BACKGROUND: Several new generation neonatal ventilators that incorporate conventional as well as high frequency ventilation (HFOV) have appeared on the market. Most of them offer the possibility to use HFOV in a volume-targeted mode, despite absence of any preclinical data. With a bench test, we evaluated the performances of 4 new neonatal HFOV devices and compared them to the SensorMedics HFOV device.
METHODS: Expiratory tidal volumes (VT) were measured for various ventilator settings and lung characteristics (ie, modifications of compliance and resistance of the system), to mimic several clinical conditions of pre-term and term infants.
RESULTS: Increasing the frequency proportionally decreased the VT for all the ventilators, although the magnitude of the decrease was highly variable between ventilators. At 15 Hz and a pressure amplitude of 60 cm H2O, the delivered VT ranged from 3.5 to 5.9 mL between devices while simulating pre-term infant conditions and from 2.6 to 6.3 mL while simulating term infant conditions. Activating the volume-targeted mode in the 3 machines that offer this mode allowed the VT to remain constant over the range of frequencies and with changes of lung mechanical properties, for pre-term infant settings only while targeting a VT of 1 mL.
CONCLUSIONS: These new generation neonatal ventilators were able to deliver adequate VT under pre-term infant, but not term infant respiratory system conditions. The clinical relevance of these findings will need to be determined by further studies.
The clinical use of high frequency oscillatory ventilation (HFOV) in pre-term infants remains controversial, as previous meta-analysis failed to show a significant difference between HFOV and conventional mechanical ventilation in mortality, bronchopulmonary dysplasia, and adverse neurological outcomes.1,2 However, some studies that initiated HFOV very early in pre-term infants with respiratory distress syndrome have shown that surfactant doses, ventilator days, and chronic lung diseases could be reduced compared with conventional mechanical ventilation.1,3,4 Therefore, the concept of using HFOV as a primary ventilation mode in infants presenting with respiratory distress syndrome of various origins remains of interest.
Accordingly, several manufacturers have introduced new neonatal HFOV devices on the market. These new devices, including the Babylog VN500 (Dräger Medical, Lübeck, Germany), the Fabian HFO (Acutronic Medical System, Hirzel, Switzerland), the SLE 5000 (SLE, South Croydon, United Kingdom), and the Leoni Plus (Heinen+Löwenstein, Bad Ems, Germany) offer both conventional mechanical ventilation and HFOV as ventilation modes. In addition, a new concept of VT targeting during HFOV has been introduced in 3 of them (Babylog VN500, Fabian HFO, and Leoni Plus), despite the absence of any clinical data on volume targeting high frequency oscillation. Recently, one experimental study investigated the stability of PCO2 during HFOV using the volume guarantee mode, showing that PCO2 remained stable with maintenance of minute ventilation during HFOV before and after surfactant depletion in piglets.5
The technical features added to these new ventilators, in particular the choice of the ventilator mode between conventional mechanical ventilation and HFOV, the measurements of VT, and the volume-targeted mode during HFOV might represent all attractive features for the clinician, but data on volume delivery accuracy and performance of these new generation HFOV devices are lacking. Most of the published data on mechanical performance of oscillatory ventilators dates back more than 10 y,6–8 and there is only one recent study, which looked into the performance of the Babylog VN500, showing a reduction of its power for generating oscillatory pressure amplitudes at higher operating frequencies.9 Furthermore, published data on stability of VT during volume-targeted ventilation are conflicting.10,11 Our group found, in a bench study12 on volume-targeted ventilation during conventional mechanical ventilation significant differences between set VT and delivered VT among the tested neonatal ventilators under changes in respiratory mechanics. Given these unknowns, we aimed to evaluate the performances of several new neonatal HFOV devices at various oscillations frequencies and in various respiratory system conditions, with and without the volume targeting mode activated.
Therefore, our hypothesis was that the new neonatal HFOV devices would perform similarly to the SensorMedics 3100A (CareFusion, San Diego, California), with and without the volume targeting mode activated, and in various lung conditions.
The clinical use of high frequency oscillatory ventilation (HFOV) in preterm infants remains controversial as meta-analysis failed to show a significant difference versus conventional ventilation with respect to mortality, bronchopulmonary dysplasia, and adverse neurological outcomes. Despite the unproven utility, a number of conventional ventilators with an HFOV option have been introduced.
