BackgroundAlthough many risk models have been tested in patients who undergo extracorporeal membrane oxygenation, few have been assessed for patients who received veno-arterial extracorporeal membrane oxygenation (VA-ECMO) support in the emergency department (ED). This study aimed to successfully predict outcomes of patients with cardiac or noncardiac failure who received VA-ECMO in the ED within 24 hours of arrival at the ED.MethodThis retrospective, observational cohort study included 154 patients, who were classified as cardiac (n = 127) and noncardiac (n = 27) patients and received VA-ECMO within 24 hours after arrival at the China Medical University Hospital ED in Taiwan between January 2009 and September 2014. We recorded mechanical ventilation settings, arterial blood gases, laboratory parameters including plasma lactate level, requirement of catecholamines, and risk scores at time of ECMO initiation. ECMO and mechanical ventilation support duration, length of stay in the hospital, and 90-day mortality data were also examined.ResultsThe overall mortality rate was 64.9 %. We used “survival after veno-arterial ECMO (SAVE)” scores to assess survival prediction in survival and nonsurvival groups, which was statistically different (–3.2 vs. –8.3, p <0.001). According to multivariate Cox proportional regression of survival, lactate (hazard ratio [HR] = 1.01, 95 % confidence interval [CI], 1.01–1.01, p <0.001) and SAVE score (HR = 0.92, [95 % CI, 0.88–0.96], p = 0.001) were independent predictors of outcome. Excellent discrimination (area under curve (AUC) = 0.843) was observed when lactate and SAVE score were combined, which we referred to as “the modified SAVE score.”ConclusionsModified SAVE scores improved outcome prediction for patients who underwent urgent VA-ECMO in the ED.

ObjectivesCurrent mortality prediction models used in the intensive care unit (ICU) have a limited role for specific diseases such as influenza, and we aimed to establish an explainable machine learning (ML) model for predicting mortality in critically ill influenza patients using a real-world severe influenza data set.Study designA cross-sectional retrospective multicentre study in TaiwanSettingEight medical centres in Taiwan.ParticipantsA total of 336 patients requiring ICU-admission for virology-proven influenza at eight hospitals during an influenza epidemic between October 2015 and March 2016.Primary and secondary outcome measuresWe employed extreme gradient boosting (XGBoost) to establish the prediction model, compared the performance with logistic regression (LR) and random forest (RF), demonstrated the feature importance categorised by clinical domains, and used SHapley Additive exPlanations (SHAP) for visualised interpretation.ResultsThe data set contained 76 features of the 336 patients with severe influenza. The severity was apparently high, as shown by the high Acute Physiology and Chronic Health Evaluation II score (22, 17 to 29) and pneumonia severity index score (118, 88 to 151). XGBoost model (area under the curve (AUC): 0.842; 95% CI 0.749 to 0.928) outperformed RF (AUC: 0.809; 95% CI 0.629 to 0.891) and LR (AUC: 0.701; 95% CI 0.573 to 0.825) for predicting 30-day mortality. To give clinicians an intuitive understanding of feature exploitation, we stratified features by the clinical domain. The cumulative feature importance in the fluid balance domain, ventilation domain, laboratory data domain, demographic and symptom domain, management domain and severity score domain was 0.253, 0.113, 0.177, 0.140, 0.152 and 0.165, respectively. We further used SHAP plots to illustrate associations between features and 30-day mortality in critically ill influenza patients.ConclusionsWe used a real-world data set and applied an ML approach, mainly XGBoost, to establish a practical and explainable mortality prediction model in critically ill influenza patients.

No abstract

No abstract

The ultrasound-guided pigtail catheter inserted by intensivists is a well-tolerated and effective method of draining all kinds of pleural effusions in critically ill patients. We suggest that pigtail catheter drainage be considered as the initial treatment of choice in the ICU.

The saturability of dFdCTP accumulation in Arm A suggests optimal delivery of gemcitabine is achieved using fixed rate infusion compared to 30-min infusion. Fixed dose rate gemcitabine is active and feasible, supporting the concept of fixed dosing rate of gemcitabine in advanced NSCLC. However, this entails a longer infusion time with associated higher costs involved.

