Proportional Assist Ventilation Andreas Schulze, Peter Schaller, Bernd Höhne, Susanne Herber-Jonat

Transcription

1 1 Proportional Assist Ventilation Andreas Schulze, Peter Schaller, Bernd Höhne, Susanne Herber-Jonat In proportional assist ventilation (PAV, also referred to as elastic and resistive unloading), the patient actually breaths spontaneously. Every breath is supported by the ventilator which increases the airway pressure. The airway pressure increase "follows" the course of the increase in the breath volume and/or breath flow during every inspiration phase. The assist airway pressure is then returned to the adjusted PEEP during the expiration phase. The pressure increase during every inspiration is proportional to the currently inspired volume and/or inspired breath flow. That means that neither the possible inspiration peak pressure nor the inspiration or expiration times are stipulated. These are defined by the patient's breathing attempt and his breathing pattern, i.e. by his biological breathing regulation system. The ventilator only "follows" the patient (the ventilator is "enslaved"). The breathing pressure curve thus varies typically in time with the changes in the spontaneous breath depth and spontaneous breathing frequency (Fig. PAV1.) 9, 13, 14 O) Paw 2 (cm H (l/min) V Time (s) Fig. 1 PAV1: Proportional assist ventilation for a preterm infant. The diagram shows the ventilation pressure signal (Paw) and the breathing gas flow signal ( V & ) over time. The measurement was taken in close proximity

2 2 to the patient at the Y-piece of the ventilation hose system respectively between the endotracheal tube and connector to the ventilation hose system. The inspiration breath flow is shown above the zero line and the expiration breath flow below it. Note the typical fluctuations in breathing frequency, breath volume, inspiration peak pressure and inspiration time. The operator only adjusts the extent of pressure gain per milliliter breath volume and/or per liter/min breathing gas flow. If this is set to "zero", the patient is breathing under CPAP, i.e. he has to do all the breathing work himself. For a positive gain adjustment, the ventilation pressure increases during inspiration so that the ventilator takes on part of the breathing work. The higher the selected gain per volume and/or flow, the lower the breathing work to be performed by the patient for ventilation. In the end, the aim is to reduce the breathing work for a patient with a diseased lung (e.g. stiff lung and/or increased airway resistance) to such an extent that it only corresponds to the breathing work necessary in a normal lung. That means that the patient can breath with a system of apparently normal mechanical breathing parameters, while the ventilator takes on the additional breathing work required because of the illness. Total breathing work for normal spontaneous breathing consists of an elastic and a resistive part. The first is based on the retraction force of the elastic fibers, which increases during inspiration in proportion to the inspired breath volume, and has to be overcome by the breathing muscles. The resistive part is based on the gas flow resistance within the airways: the faster the flow, the greater the pressure required to generate it, i.e. the necessary pressure is proportional to the magnitude of the breathing gas flow. Breathing work unloading can therefore consist of two different components. Elastic unloading. The ventilation pressure is increased in proportion to the inspiration volume. The gain of elastic unloading, i.e. volume-proportional assist, is consequently adjusted in cm H 2 O/ml (pressure per unit of inspired volume) (Fig. PAV2). 6 Resistive unloading. The ventilation pressure is increased in proportion to the breath flow. The gain of resistive unloading, i.e. flow-proportional assist, is consequently adjusted in cm H 2 O/ (l/s) (pressure per unit of inspired flow) (Fig. PAV3). 5

3 3 Breath volume Δ V Δ V Δ V Zeit Ventilation pressure PEEP Gain B Gain A Δ P Δ P Δ P CPAP: gain=0 Zeit Gain = Δ Ventilation pressure (P, cm H 2 O) Δ Breath volume (V, ml) Fig. PAV2: Diagram to show signal progression in breath volume and ventilation pressure over time for three successive spontaneous breaths with differing strength and differing speed during volume proportional assist ventilation (elastic unloading). The ventilation pressure progression follows the same "form" as breath volume progression because it is controlled proportional to the current amount of breath volume, i.e. the ratio of ventilation pressure and breath volume (ΔP/ΔV) is the same at all points in time. The higher this quotient, the greater the support for the patient's spontaneous breathing compared to the elastic retraction tendency of the lung, so that the ventilator takes on all the more elastic breathing work. At the ventilator, the operator adjusts the quotient in cm H 2 O/ml as so-called gain. The illustration shows three different settings. No support corresponds to zero gain with the patient breathing under CPAP, i.e. the ventilation pressure does not change compared to breath volume. Gain "A" is smaller than gain "B".

