Variable selection methods were poorly reported but rarely misused in major medical journals: Literature review.

Objective This work presents a review of the literature on reporting, practice and misuse of knowledge-based and data-driven variable selection methods, in five highly cited medical journals, considering recoding and interaction unlike previous reviews. Study Design and Setting Original observational studies with a predictive or explicative research question with multivariable analyses published in N. Engl. J. Med., Lancet, JAMA, Br. Med. J. and Ann. Intern. Med. between 2017 and 2019 were searched. Article screening was performed by a single reader, data extraction was performed by two readers and a third reader participated in case of disagreement. The use of data-driven variable selection methods in causal explicative questions was considered as misuse. Results 488 articles were included. The variable selection method was unclear in 234 (48%) articles, data-driven in 78 (16%) articles and knowledge-based in 176 (36%) articles. The most common data-driven methods were: Univariate selection (n = 22, 4.5%) and model comparisons or testing for interaction (n = 17, 3.5%). Data-driven methods were misused in 51 (10.5%) of articles. Conclusion Overall reporting of variable selection methods is insufficient. Data-driven methods seem to be used only in a minority of articles of the big five medical journals.

Association between hyperoxemia and mortality in patients treated by eCPR after out-of-hospital cardiac arrest.

OBJECTIVE: Assess whether elevated oxygen partial arterial pressure (PaO2) measured after the initiation of extra-corporeal cardiopulmonary resuscitation (eCPR), is associated with mortality in patients suffering from refractory out-of-hospital cardiac arrest (rOHCA). METHODS: Retrospective cohort study including rOHCA admitted to the ICU. Patients were divided into 3 groups, defined according to the PaO2 measured from arterial blood gas analysis 30 min after the initiation of eCPR. Hyperoxemia was defined as PaO2 ≥ 300 mmHg, hypoxemia as PaO2 ≤ 60 mmHg and normoxemia, as 60 < PaO2 < 300 mmHg. The main outcome was the mortality rate on day 28 after hospital admission. RESULTS: Sixty-six consecutive rOHCA, 77% male, with a mean age of 51 ±â€¯14 years, were admitted to the ICU. rOHCA were mainly due to acute coronary syndrome (67%), hypertrophic cardiomyopathy (8%) and cardiotoxic overdose (8%). Mortality at day 28 reached 61%. In the overall population, the mean PaO2 was 227 ±â€¯124 mmHg. An association between mortality and PaO2 was observed (OR = 1.01 [1.01-1.02]). The AUC for PaO2 after starting eCPR was 0.77 [0.65-0.89]. After adjustment for witnessed arrest, bystander's CPR, location, no-flow, low-flow, lactate and pH, age, and PaCO2, hyperoxemia had an ORa of 1.89 (CI95 [1.74-2.07]). CONCLUSION: We found an association between mortality and hyperoxemia in patients admitted to the ICU for rOHCA requiring eCPR. These data underline the potential toxicity of high dose of oxygen and suggest that controlled oxygen administration for these patients is crucial.

Epinephrine administration in non-shockable out-of-hospital cardiac arrest.