BackgroundFluid balance is a fundamental management of patients with sepsis, and this study aimed to investigate the impact of cumulative fluid balance on critically ill patients with influenza admitted to an intensive care unit (ICU).MethodsThis multicenter retrospective cohort study was conducted by the Taiwan Severe Influenza Research Consortium (TSIRC) which includes eight medical centers. Patients with virology-proven influenza infection admitted to ICUs between October 2015 and March 2016 were included for analysis.ResultsA total of 296 patients were enrolled (mean age: 61.4±15.6 years; 62.8% men), and 92.2% (273/296) of them required mechanical ventilation. In the survivors, the daily fluid balance was positive from day 1 to day 3, and then gradually became negative from day 4 to day 7, whereas daily fluid balance was continuously positive in the non-survivors. Using the cumulative fluid balance from day 1–4 as a cut-off point, we found that a negative cumulative day 1–4 fluid balance was associated with a lower 30-day mortality rate (log-rank test, P = 0.003). To evaluate the impact of shock on this association, we divided the patients into shock and non-shock groups. The positive correlation between negative day 1–4 fluid balance and mortality was significant in the non-shock group (log-rank test, P = 0.008), but not in the shock group (log-rank test, P = 0.396). In a multivariate Cox proportional hazard regression model adjusted for age, sex, cerebrovascular disease, and PaO2/FiO2, day 1–4 fluid balance was independently associated with a higher 30-day mortality rate (aHR 1.088, 95% CI: 1.007–1.174).ConclusionsA negative day 1–4 cumulative fluid balance was associated with a lower mortality rate in critically ill patients with influenza. Our findings indicate the critical role of conservative fluid strategy in the management of patients with complicated influenza.

BackgroundPatients with influenza complicated with pneumonia are at high risk of rapid progression to acute respiratory distress syndrome (ARDS). Prone positioning with longer duration and lung-protective strategies might reduce the mortality level in ARDS. The aim of this study is to investigate the survival predictors of prone positioning in patients with ARDS caused by influenza pneumonia.MethodsThis retrospective study was conducted by eight tertiary referral centers in Taiwan. From January 1 to March 31 in 2016, all of the patients in intensive care units with virology-proven influenza pneumonia were collected, while all of those patients with ARDS and receiving prone positioning were enrolled. Demographic data, laboratory examinations, management records, ventilator settings and clinical outcomes were collected for analysis.ResultsDuring the study period, 336 patients with severe influenza pneumonia were screened and 263 patients met the diagnosis of ARDS. Totally, 65 patients receiving prone positioning were included for analysis. The 60-day survivors had lower Acute Physiology and Chronic Health Evaluation (APACHE) II score, pneumonia severity index (PSI), creatinine level and lower rate of receiving renal replacement therapy than non-survivors (22.4 ± 8.5 vs. 29.2 ± 7.4, p = 0.003; 106.6 ± 40.9 vs. 135.3 ± 48.6, p = 0.019; 1.2 ± 0.9 mg/dL vs. 3.1 ± 3.6 mg/dL, p = 0.040; and 4% vs. 42%, p < 0.005). Multivariate Cox regression analysis identified PSI (hazard ratio 1.020, 95% confidence interval 1.009–1.032; p < 0.001), renal replacement therapy (hazard ratio 6.248, 95% confidence interval 2.245–17.389; p < 0.001), and increase in dynamic driving pressure (hazard ratio 1.372, 95% confidence interval 1.095–1.718; p = 0.006) which were independent predictors associated with 60-day mortality.ConclusionsIn the present study, in evaluating the effect of prone positioning in patients with influenza pneumonia-related ARDS, pneumonia severity index, renal replacement therapy and increase in dynamic driving pressure were associated with 60-day mortality in patients with influenza pneumonia-related ARDS receiving prone positioning.Electronic supplementary materialThe online version of this article (10.1186/s13613-018-0440-4) contains supplementary material, which is available to authorized users.