ERJ
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print March 5, 2008, 10.1183/09031936.00144807
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary data
Right arrow All Versions of this Article:
31/6/1300    most recent
09031936.00144807v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Permissions
Right arrowRequest Permissions
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fauroux, B.
Right arrow Articles by Louis, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fauroux, B.
Right arrow Articles by Louis, B.
Eur Respir J 2008; 31:1300-1307
Copyright ©ERS Journals Ltd 2008

Performance of ventilators for noninvasive positive-pressure ventilation in children

B. Fauroux1,2, K. Leroux3, G. Desmarais4,5, D. Isabey4,5, A. Clément1,2, F. Lofaso4,6 and B. Louis4,5

1 Paediatric Pulmonary Dept, Hôpital Armand Trousseau, Assistance Publique-Hôpitaux de Paris, 2 INSERM Mixed Research Unit S-719, Université Pierre et Marie Curie, 3 ADEP Assistance, Puteaux, Paris, 4 INSERM Unit 841, 5 Paris XII University, Créteil, and 6 Dept of Clinical Physiology, Hôpital Raymond Poincaré, Assistance Publique-Hôpitaux de Paris, University of Versailles Saint-Quentin-en-Yvelines, Garches, France.

CORRESPONDENCE: B. Fauroux, Assistance Publique-Hôpitaux de Paris, Hôpital Armand Trousseau, Paediatric Pulmonary Dept, Research Unit INSERM UMR S-719, Université Pierre et Marie Curie Paris 6, 28 avenue du Docteur Arnold Netter, Paris, F-75012, France. Fax: 33 144736174. E-mail: brigitte.fauroux{at}trs.aphp.fr

Keywords: Bench study, child, lung model, pressure support, trigger, volume-targeted ventilation

Received: November 2, 2007
Accepted February 7, 2008


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Support statement
 Statement of interest
 ACKNOWLEDGEMENTS
 REFERENCES
 
The aim of the present study was to evaluate the performance characteristics of all the ventilators proposed for home noninvasive positive-pressure ventilation in children in France.

The ventilators (one volume-targeted, 12 pressure-targeted and four dual) were evaluated on a bench which simulated six different paediatric ventilatory patterns. For each ventilator, the quality of the inspiratory and expiratory trigger and the ability to reach and maintain the preset pressures and volumes were evaluated with the six patient profiles.

The performance of the ventilators showed great variability, and depended upon the type of trigger (flow or pressure), type of circuit and patient profile. Differences were observed between the preset and measured airway pressure and between the tidal volume measured by the ventilator and on the bench. Leaks were associated with an inability to detect the patient’s inspiratory effort or autotriggering. No single ventilator was able to adequately ventilate the six paediatric profiles. Only a few ventilators were able to ventilate the profiles simulating the youngest patients.

A systematic paediatric bench evaluation is recommended for every ventilator proposed for home ventilation, in order to detect any dysfunction and guide the choice of the appropriate ventilator for a specific patient.

Noninvasive positive-pressure ventilation (NPPV) is increasingly used at home in children 1. NPPV may improve respiratory failure in children with neuromuscular disease 2, 3, upper airway obstruction and sleep apnoea 4, and lung diseases such as cystic fibrosis 5. These diseases concern both infants and older children, which implies that the ventilator should be able to adapt to a broad range of patient demands. Children with respiratory failure, especially the youngest ones, may develop extreme breathing patterns, which may represent a challenge for a ventilator 6. Indeed, home ventilators may not be able to adequately synchronise with patient respiratory effort 7, 8, leak compensation may be insufficient, and the triggers of assist modes and alarms are not always adapted for young children. This is explained by the fact that most ventilators have not been specifically developed for paediatric patients. However, in practice, the clinician has to deal with the available devices.

Although some studies have tested or compared home ventilators in young patients with cystic fibrosis 7, 8 or upper airway obstruction 6, no study has evaluated different types of ventilator in children with various causes of chronic respiratory insufficiency. In France, 17 ventilators are proposed for home ventilation in children. Thus, the choice of the most appropriate ventilator for a specific patient is a real challenge for the clinician. Indeed, the testing of several ventilators in every single patient is unrealistic in practice.

