Is there opportunity for automated decision-support and closed-loop control in ICU patients receiving vasopressor infusion?

Abstract

Vasopressors are administered to critically ill patients suffering from a body-wide reduction in blood circulation. In theory, if the vasopressor infusion is either too high or too low, it could be harmful to the patient. In a retrospective analysis, we investigated the degree to which today's intensive care unit (ICU) patients receive appropriate vasopressor therapy, in terms of how often the mean arterial pressure (MAP) was kept within a normative range. Using the MIMIC II database, we studied patients with minute-by-minute MAP data, sourced from the bedside monitor, who were receiving vasopressor therapy. For each record, we identified MAP samples that were out-of-range, i.e., MAP <; 60 mmHg or MAP > 100 mmHg, and grouped these into out-of-range episodes. Each out-of-range episode was categorized as either transient (<; 15 min) or sustained (≥ 15 min). Out of the 224 ICU stays, we identified 152 ICU stays (68% of ICU stays) with at least one sustained MAP out-of-range episode. In that subset, MAP was frequently out-of-range (out-of-range 18.4% of the time) due to a combination of sustained episodes of hypotension and hypertension. Compared with all ICU stays, those stays with sustained out-of-range events did not demonstrate an increased MAP variability per hour. It is possible that the out-of-range events resulted from insufficient dose-adjustment. Technologies that might continuously optimize vasopressor dosing throughout the patient's stay and thereby minimize these abnormal cardiovascular states may be worthy of further study.