BACKGROUND: Epinephrine is recommended for the treatment of non-shockable out of hospital cardiac arrest (OHCA) to obtain return of spontaneous circulation (ROSC). Epinephrine efficiency and safety remain under debate. OBJECTIVE: We propose to describe the association between the cumulative dose of epinephrine and the failure of ROSC during the first 30 min of advanced life support (ALS). METHODOLOGY: A retrospective observational cohort study using the Paris SAMU 75 registry including all non-traumatic OHCA. All OHCA receiving epinephrine during the first 30 min of ALS were enrolled. Cumulative epinephrine dose given during ALS to ROSC was retrieved from medical reports. RESULTS: Among 1532 patients with OHCA, 776 (51%) had initial non-shockable rhythm. Fifty-four patients were excluded for missing data. The mean value of cumulative dose of epinephrine was 10 ±â€¯4 mg in patients who failed to achieve ROSC (ROSC-) and 4 ±â€¯3 mg (p = 0.04) for those who achieved ROSC. ROC curve analysis indicated a cut-off point of 7 mg total cumulative epinephrine associated with ROSC- (AUC = 0.89 [0.86-0.92]). Using propensity score analysis including age, sex and no-flow duration, association with ROSC- only remained significant for epinephrine > 7 mg (p ≤10-3, OR [CI95] = 1.53 [1.42-1.65]). CONCLUSION: An association between total cumulative epinephrine dose administered during OHCA resuscitation and ROSC- was reported with a threshold of 7 mg, best identifying patients with refractory OHCA. We suggest using this threshold in this context to guide the termination of ALS and early decide on the implementation of extracorporeal life support or organ harvesting in the first 30 min of ALS.

Predicting value of prehospital body temperature for ICU admission of septic shock patients.

To specify whether an association exists between pre-hospital body temperature collected by the emergency medical services (EMS) call centre, and intensive care unit (ICU) admission of patients with septic shock. An observational study based on data collected by the EMS of Paris. All septic shocks were included. Among, the 140 calls concerning septic shock, 22 patients (16%) were admitted to ICU. The mean core temperature was 37.4±1.6°C for ICU and 38.6±1.1°C (p<4.10^-5) for non-ICU patients. Using propensity score analysis, the relative risk for ICU admission of patients with pre-hospital fever or hypothermia was 0.31 and 2 respectively. The study highlights the potential usefulness of early temperature measurement in septic shock patients to allow early proper orientation.

Triage of Septic Patients Using qSOFA Criteria at the SAMU Regulation: A Retrospective Analysis.

PURPOSE: One of the major prognostic factors in the management of sepsis is the early initiation of appropriate treatment. To serve this purpose, early identification and triage of patients are crucial steps, which are still not optimal. The objective of this study was to determine whether the quick Sequential (Sepsis-related) Organ Failure Assessment (qSOFA) score is an accurate method for prehospital triaging of septic patients. We evaluated whether the use of qSOFA criteria collected by the Service Mobile d'Urgence et de Réanimation 15 (SAMU 15) regulation call center during prehospital care would facilitate appropriate intensive care unit (ICU) admission of patients with septic syndromes. METHODS: We conducted a retrospective observational register-based study using data collected between April 01 and May 31 2011. These data are based on call registry reports of calls received by the Paris Emergency regulation call centre during prehospital management of patients. All patients with suspected infection were included in the study and evaluated using qSOFA and systemic inflammatory response syndrome (SIRS) criteria. The primary outcome was Intensive Care Unit (ICU) admission. RESULTS: Among the 30 642 reports received, 141 patients with presumed sepsis were included. Twenty-two patients were admitted to an ICU. The qSOFA and SIRS scores were the same in predicting admission to an ICU (p = 0.26). The qSOFA had a sensitivity of 75% and a specificity of 68% for ICU admission whereas the SIRS had a sensitivity of 87% and a specificity of 43%. At day 28, 12 patients (9%) had died, 5 of them in the ICU. The negative predictive value reached 93% and 94% for pre-hospital qSOFA and SIRS respectively. Pre-hospital systolic blood pressure (SBP) ≤100 mmHg was significantly associated with ICU admission (OR = 4.19 [1.89-9.84]), while all other criteria were not. CONCLUSION: The current study reports no difference between the SIRS and the qSOFA scores for prehospital triage of septic patients to predict ICU admission. Both scores have comparable, pertinent, negative predictive value for ICU admission. Nevertheless, an improved score for pre-hospital triaging is needed to predict ICU admission of septic patients.

Prehospital triage of septic patients at the SAMU regulation: Comparison of qSOFA, MRST, MEWS and PRESEP scores.