The aim of the present study was to evaluate the performance of the 17 ventilators available for home ventilation in France with the most common paediatric profiles, namely neuromuscular disease, upper airway obstruction and cystic fibrosis. In order to do this, a bench lung model that simulated the mechanical respiratory characteristics and pattern of breathing of six typical paediatric patient profiles was used.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 Support statement
 Statement of interest
 ACKNOWLEDGEMENTS
 REFERENCES
 
Patient profiles
In the present authors’ experience, approximately a third of the children treated with NPPV at home have neuromuscular diseases, a third upper airway obstruction and a final third lung diseases or other causes of chronic hypercapnic respiratory insufficiency 1. Thus, six patient profiles representing ~90% of patient profiles experienced were selected from the present authors’ NPPV cohort (table 1Go).


View this table:
[in this window]
[in a new window]

  		Table 1— Patient profiles and ventilatory modes used for the bench lung model study#

 
During routine initiation of NPPV and follow-up, breathing pattern at baseline, respiratory mechanics and respiratory output were recorded using a pneumotachograph (Fleisch No. 3; Fleisch, Lausanne, Switzerland) and a catheter-mounted pressure transducer system with two integral transducers (Gaeltec, Dunvegan, UK). Breathing pattern at baseline, i.e. when the patient was not connected to a ventilator and breathing spontaneously, was inferred by measuring the patient flow rate. Tidal volume (VT) and inspiratory time (tI) were directly deduced from this flow tracing (table 1Go).

The patient’s respiratory mechanics were inferred when the patient was connected to the ventilator via measurement of transdiaphragmatic pressure and oesophageal pressure (Poes) as previously described 5. Briefly, dynamic lung compliance (CL,dyn) was calculated as the ratio of VT to the Poes difference between the beginning and end of inspiration during quiet breathing. Individual values indicated in table 1Go were averaged on the basis of 10–20 consecutive cycles during air breathing.

Airway and lung resistance (Rrs) was calculated according to the following formula, based on the technique of Mead and Whittenberger 9.

Rrs = [(Poes,0-Poes)-(V/CL,dyn)]/V'(1)

Poes,0 is Poes at the start of inspiratory flow, V is instantaneous volume, CL,dyn is calculated for the same breath and V' is instantaneous airflow. Mean values over the inspiration were used as estimates of inspiratory Rrs (table 1Go).

The analysis of the patients’ profiles was approved by the ethics committee of Saint Antoine University Hospital (Paris, France), and patients and parents gave their informed consent.

Ventilator testing
In total, 17 ventilators were tested: 12 pressure-targeted, one volume-targeted and four capable of both modes (table 2Go). Each ventilator was tested with the six different patient profiles and with the recommended circuits. When assisted controlled ventilation (ACV) and pressure-support (PS) ventilation (PSV) were available, both modes were tested.


View this table:
[in this window]
[in a new window]

  		Table 2— Ventilators tested

 
The ventilator setting (targeted pressure or volume and positive end-expiratory pressure (PEEP)) was different for each patient profile (table 1Go). The first two patients, a 4-yr-old male with spinal muscular atrophy and a 17-yr-old male with Duchenne muscular dystrophy, had neuromuscular disease. Since both ACV and PSV may be used in such patients, ventilators able to deliver one or both modes were tested. For patient No. 3, with cystic fibrosis, ventilators able to deliver PSV and/or ACV were tested. For these first three patients, PSV or ACV with zero end-expiratory pressure (ZEEP) was chosen because of the absence of or a low (<2 cmH2O) intrinsic PEEP 5, 10. Patient No. 4 was an infant with laryngomalacia in whom only ventilators able to deliver PS with PEEP were tested. PSV with PEEP ventilators were tested in patient No. 5, who had obstructive sleep apnoea due to vocal cord paralysis. All of the PSV ventilators able to deliver ZEEP were tested in patient No. 6, who had central apnoea.

All ventilators were studied using their most sensitive inspiratory trigger that did not induce autotriggering. When possible, the highest inspiratory flow was used. For the majority of the ventilators, the expiratory trigger was set automatically. In four ventilators (GK 425ST, KnightStar 330, Vivo 40 and VPAP III ST-A), it was possible to modify the sensitivity of the expiratory trigger. In such cases, the most sensitive level that did not induce a tI inferior to the spontaneous tI was used. Where available on the same ventilator, pressure- and flow-triggering were tested. In the case of an optional integrated humidification system, the ventilator was tested with and without the humidification system. For all of the ventilators, the most recent model (year 2006) was used.

