Machine learning identifies ICU outcome predictors in a multicenter COVID-19 cohort.

BACKGROUND: Intensive Care Resources are heavily utilized during the COVID-19 pandemic. However, risk stratification and prediction of SARS-CoV-2 patient clinical outcomes upon ICU admission remain inadequate. This study aimed to develop a machine learning model, based on retrospective & prospective clinical data, to stratify patient risk and predict ICU survival and outcomes. METHODS: A Germany-wide electronic registry was established to pseudonymously collect admission, therapeutic and discharge information of SARS-CoV-2 ICU patients retrospectively and prospectively. Machine learning approaches were evaluated for the accuracy and interpretability of predictions. The Explainable Boosting Machine approach was selected as the most suitable method. Individual, non-linear shape functions for predictive parameters and parameter interactions are reported. RESULTS: 1039 patients were included in the Explainable Boosting Machine model, 596 patients retrospectively collected, and 443 patients prospectively collected. The model for prediction of general ICU outcome was shown to be more reliable to predict "survival". Age, inflammatory and thrombotic activity, and severity of ARDS at ICU admission were shown to be predictive of ICU survival. Patients' age, pulmonary dysfunction and transfer from an external institution were predictors for ECMO therapy. The interaction of patient age with D-dimer levels on admission and creatinine levels with SOFA score without GCS were predictors for renal replacement therapy. CONCLUSIONS: Using Explainable Boosting Machine analysis, we confirmed and weighed previously reported and identified novel predictors for outcome in critically ill COVID-19 patients. Using this strategy, predictive modeling of COVID-19 ICU patient outcomes can be performed overcoming the limitations of linear regression models. Trial registration "ClinicalTrials" (clinicaltrials.gov) under NCT04455451.

Utility of nociceptive flexion reflex threshold and bispectral index to predict movement responses under propofol anaesthesia.

INTRODUCTION: The nociceptive flexion reflex threshold (NFRT) is a promising tool to monitor analgesia during general anaesthesia. Clinical studies have shown that the NFRT allows to predict movement responses to painful stimuli under a combined anaesthetic regime of sedative and opioid agents. Experimental studies indicated that the NFRT is also able to predict such movement responses under an exclusively sedative regime like propofol mono-anaesthesia. Therefore, we performed this study to investigate the ability of the NFRT to predict movement responses to painful stimuli in patients during a clinical propofol mono-anaesthesia. METHODS: We investigated 140 cardiac surgery patients during their postoperative phase under propofol mono-anaesthesia. NFRT and bispectral index (BIS) were determined in each patient right before endotracheal suctioning or painful electrical test stimulation. Prediction probabilities were calculated to quantify how accurate each measure is able to predict movement responses to the stimuli. RESULTS: The 124 patients included in the analysis received a median propofol dosage of 3.2 (2.5-3.9) [median (IQR)] mg/kg/h. The included patients showed 287 movement responses after a total of 725 investigated stimuli. The prediction probabilities for positive movement responses were 0.63 (95%CI: 0.59-0.67) for the NFRT and 0.69 (95%CI: 0.65-0.73) for the BIS. CONCLUSIONS: The NFRT allows the prediction of movement responses under propofol mono-anaesthesia, which confirms its utility as a monitor to predict movement responses under general anaesthesia. The BIS allows an even more accurate prediction, although it does not reflect the physiological structures of movement suppression, but correlates closely with the dose of propofol. TRIAL REGISTRATION: German clinical trial register (DRKS00003062, Deutsches Register Klinischer Studien).

[Neuromonitoring and awareness]. / Awareness - Stellenwert des Neuromonitorings von Analgesie und Hypnose.

Awareness is a rare but in some cases severe complication of general anesthesia. In clinical practice vegetative signs such as sweating, lacrimation, tachycardia and hypertension are used to assess the anesthetic depth. Awareness however may also occur without any of these signs. Different systems based on the electroencephalogram (EEG) have been developed to monitor hypnosis. Some studies have shown that the use of EEG based monitor systems can reduce the incidence of awareness in patients of high risk. A similar reduction of the incidence was reached in studies using protocols based on concentrations of volatile anesthetics. Other systems to monitor anesthetic depth have been developed, but their influence on awareness has not yet been investigated sufficiently. Further studies are needed to clarify their impact concerning the incidence of awareness.

A missense variant in desmoglein-2 predisposes to dilated cardiomyopathy.

Familial Dilated Cardiomyopathy (FDCM) is caused by mutations in genes encoding myocardial force transduction proteins. Desmoglein-2 (DSG2) and Desmocollin-2 (DSC2) provide cellular adhesion and force transduction by cell-to-cell anchorage. To test whether perturbations of DSG2 or DSC2 exhibit a pathogenic impact on DCM pathogenesis, we sequenced both genes in 73 patients with FDCM and assessed prevalence of missense variations in matched control cohorts. We detected two missense variations in DSG2 (V55M and V919G) which were absent in 360 control alleles. Surprisingly, both variants were previously reported in patients with arrhythmogenic right ventricular cardiomyopathy. Yet, in the present study only the DSG2-V55M variant showed segregation with DCM in a family pedigree. Subsequent, analysis of 538 patients with idiopathic DCM and 617 consecutive control individuals resulted in identification of thirteen DSG2-V55M carriers with DCM, whereas only three control subjects harbored the variant. DSG2 immunostaining revealed pale structures of the intercalated disc in myocardium of one unique homozygous DSG2-V55M carrier. Furthermore, myocardial desmosomal structures were significantly shortened when compared to DCM myocardium negative for DSG2-V55M. Thus, our study identified the DSG2-V55M polymorphism as a novel risk variant for DCM associated with shortened desmosomes of the cardiac intercalated disc.