At-Risk Drinking Is Independently Associated With Acute Kidney Injury in Critically Ill Patients.

OBJECTIVES: Unhealthy use of alcohol and acute kidney injury are major public health problems, but little is known about the impact of excessive alcohol consumption on kidney function in critically ill patients. We aimed to determine whether at-risk drinking is independently associated with acute kidney injury in the ICU and at ICU discharge. DESIGN: Prospective observational cohort study. SETTING: A 21-bed polyvalent ICU in a university hospital. PATIENTS: A total of 1,107 adult patients admitted over a 30-month period who had an ICU stay of greater than or equal to 3 days and in whom alcohol consumption could be assessed. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We assessed Kidney Disease Improving Global Outcomes stages 2-3 acute kidney injury in 320 at-risk drinkers (29%) and 787 non-at-risk drinkers (71%) at admission to the ICU, within 4 days after admission and at ICU discharge. The proportion of patients with stages 2-3 acute kidney injury at admission to the ICU (42.5% vs 18%; p < 0.0001) was significantly higher in at-risk drinkers than in non-at-risk drinkers. Within 4 days and after adjustment on susceptible and predisposing factors for acute kidney injury was performed, at-risk drinking was significantly associated with acute kidney injury for the entire population (odds ratio, 2.15; 1.60-2.89; p < 0.0001) in the subgroup of 832 patients without stages 2-3 acute kidney injury at admission to the ICU (odds ratio, 1.44; 1.02-2.02; p = 0.04) and in the subgroup of 971 patients without known chronic kidney disease (odds ratio, 1.92; 1.41-2.61; p < 0.0001). Among survivors, 22% of at-risk drinkers and 9% of non-at-risk drinkers were discharged with stages 2-3 acute kidney injury (p < 0.001). CONCLUSIONS: Our results suggest that chronic and current alcohol misuse in critically ill patients is associated with kidney dysfunction. The systematic and accurate identification of patients with alcohol misuse may allow for the prevention of acute kidney injury.

Out-of-Hospital Cardiac Arrest Detection by Machine Learning Based on the Phonetic Characteristics of the Caller's Voice.

INTRODUCTION: Out-of-hospital cardiac arrest (OHCA) is a major public health issue. The prognosis is closely related to the time from collapse to return of spontaneous circulation. Resuscitation efforts are frequently initiated at the request of emergency call center professionals who are specifically trained to identify critical conditions over the phone. However, 25% of OHCAs are not recognized during the first call. Therefore, it would be interesting to develop automated computer systems to recognize OHCA on the phone. The aim of this study was to build and evaluate machine learning models for OHCA recognition based on the phonetic characteristics of the caller's voice. METHODS: All patients for whom a call was done to the emergency call center of Rennes, France, between 01/01/2017 and 01/01/2019 were eligible. The predicted variable was OHCA presence. Predicting variables were collected by computer-automatized phonetic analysis of the call. They were based on the following voice parameters: fundamental frequency, formants, intensity, jitter, shimmer, harmonic to noise ratio, number of voice breaks, and number of periods. Three models were generated using binary logistic regression, random forest, and neural network. The area under the curve (AUC) was the primary outcome used to evaluate each model performance. RESULTS: 820 patients were included in the study. The best model to predict OHCA was random forest (AUC=74.9, 95% CI=67.4-82.4). CONCLUSION: Machine learning models based on the acoustic characteristics of the caller's voice can recognize OHCA. The integration of the acoustic parameters identified in this study will help to design decision-making support systems to improve OHCA detection over the phone.

Machine learning is the key to diagnose COVID-19: a proof-of-concept study.

The reverse transcription-polymerase chain reaction (RT-PCR) assay is the accepted standard for coronavirus disease 2019 (COVID-19) diagnosis. As any test, RT-PCR provides false negative results that can be rectified by clinicians by confronting clinical, biological and imaging data. The combination of RT-PCR and chest-CT could improve diagnosis performance, but this would requires considerable resources for its rapid use in all patients with suspected COVID-19. The potential contribution of machine learning in this situation has not been fully evaluated. The objective of this study was to develop and evaluate machine learning models using routine clinical and laboratory data to improve the performance of RT-PCR and chest-CT for COVID-19 diagnosis among post-emergency hospitalized patients. All adults admitted to the ED for suspected COVID-19, and then hospitalized at Rennes academic hospital, France, between March 20, 2020 and May 5, 2020 were included in the study. Three model types were created: logistic regression, random forest, and neural network. Each model was trained to diagnose COVID-19 using different sets of variables. Area under the receiving operator characteristics curve (AUC) was the primary outcome to evaluate model's performances. 536 patients were included in the study: 106 in the COVID group, 430 in the NOT-COVID group. The AUC values of chest-CT and RT-PCR increased from 0.778 to 0.892 and from 0.852 to 0.930, respectively, with the contribution of machine learning. After generalization, machine learning models will allow increasing chest-CT and RT-PCR performances for COVID-19 diagnosis.

Monitoring Transcutaneously Measured Partial Pressure of CO2 During Intubation in Critically Ill Subjects.

BACKGROUND: The risk for severe hypoxemia during endotracheal intubation is a major concern in the ICU, but little attention has been paid to CO2 variability. The objective of this study was to assess transcutaneously measured partial pressure of CO2 ([Formula: see text]) throughout intubation in subjects in the ICU who received standard oxygen therapy, high-flow nasal cannula oxygen therapy, or noninvasive ventilation for preoxygenation. We hypothesized that the 3 methods differ in terms of ventilation and CO2 removal. METHODS: In this single-center, prospective, observational study, we recorded [Formula: see text] from preoxygenation to 3 h after the initiation of mechanical ventilation among subjects requiring endotracheal intubation. Subjects were sorted into 3 groups according to the preoxygenation method. We then assessed the link between [Formula: see text] variability and the development of postintubation hypotension. RESULTS: A total of 202 subjects were included in the study. The [Formula: see text] values recorded at endotracheal intubation, at the initiation of mechanical ventilation, and after 30 min and 1 h of mechanical ventilation were significantly higher than those recorded during preoxygenation (P < .05). [Formula: see text] variability differed significantly according to the preoxygenation method (P < .001, linear mixed model). A decrease in [Formula: see text] by > 5 mm Hg within 30 min after the start of mechanical ventilation was independently associated with postintubation hypotension (odds ratio = 2.14 [95% CI 1.03-4.44], P = .039) after adjustments for age, Simplified Acute Physiology Score II, COPD, cardiac comorbidity, the use of propofol for anesthetic induction, and minute ventilation at the start of mechanical ventilation. CONCLUSIONS: [Formula: see text] variability during intubation is significant and differs with the method of preoxygenation. A decrease in [Formula: see text] after the beginning of mechanical ventilation was associated with postintubation hypotension. (ClinicalTrials.gov registration NCT0388430.).

Doppler sonography of cerebral blood flow for early prognostication after out-of-hospital cardiac arrest: DOTAC study.

AIM: To assess the neurological prognosis of comatose survivors of cardiac arrest by early transcranial Doppler sonography (TCD). METHODS: This was a prospective study performed between May 2016 and October 2017 in a medical intensive care unit (ICU) and a cardiac ICU of a university teaching hospital. All patients older than 18 years who were successfully resuscitated from an out-of-hospital cardiac arrest (OHCA) with persistent coma after the return of spontaneous circulation (ROSC) were eligible. We excluded patients for whom OHCA was associated with traumatic brain injury, no possibility of TCD measurements, or who were dead before establishing the neurological prognosis. We measured the pulsatility index (PI) and diastolic flow velocity (DFV) of the right and left middle cerebral arteries within 12 h after ICU admission. The lowest DFV and highest PI values were used for the statistical analysis. The neurological outcome at hospital discharge was evaluated by the cerebral performance category. RESULTS: Forty-two patients were included in the final analysis: 15 had good and 27 poor neurological outcomes. The PI was higher in the poor outcome (1.49 vs. 1.12, p = 0.01) than good outcome group and the DFV was lower in the poor outcome group (17.3 cm s-1vs. 26.0 cm s-1; p = 0.01). CONCLUSION: Data provided by early TCD after ROSC are associated with neurological outcome. The use of TCD could help guide interventions to improve cerebral perfusion after ROSC in patients resuscitated from OHCA.