Abstract
Acute kidney injury (AKI) following total aortic arch replacement (TAAR) is linked to high morbidity and mortality, and timely identification remains a clinical challenge. While high serum lactate levels are associated with renal impairment, more evidence is required to establish lactate as a reliable biomarker for AKI. Therefore, this study aimed to evaluate the predictive value of postoperative serum lactate at intensive care unit (ICU) admission for AKI in TAAR patients. A retrospective analysis was conducted on adult TAAR patients admitted to the ICU of Guangdong Provincial People’s Hospital between January 2021 and December 2023. AKI was diagnosed according to Kidney Disease Improving Global Outcomes (KDIGO) criteria. The predictive value of lactate was assessed using multivariate logistic regression and area under the curve receiver operating characteristic (AUC-ROC) analysis, followed by the development and internal validation of a predictive model. Among the 356 enrolled patients, 232 (65%) developed AKI. Serum lactate levels were significantly higher in AKI patients [median (IQR): 3.90 (2.75–6.60) vs. 2.10 (1.40–3.60) mmol/L, p < 0.001]. Analysis revealed a graded association between serum lactate tertiles and AKI risk. Incorporating lactate significantly improved the model’s predictive performance, increasing the AUC-ROC from 0.757 to 0.824 (p < 0.001) and enhancing both reclassification (NRI = 0.141, p = 0.007) and discriminative ability (IDI = 0.109, p < 0.001). In conclusion, elevated serum lactate at ICU admission is independently associated with postoperative AKI, improving early risk stratification and guiding ICU management in high-risk TAAR patients.
Keywords: Acute kidney injury, lactate, intensive care unit, total aortic arch replacement, biomarkers
Introduction
Type A aortic dissection (TAAD) is a life-threatening cardiovascular disease that can be treated primarily with total aortic arch replacement (TAAR) [1]. In general, there is a high incidence of acute kidney injury (AKI) after TAAR surgery (50–70%) [2]. This complication significantly prolongs intensive care unit (ICU) and hospital stays while elevating treatment expenditures [3,4]. The pathogenesis of cardiac surgery-associated AKI (CSA-AKI) involves multifactorial mechanisms, including renal ischemia-reperfusion injury (IRI), inflammatory reactions, metabolic disturbances, and various other factors [5–7]. Given the current absence of definitive CSA-AKI therapies, it is critical to develop biomarkers to assist in timely interventions, mitigate the progression of AKI, promote renal recovery, and improve survival.
Lactate, as a product of anaerobic glycolysis, increases significantly in cases of hypoxia, stress, etc. [8]. Elevated serum lactate serves as a validated biomarker for mortality risk in critically ill patients, and recent studies have confirmed its role as an independent prognostic factor [9]. The elevation of lactate levels corresponds to a poorer prognosis [10]. In some studies, various lactate cutoff values have been proposed as indicators for predicting early aggressive intervention outcomes [10–13]. Moreover, elevated lactate levels may reflect the severity of the underlying injury and predict an adverse outcome. Among patients with newly developed AKI who are receiving continuous renal replacement therapy (CRRT), elevated serum lactate portends adverse outcomes as an independent predictor [14–16]. Lactate has associated predictive efficacy in patients with sepsis-related AKI requiring CRRT [16,17]. Recent studies have shown that elevated intraoperative lactate concentrations and transfusion requirements constitute independent predictors of postoperative AKI, significantly increasing in-hospital mortality risk [18]. Collectively, these findings highlight the promising potential of lactate as a predictor of AKI development.
However, the impact of serum lactate levels at ICU admission on post-TAAR AKI risk remains undefined. This study evaluated the association between serum lactate levels and the development of postoperative AKI in adult patients who underwent TAAR, as well as to explore the potential prognostic value of lactate in predicting AKI onset. Importantly, by incorporating variables available immediately upon ICU admission, our predictive approach is intended as an early postoperative risk stratification tool to guide timely ICU management.
Methods
Study cohort
Conducted at Guangdong Provincial People’s Hospital, this retrospective study focused on ICU patients (aged 18 years or older) undergoing TAAR from January 2021 to December 2023. Exclusion criteria included patients under 18 years old, patients diagnosed with end-stage renal disease (ESRD), those receiving renal replacement therapy (RRT) before admission, individuals with prior renal transplantation or nephrectomy, pregnant women, those whose stay in the CSICU was less than 24 h, and patients with missing admission clinical data. Clinical data were retrieved from medical records.
This study received approval from the Hospital’s Ethics Committee (approval registration number: KY2024-731-01) and adhered to the principles outlined in the Declaration of Helsinki.
Data source
Baseline characteristics of the enrolled patients were extracted from the electronic medical record system, with the following variables being included as candidate predictors: demographic characteristics, including sex, age, weight, and smoking status; preexisting clinical conditions, including hypertension, diabetes, previous cardiac surgery, chronic kidney disease, coronary heart disease, liver disease, cerebrovascular disease and hyperlipidemia; preoperative laboratory tests, including D-dimer, albumin, hemoglobin, hematocrit, baseline glomerular filtration rate (eGFR) calculated by the Chronic Kidney Disease – Epidemiology Collaborative Group (CKD-EPI) formula [19] and baseline serum creatinine; American Society of Anesthesiologists (ASA) classification; New York Heart Association (NYHA) classification; preoperative cardiovascular status, including left ventricular end-diastolic diameter (LVDD), ascending inner aortic (AA) diameter and left ventricular ejection fraction (LVEF); preoperative medication, including nephrotoxic drugs, mannitol and contrast media; perioperative medication, including norepinephrine, adrenaline and dopamine; intraoperative data, including extracorporeal circulation (CPB) time; operation time; aortic occlusion (ACC) time; volume of blood loss; urine volume; Intra-aortic balloon pump (IABP); intraoperative transfusion; emergency surgery; APACHE II score at ICU admission; and serum lactate levels at ICU admission immediately after surgery.
Definitions and outcomes
This study defined postoperative AKI occurring within 7 days after ICU admission as the primary outcome. The secondary outcomes included severe AKI, the application of RRT or extracorporeal membrane oxygenation (ECMO) during ICU stay, duration of mechanical ventilation, total hospitalization costs, hospital length of stay, ICU length of stay, and in-hospital mortality.
