Prediction of Mortality after Burn Surgery in Critically Ill Burn Patients Using Machine Learning Models

Association of inflammatory biomarkers with overall survival in burn patients: a systematic review and meta-analysis

INTRODUCTION

The inflammatory response to burn injuries can lead to organ dysfunction that ultimately results in increased mortality and morbidity. This meta-analysis was conducted to determine the efficacy of inflammatory biomarkers, including the neutrophil to lymphocyte ratio (NLR), platelet to lymphocyte ratio (PLR), procalcitonin (PCT), and C-reactive protein (CRP) as predictive tools of mortality among burn patients.

MATERIAL AND METHODS

The biomarker levels of survivors and non-survivors were consolidated according to guidelines for Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Three main databases were searched electronically: PubMed, Web of Science, and Scopus, on December 8, 2022. The Newcastle-Ottawa Quality Assessment Scale (NOS) was used to evaluate and score the methodological quality of the included studies. The standard mean difference (SMD) with a 95% confidence interval (CI) was utilized.

RESULTS

Twenty-four studies were included in our systematic review and meta-analysis, (3636 total burn patients), of whom 2878 survived. We found that deceased burn patients had elevated levels of NLR (SMD = 0.60, 95% CI; 0.19-1.00, P < 0.001), CRP (SMD = 0.80, 95% CI; 0.02-1.58, P = 0.04), and PCT (SMD = 0.85, 95% CI; 0.45-1.24, P < 0.001), compared to survivors. However, we found no association between PLR and mortality among burn patients (SMD = 0.00, 95% CI; -0.14-0.15, P < 0.001). In addition, CRP was significantly higher in non-survivors (SMD = 0.80, 95% CI; 0.02-1.58, P =0.04). Similar results were also found about PCT (SMD = 0.85, 95% CI; 0.45-1.24, P < 0.001). When we analyzed the PCT data, collected in the first 24-48 hours, we found similar results; the PCT level was significantly higher in non-survivors in the immediate postinjury-period (SMD = 0.67, 95% CI; 0.31-1.02, P < 0.001). There was no publication bias among studies on the role of NLR in burn (Egger's test P = 0.91). The based cut-off values for NLR (13), CRP (71), and PCT (1.77) yielded sensitivities of 69.2%, 100%, and 93.33%, and specificities of 76%, 72.22%, and 72.22% respectively.

DISCUSSION/CONCLUSIONS

PCT is a marker of sepsis, therefore its elevated level is presumably associated with a higher incidence and severity of sepsis among non-survivors. In addition, NLR and CRP are promising biomarkers for predicting and guiding prevention against burn deaths in clinical settings.

Machine learning models for early prediction of mortality risk in patients with burns: A single center experience

Mortality rate is considered as the most important outcome measure for assessing the severity of burn injury. A scale or model that accurately predicts burn mortality can be useful to determine the clinical course of burn injuries, discuss treatment options and rehabilitation with patients and their families, and evaluate novel, innovative interventions for the injuries. This study aimed to use machine learning models to predict the mortality risk of patients with burns after their first admission to the center and to compare the performances of these models. Overall, 1064 patients hospitalized in burn intensive care and burn service units between 2016 and 2022 were included in the study. In total, 40 parameters, including demographic characteristics and biochemical parameters of all patients, were analyzed in the study. Furthermore, the dataset was randomly divided into two clusters with 70% of the data used for artificial neural networks (ANNs) training and 30% for model success testing. The ANN model proposed in this study showed high success across all machine learning methods tried in different variants, with an accuracy of 95.92% in the test set. Machine learning models can be used to predict the mortality risk of patients with burns. This study may help validate the use of machine learning models for applications in clinical practice. Conducting multicenter studies will further contribute to the literature.

Feasibility of using red cell distribution width for prediction of postoperative mortality in severe burn patients: an association with acute kidney injury after surgery

BACKGROUND

Severe burns cause pathophysiological processes that result in mortality. A laboratory biomarker, red cell distribution width (RDW), is known as a predictor of mortality in critically-ill patients. We examined the association between RDW and postoperative mortality in severe burn patients.

METHODS

We retrospectively analyzed medical data of 731 severely burned patients who underwent surgery under general anesthesia. We evaluated whether preoperative RDW value can predict 3-month mortality after burn surgery using receiver operating characteristic (ROC) curve analysis, logistic regression, and Cox proportional-hazards regression analysis. Mortality was also analyzed according to preoperative RDW values and incidence of postoperative acute kidney injury (AKI).

RESULTS

The 3-month mortality rate after burn surgery was 27.1% (198/731). The area under the ROC curve of preoperative RDW to predict mortality after burn surgery was 0.701 (95% confidence interval [CI], 0.667-0.734; P < 0.001) with a cut-off point of 12.9. The adjusted hazard ratio in patients with RDW > 12.9 was 1.238 (95% CI, 1.138-1.347; P < 0.001). Subgroup analysis showed that the survival rate was 88.8% for the non-AKI group with RDW ≤ 12.9 and 17.6% for the AKI group with RDW > 12.9. Preoperative RDW was considered an independent risk factor for mortality (odds ratio, 1.679; 95% CI, 1.378- 2.046; P < 0.001).

CONCLUSIONS

Preoperative RDW may predict 3-month postoperative mortality in patients with severe burns, while preoperative RDW > 12.9 and postoperative AKI may further increase mortality after burn surgery.