Machine learning algorithms assist early evaluation of enteral nutrition in ICU patients.

Background: This study applied machine learning (ML) algorithms to construct a model for predicting EN initiation for patients in the intensive care unit (ICU) and identifying populations in need of EN at an early stage. Methods: This study collected patient information from the Medical Information Mart for Intensive Care IV database. All patients enrolled were split randomly into a training set and a validation set. Six ML models were established to evaluate the initiation of EN, and the best model was determined according to the area under curve (AUC) and accuracy. The best model was interpreted using the Local Interpretable Model-Agnostic Explanations (LIME) algorithm and SHapley Additive exPlanation (SHAP) values. Results: A total of 53,150 patients participated in the study. They were divided into a training set (42,520, 80%) and a validation set (10,630, 20%). In the validation set, XGBoost had the optimal prediction performance with an AUC of 0.895. The SHAP values revealed that sepsis, sequential organ failure assessment score, and acute kidney injury were the three most important factors affecting EN initiation. The individualized forecasts were displayed using the LIME algorithm. Conclusion: The XGBoost model was established and validated for early prediction of EN initiation in ICU patients.

Comparison of sequential feeding and continuous feeding on the blood glucose of critically ill patients: a non-inferiority randomized controlled trial.

BACKGROUND: Glucose control is an important aspect in managing critically ill patients. The goal of this study was to compare the effects of sequential feeding (SF) and continuous feeding (CF) on the blood glucose of critically ill patients. METHODS: A non-inferiority randomized controlled trial was adopted in this study. A total of 62 patients who were fed enteral nutritional suspension through gastric tubes were enrolled. After achieving 80% of the nutrition target calories (25 kcal·kg-1·day-1) through CF, the patients were then randomly assigned into SF and CF groups. In the SF group, the feeding/fasting time was reasonably determined according to the circadian rhythm of the human body as laid out in traditional Chinese medicine theory. The total daily dosage of the enteral nutritional suspension was equally distributed among three time periods of 7 to 9 o'clock, 11 to 13 o'clock, and 17 to 19 o'clock. The enteral nutritional suspension in each time period was pumped at a uniform rate within 2 h by an enteral feeding pump. In the CF group, patients received CF at a constant velocity by an enteral feeding pump throughout the study. Blood glucose values at five points (6:00/11:00/15:00/21:00/1:00) were monitored and recorded for seven consecutive days after randomization. Enteral feeding intolerance was also recorded. Non-inferiority testing was adopted in this study, the chi-square test or Fisher test was used for qualitative data, and the Mann-Whitney U test was used for quantitative data to determine differences between groups. In particular, a repeated measure one-way analysis of variance was used to identify whether changes in glucose value variables across the time points were different between the two groups. RESULTS: There were no significant demographic or physiological differences between the SF and CF groups (P > 0.050). The average glucose level in SF was not higher than that in CF (8.8 [7.3-10.3] vs. 10.7 [9.1-12.1] mmol/L, Z = -2.079, P for non-inferiority = 0.019). Hyperglycemia incidence of each patient was more common in the CF group than that in the SF group (38.4 [19.1-63.7]% vs. 11.8 [3.0-36.7]%, Z = -2.213, P = 0.027). Hypoglycemia was not found in either group. Moreover, there was no significant difference during the 7 days in the incidence of feeding intolerance (P > 0.050). CONCLUSIONS: In this non-inferiority study, the average blood glucose in SF was not inferior to that in CF. The feeding intolerance in SF was similar to that in CF. SF may be as safe as CF for critically ill patients. TRIAL REGISTRATION: ClinicalTrials.gov, NCT03439618; https://clinicaltrials.gov/ct2/show/record/NCT03439618.

Expression of human leukocyte antigen-G and acute rejection in patients following liver transplantation.

Acute liver rejection is one of the most severe complications that may affect the liver transplantation procedure. Thus, one of the most important focal points in the field of liver transplantation research is to discover a non-invasive or less-invasive method of diagnosing and predicting cases of acute liver rejection. In the present study, 59 tissue samples, including blood and liver tissues, were collected from patients who underwent liver transplantation between March 2005 and November 2009. The patients were divided into acute rejection and no rejection groups, the latter of which was further divided into normal and abnormal liver function groups. The samples were assayed by enzyme-linked immunosorbent assay and immunohistochemistry methods. The results were analyzed and a receiver operating characteristic (ROC) curve was plotted. The area under the ROC curve and the sensitivity and specificity of the cut-off point were analyzed statistically. The results indicated that the expression level of human leukocyte antigen-G (HLA-G) in the serum and liver samples in the acute rejection group was markedly lower than that in the no rejection group (P<0.001 and P=0.004, respectively). Furthermore, in the no rejection group, no statistically significant difference was identified in the level of HLA-G between patients with normal or abnormal liver function (P=0.0593). The area under the ROC curve was 0.805. When 2.41 U/ml HLA-G was considered as the cut-off point for the diagnosis of acute liver rejection, the sensitivity and specificity were 72.7 and 83.8%, respectively. In conclusion, in the present study, a high expression of the HLA-G was shown to correlate with a reduced occurrence of acute liver rejection. HLA-G may thus be an effective factor for the diagnosis and prediction of acute liver rejection.