4 4 Inspiration Expiration 0 Breath flow Zeit PEEP Ventilation pressure Gain B Gain A 0 Gain = CPAP: Gain=0 Δ Ventilation pressure (cm H 2 O) Δ Breath flow (L/s) Zeit Fig. PAV3: Diagram to show signal progression in breath flow and ventilation pressure over time for three successive spontaneous breaths with differing strength and differing speed during flow proportional assist ventilation (resistive unloading). The ventilation pressure progression follows the same "form" as breath flow progression because it is controlled proportional to the current amount of breath flow, i.e. the ratio of ventilation pressure and breath flow (ΔP/ΔV') is the same at all points in time. The higher this quotient, the greater the support for the patient's spontaneous breathing compared to airway resistance, so that the ventilator takes on all the more resistive breathing work. At the ventilator, the operator adjusts the quotient in cm H 2 O/ml as so-called gain. The illustration shows three different settings. No support corresponds to zero gain with the patient breathing under CPAP, i.e. the ventilation pressure does not change compared to breath volume. Gain "A" is smaller than gain "B. The two thin arrows indicate that the changes in ventilation pressure occur "practically at the same time" as the progression in the breath flow signal. Apart from the safety parameters, only three settings are adjusted for proportional assist ventilation: 1. The PEEP level. This influences the degree of functional residual capacity (FRC) The gain of elastic unloading in cm H 2 O/ml depending on the extent of restrictive lung change The gain of resistive unloading in cm H 2 O/l/s depending on the extent of obstructive lung change. 4, 5, 8

5 5 For combined use of elastic and resistive unloading, at every point in time of inspiration, the ventilation pressure curve is the sum of the pressure values generated by the two individual components. Guidelines for individual selection of mechanical breath unloading gain The simplest and most practical approach to using PAV consists of setting the gain to zero to start with, starting CPAP (or low frequency SIMV) and activating the PAV function. The gain of elastic unloading is increased slowly while at the same time, the operator observes the child's breathing pattern, the extent of thoracic contraction and the airway pressure curve on the ventilator monitor. Generally it can be expected that when the PAV has been correctly adjusted, the peak inspiration pressure lies a few cm H 2 O below the peak pressure presumed necessary for conventional ventilation. 12 A suitable unloading gain can also be detected by a reduction in thoracic contractions with regular spontaneous breathing. 2, 11 Consideration must be given to the fact that smaller children will need higher gains than larger children (in absolute terms, i.e. in cm H 2 O/ml), because compliance and breath volume are related to body weight. That means that children with a body weight < 1000g usually need more than 1 cm H 2 O/ml elastic unloading, while smaller settings are often sufficient in larger children with a comparable extent of restrictive disorder. The gain of resistive unloading can initially be set to the estimated resistance value of the endotracheal tube, to compensate at least for the increase in resistive breathing work by the child caused by the endotracheal tube (a 2.5 mm ID endotracheal tube has a resistance of about 25 cm H 2 O/l/s). When the PAV is correctly set, preterm and newborn infants typically breath "fast and flat" with breathing rates frequently reaching 60 to 100/min, and with breath volumes of around or below 5 ml/kg body weight. The inspiration times are short, usually around s. Some preterm infants reveal somewhat higher PaCO 2 volumes under PAV than with conventional ventilation with customary settings. But clinical studies have revealed that this 3, 12 difference in PaCO 2 has no statistical significance. If the gain is set too high, this can result in overcompensation of the elastic retraction force of the lung or of the airway resistance. In this case, an initial airway pressure gain tends to assume a life of its own through the cycle of positive feedback from the airway pressure gain, subsequent increase in lung volume respectively breath flow and resulting further airway pressure gain. Once a breath has begun, under these conditions this causes passive inflation of the lung (even without any further contribution from the patient) until the set upper limits for

