To estimate the mortality and length of stay in the intensive care unit (ICU) attributable to clinically important gastrointestinal bleeding in mechanically ventilated critically ill patients.Three strategies were used to estimate the mortality attributable to bleeding in two multicentre databases. The first method matched patients who bled with those who did not (matched cohort), using duration of ICU stay prior to the bleed, each of six domains of the Multiple Organ Dysfunction Score (MODS) measured 3 days prior to the bleed, APACHE II score, age, admitting diagnosis, and duration of mechanical ventilation. The second approach employed Cox proportional hazards regression to match bleeding and non-bleeding patients (model-based matched cohort). The third method, instead of matching, derived estimates based on regression modelling using the entire population (regression method). Three parallel analyses were conducted for the length of ICU stay attributable to clinically important bleeding.Sixteen Canadian university-affiliated ICUs.A total of 1666 critically ill patients receiving mechanical ventilation for at least 48 hours.We prospectively collected data on patient demographics, APACHE II score, admitting diagnosis, daily MODS, clinically important bleeding, length of ICU stay, and mortality. Independent adjudicators determined the occurrence of clinically important gastrointestinal bleeding, defined as overt bleeding in association with haemodynamic compromise or blood transfusion.Of 1666 patients, 59 developed clinically important gastrointestinal bleeding. The mean APACHE II score was 22.9 +/- 8.6 among bleeding patients and 23.3 +/- 7.7 among non-bleeding patients. The risk of death was increased in patients with bleeding using all three analytic approaches (matched cohort method: relative risk [RR]= 2.9, 95% confidence interval (CI)= 1.6-5.5; model-based matched cohort method: RR = 1.8, 95% CI = 1.1-2.9; and the regression method: RR = 4.1, 95% CI = 2.6-6.5). However, this was not significant for the adjusted regression method (RR = 1.0, 95% CI = 0.6-1.7). The median length of ICU stay attributable to clinically important bleeding for these three methods, respectively, was 3.8 days (95% CI = -0.01 to 7.6 days), 6.7 days (95% CI = 2.7-10.7 days), and 7.9 days (95% CI = 1.4-14.4 days).Clinically important upper gastrointestinal bleeding has an important attributable morbidity and mortality, associated with a RR of death of 1-4 and an excess length of ICU stay of approximately 4-8 days.

To describe the prevalence of, risk factors for, and prognostic importance of gastrointestinal (GI) bleeding and use of acid suppressants in acutely ill adult intensive care patients.We included adults without GI bleeding who were acutely admitted to the intensive care unit (ICU) during a 7-day period. The primary outcome was clinically important GI bleeding in ICU, and the analyses included estimations of baseline risk factors and potential associations with 90-day mortality.A total of 1,034 patients in 97 ICUs in 11 countries were included. Clinically important GI bleeding occurred in 2.6 % (95 % confidence interval 1.6-3.6 %) of patients. The following variables at ICU admission were independently associated with clinically important GI bleeding: three or more co-existing diseases (odds ratio 8.9, 2.7-28.8), co-existing liver disease (7.6, 3.3-17.6), use of renal replacement therapy (6.9, 2.7-17.5), co-existing coagulopathy (5.2, 2.3-11.8), acute coagulopathy (4.2, 1.7-10.2), use of acid suppressants (3.6, 1.3-10.2) and higher organ failure score (1.4, 1.2-1.5). In ICU, 73 % (71-76 %) of patients received acid suppressants; most received proton pump inhibitors. In patients with clinically important GI bleeding, crude and adjusted odds for mortality were 3.7 (1.7-8.0) and 1.7 (0.7-4.3), respectively.In ICU patients clinically important GI bleeding is rare, and acid suppressants are frequently used. Co-existing diseases, liver failure, coagulopathy and organ failures are the main risk factors for GI bleeding. Clinically important GI bleeding was not associated with increased adjusted 90-day mortality, which largely can be explained by severity of comorbidity, other organ failures and age.

