Objective Dying in the intensive care unit (ICU) has changed over the last twenty years due to increased utilization of palliative care. We sought to examine how palliative medicine (PM) integration into critical care medicine has changed outcomes in end of life including the utilization of do not resuscitate (no cardiopulmonary resuscitation but continue treatment) and comfort care orders (No resuscitation, only comfort medication). Design: Retrospective observational review of critical care patients who died during admission between two decades, 2008 to 09 and 2018 to 19. Setting: Single urban tertiary care academic medical center in Washington, D.C. Patients: Adult patients who were treated in any ICU during the admission which they died. Interventions and Measurements We sought to measure PM involvement across the two decades and its association with end of life care including do not resuscitate (DNR) and comfort care (CC) orders. Main Results: 571 cases were analyzed. Mean age was 65 ± 15, 46% were female. In univariate analysis significantly more patients received PM in 2018 to 19 (40% vs. 27%, p = .002). DNR status increased significantly over time (74% to 84%, p = .002) and was significantly more common in patients who were receiving PM (96% vs. 72%, p < 0.001). CC also increased over time (56% to 70%, p = <0.001), and was more common in PM patients (87% vs. 53%, p < 0.001). Death in the ICU decreased significantly over time (94% to 86%, p = .002) and was significantly lower in PM patients (76% vs. 96%, p < 0.001). The adjusted odds of getting CC for those receiving versus those not receiving PM were 14.51 (5.49-38.36, p < 0.001) in 2008 to 09 versus 3.89 (2.27-6.68, p < 0.001) in 2018 to 19. Conclusion: PM involvement increased significantly across a decade in our ICU and was significantly associated with incidence of DNR and CC orders as well as the decreased incidence of dying in the ICU. The increase in DNR and CC orders independent of PM over the past decade reflect intensivists delivering PM services.

Background Thromboelastography (TEG) can guide transfusion therapy in trauma and has been associated with decreased transfusion requirements. This population differs from the medical population where the most common bleeding source is gastrointestinal hemorrhage (GIB). The utility of TEG in patients with acute GIB is not well described. We sought to assess whether the use of TEG impacts blood product utilization in patients with medical GIB. Methods A retrospective study looking at all adult patients admitted with a primary diagnosis of GIB to the George Washington University Intensive Care Unit (ICU) between 01/01/2017 to 12/31/2019. The primary intervention was the use of TEG to guide blood product resuscitation in addition to standard of care (TEG arm) versus standard of care alone (non-TEG arm). Results The primary outcome was the total number of blood products utilized. Patients in the TEG arm used more blood products compared to the non-TEG arm (9.10 vs 3.60, p < 0.001). There was no difference in secondary endpoints except for an increased requirement for mechanical ventilation within the TEG arm (26.2% vs 13.4%, p  =  0.018). Conclusions The use of TEG to guide resuscitation in patients with acute GIB may be associated with increased blood product utilization without any clinical benefit to patient-centered outcomes.

No abstract

Objective To describe the sedation, narcotic and neuromuscular blockade usage in ventilated patients with COVID-19 pneumonia. Design Single-Center Retrospective Review. Setting George Washington University Hospital in Washington, D.C. Patients 62 patients with COVID-19 respiratory failure requiring mechanical ventilation admitted from March 2020 to June 2020. Intervention None. Measurements and main results Patients with COVID-19 respiratory failure required multiple sedative/narcotic infusions to achieve sedation requirements and at doses that were significantly more when compared to a general medical-surgical ICU population (represented by the MIND-USA cohort). The most common infusions were Dexmedetomadine and Propofol. Approximately 17% of our patients required a neuromuscular blockade infusion as well. Prior to intubation, narcotic utilization was stable and low. Conclusion Patients with COVID-19 respiratory failure requiring mechanical ventilation have higher sedation and narcotic requirements than general ICU patients.