Early Risk Score for Predicting Hypotension in Normotensive Patients with Non-Variceal Upper Gastrointestinal Bleedin

Comparison of Lactate Clearance with Established Risk Assessment Tools in Predicting Outcomes in Acute Upper Gastrointestinal Bleeding

Early risk stratification is mandatory in acute upper gastrointestinal bleeding (AUGIB) to guide optimal treatment. Numerous risk scores were introduced, but lack of practicability led to limited use in daily clinical practice. Lactate clearance is an established risk assessment tool in a variety of diseases, such as trauma and sepsis. Therefore, this study compares the predictive ability of pre-endoscopic lactate clearance and established risk scores in patients with AUGIB at the University Hospital of Cologne. Active bleeding was detected in 27 (25.2%) patients, and hemostatic intervention was performed in 35 (32.7%). In total, 16 patients (15%) experienced rebleeding and 12 (11.2%) died. Initially, lactate levels were elevated in 64 cases (59.8%), and the median lactate clearance was 18.7% (2.7-48.2%). Regarding the need for endoscopic intervention, the predictive ability of Glasgow Blatchford Score, pre-endoscopic Rockall score, initial lactate and lactate clearance did not differ significantly, and their area under the receiver operating characteristic curves were 0.658 (0.560-0.747), 0.572 (0.473-0.667), 0.572 (0.473-0.667) and 0.583 (0.483-0.677), respectively. Similar results were observed in relation to rebleeding and mortality. In conclusion, lactate clearance had comparable predictive ability compared to established risk scores. Further prospective research is necessary to clarify the potential role of lactate clearance as a reliable risk assessment tool in AUGIB.

Pulse pressure during the initial resuscitative period in patients with septic shock treated with a protocol-driven resuscitation bundle therapy

BACKGROUND/AIMS

Maintaining a mean arterial pressure (MAP) ≥ 65 mmHg during septic shock should be based on individual circumstances, but specific target is poorly understood. We investigated associations between time-weighted average (TWA) hemodynamic parameters during the initial resuscitative period and 28-day mortality.

METHODS

Prospectively collected data were obtained from a septic shock patient registry, according to the Sepsis-3 definition, between 2016 and 2018. The TWA systolic blood pressure, diastolic blood pressure, MAP, shock index, and pulse pressure (PP) during the first 6 hours after shock recognition were compared. Multivariable regression analysis was performed to assess associations between these parameters and 28-day mortality.

RESULTS

Of 340 patients with septic shock, 92 died. Only the median TWA PP differed between the survivors and non-survivors (39.2 mmHg vs. 43.0 mmHg, p = 0.020), whereas the other indexes did not. When PP was divided into quartiles (< 34, 34 to 40, 40 to 48, and > 48 mmHg), the mortality rate was higher in the highest quartile (41.2%). Multivariable logistic analysis revealed that PP (odds ratio [OR], 1.28; 95% confidence interval [CI], 1.012 to 1.622; p = 0.039) and PP of > 48 mmHg (OR, 2.25; 95% CI, 1.272 to 3.981; p = 0.005) were independently associated with 28-day mortality.

CONCLUSION

PP was significantly associated with 28-day mortality in patients with septic shock and MAP maintained at > 65 mmHg during the first 6 hours. Further studies are warranted to optimize strategies for maintaining PP and MAP at > 65 mmHg during the early resuscitative period.

Immature Granulocyte Count and Percentage as New Predictors of Mortality in Patients with Upper Gastrointestinal Bleeding

Aims

Early identification of patients at risk of adverse outcomes may increase the survival rates in patients with upper gastrointestinal bleeding (UGIB), but this can be difficult to predict in emergencies. The aim of the study is to evaluate immature granulocyte (IG), which can be obtained from simple hemogram tests in patients with UGIB, in terms of clinical use and as a mortality marker.