In line with the latest diagnostic criteria for cardiac surgery-related AKI, postoperative AKI was defined in accordance with the Kidney Disease: Improving Global Outcomes (KDIGO) criteria [20,21], requiring either: sCr elevation ≥0.3 mg/dl (≥26.5 µmol/L) within 48 h, sCr ≥1.5 times within 7 days, or sustained urine output <0.5 mL/kg/h over 6 h. AKI severity was staged according to KDIGO guidelines: Stage 1: sCr elevation ≥0.3 mg/dl, a 1.5–1.9-fold increase compared to baseline, or oliguria (<0.5 mL/kg/h) lasting 6–12 h. Stage 2: A 2.0–2.9-fold rise in sCr, or oliguria persisting for 12 h. Stage 3: sCr ≥3 times, sCr ≥4.0 mg/dl (≥353.6 µmol/L), RRT initiation, urine output <0.3 mL/kg/h for more than 24 h, or anuria lasting longer than 12 h.
The RRT types include CRRT, intermittent hemodialysis, and continuous ambulatory peritoneal dialysis. Stages 2 and 3 are considered severe AKI. In the ICU, urine output is routinely recorded, and sCr levels are measured daily as a standard practice. Baseline sCr was defined as the most recent measured sCr value within 7 days before surgery. In cases where multiple daily measurements were obtained, the highest reading among them was selected.
Statistical analyses
First, patients were stratified into AKI and non-AKI cohorts based on the occurrence of AKI following TAAR surgery. The clinical characteristics and outcomes of two groups were then explored comparatively. Second, to evaluate graded clinical associations, serum lactate levels at admission were categorized into tertiles, with the ranges defined as <2.4 mmol/L (tertile 1), 2.4–4.7 mmol/L (tertile 2), and >4.7 mmol/L (tertile 3), respectively. In this study, categorical variables are presented as counts and percentages. Continuous variables with non-normal distributions are described using medians and interquartile ranges, while those with normal distributions are expressed as means and standard deviations. For the comparison of categorical variables, either the chi-square test or Fisher’s exact test was employed, depending on the data characteristics. For nonnormally distributed continuous variables, between-group comparisons were performed using the Mann–Whitney U test or Kruskal–Wallis test. For normally distributed continuous variables, ANOVA and t tests were used. To assess associations between clinical outcomes and serum lactate levels (analyzed both as a continuous measure and categorized into tertiles), lactate levels were categorized as needed. Subsequently, logistic regression analysis was conducted.
Variables demonstrating univariate significance (p < 0.1) were entered into a multivariate logistic regression analysis. This process identified independent predictors and constructed the AKI prediction model (referred to as Model 1). All analyses were performed using complete-case methodology based on the 356 patients with complete data, and no imputation was applied. Afterwards, serum lactate, the selected biomarker, was integrated into Model 1, leading to the development of Model 2. To assess the value of lactate in early risk stratification for AKI, the following approaches were used in this study: 1. Receiver operating characteristic (ROC) curves were generated for both models. The area under the ROC curve (AUC-ROC) was utilized to evaluate whether lactate could enhance the accuracy of AKI prediction, while model comparison relied on the DeLong test for AUC differences [22]. 2. Calibration curves depicted the alignment between model-predicted outcomes and observed results. To evaluate the optimism in model calibration and obtain bias-corrected estimates, we performed bootstrap resampling with 1000 iterations. In each bootstrap sample, the model was refitted, and calibration was assessed. The bootstrap‑adjusted calibration curve, representing the performance corrected for overfitting, was derived by averaging the optimism across bootstrap samples and subtracting it from the apparent calibration. A 45° diagonal indicated perfect fit, and curve proximity to this benchmark assessed model goodness‑of‑fit. 3. Decision curve analysis (DCA) [23] quantified clinical utility by integrating decision-maker preferences into net benefit calculations, overcoming limitations of traditional metrics. DCA was performed across the full range of threshold probabilities from 0 to 1. The threshold probability range where each model demonstrated net benefit superior to both the ‘treat-all’ and ‘treat-none’ strategies was identified and reported. This range corresponds to clinically relevant risk levels for AKI after TAAR surgery, where intervention decisions are typically considered. Consequently, this method evaluated both models’ clinical applicability. 4. Predictive performance of two models relied on the net reclassification improvement (NRI) and integrated discrimination improvement (IDI). We used the continuous NRI, which does not require predefined risk thresholds and quantifies the extent to which the new model (Model 2) improves the correct prediction of AKI probability after incorporating the new variable. IDI reflects the overall improvement in predicting AKI probability. The 95% confidence intervals for NRI and IDI were derived using bootstrap resampling with 1000 iterations, and statistical significance was determined by whether the confidence interval excluded zero.
To assess the stability and generalizability of our models, we performed three complementary internal validation approaches. First, we conducted repeated random split‑sample validation with 100 iterations. In each iteration, the dataset was randomly partitioned into a training set (70%) and a validation set (30%). Both Model 1 and Model 2 were refitted in the training set, and their performance was evaluated in the corresponding validation set. The area under the receiver operating characteristic curve (AUC‑ROC) was calculated for each iteration, and the mean AUC with standard deviation was reported for both models in the training and validation sets. Second, to compare the overall performance of Model 1 and Model 2 on the full dataset and to obtain a robust estimate of the difference in their discriminative ability, we performed bootstrap validation with 1000 resamples. In each bootstrap sample, the AUC difference (Model 2 – Model 1) was computed, and the 95% confidence interval was derived using the bias‑corrected and accelerated (BCa) method to test whether Model 2 significantly outperforms Model 1. To further quantify potential overfitting, we performed optimism correction using bootstrap resampling with 1000 iterations. In each bootstrap sample, the model was refitted, and performance metrics were calculated both in the bootstrap sample and in the original dataset. This approach provides a more realistic assessment of model performance in future similar populations. Across all analyses, statistical significance was defined as a two-tailed P value < 0.05. All data processing utilized R version 4.4.2 and SPSS 27.0.
