Objective Fluid therapy for sepsis patients has always been a problem that puzzles clinicians, that is, knowing when patients need fluid infusion and when they need negative fluid balance. Different clinicians may have different judgment criteria and make different decisions. Recently, studies have suggested that different fluid treatment strategies can cause different clinical outcomes. This study is intended to establish and verify a model for judging the direction of fluid therapy based on machine learning. Method This study included 2705 sepsis patients from the Peking Union Medical College Hospital Intensive Care Medical Information System and Database (PICMISD) from January 2016 to April 2020. The training set and test set (January 2016 to June 2019) were randomly divided. Twenty-seven features were extracted for modeling, including 25 state features (bloc, vital sign, laboratory examination, blood gas assay and demographics), 1 action feature (fluid balance) and 1 outcome feature (ICU survival or death). SARSA was used to learn the data rules of the training set. Deep Q-learning (DQN) was used to learn the relationship between states and actions of the training set and predict the next balance. A double-robust estimator was used to evaluate the average expected reward of the test set in the deep Q-learning model. Lastly, we verified the difference between the predicted fluid therapy model and the actual treatment for the patient's prognoses, with sepsis patient data from July 2019 to April 2020 as the validation set. Results The training set and test set were extracted from the same database, and the distribution of liquid balance was similar. Actions were divided into five intervals corresponding to 0–20, 20–40, 40–60, 60–80, and 80–100% percentiles of fluid balance. The higher the reward of Q ( s, a ) calculated by SARSA from the training set, the lower the mortality rate. Deep Q-learning indicates that both fluid balance differences that are too high and too low show an increase in mortality. The more consistent the fluid balance prediction with the real result, the lower the mortality rate. The smaller the difference between the prediction and the reality, the lower the mortality rate. The double-robust estimator shows that the model has satisfactory stability. The validation set indicates that the mortality rate of patients in the “predicted negative fluid balance and actual negative fluid balance” subgroup was the lowest, which was statistically significant, indicating that the model can be used for clinical verification. Conclusion We used reinforcement learning to propose a possible prediction model for guiding the direction of fluid therapy for sepsis patients in the ICU. This model may accurately predict the best direction for fluid therapy, thereby improving patient prognosis.

Introduction Septic shock, largely caused by intestinal perforation, is the most common critical illness in intensive care unit (ICU). As an important quality control strategy in ICU, deep vein thrombosis (DVT) prevention is routinely used in the treatment of septic shock. Nevertheless, the effects of DVT prevention on septic shock are not fully revealed. This study was thus designed to investigate the effects of DVT prevention on septic shock caused by intestinal perforation in China. Methods A total of 463 hospitals were enrolled in a survey, led by the China National Critical Care Quality Control Center (China NCCQC) from January 1, 2018 to December 31, 2018. The association between DVT prevention, including pharmacological prophylaxis and mechanical prophylaxis, and outcomes, such as prognosis, complications, hospital stays, and hospitalization costs, was determined in the present study. Main Results Notably, the increased rates of DVT prevention were not associated with the onset of complications in patients with septic shock caused by intestinal perforation ( p > 0.05). In addition, even though increased DVT prevention did not affect hospital stays, it significantly decreased the discharge rates without doctor's order in patients with septic shock caused by intestinal perforation ( p < 0.05). Nevertheless, it should be noted that the rates of pharmacological prophylaxis but not mechanical prophylaxis were significantly associated with the costs of septic shock caused by intestinal perforation ( p < 0.05). Although increased total rates of DVT prevention and the rates of mechanical prophylaxis did not reduce the mortality in patients with septic shock caused by intestinal perforation, the higher frequent intervention using pharmacological prophylaxis indicated the lower mortality of these patients ( p < 0.05). Conclusions DVT prevention by any means is a safe therapeutic strategy for treating septic shock caused by intestinal perforation, and pharmacological prophylaxis reduced the mortality of patients with septic shock caused by intestinal perforation.

Introduction: Fever of unknown origin (FUO) and hemodynamic instability are complications that develop after cardiac surgery combined with cardiopulmonary bypass (CPB) for heart disease. Patients who develop fever with hemodynamic instability after cardiac surgery may have systemic inflammatory response syndrome or sepsis. Cardiopulmonary bypass (CPB) is a technique that temporarily takes over the function of the heart and lungs during cardiac surgery. Recent reports suggest that early bloodstream infections of patients undergoing CPB are due to gram-negative bacteria that are present in the intestinal flora. The theory of intestinal flora translocation has growing evidence. Intestinal ischemia-reperfusion that occurs during cardiac surgery with CPB will induce a systemic inflammatory reaction and may cause intestinal flora translocation. Does this systemic reaction cause sepsis? We therefore propose this protocol to determine whether the changes in the intestinal flora in patients after cardiac surgery with CPB are related to sepsis. Methods and Analysis: This study is a prospective observational case–control study to analyze the variation in the intestinal microflora and metabolites in patients undergoing cardiac surgery with CPB and to observe the outcomes of patients with routine clinical interventions. The control group will include healthy people without intestinal illness. Feces and blood samples will be acquired 1 day before cardiac surgery and within 24–72 h after cardiac surgery, and will be used for genomics and metabolomics analyses. Demographic data describing age, sex, main diagnosis, and past medical history and data related to the CPB time and application of antibiotics are available. Sequential (sepsis-related) organ failure assessment, infection-related laboratory items, infection site, and pathogenic microorganisms, and nutrition, and gastrointestinal function assessment are additionally recorded. Group analysis of data will be conducted according to the outcomes (sepsis vs. non-sepsis and survivors vs. non-survivors). Ethics and Dissemination: This protocol has been ethically approved by the Ethics Committee of Peking Union Medical College (ID: ZS-1612). Informed consent will be obtained before subject enrolment, and data will be stored in a secured database. The results will be submitted to international peer-reviewed journals and presented at international conferences. Trial Registration Number: {"type":"clinical-trial","attrs":{"text":"NCT04032938","term_id":"NCT04032938"}} NCT04032938 .