To examine the use of endoscopy in the UK for acute upper gastrointestinal bleeding (AUGIB) and compare with published standards. To assess the organisation of endoscopy services for AUGIB in the UK. To examine the relationship between outcomes and out of hours (OOH) service provision.Multi-centre cross sectional clinical audit.All UK hospitals accepting admissions with AUGIB.All adults (>or=16 yrs) presenting with AUGIB between 1st May and 30th June 2007.Collection A custom designed web-based reporting tool was used to collect data on patient characteristics, comorbidity and haemodynamic status at presentation to calculate the Rockall score, use and timing of endoscopy, treatment including endoscopic, rebleeding and in-hospital mortality. A mailed questionnaire was used to collect data on facilities and service organisation.Data on 6750 patients (median age 68 years) were analysed from 208 hospitals. 74% underwent inpatient endoscopy; of these 50% took place within 24 h of presentation, 82% during normal working hours and 3% between midnight and 8 am. Of patients deemed high-risk (pre-endoscopy Rockall score >or=5) only 55% were endoscoped within 24 h and 14% waited >or=72 h for endoscopy. Lesions with a high risk of rebleeding were present in 28% of patients of whom 74% received endoscopic therapy. Further bleeding was evident in 13% and mortality in those endoscoped was 7.4% (95% CI 6.7% to 8.1%). In 52% of hospitals a consultant led out of hours (OOH) endoscopy rota existed; in these hospitals 20% of first endoscopies were performed OOH compared with 13% in those with no OOH rota and endoscopic therapy was more likely to be administered (25% vs 21% in hospitals with no OOH rota). The risk adjusted mortality ratio was higher (1.21, p=0.10, (95%CI 0.96 to 1.51)) in hospitals without such rotas.This audit has found continuing delays in performing endoscopy after AUGIB and underutilisation of standard endoscopic therapy particularly for variceal bleeding. In hospitals with a formal OOH endoscopy rota patients received earlier endoscopy, were more likely to receive endoscopic therapy and may have a lower mortality.

Prophylactic endotracheal intubation (PEI) is often advocated to mitigate the risk of cardiopulmonary adverse events in patients presenting with brisk upper GI bleeding (UGIB). However, the benefit of such a measure remains controversial. Our study aimed to compare the incidence of cardiopulmonary unplanned events between critically ill patients with brisk UGIB who underwent endotracheal intubation versus those who did not.Patients aged 18 years or older who presented at Cleveland Clinic between 2011 and 2014 with hematemesis and/or patients with melena with consequential hypovolemic shock were included. The primary outcome was a composite of several cardiopulmonary unplanned events (pneumonia, pulmonary edema, acute respiratory distress syndrome, persistent shock/hypotension after the procedure, arrhythmia, myocardial infarction, and cardiac arrest) occurring within 48 hours of the endoscopic procedure. Propensity score matching was used to match each patient 1:1 in variables that could influence the decision to intubate. These included Glasgow Blatchford Score, Charleston Comorbidity Index, and Acute Physiology and Chronic Health Evaluation scores.Two hundred patients were included in the final analysis. The baseline characteristics, comorbidity scores, and prognostic scores were similar between the 2 groups. The overall cardiopulmonary unplanned event rates were significantly higher in the intubated group compared with the nonintubated group (20% vs 6%, P =.008), which remained significant (P =.012) after adjusting for the presence of esophageal varices.PEI before an EGD for brisk UGIB in critically ill patients is associated with an increased risk of unplanned cardiopulmonary events. The benefits and risks of intubation should be carefully weighed when considering airway protection before an EGD in this group of patients.Copyright © 2017 American Society for Gastrointestinal Endoscopy. Published by Elsevier Inc. All rights reserved.