Experimental bench study
Each ventilator tested was connected via its standard circuit to the first, testing, chamber of a two-chamber Michigan test lung (MII Vent Aid TTL; Michigan Instruments, Grand Rapids, MI, USA; fig. 1Go). The second, driving, chamber of the test lung was connected to a flow-rate generator that could produce various waveforms previously stored in a microcomputer. The two chambers were physically connected to each other via a small metal component that permitted the driving chamber to lift the testing chamber. The flow-rate generator, developed by INSERM Unit 841 (Créteil, France) as previously described, was built by associating pressurised air, flow-rate measurement and a servo-valve driven by a microcomputer 11. This permits continuous adjustment of the servo-valve in order to produce the desired flow for each patient profile, as indicated in the Patient profiles section. Moreover, in order to simulate the mechanical characteristics of the respiratory system of each patient, the compliance of the testing chamber was adjusted and a resistance added between the testing chamber and the ventilator tested. The compliance of the testing chambers (compliance of the respiratory system; Crs) was set according to the following formula, where Cw is the theoretical chest wall compliance, which represents ~4% of the patient’s predicted vital capacity per cmH2O, and CL is the lung compliance corresponding to the patient’s CL,dyn.


Figure 1
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 1— Lung bench model used for the study. V': flow; P: pressure.

 
1/Crs = 1/Cw+1/CL(2)

The resistance was a parabolic airway resistor, Pneuflo® airway resistor Rp5, Rp20, Rp50 or Rp200 (Michigan Instruments). For each profile, the resulting breathing effort generated in the bench test was characterised by the inspiratory airway occlusion pressure at 0.1 s (P0.1), and by the inspiratory volume and flow 0.1 s after initiation of a spontaneous breath (V0.1 and V'0.1, respectively; table 1Go). P0.1 was inferred when the ventilator under test and its circuit (fig. 1Go) were replaced by a rigid stopper, and V0.1 and V'0.1 were inferred when the lung test was opened to the atmosphere. A leak valve was added to simulate leaks that could occur through a mask during NPPV, which permitted the testing of an increasing leak.

Airway pressure (Paw) and flow were measured at the end of the ventilator circuit using, respectively, a pressure differential transducer (Validyne DP 45±56 cmH2O; Validyne Northridge, CA, USA) and a pneumotachograph (Fleisch No. 2) associated with a pressure differential transducer (Validyne DP 45±3.5 cmH2O). The leak flow was measured using a second pneumotachograph. Calibration of pressure and flow was performed before each test. Signals were digitised at 200 Hz by an analogue/digital system (MP100; Biopac Systems, Goleta, CA, USA) and recorded on a microcomputer for further analysis.

As is generally the case, the following parameters were computed from each pressure and/or flow trace: PEEP, PS for PSV, measured VT (VT,m), and VT indicated by the ventilator (VT,V). The sensitivity of the inspiratory trigger was evaluated from the trigger time delay ({Delta}t; time between onset of inspiratory effort and point of minimum Paw) and the trigger pressure ({Delta}P; pressure swing between baseline pressure and minimum Paw) 7. The sensitivity of the expiratory trigger was evaluated as the difference between the patient’s tI during spontaneous breathing and the tI during NPPV. The pressurisation slope was calculated for 150 ms from the time of minimum Paw. Each parameter was averaged on the basis of 30 respiratory cycles.

