OBJECTIVES: Abdominal aortic aneurysm is a pathological condition in which the abdominal aorta is dilated beyond 3.0 cm. The surgical options include open surgical repair (OSR) and endovascular aneurysm repair (EVAR). Prediction of acute kidney injury (AKI) after OSR is helpful for decision-making during the postoperative phase. To find a more efficient method for making a prediction, this study aims to perform tests on the efficacy of different machine learning models. METHODS: Perioperative data of 80 OSR patients were retrospectively collected from January 2009 to December 2021 at Xiangya Hospital, Central South University. The vascular surgeon performed the surgical operation. Four commonly used machine learning classification models (logistic regression, linear kernel support vector machine, Gaussian kernel support vector machine, and random forest) were chosen to predict AKI. The efficacy of the models was validated by five-fold cross-validation. RESULTS: AKI was identified in 33 patients. Five-fold cross-validation showed that among the 4 classification models, random forest was the most precise model for predicting AKI, with an area under the curve of 0.90±0.12. CONCLUSIONS Machine learning models can precisely predict AKI during early stages after surgery, which allows vascular surgeons to address complications earlier and may help improve the clinical outcomes of OSR.
Humans ; Aortic Aneurysm, Abdominal/complications* ; Endovascular Procedures/methods* ; Retrospective Studies ; Blood Vessel Prosthesis Implantation/adverse effects* ; Acute Kidney Injury/etiology* ; Machine Learning ; Treatment Outcome ; Postoperative Complications/etiology* ; Risk Factors

OBJECTIVE: To establish a prediction model of acute kidney injury (AKI) in moderate and severe burn patients, so as to provide basic research evidence for early identification of burn-related AKI. METHODS: Patients who were admitted to the department of plastic burn surgery of the Affiliated Hospital of Southwest Medical University from November 2018 to January 2021 were selected, and their clinical characteristics, laboratory examinations and other indicators were recorded. Multivariate Logistic regression analysis was used to screen out the risk factors of AKI related to moderate and severe burns, and R software was used to establish the nomogram of moderate and severe burn patients complicated with AKI. The Bootstrap method model was used for internal verification by repeating sample for 1 000 times. Consistency index and calibration curve were used to evaluate the accuracy of the model, and the receiver operator characteristic curve (ROC curve) and the area under the curve (AUC) were used to evaluate the prediction efficiency, decision curve analysis (DCA) was used to evaluate the clinical utility of the model. RESULTS: A total of 186 patients with moderate and severe burn were included, among which 54 patients suffered from AKI, and the incidence rate was 29.03%. Multivariate Logistic regression analysis showed that the total burn surface area [TBSA; odds ratio (OR) = 1.072, 95% confidence interval (95%CI) was 1.031-1.115, P = 0.001], estimated glomerular filtration rate (eGFR; OR = 0.960, 95%CI was 0.931-0.990, P = 0.010), neutrophil (NEU; OR = 1.190, 95%CI was 1.021-1.386, P = 0.026), neutrophil/lymphocyte ratio (NLR; OR = 0.867, 95%CI was 0.770-0.977, P = 0.019), D-dimer (OR = 4.603, 95%CI was 1.792-11.822, P = 0.002) were the risk factors for patients with moderate and severe burn complicated with AKI. Taking the above indexes as predictive factors, a nomogram prediction model was established, the ROC curve was plotted with AUC of 0.998 (95%CI was 0.988-1.000). Optimum threshold of ROC curve was -0.862, the sensitivity was 98.0% and the specificity was 98.2%, and the consistency index was 0.998 (95%CI was 0.988-1.000). The calibration curve showed that the prognostic nomogram model was accurate, DCA showed that most patients can benefit from this model. CONCLUSIONS The burned patients with higher TBSA, NEU, NLR, D-dimer and lower eGFR tend to suffer from AKI. The nomogram based on the above five risk factors has high accuracy and clinical value, which can be used as a predictive tool to evaluate the risk of AKI in moderate and severe burn patients.
Humans ; Prognosis ; Nomograms ; Retrospective Studies ; Burns/complications* ; Acute Kidney Injury/etiology* ; ROC Curve

