Predicting Outcomes in Trauma Patients With Cirrhosis Using Model for End-Stage Liver Disease Score: A Retrospective Study.

BACKGROUND: Liver cirrhosis is associated with increased mortality in trauma victims. We stratified the impact of cirrhosis on trauma mortality by Model for End-stage Liver Disease (MELD) score. METHODS: Trauma center database was accessed for patients with established diagnosis of cirrhosis presenting 2014 - 2018, matched to control patients without cirrhosis in a 2:1 ratio by age, sex, and TRISS. Primary outcome was mortality, secondary outcomes were length of stay, intensive care unit days, and ventilator dependent days. RESULTS: Cirrhosis was present in 182 (1.5%) trauma patients. Mortality difference between 12 (7%) deaths in cirrhosis cohort versus 14 (4%) in control was not statistically significant (p = 0.38). No difference was found in secondary outcomes. Categorization of cirrhosis severity by MELD score range (MELD 6-7, 8-10, 11-14, 15-20, 21-30) showed a 1.9 fold increase in the odds of mortality for every increase in MELD score category (OR = 1.91, p = 0.03, 95% CI = 1.08 - 3.37). CONCLUSION: Mortality effects of cirrhosis in trauma patients can be estimated using MELD score.

Variations in insulin requirements can be an early indicator of sepsis in burn patients.

INTRODUCTION: A >25% increase in daily insulin dosing is suggestive of possible sepsis in burn patients, however, no conclusive evidence is available regarding the time point at which insulin dosing begins to increase. The purpose of this study is to determine the exact time point at which the insulin requirement increases among non-diabetic burn patients with sepsis. METHODS: A retrospective chart review in non-diabetic burn patients with ≥20% total body surface area burned (TBSA) during 2010-2018 who received a blood culture for suspected sepsis. Absolute insulin dosing at intervals (0, 24, 48, 72, and 96 h prior to blood culture) were Box-Cox transformed and compared vs.-96 h reference using mixed-effects models accounting for within-patient dependencies. RESULTS: Fifty-eight patients (84% males, age 44 ± 17 years, TBSA% 49 ± 17.5) were included. When cube root of daily insulin dosing was regressed on each time point in a mixed-effects model, statistically significant increase in insulin dosing compared to baseline was observed for -48 (p = 0.018), -24 (p = 0.011), and 0 h (p = 0.008). CONCLUSION: Daily insulin dosing increases 48 h prior to development of other clinical signs of sepsis and can be used as a sensitive early marker.