© 2003 American Thoracic Society
Ventilatory Assist Driven by Patient DemandDepartment of Critical Care St. Michael's Hospital Toronto, Ontario, Canada One of the major aims of mechanically ventilating spontaneously breathing patients is to satisfy the patient's respiratory demand, both in terms of timing (initiation and termination of the assist) and of delivering adequate levels of assist throughout each breath (1). Therefore, by definition, modes of mechanical ventilation that deliver fixed tidal volumes, flow rates, and/or pressures cannot be considered "demand-controlled" ventilator assistance, and are prone to interfering with the patient's own respiratory control (2). In 1970, Huszczuk (3) published an article entitled "A Respiratory pump controlled by phrenic nerve activity," describing a device used to keep airway pressure proportional to a command signal. This type of respiratory demand-driven ventilator was useful in studies on control of breathing, but because it was so invasive it was limited to studies in animals. In 1992, Younes (4) described proportional-assist ventilation, in which the ventilator amplifies patient instantaneous effort throughout inspiration while leaving the patient with full control of breathing pattern. Proportional-assist ventilation is based on an elegant method of predicting respiratory load using the equation of motion, provided that flow, volume, and respiratory system elastance and resistance are adequately measured. In 1999, a complementary system using diaphragmatic electrical activity was described for neurally adjusted ventilatory assist (5). Literature suggests that these modes of demand-driven ventilator assistance can unload the respiratory muscles, improve patient–ventilator synchrony, allow a variable breathing pattern, and improve patient comfort, vis-à-vis pressure support ventilation (6). In this issue of AJRCCM (pp. 760–769), Sharshar and coworkers (7) describe for the first time a mechanical ventilator system that triggers and delivers ventilator assistance in proportion to the transdiaphragmatic pressure, the latter obtained via sensors located in the esophagus and stomach. In terms of "proximity" to the respiratory centers, this system uses a control signal that is more proximal than flow/volume signals (and thus is not affected by respiratory system mechanics) and more distal than diaphragmatic activation (and thus is affected by respiratory muscle function) (5). Sharshar and coworkers (7) describe their system "as an external inspiratory muscle driven by the diaphragmatic mechanical output." This statement nicely summarizes the features of the new generation of ventilators, in which the caregiver can control the work share between the respiratory muscles and the ventilator output, without the ability to control delivered pressure, volume, or respiratory rate. As evidence of how well transdiaphragmatic pressure-controlled ventilator assistance is integrated with neural reflexes, Sharshar and coworkers (7) show that, using this approach, breathing patterns were similar to spontaneous breathing (without assistance) during normocapnia, and at different levels of hypercapnia. Sharshar and coworkers (7) demonstrate that the transdiaphragmatic pressure trigger frequently precedes flow triggering, confirming previous observations of Parthasarathy and coworkers (8). As pointed out by the authors (7) however, the influence of cardiac artifacts on transdiaphragmatic pressure triggering in patients with acute respiratory failure in the supine position remains to be evaluated. In terms of triggering of ventilator assistance, an important feature of using of transdiaphragmatic pressure, pointed out by Sharshar and coworkers (7), is that it eliminates trigger delays due to intrinsic positive end-expiratory pressure, but without the risk of false triggering due to expiratory muscle activity, which would be the case when using esophageal pressure (9). On the same topic, because transdiaphragmatic pressure is gastric pressure minus esophageal pressure, diaphragmatic work of breathing, by definition, includes an abdominal component. Consequently, systems delivering assistance in response to diaphragmatic activation or transdiaphragmatic pressure respond to abdominal loads, which can be important when abdominal compliance is reduced. New trends in mechanical ventilation favor noninvasive positive-pressure ventilation, which reduces the complications associated with endotracheal intubation, such as nosocomial infections (10). However, there are limitations to successful implementation of noninvasive ventilation, mainly because of poor patient–ventilator interaction due to leaks and mechanical components of respiratory system, circuit, and interface (11). These are factors that do not affect transdiaphragmatic pressure and neural activation as controller signals. Given the complications of endotracheal intubation, it appears reasonable to examine the possible advantage of trading tracheal for esophageal invasiveness, using noninvasive interfaces. If improved patient–ventilator interaction improves tolerance to mechanical ventilation it should help to limit the use of sedation, paralysis, and controlled modes of mechanical ventilation. Studies have shown that severe diaphragmatic atrophy develops within 1 day in mechanically ventilated animals that do not breathe spontaneously (12). Therefore, it is important to ensure that patients who are judged capable of maintaining spontaneous breathing, actually do maintain reasonable breathing activity. The first principle of demand-driven proportional ventilatory assistance is that the inspiratory muscles be active both at the onset and throughout the entire breath, otherwise no assistance is delivered. Although patient-triggered mechanical ventilation was initially designed to maintain spontaneous breathing, conventional ventilator systems cannot guarantee to what extent muscles remain active after the trigger phase. Twelve years ago, Kimura and coworkers (13) demonstrated that during high levels of assistance in the pressure-support mode only the triggering phase is associated with muscle effort. Today, flow-based trigger systems have increased sensitivity to such levels that (in the absence of leaks and increased airway resistance) cardiac-induced fluctuations in the flow pattern can trigger ventilator assistance (14). Through both enhanced monitoring capabilities and the intrinsic need for transdiaphragmatic pressure generation, the system of proportional assist described by Sharshar and coworkers (7) guarantees that patients cannot be overassisted, and thus provides built-in protection against inspiratory muscle disuse atrophy. Use of ventilator modes that allow the patient's own afferent receptors to control the degree of lung distention may be a preferred strategy to keep the airways from collapsing or to avoid excessive lung volumes, and thus protect from ventilator-induced lung injury (15). In the supine position, recruitment of dependent lung regions is compromised during control-mode ventilation in anesthetized and paralyzed patients relative to spontaneously breathing patients (16). Although future studies are still required to determine whether or in what patient groups this new generation of patient demand-controlled ventilators will be of advantage, it is certain that demand-driven mechanical ventilation will play an important role in the future of mechanical ventilation. FOOTNOTES Conflict of Interest Statement: C.S. is consultant for Siemens Elema (Solna, Sweden) through NeuroVent Research, Inc., an R&D company in which Sinderby is shareholder. NeuroVent Research, Inc., has no sales activities other than for developing equipment for research collaborations in which Dr. Sinderby participates. The consulting is to transfer technology to Siemens Elema for EMG-controlled ventilation. Dr. Sinderby has no financial/commercial interest or experience in using diaphragm pressure to control mechanical ventilation (the field to which the editorial pertains). REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
