Search

Motivation Acute kidney injury (AKI) is a syndrome that affects a large fraction of all critically ill patients, and early diagnosis to receive adequate treatment is as imperative as it is challenging to make early. Consequently, machine learning approaches have been developed to predict AKI ahead of time. However, the prevalence of AKI is often underestimated in state-of-the-art approaches, as they rely on an AKI event annotation solely based on creatinine, ignoring urine output.   We construct and evaluate early warning systems for AKI in a multi-disciplinary ICU setting, using the complete KDIGO definition of AKI. We propose several variants of gradient-boosted decision tree (GBDT)-based models, including a novel time-stacking based approach. A state-of-the-art LSTM-based model previously proposed for AKI prediction is used as a comparison, which was not specifically evaluated in ICU settings yet. Results We find that optimal performance is achieved by using GBDT with the time-based stacking technique (AUPRC = 65.7%, compared with the LSTM-based model's AUPRC = 62.6%), which is motivated by the high relevance of time since ICU admission for this task. Both models show mildly reduced performance in the limited training data setting, perform fairly across different subcohorts, and exhibit no issues in gender transfer.   Following the official KDIGO definition substantially increases the number of annotated AKI events. In our study GBDTs outperform LSTM models for AKI prediction. Generally, we find that both model types are robust in a variety of challenging settings arising for ICU data. Availability and implementation The code to reproduce the findings of our manuscript can be found at: https://github.com/ratschlab/AKI-EWS [ABSTRACT FROM AUTHOR]

Purpose Mortality in circulatory shock is high. Enhanced resolution of shock may improve outcomes. We aim to determine whether adding hemodynamic monitoring with continual transesophageal echocardiography (hTEE) to usual care accelerates resolution of hemodynamic instability. Methods 550 patients with circulatory shock were randomly assigned to four groups stratified using hTEE (hTEE vs usual care) and assessment frequency (minimum every 4 h vs 8 h). Primary outcome was time to resolution of hemodynamic instability, analyzed as intention-to-treat (ITT) analysis at day 6 and in a predefined secondary analysis at days 3 and 28. Results Of 550 randomized patients, 271 with hTEE and 274 patients with usual care were eligible and included in the ITT analysis. Time to resolution of hemodynamic instability did not differ within the first 6 days [hTEE vs usual care adjusted sub-hazard ratio (SHR) 1.20, 95% confidence interval (CI) 0.98-1.46, p = 0.067]. Time to resolution of hemodynamic instability during the 72 h of hTEE monitoring was shorter in patients with TEE (hTEE vs usual care SHR 1.26, 95% CI 1.02-1.55, p = 0.034). Assessment frequency had no influence. Time to resolution of clinical signs of hypoperfusion, duration of organ support, length of stay and mortality in the intensive care unit and hospital, and mortality at 28 days did not differ between groups. Conclusions In critically ill patients with shock, hTEE monitoring or hemodynamic assessment frequency did not influence resolution of hemodynamic instability or mortality within the first 6 days. Trial registration and statistical analysis plan ClinicalTrials.gov Identifier: NCT02048566., Author(s): Tobias M. Merz [sup.1], Luca Cioccari [sup.1], Pascal M. Frey [sup.2], Andreas Bloch [sup.1], David Berger [sup.1], Bjoern Zante [sup.1], Stephan M. Jakob [sup.1], Jukka Takala [sup.1] Author Affiliations: [...]

Author(s): Philippe Vignon [sup.1] [sup.2] [sup.3], Tobias M. Merz [sup.4], Antoine Vieillard-Baron [sup.5] [sup.6] Author Affiliations: (1) 0000 0001 1481 5225, grid.412212.6, Medical-surgical Intensive Care Unit, University Hospital Dupuytren, , [...]

Purpose Rapid assessment and intervention is important for the prognosis of acutely ill patients admitted to the emergency department (ED). The aim of this study was to prospectively develop and validate a model predicting the risk of in-hospital death based on all available information available at the time of ED admission and to compare its discriminative performance with a non-systematic risk estimate by the triaging first health-care provider. Methods Prospective cohort analysis based on a multivariable logistic regression for the probability of death. Results A total of 8,607 consecutive admissions of 7,680 patients admitted to the ED of a tertiary care hospital were analysed. Most frequent APACHE II diagnostic categories at the time of admission were neurological (2,052, 24 %), trauma (1,522, 18 %), infection categories [1,328, 15 %; including sepsis (357, 4.1 %), severe sepsis (249, 2.9 %), septic shock (27, 0.3 %)], cardiovascular (1,022, 12 %), gastrointestinal (848, 10 %) and respiratory (449, 5 %). The predictors of the final model were age, prolonged capillary refill time, blood pressure, mechanical ventilation, oxygen saturation index, Glasgow coma score and APACHE II diagnostic category. The model showed good discriminative ability, with an area under the receiver operating characteristic curve of 0.92 and good internal validity. The model performed significantly better than non-systematic triaging of the patient. Conclusions The use of the prediction model can facilitate the identification of ED patients with higher mortality risk. The model performs better than a non-systematic assessment and may facilitate more rapid identification and commencement of treatment of patients at risk of an unfavourable outcome., Author(s): Michael Coslovsky [sup.1], Jukka Takala [sup.2], Aristomenis K. Exadaktylos [sup.3], Luca Martinolli [sup.3], Tobias M. Merz [sup.2] Author Affiliations: (1) grid.5734.5, 0000000107265157, Clinical Trials Unit, Institute of Social and [...]

Background: Septic shock is associated with decreased vasopressor responsiveness. Experimental data suggest that central alpha2-agonists like dexmedetomidine (DEX) increase vasopressor responsiveness and reduce catecholamine requirements in septic shock. However, DEX may also cause hypotension and bradycardia. Thus, it remains unclear whether DEX is hemodynamically safe or helpful in this setting.Methods: In this post hoc subgroup analysis of the Sedation Practice in Intensive Care Evaluation (SPICE III) trial, an international randomized trial comparing early sedation with dexmedetomidine to usual care in critically patients receiving mechanical ventilation, we studied patients with septic shock admitted to two tertiary ICUs in Australia and Switzerland. The primary outcome was vasopressor requirements in the first 48 h after randomization, expressed as noradrenaline equivalent dose (NEq [μg/kg/min] = noradrenaline + adrenaline + vasopressin/0.4).Results: Between November 2013 and February 2018, 417 patients were recruited into the SPICE III trial at both sites. Eighty-three patients with septic shock were included in this subgroup analysis. Of these, 44 (53%) received DEX and 39 (47%) usual care. Vasopressor requirements in the first 48 h were similar between the two groups. Median NEq dose was 0.03 [0.01, 0.07] μg/kg/min in the DEX group and 0.04 [0.01, 0.16] μg/kg/min in the usual care group (p = 0.17). However, patients in the DEX group had a lower NEq/MAP ratio, indicating lower vasopressor requirements to maintain the target MAP. Moreover, on adjusted multivariable analysis, higher dexmedetomidine dose was associated with a lower NEq/MAP ratio.Conclusions: In critically ill patients with septic shock, patients in the DEX group received similar vasopressor doses in the first 48 h compared to the usual care group. On multivariable adjusted analysis, dexmedetomidine appeared to be associated with lower vasopressor requirements to maintain the target MAP.Trial Registration: The SPICE III trial was registered at ClinicalTrials.gov ( NCT01728558 ). [ABSTRACT FROM AUTHOR]