023;27(3):799–816. doi: 10.1007/s11325-022-02662-7. [DOI] [PubMed] [Google Scholar]
- 25.Klinkmann G, Waterstradt K, Klammt S, et al. Exploring albumin functionality assays: a pilot study on sepsis evaluation in intensive care medicine. Int J Mol Sci. 2023;24(16):12551. doi: 10.3390/ijms241612551. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Schupp T, Behnes M, Rusnak J, et al. Does albumin predict the risk of mortality in patients with cardiogenic shock. Int J Mol Sci. 2023;24(8):7375. doi: 10.3390/ijms24087375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hansrivijit P, Yarlagadda K, Cheungpasitporn W, et al. Hypoalbuminemia is associated with increased risk of acute kidney injury in hospitalized patients: A meta-analysis. J Crit Care. 2021;61:96–102. doi: 10.1016/j.jcrc.2020.10.013. [DOI] [PubMed] [Google Scholar]
- 28.Chen Y, Yang K, Wu B, et al. Association between lactate/albumin ratio and mortality in patients with heart failure after myocardial infarction. ESC Heart Fail. 2023;10(3):1928–1936. doi: 10.1002/ehf2.14359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shi X, Zhong L, Lu J, et al. Clinical significance of the lactate-to-albumin ratio on prognosis in critically ill patients with acute kidney injury. Ren Fail. 2024;46(1):2350238. doi: 10.1080/0886022X.2024.2350238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Griffin BR, Liu KD, Teixeira JP.. Critical care nephrology: core curriculum 2020. Am J Kidney Dis. 2020;75(3):435–452. doi: 10.1053/j.ajkd.2019.10.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Niu D, Wu Y, Lei Z, et al. Lactic acid, a driver of tumor-stroma interactions. Int Immunopharmacol. 2022;106:108597. doi: 10.1016/j.intimp.2022.108597. [DOI] [PubMed] [Google Scholar]
- 32.Gao Y, Zhou H, Liu G, et al. Tumor microenvironment: lactic acid promotes tumor development. J Immunol Res. 2022;2022:3119375–3119378. doi: 10.1155/2022/3119375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Broskey NT, Pories WJ, Jones TE, et al. The association between lactate and muscle aerobic substrate oxidation: is lactate an early marker for metabolic disease in healthy subjects? Physiol Rep. 2021;9(3):e14729. doi: 10.14814/phy2.14729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Oster HS, Dolev Y, Kehat O, et al. Serum hypoalbuminemia is a long-term prognostic marker in medical hospitalized patients, irrespective of the underlying disease. J Clin Med. 2022;11(5):1207. doi: 10.3390/jcm11051207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Zeng Z, Huang R, Lin H, et al. Serum lactate is an indicator for short-term and long-term mortality in patients with acute pancreatitis. Dig Dis Sci. 2024;69(6):2223–2234. doi: 10.1007/s10620-024-08419-4. [DOI] [PubMed] [Google Scholar]
- 36.Nie Y, Zhang Y, Liu LX, et al. Serum lactate level predicts short-term and long-term mortality of hbv-aclf patients: a prospective study. Ther Clin Risk Manag. 2020;16:849–860. doi: 10.2147/TCRM.S272463. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Wang D, Luo C, Li Q, et al. Association between lactate/albumin ratio and all-cause mortality in critical patients with acute myocardial infarction. Sci Rep. 2023;13(1):15561. doi: 10.1038/s41598-023-42330-8. [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
Publicly available datasets were analyzed in this study. The data is accessible at https://physionet.org/content/mimiciv/2.0/. The datasets utilized and examined during this study can be obtained from the corresponding author upon reasonable request.