In order to facilitate interpretation of the results and guide the reader, the performances of the ventilators are presented qualitatively as follows. The inspiratory trigger was considered appropriate for a {Delta}t of ≤100 ms and {Delta}P of ≤ -1.0 cmH2O 12, acceptable for a {Delta}t of 100–150 ms and {Delta}P of -1.5–0 cmH2O or a {Delta}t of 0–150 ms and {Delta}P of -1.5– -1.0 cmH2O, and inappropriate if the ventilator did not detect the inspiratory effort or in the case of autotriggering. The coping of the ventilator with leaks was ranked as follows: 1) relatively insensitive to a leak (no triggering or autotriggering for a leak of ≥40 L·min–1); 2) moderately sensitive to a leak (no triggering or autotriggering for a leak of 10–40 L·min–1); and 3) very sensitive to a leak (no triggering or autotriggering for a leak of ≤10 L·min–1). The performance of each ventilator is also presented qualitatively as follows: appropriate for a VT,m of less than the required VT±10% for ACV, and for a measured PS (PSm) of less than the required PS±10% and pressurisation slope of ≥60 cmH2O·s–1 for PSV; acceptable for a VT,m of less than the required VT±15% for ACV, and a PSm of less than the required PS±15% and pressurisation slope of ≥40 cmH2O·s–1 for PSV; and inappropriate for nondetection of the inspiratory effort or autotriggering, or for a VT,m of ≥VT±15% for ACV, or a PSm of at least the required PS±15% and pressurisation slope <40 cmH2O·s–1 for PSV.


   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 Support statement
 Statement of interest
 ACKNOWLEDGEMENTS
 REFERENCES
 
Except in three cases, Smartair+ in the patient with cystic fibrosis (profile No. 3), Vivo 40 in the patient with vocal cord paralysis (profile No. 5), and VS Ultra double-circuit pressure trigger in the patient with central apnoea (profile No. 6), very close results were found with and without the humidification system. Therefore, the results are presented as the means obtained with and without humidification.

The complete data concerning the performance of each ventilator for the six different patient profiles are given in the supplementary material (online tables 1–6). For the patient with spinal muscular atrophy, all of the seven ventilators that had a compatible mode had inappropriate triggers (table 3Go). For the adolescent with Duchenne muscular dystrophy, only two ventilators, Eole 3 with flow trigger and Legendair, had an appropriate inspiratory trigger in the ACV mode. However, the Eole 3 was very sensitive, and the Legendair moderately sensitive, to leaks. Of the five other ventilators that had a compatible mode, all had an inappropriate trigger. For the patient with cystic fibrosis, only four ventilators had an appropriate trigger: Eole 3 with flow trigger; Legendair in both modes; Smartair+ with simple circuit; and VS Ultra in PSV mode with pressure trigger and double circuit, and in ACV mode with simple circuit. The VS Ultra had an acceptable trigger when in PSV mode with simple circuit and in ACV mode with double circuit and pressure trigger. However, the Elisée 150, NEFTIS 2 and VS Integra all had inappropriate triggers. None of the 16 PSV ventilators were able to detect the inspiratory effort of the infant with laryngomalacia. For the patient with vocal cord paralysis, four ventilators had an appropriate trigger: iSleep 22, KnightStar 330, VS Integra with simple circuit and VS Serena. However, these ventilators were either moderately sensitive (iSleep 22 and VS Serena) or very sensitive (KnightStar 330 and VS Integra with simple circuit) to leaks. For this patient profile, eight ventilators had an acceptable trigger: GK 425ST, Harmony 2, Legendair, NEFTIS 2, Synchrony, Synchrony 2, VPAP III ST-A and VS Ultra with simple circuit or double circuit with pressure trigger. Only four ventilators had an inappropriate trigger: Elisée 150, Smartair, Vivo 40 and VPAP III ST-A. Four ventilators were relatively insensitive to leaks (GK 425ST, Harmony 2, Synchrony, and Synchrony 2) and three were moderately sensitive to leaks (iSleep 22, VPAP III ST-A and VS Serena), the other five being very sensitive to leaks. None of the six ventilators that had a compatible mode had an appropriate trigger for the patient with central apnoea, although three ventilators had an acceptable trigger: Legendair, Smartair with simple circuit and VS Ultra. However, none of these ventilators coped adequately with leaks. The Elisée 150, NEFTIS 2 and VS Integra had inappropriate triggers.


View this table:
[in this window]
[in a new window]

  		Table 3— Trigger performances of the ventilators according to the six patient profiles

 
The quality of the expiratory triggers is presented in the supplementary material (online table 7). The major observation is that the performance of the expiratory triggers varies according to ventilator and also to patient profile. Only the KnightStar 330 and the Legendair were able to detect the expiratory effort of patient No. 4, the infant with laryngomalacia. The expiratory trigger of the Elisée 150 was good in patient No. 2 (Duchenne muscular dystrophy) but much less so in patients No. 3 (cystic fibrosis) and 5 (vocal cord paralysis).