Humans ; Retrospective Studies ; Cardiopulmonary Bypass/adverse effects* ; Coronary Artery Bypass/adverse effects* ; Coronary Artery Disease/surgery* ; Acute Kidney Injury/etiology* ; Postoperative Complications ; Risk Factors ; Treatment Outcome

OBJECTIVE: To investigate the predictive value of pulse infusion index (PPI) in the short-term prognosis of patients with sepsis-induced acute kidney injury (AKI). METHODS: A retrospective cohort study was conducted. The clinical data of patients with sepsis-induced AKI admitted to intensive care unit (ICU) of the First Affiliated Hospital of Soochow University from July 2021 to December 2022 were enrolled. The basic information of the patients were collect, including age, gender, site of infection, underlying disease, mean arterial pressure (MAP) and heart rate (HR) at admission, as well as the use of mechanical ventilation and vasoactive drugs, and norepinephrine (NE) dosage. Laboratory indicators, sequential organ failure assessment (SOFA) score and PPI within 24 hours of admission were also recorded, and the patient's prognosis during ICU hospitalization was also recorded. The differences in clinical data between the patients of two groups with different prognosis were compared. Spearman correlation method was used to analyze the correlation between PPI and SOFA score. Binary multivariate Logistic regression analysis was used to screen independent risk factors for death during ICU hospitalization in sepsis patients with AKI. Receiver operator characteristic curve (ROC curve) was plotted to evaluate the predictive value of PPI for the short-term prognosis of patients with sepsis-induced AKI. RESULTS: A total of 102 patients with sepsis-induced AKI were enrolled, of which 70 patients in the survival group and 32 patients in the death group, with ICU mortality of 31.4. Compared with the survival group, SOFA score, HR, procalcitonin (PCT), serum creatinine (SCr), and NE dosage in the death group were significantly increased [SOFA score: 11.22±2.48 vs. 8.56±2.01, HR (bpm): 103.80±12.47 vs. 97.41±9.73, PCT (μg/L): 9.22 (5.24, 17.84) vs. 6.19 (3.86, 7.71), SCr (μmol/L): 163.2 (104.7, 307.9) vs. 125.5 (89.3, 221.0), Lac (mmol/L): 2.81 (1.95, 4.22) vs. 2.13 (1.74, 2.89), NE usage (μg×kg^-1×min^-1): 0.7 (0.4, 1.1) vs. 0.5 (0.2, 0.6), all P < 0.05], while PPI was significantly lower than that in survival group [0.83 (0.42, 1.55) vs. 1.70 (1.14, 2.20), P < 0.01]. Spearman correlation analysis showed that based on SOFA score, PPI was closely related to the severity of patients with sepsis-induced AKI (r = -0.328, P < 0.05). Binary multivariate Logistic regression analysis showed that PPI [odds ratio (OR) = 0.590, 95% confidence interval (95%CI) was 0.361-0.966, P = 0.002], SOFA score (OR = 1.406, 95%CI was 1.280-1.545, P < 0.001), PCT (OR = 2.061, 95%CI was 1.267-3.350, P = 0.006) were independent risk factors of the short-term prognosis of patients with sepsis-induced AKI. ROC curve analysis showed that the area under the ROC curve (AUC) of PPI for death during ICU hospitalization in patients with sepsis-induced AKI was 0.779 (95%CI was 0.686-0.855, P < 0.001), which superior to PCT (AUC = 0.677, 95%CI was 0.577-0.766, P = 0.004), and similar to SOFA score (AUC = 0.794, 95%CI was 0.703-0.868, P < 0.001). When the cut-off value of PPI was 0.72, the sensitivity was 50.0%, and the specificity was 97.1%. CONCLUSIONS PPI has a good predictive value for the short-term prognosis of patients with sepsis-induced AKI during ICU hospitalization.
Humans ; Heart Rate ; Retrospective Studies ; ROC Curve ; Sepsis/complications* ; Prognosis ; Procalcitonin ; Acute Kidney Injury/etiology* ; Intensive Care Units