Results
Baseline characteristics and results
Between January 2021 and December 2023, 407 TAAR patients were initially recruited. After excluding 2 patients under 18 years of age, 3 with prior renal transplantation or nephrectomy recipients, 6 patients who developed ESRD or RRT, 2 with ICU stays under 24 h, and 38 with missing admission clinical data, a final cohort of 356 adult patients was included in the analysis. Over a 7-day continuous observation period, 232 patients (65%) developed AKI, among whom 125 (54%) progressed to severe AKI during hospitalization (Figure 1). Baseline characteristics and predicted outcomes of the patients are summarized in Table 1. Compared with patients without AKI, those who developed AKI tended to be older and had a higher proportion of males. Additionally, the AKI group showed a higher prevalence of hypertension, diabetes, and a history of emergency surgery. Patients with AKI tend to have higher D-dimer levels and lower baseline eGFRs than patients without AKI. Patients with AKI have greater blood loss, less urine output, and greater transfusion requirements. Patients with AKI tend to have higher APACHE II scores and serum lactate levels at ICU admission. In the AKI group, adverse outcomes, including higher RRT use, prolonged mechanical ventilation, higher mortality, higher total hospital costs, and extended ICU and hospital stays, were more likely to occur.
Figure 1.
Flowchart for patient selection.
Table 1.
Patient characteristics by AKI status.
| Characteristic | All patients (n = 356) | Non-AKI (N = 124) | AKI (N = 232) | P value |
|---|---|---|---|---|
| Male, n (%) | 282 (79) | 92 (74) | 190 (82) | 0.088 |
| Age, years | 53 (44,61) | 51 (41,58) | 54 (45,61) | 0.050 |
| Weight, kg | 70 (63,76) | 70 (62,77) | 70 (65,76) | 0.471 |
| Smoker, n (%) | 122 (34) | 49 (40) | 73 (32) | 0.159 |
| Preexisting clinical conditions, n (%) | ||||
| Hypertension | 220 (62) | 63 (51) | 157 (68) | 0.002 |
| Diabetes | 14 (4) | 2 (2) | 12 (5) | 0.174 |
| Previous cardiac surgery | 22 (6) | 6 (5) | 16 (7) | 0.591 |
| Chronic kidney disease | 6 (2) | 2 (2) | 4 (2) | 1.000 |
| Coronary artery disease | 12 (3) | 9 (7) | 3 (1) | 0.005 |
| Liver disease | 15 (4) | 7 (6) | 8 (3) | 0.480 |
| Cerebrovascular disease | 13 (4) | 6 (5) | 7 (3) | 0.564 |
| Hyperlipidemia | 86 (24) | 23 (19) | 63 (27) | 0.093 |
| Preoperative laboratory tests | ||||
| Hemoglobin, g/L | 127 (113,139) | 126 (113,138) | 127 (114,139) | 0.851 |
| Hematocrit, % | 0.36 (0.31,0.40) | 0.36 (0.32,0.40) | 0.35 (0.30,0.40) | 0.211 |
| D-dimer, ng/mL | 5475 (2135,16145) | 3260 (1715,10112) | 6620 (2465,19550) | <0.001 |
| Albumin, g/L | 37.86 (35.38,40.54) | 37.62 (35.46,40.66) | 38.08 (35.38,40.52) | 0.691 |
| Baseline eGFR, mL/min/1.73 m2 | 92.24 (67.95,103.27) | 97.02 (80.09,108.14) | 85.84 (63.89,100.71) | <0.001 |
| Baseline serum creatinine, μmol/L | 80.40 (65.87,103.56) | 75.29 (60.53,93.62) | 84.73 (69.19,111.69) | <0.001 |
| Preoperative cardiovascular status | ||||
| LVDD, mm | 47 (43,51) | 47 (44,51) | 47 (43,51) | 0.580 |
| LVEF, % | 64 (61,67) | 65 (62,67) | 64 (60,67) | 0.111 |
| AA diameter, mm | 42 (38,45) | 42 (37,45) | 42 (38,45) | 0.539 |
| ASA ≥ III grade, n (%) | 353 (99) | 122 (99) | 231 (99) | 0.580 |
| NYHA ≥ III grade, n (%) | 185 (52) | 63 (51) | 122 (53) | 0.835 |
| Preoperative medication, n (%) | ||||
| Nephrotoxic drugsa | 178 (50) | 52 (42) | 126 (54) | 0.035 |
| Mannitol | 297 (83) | 101 (82) | 196 (85) | 0.560 |
| Contrast media | 16 (5) | 4 (3) | 12 (5) | 0.565 |
| Perioperative medication, n (%) | ||||
| Norepinephrine | 89 (25) | 25 (20) | 64 (28) | 0.158 |
| Adrenaline | 249 (70) | 87 (70) | 162 (70) | 1.000 |
| Dopamine | 86 (24) | 20 (16) | 66 (28) | 0.014 |
| Intraoperative information | ||||
| CPB time, min | 224 (191,268) | 207 (183,238) | 235 (196,285) | <0.001 |
| ACC time, min | 117 (96,146) | 114 (93,136) | 121 (98,150) | 0.041 |
| IABP, n (%) | 3 (1) | 1 (1) | 2 (1) | 1.000 |
| Operation time, min | 400 (360,465) | 400 (343,435) | 415 (377,484) | <0.001 |
| Volume of blood loss, mL | 400 (300,500) | 400 (300,500) | 400 (300,500) | 0.019 |
| Urine volume, mL | 1300 (700,1850) | 1500 (1000,2225) | 1200 (600,1600) | <0.001 |
| Intraoperative transfusion | ||||
| Received RBC, Units | 4 (0,6) | 4 (0,5) | 4 (0,8) | 0.012 |
| Received FFP, mL | 350 (0,400) | 0 (0,400) | 400 (0,600) | <0.001 |
| Received PLT, Units | 1 (1,2) | 1 (1,1) | 1 (1,2) | 0.002 |
| Crystalloid, mL | 0 (0,250) | 0 (0,250) | 0 (0,250) | 0.251 |
| Colloid, mL | 800 (500,1000) | 800 (500,1000) | 800 (500,1000) | 0.103 |
| Emergent surgery, n (%) | 267 (75) | 86 (70) | 181 (78) | 0.072 |
| APACHE II score at ICU admission | 17 (12,23) | 13 (11,18) | 19 (15,25) | <0.001 |
| Serum lactate at ICU admission, mmol/L | 3.40 (2.10,5.35) | 2.10 (1.40,3.60) | 3.90 (2.75,6.60) | <0.001 |
| Outcomes | ||||
