OBJECTIVE:To establish a nomogram prediction model for septic shock in patients with sepsis based on serum sphingosine-1-phosphate (S1P), microRNA-155-5p (miR-155-5p) and soluble suppression of tumorigenicity 2 protein (sST2), and evaluate its predictive value. METHODS:A prospective observational study was conducted. A total of 400 patients with sepsis admitted to Liaocheng Second People's Hospital from January 2022 to January 2025 were enrolled. The patients were randomly divided into the training set and the validation set at a ratio of 7 : 3. General clinical data were collected. Arterial blood gas analysis, routine blood test, liver and renal function tests, and inflammatory indicators were measured upon admission. Enzyme linked immunosorbent assay was used to measure serum levels of S1P and sST2. Real-time fluorescence quantitative reverse transcription-polymerase chain reaction was applied to detect the expression of serum miR-155-5p. Disease severity and organ dysfunction were also assessed. Patients were divided into sepsis group and septic shock group according to the occurrence of septic shock. Multivariate Logistic regression analysis was performed to screen independent influencing factors for septic shock, and a nomogram prediction model was constructed. Bootstrap resampling, receiver operator characteristic curve (ROC curve), calibration curve and decision curve analysis were used to validate and evaluate the model. RESULTS:A total of 400 patients with sepsis were enrolled. Septic shock developed in 94 (33.6%) of the 280 patients in the training set and 44 (36.7%) of the 120 patients in the validation set. There were no statistically significant differences in baseline clinical characteristics between the training set and the validation set (all P>0.05). In the training set, compared with the sepsis group, the septic shock group had a higher proportion of patients aged over 65 years (60.6% vs. 45.2%), and increased levels of lactic acid (mmol/L: 1.84±0.31 vs. 1.72±0.25), procalcitonin [PCT (μg/L): 11.23±3.15 vs. 8.54±2.57], interleukin-6 [IL-6 (ng/L): 138.37±36.12 vs. 112.54±31.73], Sequential Organ Failure Assessment (SOFA: 10.34±3.15 vs. 8.26±2.83), Acute Physiology and Chronic Health Evaluation II (APACHE II: 26.43±6.32 vs. 23.57±5.29), sST2 (μg/L: 615.73±117.85 vs. 473.68±84.37) and miR-155-5p (miR-155-5p/U6: 1.74±0.36 vs. 1.35±0.28), the level of S1P was significantly decreased (μg/L: 67.38±18.31 vs. 87.54±24.67), all differences were statistically significant (all P<0.05). Multivariate Logistic regression analysis indicated that age over 65 years old [odds ratio (OR)=1.425, 95% confidence interval (95%CI) was 1.050-1.935], elevated lactic acid (OR=1.538, 95%CI was 1.023-2.312), elevated PCT (OR=1.624, 95%CI was 1.049-2.514), higher SOFA score (OR=1.674, 95%CI was 1.107-2.531), higher APACHE II score (OR=2.547, 95%CI was 1.460-4.440), decreased S1P (OR=0.615, 95%CI was 0.405-0.934), elevated sST2 (OR=2.017, 95%CI was 1.270-3.203) and elevated miR-155-5p (OR=1.628, 95%CI was 1.072-2.472) were independent influencing factors for septic shock in sepsis patients (all P<0.05). A nomogram was established based on the above factors. ROC curve analysis showed that the area under the curve (AUC) was 0.865 (95%CI was 0.819-0.903) in the training set and 0.841 (95%CI was 0.793-0.882) in the validation set. The model presented good calibration, favourable net clinical benefit and acceptable threshold probability in both sets. CONCLUSIONS:The nomogram prediction model established based on serum S1P, sST2 and miR-155-5p has good predictive performance for septic shock in patients with sepsis, which can provide a reference for clinical decision-making.
OBJECTIVE:To investigate the incidence and severity of acute kidney injury (AKI) in neurocritical patients at high altitude and its impact on prognosis. METHODS:A retrospective analysis was conducted. The clinical data of adult critically ill neurological patients admitted to the critical care medicine of People's Hospital of Xizang Autonomous Region from January 1, 2022, to February 29, 2024 were enrolled. AKI was diagnosed according to the Kidney Disease: Improving Global Outcomes (KDIGO) serum creatinine criteria. The incidence, underdiagnosis rate and severity of AKI within 7 days after intensive care unit (ICU) admission were analyzed. The primary endpoint was in-hospital adverse outcome, defined as death or withdrawal of treatment during hospitalization. The secondary endpoint included persistent deterioration of renal function at discharge, length of ICU stay, total hospital stay and hospitalization costs. Univariate and multivariate Logistic regression analyses were performed to examine the relationship between AKI and adverse outcomes in patients with neurocritical illness. RESULTS:Among the 390 critically ill neurological patients at high altitude, the incidence of AKI was 21.3% (83/390), with an underdiagnosis rate of 50.6% (42/83). Among the 83 patients with AKI, 42 (50.6%) had stage 1 AKI, 21 (25.3%) stage 2, and 20 (24.1%) stage 3; moderate to severe AKI (stage 2-3) accounted for 49.4% (41/83). In addition, most AKI cases (83.1%) developed within 3 days after ICU admission. During hospitalization, 102 patients experienced adverse outcomes, including 65 deaths and 37 cases of treatment withdrawal. The prevalence of AKI was significantly higher in patients with in-hospital adverse outcome than that in patients with favorable prognosis (40.2% vs. 14.6%, P<0.05). Multivariate Logistic regression analysis demonstrated that AKI was an independent risk factor for in-hospital adverse outcome in neurocritical patients at high altitude [odds ratio (OR)=3.037, 95% confidence interval (95%CI) was 1.724-5.351, P<0.001]. After multivariate adjustment, the risk of adverse outcomes increased with the severity of AKI (stage 1 AKI: OR=2.734, 95%CI was 1.313-5.693, P=0.007; stage 2-3 AKI: OR=3.408, 95%CI was 1.609-7.218, P=0.001). The incidence of adverse outcome increased progressively with the aggravation of AKI severity, which was 42.9% (18/42) in stage 1, 47.6% (10/21) in stage 2, and 65.0% (13/20) in stage 3. In addition, compared with non-AKI patients, AKI patients had longer length of ICU stay [days: 14 (8, 22) vs. 11 (5, 16)], higher hospitalization costs [ten thousand yuan: 14.85 (8.88, 23.41) vs. 11.94 (7.03, 16.91)], and a higher rate of persistent renal function deterioration at discharge (27.7% vs. 2.3%), with all differences being statistically significant (all P<0.05). CONCLUSIONS:AKI exhibits a high incidence and common underdiagnosis in patients with high-altitude critical neurological illness. It mainly occurs at the early stage of ICU admission, with a high proportion of stage 2-3 cases. AKI is an independent risk factor for in-hospital adverse outcomes in such patients, and the risk of adverse outcomes increases with the severity of AKI stage.
OBJECTIVE:To explore the clinical value of dynamic bedside ultrasound monitoring of optic nerve sheath diameter (ONSD) in noninvasive and real-time evaluation of intracranial pressure (ICP)-lowering efficacy of mannitol in patients with spontaneous cerebral hemorrhage combined with intracranial hypertension, so as to provide objective bedside evidence for individualized mannitol administration. METHODS:A prospective self-controlled clinical trial was performed. Adult patients with spontaneous cerebral hemorrhage and intracranial hypertension who were diagnosed by cranial CT, underwent lumbar drainage, and had an ICP>15 mmHg (1 mmHg=0.133 kPa) and required mannitol dehydration treatment, and were admitted to the department of critical care medicine of Shizuishan Second People's Hospital from March 2024 to December 2025 were enrolled. All subjects received intravenous infusion of 20% mannitol 0.5 g/kg, which was completed within 30 minutes. High-frequency bedside ultrasound was used to measure ONSD simultaneously before and at 0.5, 1, 3, 6, 8, and 12 hours after treatment, and invasive ICP was recorded through the lumbar drainage tube. The trend of changes in ONSD and ICP was observed before and after treatment. Pearson correlation analysis was used to analyze the correlation between decrease in ONSD (ΔONSD) and the decrease in ICP (ΔICP) at 1 hours after treatment; intraclass correlation coefficient (ICC) was adopted to assess measurement repeatability; Bland-Altman analysis was used to evaluate the consistency between ONSD reduction rate and ICP reduction rate. The effective treatment of mannitol was judged by a decrease rate of ≥10% in ICP 1 hour after treatment. The receiver operator characteristic curve (ROC curve) was drawn to analyze the predictive efficiency of ΔONSD on the efficacy of mannitol 1 hour after treatment. RESULTS:Among the enrolled 59 patients, 35 male and 24 were female, with a mean age of (58.4±10.7) years old. Both ONSD and ICP decreased significantly at all time points after mannitol administration compared with baseline (F values were 25.634, 28.915, respectively, both P<0.05), and the pressure-lowering plateau was observed from 1 to 6 hours post-treatment. Pearson correlation analysis showed a positive correlation between ΔONSD and ΔICP 1 hour after treatment (r=0.695, P<0.001), and their reduction rates showed favorable consistency with 95% limit of agreement ranging from -8.5% to 9.1%. Among the 59 patients, 42 were effective and 17 were ineffective in short-term treatment with mannitol. ROC curve analysis showed that the area under the curve (AUC) for predicting the efficacy of mannitol using ΔONSD at 1 hour after treatment was 0.892, with a 95% confidence interval (95%CI) of 0.815-0.969; with the optimal cut-off value of 0.48 mm, sensitivity of 88.1%, specificity of 81.0% and Youden index of 0.691. Subgroup analysis showed that AUC predicted by ΔONSD at 1 hour after treatment were 0.876 and 0.903 in subgroups with baseline ICP of 15-20 mmHg (32 cases) and >20 mmHg (27 cases), respectively. The intra-observer ICC was 0.935 and inter-observer ICC was 0.912. CONCLUSIONS:Dynamic bedside ultrasound monitoring of ONSD can noninvasively and in real time reflect the ICP-lowering effect of mannitol, which is highly consistent with the changing trend of invasive ICP. ΔONSD≥0.48 mm at 1 hour after medication can serve as a quantitative bedside indicator to predict short-term mannitol efficacy, worthy of clinical popularization.
OBJECTIVE:To develop a stage prediction model for sepsis-associated acute kidney injury (SA-AKI) based on interpretable machine learning, thereby providing support for individualized clinical intervention. METHODS:Adult patients with sepsis were selected from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database and the eICU-Collaborative Research Database (eICU-CRD). Intensive care unit (ICU) admission was defined as time zero. Demographic data were extracted, along with clinical variables collected within 24 hours after ICU admission, including vital signs, laboratory parameters, organ function scores, and therapeutic interventions. The occurrence of acute kidney injury (AKI) and its highest stage within 7 days after ICU admission were used as the study outcomes. Data from the MIMIC-IV database were divided into training and internal validation sets in a 7 : 3 ratio, stratified by AKI stage, while data from eICU-CRD were used as the external validation set. LASSO regression and multivariable ordinal Logistic regression were used to select features associated with SA-AKI stage. Six machine learning models were developed, including eXtreme Gradient Boosting, Light Gradient Boosting Machine (LightGBM), random forest, K-nearest neighbors, support vector machine, and logistic regression. The optimal model was selected based on the area under the receiver operator characteristic curve (AUC), accuracy, F1-score, precision, specificity, and sensitivity. Calibration curve and decision curve analysis (DCA) were used to evaluate model performance, and SHapley Additive exPlanations (SHAP) were used to interpret the optimal model. RESULTS:A total of 7 415 and 2 727 patients were included from the MIMIC-IV and eICU-CRD databases, respectively. Thirteen features were selected using LASSO regression and multivariable ordinal Logistic regression, including 24-hour urine output [odds ratio (OR)=0.183, 95% confidence interval (95%CI) was 0.168-0.200, P<0.001], body mass index (OR=2.016, 95%CI was 1.887-2.155, P<0.001), invasive mechanical ventilation (OR=1.395, 95%CI was 1.315-1.478, P<0.001), congestive heart failure (OR=1.405, 95%CI was 1.321-1.494, P<0.001), male (OR=1.088, 95%CI was 1.028-1.152, P=0.004), mean serum potassium level (OR=1.109, 95%CI