Prediction algorithm for ICU mortality and length of stay using machine learning.

Machine learning can predict outcomes and determine variables contributing to precise prediction, and can thus classify patients with different risk factors of outcomes. This study aimed to investigate the predictive accuracy for mortality and length of stay in intensive care unit (ICU) patients using machine learning, and to identify the variables contributing to the precise prediction or classification of patients. Patients (n = 12,747) admitted to the ICU at Chiba University Hospital were randomly assigned to the training and test cohorts. After learning using the variables on admission in the training cohort, the area under the curve (AUC) was analyzed in the test cohort to evaluate the predictive accuracy of the supervised machine learning classifiers, including random forest (RF) for outcomes (primary outcome, mortality; secondary outcome, length of ICU stay). The rank of the variables that contributed to the machine learning prediction was confirmed, and cluster analysis of the patients with risk factors of mortality was performed to identify the important variables associated with patient outcomes. Machine learning using RF revealed a high predictive value for mortality, with an AUC of 0.945 (95% confidence interval [CI] 0.922-0.977). In addition, RF showed high predictive value for short and long ICU stays, with AUCs of 0.881 (95% CI 0.876-0.908) and 0.889 (95% CI 0.849-0.936), respectively. Lactate dehydrogenase (LDH) was identified as a variable contributing to the precise prediction in machine learning for both mortality and length of ICU stay. LDH was also identified as a contributing variable to classify patients into sub-populations based on different risk factors of mortality. The machine learning algorithm could predict mortality and length of stay in ICU patients with high accuracy. LDH was identified as a contributing variable in mortality and length of ICU stay prediction and could be used to classify patients based on mortality risk.

Interhospital transportation of a COVID-19 patient undergoing veno-venous extracorporeal membrane oxygenation by helicopter.

Some coronavirus disease 2019 (COVID-19) patients develop rapidly progressive acute respiratory distress syndrome and require veno-venous extracorporeal membrane oxygenation (V-V ECMO). A previous study recommended the transfer of ECMO patients to ECMO centers. However, because of the pandemic, a limited number of ECMO centers are available for patient transfer. The safe long-distance interhospital transport of these patients is a concern. To minimize transportation time, helicopter use is a suitable choice. We report the first case of a COVID-19 patient on V-V ECMO, transferred to our ECMO center by helicopter. A 45-year-old man with rheumatoid arthritis history, treated with immunosuppressants, presented with fever and sore throat. He was diagnosed with COVID-19 following a positive severe acute respiratory syndrome coronavirus 2 polymerase chain reaction test result and was subsequently prescribed favipiravir. However, his respiratory failure progressively worsened. On day 10 of hospitalization at the previous hospital, he was intubated, and we received a request for ECMO transport on the next day. The ECMO team, who wore personal protective equipment (N95 respirators, gloves, gowns, and face shields), initiated V-V ECMO in the referring hospital and safely transported the patient by helicopter. The flight time was 7 min. He was admitted to the intensive care unit of our hospital and received tocilizumab. He was discharged on hospital day 31 with no significant sequelae. In this case report, we discuss important factors for the safe and appropriate interhospital transportation of COVID-19 patients on ECMO as well as staff and patient safety during helicopter transportation.

Interleukin-6 as a diagnostic marker for infection in critically ill patients: A systematic review and meta-analysis.

BACKGROUND: The ability of blood levels of interleukin (IL)-6 to differentiate between infection and non-infection in critically ill patients with suspected infection is unclear. We assessed the diagnostic accuracy of serum IL-6 levels for the diagnosis of infection in critically ill patients. METHODS: We systematically searched the PubMed, MEDLINE, Cochrane Resister of Controlled Trials, Cochrane Database of Systematic Reviews, CINAHL, and Igaku Chuo Zasshi databases for studies published from 1986 to August 2016 that evaluated the accuracy of IL-6 levels for the diagnosis of infection. We constructed 2 × 2 tables and calculated summary estimates of sensitivity and specificity using a bivariate random-effects model. RESULTS: The literature search identified 775 articles, six of which with a total of 527 patients were included according to the predefined criteria. The pooled sensitivity, specificity, and diagnostic odds ratio were 0.73 (95% confidence interval [CI], 0.61-0.82), 0.76 (95% CI, 0.61-0.87), and 2.31 (95% CI, 1.20-3.48), respectively. The area under the curve (AUC) of the summary receiver operator characteristic (SROC) curve was 0.81 (95% CI, 0.78-0.85). In the secondary analysis of two studies with a total of 263 adult critically ill patients with organ dysfunction, the pooled sensitivity, specificity, and diagnostic odds ratio were 0.81 (95% CI, 0.75-0.86), 0.77 (95% CI, 0.67-0.84), and 2.87 (95% CI 2.15-3.60), respectively. CONCLUSIONS: Blood levels of IL-6 have a moderate diagnostic value and a potential clinical utility to differentiate infection in critically ill patients with suspected infection.

Empiric Antibiotic Therapy for Severe Sepsis and Septic Shock.

BACKGROUND: A retrospective study was conducted to investigate the validity and the effectiveness of early empiric antibiotic and de-escalation therapy for the treatment of severe sepsis and septic shock patients in the intensive care unit (ICU). METHODS: Patients admitted to the ICU at Chiba University Hospital from January 1, 2010, to December 31, 2012, for the treatment of severe sepsis or septic shock were selected for analysis. RESULTS: One-hundred and ten patients were enrolled for the analysis. Carbapenems were selected most frequently (57.3%), followed by cephalosporins (22.7%), and penicillins (21.8%). Empiric antibiotic therapy was appropriate for 85 (77.3%) patients. Mortality rates for patients with inappropriate empiric therapy was 36.8%, whereas mortality rates for patients with appropriate empiric therapy was 17.5%. Among the patients with appropriate empiric antibiotic administration, de-escalation was associated with lower mortality rates of 5.0% for severe sepsis and 9.7% for septic shock patients. The mortality rates for the no de-escalation group were 19.0% and 35.7%, respectively. CONCLUSION: Empiric antibiotic therapy was acceptable for severe sepsis and septic shock patients treated in the ICU. The appropriate selection of empiric antibiotics was related to a greater rate of de-escalation and better survival. The risk of multi-drug-resistant bacterial infections was not as high as expected, but will need further attention in the future.