16 Clinical Outcomes in Patients with Diabetes, Prediabetes, and No Diabetes Following Acute Burn Injury

Abstract

Abstract Introduction Based on the National Burn Repository Data (2002–2011), the incidence of diabetes is increasing among adult patients with burn injury. Diabetes has been associated with increased morbidity and mortality following burn injury. There is limited data related to prediabetes in burn injury, and no studies to date have compared clinical outcomes in patients without diabetes, with prediabetes, and with diabetes. Therefore, this study aimed to compare the impact of diabetes status, defined as no diabetes (ND), prediabetes (PD), or diabetes (DM), on clinical outcomes after burn injury. Methods A single-center, retrospective cohort study of adult patients admitted for initial management of burn injury between January 1, 2009 and February 28, 2019 who had a hemoglobin A1c (HbA1c) upon admission was performed. Patients with chemical or electrical injury, isolated smoke inhalation injury, total body surface area (TBSA) burn injury < 1%, patients who expired or transitioned to comfort care within 24 hours of admission, or were incarcerated or pregnant were excluded. Patients were categorized as ND, PD, or DM based on their past medical history and admission HbA1c. The primary outcome was LOS/%TBSA burn. Secondary outcomes included mortality, LOS, discharge disposition, re-grafting of same site, and amputations. Propensity score methods were used to account for imbalance between groups. Separate propensity score weights were calculated for analysis of DM versus ND and DM versus PD. The primary outcome LOS/%TBSA burn was compared between groups using weighted linear models. Since the outcome was right-skewed, LOS/%TBSA burn was log-transformed for analysis, yielding estimates of the ratio of geometric means. Results A total of 1137 patients were included (710 ND, 191 PD, 236 DM). Patients were predominately male (70.3% ND, 67.4% PD, 70.2% DM) and Caucasian (85.5% ND, 73.8% PD, 82.1% DM). The mean BMI was highest in DM (31.8 kg/m2) and lowest in ND (27.8 kg/m2). After inverse probability weighing to adjust for potentially confounding factors, DM had longer LOS/%TBSA burn than patients with ND (ratio of geometric means (95% CI) = 1.61 (1.21, 2.15), p = 0.001) or PD (ratio (95% CI) = 1.53 (1.14, 2.05), p =0.01). No differences in secondary outcomes were observed between the DM and ND groups and between DM and PD groups with the exception of amputations (2.8% DM vs. 0.7% ND, p=0.06; 2.6% DM vs. 0% PD, p=0.04). Conclusions Diabetes in the burn patient significantly increases LOS/%TBSA index and amputation rates compared to both PD and ND. Applicability of Research to Practice Diabetes may contribute to increased cost of care in burn injury. Early identification of diabetes status may aid management of this patient population.