What this paper contributes to our knowledge
Neonatal HFOV ventilators deliver adequate oscillation tidal volume (1 mL/kg) under bench test conditions simulating respiratory system mechanics of a pre-term infant and maintain tidal volume constant in the volume-targeted mode. However, not all of the ventilators could deliver adequate oscillation tidal volume (4 mL/kg) for term infant conditions.
We studied 4 neonatal ventilators offering both HFOV and conventional mechanical ventilation modes: the Babylog VN500, the Fabian HFO, the SLE 5000, and the Leoni Plus (Table 1). As reference device for comparisons, we used the well-established SensorMedics 3100A.
The ventilators were equipped with the same commercially available standard neonatal respiratory circuit (single-use neonatal respiratory circuit, Fisher & Paykel Healthcare, Auckland, New Zealand) with the exception of the SLE 5000 and the SensorMedics 3100A, which can only operate a unique circuit supplied by the manufacturer. The humidification chamber was kept in line but run dry with no heating. The continuous bias flow was (1) automatically modulated for the Babylog VN500, (2) delivered at a constant flow at 8 L/min for the SLE 5000 and 7 L/min for the Leoni Plus, (3) set at 20 L/min at the maximal value possible for the Fabian HFOV, and (4) set at 20 L/min per our guidelines for the SensorMedics 3100A.
The ventilators were connected to the 8.0-mm internal diameter tube of a neonatal test lung (model 1601, Michigan Instruments, Grand Rapids, Michigan) (Fig. 1). Compliance was adjusted on the test lung, and airway resistance was simulated using different airway resistors, which allowed comparison of the performance of the ventilators in various respiratory system conditions. To simulate the mechanical proprieties of the respiratory systems of a full-term infant (with an assumed body weight of 4 kg), we used a resistance of 70 cm H2O/L/s with compliance of 4 mL/cm H2O; to simulate a pre-term infant (with an assumed body weight of 1 kg), we used a resistance of 100 cm H2O/L/s with compliance of 2 mL/cm H2O as a baseline settings. The expiratory VT was measured with the Florian neonatal respiratory monitor (Acutronic Medical System), which has been validated for measurements of very small VT during HFOV.13 The Florian's flow sensor was positioned at the Y-piece and was calibrated according to the manufacturer's instructions before each test series. Mean airway pressure (̄Paw) and oscillation pressure amplitude (ΔPvent) were measured by the ventilator's transducer inserted between the ventilator and the Y-piece and recorded as displayed by the ventilator. For each of the 5 ventilators, the same operator recorded single measurements manually after the system had reached a steady state, with no more changes in the readout after a minimum of 1 min.
We compared the expiratory VT at different oscillatory frequencies (5, 10, and 15 Hz) and at increasing ΔPvent (30, 45, 60 cm H2O and maximal ΔPvent offered by the ventilator) in full-term and pre-term infant settings. The ̄Paw was set at 15 cm H2O, and the I-E ratio was set at 1:1. To simulate restrictive lung disease, the compliance was modified from 4 to 2 mL/cm H2O (full-term) and from 2 to 1 mL/cm H2O (pre-term). To simulate obstructive disease, resistance was modified from 70 to 200 cm H2O/L/s (full-term) and from 100 to 200 cm H2O/L/s (pre-term). To evaluate the effectiveness of the volume-targeted mode in maintaining the desired VT, the test series was repeated with the volume-targeted mode activated in 3 of the ventilators offering this option. We used a targeted VT of 4 mL under full-term settings and 1 mL under pre-term infant settings. These target tidal volumes during oscillation have been chosen based on our own clinical observations (with most often observed VT, which resulted in normocapnia, in the range 0.5–1.5 mL/kg body weight at an oscillation frequency of 15 Hz) over the years with the use of the SensorMedics 3100A device in combination with a neonatal respiratory monitor (Florian). The Florian has been validated for VT measurements during HFOV by Scalfaro et al.13
Influence of the Ventilator Properties on Delivered VT (Volume-Targeted Mode Inactivated)
While simulating full-term infant conditions, all ventilators demonstrated a linear increase of the VT at 5 and 10 Hz following ΔPvent increase (Fig. 2A). However, at 15 Hz and a ΔPvent set at 60 cm H2O, we observed heterogeneity in the delivered VT, which ranged from 2.6 to 6.3 mL. In particular, the Babylog VN500 was unable to increase the VT because ΔPvent could not be increased beyond 36 cm H2O at this oscillation frequency. While simulating pre-term infant conditions, we observed larger variability in the delivered VT between devices (Fig. 2B), in particular at 15 Hz, with VT ranging from 3.5 to 5.9 mL with a pressure amplitude set at 60 cm H2O. Again, the Babylog VN500 delivered lower VT at 5 Hz and 15 Hz than the other ventilators and was unable to increase its maximal ΔPvent beyond 55 cm H2O at a frequency of 10 Hz and beyond 33 cm H2O at a frequency of 15 Hz. Notably, the performance of the Babylog VN500 was unaffected by the type of the ventilator circuit used: Fisher Paykel Healthcare or Hyrtel (Dräger Medical), the latter being the only one recommended by the manufacturer.