6 6 airway pressure or breath volume are reached. Proportional assist ventilation then changes into assist/control. 10 Overcompensation (excessive gain) for elastic unloading is revealed by the airway pressure curve. At the start of inspiration, the pressure increases rapidly in the automatic mode until the upper airway pressure limit setting is reached (or the default limit for breath volume) (Fig. PAV4). In the Stephanie infant ventilator, the airway pressure is then immediately unloaded to the PEEP setting (in the new software version, the airway pressure initially remains on the level of the upper pressure limit and is only reduced when the patient begins the expiration phase or at the end of a maximum time period set to 0.7 s). The greater the overcompensating gain, the faster the increase in insufflatory pressure. O) Paw 2 (cm H Upper airway pressure limit (L/min) V Time (s) Fig. PAV4: Proportional assist ventilation in a preterm infant with well adjusted elastic unloading (left) and elastic overcompensation (right). At the start of a spontaneous breath or when there is an artifact on the flow signal, which the unit interprets as the start of a breath, in overcompensation the airway pressure quickly increases automatically through to the upper pressure limit setting. Once the airway pressure has then been reduced to the PEEP setting, the cycle can be repeated and finally cause excessive ventilation. If overcompensation is detected, the unloading gain should be reduced to avoid excessive ventilation. Overcompensation applies in elastic unloading when the adjusted ventilation pressure gain per breath volume unit (in cm H 2 O/ml) is greater than the simultaneous increase in elastic retraction force of the lung per volume unit. An elastic unloading gain which was adjusted at one point in time to an appropriate setting can therefore become overcompensating

7 7 unloading when there is an improvement in lung compliance. The gain should therefore be reduced appropriately when the lung disease improves. The upper airway pressure limit setting protects the patient from excessive pressure also in the following situations under proportional assist ventilation: Sighing breaths or hiccoughs with high breath volumes which could cause increased pressures under persistent proportional airway pressure gain (Fig. PAV5) Larger leak in the endotracheal tube. The leak flow simulates a larger inspiration volume without this actually reaching the lung. O) Paw (cm 2 H Upper airway pressure limit (L/min) V sigh Time (s) Fig. PAV5: Preterm infant under proportional assist ventilation. The airway pressure also increases proportional to the large volume of a sigh but not beyond the upper airway pressure limit setting. Here it is reduced to the PEEP. Backup ventilation for apnea. During proportional assist ventilation, in the case of apnea, conventional, adequately effective passive ventilation must start in good time (Fig. PAV6). The previous software for the Stephanie infant ventilator presumes apnea when the breath flow signal follows the zero line, but also when all breath volumes ascertained according to the flow signal are so small that they remain under a stipulated minimum. The default for this lower limit for detecting a change in breath volume is set to 1 ml but can be changed (menu "Options" > "Minimum VT" > ). In this case, the breath volume measured during expiration is measured because only this certainly represents the volume which was

8 8 actually in the lung (by contrast, the volume measured during inspiration can contain a leak component which has no effect on the exchange of gases). Passive conventional ventilation then begins after a set delay period (called "apnea time" in the new software). The default delay time which is normally set to 2 s can be changed in the menu "Options" > "Backup delay time" > "Time in seconds; select from 1, 2, 4, 8, 16 seconds". O) Paw 2 (cm H backup backup ventilation (L/min) V Time (s) Fig. PAV6: Passive, conventional mechanical ventilation (*, backup ventilation) is triggered on ascertaining apnea under proportional assist ventilation. The first spontaneous breath to be identified as such during backup ventilation terminates conventional ventilation and returns to proportional assist ventilation (older software version). In the case of apnea under PAV, arterial hemoglobin O 2 saturation falls faster depending on how low the child's functional residual capacity is, i.e. the less oxygen is held in the lung. The setting for the backup delay time (the new software version calls it "apnea time" or "backup begin after apnea time" which can be adjusted in steps of 0.5 seconds) should therefore be adjusted to shorter intervals for more critical pulmonary diseases. Backup ventilation is adjusted by the operator as normal for conventionally controlled ventilation with regard to inspiration peak pressure (Pmax setting adjustment), inspiration and expiration time and the form of inspiration pressure gain (linear, sinusoidal or quasi rectangular gain). The settings should be adjusted so that any possible oxygen de-saturation occurring after the backup delay time (new software: apnea time) can be rectified without resulting in increased hyperventilation of the child. Hyperventilation could result in total suppression of

