Evaluating noninvasive ventilation using a monitoring system coupled to a ventilator: a bench-to-bedside study

C. Rabec, M. Georges, N. K. Kabeya, N. Baudouin, F. Massin, O. Reybet-Degat, P. Camus
European Respiratory Journal 2009 34: 902-913; DOI: 10.1183/09031936.00170508

Abstract

Empirically determined noninvasive ventilation (NIV) settings may not achieve optimal ventilatory support. Some ventilators include monitoring modules to assess ventilatory quality. We conducted a bench-to-bedside study to assess the ventilatory quality of the VPAPTMIII-ResLinkTM (ResMed, North Ryde, Australia).

We tested the accuracy of minute ventilation (MV) and leak calculations given by VPAPTMIII-ResLinkTM compared to those measured by a bench model at varied leak levels and ventilator settings. We systematically assessed NIV efficacy using this system from 2003 to 2006. Ventilation was considered inadequate if leak (>24 L·min−1), continuous desaturation (>30% of the trace) or desaturation dips (>3%) were present.

On the bench test, both methods were highly correlated (r = 0.947, p>0.0001 and r = 0.959, p<0.0001 for leak and MV, respectively). We performed 222 assessments in 169 patients (aged 66.42±16 yrs, 100 males). Abnormalities were detected on 147 (66%) out of 222 occasions. Leak was the most common abnormality (34.2%) followed by desaturation dips (23.8%). The most effective therapeutic solutions were a chin strap if leak was detected (61.2%) and expiratory positive airway pressure increase for desaturation dips (59.5%). In 15.7% of cases, when abnormalities persisted, a polygraphy was performed.

The systematic use of this device enables NIV to be optimised, limiting the indication of sleep studies to complex cases.

Noninvasive ventilation (NIV) has been demonstrated to be effective treatment in respiratory failure. In some cases, however, the clinical results may be less than expected, despite using the correct technique. When NIV is initiated, the ventilator settings are determined empirically based on clinical evaluation of underlying disease, patient tolerance when awake, and diurnal blood gas variations 1. However, NIV is predominantly applied at night 2, when profound ventilatory changes may occur, particularly in patients with respiratory failure 3. Such changes include modifications of ventilatory control, upper airway patency and respiratory muscle recruitment. Consequently, modulating NIV settings during the day and underestimating these physiological differences may lead to suboptimal patient–ventilator interaction that reduces NIV efficacy. Moreover, NIV uses a non-hermetic technique, which poses the possible risk of leak. Leakage may be absent or minimal when the patient is awake and may worsen during sleep as a result of the loss of voluntary control and decreased muscle tone. Thus, NIV settings chosen empirically on daytime evaluation may not predict optimal nocturnal ventilatory support. Consequently, NIV effectiveness may be more correctly assessed by sleep studies than through daytime assessment 1. Ideally this requires complete polysomnography (PSG) or ventilatory polygraphy (PG) at the time of initiation to NIV. However, it is not technically feasible to perform repeated long-term assessment by PG/PSG in these patients.

Recently, NIV has been provided by portable ventilators capable of meeting high ventilatory demands 4. Some of these ventilators include monitoring modules that can assess various ventilatory parameters, storing the data in memory for subsequent analysis. Such technology and its clinical usefulness merits critical analysis.

We wanted to assess one of these systems, which consisted of a ventilator (VPAPTMIII) and its monitoring module (ResLinkTM). We assessed the accuracy of the parameters measured by this device in a bench model. In addition, we investigated the results of using the system by systematically assessing the efficacy of the NIV in different clinical situations.

MATERIALS AND METHODS

The VPAPTMIII-ResLinkTM (ResMed, North Ryde, Australia) is a bi-level ventilator that includes a monitoring system which provides trend data on minute ventilation (MV) (defined as estimated MV received by the patient after the removal of leak), respiratory rate, air leaks (above “intentional” ones), arterial oxygen saturation measured by pulse oximetry (Sp,O2) and heart rate; all of which are analysed by the software package AutoScan 5.7TM (ResMed). The monitoring module, ResLinkTM, obtains the information from signals received at the mask and only requires an additional connection to a pulse oximeter (included). The oximeter has a sampling frequency of 1 Hz and an average of the Sp,O2 value is made over the last four beats. For the oxygen desaturation index, the averaging window is 5 s in duration. The data can be stored on a SmartMedia card (Sandisk, Milpita, CA, USA) and transferred to a computer for simultaneous onscreen viewing.

Bench study

In an artificial lung model, we tested the accuracy of the MV and leak estimation calculated by ResLinkTM, and compared these results with a reference system.