Effect of Lung Mechanical Proprieties on Delivered VT
When compliance was changed from 4 to 2 mL/cm H2O in full-term infant setting (Table 2), delivered VT remained similar in all the ventilators and for all the frequencies tested, except for 15 Hz. At 15 Hz, the Babylog VN500 and the Leoni Plus responded to a decrease in compliance by an increase of the delivered VT. In pre-term infant settings (Table 3), a compliance change from 2 to 1 mL/cm H2O at 5 Hz resulted in a decrease in VT, whereas at 15 Hz, the VT remained similar for all the ventilators with the exception of the Fabian HFO (which increased its VT). The increase in resistance both in pre-term and full-term infant settings resulted in a decrease of the delivered VT for all the ventilators, regardless of compliance.
Influence of Volume-Targeted Mode on VT After Modification of Lung Mechanical Properties
While simulating a full-term infant at 5 and 10 Hz and with a volume target set at 4 mL, all 3 ventilators compensated for changes in compliance and resistance and kept the targeted volume constant (Fig. 3A). However, at 15 Hz, none of the ventilators could fully compensate for the increase in resistance, resulting in a substantial loss of VT. While simulating a pre-term infant, with a targeted volume of 1 mL, all 3 ventilators compensated for changes in lung compliance and resistance and delivered the targeted VT when the frequency was set at 15 Hz (Fig. 3B). However, below 15 Hz, all the ventilators delivered a VT that was higher than the pre-set 1 mL.
We observed large variability in performance between the ventilators, in particular at high oscillatory frequency (ie, 15 Hz). The activation of the volume-targeted mode in pre-term infant settings allowed the VT to remain constant, although not always matching the set VT, during changes of the respiratory system mechanical properties.
As shown previously,6–8,14 the principal factors influencing the mechanical performance of HFOV ventilators are frequency, inspiratory time, and load charge. Although all ventilators delivered lower VT with increasing frequencies, there were marked differences in the extent of VT decrease between ventilators at an oscillatory frequency of 15 Hz. Indeed, the new generation of neonatal HFOV ventilators (Babylog VN500, Fabian HFO, and Leoni Plus) delivered lower VT than the 3100A and SLE 5000 with the same pressure amplitude settings, both under pre-term infant and full term infant settings. As the CO2 elimination during HFOV is proportional to f·VT,2,15,16 the difference in performance observed under our test conditions may be even more significant in clinical settings.
There is no consensus about the ideal VT for HFOV, but studies in animals17,18 and pre-term neonates19 have shown that a VT of 2–3 mL/kg allows optimal gas exchange during HFOV. Although the use of these new generation neonatal ventilators is intended for the neonatal population unlike the 3100A (a ventilator that can be used with patients larger than neonates), some of them might not have sufficient power to deliver required VT for adequate gas exchange, especially when used in full-term infants at high oscillatory frequencies.
Differences in the new HFOV ventilator design limiting maximal oscillatory flows may account for the variability in performance observed in conditions characterized by reduced flows toward the patient (ie, at higher frequencies and lower inspiratory time). The SLE 5000, whose operating principle includes a continuous bias flow with 3 opposing jet flows, was the most powerful neonatal ventilator in term of volume delivery, with a performance similar to the 3100A. On the other hand, the Babylog VN500 presented important oscillatory flow limitations at high frequency settings as has also been reported in a recent bench study by John et al,9 and these were similar to those seen in the older Babylog 8000.6,7 This feature may be explained by the specific design of the Babylog VN500 that incorporates some of the technical concepts of the Babylog 8000 (ie, a Venturi multi-valve system on the expiratory side) combined with a new pulsed inspiratory flow. Because the flow is set automatically based on the frequency and the ̄Paw, an insufficient bias flow may explain the power limitation of this ventilator seen in our study.
