Application of artificial neural network in daily prediction of bleeding in ICU patients treated with anti-thrombotic therapy.

OBJECTIVES: Anti-thrombotic therapy is the basis of thrombosis prevention and treatment. Bleeding is the main adverse event of anti-thrombosis. Existing laboratory indicators cannot accurately reflect the real-time coagulation function. It is necessary to develop tools to dynamically evaluate the risk and benefits of anti-thrombosis to prescribe accurate anti-thrombotic therapy. METHODS: The prediction model,daily prediction of bleeding risk in ICU patients treated with anti-thrombotic therapy, was built using deep learning algorithm recurrent neural networks, and the model results and performance were compared with clinicians. RESULTS: There was no significant statistical discrepancy in the baseline. ROC curves of the four models in the validation and test set were drawn, respectively. One-layer GRU of the validation set had a larger AUC (0.9462; 95%CI, 0.9147-0.9778). Analysis was conducted in the test set, and the ROC curve showed the superiority of two layers LSTM over one-layer GRU, while the former AUC was 0.8391(95%CI, 0.7786-0.8997). One-layer GRU in the test set possessed a better specificity (sensitivity 0.5942; specificity 0.9300). The Fleiss' k of junior clinicians, senior clinicians, and machine learning classifiers is 0.0984, 0.4562, and 0.8012, respectively. CONCLUSIONS: Recurrent neural networks were first applied for daily prediction of bleeding risk in ICU patients treated with anti-thrombotic therapy. Deep learning classifiers are more reliable and consistent than human classifiers. The machine learning classifier suggested strong reliability. The deep learning algorithm significantly outperformed human classifiers in prediction time.

Efficacy and safety of awake prone positioning in the treatment of non-intubated spontaneously breathing patients with COVID-19-related acute respiratory failure: A systematic review and meta-analysis.

Background: Since the beginning of the coronavirus disease 2019 (COVID-19) pandemic, prone positioning has been widely applied for non-intubated, spontaneously breathing patients. However, the efficacy and safety of prone positioning in non-intubated patients with COVID-19-related acute hypoxemic respiratory failure remain unclear. We aimed to systematically analyze the outcomes associated with awake prone positioning (APP). Methods: We conducted a systematic literature search of PubMed/MEDLINE, Cochrane Library, Embase, and Web of Science from January 1, 2020, to June 3, 2022. This study included adult patients with acute respiratory failure caused by COVID-19. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed, and the study quality was assessed using the Cochrane risk-of-bias tool. The primary outcome was the reported cumulative intubation risk across randomized controlled trials (RCTs), and the effect estimates were calculated as risk ratios (RRs; 95% confidence interval [CI]). Results: A total of 495 studies were identified, of which 10 fulfilled the selection criteria, and 2294 patients were included. In comparison to supine positioning, APP significantly reduced the need for intubation in the overall population (RR=0.84, 95% CI: 0.74-0.95). The two groups showed no significant differences in the incidence of adverse events (RR=1.16, 95% CI: 0.48-2.76). The meta-analysis revealed no difference in mortality between the groups (RR=0.93, 95% CI: 0.77-1.11). Conclusions: APP was safe and reduced the need for intubation in patients with respiratory failure associated with COVID-19. However, it did not significantly reduce mortality in comparison to usual care without prone positioning.

Association between input/weight ratio and acute kidney injury in obese critical ill patients: a propensity analysis of multicenter clinical databases.

Rehydration volume may be underestimated in obese critically ill patients, which can lead to acute kidney injury (AKI). This study aimed to investigate the association between input/weight ratio (IWR) and AKI risk in obese critical patients. This retrospective observational study analyzed data from three large open databases. Patients were divided into lean and obese groups and matched 1:1 based on age, sex, APACHE II score, SOFA score, sepsis status, mechanical ventilation status, renal replacement therapy status, and hospital type. The exposure of interest was the mean IWR during the first three ICU admission days. The primary outcome was the incidence of AKI within 28 days after ICU admission. Cox regression analysis was used to evaluate the association between IWR and AKI risk. A total of 82,031 eligible patients were included in the study, with 25,427 obese patients matched with 25,427 lean patients. The IWRs were significantly lower in the obese groups in both the unmatched cohort (35.85 ± 19.05 vs. 46.01 ± 30.43 ml/kg, p < 0.01) and the matched cohort (36.13 ± 19.16 vs. 47.34 ± 31.13 ml/kg, p < 0.01). An increase in IWR was significantly associated with decreased creatinine levels, increased urine output and a lower AKI risk. The interaction terms of IWR and obesity were significantly associated with decreased AKI incidence in both the unmatched cohort (hazard ratio [HR] = 0.97, 95% CI 0.96-0.97, p < 0.01) and the matched cohort (HR = 0.97, 95% CI 0.96-0.97, p < 0.01). Inadequate rehydration of patients with obesity may contribute to an increased risk of AKI in patients with obesity. These results highlight the need for better rehydration management in patients with obesity.

Paraquat is an agonist of STIM1 and increases intracellular calcium levels.

