10 A Multi-center Study Analyzing Association of Vitamin D Deficiency and Replacement with Infectious Outcomes in Patients with Burn Injuries

Abstract

Abstract Introduction Vitamin D (25OHD) deficiency has been associated with poor outcomes in intensive care populations. A recent single-center, burn study found a high incidence of 25OHD deficiency. A difference was noted in infectious complications, but was underpowered. The primary objective of this multi-center study was to determine if 25OHD deficiency is associated with infectious outcomes in adult burn patients. Methods Adult patients were eligible for inclusion in this 7 center, retrospective study if admitted January 1, 2016 - July 25, 2019 and had a 25OHD concentration drawn within the first 7 days of admission. Patients were excluded if admitted for a non-burn injury, had total body surface area (TBSA) burned of less than 5%, a readmission, pregnant, incarcerated, or made comfort care or expired within 48 hours of admission. Expecting a 3:1 enrollment, goal was at least 250 total patients to be appropriately powered (β = 0.2; α = 0.05) to detect a 33% difference in composite infectious outcome (bacteremia, pneumonia, urinary tract infection, wound infection, graft loss, or death) between patients with 25OHD deficiency (< 20 ng/mL) and control (≥ 20 ng/mL). Generalized linear mixed modelling was used to control for center effect, % TBSA, age, and presence of inhalation injury and find the most predictive model. Results A total of 1147 patients were initially included. After exclusions, 234 (56.8%) in the deficient and 178 in the control group remained. Patients in the control group had their concentration drawn earlier (p < 0.001), were more likely to be male (p = 0.006), Caucasian (p < 0.001), lower body mass index (p = 0.009), lower % TBSA burn (p = 0.002), and taking a 25OHD supplement prior to admission (p < 0.001). Deficient patients were more likely to have an infectious outcome (52.1% vs 36.0%, p = 0.002), acute kidney injury requiring renal replacement therapy (p = 0.009), less ventilator free days in the first 28 days (p < 0.001), and more days requiring vasopressors (p = 0.008). After controlling for center, % TBSA, age, and inhalation injury the best model also included presence of deficiency (odds ratio = 2.425 [1.035 - 1.252]), days until 25OHD supplement initiation (1.139 [1.035 - 1.252]), and choice of cholecalciferol over ergocalciferol 2.112 [1.151 - 3.877]). Conclusions Dilution concerns were controlled by including %TBSA in the regression model. Even if low 25OHD concentrations were an acute reaction to burn injury and not representative of true deficiency, low concentrations and delay in supplementation were independently associated with increased risk of an infectious outcome.