| Severe AKI, n (%) | 125 (35) | 0 (0) | 125 (54) | <0.001 |
| RRT during ICU stay, n (%) | 42 (12) | 2 (1.6) | 40 (17) | <0.001 |
| ECMO during ICU stay, n (%) | 5 (1.4) | 0 (0) | 5 (2.2) | 0.200 |
| Mechanical ventilation, hours | 64.38 (23.16,136.34) | 31.15 (18.07,96.56) | 94.84 (37.93,160.50) | <0.001 |
| Hospital length of stay, days | 17.51 (13.95,22.82) | 15.14 (12.05,19.32) | 18.45 (15.23,23.90) | <0.001 |
| ICU length of stay, hours | 145.03 (90.97,227.53) | 116.58 (69.00,157.68) | 172.89 (103.93,257.98) | <0.001 |
| Total hospital cost, CNY | 231510.72 (201825.79,277769.93) | 210667.54 (186729.55,235618.37) | 251133.80 (213067.03,313807.26) | <0.001 |
| Hospital mortality, n (%) | 8 (2.2) | 0 (0) | 8 (3.4) | 0.055 |
Categorical variables are presented as counts (percentages). Continuous variables are expressed as means (± standard deviation) or medians (25th–75th percentiles). When P values for continuous variables are calculated using the Mann-Whitney U test, this test compares distributions rather than medians alone. Identical medians and interquartile ranges may still be associated with significant P values if the overall distributions differ. AKI, acute kidney injury; eGFR, estimated glomerular filtration rate; ICU, intensive care unit; LVDD, left ventricular end-diastolic dimension; LVEF, left ventricular ejection fraction; AA ascending aorta; ASA, American Society of Anesthesiologists; NYHA, New York Heart Association; aIncludes any of the following medications administered within 5 days prior to surgery: nonsteroidal anti-inflammatory drug, immunosuppressant, aminoglycoside, vancomycin, acyclovir, amphotericin or colistin; RBC, red blood cell; FFP, fresh frozen plasma; PLT, platelets; CPB, cardiopulmonary bypass time; ACC aortic cross-clamping; IABP intra-aortic balloon pump; APACHE, Acute Physiology and Chronic Health Evaluation; RRT, renal replacement therapy; ECMO, extracorporeal membrane oxygenation.
Serum lactate levels and AKI
Univariable and multivariable logistic regression analyses assessed associations between serum lactate levels and the risk of AKI (Table 2). Elevated lactate demonstrated a significant independent relationship with AKI onset (adjusted OR 1.530, 95% CI: 1.304–1.794, p < 0.001). Patients were stratified into tertiles on the basis of their serum lactate levels at ICU admission. Stratification revealed progressively increasing AKI incidence: The lowest tertile exhibited a 36% incidence, while it reached to 75% in the middle tertile (adjusted OR 6.945, 95% CI: 3.508–13.750, p < 0.001). In the highest tertile, it reached to 85% (adjusted OR 9.812, 95% CI: 4.427–21.748, p < 0.001). Using the lowest tertile as reference, the risk gradient corresponded with rising lactate concentrations (p < 0.001).
Table 2.
Logistic regression analyses of serum lactate levels at ICU admission and AKI.
| Lactate | Unadjusted |
Adjusted* |
||
|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |
| Continuous | 1.552 (1.356,1.776) | <0.001 | 1.530 (1.304,1.794) | <0.001 |
| Tertiles | ||||
| T1 | REF | REF | ||
| T2 | 5.302 (3.052,9.212) | <0.001 | 6.945 (3.508,13.750) | <0.001 |
| T3 | 9.981 (5.278,18.873) | <0.001 | 9.812 (4.427,21.748) | <0.001 |
AKI, acute kidney injury; OR, odds ratio; CI, confidence interval; REF, reference; *Adjusted for male, age, hypertension, coronary artery disease, hyperlipidemia, D-dimer, baseline eGFR, nephrotoxic drugs, dopamine, received RBC, received FFP, received PLT, CPB time, ACC time, operation time, emergent surgery, volume of blood loss, urine volume, APACHE II score at ICU admission. eGFR, estimated glomerular filtration rate; RBC, red blood cell; FFP, fresh frozen plasma; PLT, platelets; CPB, cardiopulmonary bypass time; ACC aortic cross-clamping; APACHE, Acute Physiology and Chronic Health Evaluation.
Moreover, when serum lactate levels were considered as continuous variables, elevated lactate levels were associated with the occurrence of both AKI (p < 0.001) and severe AKI (p < 0.001), as detailed in Table 3. Multivariate-adjusted analyses confirmed that serum lactate levels measured at ICU admission served as independent risk factors for both AKI (p < 0.001) and severe AKI (p = 0.003) in patients who underwent TAAR. Specifically, for each 1 mmol/L rise in serum lactate, the risk of developing AKI increased by a factor of 1.530 (adjusted OR: 1.530; 95% CI: 1.304–1.794), while the risk of severe AKI rose by 1.151 times (adjusted OR: 1.151; 95% CI: 1.050–1.261). In addition, when serum lactate levels were considered as categorical variables, compared to the lowest tertile cohort, patients in the highest tertile exhibited a 9.812-fold elevated AKI incidence (adjusted OR, 9.812; 95% CI, 4.427–21.748, p < 0.001) and a 3.597-fold higher severe AKI probability (adjusted OR, 3.597; 95% CI, 1.736–7.453, p < 0.001).
Table 3.