9 9 spontaneous breathing activity so that it is not possible to return to PAV. The first spontaneous breath to be identified as such by the unit during backup ventilation terminates conventional ventilation and returns to proportional assist ventilation (older software version). The expiration time set for backup ventilation should therefore be long enough (even when using the new software) to allow the patient to take spontaneous breaths between backup insufflation. New software version with improved forms of backup ventilation under PAV. In preterm infants, the first spontaneous breaths after apnea are usually smaller and less regular than during a preceding period with sufficient spontaneous breathing. If backup ventilation is switched off completely during the first breath after apnea, as was the case with the previous software version, the initial spontaneous breathing frequently transforms into apnea again. This can occur particularly for existing or progressive oxygen de-saturation which in turn depress spontaneous breathing in preterm infants (paradox hypoxia in the newborn) (Fig. PAV7). Fig. PAV7. PAV showing the breath flow signal (V ) and ventilation pressure (Paw) in a preterm infant with 800 g birthweight with bouts of apnea. The backup ventilation (previous software version) is triggered after the first apnea phase but is switched off again completely after two ventilation strokes (asterisks) because of small spontaneous breathing attempts which are not sufficient to stop the incipient fall in oxygen saturation (SpO 2 ). They rapidly become apnea again. This repeats several times, with a constant decrease in oxygen saturation. The new software version therefore contains a minimum time period for full backup ventilation ("backup time") together with only gradual reduction of the backup ventilation frequency on the return of spontaneous breathing(fig. PAV8).

10 10 During the return to spontaneous breathing after apnea, it is therefore advisable to warrant additional breathing support for a limited period of time. 1 The new software version contains the following standard functions for backup ventilation: 1. It begins with full backup ventilation as soon as no spontaneous breathing was detected for a stipulated period of time. This period of time stipulated in the unit is called "apnea time". The default setting is 4 seconds and can be modified by the operator in steps of 0.5 seconds in the range from 0.5 to 15 seconds in the menu "Apnea time > backup begins after... seconds". Spontaneous breathing is detected only by the trigger function, i.e. depending on the adjusted trigger limit from the flow signal from inspiration. 2. Backup ventilation is pressure-controlled, time-limited and synchronized if spontaneous breaths occur. The ventilation parameters are adjusted on the front panel of the unit at the knobs for insufflation pressure (P max ), insufflation time and expiration time (frequency) etc. 3. Once backup ventilation has been triggered, it runs initially for a set minimum period of time, i.e. it cannot be interrupted again by one single recurrence of the trigger. This minimum period of time is called "backup time". The default setting is 10 seconds and can be modified by the operator in steps of 5 seconds in the range from 5 to 60 seconds in the "Backup time" menu. 4. The end of backup ventilation is initiated when the trigger is activated by spontaneous inspiration after the end of the backup period. Backup ventilation is brought to a gradual close (Fig. PAV8) by reducing the backup ventilation frequency step by step, in each case after the end of a time period which is the same as the "backup time" setting. The ventilation frequency set on the unit which constitutes the initial backup frequency is reduced in three steps of 50%, 33% and finally 20% of the initial frequency until the backup support ceases completely. The initial form of complete backup ventilation starts up again immediately if "apnea" is ascertained during the weaning phase, i.e. if the trigger is not activated during a period of time corresponding to the stipulated "apnea time".

11 11 Fig. PAV8: The frequency of backup ventilation (asterisks) is gradually reduced as spontaneous breathing returning after apnea gets increasingly stronger. Spontaneous breathing is supported by PAV already in this weaning phase. Preterm infant weighing 840 g under PAV; new software version. The stated standard functions for backup ventilation can be supplemented by the Stefanie unit thanks to the new software which now enables it to accept a pulse oximetry signal (from the NEOSID Nova unit, F. Stephan GmbH, Massimo Technology) to detect oxygen de-saturation and optimize backup ventilation accordingly. Controlling backup ventilation just on the basis of mechanical ventilation criteria is not capable of providing a fully satisfactory solution in certain situations. Oxygen de-saturation resulting from hypoventilation or other causes can precede apnea and even cause or prolong such conditions (paradox hypoxia reaction in the newborn) (Fig. PAV9).