Concerning the performance of the ventilators, for patient No. 1 with spinal muscular atrophy, only the Elisée 150 in the ACV mode with a simple circuit had an appropriate performance (table 4Go). The performance of the NEFTIS 2 and VS Ultra with a double circuit and flow trigger were acceptable, while four ventilators had an inappropriate performance (Eole 3, Legendair, Smartair+, and VS Integra). The Elisée 150, the Legendair in the ACV mode, and the VS Ultra in the PSV mode with a simple circuit or a double circuit with flow trigger had an appropriate performance. The performance of the VS Ultra in the ACV mode with a double circuit and flow trigger was acceptable, whereas the performance of the NEFTIS 2, the Smartair+ and the VS Integra was inappropriate. For the patient with cystic fibrosis, only three ventilators had an appropriate performance: Eole 3, NEFTIS 2, and VS Ultra with the two modes with a simple circuit or a double circuit with flow trigger. The KnightStar 330 was the only ventilator having an acceptable performance in the infant with laryngomalacia. For the patient with vocal cord paralysis, two ventilators had an acceptable performance: VPAP III ST-A and VS Ultra with a simple circuit or with a double circuit and pressure trigger. All the other devices with a compatible mode had inappropriate performances. For the patient with central apnoea, the Elisée 150 with a double circuit and the VS Ultra with a double circuit and flow trigger had appropriate performances, whereas the other four ventilators that had a compatible mode had inappropriate performances.


View this table:
[in this window]
[in a new window]

  		Table 4— Performance of the ventilators according to the six patient profiles

 

   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
Support statement
 Statement of interest
 ACKNOWLEDGEMENTS
 REFERENCES
 
The current study is the first to provide a bench test evaluation of the performance of a broad range of home ventilators, none of which were primarily developed for children, for six different paediatric patient profiles according to a strict protocol. The major findings of the present study can be summarised as follows: 1) no ventilator is perfect and able to adequately ventilate the six different patient profiles; 2) the performance of the ventilators was very heterogeneous and depended upon the type of trigger and circuit and, most importantly, upon the characteristics of the patient; and 3) the sensitivity of the inspiratory triggers of most of the ventilators was insufficient for infants.

Paediatric specificities
The present study confirms the limitations of the ventilators currently available for home ventilation in children. Numerous ventilators were unable to respond adequately to the patient’s demands. Several paediatric specificities may explain these difficulties. For example, the patient’s inspiratory effort may be too low, or lower than that of adults, reducing the ability of the ventilator to detect the onset of inspiration. For the six patient profiles, P0.1 in the lung model ranged 0.4–4.3 cmH2O. This is in agreement with values reported in the literature for adults 13. A recent study observed that inspiratory effort, evaluated by P0.1, was higher in children with neuromuscular disease than in healthy controls 14. However, when each patient’s P0.1 was assessed in terms of the number of ventilators detecting the patient’s inspiratory effort, it was observed that the patients who exhibited the lowest P0.1 were also those in whom the majority of the ventilators were unable to detect the patient’s inspiratory effort. Indeed, only 39% of the ventilators were able to detect the inspiratory effort of patient No. 1 (online table 1Go), who had a P0.1 of 0.94 cmH2O, and only 9% of the ventilators were able to detect the inspiratory effort of patient No. 4 (online table 4Go), who had a P0.1 of 0.4 cmH2O. This suggests that the inspiratory effort generated by the youngest children may be too small to be detected by the majority of ventilators. Moreover, these two patients also exhibited the lowest V0.1 and V'0.1 during spontaneous breathing (with V0.1 of 5.8 and 1.3 mL, and V'0.1 of 17 and 17 mL·s–1, for patients No. 1 and 4, respectively). This implies that a ventilator with a trigger based upon a flow signal should be able to detect a flow and/or volume inferior to these values in order to generate an adequate {Delta}t. In practice, the use of a high back-up rate, i.e. equivalent to or two or three breaths below the patient’s spontaneous respiratory frequency, may overcome the problems associated with an inadequate inspiratory trigger. Such a setting is recommended for patients with neuromuscular disease 15.

