Explore the vast range of titles available.

The benefits of cardiac surgery are sometimes difficult to predict and the decision to operate on a given individual is complex. Machine Learning and Decision Curve Analysis (DCA) are recent methods developed to create and evaluate prediction models. We conducted a retrospective cohort study using a prospective collected database from December 2005 to December 2012, from a cardiac surgical center at University Hospital. The different models of prediction of mortality in-hospital after elective cardiac surgery, including EuroSCORE II, a logistic regression model and a machine learning model, were compared by ROC and DCA. Of the 6,520 patients having elective cardiac surgery with cardiopulmonary bypass, 6.3% died. Mean age was 63.4 years old (standard deviation 14.4), and mean EuroSCORE II was 3.7 (4.8) %. The area under ROC curve (IC95%) for the machine learning model (0.795 (0.755-0.834)) was significantly higher than EuroSCORE II or the logistic regression model (respectively, 0.737 (0.691-0.783) and 0.742 (0.698-0.785), p < 0.0001). Decision Curve Analysis showed that the machine learning model, in this monocentric study, has a greater benefit whatever the probability threshold. According to ROC and DCA, machine learning model is more accurate in predicting mortality after elective cardiac surgery than EuroSCORE II. These results confirm the use of machine learning methods in the field of medical prediction.

BackgroundHigh-density lipoproteins (HDLs), particles characterized by their reverse cholesterol transport function, display pleiotropic properties, including anti-inflammatory and antioxidant functions. Moreover, all lipoproteins (HDLs but also low-density lipoproteins (LDLs)) neutralize lipopolysaccharides, leading to increased bacterial clearance. These two lipoproteins decrease during sepsis, and an association between low lipoprotein levels and poor outcome was reported. The goals of this study were to characterize the lipid profile of septic patients hospitalized in our intensive care unit (ICU) and to determine the relationship with the outcome.MethodsA prospective observational study was conducted in a university hospital ICU. All consecutive patients admitted for septic shock or sepsis were included. Total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglyceride levels were assessed at admission (day 1), at day 3, and at ICU discharge. When available, a prehospitalization lipid profile collected prior to the patient’s hospitalization was compiled. Short-term and 1-year prognostic outcomes were prospectively assessed.ResultsA total of 205 patients were included. We found a decrease in HDL-C concentration between previous values and those at admission, followed by an additional decrease at day 3. At ICU discharge, the concentration was higher than that at day 3 but did not reach the concentration measured prior to hospitalization (prior HDL-C = 1.22 (1.04–1.57) mmol/l; day 1 HDL-C = 0.44 (0.29–0.70) mmol/l; day 3 HDL-C = 0.30 (0.25–0.48) mmol/l; and HDL-C at discharge = 0.65 (0.42–0.82) mmol/l). A similar trend was found for LDL-C (prior LDL-C = 2.7 (1.91–3.33) mmol/l; day 1 LDL-C = 1.0 (0.58–1.50) mmol/l; day 3 LDL-C = 1.04 (0.64–1.54) mmol/l; and LDL-C at discharge = 1.69 (1.26–2.21) mmol/l). Mixed models for repeated measures of lipoprotein concentrations showed a significant difference in HDL-C and LDL-C concentrations over time between survivors and nonsurvivors at day 28. An HDL-C concentration at admission of less than 0.4 mmol/l was associated with increased mortality at day 28 (log-rank test, p = 0.034) but not at 1 year (log-rank test, p = 0.24). An LDL-C concentration at admission of less than 0.72 mmol/l was associated with increased mortality at day 28 and at 1 year (log-rank test, p < 0.001 and p = 0.007, respectively). No link was found between prior lipid profile and mortality.ConclusionsWe showed no relationship between the prehospitalization lipid profile and patient outcome, but low lipoprotein levels in the ICU were strongly associated with short-term mortality.