12 12 Fig. PAV9: In spite of regular and strong spontaneous breathing, a fall in oxygen saturation occurs, followed by apnea. This sequence corresponds to so-called paradox hypoxia reaction in the newborn where hypoxia is a secondary cause of breath depression. Here the backup ventilation was controlled purely by mechanical ventilation criteria. When measured values fall below a stipulated pulse oximetry limit, the new software is capable of initiating backup ventilation at this early stage, thus anticipating breath depression in these situations. It is therefore appropriate and, according to clinical tests, effective to start beginning backup ventilation every time the measured values fall below a set lower pulse oximetry limit. 1 The gradual reduction in backup ventilation does not start until this limit value has been exceeded again, at least the set backup period has finished and spontaneous inspiration has been detected. Another new option in the PAV mode for the "Stephanie" unit consists in parallel additional support for spontaneous breathing during periods with low tidal volume. This form of additional support is called "minimum volume guarantee". The user can adjust a minimum target tidal volume for spontaneous breaths using the tidal volume knob (V T ). If the volume falls below this level, the unit gradually adds more ventilation pressure with every following breath in addition to the pressure generated by the PAV function until the target volume is reached. If this target volume is exceeded later by increasingly stronger inspiration efforts, then the additionally supporting inspiration pressure is decreased gradually and then reduced completely so that the PAV function on its own reaches or exceeds the target volume. When the "minimum volume guarantee" function is activated and the tidal volume falls below the set target, the inspiration ventilation pressure no longer progresses just proportional to the inspiration diaphragm muscle effort (i.e. endogenous

13 13 breath drive expressed in the volume and flow signal of spontaneous breathing) but is then out of proportion to the effort. A minimum tidal volume is then continuously enforced unless the spontaneous breathing is so weak that the inspiration flow no longer reaches the adjusted trigger limit line, i.e. spontaneous breaths are no longer detected so that pressurecontrolled backup ventilation starts at the end of the apnea time, as described above. Up to now, the "minimum volume guarantee" function has not been studied in animal experiments or clinical studies. According to theoretical model tests, it effectively prevents longer periods of breaths below the set target level without destabilizing spontaneous breathing activity during PAV. Small preterm infants develop a fast, flat breathing pattern under PAV. It is not known how the "minimum volume guarantee" function modifies this endogenous breathing pattern, particularly depending on the level selected for the minimum volume, or whether for example there is any influence on the arterial CO 2 Wert and its fluctuation. The problem of endotracheal tube leaks under PAV. A leak between the endotracheal tube and the trachea causes breath to flow without contributing to the breath volume. But the "leak flow" passes the flow sensor during inspiration (Fig. PAV10) so that it is initially measured in with the inspiration flow. Leak flow can therefore "simulate" inspiration. Fig. PAV1: "Leak flow" between endotracheal tube and trachea (green, bent arrows) flows initially through the flow sensor and endotracheal tube but then it leaks out without reaching the lung as effective breath volume. The volume of the leak therefore "simulates" inspiration volume at the flow sensor. Under proportional assist ventilation, a leak flow can therefore result in an increase in airway pressure which is out of proportion to the increase in actual lung volume (respectively to the actual breath flow in the lung). A large leak of this kind can hinder proportional assist ventilation, particularly when the leak also varies in size. For an excessive leak, the airway

14 14 pressure changes rapidly up and down between PEEP and the set upper pressure limit. A small leak typically only occurs at the end of the inspiration phase when the airway pressure has increased to such an extent that expansion of the trachea causes a sudden opening at the endotracheal tube, which then closes again during expiration under lower airway pressure. In this case, airway pressure under proportional assist ventilation initially shows typical progression but then suddenly increases (together with the flow signal) at the end of the inspiration phase. The software of the Stefanie ventilator estimates the size of the leak flow using an algorithm and corrects the measured flow signal internally up to a certain degree. This permits PAV to continue in spite of small to medium leaks at the endotracheal tube (Fig. PAV11). O) Paw 2 (cm H V T (ml) Time (s) Fig. PAV11: Proportional assist ventilation in a 12 week old preterm infant with severe chronic lung disease. The diagram (bottom half) shows the volume signal measured between endotracheal tube and ventilation tube system (V T, obtained as integral of the flow signal over time). The volume measured during inspiration is far larger than the expiration volume because of the leak at the endotracheal tube. The volume signal is reset to the zero limit by the unit software at the start of every inspiration phase because otherwise it would increase constantly with every leak. In this illustrated case, the leak is > 50% of the breath volume, but proportional assist ventilation is still possible. The size of the endotracheal tube and the amount of ventilation pressure define the size of the leak. It usually increases clearly, the longer ventilation lasts.

