Novel nomogram for predicting acute kidney injury after cardiac surgery

Introduction

Background

Acute kidney injury (AKI) is a formidable and frequent complication following cardiac surgery that imposes a substantial burden on the outcomes of patients and the healthcare systems in which they are treated (1). While the incidence of postoperative AKI varies, it remains alarmingly high, particularly following complex surgical procedures; furthermore, its occurrence is strongly associated with an increased risk of morbidity, prolonged hospitalization periods, and elevated mortality rates (2). In the specific context of elective adult cardiac surgery, which encompasses both valve replacement and coronary artery bypass grafting (CABG), the development of AKI represents a critical juncture that can derail recovery and compromise long-term renal function (3).

The pathophysiology of AKI is multifactorial in nature and involves a complex interplay of patient-specific susceptibilities, systemic inflammatory responses triggered by surgical trauma and cardiopulmonary bypass, hemodynamic instability, and exposure to nephrotoxic insults (4). Considering the elective nature of these procedures in this population, there is a pivotal window for preoperative risk stratification, and the ability to accurately predict which patients exhibit the highest risk for developing AKI could enable the implementation of targeted perioperative management strategies to potentially mitigate this serious complication (5). Consequently, there has been growing clinical and research interest in the development and validation of robust clinical prediction models for postoperative AKI in patients undergoing elective cardiac surgery.

Rationale and knowledge gap

Given the complex array of risk factors associated with AKI, there is a clear need for tools that can synthesize relevant information to generate practical and clinically useful estimates of individual patient risk. Several risk scores have been developed to predict the occurrence of AKI after cardiac surgery, with the Cleveland Clinic Score being one of the most widely recognized tools for preoperative assessments (6); however, the predictive utility of these scores can vary across surgical populations; while it was particularly useful for predicting the need for renal replacement therapy, its performance differed when applied to specific procedure types, such as isolated CABG vs. aortic valve surgery, suggesting a need for model refinement or procedure-specific adjustments (7). These findings highlight the fact that many existing scores may not fully capture or apply proper weighting to the diverse set of relevant predictive factors, and such scores may lack validation in heterogeneous surgical cohorts.

Objective

Therefore, this study aimed to optimize perioperative management strategies for adult patients undergoing elective cardiac surgery. A comprehensive panel of predictive factors encompassing demographic characteristics, comorbidities, preoperative laboratory indices, and detailed intraoperative variables was systematically investigated, and logistic regression analyses were conducted to develop a clinically actionable nomogram for the refined prediction of postoperative AKI. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0698/rc).

Methods

Study design

This retrospective study was approved by the Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine (No. 069/2017, 11/22/2017). The requirement for informed consent was waived owing to the fact that this study was based solely on the analysis of anonymized medical record data obtained from prior clinical consultations and did not pose any risk to the included patients. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Population

Patients aged ≥18 years who underwent cardiopulmonary bypass surgery (including single valve replacement, multivalve replacement, and valve replacement combined with CABG) at The Second Affiliated Hospital of Zhejiang University School of Medicine between September 1, 2013 and December 31, 2024 were eligible for inclusion. We validated externally with data from West China Hospital, spanning June 2012 to June 2017. Those who underwent emergency surgery, total aortic arch replacement surgery, or cardiac tumor resection were excluded, as were individuals requiring massive perioperative blood transfusions (>10 U within 24 h) and those lacking more than 10% of the information required to determine the outcome, as these factors were likely to have affected the statistical analyses.

Management of perioperative anesthesia

Total intravenous anesthesia was induced using a combination of midazolam, sufentanil, etomidate, and cisatracurium. Anesthesia was maintained through continuous infusion of propofol and remifentanil via an infusion pump, whereas sufentanil and cisatracurium were administered intermittently. The decision to use sevoflurane as an inhalational anesthetic was left to the discretion of the attending anesthesiologist, depending on the intraoperative conditions and surgical progression in each case.

Data collection and candidate variable selection

Data were obtained from the electronic medical record system and DoCare anesthesia system of The Second Affiliated Hospital of Zhejiang University School of Medicine and West China Hospital, predominantly including information pertaining to demographic and perioperative clinical variables. To comprehensively incorporate factors affecting the requirement for cardiac surgery-associated AKI (CSA-AKI) into the model, marquee variables were identified based on a combination of a literature search of the EMBASE, PubMed, and Cochrane databases, as well as the clinical expertise of the authors.