Association between lactate and postoperative outcomes.
| Lactate |
Odds ratio (95% CI) |
||||||
|---|---|---|---|---|---|---|---|
| Outcomes | Model | OR (95%CI) | P value | T1 (<2.4 mmol/L) |
T2 (2.4–4.7 mmol/L) |
T3 (>4.7 mmol/L) |
P value |
| AKI | N events/N participants | 232/356 | N/A | 43/119 | 93/124 | 96/113 | N/A |
| Univariable | 1.552 (1.356,1.776) | <0.001 | Reference | 5.302 (3.052,9.212) | 9.981 (5.278,18.873) | <0.001 | |
| Multivariable | 1.530 (1.304,1.794) | <0.001 | Reference | 6.945 (3.508,13.750) | 9.812 (4.427,21.748) | <0.001 | |
| Severe AKI | N events/N participants | 125/356 | N/A | 19/119 | 49/124 | 57/113 | N/A |
| Univariable | 1.230 (1.138,1.329) | <0.001 | Reference | 3.439 (1.871,6.319) | 5.357 (2.900,9.896) | <0.001 | |
| Multivariable | 1.151 (1.050,1.261) | 0.003 | Reference | 2.994 (1.493,6.003) | 3.597 (1.736,7.453) | <0.001 | |
Odds ratio (95% confidence intervals) from multivariable logistic regression models were adjusted for male, age, hypertension, coronary artery disease, hyperlipidemia, D-dimer, baseline eGFR, nephrotoxic drugs, dopamine, received RBC, received FFP, received PLT, CPB time, ACC time, operation time, emergent surgery, volume of blood loss, urine volume, APACHE II score at ICU admission. AKI, acute kidney injury; OR, odds ratio; CI, confidence interval; eGFR, estimated glomerular filtration rate; RBC, red blood cell; FFP, fresh frozen plasma; PLT, platelets; CPB, cardiopulmonary bypass time; ACC aortic cross-clamping; APACHE, Acute Physiology and Chronic Health Evaluation.
Relative contribution of serum lactate to the clinical risk model
Table 4 presents the independent predictive factors for AKI identified through logistic regression analysis. Logistic regression analysis included 20 candidate variables (detailed in Table 2 footnotes). Through multivariate logistic regression, the 6 most optimal variables were selected for inclusion. The predictors selected were age, preexisting hypertension, intraoperative urine output, intraoperative platelet transfusion, serum lactate level at ICU admission and APACHE II score at ICU admission. The established risk factors incorporated into Model 1 comprised age, preexisting hypertension, intraoperative urine output, intraoperative platelet transfusion, and APACHE II score at ICU admission. Model 2 was then developed by adding the serum lactate level to this baseline model (Model 1). Notably, the odds ratios for urine output in Table 4 represent the change in AKI odds per 100 mL increase in intraoperative urine output.
Table 4.
Multivariable logistic regression analysis for AKI.
| Prediction model and component | OR | 95% CI | P value |
|---|---|---|---|
| Model 1 | |||
| Age | 0.995 | 0.974,1.016 | 0.620 |
| Hypertension | 1.988 | 1.198,3.300 | 0.008 |
| Received PLT, u | 1.208 | 1.038,1.407 | 0.015 |
| Urine volume (per 100 mL) | 0.971 | 0.949,0.992 | 0.008 |
| APACHE II score at ICU admission | 1.143 | 1.091,1.197 | <0.001 |
| Model 2 | |||
| Age | 1.011 | 0.987,1.034 | 0.373 |
| Hypertension | 1.780 | 1.036,3.058 | 0.037 |
| Received PLT, u | 1.214 | 1.039,1.420 | 0.015 |
| Urine volume (per 100 mL) | 0.972 | 0.950,0.994 | 0.015 |
| APACHE II score at ICU admission | 1.113 | 1.059,1.169 | <0.001 |
| Serum lactate levels at ICU admission, mmol/L | 1.515 | 1.306,1.759 | <0.001 |
Urine volume was scaled to 100 mL units. The ORs of 0.971 (Model 1) and 0.972 (Model 2) represent the change in AKI odds for each 100 mL increase in intraoperative urine output, corresponding to approximately a 2.9% and 2.8% reduction in AKI risk, respectively. AKI, acute kidney injury; OR, odds ratio; CI, confidence interval; PLT, platelets; APACHE, Acute Physiology and Chronic Health Evaluation; REF, reference.
To more comprehensively explore the clinical value of incorporating serum lactate levels into Model 1, a series of validation analyses were conducted. Figure 2A presents ROC curves for Model 1 and Model 2. The lactate cutoff point for the most predictive occurrence of AKI was 2.450 mmol/L. Model 1 showed a reasonable ability to predict postoperative AKI, with an AUC-ROC of 0.757. When serum lactate was added, the AUC-ROC increased significantly to 0.824. DeLong testing confirmed a significant difference (p < 0.001), as detailed in Table 5. Both models maintained high calibration accuracy (Figure 2B,C), while DCA revealed superior net benefits for Model 2 (0.30–0.95) within a wider threshold probability range than Model 1 (0.38–0.95), outperforming treat-all or no-treat strategies (Figure 2D,E). Reclassification metrics including NRI and IDI further validated Model 2’s clinical utility. (Table 5). The findings indicated that the NRI was 0.141 (95% CI, 0.039–0.243). This suggests that incorporating lactate into Model 1 generated a statistically significant improvement in AKI case reclassification (p = 0.007). Likewise, the IDI was 0.109 (95% CI, 0.076 to 0.142), which demonstrates that adding lactate to Model 1 resulted in a significant improvement in AKI prediction relative to Model 1 on its own (p < 0.001).
Figure 2.
ROC analysis, calibration curves and DCA for Model 1 and Model 2 in predicting postoperative AKI. Plot (A) presents the ROC analysis results of Model 1 and Model 2. The AUC-ROC for Model1 and Model2 were 0.757 (95% CI: 0.708–0.807, p < 0.001) and 0.824 (95% CI: 0.781–0.867, p < 0.001). Plots (B) and (C) display the calibration curves of Model 1 and Model 2, respectively. Calibration curves depicted the alignment between model-predicted outcomes and observed results, where a 45° diagonal indicated perfect fit. The thin dashed line and solid line denote the actual prediction curve and the bootstrap-adjusted curve, respectively. A closer fit of the solid line to the 45° line indicates better predictive accuracy of the model. Plots (D) and (E) show DCA of Model 1 and Model 2 for predicting postoperative AKI. The two curves are compared to the curves of treating none and all patients. AKI, acute kidney injury; AUC, area under the receiver operating characteristic curve; CI, confidence interval; DCA, decision curve analysis; ROC, receiver operating characteristic.
Table 5.
