Over 357,500 structure fires occur in the United States annually resulting in about 2,710 civilian deaths. The combustion of plastic, vinyl, wood, silk, and petroleum releases toxic hydrogen cyanide gas. Empiric treatment in the prehospital and hospital setting with hydroxocobalamin has been advocated. Without clear guidelines overuse can occur. The objective of this quality improvement initiative was to evaluate hydroxocobalamin’s use for inhaled cyanide toxicity after implementing a smart order-set and care pathway. This is a retrospective, pre (2021-2022) post-group (2023-2024) review of patients with inhalation injury admitted to one verified burn center. Specific criteria for the ordering of hydroxocobalamin for cyanide toxicity were implemented in 2022. Patients without smoke inhalation were excluded. Baseline demographics, laboratory values and medication administration records were evaluated. Analysis of data included: Chi square, Mann Whitney U test, and logistic regression. 120 patients were diagnosed with an inhalation injury and of those 74 patients received hydroxocobalamin prehospital or in the ED (32 patients pre-group, 42 post-group). 46% of the patients received the medication prior to arrival at the burn center. 11 vs 14 patients experienced cardiac arrest in the pre vs post group. 21 (66%) and 20 (48%) of patients in the pre vs post group did not meet criteria to receive the antidote. Administration after 2 hours was the most common reason for inappropriate use in both groups. There was no difference in the mortality rate between groups (p=0.448). There was no difference in hospital or ICU length of stays, or duration of mechanical ventilation. Based on regression analysis, CPR, facial burns, and age were significant predictors for mortality (p=0.002; p= 0.016; p=0.032). The risk of death was increased by 8 times if the patient experienced a cardiac arrest, 4 times if there were facial burns and for every year of age the risk increased by 1 time in patients that received hydroxocobalamin. Using a smart order-set, along with continued education, will improve compliance with appropriate treatment for suspected cyanide toxicity. The avoidance of inappropriate use of this expensive antidote provides cost savings to the health care system. Future studies to evaluate the effectiveness of using an evidence-based approach are needed. Smart order-sets can help guide prescribers in ordering and ensuring compliance with guidelines/protocols. N/A
Organizations frequently disseminate safety messaging to convey crucial injury prevention information. However, a systematic evaluation of these messages for accuracy, relevance, clarity, accessibility, and efficacy is often lacking or conducted with insufficient rigor. To address this gap, a national group of subject matter experts (SMEs) from diverse fields collaborated to develop an evidence-based methodology and simple user-friendly tool for evaluating burn prevention and fire safety messages. In 2021, eleven SMEs gathered under a FEMA-funded two-year grant project. Extensive reviews were conducted of medical, prevention, behavioral science, psychology, public health, and education literature to compile key findings and approaches. The identified components of evidence-informed effective messaging were utilized to create an online messaging reference guide and a 29-question self-evaluation tool (IMPAC-Tool). The online questionnaire-based format of IMPAC-Tool auto-scored the message being assessed and provided specific recommendations for improvement. The reference guide included the supporting evidence, examples, additional resources, and tips to assist users. Pilot testing was conducted via a national stakeholder focus group instructed to use IMPAC-Tool to evaluate safety messages currently embedded in their prevention programs. Testing results and feedback were utilized to revise and improve the tool. In the pilot test of key stakeholders (n=15), 90% of respondents stated the tool questions were helpful in evaluating and improving their safety messages. Based on feedback, minor revisions were made to the tool for clarity. A beta test was conducted via a large national community risk reduction network (n=56 organizations), with 13 providing focus group responses and 29 completing the tool evaluation. Half of the messages evaluated in IMPAC-Tool