Selection of candidate variables

Based on previous literature and the recommendations of the clinical experts, 30 preoperative candidate variables were initially selected for evaluation. To adequately select the factors associated with the primary outcome, all potentially influential preoperative and intraoperative variables identified based on the literature search and clinical expertise of the authors were incorporated into the least absolute shrinkage and selection operator (LASSO) regression analysis. Categorical variables were transformed into factors, and optimal λ values were selected using ten-fold cross-validation. Researchers received training prior to data collection to clarify the scope of the study. Upon completion of data collection, the data were verified and archived by two researchers who had no prior involvement in the study.

Primary outcome

The primary outcome was the occurrence of AKI, diagnosed based on serum creatinine (SCr) levels according to the Kidney Disease: Improving Global Outcomes criteria (stage 1, SCr increase of ≥0.3 mg/dL within 48 h or 1.5–1.9 times the baseline within 7 days; stage 2, increase of 2.0–2.9 times baseline; and stage 3, SCr increase of ≥3.0 times baseline, SCr ≥4.0 mg/dL) or the initiation of renal replacement therapy (2). Urine output was not used as a criterion because it can be strongly influenced by perioperative diuretic administration and other clinical interventions, potentially confounding its association with postoperative outcomes (8).

Statistical analysis

The dataset was randomly divided into a training cohort and a validation cohort at a ratio of 7:3, and the variables were compared. Continuous variables with a normal distribution are presented as means ± standard deviations, whereas non-normally distributed data are presented as medians (interquartile ranges). In the univariate analysis, either the Chi-squared test or Fisher’s exact test was used to analyze categorical variables, whereas the Student’s t-test or rank-sum test was used to examine continuous variables. In the training cohort, LASSO logistic regression was used for the multivariate analysis to screen for independent risk factors and construct a nomogram for predicting the occurrence of AKI. The performance of the model was assessed using receiver operating characteristic (ROC) curve and calibration curve analyses, with the area under the receiver operating characteristic curve (AUC) ranging from 0.5 (non-discriminant) to 1 (complete discriminant). A calibration plot was used to assess the calibration of the prediction model, which reflects the agreement between predicted probabilities and observed event rates. A decision curve analysis (DCA) was performed to determine the net benefit threshold of the prediction. Statistical significance was set at P<0.05. All statistical analyses were performed using R (version 4.2.2).

Results

Patient characteristics

Baseline demographic and clinical characteristics of the study population (divided into a training cohort (n=4,805), an internal validation cohort (n=2,059), and an external validation set (n=6,819), are presented in Table 1 and Figure S1. The cohorts were generally well-balanced across most variables; however, a slight but statistically significant difference was observed in the sex distribution between the cohorts (P=0.03), with males constituting 53.9% and 56.8% of the training and test cohorts, respectively. The mean age of the patients was 59±12 years in both cohorts, with no statistically significant difference (P=0.51). Table 1 summarizes intraoperative and postoperative vasopressor doses (3.16±7.07 vs. 3.46±16.19), fluid balance (2,442±1,044 vs. 2,436±1,063), the lowest intraoperative hemoglobin (83±19 vs. 84±20), and the rate of postoperative vasoactive medication use [126 cases (3.4%) vs. 46 cases (2.9%), P=0.40].

No significant differences were observed in terms of comorbidities or indicators of preoperative health status. The American Society of Anesthesiologists (ASA) physical status classification was similar between cohorts (P=0.31), with the vast majority of patients (>82%) classified as ASA III. The prevalence of hypertension was approximately 65% and 64% in the training and test cohorts, respectively (P=0.24). Approximately 8% of the patients in both groups had undergone prior surgery (P=0.51). The prevalence of chronic obstructive pulmonary disease (COPD) was very low (0.5–0.6%, P=0.69), as was the prevalence of stroke (2.8% in both cohorts, P>0.99), diabetes with complications (3.8% in both cohorts, P=0.99), and kidney disease (2.3% vs. 2.8%, P=0.22).