Materials and methods

The patients diagnosed with UGIB between March 1, 2019, and September 30, 2019, were evaluated retrospectively. Demographic characteristics, causes of hemorrhage, clinical presentations, hemogram, and biochemistry values at ED admission and 30-day mortality status of the patients were examined. We divided the patients into groups according to their mortality status, and the groups were compared among themselves in terms of parameters.

Results

A total of 213 patients who met the inclusion criteria were included in the study. Of these patients, 139 (65.3%) were male and the mean age was 65.05 ± 16.7 years. Fifteen (7%) of them were in the nonsurvival group, while 198 (93%) were in the survival group. The efficacy of both the IG count (IGC) and IG% in predicting mortality was statistically significant (p = 0.002, p = 0.008, respectively). The sensitivity and specificity for the IGC were found as 60% and 84.4; for the IG%, they were found as 66.7% and 75.7%, respectively.

Conclusion

IGC and IG% are independent risk factors for the 30-day mortality status. These measurements are obtained from simple hemogram tests and may be useful for the evaluation of mortality in patients with UGIB.

How to cite this article

Bedel C, Korkut M, Avcı A, Uzun A. Immature Granulocyte Count and Percentage as New Predictors of Mortality in Patients with Upper Gastrointestinal Bleeding. Indian J Crit Care Med 2020;24(9):794-798.

Comparative Analysis on Machine Learning and Deep Learning to Predict Post-Induction Hypotension

Hypotensive events in the initial stage of anesthesia can cause serious complications in the patients after surgery, which could be fatal. In this study, we intended to predict hypotension after tracheal intubation using machine learning and deep learning techniques after intubation one minute in advance. Meta learning models, such as random forest, extreme gradient boosting (Xgboost), and deep learning models, especially the convolutional neural network (CNN) model and the deep neural network (DNN), were trained to predict hypotension occurring between tracheal intubation and incision, using data from four minutes to one minute before tracheal intubation. Vital records and electronic health records (EHR) for 282 of 319 patients who underwent laparoscopic cholecystectomy from October 2018 to July 2019 were collected. Among the 282 patients, 151 developed post-induction hypotension. Our experiments had two scenarios: using raw vital records and feature engineering on vital records. The experiments on raw data showed that CNN had the best accuracy of 72.63%, followed by random forest (70.32%) and Xgboost (64.6%). The experiments on feature engineering showed that random forest combined with feature selection had the best accuracy of 74.89%, while CNN had a lower accuracy of 68.95% than that of the experiment on raw data. Our study is an extension of previous studies to detect hypotension before intubation with a one-minute advance. To improve accuracy, we built a model using state-of-art algorithms. We found that CNN had a good performance, but that random forest had a better performance when combined with feature selection. In addition, we found that the examination period (data period) is also important.

Prediction of Adverse Events in Stable Non-Variceal Gastrointestinal Bleeding Using Machine Learning

Clinical risk-scoring systems are important for identifying patients with upper gastrointestinal bleeding (UGIB) who are at a high risk of hemodynamic instability. We developed an algorithm that predicts adverse events in patients with initially stable non-variceal UGIB using machine learning (ML). Using prospective observational registry, 1439 out of 3363 consecutive patients were enrolled. Primary outcomes included adverse events such as mortality, hypotension, and rebleeding within 7 days. Four machine learning algorithms, namely, logistic regression with regularization (LR), random forest classifier (RF), gradient boosting classifier (GB), and voting classifier (VC), were compared with the Glasgow-Blatchford score (GBS) and Rockall scores. The RF model showed the highest accuracies and significant improvement over conventional methods for predicting mortality (area under the curve: RF 0.917 vs. GBS 0.710), but the performance of the VC model was best in hypotension (VC 0.757 vs. GBS 0.668) and rebleeding within 7 days (VC 0.733 vs. GBS 0.694). Clinically significant variables including blood urea nitrogen, albumin, hemoglobin, platelet, prothrombin time, age, and lactate were identified by the global feature importance analysis. These results suggest that ML models will be useful early predictive tools for identifying high-risk patients with initially stable non-variceal UGIB admitted at an emergency department.