The artificial lung is composed of two 40-L Nalgene® translucent polycarbonate canisters (Bioblock Scientific, Illkirch, France). Details on the lung prototype can be seen in figure 1⇓ of the supplementary data. The ventilator and monitoring module were connected to the lung model by a 60-cm long section of tubing. We included the following in the circuit. 1) A variable-opening valve providing a variable leak of 0–180 L·min−1 joined by a Y-connection to a first pneumotachograph (RT200TM; Timeter Instruments, St Louis, MO, USA). 2) A second flow-measurement system located downstream and capable of measuring the actual MV. A ventilator-testing device (VentestTM; Soderel, Heillecourt, France) was used to measure MV and calculate compliance and resistance. To simulate physiological ventilation conditions, a leak device corresponding to the intentional leak of the mask stated in the system was included (“calibrated leak”). With the VPAPTMIII-ResLinkTM, the type of mask and circuit length could be registered, which enables the device to “know” the value of the intentional leak and subtract it from the estimated leak. We used an intentional leak corresponding to a standard mask (25 L·min−1 at 10 cmH2O), and the length of the circuit was 3 m. These were the values registered in the system. The protocol consisted of a total of 48 recordings of 10 min each, at leak levels of 0, 18, 24 and 30 L·min−1 (when the apparatus was tested in a bi-level mode, this value corresponded to the maximum leak during the inspiratory phase) at different ventilatory settings (fig. 1⇓), resulting in varying compliance and resistance as follows. 1) Compliance: 50 mL·cmH2O−1, resistance: 4 cmH2O·L−1·s−1 (physiological); 2) compliance: 50 mL·cmH2O−1, resistance: 15 cmH2O·L−1·s−1 (increased resistance); and 3) compliance: 25 mL·cmH2O−1, resistance: 4 cmH2O·L−1·s−1 (reduced compliance). ResLinkTM estimates averaged values while the pneumotachograph measures instant flow in inspiration and expiration. Therefore, we expressed a mean leak for pneumotachograph measures. This was the maximum instantaneous leak in inspiration and expiration weighted to the duration of inspiratory and expiratory phases. The device was set in the “T-mode” with an inspiration/expiration ratio of 1:2.

Fig. 1—
Fig. 1—

Schematic representation of the bench model. The protocol consisted of a total of 48 recordings of 10 min each, at leak levels of 0, 18, 24 and 30 L·min−1. The different ventilatory settings are shown. The device was set in the “T-mode” with an inspiration/expiration (I/E) ratio of 1:2. CPAP: continuous positive airway pressure; IPAP: inspiratory positive airway pressure; EPAP: expiratory positive airway pressure; RR: respiratory rate; R: resistance; C: compliance. RT200TM: Timeter Instruments, St Louis, MO, USA. VentestTM: Soderel, Heillecourt, France. VPAPTMIII-ResLinkTM: ResMed, North Ryde, Australia.

Data provided by the VPAPTMIII-ResLinkTM were then recorded on a SmartCard, which was then transferred to a computer and visualised via the AutoScan 5.7TM program in order to compare them with the reference values.

Clinical study

We included patients recruited between 2003 and 2006 treated with NIV using VPAPTMIII, either in the acute setting (acute respiratory failure (ARF) group) or electively, in patients for whom home MV was indicated (chronic respiratory failure (CRF) group).

The efficacy of ventilation was monitored using ResLinkTM under ventilation conditions (parameters and interfaces) present at the time of assessment. In patients discharged home with NIV, another trace was performed 4–6 months later to confirm NIV efficacy (long-term mechanical ventilation (LTMV) group) under the parameters used at home. The ResLinkTM criteria used to define ventilation as ineffective were as follows: 1) leakage >24 L·min−1 for >20% of the trace duration; 2) continuous desaturation (>30% of the trace with Sp,O2 <90%), whether or not accompanied by a simultaneous reduction of MV (>10% of reduction of MV compared to baseline) estimated by the ResLinkTM, in the absence of significant leak; and 3) cumulated desaturation dips (>3%) for >10% of the trace duration. If these criteria were not met then the ventilation was considered effective. The data were analysed using the AutoScan 5.7TM software package. If one of these abnormality criteria were detected, one or more conditions were modified to optimise ventilation and another recording was performed. This procedure was repeated until the best possible results were obtained. When the trace could not be corrected despite these changes, we performed PG/PSG under ventilation. This PG/PSG included, at least, pneumotachograph-based airflow, mask pressure, thoraco-abdominal movements and Sp,O2 recordings. Typical ReslinkTM traces of the various abnormalities are shown in figure 2⇓. Arterial blood gases (ABG) during spontaneous ventilation were measured under the conditions of oxygenation at the time of assessment. In the ARF and CRF patient groups, ABG were sampled before the start of ventilation and in the LTMV group during the first hour in the morning after the patient was disconnected from the ventilator.

