Non-invasive stroke volume measurement during acute burn fluid resuscitation [abstract #132]
- View All
Introduction: Effective fluid resuscitation is one of the cornerstones of modern burn care, but determining optimal resuscitation can be challenging. Traditional endpoints used have been urine output (UO) and mean blood pressure (MAP). Studies using cardiac output and index, oxygen delivery and consumption, and base deficit and lactate have resulted in increased IV fluid (IVF) administration without improvements in burn outcomes. Changes in Stroke Volume Index (SVI) and Stoke Volume Variation (SVV) in the hypovolemic individual (SVI < 33 and a SVV> 12%) have been shown to predict fluid responsiveness in some patient populations but has not yet been studied in burn patients. The purpose of this study was to determine if SVI and SVV, measured by a non-invasive cardiac output monitor (NICOM), correlates with traditional endpoints to predict the need for increased IVF during resuscitation. Methods: This was an IRB approved, prospective, blinded, observational study of patients >= 18 years old with >20% TBSA burns. The NICOM device was applied during the acute resuscitation and hemodynamic data were recorded every minute, stored on a memory device, and analyzed using standard Receiver Operating Curve (ROC) methods. Our standardized Nurse Driven Resuscitation Protocol, which titrates IVF based on hourly UO, MAP, and central venous pressure was used. Median SVI and SVV measurements in the hour prior to a patient recording of low UO were analyzed to determine their ability to predict such episodes. A ROC curve was plotted with confidence bounds, as well as the area under the curve (AUC) and its confidence interval. Potential thresholds for SVI and SVV were assessed for sensitivity and specificity. Results: This study included 20 patient resuscitations. In spite of chest wall burns, we were able to successfully obtain measurements from the skin sensors on all patients. SVI ROC analysis estimated an AUC of 67.6% (95% CI: 61.3%-73.8%), representing a moderate-low ability to predict an upcoming low UO. SVI values between 22.5 and 28.5 had both sensitivity and specificity >50%, with specificity reaching 75% for an SVI of 22. The SVV ROC analysis showed a low ability to predict low urine output (AUC=60.3% [53.7%-66.9%]. Conclusions: SVI below 33 does not predict need for increased IVF in the burn patient. A SVI < 22 would intervene early in half, while overtreating a quarter of the patients. Despite evidence in the literature to support the use of SVV in fluid resuscitation, SVV was not able to predict patients with low upcoming urine outputs as well as SVI in this patient population.