Predictor selection

The candidate predictors incorporated into the original model included sex, age, body mass index (BMI), smoking history, ASA classification, drinking history, hypertension, surgical history, COPD, stroke, diabetes with complications, creatinine, preoperative hemoglobin level, platelet count, white blood cell count, international normalized ratio, activated partial thromboplastin time (APTT), total intraoperative red blood cell transfusion volumes, surgical characteristics, intraoperative/postoperative vasopressor dosage, total fluid volume during surgery, nadir hemoglobin level during surgery, postoperative vasopressor rate, preoperative kidney disease severity, preoperative ejection fraction and bleeding; these were subsequently reduced to five potential predictors via the LASSO regression analysis performed in the training cohort. A cross-validated error plot of the LASSO regression model is also shown in Figure 1. The most regularized and parsimonious model, with a cross-validated error within one standard error of the minimum, included five variables.

Figure 1 Screening of variables for incorporation into the predictive model. (A) The binomial deviance plot shows the optimal λ value selected by cross-validation. (B) The LASSO coefficient profiles of the candidate variables. The dashed vertical line indicates the optimal λ value determined by cross-validation. LASSO, least absolute shrinkage and selection operator.

As shown in Figure S2, the ROC curve analysis of the aforementioned variables yielded AUC values almost greater than 0.6. The ROC curve analysis for predicting AKI demonstrated that preoperative hemoglobin exhibited the highest discriminatory capacity, with an AUC of 0.770 [95% confidence interval (CI): 0.750–0.791], followed by bleeding, with an AUC of 0.747 (95% CI: 0.726–0.768). In contrast, age, creatinine, and total intraoperative RBCs exhibited relatively lower discriminatory capacities, with AUC values of 0.592 (95% CI: 0.568–0.616), 0.664 (95% CI: 0.638–0.689), and 0.686 (95% CI: 0.665–0.707), respectively.

Development of the predictive model for AKI

The multivariable logistic regression analysis in the training cohort identified several independent predictors of AKI. For example, older age was significantly associated with an increased risk of AKI [odds ratio (OR) =1.03; 95% CI: 1.02–1.04; P<0.001]. Compared to that of patients without anemia, those with such complications exhibited a significantly higher likelihood of developing AKI (OR =0.98; 95% CI: 0.98–0.98; P<0.001). Creatinine was also a predictor, conferring a markedly elevated risk of AKI development (OR =1.03; 95% CI: 1.02–1.03; P<0.001). Greater total intraoperative red blood cell transfusion volumes (OR =1.07; 95% CI: 1.02–1.12; P<0.001) and increased perioperative bleeding (OR =1.00; 95% CI: 1.00–1.00; P<0.001) were also independently associated with AKI (Table 2).

As shown in Figure 2, the final logistic model included five independent predictors (age, preoperative hemoglobin, baseline creatinine level, total intraoperative red blood cell transfusion volume, and bleeding) and was developed as a simple-to-use nomogram.

Figure 2 Nomogram of the predictive model for evaluating the risk of AKI in patients undergoing cardiac surgery. The nomogram incorporates five predictors: age, preoperative hemoglobin, creatinine, total intraoperative RBC transfusion, and bleeding volume. For each variable, the corresponding points are determined by drawing a vertical line to the “Points” scale. The total points are summed and located on the “Total points” scale, which corresponds to the predicted probability of AKI on the “Risk of outcome” scale. AKI, acute kidney injury; RBC, red blood cell.

The AUCs of the model in the different cohorts are shown in Figure 3. The ROC curve analysis of the prediction model demonstrated robust discriminative performance across both cohorts. More specifically, in the training cohort, the AUC was 0.880 (95% CI: 0.867–0.894), and the model maintained its performance in the independent validation cohort, achieving an AUC of 0.883 (95% CI: 0.863–0.904). The AUC for the external validation set is 0.690, with a 95% CI of 0.671–0.709.

Figure 3 Assessment of the outcome risk using the predictive model. The AUC values with 95% CIs are presented for each cohort. The diagonal gray line represents the reference line of no discrimination (AUC =0.5). AUC, area under the ROC curve; CI, confidence interval; ROC, receiver operating characteristic.

Model evaluation

Calibration plots of the nomogram in the different cohorts are shown in Figure 4A-4C, which demonstrated a good correlation between the observed and predicted occurrence of AKI. The original nomogram retained its validity for use in the validation sets, and the calibration curve of this model was relatively close to the ideal curve, indicating that the predicted results were consistent with the actual findings.

Figure 4 Assessment of predictive model performance. (A-C) Calibration curves illustrating the agreement between predicted probabilities and observed outcomes in the testing set (A), internal validation set (B), and external validation set (C). (D-F) DCA for the corresponding models in the testing set (D), internal validation set (E), and external validation set (F). The region where the model’s curve lies above the “All” and “None” lines indicates the range of threshold probabilities at which the model provides a positive net clinical benefit. DCA, decision curve analysis.