High-density lipoproteins (HDLs) are synthesized by the liver and display endothelioprotective properties, including anti-inflammatory, antiapoptotic, antithrombotic and antioxidant effects. In both septic and chronic liver failure patients, a low HDL cholesterol (HDL-C) concentration is associated with overmortality. Whereas sepsis-associated liver dysfunction is poorly defined, the aim of this study was to characterize the relationship between liver dysfunction, lipoprotein concentrations and mortality in septic patients in the intensive care unit (ICU). A prospective observational study was conducted in a university hospital ICU. All consecutive patients admitted for septic shock or sepsis were included. Total cholesterol, HDL-C, low-density lipoprotein-cholesterol (LDL-C), and triglyceride levels were assessed at admission. Sepsis-associated liver dysfunction was defined as a serum bilirubin[greater than or equal to] 2N or aspartate aminotransferase/alanine aminotransferase concentrations [greater than or equal to] 2N. Short-term and one-year prognostic outcomes were prospectively assessed. A total of 219 septic patients were included, and 15% of them presented with sepsis-associated liver dysfunction at admission. Low concentrations of lipoproteins were associated with mortality at Day 28 in the overall population. Sepsis-associated liver dysfunction at admission was associated with overmortality. In this subgroup, patients had a lower HDL-C concentration than patients without hepatic dysfunction (HDL-C = 0.31 [0.25, 0.55] mmol/L vs. 0.48 [0.29, 0.73] mmol/L, p = 0.0079) but there was no relationship with the outcome. Interestingly, no correlation was observed between lipoprotein concentrations and liver dysfunction markers. Sepsis-associated liver dysfunction at ICU admission is strongly associated with overmortality and is associated with a lower HDL-C concentration. However, in this subgroup of patients, HDL-C concentration had no relationship with mortality. Further exploratory studies are needed to better understand the interaction between lipoproteins and liver dysfunction during sepsis.

Background Temporal changes in the microbiological resistance profile have been reported in several life-threatening infections. However, no data have ever assessed this issue in postoperative peritonitis (POP). Our purpose was to assess the rate of multidrug-resistant organisms (MDROs) in POP over a two-decade period and to analyse their influence on the adequacy of empirical antibiotic therapy (EAT). Methods This retrospective monocentric analysis (1999–2019) addressed the changes over time in microbiologic data, including the emergence of MDROs and the adequacy of EAT for all intensive care unit adult patients treated for POP. The in vitro activities of 10 antibiotics were assessed to determine the most adequate EAT in the largest number of cases among 17 antibiotic regimens in patients with/without MDRO isolates. Our primary endpoint was to determine the frequency of MDRO and their temporal changes. Our second endpoint assessed the impact of MDROs on the adequacy of EAT per patient and their temporal changes based on susceptibility testing. In this analysis, the subgroup of patients with MDRO was compared with the subgroup of patients free of MDRO. Results A total of 1,318 microorganisms were cultured from 422 patients, including 188 (45%) patients harbouring MDROs. The growing proportions of MDR Enterobacterales were observed over time ( p  = 0.016), including ESBL-producing strains ( p  = 0.0013), mainly related to Klebsiella spp ( p  < 0.001). Adequacy of EAT was achieved in 305 (73%) patients. Decreased adequacy rates were observed when MDROs were cultured [ p  = 0.0001 vs. MDRO-free patients]. Over the study period, decreased adequacy rates were reported for patients receiving piperacillin/tazobactam in monotherapy or combined with vancomycin and imipenem/cilastatin combined with vancomycin ( p  < 0.01 in the three cases). In patients with MDROs, the combination of imipenem/cilastatin + vancomycin + amikacin or ciprofloxacin reached the highest adequacy rates (95% and 91%, respectively) and remained unchanged over time. Conclusions We observed high proportions of MDRO in patients treated for POP associated with increasing proportions of MDR Enterobacterales over time. High adequacy rates were only achieved in antibiotic combinations involving carbapenems and vancomycin, while piperacillin/tazobactam is no longer a drug of choice for EAT in POP in infections involving MDRO. Graphical Abstract

High-density lipoproteins (HDLs) are synthesized by the liver and display endothelioprotective properties, including anti-inflammatory, antiapoptotic, antithrombotic and antioxidant effects. In both septic and chronic liver failure patients, a low HDL cholesterol (HDL-C) concentration is associated with overmortality. Whereas sepsis-associated liver dysfunction is poorly defined, the aim of this study was to characterize the relationship between liver dysfunction, lipoprotein concentrations and mortality in septic patients in the intensive care unit (ICU). A prospective observational study was conducted in a university hospital ICU. All consecutive patients admitted for septic shock or sepsis were included. Total cholesterol, HDL-C, low-density lipoprotein-cholesterol (LDL-C), and triglyceride levels were assessed at admission. Sepsis-associated liver dysfunction was defined as a serum bilirubin[greater than or equal to] 2N or aspartate aminotransferase/alanine aminotransferase concentrations [greater than or equal to] 2N. Short-term and one-year prognostic outcomes were prospectively assessed. A total of 219 septic patients were included, and 15% of them presented with sepsis-associated liver dysfunction at admission. Low concentrations of lipoproteins were associated with mortality at Day 28 in the overall population. Sepsis-associated liver dysfunction at admission was associated with overmortality. In this subgroup, patients had a lower HDL-C concentration than patients without hepatic dysfunction (HDL-C = 0.31 [0.25, 0.55] mmol/L vs. 0.48 [0.29, 0.73] mmol/L, p = 0.0079) but there was no relationship with the outcome. Interestingly, no correlation was observed between lipoprotein concentrations and liver dysfunction markers. Sepsis-associated liver dysfunction at ICU admission is strongly associated with overmortality and is associated with a lower HDL-C concentration. However, in this subgroup of patients, HDL-C concentration had no relationship with mortality. Further exploratory studies are needed to better understand the interaction between lipoproteins and liver dysfunction during sepsis.