PURPOSE: A couple of scoring systems have been developed for risk stratification of septic patients. Their performance in the management of out-of-hospital initial care delivery is not documented. This study try to evaluate the predictive ability of Quick Sequential Organ Failure Assessment (qSOFA), Robson Screening Tool (RST), Modified Early Warning Score (MEWS) and Prehospital Early Sepsis Detection (PRESEP) scores on out of-hospital triage of septic patients, to predict intensive care unit (ICU) admission. METHODS: A retrospective study using call records received by the SAMU 15 regulation call centre including all patients with presumed septic shock was performed. The primary outcome was the admission to the ICU. RESULTS: Among the 47 000 reports received, 37 patients with presumed septic shock were included. Twenty-two patients (59%) were admitted to ICU. AUCs of qSOFA, RST, MEWS and PRESEP scores were respectively 0.40 [0.22-0.59], 0.60 [0.43-0.78], 0.66 [0.47-0.85] and 0.67 [0.51-0.84]. RST outperformed PRESEP, MEWS and qSOFA for sensitivity (1, 0.92, 0.85 and 0.62 respectively). MEWS showed better specificity than PRESEP, MRST and qSOFA (0.33, 0.29, 0.16 and 0.16). MEWS showed comparable positive predictive value than PRESEP and outperformed MRST and qSOFA (0.41, 0.41, 0.39 and 0.29 respectively). Negative predictive value of MRST outperformed PRESEP, MEWS and qSOFA (1, 0.88, 0.80 and 0.44 respectively). CONCLUSION: Our findings suggest that screening patients at SAMU 15 regulation call centre using qSOFA, MRST, MEWS and PRESEP scores to predict ICU admission is irrelevant. Development of a specific scoring system for out-of-hospital triage of septic patients is needed.

Toxicodynetics in nordiazepam and oxazepam overdoses.

OBJECTIVES: Toxicodynetics aims at defining the time-course of major clinical events in drug overdose. We report the toxicodynetics in mono-intoxications with oxazepam and nordiazepam. METHODS: Cases of oxazepam or nordiazepam overdoses collected at the Paris poison control centre from 1999 to 2014 on the basis of self-report. A particular attention was paid to eliminate the concomitant alcohol or psychotropic co-ingestions. The toxicodynetic parameters were assessed as previously described. Results are expressed using 10-90 percentiles. In adults, the dose was normalized (TI, toxic Index) by dividing the supposed ingested dose by the maximal recommended dose. RESULTS: Two hundred and fifty-one and 74 cases of oxazepam and nordiazepam poisonings were included, respectively. The Emax for oxazepam and nordiazepam were sleepiness or obtundation in 106 and 36 cases, respectively. Coma was used to qualify only one oxazepam overdose. The median delay in onset of the Emax was 1.5h (0.33-15) in nordiazepam and 4h (0.5-15) in oxazepam overdose. In both overdoses, the onset of Emax occurred on an "on-off" mode. In adults, the greatest TIs in nordiazepam and oxazepam overdoses were 45 and 26.7, respectively. The TI in the oxazepam-induced coma was 26.7, the largest dose. CONCLUSION: Data collected in PCC allow determining a number of toxicodynetic parameters. Toxicodynetics showed that nordiazepam is not a cause of coma even in large overdose while oxazepam causes coma only at a very high dose. Deep coma in nordiazepam overdose whatever the dose and deep coma in overdose with oxazepam involving TI less than 20 result from unrecognized drug-drug interaction.

Base excess and lactate as prognostic indicators for patients treated by extra corporeal life support after out hospital cardiac arrest due to acute coronary syndrome.