The DCA curves related to the nomogram are shown in Figure 4D-4F. A high-risk threshold probability indicates the chance of observing significant discrepancies in the model’s predictions when clinicians encounter major flaws when using a nomogram for diagnostic and decision-making purposes. The DCA findings demonstrated that the nomogram developed in the present study offers substantial net benefits when used in clinical applications.

Sample size calculation

To determine whether the study was sufficiently powered, a post hoc calculation of the required sample size was conducted. Considering 20–30% of cases had to have developed AKI in the model cohort, 30 potential variables, and Nagelkerke’s R²=0.29, at least 1,056 participants were required to ensure that the statistical analyses were sufficiently robust; this corresponded to at least 44 events per predictor variable. Given that the model development cohort comprised 6,864 patients, the sample size was confirmed to be sufficient.

Discussion

Key findings

The development and validation of a nomogram to predict the occurrence of postoperative AKI in adults undergoing elective cardiac valve and bypass surgeries yielded several critical findings. The multivariable logistic regression analysis identified age, preoperative hemoglobin, preoperative SCr levels, intraoperative blood transfusion volume, and perioperative bleeding as independent predictors of AKI; among these factors, intraoperative blood transfusion volume emerged as the strongest predictor of postoperative AKI, underscoring the profound vulnerability of this patient population. The predictive model showed strong and consistent accuracy across the training, internal validation, and external validation sets. Thus, a complex array of preoperative and intraoperative variables was effectively translated into a nomogram that facilitates individualized risk assessment.

Strengths and limitations

The visual and quantitative attributes of the risk-benefit chart effectively convey potential risks to both patients and the multidisciplinary medical team involved in their care, thereby enhancing the processes of informed consent and shared decision-making (9). In the preoperative phase, when risk factors such as preoperative anemia or elevated baseline creatinine levels are identified, immediate proactive measures can be implemented, including aggressive correction of anemia, postponement of surgery, and consideration of pharmacological interventions aimed at renal protection (10). Patients classified as high-risk through the risk-scorecard necessitate intensified monitoring of fluid balance and the implementation of renal protection strategies to avert secondary injury, which may help mitigate the progression of AKI and its severe repercussions, including increased mortality and the potential development of chronic kidney disease (11).

Despite the strengths of this study, several limitations must be acknowledged. The model also relies on clinically available variables, which is a pragmatic strength; however, it does not incorporate novel biomarkers that may facilitate earlier or more specific detection of renal stress. For example, some studies have explored the utility of cystatin C, kidney injury molecule 1, and oxidative stress markers such as nuclear factor erythroid 2-related factor 2 in predicting AKI occurrence (12); the integration of such biomarkers, perhaps into a dynamic prediction model, could further improve predictive accuracy and clinical timelines (13). Furthermore, while a comprehensive set of predictors was included in the model development, unmeasured confounders such as the use of specific anesthetic agents, detailed fluid management strategies, and genetic predispositions could influence the risk of AKI development (14). Finally, the nomogram predicts the occurrence of AKI; however, it does not stratify its severity or the need for renal replacement therapy, which are critical outcomes that influence clinical management strategies and resource planning. Future prospective studies with larger sample sizes are needed to conduct Kidney Disease: Improving Global Outcomes stage-based stratification and explore the predictive value for continuous renal replacement therapy (CRRT) requirement, so as to further optimize and upgrade the model.

Comparison with similar research

The present findings are consistent with and build upon the existing literature on AKI prediction in patients undergoing cardiac surgery. The association of older age with increased AKI risk is a well-recognized phenomenon that has been attributed to age-related decline in the glomerular filtration rate, increased vascular stiffness, and reduced regenerative capacity of tubular epithelial cells (15). Furthermore, bleeding and blood transfusions contribute to the likelihood of AKI occurrence through several mechanisms; for example, hemorrhagic shock leads to renal hypoperfusion and ischemic tubular injury, whereas transfusions introduce stored red blood cells that may be less deformable and release free hemoglobin and iron, thereby promoting oxidative stress and direct tubular toxicity, a process akin to the subclinical hemolysis observed after certain medical procedures (16). The present results consolidate these observations by directly quantifying the incremental risk per unit of transfused red blood cells. The model’s performance, as evidenced by an AUC exceeding 0.80, suggests that it may be superior to some existing tools such as the Cleveland Clinic Score, which demonstrated a more moderate predictive ability (AUC =0.630) in a recent study evaluating its performance across elective cardiac surgeries (17).