Discrimination and reclassification of the combination of serum lactate for predicting AKI.
| AUC-ROC (95% CI) |
P value* | NRI (95% CI)* | P value* | IDI (95% CI)* | P value* | |
| Model 1 | 0.757 (0.708,0.807) | Reference | Reference | |||
| Model 2 | 0.824 (0.781,0.867) | <0.001 | 0.141 (0.039,0.243) | 0.007 | 0.109 (0.076,0.142) | <0.001 |
Model 1 included age, hypertension, received PLT, urine volume and APACHE II score at ICU admission. Model 2 included age, hypertension, received PLT, urine volume, APACHE II score at ICU admission and serum lactate levels at ICU admission. AKI, acute kidney injury; CI, confidence interval; PLT, platelets; APACHE, Acute Physiology and Chronic Health Evaluation; NRI, net reclassification improvement; IDI, integrated discrimination improvement; *Comparisons between Model 1 and Model 2.
The repeated random split‑sample validation (100 iterations) demonstrated consistent performance of both models. For Model 2 (the lactate‑enhanced model), the mean AUC in the training sets was 0.828 (SD = 0.015), and the mean AUC in the validation sets was 0.806 (SD = 0.037), indicating good discrimination and stability across different data partitions. For Model 1, the mean training AUC was 0.763 (SD = 0.017) and the mean validation AUC was 0.736 (SD = 0.042). Bootstrap validation with 1000 resamples further confirmed that Model 2 significantly outperformed Model 1. The 95% BCa confidence interval for the AUC difference (Model 2 – Model 1) was 0.0295 to 0.1034, and the bootstrap p‑value was <0.001, indicating that the improvement in discriminative ability with the addition of lactate was highly statistically significant and robust to data sampling variability. Optimism‑corrected bootstrap validation further confirmed the robustness of Model 2 against overfitting. With 1000 resamples, the apparent AUC was 0.824 and the optimism‑corrected AUC was 0.823. The negligible difference between the apparent and optimism‑corrected AUCs (0.824 vs. 0.823) indicates minimal overfitting, supporting the stability and reliability of our model.
Discussion
This study examined the association between serum lactate levels at ICU admission and postoperative AKI in adult patients who underwent TAAR surgery. The findings revealed that serum lactate elevation exhibited a significant association with AKI incidence after TAAR. Moreover, the OR of AKI gradually increased as the serum lactate level at ICU admission gradually increased. This correlation remained significant even after multivariable adjustment. In addition, incorporating serum lactate levels as a variable improved the predictive power for AKI. Consequently, serum lactate levels at ICU admission offer clinical utility as a valuable biomarker for AKI detection in TAAR surgical management. It is important to clarify that our model is intended as an early postoperative risk stratification tool to guide timely ICU management.
AKI is a well-recognized complication following TAAR and strongly associates with unfavorable prognoses [24,25]. Research indicates that aortic surgeries pose greater risks compared to other cardiac procedures such as valve replacement and coronary artery bypass grafting, especially given the high incidence of AKI [26–28]. In the current study, the incidence of AKI after TAAR, as defined by the KDIGO criteria, was 65.2%. This high incidence was associated with increased mortality, longer ICU stays, and higher medical costs. These findings are consistent with those of previous studies [29], highlighting that a high AKI incidence is a risk factor for adverse outcomes. Therefore, constructing models using biomarkers and relevant clinical variables to identify AKI as early as possible is highly important.
It is worth noting that intraoperative urine output, while reflecting renal perfusion during surgery, does not preclude the development of postoperative AKI. Several factors may explain this apparent discrepancy. First, intraoperative diuresis is often maintained by aggressive fluid resuscitation and the use of diuretics, which can mask underlying tubular injury. Second, ischemia-reperfusion injury may have already occurred during surgery, but its functional consequences (decreased glomerular filtration rate and oliguria) may not manifest until hours later. Third, many patients in our cohort developed non-oliguric AKI, diagnosed by serum creatinine elevation despite preserved urine output.
Serum lactate detection is a relatively simple and rapid method, making it widely applicable in clinical settings, particularly for routine monitoring of high-risk patients. Compared with several complex biomarker tests, lactate testing is cost effective, delivers results quickly, and can be conducted at the bedside, thus facilitating timely clinical decision-making [30]. Unlike single hemodynamic parameters such as blood pressure or fluid status, which provide only indirect snapshots of perfusion, lactate offers an integrative measure of the adequacy of tissue perfusion and oxygenation at the cellular level. Furthermore, while inotrope requirement reflects clinician-driven treatment decisions based on perceived hemodynamic instability, lactate provides a direct and objective assessment of the physiological consequences of that instability. The potential of lactate in predicting AKI has been preliminarily investigated in a small number of studies. Some research has proposed different lactate thresholds as predictive indicators for early positive recovery [10–13]. Moreover, elevated lactate levels may reflect the severity of potential injuries and predict adverse outcomes. In patients undergoing TAAR, several factors may contribute to elevated lactate at ICU admission. These include intraoperative factors such as prolonged cardiopulmonary bypass, deep hypothermic circulatory arrest, and ischemia-reperfusion injury [18]; postoperative hemodynamic instability including low cardiac output syndrome or hypovolemia [31]; and the systemic inflammatory response triggered by surgical trauma and cardiopulmonary bypass [9]. Considering the simplicity and efficiency of serum lactate testing, we evaluated the potential value of lactate at ICU admission as a promising biomarker for post-TAAR AKI. In our study, serum lactate levels exhibited a significant correlation with postoperative AKI incidence, persisting after multivariable adjustment. The AUC-ROC value of lactate for postoperative AKI diagnosis was 0.768, indicating moderate diagnostic accuracy and potential clinical utility.
In this study, the focus was not on the role of a single biomarker. Instead, we constructed a multivariate logistic regression prediction model for AKI occurrence, with lactate included as one of the variables. Our findings were consistent with prior literature, which indicated that postoperative AKI was associated with factors such as age, preexisting hypertension, intraoperative urine output, and APACHE II score at ICU admission [31–33]. Additionally, platelet transfusion usually implies a significant loss of blood, which is invariably accompanied by the transfusion of other blood products. Multiple studies have confirmed that blood product transfusion is associated with an increased risk of AKI [34,35]. Model 1 yielded an AUC-ROC of 0.757, reflecting a moderate level of diagnostic accuracy. This performance was not superior to that of previously established prediction models [36,37]. We found that when lactate level was combined with Model 1 to form Model 2, the prediction of AKI improved, and the AUC-ROC value significantly increased to 0.824, indicating good diagnostic accuracy. In addition, the results of the NRI (NRI = 0.141, p = 0.007) and IDI (IDI = 0.109, p < 0.001) further confirmed that Model 2 could significantly enhance risk reclassification compared with Model 1, enabling prompt AKI detection in perioperative care. Thus, the data from this study provide evidence that serum lactate is independently associated with TAAR-associated AKI and enhances early risk stratification. Although renal injury may already be evolving upon ICU arrival, full AKI manifestation often takes hours to days, providing a window for early intervention. Our model offers an immediate quantitative risk estimate. Patients with lactate >4.7 mmol/L have a nearly 10‑fold higher AKI risk and could be targeted for enhanced monitoring and goal‑directed therapy. This risk‑stratified approach goes beyond routine care, enabling timely interventions that may mitigate AKI progression.
