Performance of BEAMS risk of death score for mortality prediction in Australian and New Zealand burns patients

Association of creatinine-albumin ratio with 28-day mortality in major burned patients: A retrospective cohort study

BACKGROUND

Serum creatinine (Cr) and Albumin (Alb) have emerged as prognostic factors for mortality in many diseases including burned patients. However, few studies report the relationship between Cr/Alb ratio and major burned patients. The purpose of this study is to make evaluation of efficacy of Cr/Alb ratio in predicting 28-day mortality in major burned patients.

METHOD

Based on a local largest tertiary hospital in South of China, we retrospectively analyzed data of 174 patients with total burn area surface (TBSA) ≥ 30% from January 2010 to December 2022. Receiver operating characteristic curve (ROC), logistic analysis, and Kaplan-Meier analysis were performed to evaluate the association between Cr/Alb ratio and 28-day mortality. Integrated discrimination improvement (IDI), and net reclassification improvement (NRI) were used to estimate the improvements in new model performance.

RESULTS

28-day mortality rate was 13.2% (23/174) in burned patients. Cr/Alb on admission at level of 3.340μmol/g showed the best discrimination between survivors and non-survivors after admission at 28 days. The result of multivariate logistic analysis suggested that age (OR, 1.058 [95%CI 1.016-1.102]; p = 0.006), higher FTSA (OR, 1.036 [95%CI 1.010-1.062]; p = 0.006), and higher level of Cr/Alb ratio (OR, 6.923 [95CI% 1.743-27.498]; p = 0.006) were independently associated with 28 day-mortality. A regression model was constructed by logit(p) = 0.057 *Age + 0.035 *FTBA + 1.935 * Cr/Alb - 6.822. The model showed a better discrimination and risk reclassification compared with ABSI and rBaux score.

CONCLUSIONS

High Cr/Alb ratio at admission is a herald of poor outcome. The model generated from multivariate analysis could serve as an alternative prediction tool among major burned patients.

Improvement of Predictive Scores in Burn Medicine through Different Machine Learning Approaches

The mortality of severely burned patients can be predicted by multiple scores which have been created over the last decades. As the treatment of burn injuries and intensive care management have improved immensely over the last years, former prediction scores seem to be losing accuracy in predicting survival. Therefore, various modifications of existing scores have been established and innovative scores have been introduced. In this study, we used data from the German Burn Registry and analyzed them regarding patient mortality using different methods of machine learning. We used Classification and Regression Trees (CARTs), random forests, XGBoost, and logistic regression regarding predictive features for patient mortality. Analyzing the data of 1401 patients via machine learning, the factors of full-thickness burns, patient's age, and total burned surface area could be identified as the most important features regarding the prediction of patient mortality following burn trauma. Although the different methods identified similar aspects, application of machine learning shows that more data are necessary for a valid analysis. In the future, the usage of machine learning can contribute to the development of an innovative and precise predictive score in burn medicine and even to further interpretations of relevant data regarding different forms of outcome from the German Burn registry.

Comparative Analysis of Composite Mortality Prediction Scores in Intensive Care Burn Patients

Multiple outcome scoring models have been used in predicting mortality in burn patients. In this study, we compared the accuracy of five established models in predicting outcomes in burn patients admitted to the intensive care unit and assessed risk factors associated with mortality. Intensive care burn patients admitted between March 2007 and December 2020 with total body surface area (TBSA) affected ≥ 10% were analyzed. Multivariate analysis was conducted to examine variables associated with mortality. The ABSI, Ryan, BOBI, revised Baux and BUMP scores were analyzed by receiver operating characteristics. A total of 617 patients were included. Morality was 14.4%, with non-survivors being significantly older, male, and having experienced domestic burns. Multivariate analysis identified age, TBSA, full-thickness burns and renal insufficiency as independent mortality predictors. The BUMP score presented the highest mortality prognostication rate, followed by ABSI, revised Baux, BOBI and Ryan scores. BUMP, ABSI and revised Baux scores displayed AUC values exceeding 90%, indicating excellent prognostic capabilities. The BUMP score showed the highest accuracy of predicting mortality in intensive care burn patients and outperformed the most commonly used ABSI score in our cohort. The older models displayed adequate predictive performance and accuracy compared with the newest model.