Background Bacterial respiratory infections (BRI) are major complications contributing to increased morbidity and mortality after lung transplantation (LT). This study analyzed epidemiology and outcome of 175 consecutive patients developing BRI in ICU after LT between 2006 and 2012. Methods Three situations were described: colonization determined in donors and recipients, pneumonia and tracheobronchitis during the first 28 postoperative days. Severity score, demographic, bacteriologic and outcome data were collected. Results 26% of donors and 31% of recipients were colonized. 92% of recipients developed BRI, including at least one episode of pneumonia in 19% of recipients. Only 21% of recipients developed BRI with an organism cultured from the donor’s samples, while 40% of recipients developed BRI with their own bacteria cultured before LT. Purulent sputum appears to be an important factor to discriminate tracheobronchitis from pneumonia. When compared to patients with tracheobronchitis, those with pneumonia had longer durations of mechanical ventilation (13 [3–27] vs 3 [29], p  = 0.0005) and ICU stay (24 [16–34] vs 14 [9-22], p  = 0.002). Pneumonia was associated with higher 28-day (11 (32%) vs 9 (7%), p  = 0.0004) and one-year mortality rates (21 (61%) vs 24 (19%), p  ≤ 0.0001). Conclusions These data confirm the high frequency of BRI right from the early postoperative period and the poor prognosis of pneumonia after LT.

PurposeThe management of chest stab wounds necessitates to perform an efficient imaging strategy. Compared to chest X-ray, computed tomography (CT) scan has a higher sensitivity. Nevertheless, the utility of diagnosing occult injuries remains controversial. Previous studies reported very different rates of management modifications induced by CT-scan. Indeed, no study specifically addressed the issue of ruling out traumatic diaphragmatic injury (TDI) in the specific population of chest stab trauma. The aim of the study was to evaluate the rate of thoracic procedures induced or guided by the results of thoracic CT-scan in the specific population of chest stab wounds. Secondary objective was to evaluate the utility of CT-scan for the diagnosis of TDI.MethodsWe conducted a prospective observational study. All consecutive patients referred to the acute care unit were included. We recorded the general characteristics of patients, the localization of wounds, all imaging tests, the final injury diagnosis, and the patients’ management. We compared patients with modifications of management induced by CT-scan results to other patients. We evaluated the performance of CT-scan for the diagnosis of TDI by calculating its sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV).Results153 patients were included. There were 92 patients with normal chest X-ray. 67 of them received a CT-scan. 34 (51%) patients had an abnormal CT-scan, including 19 (21%) patients with thoracic new findings, with 3 (4.5%) modification of management. There were 50 patients who had an abnormal chest X-ray. 31 of them received a CT-scan, and 31 (100%) had an abnormal CT-scan, including 19 thoracic new findings, with 11 (36%) modifications of management. The diagnostic performance of CT-scan for TDI was: sensitivity 50%; specificity 95%; NPV 72%; PPV 88%.ConclusionsIn chest stab trauma, CT-scans may be unnecessary outside the thoracoabdominal zone when chest X-ray is normal. In other cases, CT-scan seems to have an impact on the decision-making. In case of thoracoabdominal wounds, CT-scan helps to detect intra-abdominal injuries. The performance of CT-scan to diagnose TDI is not high enough to reliably rule out all TDI.