The patient with central apnoea should, theoretically, have been ventilated using a controlled mode. However, such patients may take some spontaneous breaths. Thus, in order to increase the comfort of NPPV and favour the synchronisation of the patient with the ventilator, a spontaneous mode with a back-up rate slightly below the spontaneous respiratory frequency of the patient may be used, permitting the evaluation of the inspiratory trigger in this patient.

These limitations of the ventilators observed in the present study with simulated paediatric patterns were not completely unexpected, since few devices have been specifically developed for children. In addition, the majority of the manufacturers (12 out of 17) do not implicitly recommend ventilation of the youngest children with their ventilator (with the ventilators being denoted adult/child, not for newborn, or >30 kg). The quality of the inspiratory triggers may also limit the performance of ventilators. Nevertheless, due to the lack of information disclosed by the manufacturers concerning the principle and algorithms used for the inspiratory trigger, it is difficult to understand why one ventilator seems to exhibit a better trigger than another. With a classical pressure trigger, a closed system is mandatory in order to facilitate the generation of a differential pressure. For example, in the case of the Eole 3 pressure trigger, no large decrease in Paw was observed during the patient’s inspiratory effort while an inspiratory flow signal was detected. This confirms that it is an open system, which is one explanation for the lack of detection of the inspiratory effort observed with this ventilator. With a trigger based upon flow signal, the system should be open. One of the major problems of such a trigger is the take-up of the leak. Nevertheless the present results do not suggest that a simple circuit plus leak permitted a better or worse inspiratory trigger than ventilation without leak (with a simple or double circuit). In the case of a flow trigger, the ventilator should be able to detect very low flows, especially in young children who have the smallest VT. Significant differences with regard to the expiratory triggers were also observed. These results are in agreement with clinical results, which showed that the sensitivity of the expiratory triggers may be insufficient for infants requiring NPPV for severe upper airway obstruction 6.

Characteristics of the ventilators
Ventilators are becoming more sophisticated and tend to continuously integrate new options and measures. A large number of ventilators are able to deliver different ventilatory modes, such as PSV, with or without PEEP, as well as ACV. Different circuits (simple, double or leak) and triggers (pressure or flow) may be available on the same ventilator. The present study clearly shows that the performance of a ventilator may vary according to the ventilatory mode or type of trigger and circuit. Indeed, the quality of the inspiratory trigger varied among the different ventilators, and also for a specific ventilator, according to patient profile. For example, the {Delta}t of the NEFTIS 2 was shorter in patient No. 6, who showed high inspiratory effort, than in patient No. 1, who showed low inspiratory effort (0.15 versus 0.28 s, respectively; online tables 1Go and 6). It is important that the clinician who chooses the device is aware of these differences, which are rarely specified by the manufacturer.

Some ventilators, such as the Legendair, showed a low pressurisation slope and stability index, which signifies that the ventilator is not able to reach the preset pressure within a minimal time frame. Most ventilators measure physiological variables, such as VT or Paw. Significant differences were observed for almost all of the ventilators between the results shown on the ventilator and the values measured on the bench. This may be explained by the fact that most of these variables are estimated by software incorporated inside the ventilator. Since NPPV is leak ventilation, the VT,V represents the volume of air generated by the device. On the bench, the VT,m was measured by a pneumotachograph inserted between the circuit and the interface. Thus, this measure was closer to the patient and more accurately reflects the VT received by the patient in the case of calibrated leak ventilation. However, differences between the VT set on the ventilator and the VT measured by the ventilator and by a pneumotachograph have also been observed previously with other ventilatory modes 16. It should be noted that less discrepancy was observed for Paw.

The ability of a ventilator to compensate for additional leaks is important in the case of NPPV. Therefore, the effect of an additional leak in the inspiratory circuit was tested for every ventilator. Most of the ventilators were unable to cope with additional leaks, which resulted in autotriggering or an inability to detect the patient’s inspiratory effort.

Advantages and limitations of the study
The responses of several devices to identical mechanical properties of the respiratory system and patterns of inspiratory flow contour were compared, which is not possible in a clinical study given the variability of these respiratory parameters. In addition, given the number of ventilators available for testing, it would be unreasonable to ask children to perform such a study.

One limitation of the bench is that the resistance added by the test system may be more representative of upper airway obstruction, as encountered in the patients with laryngomalacia and vocal cord paralysis, than small airway disease, such as encountered in the patient with cystic fibrosis.