The use of volume target modes during conventional mechanical ventilation is well documented in neonates and has been shown to reduce the incidence of death, bronchopulmonary dysplasia, and severe intraventricular hemorrhage compared with pressure limited ventilation.20 Despite those obvious benefits, several reports demonstrated inaccuracies in the delivered VT during volume target ventilation with conventional mechanical ventilation with volume overshooting upon rapid change in respiratory mechanics.10,12
During HFOV, VT represents the principal determinant for CO2 removal and several reports found important fluctuation in the delivered VT during HFOV.21 In absence of direct VT monitoring on the old generation HFOV such as the SensorMedics 3100A, the clinician had to rely on the clinical exam as well as on repeated blood gas measurements to make changes in the ventilator settings which exposed the patients at risk for variation in PaCO2.
In this perspective, the incorporation of a flow sensor to measure VT into these new generation of neonatal HFOV represents a potentially useful improvement for the clinician although some inaccuracies in the measurements of VT has been reported.22 In addition, the possibility to combine HFOV with a volume-targeted mode on these ventilators will keep VT constant after change in lung mechanics as demonstrated recently in an animal model of surfactant deficiency.5
The activation of the volume-targeted mode under high resistance conditions allowed the ventilators to maintain constant small targeted VT (1 mL) but failed to maintain larger targeted VT (4 mL) at 15 Hz. Furthermore, at lower frequencies with a lower target VT of 1 mL, all 3 ventilators delivered a higher-than-set VT. Although there might theoretically be some potential advantages of using a guaranteed volume during HFOV, especially for small pre-term neonates, some precautions must be taken when this option is activated at a low frequency with a small VT setting, as it may result in undesired overventilation for very small pre-term babies. Therefore, experimental studies are needed to rule out our concerns before clinical studies can be conducted to determine the real utility of this new mode.
The primary limitation of our study is the experimental setup, which differs from the previous bench evaluation of neonatal HFOV ventilators.6,8,23 First, we used a test lung with airway resistance simulated using different sized airway resistors. This system allowed us to test different changes in respiratory mechanics but cannot completely reproduce the complexity of the neonatal respiratory system. Second, the delivered VT was directly measured using a hot-wire flow sensor, which may have added further resistance and could have attenuated the VT delivered by the ventilators. Although this might have induced a measurement bias, it could only be very small, because the resistance of the Florian anemometer is only 0.55 Pa/mL with a V̇ of 100 mL/s, which would result theoretically in an additional reduction in VT of only 0.3 mL.13
The record of single manual measurements on single ventilators constitutes another limitation, which may have affected the accuracy of the results. However, the record of measurements after 1 min steady state and the use of functional ventilators provided by their respective manufacturer and tested as per manufacturers' protocol before each test series helped to reduce the impact of these limitations on the different ventilators' characteristics described in this study. Because of these limitations, our results might not be directly applied to the clinical setting but emphasize the need for further clinical and animal studies to evaluate the performance and ventilation efficiency of the new generation neonatal HFOV ventilators.
In summary, all of the new generation neonatal HFOV ventilators can deliver adequate oscillation VT (1 mL/kg) under bench test conditions simulating various respiratory system mechanics of a pre-term infant, and do keep the VT constant in various respiratory system conditions upon activation of the volume-targeted mode. However, not all of the ventilators could deliver adequate oscillation VT (4 mL/kg) for term infant conditions. Furthermore, there was heterogeneity between devices in the VT delivery at higher frequencies. Therefore, further studies are needed to determine the clinical importance of our findings from the bench.
We thank Gustavo Matute-Bello MD (Division of Pulmonary & Critical Care Medicine, University of Washington, Seattle, Washington) for the review of this manuscript and the useful comments he made.
- Correspondence: Peter C Rimensberger MD, Neonatal and Pediatric Intensive Care Unit, Department of Pediatrics, University Hospital of Geneva, Rue Willy Donzé 6, 1211 Geneva 14, Switzerland. E-mail: .
The authors have disclosed no conflicts of interest.
- Copyright © 2015 by Daedalus Enterprises