Fig. 2—
Fig. 2—

VPAPTMIII-ResLinkTM System (ResMed, North Ryde, Australia) characteristic traces (2 h per page). a) Normal trace; b) trace showing leaks >0.40 L·s−1 followed by drop in arterial oxygen saturation measured by pulse oximetry (Sp,O2); c) trace showing continuous desaturation accompanied by a drop in minute ventilation (MV) as estimated by the ventilator (without leaks); d) trace showing desaturation dips (see oscillations in respiratory rate (RR) and MV as a reflection of desaturation dips).

Statistical analysis

Values are shown as mean±sd. The values for MV and leak obtained by the two methods in the bench study were correlated by linear regression using Pearson's correlation coefficient. We also calculated the bias (where d is the mean difference between reference bench measures and the ResLinkTM estimated values) and the limits of agreement between the parameters as described by Bland and Altman 5. For the clinical study, data were compared using the unpaired t-test and Mann–Whitney U-test. Differences between the groups were assessed using ANOVA. A Chi-squared analysis was performed to study categorical data. A p-value of 0.05 was considered as significant.

RESULTS

Bench study

The correlations between values obtained by pneumotachograph and those estimated by the VPAPTMIII-ResLinkTM are shown in figure 3a⇓ and b. Both sets of measurement were highly correlated (leak: r = 0.947 and p>0.0001; MV: r = 0.959 and p<0.0001), as was the correlation for data subgroups (for both MV and leak) but the correlation of leak was not as good at low leak levels (at 18 L·min−1: r = 0.853 and p<0.001). Figure 3c⇓ and d shows Bland–Altman plots of the agreement between pneumotachograph-measured leak and MV and ResLinkTM values. For all observed values of leak and MV, the mean observed bias was 0.736±1.68 and 0.077±0.97 L·min−1, respectively. The limits of agreement between both techniques were -4.04 and 2.57 L·min−1 for leak and -1.82 and 1.97 L·min−1 for MV. Again the agreement was less for low leak values (<22 L·min−1).

Fig. 3—
Fig. 3—

Pearson correlation between values of measured a) leaks (RT200TM; Timeter Instruments, St Louis, MO, USA) and b) minute ventilation (MV) (VentestTM; Soderel, Heillecourt, France) and those estimated by the VPAPTMIII-ResLinkTM system (ResMed, North Ryde, Australia). a) r = 0.947, p<0.0001. b) r = 0.959, p<0.0001. Bland–Altman plots for the differences against mean for c) leaks (n = 48) and d) MV (n = 45). ––––: mean; ······: ±1.96 sd from the mean.

Bedside study

NIV efficacy was assessed for 222 different hospital admissions in 169 patients (aged 66.42±16 yrs, 100 males). During each admission one or more traces were performed. The total number of recordings performed was 542. Amongst them, 500 traces were useful for analysis (2.25±1.45 traces·admission−1) and 42 (8.4%) traces were not suitable for analysis because of technical problems. 69 (31.08%) assessments were performed in patients in the ARF group (180 valid traces, 2.6±1.72 traces·patient−1), 53 (23.8%) in the CRF group (147 valid traces, 2.77±1.75 traces·patient−1) and 100 (44.59%) in those from the LTMV group (173 valid traces, 1.73±0.95 traces·patient−1). Oxygen therapy was added on 112 occasions to ensure diurnal arterial oxygen tension (Pa,O2) >60 mmHg (69 out of 69 in ARF, 20 out of 53 in CRF and 23 out of 100 in LTMV). The difference between the numbers of traces per admission was significant between ARF and LTMV and between CRF and LTMV, but not between ARF and CRF (table 1⇓). 50 patients were assessed on two different occasions and two patients on three occasions. A nasal mask (NM) was used on 167 (75.2%) occasions and a facial mask (FM) on 55 (24.8%). Patients in the ARF group are significantly older than the patients in the other groups.

View this table:
Table 1—

Ventilation settings, arterial blood gases and number of studies performed, classified by clinical situation

Table 2⇓ illustrates the classification of patients according to the categories defined in the Eurovent survey 6. Patient characteristics, ventilatory parameters and ABG values are shown in table 1⇑. There were no significant differences in ventilatory settings for the three groups. However, differences in the types of interfaces used were observed, with a gradual increase in the use of NMs compared to FMs as patients progressed from an acute to a more stable situation (nasal mask ARF<CRF<LTMV) (p<0.01).