Another limitation of the present study was that the six patients were recorded during wakefulness and not during sleep. Sleep may be associated with both upper airway and inspiratory effort instability. Thus, the mechanical output occurring during spontaneous respiratory drive, i.e. the inspiratory flow or airway depression that the ventilator has to detect in order to synchronise the ventilatory assistance to the patient’s inspiratory effort, may be less easy to detect during sleep. Recording the patients during sleep was refrained from, although NPPV is generally performed during sleep, since NPPV is initially started and adapted during wakefulness, before being tested during sleep. In addition, typical patient profiles were used, but, in clinical practice, the presence of several factors favouring nocturnal hypoventilation is a common situation, such as the association of obesity and upper airway obstruction in patients with Duchenne muscular dystrophy. It was not possible to integrate such mixed pathologies in the present bench model. It was also not possible to include dynamic modifications, such as upper airway obstruction and decrease in respiratory drive during sleep. If there is confidence that the ventilators that were unable to detect the simulated respiratory efforts would also be unable to detect respiratory efforts under real-life conditions, it cannot be ascertained that the ventilators considered appropriate by the bench study were effectively appropriate under real-life conditions. Therefore, the present study only permits preselection of ventilatory devices which can be reasonably tested in a paediatric patient, and cannot exclude a clinical evaluation before considering that a ventilator is really appropriate for a child.

Nevertheless, a systematic comparison of bench data with in vivo data is lacking. However, for most typical situations, the in vitro results are in agreement with in vivo patient tracings. Indeed, the lack of detection of the patient’s inspiratory and expiratory effort by the majority of the bilevel devices in infants and young children has been previously observed 6. The insufficient sensitivity of the inspiratory trigger of the Eole 3 XLS has also been observed in young patients with cystic fibrosis 7. Moreover, the majority of the problems encountered with the various ventilators during the bench testing have been observed in patients 6.

Practical recommendations
The present results underline the necessity of a systematic bench evaluation of all ventilators proposed for NPPV in children. This evaluation should ideally include assessment of the quality of the inspiratory (pressure and/or flow) trigger and the ability of the ventilator to reach and maintain the preset volume or pressure, as well as to cope with leaks. However, for some patients, such as patients with neuromuscular disease or central apnoea, effort-independent modes are recommended, precluding the evaluation of the inspiratory trigger. This bench evaluation should be followed by a clinical evaluation in the patients for whom the ventilator has shown good or acceptable performances, as defined by specific criteria, e.g. those proposed in the present study.

The choice of a ventilator for a specific patient depends upon the patient’s characteristics (underlying disease, age and weight), the ventilatory mode to be used and the performance of the ventilator. Other ventilator characteristics, not evaluated in the present study, such as the accuracy of the alarms and the possibility of humidification or additional oxygen therapy, should also be taken into account. Finally, ergonomics, such as transportability and internal battery, are important in clinical use. However, ultimate efficacy must be checked in each individual case by daytime performance and comfort, associated with overnight control.

Conclusion
The present bench study, which, for the first time, evaluated 17 home ventilators for the six most common paediatric profiles, shows that the performance of the ventilators varied according to not only ventilator characteristics (type of circuit and trigger) but, most importantly, also patient profile, including age and weight, as well as underlying disease. Even if different modes and different ventilators may be used in a specific patient, a systematic bench evaluation, coupled to a clinical in vivo evaluation, is recommended for all ventilators proposed for home noninvasive positive-pressure ventilation in children.


   Support statement
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Support statement
Statement of interest
 ACKNOWLEDGEMENTS
 REFERENCES
 
B. Fauroux is supported by the Association Française contre les Myopathies (Evry, France), the Assistance Publique-Hôpitaux de Paris (Paris, France), the Institut National de la Santé et de la Recherche Médicale (INSERM), Legs Poix (Paris) and the Université Pierre et Marie Curie Paris 6 (Paris).


   Statement of interest
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Support statement
 Statement of interest
ACKNOWLEDGEMENTS
 REFERENCES
 
None declared.


   ACKNOWLEDGEMENTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Support statement
 Statement of interest
 ACKNOWLEDGEMENTS
REFERENCES
 
The authors would like to thank E. Cohen, C. Justine and C. Boniface (INSERM Mixed Research Unit S-719, Université Pierre et Marie Curie, Paris, France) for their excellent technical assistance.


