312 Emergency Physician Accuracy in Estimating Volume Responsive Shock Using the “CURVES” Questionnaire

Abstract

Excess fluid administration and inadequate or delayed fluid resuscitation carry increased morbidity and mortality. Intensive care unit studies have shown that approximately 50% of patients given fluid boluses were truly volume responsive when compared to dynamic changes in cardiac output or stroke volume. In ED patients with undifferentiated shock, emergency physician accuracy in estimating volume responsiveness when compared to dynamic changes in cardiac output or stroke volume is unknown. To assess emergency physician accuracy in estimating volume responsiveness in patients with shock. This is a pilot study analyzing patients from a prospective randomized controlled trial of fluid optimization in the ED (“EFFORT”-ED Flow-directed Fluid Optimization Resuscitation Trial; ClinicalTrials.gov ID: NCT01128413). The study is IRB approved with informed consent. Enrollment began in August, 2010 and continues at an urban, academic tertiary care medical center with 90,000 ED visits/year. Inclusion criteria are adults (≥ 18 years) with either of the following: hypotension (SBP ≤90 or mean arterial pressure ≤65) despite 20ml/kg fluids; vasopressor use; or lactate ≥ 2.5 mmol/L. Patients are randomized to receive IV fluids via 1 of 2 arms: 1) routine care per clinician discretion 2) fluid optimization with 500ml normal saline boluses if the cardiac index or stroke volume index increases ≥ 15% with passive leg raise testing (PLRT). Patients are monitored until ED departure or a maximum of 6 hours using continuous noninvasive cardiac output monitoring (Cheetah NICOM®). During this time, passive leg raise testing is performed immediately before and after each fluid bolus or at 30-minute intervals if on a locked or maintenance fluid rate. With each passive leg raise testing, a CURVES (CardiovascUlar Response and Volume Estimation in Shock) questionnaire is administered separately to attending and resident physicians caring for the patients. Clinicians are blinded to the results of the passive leg raise testing and NICOM® monitoring. The CURVES questionnaire is a newly developed 5-item survey for this study which queries clinicians for the patient's potential shock etiology and estimations of volume responsiveness at that point in time (ie, responsive or unresponsive to further fluid administration), intravascular volume status, and cardiac index. With each questionnaire, clinicians are given the current vital signs, total volume administered, urine output, and the option for bedside re-examination, lab review, or the use of any other modality used in their routine care (ie, ultrasound, CVP, Scv02). Descriptive statistical analysis was performed using Microsoft® Excel. 11/11 per protocol patients were analyzed. The median age was 58 years (range 20-90) with 27% female. A total of 79 passive leg raise testings and 118 CURVES were acquired. 50/79 (63.3%) passive leg raise testing demonstrated volume responsiveness changes in cardiac index or stroke volume index. Attending emergency physician accuracy in estimating volume responsiveness was 68% with a sensitivity of 89% and specificity of 32%. Resident accuracy was 41% with a sensitivity of 45% and specificity of 35%. This pilot study suggests that emergency physician accuracy in estimating volume responsiveness in the setting of shock is poor when compared to passive leg raise testing changes in cardiac index or stroke volume. Routine clinical estimations of volume responsiveness not guided by dynamic changes in cardiac index or stroke volume index may lead to inadequate or excess fluid administration. The impact of ED fluid administration guided by dynamic changes in cardiac index or stroke volume index on morbidity and mortality is unknown.