Objective: To investigate the changes of high density lipoprotein cholesterol (HDL-C) in sepsis patients and its early predictive value for secondary acute kidney injury (AKI) in such patients. Methods: A retrospective case series study was conducted. From June 2019 to June 2021, 232 sepsis patients who met the inclusion criteria were admitted to the Second Hospital of Hebei Medical University, including 126 males and 106 females, aged 24 to 71 years. According to whether complicating secondary AKI, the patients were divided into non-AKI group (n=158) and AKI group (n=74). Data of patients between the two groups were compared and statistically analyzed with independent sample t test or chi-square test, including the sex, age, body mass index (BMI), body temperature, heart rate, primary infection site, combined underlying diseases, acute physiology and chronic health evaluation Ⅱ (APACHE Ⅱ) score and sepsis-related organ failure assessment (SOFA) score at admission, and the serum levels of C-reactive protein (CRP), procalcitonin, creatinine, cystatin C, and HDL-C measured at diagnosis of sepsis. The multivariate logistic regression analysis was performed on the indicators with statistically significant differences between the two groups to screen the independent risk factors for developing secondary AKI in 232 sepsis patients, and the joint prediction model was established based on the independent risk factors. The receiver operating characteristic (ROC) curve of the independent risk factors and the joint prediction model predicting secondary AKI in 232 sepsis patients were drawn, and the area under the curve (AUC), the optimal threshold, and the sensitivity and specificity under the optimal threshold were calculated. The quality of the above-mentioned AUC was compared by Delong test, and the sensitivity and specificity under the optimal threshold were compared using chi-square test. Results: The sex, age, BMI, body temperature, heart rate, primary infection site, combined underlying diseases, and CRP level of patients between the two groups were similar (P>0.05). The procalcitonin, creatinine, cystatin C, and scores of APACHE Ⅱ and SOFA of patients in AKI group were all significantly higher than those in non-AKI group (with t values of -3.21, -16.14, -12.75, -11.13, and -12.88 respectively, P<0.01), while the HDL-C level of patients in AKI group was significantly lower than that in non-AKI group (t=6.33, P<0.01). Multivariate logistic regression analysis showed that creatinine, cystatin C, and HDL-C were the independent risk factors for secondary AKI in 232 sepsis patients (with odds ratios of 2.45, 1.68, and 2.12, respectively, 95% confidence intervals of 1.38-15.35, 1.06-3.86, and 0.86-2.56, respectively, P<0.01). The AUCs of ROC curves of creatinine, cystatin C, HDL-C, and the joint prediction model for predicting secondary AKI in 232 sepsis patients were 0.69, 0.79, 0.89, and 0.93, respectively (with 95% confidence intervals of 0.61-0.76, 0.72-0.85, 0.84-0.92, and 0.89-0.96, respectively, P values all below 0.01); the optimal threshold were 389.53 μmol/L, 1.56 mg/L, 0.63 mmol/L, and 0.48, respectively; the sensitivity under the optimal threshold were 76.6%, 81.4%, 89.7%, and 95.5%, respectively; the specificity under the optimal threshold values were 78.6%, 86.7%, 88.6%, and 96.6%, respectively. The AUC quality of cystatin C was significantly better than that of creatinine (z=2.34, P<0.05), the AUC quality and sensitivity and specificity under the optimal threshold of HDL-C were all significantly better than those of cystatin C (z=3.33, with χ² values of 6.43 and 7.87, respectively, P<0.01) and creatinine (z=5.34, with χ² values of 6.32 and 6.41, respectively, P<0.01); the AUC quality and sensitivity and specificity under the optimal threshold of the joint prediction model were all significantly better than those of creatinine, cystatin C, and HDL-C (with z values of 6.18, 4.50, and 2.06, respectively, χ² values of 5.31, 7.23, 3.99, 6.56, 7.34, and 4.00, respectively, P<0.05 or P<0.01). Conclusions: HDL-C level in sepsis patients with secondary AKI is significantly lower than that in patients without secondary AKI. This is an independent risk factor for secondary AKI in sepsis patients with a diagnostic value being superior to that of creatinine and cystatin C. The combination of the aforementioned three indicators would have higher predicative valuable for secondary AKI in sepsis patients.
Acute Kidney Injury/etiology* ; Adult ; Aged ; Cholesterol, HDL ; Female ; Humans ; Male ; Middle Aged ; Prognosis ; ROC Curve ; Retrospective Studies ; Sepsis/diagnosis* ; Young Adult