Improving the Intensive Care Unit (ICU) management network and building cost-effective and well-managed healthcare systems are high priorities for healthcare units. Creating accurate and explainable mortality prediction models helps identify the most critical risk factors in the patients’ survival/death status and early detect the most in-need patients. This study proposes a highly accurate and efficient machine learning model for predicting ICU mortality status upon discharge using the information available during the first 24 hours of admission. The most important features in mortality prediction are identified, and the effects of changing each feature on the prediction are studied. We used supervised machine learning models and illness severity scoring systems to benchmark the mortality prediction. We also implemented a combination of SHAP, LIME, partial dependence, and individual conditional expectation plots to explain the predictions made by the best-performing model (CatBoost). We proposed E-CatBoost, an optimized and efficient patient mortality prediction model, which can accurately predict the patients’ discharge status using only ten input features. We used eICU-CRD v2.0 to train and validate the models; the dataset contains information on over 200,000 ICU admissions. The patients were divided into twelve disease groups, and models were fitted and tuned for each group. The models’ predictive performance was evaluated using the area under a receiver operating curve (AUROC). The AUROC scores were 0.86 [std:0.02] to 0.92 [std:0.02] for CatBoost and 0.83 [std:0.02] to 0.91 [std:0.03] for E-CatBoost models across the defined disease groups; if measured over the entire patient population, their AUROC scores were 7 to 18 and 2 to 12 percent higher than the baseline models, respectively. Based on SHAP explanations, we found age, heart rate, respiratory rate, blood urine nitrogen, and creatinine level as the most critical cross-disease features in mortality predictions.

The American College of Gastroenterology recommends early risk stratification in patients presenting with upper GI bleeding (UGIB). The AIMS65 score is a risk stratification score previously validated to predict inpatient mortality. The aim of this study was to validate the AIMS65 score as a predictor of inpatient mortality in patients with acute UGIB and to compare it with established pre- and postendoscopy risk scores.ICD-10 (International Classification of Diseases, Tenth Revision) codes identified patients presenting with UGIB requiring endoscopy. All patients were risk stratified by using the AIMS65, Glasgow-Blatchford score (GBS), pre-endoscopy Rockall, and full Rockall scores. The primary outcome was inpatient mortality. Secondary outcomes were a composite endpoint of inpatient mortality, rebleeding, and endoscopic, radiologic, or surgical intervention; blood transfusion requirement; intensive care unit (ICU) admission; rebleeding; and hospital length of stay. The area under the receiver-operating characteristic curve (AUROC) was calculated for each score.Of the 424 study patients, 18 (4.2%) died and 69 (16%) achieved the composite endpoint. The AIMS65 score was superior to both the GBS (AUROC, 0.80 vs 0.76, P <.027) and the pre-endoscopy Rockall score (0.74, P =.001) and equivalent to the full Rockall score (0.78, P =.18) in predicting inpatient mortality. The AIMS65 score was superior to all other scores in predicting the need for ICU admission and length of hospital stay. AIMS65, GBS, and full Rockall scores were equivalent (AUROCs, 0.63 vs 0.62 vs 0.63, respectively) and superior to pre-endoscopy Rockall (AUROC, 0.55) in predicting the composite endpoint. GBS was superior to all other scores for predicting blood transfusion.The AIMS65 score is a simple risk stratification score for UGIB with accuracy superior to that of GBS and pre-endoscopy Rockall scores in predicting in-hospital mortality and the need for ICU admission.Crown Copyright © 2016. Published by Elsevier Inc. All rights reserved.

The ICD-9-CM adaptation of the Charlson comorbidity score has been a valuable resource for health services researchers. With the transition into ICD-10 coding worldwide, an ICD-10 version of the Deyo adaptation was developed and validated using population-based hospital data from Victoria, Australia.The algorithm was translated from ICD-9-CM into ICD-10-AM (Australian modification) in a multistep process. After a mapping algorithm was used to develop an initial translation, these codes were manually examined by the coding experts and a general physician for face validity. Because the ICD-10 system is country specific, our goal was to keep many of the translated code at the three-digit level for generalizability of the new index.There appears to be little difference in the distribution of the Charlson Index score between the two versions. A strong association between increasing index scores and mortality exists: the area under the ROC curve is 0.865 for the last year using the ICD-9-CM version and remains high, at 0.855, for the ICD-10 version.This work represents the first rigorous adaptation of the Charlson comorbidity index for use with ICD-10 data. In comparison with a well-established ICD-9-CM coding algorithm, it yields closely similar prevalence and prognosis information by comorbidity category.

