Multicenter study of noninvasive monitoring systems as alternatives to invasive monitoring of acutely ill emergency patients.

ABSTRACT

BACKGROUND: Recent reports showed lack of effectiveness of pulmonary artery catheterization in critically ill medical patients and relatively late-stage surgical patients with organ failure. Since invasive monitoring requires critical care environments, the early hemodynamic patterns may have been missed. Ideally, early noninvasive hemodynamic monitoring systems, if reliable, could be used as the "front end" of invasive monitoring to supply more complete descriptions of circulatory pathophysiology. OBJECTIVES: To evaluate the accuracy and reliability of noninvasive hemodynamic monitoring consisting of a new bioimpedance method for estimating cardiac output combined with arterial BP, pulse oximetry, and transcutaneous PO2 and PCO2; we compared this system of noninvasive monitoring with simultaneous invasive measurements to evaluate circulatory deficiencies in acutely ill patients shortly after hospital admission where invasive monitoring was not readily available. We also preliminarily explored early differences in temporal hemodynamic patterns of survivors and nonsurvivors. DESIGN AND SETTING: Prospective comparison of simultaneous invasive and noninvasive measurements of circulatory function with retrospective analysis of data in university-run county hospitals, university hospitals and affiliated teaching hospitals, and a community private hospital. PATIENTS We studied 680 patients, including 139 severely injured or hemorrhaging patients in the emergency department (ED), 129 medical (nontrauma) patients on admission to the ED, 274 high-risk surgical patients intraoperatively, and 138 patients recently admitted to the ICU. RESULTS: A new noninvasive impedance device provided cardiac output estimations under conditions in which invasive thermodilution measurements were not usually applied. There were 2,192 simultaneous bioimpedance and thermodilution cardiac index measurements; the correlation coefficient, r = 0.85, r2 = 0.73, p < 0.001. The precision and bias was -0.124+/-0.75 L/min/m2. Both invasive and noninvasive monitoring systems provide similar information and identified episodes of hypotension, low cardiac index, arterial hemoglobin desaturation, low transcutaneous O2, high transcutaneous CO2, and low oxygen consumption before and during initial resuscitation. The limitations of noninvasive systems were described. CONCLUSIONS: Noninvasive monitoring systems gave continuous displays of physiologic data that provided information allowing early recognition of low flow and poor tissue perfusion that were more pronounced in the nonsurvivors. Noninvasive systems may be acceptable alternatives where invasive monitoring is not available.