scored < 10% (< 2 points out of 17) in quality. Qualitative results showed IMPAC-Tool prompted them to rethink their message, revise it to be action-focused, change from “don’t” to “do” messaging, and incorporate special needs, cultural, and accessibility considerations. The free online IMPAC-Tool was publicly launched in Fall 2023, with an additional 45 responses from injury, trauma, and burn prevention fields. All injury and violence prevention programs can employ the evidence-based IMPAC-Tool to evaluate the efficacy, potential effectiveness, and specific areas for enhancement of their safety messages. This project provides a foundation for future studies to evaluate the effectiveness of evidence-based messaging in directly preventing injuries and deaths. Federal Emergency Management Agency (FEMA) Fire Prevention and Safety Grant EMW -2020-FP00029
Changes in drug metabolism, clearance, and volume of distribution pose unique challenges in effectively dosing antifungal agents in the burn patient. Therapeutic drug monitoring (TDM) can aid appropriate dosing despite. challenges with the timeliness of lab turnaround. This review evaluates dosing strategies, adverse events, MIC of fungal pathogens, and outcomes of fungal infections. A retrospective chart review of patients admitted to a burn center and treated with systemic antifungal agents from 2014 to 2024 was conducted. Baseline demographics, laboratory and culture data, medication administration records, drug-drug interactions, and rates of adverse events were evaluated. The primary outcomes evaluate microbiological cure, voriconazole and posaconazole levels, length of stay, and hospital mortality. Data analysis includes descriptive statistics. Ninety-five patients with culture-positive fungal infections were identified. The median (IQR) TBSA was 47(40) %. Voriconazole was used in seventy-six patients, posaconazole in eleven patients (salvage therapy), and fourteen were treated with a combination of either voriconazole/posaconazole with liposomal amphotericin B. (Table 1) Of the 50 patients with TDM levels the median time to result was 8 days. Seventy-two voriconazole patients received medications that were inhibitors or substrates of CYP2C19 which may have affected voriconazole levels. Patients in the voriconazole and posaconazole group were exposed to a median (IQR) of 5(2) and 4(3) antibiotic regimens prior to antifungal therapy. (Table 2) Less than 30% of the initial fungal isolates were susceptible to echinocandins or fluconazole. Most organisms were molds (including Aspergillus species, Fusarium, Mucor/Rhizopus species, and various dematiaceous molds). Microbiological cure rate was 31(40.8%) and 6(54.5%) in the voriconazole and posaconazole regimens. Overall hospital mortality was 44%. The standard 4 mg/kg dose of voriconazole resulted in subtherapeutic initial levels 15% of the times. Subsequent voriconazole levels during treatment were therapeutic only 50% of the time. About half of the voriconazole patients did not have levels drawn. The levels may have been affected by the PK/PD alterations in burn patients or variability of the metabolism by CYP2C19. The turnaround time for antifungal levels delayed adjusting medication in a timely manner. More studies are needed regarding burn fungal infections, pathogens, appropriate dosing strategies and outcomes associated with MIC of the various fungal pathogens to standard therapies. N/A
e23352 Background: Portal vein thrombosis (PVT)—the formation of a thrombus within the portal vein or its branches (superior mesenteric and splenic veins)—arises from local and systemic prothrombotic factors, including cirrhosis, hepatobiliary malignancies, myeloproliferative disorders, and inflammatory abdominal diseases. While PVT occurs in less than 1% of the general population, its prevalence rises to 20% in advanced liver disease, particularly among liver transplant candidates. Anticoagulation is central to PVT management, with warfarin traditionally used for prophylaxis and treatment. However, direct oral anticoagulants (DOACs) have garnered interest as a more convenient and potentially safer alternative. This study reviews the efficacy, safety, and practicality of warfarin versus DOACs for PVT prevention in cirrhotic patients, those with liver cancer, prior liver transplantation, or splenectomy, evaluating