MADRID – In patients managed on mechanical ventilation in an intensive care unit following cardiac surgery, a fully automated system provides more reliable ventilatory support than highly experienced ICU nurses, suggest results of a randomized trial.
The study’s control group received usual care, which means that nurses adjusted mechanical ventilation manually in response to respiratory rate, tidal volume, positive end-respiratory pressure (PEEP), and other factors to maintain ventilation within parameters associated with safe respiration. The experimental group was managed with a fully automated closed-loop system to make these adjustments without any nurse intervention.
For those in the experimental group “the proportion of time in the optimal zone was increased and the proportion of time in the unsafe zone was decreased” relative to those randomized to conventional nursing care, Marcus J. Schultz, MD, reported at the annual congress of the European Respiratory Society.
Conducted at a hospital with an experienced ICU staff, the study had a control arm that was managed by “dedicated nurses who, I can tell you, are very eager to provide the best level of care possible,” said Dr. Schultz, professor of experimental intensive care, University of Amsterdam, the Netherlands..
The investigator-initiated POSITiVE trial randomized 220 cardiac surgery patients scheduled to receive postoperative mechanical ventilation in the ICU. Exclusions included those with class III or higher chronic obstructive pulmonary disease (COPD), a requirement for extracorporeal membrane oxygenation (ECMO), or a history of lung surgery.
The primary endpoint was the proportion of time spent in an optimal zone, an acceptable zone, or a dangerous zone of ventilation based on predefined values for tidal volume, maximum airway pressure, end-tidal CO2, and oxygen saturation (SpO2).
The greatest between-group difference was seen in the proportion of time spent in the optimal zone. This climbed from approximately 35% in the control arm to slightly more than 70% in the experimental arm, a significant difference. The proportion of time in the dangerous zone was reduced from approximately 6% in the control arm to 3% in the automated arm. On average nurse-managed patients spent nearly 60% of the time in the acceptable zone versus less than 30% of those in the automated experimental arm.
A heat map using green, yellow, and red to represent optimal, acceptable, and dangerous zones, respectively, for individual participants in the trial provided a more stark global impression. For the control group, the heat map was primarily yellow with scattered dashes of green and red. For the experimental group, the map was primarily green with dashes of yellow and a much smaller number of red dashes relative to the control group.
In addition, the time to spontaneous breathing was 38% shorter for those randomized to automated ventilation than to conventional care, a significant difference.
There are now many devices marketed for automated ventilation, according to Dr. Schultz. The device used in this study was the proprietary INTELLiVENT-ASV system, marketed by Hamilton Medical, which was selected based on prior satisfactory experience. Although not unique, this system has sophisticated software to adjust ventilation to reach targets set by the clinician on the basis of information it is receiving from physiologic sensors for such variables as respiratory rate, tidal volume, and inspiratory pressure.
“It is frequently adjusting the PEEP levels to reach the lowest driving pressure,” said Dr. Schultz. Among its many other features, it also “gives spontaneous breathing trials automatically.”
Uncomplicated patients were selected purposefully to test this system, but Dr. Schultz said that a second trial, called POSITiVE 2, is now being planned that will enroll more complex patients. Keeping complex patients within the optimal zone as defined by tidal volume and other critical variables has the potential to reduce the lung damage that is known to occur when these are not optimized.
“Applying safe ventilatory support in clinical practice remains a serious challenge and is extremely time consuming,” Dr. Schultz said. He reported that fully automated ventilation appears to be reliable, and “it takes out the human factor” in regard to diligence in monitoring and potential for error.
Overall, these results support the potential for a fully automated system to improve optimal ventilatory support, reduce risk of lung injury, and reduce staffing required for monitoring of mechanical ventilation, according to Dr. Schultz.
Relative costs were not evaluated in this analysis, but might be another factor relevant to the value of fully automated ventilation in ICU patients.