Explanations of findings

The pathophysiological pathways linking the predictors of AKI identified in the present study are multifaceted and interconnected. Preoperative SCr concentrations represent a state of reduced functional nephron mass and often concomitant microvascular rarefaction, leaving the kidneys extremely susceptible to further injury (18). We categorized kidney disease into chronic kidney disease and dialysis-dependent severe renal disease, using preoperative SCr as a continuous measure of severity. This stratification significantly influenced the nomogram model. Dialysis-dependent patients scored highest, followed by those with chronic kidney disease. Higher creatinine levels correlated with increased risk scores, indicating a higher likelihood of postoperative AKI (19). Including renal disease severity and creatinine levels improved the model’s accuracy and reliability. Baseline renal function was a key factor for personalized risk assessment, enhancing the model’s clinical relevance and usability. This cascade, coupled with potential episodes of hypotension experienced during cardiopulmonary bypass procedures, creates a perfect storm that promotes AKI development, particularly in patients with an already compromised renal reserve (20,21).

Implications and actions needed

Future research should build on the foundation established in this study in several ways. Validation in prospective, multicenter studies is the immediate next step toward establishing the nomogram’s reliability across varied clinical settings (22). From an interventional perspective, randomized controlled trials are required to evaluate whether targeted care bundles triggered by high nomogram scores can effectively prevent AKI or mitigate its severity; such bundles could include specific hemodynamic optimization protocols, personalized transfusion algorithms, or the prophylactic administration of renoprotective agents in rigorously defined high-risk groups (23).

Conclusions

In summary, understanding the long-term renal trajectory of high-risk patients identified using this nomogram is crucial. Researchers should examine the link between a high predicted risk, the occurrence of postoperative AKI, and the subsequent development of chronic kidney disease, as the transition from AKI to chronic kidney disease is a major clinical concern with significant public health implications. Exploring these future directions could help optimize the predictive utility of this clinical tool, ultimately contributing to improved perioperative outcomes in patients undergoing complex cardiac surgery.

Acknowledgments

We thank all stuff from The Second Affiliated Hospital of Zhejiang University School of Medicine for help in data collection and their support. Finally, we wish to thank the English editors of Editage (https://www.editage.cn/).

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0698/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0698/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0698/prf

Funding: This study was supported by the Leading Health Talents of Zhejiang Province, Zhejiang Health Office [No. 18(2020)], the Key Research and Development Program of Zhejiang Province, and the National Clinical Key Specialty Construction Project of China 2021 (No. 2021-LCZDZK-01).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0698/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine (No. 069/2017, 11/22/2017). The requirement for informed consent was waived by the Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine due to the retrospective nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Cerda J, Kashani K, Ostermann M, et al. The global epidemiology of acute kidney injury: challenges and opportunities. Nat Rev Nephrol 2026;22:179-98. [Crossref] [PubMed]
2. Cheruku SR, Raphael J, Neyra JA, et al. Acute Kidney Injury after Cardiac Surgery: Prediction, Prevention, and Management. Anesthesiology 2023;139:880-98. [Crossref] [PubMed]
3. Demirjian S, Bashour CA, Shaw A, et al. Predictive Accuracy of a Perioperative Laboratory Test-Based Prediction Model for Moderate to Severe Acute Kidney Injury After Cardiac Surgery. JAMA 2022;327:956-64. [Crossref] [PubMed]
4. Nguyen M, Alvarez M, Evezard C, et al. Impaired endotoxin inactivation, rather than gut translocation, is associated with organ injury in cardiac surgery with cardiopulmonary bypass: An ancilliary analysis of a randomised control trial. Eur J Anaesthesiol 2026; Epub ahead of print. [Crossref]
5. Kim JH, Cheon BR, Kim MG, et al. Harnessing Machine Learning for Prediction of Postoperative Pulmonary Complications: Retrospective Cohort Design. J Clin Med 2023;12:5681. [Crossref] [PubMed]
6. Tseng PY, Chen YT, Wang CH, et al. Prediction of the development of acute kidney injury following cardiac surgery by machine learning. Crit Care 2020;24:478. [Crossref] [PubMed]
7. Johnston A, Mesana TG, Lee DS, et al. Sex Differences in Long-Term Survival After Major Cardiac Surgery: A Population-Based Cohort Study. J Am Heart Assoc 2019;8:e013260. [Crossref] [PubMed]
8. Vaara ST, Ostermann M, Selander T, et al. Protocol and statistical analysis plan for the REstricted fluid therapy VERsus Standard trEatment in Acute Kidney Injury-REVERSE-AKI randomized controlled pilot trial. Acta Anaesthesiol Scand 2020;64:831-8. [Crossref] [PubMed]
9. Riera R, de Oliveira Cruz Latorraca C, Padovez RCM, et al. Strategies for communicating scientific evidence on healthcare to managers and the population: a scoping review. Health Res Policy Syst 2023;21:71. [Crossref] [PubMed]
10. Lai YC, Yu TS, Hung KC, et al. Preoperative hemoglobin level and risk of new-onset atrial fibrillation and flutter after total joint arthroplasty: a retrospective analysis. Sci Rep 2025;16:3343. [Crossref] [PubMed]
11. Adejumo OA, Edeki IR, Ojo OE, et al. Incidence of acute kidney injury and associated mortality in Nigeria: systematic review and meta-analysis. BMC Nephrol 2026;27:160. [Crossref] [PubMed]
12. Wang Y, Zheng X, Yang Z, et al. Cystatin C for predicting acute kidney injury in critically ill children with bacterial infections: a retrospective cohort study. BMJ Paediatr Open 2026;10:e004152. [Crossref] [PubMed]
13. Wang L, Zhou D, Xu P. A Deep Learning Model for Dynamic Prediction of Acute Kidney Injury in Heart Failure Patients. Can J Cardiol 2026;42:1605-15. [Crossref] [PubMed]
14. Mazhar F, Fu EL, Faucon AL, et al. Systemic Inflammation and the Risks of Adverse Kidney Outcomes in Adults With Atherosclerotic Cardiovascular Disease. Am J Kidney Dis 2025;86:314-323.e1. [Crossref] [PubMed]
15. Zhang T, McArdle Z, Moore BK, et al. Sex-dependent differences in the progression of renal injury and fibrosis following ischemic acute kidney injury. Clin Sci (Lond) 2026;140:275-90. [Crossref] [PubMed]
16. Kuraoka S, Nozoe M, Mannoji H, et al. Differential subclinical hemolysis after pulsed field ablation using the FARAPULSE pentaspline catheter vs the PulseSelect circular multi-electrode array catheter. Heart Rhythm O2 2025;6:1911-8. [Crossref] [PubMed]
17. Kozioł M, Kyslyi V, Sobczyk D, et al. Evaluation of the Cleveland Clinic Score for predicting acute kidney injury across different elective cardiac surgeries-a retrospective study. PeerJ 2026;14:e20533. [Crossref] [PubMed]
18. Jiang B, Zhen G, Yang H, et al. Normalization of elevated preoperative serum creatinine and acute kidney injury after cardiac surgery: a retrospective cohort study. Sci Rep 2025;15:27933. [Crossref] [PubMed]
19. Zhao P, Feng D, Wang H, et al. Safety, tolerability, pharmacokinetic, and pharmacodynamic of CZ1S injection for unilateral brachial plexus blockade in healthy Chinese adults: Double-blind, randomized, positive-controlled Phase I study. J Pharm Sci 2025;114:1245-53. [Crossref] [PubMed]
20. Deng YH, Liu Q, Luo XQ. The gut-kidney axis in pediatric acute kidney injury: a review of pathophysiological mechanisms and therapeutic frontiers. Pediatr Nephrol 2026; Epub ahead of print. [Crossref]
21. Jin Y, Zhang SJ, Zhuang S, et al. Microbiota-gut-kidney axis in health and renal disease. Int J Biol Sci 2026;22:750-70. [Crossref] [PubMed]
22. Gould DW, Doidge J, Sadique MZ, et al. Heparin versus citrate anticoagulation for continuous renal replacement therapy in intensive care: the RRAM observational study. Health Technol Assess 2022;26:1-58. [Crossref] [PubMed]
23. Peng K, McIlroy DR, Bollen BA, et al. Society of Cardiovascular Anesthesiologists Clinical Practice Update for Management of Acute Kidney Injury Associated With Cardiac Surgery. Anesth Analg 2022;135:744-56. [Crossref] [PubMed]