current evidence and clinical guidelines. Methods: Patients with NAFLD, liver cell carcinoma, liver transplant history, cirrhosis, thrombophilia, portal hypertension, splenectomy, or noninfective colitis were included. Two cohorts were analyzed: DOACs (apixaban/rivaroxaban) vs. warfarin. Propensity score matching balanced groups for age, sex, type 2 diabetes, hypertension, chronic kidney disease, and atrial fibrillation. The primary outcome was PVT incidence at 6 months, 1 year, and 5 years. Patients with prior PVT were excluded. Results: At 6 months, 620 warfarin-treated and 525 DOAC-treated patients developed PVT. The risk ratio was 1.191 (95% CI: 1.06–1.337), and mean PVT incidence was higher in the warfarin group (4.202) than the DOAC group (2.72) (p < 0.0001). At 1 year, 686 warfarin-treated vs. 607 DOAC-treated patients developed PVT (risk ratio: 1.14, 95% CI: 1.022–1.271). Mean incidence remained higher in warfarin users (5.211) vs. DOAC users (3.277) (p < 0.0001). At 5 years, 893 warfarin-treated vs. 867 DOAC-treated patients developed PVT (risk ratio: 1.039, 95% CI: 0.947–1.14). Mean incidence was higher in warfarin users (7.572) vs. DOAC users (4.569) (p < 0.0001). Conclusions: DOACs showed superior efficacy in preventing PVT over 6 months and 1 year, with fewer complications, reduced bleeding risks, and no need for routine monitoring, offering practical advantages. These findings support the need for updated clinical guidelines to reflect the evolving role of DOACs in PVT prophylaxis. Time Point Incidence Rate per 100 PY (Warfarin, 95% CI) Incidence Rate per 100 PY (DOAC, 95% CI) Hazard Ratio (95% CI, p-Value) 6 Months 5.1 (4.7-5.5) 3.9 (3.5-4.3) 1.22 (1.08-1.37, 0.001) 1 Year 4.8 (4.4-5.2) 3.7 (3.3-4.1) 1.16 (1.04-1.30, 0.003) 5 Years 3.2 (2.9-3.5) 2.9 (2.6-3.2) 1.07 (0.97-1.17, 0.169)
Large language models (LLMs) are artificial intelligence tools that are trained to summarize, translate text, and respond to user questions that have grown in popularity since their release. While they have great potential to improve access to medical knowledge, there is also potential to spread misinformation due to lack of adequate accountability and review. Therefore, we sought to investigate the ability of LLMs to accurately and thoroughly answer basic burn wound care questions. We created a standardized list of prompts covering basic at-home burn first aid questions. This list encompassed what steps to take at home after a burn, what to avoid, and how to assess severity of burn and determine when one should seek the care of a doctor. These questions were then input into six different popular LLMs: ChatGPT, Claude, Perplexity, Copilot, Jasper, and Gemini. The responses were then collected and compiled by our team for review. Each set of responses was directly compared to the American Burn Association’s Recommendations for Initial First Aid Treatment of Minor Burns for accuracy (not containing any contradictory information to published guidelines) and thoroughness (containing all aspects of published guidelines). While there were variations in the length and style of answers amongst the LLMs, all models responded with accurate description of burn depth and advice that aligned with the tenets of first aid treatment for minor burns including cooling the burn with cool but not cold water and avoiding ice, covering the burn with sterile bandage, and avoiding butter and home remedies. Only one LLM (ChatGPT) did not specify removing jewelry, rings, and clothing. All LLMs also gave accurate recommendations regarding when to seek medical attention, with only one (Gemini) not including burns over major joints. All LLMs recommended seeking advice from a healthcare provider with each prompt. References were given in only two LLMs (Copilot and Perplexity). The future of LLMs continues to develop but their growing use by both patients and medical providers is evident. Of the included LLMs, all provided data consistent with first aid guidelines for care of minor wounds without major inconsistencies or misinformation. While to be used with caution, the potential for LLMs to facilitate access to medical knowledge is great. The integration of LLMs into medical practice is inevitable. We must continue to investigate the applicability and accuracy as both providers and patients increasingly utilize them. N/A