The Model for End-stage Liver Disease (MELD) was initially created to predict survival in patients with complications of portal hypertension undergoing elective placement of transjugular intrahepatic portosystemic shunts. The MELD which uses only objective variables was validated subsequently as an accurate predictor of survival among different populations of patients with advanced liver disease. The major use of the MELD score has been in allocation of organs for liver transplantation. However, the MELD score has also been shown to predict survival in patients with cirrhosis who have infections, variceal bleeding, as well as in patients with fulminant hepatic failure and alcoholic hepatitis. MELD may be used in selection of patients for surgery other than liver transplantation and in determining optimal treatment for patients with hepatocellular carcinoma who are not candidates for liver transplantation. Despite the many advantages of the MELD score, there are approximately 15%-20% of patients whose survival cannot be accurately predicted by the MELD score. It is possible that the addition of variables that are better determinants of liver and renal function may improve the predictive accuracy of the model. Efforts at further refinement and validation of the MELD score will continue.

1: ESGE recommends in patients with acute upper gastrointestinal hemorrhage (UGIH) the use of the Glasgow-Blatchford Score (GBS) for pre-endoscopy risk stratification. Patients with GBS ≤ 1 are at very low risk of rebleeding, mortality within 30 days, or needing hospital-based intervention and can be safely managed as outpatients with outpatient endoscopy.Strong recommendation, moderate quality evidence. 2: ESGE recommends that in patients with acute UGIH who are taking low-dose aspirin as monotherapy for secondary cardiovascular prophylaxis, aspirin should not be interrupted. If for any reason it is interrupted, aspirin should be re-started as soon as possible, preferably within 3-5 days.Strong recommendation, moderate quality evidence. 3: ESGE recommends that following hemodynamic resuscitation, early (≤ 24 hours) upper gastrointestinal (GI) endoscopy should be performed. Strong recommendation, high quality evidence. 4: ESGE does not recommend urgent (≤ 12 hours) upper GI endoscopy since as compared to early endoscopy, patient outcomes are not improved. Strong recommendation, high quality evidence. 5: ESGE recommends for patients with actively bleeding ulcers (FIa, FIb), combination therapy using epinephrine injection plus a second hemostasis modality (contact thermal or mechanical therapy). Strong recommendation, high quality evidence. 6: ESGE recommends for patients with an ulcer with a nonbleeding visible vessel (FIIa), contact or noncontact thermal therapy, mechanical therapy, or injection of a sclerosing agent, each as monotherapy or in combination with epinephrine injection. Strong recommendation, high quality evidence. 7 : ESGE suggests that in patients with persistent bleeding refractory to standard hemostasis modalities, the use of a topical hemostatic spray/powder or cap-mounted clip should be considered. Weak recommendation, low quality evidence. 8: ESGE recommends that for patients with clinical evidence of recurrent peptic ulcer hemorrhage, use of a cap-mounted clip should be considered. In the case of failure of this second attempt at endoscopic hemostasis, transcatheter angiographic embolization (TAE) should be considered. Surgery is indicated when TAE is not locally available or after failed TAE. Strong recommendation, moderate quality evidence. 9: ESGE recommends high dose proton pump inhibitor (PPI) therapy for patients who receive endoscopic hemostasis and for patients with FIIb ulcer stigmata (adherent clot) not treated endoscopically. (A): PPI therapy should be administered as an intravenous bolus followed by continuous infusion (e. g., 80 mg then 8 mg/hour) for 72 hours post endoscopy. (B): High dose PPI therapies given as intravenous bolus dosing (twice-daily) or in oral formulation (twice-daily) can be considered as alternative regimens.Strong recommendation, high quality evidence. 10: ESGE recommends that in patients who require ongoing anticoagulation therapy following acute NVUGIH (e. g., peptic ulcer hemorrhage), anticoagulation should be resumed as soon as the bleeding has been controlled, preferably within or soon after 7 days of the bleeding event, based on thromboembolic risk. The rapid onset of action of direct oral anticoagulants (DOACS), as compared to vitamin K antagonists (VKAs), must be considered in this context.Strong recommendation, low quality evidence.European Society of Gastrointestinal Endoscopy. All rights reserved.