Paraquat (PQ) is an efficient herbicide but leads to high mortality with no antidote in mammals. PQ produces reactive oxygen species (ROS), leading to epithelial-mesenchymal transition (EMT) for pulmonary fibrosis in type II alveolar (AT II) cells. Intriguingly, strategies reducing ROS exhibit limited therapeutic effects, indicating other targets existing for PQ toxicity. Herein we report that PQ is also an agonist for STIM1 that increases intracellular calcium levels. Particularly, PQ promotes STIM1 puncta formation and association with TRPC1 or ORAI for extracellular calcium entry and thus intracellular calcium influx. Further studies reveal the importance of P584&Y586 residues in STIM1 for PQ association that facilitates STIM1 binding to TRPC1. Consequently, the STIM1-TRPC1 route facilitates PQ-induced EMT for pulmonary fibrosis as well as cell death. Our results demonstrate that PQ is an agonist of STIM1 that induces extracellular calcium entry, increases intracellular calcium levels, and thus promotes EMT in AT II cells.

[Comparative analysis of clinical features between severe coronavirus disease 2019 and severe community acquired pneumonia].

OBJECTIVE: To compare and analyze the clinical features of patients with severe coronavirus disease 2019 (sCOVID-19) and severe community acquired pneumonia (sCAP) who meet the diagnostic criteria for severe pneumonia of the Infectious Diseases Society of America/American Thoracic Society (IDSA/ATS). METHODS: A retrospective comparative analysis of the clinical records of 116 patients with sCOVID-19 admitted to the department of critical care medicine of Wuhan Third Hospital from January 1, 2020 to March 31, 2020 and 135 patients with sCAP admitted to the department of critical care medicine of Shanghai First People's Hospital from January 1, 2010 to December 31, 2017 was conducted. The basic information, diagnosis and comorbidities, laboratory data, etiology and imaging results, treatment, prognosis and outcome of the patients were collected. The differences in clinical data between sCOVID-19 and sCAP patients were compared, and the risk factors of death were analyzed. RESULTS: The 28-day mortality of sCOVID-19 and sCAP patients were 50.9% (59/116) and 37.0% (50/135), respectively. The proportion of arterial partial pressure of oxygen/fraction of inspired oxygen (PaO2/FiO2) ≤ 250 mmHg (1 mmHg ≈ 0.133 kPa) in sCOVID-19 patients was significantly higher than that of sCAP [62.1% (72/116) vs. 34.8% (47/135), P < 0.01]. The possible reason was that the proportion of multiple lung lobe infiltration in sCOVID-19 was significantly higher than that caused by sCAP [94.0% (109/116) vs. 40.0% (54/135), P < 0.01], but the proportion of sCOVID-19 patients requiring mechanical ventilation was significantly lower than that of sCAP [45.7% (53/116) vs. 60.0% (81/135), P < 0.05]. Further analysis of clinical indicators related to patient death found that for sCOVID-19 patients PaO2/FiO2, white blood cell count (WBC), neutrophils (NEU), neutrophil percentage (NEU%), neutrophil/lymphocyte ratio (NLR), total bilirubin (TBil), blood urea nitrogen (BUN), albumin (ALB), Ca2+, prothrombin time (PT), D-dimer, C-reactive protein (CRP) and other indicators were significantly different between the death group and the survival group, in addition, the proportion of receiving mechanical ventilation, gamma globulin, steroid hormones and fluid resuscitation in death group were higher than survival group. Logistic regression analysis showed that the need for mechanical ventilation, NLR > 10, TBil > 10 µmol/L, lactate dehydrogenase (LDH) > 250 U/L were risk factors for death at 28 days. For sCAP patients, there were significant differences in age, BUN, ALB, blood glucose (GLU), Ca2+ and D-dimer between the death group and the survival group, but there was no significant difference in treatment. Logistic regression analysis showed that BUN > 7.14 mmol/L and ALB < 30 g/L were risk factors for 28-day death of sCAP patients. CONCLUSIONS: The sCOVID-19 patients in this cohort have worse oxygen condition and symptoms than sCAP patients, which may be due to the high proportion of lesions involving the lungs. The indicators of the difference between the death group and the survival group were similar in sCOVID-19 and sCAP patients. It is suggested that the two diseases have similar effects on renal function, nutritional status and coagulation function. But there were still differences in risk factors affecting survival. It may be that sCOVID-19 has a greater impact on lung oxygenation function, inflammatory cascade response, and liver function, while sCAP has a greater impact on renal function and nutritional status.

[Relationship between blood electrolytes and prognosis of patients with severe coronavirus disease 2019].