Although the number of deaths in our cohort was low (n = 8), precluding any adjusted analysis, we observed a trend toward higher mortality in patients who developed AKI compared to those without AKI. This finding is consistent with the well-established association between AKI and increased mortality in cardiac surgery patients, and underscores the clinical relevance of early AKI prediction in this high-risk population.
Limitations
Several limitations affect this study. First, the monocentric retrospective design and constrained cohort size may limit generalizability. Second, the exclusion of 38 patients due to missing admission clinical data represents a potential source of selection bias. However, given the small proportion of excluded patients and the absence of significant differences in key characteristics between included and excluded patients, any potential selection bias is likely to be minimal. Third, although internal validation indicated reliability, multi-center external validation in in dependent sample populations is still needed to confirm its general applicability. Finally, we measured lactate only at ICU admission and did not perform dynamic monitoring of its changes over time. Therefore, further large-sample prospective multi-center trials are required to verify the study results.
Conclusions
In TAAR patients, higher serum lactate levels at ICU admission are associated with a heightened risk of developing AKI. Moreover, incorporating serum lactate levels into a clinical risk model based on variables available at ICU admission significantly enhances early postoperative risk stratification, providing a practical tool for guiding immediate ICU management in this high-risk population.
Funding Statement
This study was supported by the National Natural Science Foundation of China (82172162 to Chunbo Chen).
Ethics approval and consent to participate
Following the Declaration of Helsinki, this study was approved by the Ethics Committee of Guangdong Provincial People’s Hospital (Registration No.: KY2024-731-01). The requirement for written informed consent was deemed exempt by the Ethics Committee of Guangdong Provincial People’s Hospital because the study used deidentified data extracted from routine clinical records. This article does not contain any studies with human participants or animals performed by any of the authors.
Disclosure statement
The authors declare no competing interests.
Availability of data and materials
The data that support the findings of this study are available from the corresponding author, C.C., upon reasonable request.
References
- 1.Criado FJ. Aortic dissection: a 250-year perspective. Tex Heart I J. 2011;38(6):694–700. [PMC free article] [PubMed] [Google Scholar]
- 2.Fang M, Li J, Fang H, et al. Prediction of acute kidney injury after total aortic arch replacement with serum cystatin C and urine N-acetyl-beta-d-glucosaminidase: a prospective observational study. Clin Chim Acta. 2023;539:105–113. doi: 10.1016/j.cca.2022.12.002. [DOI] [PubMed] [Google Scholar]
- 3.Lau D, Pannu N, James MT, et al. Costs and consequences of acute kidney injury after cardiac surgery: a cohort study. J Thorac Cardiovasc Surg. 2021;162(3):880–887. doi: 10.1016/j.jtcvs.2020.01.101. [DOI] [PubMed] [Google Scholar]
- 4.Li Y, Jiang X, Chen J, et al. Evaluation of the contribution of gut microbiome dysbiosis to cardiac surgery-associated acute kidney injury by comparative metagenome analysis. Front Microbiol. 2023;14:1119959. doi: 10.3389/fmicb.2023.1119959. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Xie T, Xin Q, Zhang X, et al. Construction and validation of a nomogram for predicting survival in elderly patients with cardiac surgery. Front Public Health. 2022;10:972797. doi: 10.3389/fpubh.2022.972797. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Liu N, Li D, Liu D, et al. FOSL2 participates in renal fibrosis via SGK1-mediated epithelial-mesenchymal transition of proximal tubular epithelial cells. J Transl Int Med. 2023;11(3):294–308. doi: 10.2478/jtim-2023-0105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Jhaveri KD, Saratzis AN, Wanchoo R, et al. Endovascular aneurysm repair (EVAR)- and transcatheter aortic valve replacement (TAVR)-associated acute kidney injury. Kidney Int. 2017;91(6):1312–1323. doi: 10.1016/j.kint.2016.11.030. [DOI] [PubMed] [Google Scholar]
- 8.Gomez H, Kellum JA.. Lactate in sepsis. JAMA-J Am Med Assoc. 2015;313(2):194–195. doi: 10.1001/jama.2014.13811. [DOI] [PubMed] [Google Scholar]
- 9.Singer AJ, Taylor M, Domingo A, et al. Diagnostic characteristics of a clinical screening tool in combination with measuring bedside lactate level in emergency department patients with suspected sepsis. Acad Emerg Med. 2014;21(8):853–857. doi: 10.1111/acem.12444. [DOI] [PubMed] [Google Scholar]
- 10.Nichol AD, Egi M, Pettila V, et al. Relative hyperlactatemia and hospital mortality in critically ill patients: a retrospective multi-centre study. Crit Care. 2010;14(1):R25. doi: 10.1186/cc8888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jansen TC, van Bommel J, Schoonderbeek FJ, et al. Early lactate-guided therapy in intensive care unit patients: a multicenter, open-label, randomized controlled trial. Am J Respir Crit Care Med. 2010;182(6):752–761. doi: 10.1164/rccm.200912-1918OC. [DOI] [PubMed] [Google Scholar]
- 12.Puskarich MA, Trzeciak S, Shapiro NI, et al. Whole blood lactate kinetics in patients undergoing quantitative resuscitation for severe sepsis and septic shock. Chest. 2013;143(6):1548–1553. doi: 10.1378/chest.12-0878. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wacharasint P, Nakada T, Boyd JH, et al. Normal-range blood lactate concentration in septic shock is prognostic and predictive. Shock. 2012;38(1):4–10. doi: 10.1097/SHK.0b013e318254d41a. [DOI] [PubMed] [Google Scholar]