OBJECTIVE: To examine whether values of arterial base excess or lactate taken 3 h after starting ECLS indicate poor prognosis and if this can be used as a screening tool to follow Extra Corporeal Life Support after Out Hospital Cardiac Arrest due to acute coronary syndrome. DESIGN: Single Centre retrospective observational study. SETTING: University teaching hospital general adult intensive care unit. PATIENTS: 15 consecutive patients admitted to the intensive care unit after refractory Out Hospital Cardiac Arrest due to acute coronary syndrome treated by Extra Corporeal Life Support. INTERVENTIONS: Arterial base excess and lactate concentrations were measured immediately after starting ECLS and every 3 h after. RESULTS: Both base excess and arterial lactate measured 3 h after starting ECLS effectively predict multi-organ failure occurrence and mortality in the following 21 h (area under the curve on receiver operating characteristic analysis of 0.97, 0.95 respectively). The best predictive values were obtained with a base excess level measured 3 h after starting ECLS of less than -10 mmol/l and lactate concentrations greater than 12 mmol/l. The combination of these two markers measured 3 h after starting ECLS predicted multiorgan failure occurrence and mortality in the following 21 h with a sensitivity of 70% and a specificity of 100%. CONCLUSIONS: Combination of base excess and lactate, measured 3 h after starting ECLS, can be used to predict multiorgan failure occurrence and mortality in the following 21 h in patients admitted to an intensive care unit for refractory Out Hospital Cardiac Arrest due to acute coronary syndrome treated by Extra Corporeal Life Support. These parameters can be obtained simply and rapidly and help in the decision process to continue ECLS for refractory CA.

[Family's and/or relatives expectations concerning a patient emergently hospitalized into an intensive care unit after out-of-hospital medical care of by a samu team]. / Enquête sur les attentes de la famille et/ou des proches d'un patient hospitalisé en réanimation après prise en charge extrahospitalière par une équipe du samu.

INTRODUCTION: Emergency admission of a patient into an intensive care unit (ICU) is a source of stress for family and/or relatives. Expectations of family and/or relatives are important endpoints for the medical and paramedical team in charge of the patient, to better answer to their questions. OBJECTIVE: The aim of this study was to determine family's and/or relatives expectations concerning a patient emergently hospitalized into an ICU after out-of-hospital medical care of by a samu team. MATERIALS AND METHOD: We performed a survey using a questionnaire sent by email to 500 randomly chosen individuals from the French population. RESULTS: We received 220 responses (44 %). Family and/or relatives expectations are different about short, medium and long terms. Elements perceived as the most important in the short term are severity, diagnosis and prognosis, whereas in the medium and long terms they are hospital-stay and potential sequels. Medical language used is considered as too much complex for more than half of respondents. In case of foreseeable unfavourable outcome, 90 % of respondents would like to receive immediate information using a simple and brief language. At last, nearly two thirds of respondents had been really exposed to such a situation before, and assessed the quality of information received as moderate, with a score of 5/10. CONCLUSION: Information of families and/or relatives of a patient hospitalized in ICU is essential. Their expectations concern short, medium and long terms. At last, most of them prefers that information would be delivered by a physician, and using a simple and brief language.

[Haemorrhagic shock after severe abdominal trauma in children: knowing when to change from conservative to surgical management]. / Choc hémorragique après traumatisme abdominal grave chez l'enfant : savoir opter pour l'attitude interventionnelle chirurgicale.

Abdominal vascular injuries following a serious falling out are quite rare in children. They can lead to haemorrhagic shock whose etiological diagnosis may be difficult in children in case of multiple trauma. The current management of abdominal injuries in the child is usually conservative, surgery being indicated in haemodynamically unstable patients. We report the case of a 7-year-old girl who presented with abdominal trauma with rupture of the hepatic artery and shredding of the splenic vein following a falling out of 10 meters. Aggressive resuscitation associated with early laparotomy for haemostasis, contrary to usual practices advocated in such a context, have helped control the hemorrhagic shock and stabilize the haemodynamic status of the child. The subsequent evolution was favourable, with full recovery. While a conservative attitude usually prevails in the management of traumatic intra abdominal bleeding in children an interventional attitude with emergency surgery must be sometimes considered.