   FOOTNOTES
 
This article has supplementary material accessible from www.erj.ersjournals.com


   REFERENCES
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Support statement
 Statement of interest
 ACKNOWLEDGEMENTS
 REFERENCES
 

  1. Fauroux B, Boffa C, Desguerre I, Estournet B, Trang H. Long-term noninvasive mechanical ventilation for children at home: a national survey. Pediatr Pulmonol 2003;35:119–125.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
  2. Simonds A, Muntoni F, Heather S, Fielding S. Impact of nasal ventilation on survival in hypercapnic Duchenne muscular dystrophy. Thorax 1998;53:949–952.[Abstract/Free Full Text]
  3. Mellies U, Ragette R, Dohna Schwake C, Boehm H, Voit T, Teschler H. Long-term noninvasive ventilation in children and adolescents with neuromuscular disorders. Eur Respir J 2003;22:631–636.[Abstract/Free Full Text]
  4. Marcus CL, Rosen G, Ward SL, et al. Adherence to and effectiveness of positive airway pressure therapy in children with obstructive sleep apnea. Pediatrics 2006;117:e442–e451.[Abstract/Free Full Text]
  5. Fauroux B, Pigeot J, Isabey D, Harf A, Clément A, Lofaso F. In vivo physiological comparison of two ventilators used for domiciliary ventilation in children with cystic fibrosis. Crit Care Med 2001;29:2097–2105.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
  6. Essouri S, Nicot F, Clément A, et al. Noninvasive positive pressure ventilation in infants with upper airway obstruction: comparison of continuous and bilevel positive pressure. Intensive Care Med 2005;31:574–580.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
  7. Fauroux B, Louis B, Hart N, et al. The effect of back-up rate during non-invasive ventilation in young patients with cystic fibrosis. Intensive Care Med 2004;30:673–681.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
  8. Fauroux B, Nicot F, Essouri S, et al. Setting of pressure support in young patients with cystic fibrosis. Eur Respir J 2004;24:624–630.[Abstract/Free Full Text]
  9. Mead J, Whittenberger J. Physical properties of the human lung measured during spontaneous respiration. J Appl Physiol 1953;5:779–796.[Free Full Text]
  10. Hart N, Polkey MI, Clément A, et al. Changes in pulmonary mechanics with increasing disease severity in children and young adults with cystic fibrosis. Am J Respir Crit Care Med 2002;166:61–66.[Abstract/Free Full Text]
  11. Lofaso F, Desmarais G, Leroux K, et al. Bench evaluation of flow limitation detection by automated continuous positive airway pressure devices. Chest 2006;130:343–349.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
  12. Richard JC, Carlucci A, Breton L, et al. Bench testing of pressure support ventilation with three different generations of ventilators. Intensive Care Med 2002;28:1049–1057.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
  13. Tobin MJ. Respiratory monitoring in the intensive care unit. Am Rev Respir Dis 1988;138:1625–1642.[Web of Science][Medline] [Order article via Infotrieve]
  14. Mulreany LT, Weiner DJ, McDonough JM, Panitch HB, Allen JL. Noninvasive measurement of the tension–time index in children with neuromuscular disease. J Appl Physiol 2003;95:931–937.[Abstract/Free Full Text]
  15. Clinical indications for noninvasive positive pressure ventilation in chronic respiratory failure due to restrictive lung disease, COPD, and nocturnal hypoventilation–a consensus conference report. Chest 1999;116:521–534.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
  16. Lofaso F, Fodil R, Lorino H, et al. Inaccuracy of tidal volume delivered by home mechanical ventilators. Eur Respir J 2000;15:338–341.[Abstract]



This article has been cited by other articles:


Home page
 Eur Respir JHome page
R. Farre and D. Navajas
Quality control: a necessary, but sometimes overlooked, tool for improving respiratory medicine
Eur. Respir. J., April 1, 2009; 33(4): 722 - 723.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary data
Right arrow All Versions of this Article:
31/6/1300    most recent
09031936.00144807v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Permissions
Right arrowRequest Permissions
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fauroux, B.
Right arrow Articles by Louis, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fauroux, B.
Right arrow Articles by Louis, B.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS