Prehospital triage of acute aortic syndrome using a machine learning algorithm.

BACKGROUND: Acute aortic syndrome (AAS) comprises a complex and potentially fatal group of conditions requiring emergency specialist management. The aim of this study was to build a prediction algorithm to assist prehospital triage of AAS. METHODS: Details of consecutive patients enrolled in a regional specialist aortic network were collected prospectively. Two prediction algorithms for AAS based on logistic regression and an ensemble machine learning method called SuperLearner (SL) were developed. Undertriage was defined as the proportion of patients with AAS not transported to the specialist aortic centre, and overtriage as the proportion of patients with alternative diagnoses but transported to the specialist aortic centre. RESULTS: Data for 976 hospital admissions between February 2010 and June 2017 were included; 609 (62·4 per cent) had AAS. Overtriage and undertriage rates were 52·3 and 16·1 per cent respectively. The population was divided into a training cohort (743 patients) and a validation cohort (233). The area under the receiver operating characteristic (ROC) curve values for the logistic regression score and the SL were 0·68 (95 per cent c.i. 0·64 to 0·72) and 0·87 (0·84 to 0·89) respectively (P < 0·001) in the training cohort, and 0·67 (0·60 to 0·74) and 0·73 (0·66 to 0·79) in the validation cohort (P = 0·038). The logistic regression score was associated with undertriage and overtriage rates of 33·7 (bootstrapped 95 per cent c.i. 29·3 to 38·3) and 7·2 (4·8 to 9·8) per cent respectively, whereas the SL yielded undertriage and overtriage rates of 1·0 (0·3 to 2·0) and 30·2 (25·8 to 34·8) per cent respectively. CONCLUSION: A machine learning prediction model performed well in discriminating AAS and could be clinically useful in prehospital triage of patients with suspected AAS.

ANTECEDENTES: Los síndromes aórticos agudos (aortic acute syndromes, AAS) constituyen un grupo complejo y potencialmente letal de entidades que requieren un tratamiento especializado en emergencias. El objetivo de este estudio fue construir un algoritmo de predicción para ayudar a la selección prehospitalaria de los AAS. MÉTODOS: Se recogieron prospectivamente una serie de pacientes consecutivos inscritos en una red regional especializada en patología aórtica. Se desarrollaron dos algoritmos de predicción para AAS basados en una regresión logística y en un método de aprendizaje automático denominado Super Learner (SL). Undertriage (infra-selección) se definió como la proporción de pacientes con AAS no transportados al centro especializado en patología aórtica y el overtriage (sobre-selección) como la proporción de pacientes con diagnósticos alternativos al AAS pero transportados al centro especializado en patología aórtica. RESULTADOS: Se incluyeron los datos de 976 ingresos hospitalarios entre febrero de 2010 y junio de 2017, con 609 (62,4%) AAS. Las tasas de overtriage y undertriage fueron del 52,3% y del 16,1%, respectivamente. La población se dividió en una cohorte de entrenamiento (n = 743) y en una cohorte de validación (n = 233). El área bajo la curva ROC para la puntuación de regresión logística y el SL fueron de 0,68 (0,64, 0,72) y de 0,87 (0,84, 0,89), respectivamente (P < 0,001) en la cohorte de entrenamiento, y de 0,67 (0,60, 0,74) y de 0,73 (0,66, 0,79) en la cohorte de validación (P = 0,038). La puntuación de regresión logística se asoció con tasas de undertriage y overtriage de 33,7% (i.c. del 95% bootstrapped 29,3%, 38,3%) y de 7,2% (4,8%, 9,8%), respectivamente, mientras que el SL presentó tasas de undertriage y overtriage de 1,0% (0,3%, 2,0%) y de 30,2% (25,8%, 34,8%), respectivamente. CONCLUSIÓN: El modelo de predicción de aprendizaje automático funcionó bien para discriminar AAS y podría ser clínicamente útil en la selección prehospitalaria de pacientes con sospecha de síndrome aórtico agudo.

Major traumatic complications after out-of-hospital cardiac arrest: Insights from the Parisian registry.

AIM: Due to collapse and cardiopulmonary resuscitation (CPR) maneuvers, major traumatic injuries may complicate the course of resuscitation for out-of-hospital cardiac arrest patients (OHCA). Our goals were to assess the prevalence of these injuries, to describe their characteristics and to identify predictive factors. METHODS: We conducted an observational study over a 9-year period (2007-2015) in a French cardiac arrest (CA) center. All non-traumatic OHCA patients admitted alive in the ICU were studied. Major injuries identified were ranked using a functional two-level scale of severity (life-threatening or consequential) and were classified as CPR-related injuries or collapse-related injuries, depending of the predominant mechanism. Factors associated with occurrence of a CPR-related injury and ICU survival were identified using multivariable logistic regression. RESULTS: A major traumatic injury following OHCA was observed in 91/1310 patients (6.9%, 95%CI: 5.6, 8.3%), and was classified as a life-threatening injury in 36% of cases. The traumatic injury was considered as contributing to the death in 19 (21%) cases. Injuries were related to CPR maneuvers in 65 patients (5.0%, (95%CI: 3.8, 6.1%)). In multivariable analysis, age [OR 1.02; 95%CI (1.00, 1.04); p = 0.01], male gender [OR 0.53; 95%CI (0.31, 0.91); p = 0.02] and CA occurring at home [OR 0.54; 95%CI (0.31, 0.92); p = 0.02] were significantly associated with the occurrence of a CPR-related injury. CPR-related injuries were not associated with the ICU survival [OR 0.69; 95%CI (0.36, 1.33); p = 0.27]. CONCLUSIONS: Major traumatic injuries are common after cardiopulmonary resuscitation. Further studies are necessary to evaluate the interest of a systematic traumatic check-up in resuscitated OHCA patients in order to detect these injuries.

Propensity score estimators for the average treatment effect and the average treatment effect on the treated may yield very different estimates.

OBJECTIVE: Propensity score matching is typically used to estimate the average treatment effect for the treated while inverse probability of treatment weighting aims at estimating the population average treatment effect. We illustrate how different estimands can result in very different conclusions. STUDY DESIGN: We applied the two propensity score methods to assess the effect of continuous positive airway pressure on mortality in patients hospitalized for acute heart failure. We used Monte Carlo simulations to investigate the important differences in the two estimates. RESULTS: Continuous positive airway pressure application increased hospital mortality overall, but no continuous positive airway pressure effect was found on the treated. Potential reasons were (1) violation of the positivity assumption; (2) treatment effect was not uniform across the distribution of the propensity score. From simulations, we concluded that positivity bias was of limited magnitude and did not explain the large differences in the point estimates. However, when treatment effect varies according to the propensity score (E[Y(1)-Y(0)|g(X)] is not constant, Y being the outcome and g(X) the propensity score), propensity score matching ATT estimate could strongly differ from the inverse probability of treatment weighting-average treatment effect estimate. We show that this empirical result is supported by theory. CONCLUSION: Although both approaches are recommended as valid methods for causal inference, propensity score-matching for ATT and inverse probability of treatment weighting for average treatment effect yield substantially different estimates of treatment effect. The choice of the estimand should drive the choice of the method.

Extracorporeal life support (ECLS) for refractory cardiac arrest after drowning: an 11-year experience.

AIM: Neuroprotective effects of hypothermia may explain surprisingly high survival rates reported after drowning in cold water despite prolonged submersion. We described a cohort of refractory hypothermic cardiac arrests (CA) due to drowning treated by extracorporeal life support (ECLS) and aimed to identify criteria associated with 24-h survival. METHODS: Eleven-year period (2002-2012) retrospective study in the surgical intensive care unit (ICU) of a tertiary hospital (European Hospital Georges Pompidou, Paris, France). All consecutive hypothermic patients admitted for refractory CA after drowning in the Seine River were included. Patients with core temperature below 30°C and submersion duration of less than 1h were potentially eligible for ECLS resuscitation. RESULTS: Forty-three patients were admitted directly to the ICU during the study period. ECLS was initiated in 20 patients (47%). Among these 20 patients, only four (9%) survived more than 24h. A first hospital core temperature ≤26°C and a potassium serum level between 4.2 and 6mM at hospital admission have a sensitivity of 100% [95%CI: 28-100%] and a specificity of 100% [95%CI: 71-100%] to discriminate patients who survived more than 24h. Overall survival at ICU discharge and at 6-months was 5% [95%CI: 1-16%] (two patients). CONCLUSIONS: Despite patient hypothermia and aggressive resuscitation with ECLS, the observed survival rate is low in the present cohort. Like existing algorithms for ECLS management in avalanche victims, we recommend to use first core temperature and potassium serum level to indicate ECLS for refractory CA due to drowning.

[Usefulness of Bispectral Index (BIS) monitoring for early detection of cerebral hypoperfusions]. / Intérêt du monitorage de l'index bispectral (BIS) pour la détection précoce des hypoperfusions cérébrales.

OBJECTIVES: The goal of the study was to assess whether clinically significant cerebral hypoperfusion in awake patients would be associated with some alterations in the values of the bispectral index (BIS) monitoring. STUDY DESIGN: Observational study. POPULATION AND METHODS: We monitored the BIS during endovascular carotid artery occlusion testing in awake patients. RESULTS: Twenty-eight patients were included. Twenty-one adequately tolerated the procedure. Their BIS value remained stable throughout the procedure. Four patients had poor angiographic tolerance, but no clinical symptoms. Their BIS value slightly decreased during the test (minimal BIS: 83 [79-87]). Three patients had poor clinical and angiographic tolerance of the occlusion. They all experienced an immediate and dramatic decrease in their BIS value (minimal BIS: ipsilateral to clamping: 50 [45-60]; contralateral to clamping: 48 [45-52]). In all patients, the clinical symptoms and the BIS normalized after deflating the occlusion balloon. CONCLUSION: In awake patients, the observed values of the BIS monitoring seem to be associated with clinically relevant cerebral hypoperfusion.

[Out-of-hospital assessment of elderly patients' preference for ICU care]. / Évaluation préhospitalière du désir d'accès aux soins de réanimation chez les sujets âgés.

OBJECTIVE: To estimate the adequacy between elderly patients' preference for ICU care when treated for a life-threatening pathology, and the strategy proposed by the medical team on scene. STUDY DESIGN: Prospective, observational study. PATIENTS AND METHODS: All patients older than 80 treated out-of-hospital for a life threatening pathology were included, except in case of language barrier, or when patients were unable to answer and absence of next-of-kin. The results of the questionnaire on quality of life and patients' preference concerning ICU care were compared to the responses provided blindly by the medical team. RESULTS: Fifty-five patients were included. Quality of life as expressed by the patients was 7 (5-10) and by the physician 7 (6-8) (P=0.69). Thirty-six patients (65%) expressed the wish to be resuscitated, while ICU admission would have been proposed for 44 patients (80%) by the doctors (P=0.01). Among the 14 patients reluctant to ICU admission, 11 would have been proposed for ICU admission. In multivariate analysis, age (OR: 1.55 [1.04-2.32], P=0.03) and history of neurological pathology (OR: 11,91 [5.68->100], P=0.04) were associated with such an inadequacy. CONCLUSION: The inadequacy between elderly patients' preferences and doctors' opinion concerning ICU cares is frequent. The present results support a more systematic collection of patients' preferences when treated on scene for a life-threatening pathology.

[From iatrogenesis to medical errors: review of the literature and analytical approach]. / De l'iatrogénèse aux erreurs médicales : mise au point et approche analytique.

Iatrogenesis and medical errors have been increasingly studied over the past years. Because of the lack of consensus concerning the definitions, it remains difficult to draw general conclusions from the published. Moreover, it is still likely to be underestimated because of underreporting. This review aims at evaluating the overall incidence of iatrogenesis and medical errors in anaesthesia and intensive care and at discussing the strategies to prevent these incidents, at the individual or systemic level.

Neurosurgery and elderly: analysis through the years.

The aging of the population in westernized countries constitutes an important issue for the health systems struggling with limited resources and increasing costs. Morbidity and mortality rates reported for neurosurgical procedures in the elderly vary widely. The lack of data on risk benefit ratios may result in challenging clinical decisions in this expanding group of patients. The aim of this paper is to analyze the elderly patients cohort undergoing neurosurgical procedures and any trend variations over time. The medical records of elderly patients (defined as an individual of 70 years of age and over) admitted to the Neurosurgical and Neuro-ICU Departments of a major University Hospital in Paris over a 25-year period were retrospectively reviewed. The analysis included: (1) number of admissions, (2) percentage of surgically treated patients, (3) type of procedures performed, (4) length of hospital stay, and (5) mortality. The analysis showed a progressive and significant increase in the proportion of elderly presenting for neurosurgical elective and/or emergency procedures over the last 25 years. The number of procedures on patients over 70 years of age increased significantly whereas the mortality dropped. Though the length of hospital stay was reduced, it remained significantly higher than the average stay. The types of procedures also changed over time with more craniotomies and endovascular procedures being performed. Age should not be considered as a contraindication for complex procedures in neurosurgery. However, downstream structures for postoperative elderly patients must be further developed to reduce the mean hospital stay in neurosurgical departments because this trend is likely to continue to grow.

A randomized study of out-of-hospital continuous positive airway pressure for acute cardiogenic pulmonary oedema: physiological and clinical effects.

AIMS: In acute cardiogenic pulmonary oedema (ACPE), continuous positive airway pressure (CPAP) added to medical treatment improves outcome. The present study was designed to assess the benefit of CPAP as a first line treatment of ACPE in the out-of-hospital environment. METHODS AND RESULTS: The protocol lasted 45 min, divided into three periods of 15 min. Patients with ACPE were randomly assigned in two groups: 1/Early CPAP (n = 63): CPAP alone (T0-T15); CPAP + medical treatment (T15-T30); medical treatment alone (T30-T45) and 2/Late CPAP (n = 61): medical treatment alone (T0-T15); medical treatment + CPAP (T15-T30); medical treatment alone (T30-T45). Primary endpoint: effect of early CPAP on a dyspnoea clinical score and on arterial blood gases. Secondary endpoints: incidence of tracheal intubation, inotropic support, and in-hospital mortality. T0-T15: CPAP alone had a greater effect than medical treatment on the clinical score (P = 0.0003) and on PaO(2) (P = 0.0003). T15-T30: adding CPAP to medical treatment (late CPAP group) improved clinical score and blood gases and the two groups were no longer different at T30. T30-T45: in both groups, CPAP withdrawal worsened clinical score. Six patients in 'early CPAP' group vs. 16 in 'late CPAP' group were intubated [P = 0.01, odds-ratio: OR = 0.30 (0.09-0.89)]. Dobutamine was used only in the 'late CPAP' group (n = 5), (P = 0.02). Hospital death was higher in 'late CPAP' group (n = 8) than in 'early CPAP' group (n = 2) [P = 0.05, OR = 0.22 (0.04-1.0)]. CONCLUSION: When compared to usual medical care, immediate application of CPAP alone in out-of-hospital treatment of ACPO is significantly better improving physiological variables and symptoms and significantly reduces tracheal intubation incidence and in-hospital mortality.

Diastolic heart failure in anaesthesia and critical care.

Diastolic heart failure is an underestimated pathology with a high risk of acute decompensation during the perioperative period. This article reviews the epidemiology, risk factors, pathophysiology, and treatment of diastolic heart failure. Although frequently underestimated, diastolic heart failure is a common pathology. Diastolic heart failure involves heart failure with preserved left ventricular (LV) function, and LV diastolic dysfunction may account for acute heart failure occurring in critical care situations. Hypertensive crisis, sepsis, and myocardial ischaemia are frequently associated with acute diastolic heart failure. Symptomatic treatment focuses on the reduction in pulmonary congestion and the improvement in LV filling. Specific treatment is actually lacking, but encouraging data are emerging concerning the use of renin-angiotensin-aldosterone axis blockers, nitric oxide donors, or, very recently, new agents specifically targeting actin-myosin cross-bridges.

Feasibility of closed-loop titration of propofol guided by the Bispectral Index for general anaesthesia induction: a prospective randomized study.

BACKGROUND: This study was designed to evaluate the feasibility of propofol infusion by a closed-loop system for the titration of anaesthetic induction guided by Bispectral Index. METHODS: Forty patients were prospectively and randomly allocated into two groups: the target control infusion (TCI) group, where propofol titration was performed manually guided by the Bispectral Index using a commercial pharmacokinetic model (Diprifusor device) and the closed-loop group where titration was performed using a proportional differential algorithm. For both groups, the objective was to achieve a Bispectral Index of 50. Remifentanil TCI was infused at a target of 2 ng mL-1 and was maintained constant throughout the study. Feasibility of automatic induction was evaluated with performance error and haemodynamic data. RESULTS: Bispectral Index overshoot (-9 +/- 13% vs. -16 +/- 20%, P = 0.035) and mean duration of induction (381 +/- 106 s vs. 490 +/- 131 s, P = 0.004) were lower in the closed-loop group than in the TCI group. Haemodynamic data were similar between groups with a similar use of ephedrine bolus. CONCLUSION: The system was able to allow induction clinically for all patients. Automated titration guided by Bispectral Index for propofol infusion was feasible without increase in haemodynamic adverse effects.