Background Some guidelines advocate for managing patients with penetrating thoracic wounds in trauma centres with cardiothoracic surgery. This systematic approach is questionable. Only 15% of these patients require surgery. It is known that clinical examination fails to detect hemopneumothorax in penetrating trauma. However, no studies have evaluated the combined diagnostic performance of vital signs and the clinical evaluation of wounds. The clinical characteristics of wounds have not been investigated. We aimed to evaluate the ability of combinations of pre-hospital signs to rule out invasive chest stab trauma. Methods This was a prospective observational study. All consecutive adult patients hospitalized in the perioperative acute care unit of a tertiary university hospital were included. Injury diagnoses were provided by exploratory surgery and imaging tests. Patients with a final diagnosis of invasive wounds (IWs) and patients with only superficial wounds were compared. Data regarding management and outcome were analysed. Results A total of 153 patients were included. After imaging or surgery, 58 (38%) patients were diagnosed with only superficial wounds, and 95 (62%) were diagnosed with thoracic or abdominal IWs. The false-negative rate of pre-hospital evaluations in the diagnosis of IWs was 42% [31–51] IQR25–75 . In stable patients, pre-hospital data could not rule out IWs, with a negative predictive value of 58% and a positive predictive value of 70%. Twenty-nine (19%) patients required early emergent cardiothoracic surgery. Among these patients, 8 (28%) had no evidence of IWs in the pre-hospital period. Among the 59 patients without pre-hospital signs of IWs, 19 (33%) underwent at least one emergent procedure. Conclusions The combination of pre-hospital vital signs, visual evaluation of wounds, and physical examination failed to rule out IWs in patients with chest stab wounds. This implies that caution is needed in triage decision-making.

BackgroundConflicting studies exist about the effectiveness of cardiopulmonary resuscitation (CPR) on a dental chair. In some situations, dental surgeons are obliged to perform CPR with the patient on the chair. Feedback devices are supposed to guide the compression depth in order to improve the quality of CPR, but some devices are based on an accelerometer that can theoretically report erroneous results because of the lack of rigidity of a dental chair.ObjectiveThe aim of this study was to evaluate the accuracy of these devices to guide chest compressions on a dental chair.MethodsA prospective, randomised, crossover, equivalence/non-inferiority study was conducted to compare the values of compression depths provided by the feedback device (Real CPR Help®, delivered by Zoll© Medical Corporation, Chelmsford, MA, USA) with the real measurements provided by the manikin (Resusci Anne® Advanced SkillTrainer, Laerdal Medical AS©, Norway). Chest-compression-only CPR was performed by 15 Basic Life Support instructors who carried out two rounds of continuous CPR for 2 min each. Data were analysed with a correlation test, a Bland–Altman method and a Wilcoxon test. Statistical significance was defined as p<0.05.ResultsA significant difference was found between the mean depths of compression measured by the feedback device and the manikin on a dental chair and on the floor (p<0.0001). The feedback device overestimated the depth of chest compressions, and Bland–Altman analysis demonstrated poor agreement.ConclusionThis study indicates that feedback devices with accelerometer technology are not sufficiently reliable to ensure adequate chest compression on dental chairs.

Anticoagulation in patients with mechanical heart valves (MHV) is associated with a risk of major bleeding episodes (MBE). In case of MBE, anticoagulant interruption is advocated. However, there is lack of data regarding the thrombo-embolic events (TE) risk associated with anticoagulant interruption. The main objective of the study was to evaluate the rate and risk factors of 6-months of TEs in patients with MHV experiencing MBE. This observational study was conducted over a 13-year period. Adult patients with a MHV presenting with a MBE were included. The main study endpoint was 6-month TEs, defined by clinical TEs or an echocardiographic documented thrombosis, occurring during an ICU stay or within 6-months. Thromboembolic events were recorded at ICU discharge, and 6 months after discharge. Seventy-nine MBEs were analysed, the rate of TEs at 6-months was 19% CI [11-29%]. The only difference of presentation and management between 6-month TEs and free-TE patients was the time without effective anticoagulation (TWA). The Receiver Operator Characteristic curve identified the value of 122 h of TWA as a cut-off. The multivariate analysis identified early bleeding recurrences (OR 3.62, 95% CI [1.07–12.25], p = 0.039), and TWA longer than 122 h (OR 4.24, 95% CI [1.24–14.5], p = 0.021), as independent risk factors for 6-month TEs. A higher rate of TE was associated with anticoagulation interruption longer than 5 days and early bleeding recurrences. However, the management should still be personalized and discussed for each case given the heterogeneity of causes of MBE and possibilities of haemostatic procedures.