15 15 Supplementary remarks on safety settings. The operator must adjust the following in every case. The upper limit for obtainable airway pressure (at the 'Pmax' know). This pressure limit should be a few cm H 2 O above the inspiration peak pressure resulting during PAV with well adjusted gain. This also defines the inspiration peak pressure used during backup ventilation. The knob for breath volume on the Stefanie respirator acts as an additional safeguard during PAV to prevent excessive airway pressure. The airway pressure (for elastic unloading) is not increased any further on reaching the breath volume adjusted by the VT knob. This setting should therefore be about 5 10 ml higher than the expected "normal" breath volume for PAV (about 4-6 ml/kg). If the VT setting is adjusted too low by mistake, the unit will not build up sufficient inspiration pressure if any, in spite of the presence of spontaneous breathing (knob turned right round to the left). Guaranteed by the unit without additional settings The maximum possible length of the inspiration time is limited by the software. If insufflation lasts longer than 0.7 s, the airway pressure is reduced to the PEEP. LITERATURE 1. Herber-Jonat S, Rieger-Fackeldey E, Hummler H, Schulze A. Adaptive mechanical backup ventilation for preterm infants on respiratory assist modes. Intensive Care Med, eingereicht Musante G, Schulze A, Gerhardt T, Everett R, Claure N, Schaller P, Bancalari E. Respiratory mechanical unloading decreases thoraco-abdominal asynchrony and chest wall distortion in preterm infants. Pediatr Res 2001; 49: Rieger-Fackeldey E, Gerhardt T, Claure N, Everett R, Schulze A, Bancalari E. Crossover comparison of proportional assist ventilation and synchronized intermittent mandatory ventilation in very low birthweight infants with chronic lung disease. Pediatr Res 2001; 49:281A. 4. Schulze A, Schaller P, Dinger J, Winkler U, Gmyrek D. A method of calculating total respiratory system compliance from resonant frequency: validity in a rabbit model. Pediatr Res 1990; 28:

16 16 5. Schulze A, Schaller P, Gehrhardt B, Mädler HJ, Gmyrek D. An infant ventilator technique for resistive unloading during spontaneous breathing. Results in a rabbit model of airway obstruction. Pediatr Res 1990; 28: Schulze A, Schaller P, Jonzon A, Sedin G. Assisted mechanical ventilation using elastic unloading: A study in cats with normal and injured lungs. Pediatr Res 1993; 34: Schulze A, Schaller P, Töpfer A, Kirpalani H. Resistive and elastic unloading to assist spontaneous breathing does not change functional residual capacity. Pediatr Pulmonol 1993; 16: Schulze A, Jonzon A, Schaller P, Sedin G. Effects of ventilator resistance and compliance on phrenic nerve activity in spontaneously breathing cats. Am J Respir Crit Care Med 1996; 153: Schulze A, Schaller P. Assisted mechanical ventilation using resistive and elastic unloading. Semin Neonatol 1997; 2: Schulze A, Rich W, Schellenberg L, Schaller P, Heldt GP. Effects of different gain settings during assisted mechanical ventilation using respiratory unloading in rabbits. Pediatr Res 1998; 44: Schulze A, Suguihara C, Gerhardt T, Schaller P, Claure N, Everett R, Bancalari E. Effects of respiratory mechanical unloading on thoracoabdominal motion in meconium-injured piglets and rabbits. Pediatr Res 1998; 43: Schulze A, Gerhardt T, Musante G, Schaller P, Claure N, Everett R, Gomez-Marin O, Bancalari E. Proportional assist ventilation in low birth weight infants with acute respiratory disease. A comparison to assist/control and conventional mechanical ventilation. J Pediatr 1999; 135: Schulze A, Bancalari E. Proportional assist ventilation in infants. Clin Perinatol 2001; 28: Schulze A. Respiratory mechanical unloading and proportional assist ventilation in infants. Acta Paediatr Suppl 2002; 91:19-22.