OBJECTIVE: To analyze the relationship between blood electrolytes and the prognosis of patients with severe coronavirus disease 2019 (COVID-19) and to provide assistance for clinical decision-making. METHODS: The clinical data of patients with severe COVID-19 admitted to intensive care unit (ICU) of the Wuhan Third Hospital by the Shanghai aid-Hubei medical team from January 21 to March 4, 2020 were collected. Excluding ineligible patients, 110 patients were finally enrolled. The patients' gender, age, temperature, heart rate, systolic and diastolic blood pressure, clinical symptoms at admission, time of symptom onset, duration of fever, and relevant indicators at admission to ICU (including blood potassium, chloride, sodium, calcium, phosphorus, and magnesium, etc.) and prognosis were analyzed. The patients were grouped by blood potassium or calcium levels or blood potassium/calcium ratio. The Kaplan-Meier survival curves were used to analyze the survival of patients in each group. The relationship between the potassium/calcium ratio and the prognosis was analyzed using restricted cubic spline plots. The relationship between each index in the different models and the prognosis was analyzed using Cox regression models. RESULTS: Among 110 severe COVID-19 patients, 78 cases survived, and 32 cases died. Compared with the surviving group, patients in the death group had higher blood potassium levels [mmol/L: 4.25 (3.80, 4.65) vs. 3.90 (3.60, 4.20), P < 0.05] and lower blood calcium levels (mmol/L: 2.00±0.14 vs. 2.19±0.18, P < 0.05). The Kaplan-Meier survival curves showed that patients in the potassium > 4.2 mmol/L group had a worse prognosis than the potassium < 3.8 mmol/L group and the potassium 3.8-4.2 mmol/L group (P = 0.011), patients in the calcium > 2.23 mmol/L group had a better prognosis than the calcium < 2.03 mmol/L group and the calcium 2.03-2.23 mmol/L group, and the lower calcium group had a worse prognosis (P = 0.000 15). Cox regression analysis showed that the hazard ratio (HR) of blood potassium and calcium were 2.08 and 0.01, respectively, in model 1 (single blood potassium or calcium) and in model 2 (model 1 plus age and gender), the HR of blood potassium and calcium were 1.98 and 0.01 respectively, which were significantly associated with patient prognosis (all P < 0.05). Patients in the group with the potassium/calcium ratio > 1.9 had higher blood potassium levels and a higher proportion of mechanical ventilation, lower calcium levels and lower proportion of survival, and longer time of ICU admission compared with the groups with the potassium/calcium ratio < 1.7 and 1.7-1.9. The Kaplan-Meier survival curves showed that the survival rate of the potassium/calcium ratio > 1.9 group was the lowest (P < 0.000 1), and there was no statistically significant difference in survival between the potassium/calcium ratio < 1.7 group and the potassium/calcium ratio 1.7-1.9 group. A restricted cubic spline plot corrected for age and gender showed that patients in the potassium/calcium ratio > 1.8 group had HR values > 1. Cox regression analysis corrected for other indicators showed that the potassium/calcium ratio was still associated with patient prognosis (HR = 4.85, P = 0.033). CONCLUSION: Blood potassium, calcium, and the potassium/calcium ratio at ICU admission are related to the prognosis of patients with severe COVID-19, and the potassium/calcium ratio is an independent risk factor for the death of patients. The higher the potassium/calcium ratio, the worse the prognosis of patients.

Relationship between early serum sodium and potassium levels and AKI severity and prognosis in oliguric AKI patients.

PURPOSE: Acute kidney injury (AKI) is a common organ dysfunction in ICU and up to now there is no good way to predict the AKI progression and patient prognosis. Blood electrolyte tests are common in ICU, but there are few studies on early blood electrolytes and the AKI progression and patient prognosis. Therefore, we concentrated on the serum sodium and potassium levels before AKI diagnosis and evaluated the relationship between serum sodium and potassium levels and the severity and prognosis of AKI. METHODS: This study included data of all patients from the MIMIC-III. We used the urine output criteria in the KDIGO as diagnostic criteria for oliguric AKI. Patients admitted to the ICU several times only included their initial ICU admission results. Patients younger than 18 years old, diagnosed with AKI stage 3, ICU stays less than 24 h or without corresponding laboratory results or data were excluded. The included patients were divided into four groups based on the interquartile range of serum sodium and potassium. We evaluated the serum sodium and potassium levels before AKI diagnosis and AKI severity and prognosis through retrospective analysis. RESULTS: Patients with serum potassium > 4.6 mmol/L were more likely to progress to AKI stage 3 or death than patients with serum potassium ≤ 4.6 mmol/L (overall p < 0.0001). Patients with sodium < 137 mmol/L or > 141 mmol/L had a higher risk of progressing to AKI stage 3 (overall p = 0.00023) and risk of death (overall p < 0.0001) than other patients. In the Cox regression model, after adjusting for age, sex, and BMI, serum sodium or potassium were associated with AKI progression and prognosis (p < 0.01). After continuing to adjust for comorbidities, serum potassium was still associated with AKI progression and prognosis (p < 0.01), but serum sodium was only associated with prognosis (p = 0.027). After adjusting for other indicators, there was no statistically significant correlation between serum sodium or potassium and AKI progression and prognosis. After adjusting for serum sodium or potassium, the corresponding results were not significantly different from those before adjustment. CONCLUSION: This study found that abnormal serum sodium or potassium levels before AKI diagnosis were more likely to lead to AKI progression and poor prognosis, of which lower serum sodium and higher serum potassium were more likely to progress to AKI stage 3 or death.