was 1.044-1.178, P<0.001), mean activated partial thromboplastin time (OR=1.212, 95%CI was 1.141-1.288, P<0.001), maximum lactate level (OR=1.373, 95%CI was 1.282-1.470, P<0.001), mean body temperature (OR=0.875, 95%CI was 0.823-0.930, P<0.001), minimum pulse oxygen saturation (OR=0.811, 95%CI was 0.763-0.862, P<0.001), diabetes mellitus (OR=1.131, 95%CI was 1.067-1.199, P<0.001), maximum total bilirubin level (OR=1.345, 95%CI was 1.246-1.451, P<0.001), and vasoactive drug use (OR=1.160, 95%CI was 1.092-1.233, P<0.001). Compared with other models, LightGBM was identified as the best-performing model, and its predicted probabilities were well calibrated. The model achieved AUCs of 0.872 and 0.851 for predicting SA-AKI stage in the internal and external validation sets. In the internal validation set, the AUCs for predicting non-AKI and AKI stages 1, 2 and 3 were 0.909, 0.837, 0.838, and 0.904, respectively, while the corresponding AUCs in the external validation set were 0.837, 0.818, 0.815, and 0.937. The DCA results showed that the highest net benefit was achieved within a threshold range of 0.1-0.8 for patients with non-AKI and AKI stage 2 and stage 3, and within a threshold range of 0.1-0.4 for patients with AKI stage 1. Differences in feature importance across SA-AKI stages were revealed by SHAP analysis. The 24-hour urine output was ranked first in most stages, whereas invasive mechanical ventilation was identified as the most important feature among non-AKI patients in the external validation set. CONCLUSIONS:The LightGBM-based model for predicting SA-AKI stages was successfully developed, and key features affecting each stage were identified using SHAP, providing an interpretable tool for the clinical identification of different SA-AKI stages.
OBJECTIVE:To investigate the relationship between serum G protein-coupled receptor kinase 2 (GRK2) level and the occurrence of sepsis-associated liver injury (SALI). METHODS:A prospective observational study was conducted. Patients with sepsis admitted to the department of intensive care medicine of Northern Jiangsu People's Hospital Affiliated to Yangzhou University from January to December 2024 were enrolled. Baseline data were collected. Peripheral venous blood samples were obtained within 24 hours to measure laboratory indicators and detect serum GRK2 level using enzyme linked immunosorbent assay (ELISA). Patients were divided into SALI and non-SALI groups based on whether SALI occurred at enrollment. Logistic regression analysis was used to identify independent risk factors for SALI, and the identify value of serum GRK2 for SALI was evaluated using receiver operator characteristic curve (ROC curve). Internal validation was performed via Bootstrap resampling with 1 000 repetitions. RESULTS:A total of 144 patients with sepsis were finally enrolled, including 82 in the non-SALI group and 62 in the SALI group. The level of serum GRK2 was significantly higher in the SALI group than that in the non-SALI group [ng/L: 808.78 (766.70, 832.14) vs. 586.40 (553.43, 614.19), P<0.05]. Additionally, compared with the non-SALI group, the SALI group showed significantly higher body mass index (BMI), proportion of smoking history, 28-day mortality, lactate, alanine aminotransferase, aspartate aminotransferase, total bilirubin, indirect bilirubin, and direct bilirubin, prothrombin time and international normalized ratio (all P<0.05). The indicators with statistically significant differences in univariate Logistic regression analysis, such as BMI, smoking history, GRK2, and lactate, were included in the multivariate Logistic regression analysis. The results showed that, elevated GRK2 [odds ratio (OR)=1.024, 95% confidence interval (95%CI) was 1.016-1.033, P<0.001], elevated lactate (OR=1.674, 95%CI was 1.160-2.418, P=0.006), and high BMI (OR=1.243, 95%CI was 1.032-1.498, P=0.022) were independent risk factors for SALI. ROC curve analysis showed that the area under the curve (AUC) of GRK2 for identifying SALI was 0.940, with a 95%CI of 0.893-0.988; the sensitivity was 0.96 and the specificity was 0.89 when the optimal cut-off value of 692.52 ng/L. The AUC of GRK2 combined with lactate was 0.946, with a 95%CI of 0.903-0.989. After internal validation, the AUC of GRK2 for identifying SALI was 0.940, with a 95%CI of 0.893-0.988. CONCLUSIONS:Elevated serum GRK2 was positively correlated with the occurrence of SALI and may serve as a potential biomarker for assisting in the diagnosis of SALI.