PURPOSE: To investigate the clinical value of urine interleukin-18 (IL-8), neutrophil gelatinase-associated lipocalin (NGAL) and kidney injury molecule-1 (KIM-1) for the early diagnosis of acute kidney injury (AKI) in patients with ureteroscopic lithotripsy (URL) related urosepsis. METHODS: A retrospective study was carried out in 157 patients with urosepsis after URL. The patients were divided into AKI group and non-AKI group according to the Kidigo guideline and urine IL-8, NGAL and KIM-1 levels were detected by enzyme-linked immunosorbent assay at 0, 4, 12, 24 and 48 h after the surgery. Receiver operating characteristic curve (ROC) was used to evaluate the diagnostic value of these three biomarkers for postoperative AKI. RESULTS: The level of urine IL-8, NGAL and KIM-1 in AKI group was significantly higher than that in non-AKI group at 4, 12, 24 and 48 h (p < 0.01). The ROC analysis showed the combined detection of urine IL-8, NGAL and KIM-1 at 12 h had a larger area under curve (AUC) than a single marker (0.997, 95% CI: 0.991-0.998), and the sensitivity and specificity were 98.2% and 96.7%, respectively. Pearson correlation analysis showed that the levels of urine NGAL at 4, 12, 24 and 48 h in AKI patients were positively correlated with the levels of urine KIM-1 and IL-18 (p < 0.01). CONCLUSION AKI could be quickly recognized by the elevated level of urine IL-8, NGAL and KIM-1 in patients with URL-related urosepsis. Combined detection of the three urine biomarkers at 12 h after surgery had a better diagnostic performance, which may be an important reference for the early diagnosis of AKI.
Acute Kidney Injury/etiology* ; Biomarkers ; Early Diagnosis ; Hepatitis A Virus Cellular Receptor 1 ; Humans ; Interleukin-18 ; Interleukin-8 ; Lipocalin-2 ; Lithotripsy ; Retrospective Studies ; Ureteroscopy

BACKGROUND: Acute kidney injury (AKI) is a common and serious complication following lung transplantation (LTx), and it is associated with high mortality and morbidity. This study assessed the incidence of AKI after LTx and analyzed the associated perioperative factors and clinical outcomes. METHODS: This retrospective study included all adult LTx recipients at the China-Japan Friendship Hospital in Beijing between March 2017 and December 2019. The outcomes were AKI incidence, risk factors, mortality, and kidney recovery. Multivariate analysis was performed to identify independent risk factors. Survival analysis was presented using the Kaplan-Meier curves. RESULTS: AKI occurred in 137 of the 191 patients (71.7%), with transient AKI in 43 (22.5%) and persistent AKI in 94 (49.2%). AKI stage 1 occurred in 27/191 (14.1%), stage 2 in 46/191 (24.1%), and stage 3 in 64/191 (33.5%) of the AKI patients. Renal replacement therapy (RRT) was administered to 35/191 (18.3%) of the patients. Male sex, older age, mechanical ventilation (MV), severe hypotension, septic shock, multiple organ dysfunction (MODS), prolonged extracorporeal membrane oxygenation (ECMO), reintubation, and nephrotoxic agents were associated with AKI (P < 0.050). Persistent AKI was independently associated with pre-operative pulmonary hypertension, severe hypotension, post-operative MODS, and nephrotoxic agents. Severe hypotension, septic shock, MODS, reintubation, prolonged MV, and ECMO during or after LTx were related to severe AKI (stage 3) (P < 0.050). Patients with persistent and severe AKI had a significantly longer duration of MV, longer duration in the intensive care unit (ICU), worse downstream kidney function, and reduced survival (P < 0.050). CONCLUSIONS AKI is common after LTx, but the pathogenic mechanism of AKI is complicated, and prerenal causes are important. Persistent and severe AKI were associated with poor short- and long-term kidney function and reduced survival in LTx patients.
Acute Kidney Injury/etiology* ; Aged ; Humans ; Incidence ; Lung Transplantation/adverse effects* ; Male ; Renal Replacement Therapy ; Retrospective Studies ; Risk Factors

