Acute kidney injury (AKI) is a common and serious organ dysfunction in critically ill children. Early identification and prediction of AKI are of great significance. However, current AKI criteria are insufficiently sensitive and specific, and AKI heterogeneity limits the clinical value of AKI biomarkers. This study aimed to establish and validate an explainable prediction model based on the machine learning (ML) approach for AKI, and assess its prognostic implications in children admitted to the pediatric intensive care unit (PICU). This multicenter prospective study in China was conducted on critically ill children for the derivation and validation of the prediction model. The derivation cohort, consisting of 957 children admitted to four independent PICUs from September 2020 to January 2021, was separated for training and internal validation, and an external data set of 866 children admitted from February 2021 to February 2022 was employed for external validation. AKI was defined based on serum creatinine and urine output using the Kidney Disease: Improving Global Outcome (KDIGO) criteria. With 33 medical characteristics easily obtained or evaluated during the first 24 h after PICU admission, 11 ML algorithms were used to construct prediction models. Several evaluation indexes, including the area under the receiver-operating-characteristic curve (AUC), were used to compare the predictive performance. The SHapley Additive exPlanation method was used to rank the feature importance and explain the final model. A probability threshold for the final model was identified for AKI prediction and subgrouping. Clinical outcomes were evaluated in various subgroups determined by a combination of the final model and KDIGO criteria. The random forest (RF) model performed best in discriminative ability among the 11 ML models. After reducing features according to feature importance rank, an explainable final RF model was established with 8 features. The final model could accurately predict AKI in both internal (AUC = 0.929) and external (AUC = 0.910) validations, and has been translated into a convenient tool to facilitate its utility in clinical settings. Critically ill children with a probability exceeding or equal to the threshold in the final model had a higher risk of death and multiple organ dysfunctions, regardless of whether they met the KDIGO criteria for AKI. Our explainable ML model was not only successfully developed to accurately predict AKI but was also highly relevant to adverse outcomes in individual children at an early stage of PICU admission, and it mitigated the concern of the "black-box" issue with an undirect interpretation of the ML technique. The National Natural Science Foundation of China, Jiangsu Province Science and Technology Support Program, Key talent of women's and children's health of Jiangsu Province, and Postgraduate Research & Practice Innovation Program of Jiangsu Province.

Background: Hyperlactatemia upon admission is a documented risk factor for mortality in critically ill adult patients. However, the predictive significance of a single lactate measurement at admission for mortality in the general population of critically ill children remains uncertain. This study evaluated the predictive value of blood lactate levels at admission and determined the cut-off values for predicting in-hospital mortality in the critically ill pediatric population. Methods: We enrolled 1109 critically ill children who were admitted to a pediatric intensive care unit between July 2008 and December 2010. Arterial blood samples were collected in the first 2 hours after admission, and the lactate levels were determined. The Pediatric Risk of Mortality III (PRISM III) scores were calculated during the first 24 hours after admission. Results: Of the 1109 children admitted, 115 (10.4%) died in the hospital. The median (interquartile range) blood lactate level in critically ill children was 3.2 mmol/l (2.2-4.8). Among the children, 859 (77.5%) had a lactate concentration >2.0 mmol/l. The blood lactate level upon admission was significantly associated with mortality (odds ratio [OR] = 1.38; 95% confidence interval [CI], 1.30-1.46; p <0.001), even after adjustment for age, gender, and illness severity assessed by PRISM III (OR = 1.27; p <0.001). Multivariate regression analysis showed that a high blood lactate level (OR = 1.17; 95% CI, 1.07-1.29; p = 0.001), a high PRISM III score (OR = 1.15; 95% CI, 1.11-1.20; p <0.001), and a low serum albumin (OR =0.92; 95% CI, 0.88-0.96; p <0.001) were independent risk factors for mortality in critically ill children. Blood lactate achieved an area under-the-receiver-operating-characteristic curve (AUC) of 0.79 (p <0.001) for predicting mortality that was similar to that of PRISM III (AUC = 0.82; p <0.001). The p-value for a comparison of both AUCs was 0.318. Blood lactate displayed a sensitivity of 61% and a specificity of 86% in predicting mortality at the optimal cut-off value of 5.55 mmol/l, and the positive and negative likelihood ratios were 4.5 and 0.45, respectively. Conclusions: A high blood lactate level at admission is independently associated with and predictive of in-hospital mortality in the general population of critically ill children.

To evaluate whether early and acquired daily fluid overload (FO), as well as fluctuations in fluid accumulation, were associated with adverse outcomes in critically ill children with severe sepsis. This study enrolled 202 children in a pediatric intensive care unit (PICU) with severe sepsis. Early fluid overload was defined as ≥5% fluid accumulation occurring in the first 24 hours of PICU admission. The maximum daily fluid accumulation ≥5% occurring during the next 6 days in patients with at least 48 hours of PICU stay was defined as PICU-acquired daily fluid overload. The fluctuation in fluid accumulation was calculated as the difference between the maximum and the minimum daily fluid accumulation obtained during the first 7 days after admission. Of the 202 patients, 61 (30.2%) died during PICU stay. Among all patients, 41 (20.3%) experienced early fluid overload, including 9 with a FO ≥10%. Among patients with at least 48 hours of PICU stay (n = 154), 36 (23.4%) developed PICU-acquired daily fluid overload, including 2 with a FO ≥10%. Both early fluid overload (AOR = 1.20; 95% CI 1.08–1.33; P = 0.001; n = 202) and PICU-acquired daily fluid overload (AOR = 5.47 per log increase; 95% CI 1.15–25.96; P = 0.032; n = 154) were independent risk factors associated with mortality after adjusting for age, illness severity, etc. However, fluctuations in fluid accumulation were not associated with mortality after adjustment. Length of PICU stay increased with greater fluctuations in fluid accumulation in all patients with at least 48 hours of PICU stay (FO <5%, 5%-10% vs. ≥10%: 4 [3–8], 7 [4–11] vs. 10 [6–16] days; P <0.001; n = 154) and in survivors (4 [3–8], 7 [5–11] vs. 10 [5–15] days; P <0.001; n = 121). Early fluid overload achieved an area under-the-receiver-operating-characteristic curve of 0.74 (95% CI 0.65–0.82; P <0.001; n = 202) for predicting mortality in patients with severe sepsis, with a sensitivity of 67.2% and a specificity of 80.1% at the optimal cut-off value of 2.65%. Both early and acquired daily fluid overload were independently associated with PICU mortality in children with severe sepsis.