- 14.Allegretti AS, Steele DJR, David-Kasdan JA, et al. Continuous renal replacement therapy outcomes in acute kidney injury and end-stage renal disease: a cohort study. Crit Care. 2013;17(3):R109. doi: 10.1186/cc12780. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gisewhite S, Stewart IJ, Beilman G, et al. Urinary metabolites predict mortality or need for renal replacement therapy after combat injury. Crit Care. 2021;25(1):119. doi: 10.1186/s13054-021-03544-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Järvisalo MJ, Hellman T, Uusalo P.. Mortality and associated risk factors in patients with blood culture positive sepsis and acute kidney injury requiring continuous renal replacement therapy-a retrospective study. PLoS One. 2021;16(4):e0249561. doi: 10.1371/journal.pone.0249561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Passos RDH, Ramos JGR, Gobatto A, et al. Lactate clearance is associated with mortality in septic patients with acute kidney injury requiring continuous renal replacement therapy: a cohort study. Medicine (Baltimore). 2016;95(40):e5112. doi: 10.1097/MD.0000000000005112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lyu Y, Liu Y, Xiao X, et al. High level of intraoperative lactate might predict acute kidney injury in aortic arch surgery via minimally invasive approach in patients with type A dissection. Front Cardiovasc Med. 2023;10:1188393. doi: 10.3389/fcvm.2023.1188393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. doi: 10.7326/0003-4819-150-9-200905050-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Thomas ME, Blaine C, Dawnay A, et al. The definition of acute kidney injury and its use in practice. Kidney Int. 2015;87(1):62–73. doi: 10.1038/ki.2014.328. [DOI] [PubMed] [Google Scholar]
- 21.Kellum JA, Lameire N, KDIGO AKI Guideline Work Group . Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1). Crit Care. 2013;17(1):204. doi: 10.1186/cc11454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.DeLong ER, DeLong DM, Clarke-Pearson DL.. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837–845. [PubMed] [Google Scholar]
- 23.Vickers AJ, Elkin EB.. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making. 2006;26(6):565–574. doi: 10.1177/0272989X06295361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ko T, Higashitani M, Sato A, et al. Impact of acute kidney injury on early to long-term outcomes in patients who underwent surgery for type A acute aortic dissection. Am J Cardiol. 2015;116(3):463–468. doi: 10.1016/j.amjcard.2015.04.043. [DOI] [PubMed] [Google Scholar]
- 25.Hobson CE, Yavas S, Segal MS, et al. Acute kidney injury is associated with increased long-term mortality after cardiothoracic surgery. Circulation. 2009;119(18):2444–2453. doi: 10.1161/CIRCULATIONAHA.108.800011. [DOI] [PubMed] [Google Scholar]
- 26.Lei G, Wang G, Liu Q, et al. Single-stage hybrid aortic arch repair is associated with a lower incidence of postoperative acute kidney injury than conventional aortic surgery. J Cardiothorac Vasc Anesth. 2019;33(12):3294–3300. doi: 10.1053/j.jvca.2019.05.024. [DOI] [PubMed] [Google Scholar]
- 27.Zhou H, Wang G, Yang L, et al. Acute kidney injury after total arch replacement combined with frozen elephant trunk implantation: incidence, risk factors, and outcome. J Cardiothorac Vasc Anesth. 2018;32(5):2210–2217. doi: 10.1053/j.jvca.2018.02.026. [DOI] [PubMed] [Google Scholar]
- 28.Jiang X, Li Z, Pan C, et al. The role of serum magnesium in the prediction of acute kidney injury after total aortic arch replacement: a prospective observational study. J Med Biochem. 2024;43(4):574–586. doi: 10.5937/jomb0-48779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Zhao H, Pan X, Gong Z, et al. Risk factors for acute kidney injury in overweight patients with acute type A aortic dissection: a retrospective study. J Thorac Dis. 2015;7(8):1385–1390. doi: 10.3978/j.issn.2072-1439.2015.07.19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Gao B, Wang S, Jiang S.. The occurrence mechanism, assessment, and non-pharmacological treatment of dyspnea. Med Rev (2021). 2024;4(5):395–412. doi: 10.1515/mr-2024-0006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.He L, Liang S, Liang Y, et al. Defining a postoperative mean arterial pressure threshold in association with acute kidney injury after cardiac surgery: a prospective observational study. Intern Emerg Med. 2023;18(2):439–448. doi: 10.1007/s11739-022-03187-3. [DOI] [PubMed] [Google Scholar]
- 32.Neugarten J, Sandilya S, Singh B, et al. Sex and the risk of AKI following cardio-thoracic surgery: a meta-analysis. Clin J Am Soc Nephrol. 2016;11(12):2113–2122. doi: 10.2215/CJN.03340316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Honore PM, Kugener L, Redant S, et al. Recovery from acute kidney injury is a potent predictor of survival and good neurological outcome at discharge after out-of-hospital cardiac arrest: fact or fiction. J Transl Int Med. 2021;9(1):57–58. doi: 10.2478/jtim-2021-0018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Goren O, Matot I.. Perioperative acute kidney injury. Br J Anaesth. 2015;115 Suppl 2:ii3–i14. doi: 10.1093/bja/aev380. [DOI] [PubMed] [Google Scholar]
- 35.Nellis ME, Mandel A, Langer Y, et al. The association between platelet transfusion and acute kidney injury following Fontan surgery. World J Pediatr Congenit Heart Surg. 2025;16(5):668–673. doi: 10.1177/21501351241313318. [DOI] [PubMed] [Google Scholar]
- 36.Hou Y, Deng Y, Hu L, et al. Assessment of 17 clinically available renal biomarkers to predict acute kidney injury in critically ill patients. J Transl Int Med. 2021;9(4):273–284. doi: 10.2478/jtim-2021-0047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Hu L, Gao L, Zhang D, et al. The incidence, risk factors and outcomes of acute kidney injury in critically ill patients undergoing emergency surgery: a prospective observational study. BMC Nephrol. 2022;23(1):42. doi: 10.1186/s12882-022-02675-0. [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
The data that support the findings of this study are available from the corresponding author, C.C., upon reasonable request.