With the advancement of disease treatments, the number of patients undergoing surgery worldwide is increasing. However, many patients still experience severe perioperative complications. Perioperative hypotension is one of the common side effects during surgery. Physiologically, perioperative hypotension can lead to insufficient perfusion of important organs and result in acute and chronic irreversible organ injury, which cause serious consequences for the patient's postoperative hospitalization and even the long-term outcome. Therefore, in order to optimize perioperative circulation management and improve the quality of life for patients after surgery, it is of great importance to investigate the relationship between perioperative hypotension and postoperative myocardial injury, ischemic stroke, postoperative delirium, acute kidney injury, and postoperative mortality. Individualized circulation management and reasonable application of vasoactive drugs may be the key point to early prevention and correct treatment of perioperative hypotension, which is of great significance for reducing perioperative related morbidity and mortality and improving the prognosis for the surgical patients.
Acute Kidney Injury/etiology* ; Humans ; Hypotension/etiology* ; Postoperative Complications/etiology* ; Quality of Life

Acute Kidney Injury/etiology* ; Humans ; Platelet Count ; Rhabdomyolysis/etiology* ; Risk Factors

OBJECTIVE: To explore the influence of intraoperative urine volume on postoperative acute kidney injury (AKI) and the independent risk factors of AKI. METHODS: This was a retrospective cohort study recruiting patients who received selective pulmonary resection under general anesthesia in Peking University First Hospital from July, 2017 to June, 2019. The patients were divided into the AKI group and the control group according to whether they developed postoperative AKI or not. Firstly, univariate analysis was used to analyze the relationship between perioperative variables and postoperative AKI. Secondly, receiver operating characteristic (ROC) curve was used to explore the predictive value of intraoperative urine output for postoperative AKI. The nearest four cutoff values [with the interval of 0.1 mL/(kg·h)] at maximum Youden index were used as cutoff values of oliguria. Then univariate analysis was used to explore the relationship between oliguria defined by these four cutoff values and the risk of AKI. And the cutoff value with maximum OR was chosen as the threshold of oliguria in this study. Lastly, the variables with P < 0.10 in the univariate analysis were selected for inclusion in a multivariate Logistic model to analyze the independent predictors of postoperative AKI. RESULTS: A total of 1 393 patients were enrolled in the study. The incidence of postoperative AKI was 2.2%. ROC curve analysis showed that the area under curve (AUC) of intraoperative urine volume used for predicting postoperative AKI was 0.636 (P=0.009), and the cutoff value of oliguria was 0.785 mL/(kg·h) when Youden index was maximum (Youden index =0.234, sensitivity =48.4%, specificity =75.0%). Furthermore, 0.7, 0.8, 0.9, 1.0 mL/(kg·h) and the traditional cutoff value of 0.5 mL/(kg·h) were used to analyze the influence of oliguria on postoperative AKI. Univariate analysis showed that, when 0.8 mL/(kg·h) was selected as the threshold of oliguria, the patients with oliguria had the most significantly increased risk of AKI (AKI group 48.4% vs. control group 25.3%, OR=2.774, 95%CI 1.357－5.671, P=0.004). Multivariate regression analysis showed that intraoperative urine output < 0.8 mL/(kg·h) was one of the independent risk factors of postoperative AKI (OR=2.698, 95%CI 1.260－5.778, P=0.011). The other two were preoperative hemoglobin ≤120.0 g/L (OR=3.605, 95%CI 1.545－8.412, P=0.003) and preoperative estimated glomerular filtration rate < 30 mL/(min·1.73 m²) (OR=11.009, 95%CI 1.813－66.843, P=0.009). CONCLUSION Oliguria is an independent risk fact or of postoperative AKI after pulmonary resection, and urine volume < 0.8 mL/(kg·h) is a possible screening criterium.
Acute Kidney Injury/etiology* ; Humans ; Lung ; Oliguria/etiology* ; Postoperative Complications/etiology* ; Postoperative Period ; Retrospective Studies ; Risk Factors