Background Urinary interleukin-18 and cystatin-C are suggested to be biomarkers for predicting acute kidney injury (AKI). The aims of this study are to examine whether the urinary concentrations of interleukin-18 and cystatin-C vary with gestational age and other factors in non-AKI control neonates, and to determine whether urinary interleukin-18 and cystatin-C can predict AKI development in non-septic critically ill neonates, independently of potential confounders. Methods We enrolled 62 non-septic critically ill neonates. Urine was collected every 48–72 h during the first 10 days of life. Results Urinary concentration of cystatin-C, but not interleukin-18, decreased with increasing gestational age and body weight, but not with increasing postnatal age in non-AKI control neonates. Both urinary interleukin-18 and cystatin-C were associated with AKI, even after controlling for gestational and postnatal age, birth weight, gender, Apgar score and the score for neonatal acute physiology in non-septic critically ill neonates. Urinary interleukin-18 and cystatin-C had odds ratios of 2.27 and 2.07, and achieved the area under-the-receiver-operating-characteristic curve of 0.72 and 0.92, respectively, for predicting AKI. Conclusions The urinary concentration of cystatin-C, but not interleukin-18, may decrease with increasing renal maturity. Both urinary interleukin-18 and cystatin-C are independently predictive of AKI in non-septic critically ill neonates.

Background: Survivin, a member of the family of inhibitor of apoptosis proteins, functions as a key regulator of mitosis and programmed cell death. YM155, a novel molecular targeted agent, suppresses survivin, which is overexpressed in many tumor types. The aim of this study was to determine the antitumor activity of YM155 in SK-NEP-1 cells. Methods: SK-NEP-1 cell growth in vitro and in vivo was assessed by MTT and nude mice experiments. Annexin V/propidium iodide staining followed by flow cytometric analysis was used to detect apoptosis in cell culture. Then gene expression profile of tumor cells treated with YM155 was analyzed with real-time PCR arrays. We then analyzed the expression data with MEV (Multi Experiment View) cluster software. Datasets representing genes with altered expression profile derived from cluster analyses were imported into the Ingenuity Pathway Analysis tool. Results: YM155 treatment resulted in inhibition of cell proliferation of SK-NEP-1cells in a dose-dependent manner. Annexin V assay, cell cycle, and activation of caspase-3 demonstrates that YM155 induced apoptosis in SK-NEP-1 cells. YM155 significantly inhibited growth of SK-NEP-1 xenografts (YM155 5 mg/kg: 1.45 ± 0.77 cm³; YM155 10 mg/kg: 0.95 ± 0.55 cm³) compared to DMSO group (DMSO: 3.70 ± 2.4 cm³) or PBS group cells (PBS: 3.78 ± 2.20 cm³, ANOVA P < 0.01). YM155 treatment decreased weight of tumors (YM155 5 mg/kg: 1.05 ± 0.24 g; YM155 10 mg/kg: 0.72 ± 0.17 g) compared to DMSO group (DMSO: 2.06 ± 0.38 g) or PBS group cells (PBS: 2.36 ± 0.43 g, ANOVA P < 0.01). Real-time PCR array analysis showed between Test group and control group there are 32 genes significantly up-regulated and 54 genes were significantly down-regulated after YM155 treatment. Ingenuity pathway analysis (IPA) showed cell death was the highest rated network with 65 focus molecules and the significance score of 44. The IPA analysis also groups the differentially expressed genes into biological mechanisms that are related to cell death, cellular function maintenance, cell morphology, carbohydrate metabolism and cellular growth and proliferation. Death receptor signaling (3.87E-19), TNFR1 signaling, induction of apoptosis by HIV1, apoptosis signaling and molecular mechanisms of cancer came out to be the top four most significant pathways. IPA analysis also showed top molecules up-regulated were BBC3, BIRC3, BIRC8, BNIP1, CASP7, CASP9, CD5, CDKN1A, CEBPG and COL4A3, top molecules down-regulated were ZNF443, UTP11L, TP73, TNFSF10, TNFRSF1B, TNFRSF25, TIAF1, STK17A, SST and SPP1, upstream regulator were NR3C1, TP53, dexamethasone , TNF and Akt. Conclusions: The present study demonstrates that YM155 treatment resulted in apoptosis and inhibition of cell proliferation of SK-NEP-1cells. YM155 had significant role and little side effect in the treatment of SK-NEP-1 xenograft tumors. Real-time PCR array analysis firstly showed expression profile of genes dyes-regulated after YM155 treatment. IPA analysis also represents new molecule mechanism of YM155 treatment, such as NR3C1 and dexamethasone may be new target of YM155. And our results may provide new clues of molecular mechanism of apoptosis induced by YM155.