View this table:
Table 2—

Classification of disease categories for patients included in the study according to the Eurovent survey6

Results of the ResLink monitoring traces

Abnormalities were detected on 147 (66%) of the 222 occasions. In the group as a whole, 129 (76.3%) out of 169 patients presented with abnormalities in some of the assessments, most commonly in the ARF (53 (76.8%) out of 69 occasions) and CRF (41 (77.3%) out of 53 occasions) groups compared to the LTMV group (53 (53.0%) out of 100 occasions). The differences in the number of abnormalities were significant between ARF and LTMV and between CRF and LTMV (p<0.05) but not between ARF and CRF. These abnormalities were significantly more common when using a NM in ARF (p<0.02) but not in CRF or LTMV. The proportion of patients with suboptimal ventilation was similar, independent of the disease category or ventilatory parameters. There were no differences regarding subjective perception of quality of sleep as evaluated by a categorical scale (poor-fair-good) between patients with and without abnormalities.

Leak was the most common abnormality in the overall population (76 (34.2%) occasions) and also in the ARF (30 (43.4%) occasions) and CRF (21 (39.6%) occasions) groups. In the LTMV group, leak was less common (25 (25%) occasions), and the number of cases was significantly smaller compared with ARF (p<0.02) but not with CRF. Leak seems to be unrelated to inspiratory positive airway pressure (IPAP) levels and was significantly more common in the overall population when a NM was used (p<0.01). The second most common abnormality was 3% desaturation dips that were observed on 53 (23.8%) occasions with no significant differences between the three groups. A significantly greater prevalence of desaturation dips was observed with FM than NM (p<0.04), but only the ARF group contributed to this difference. Finally, episodes of continuous desaturation were observed on 18 (8.1%) occasions with no significant differences between the three groups, and a similar prevalence of these events with both interfaces. On 10 occasions (4.5% of the total) the desaturation episodes were accompanied by a significant reduction in MV.

Figures 4⇓, 5⇓ and 6⇓ illustrate the results of the recordings, categorised by clinical situation, and subsequent therapeutic modifications. The results classified by type of abnormality and interface can be seen in table 3⇓.

Fig. 4—
Fig. 4—

Monitoring results classified by clinical situation and subsequent therapeutic modifications in the acute setting group. ARF: acute respiratory failure; EPAP: expiratory positive airway pressure; IPAP: inspiratory positive airway pressure.

Fig. 5—
Fig. 5—

Monitoring results classified by clinical situation and subsequent therapeutic modifications in the elective ventilation group. CRF: chronic respiratory failure; EPAP: expiratory positive airway pressure; IPAP: inspiratory positive airway pressure.

Fig. 6—
Fig. 6—

Monitoring results classified by clinical situation and subsequent therapeutic modifications in patients evaluated at a steady state. LTMV: long-term mechanical ventilation; EPAP: expiratory positive airway pressure; IPAP: inspiratory positive airway pressure.

View this table:
Table 3—

Monitoring results: classified by type of abnormality for each interface and effective therapeutic modifications

Comparison between patients with normal and abnormal traces

Relationship between ABG and monitoring results

Table 4⇓ shows a comparison of ABG between patients with normal and abnormal ResLinkTM traces. These were sampled prior to the start of the ventilation in ARF and CRF and on spontaneous ventilation after disconnecting the ventilator in LTMV. When ABG values in LTMV patients with abnormal traces were compared according to the type of ResLinkTM abnormality, we did not find significant differences between the three groups of abnormalities as regards arterial carbon dioxide tension (Pa,CO2) and pH. Pa,O2 was significantly worse in patients with an abnormal trace, with continuous desaturation being associated with the lowest Pa,O2, followed by the desaturation dips, and the group with leak having the least impact.

View this table:
Table 4—

Comparison in terms of arterial blood gases between patients with normal and abnormal traces

Additional data concerning the LTMV group

Comparative mean compliance

There were no differences in terms of daily duration of ventilator use between patients with normal and abnormal ResLinkTM traces (7.92±2.2 and 7.90±2.0 h·day−1, respectively).

Modifications in ventilation as a result of the monitoring

Figures 4⇑, 5⇑ and 6⇑ and table 3⇑ illustrate the modifications made as a result of monitoring, classified by situation and interface, respectively. If leaks were detected, the most common approach was to add a chin strap (49 (64%) out of 76 patients ventilated with a NM), which was effective in 30 (61.2%) patients. The efficacy of this was similar in the three clinical situations. In the case of desaturation dips, the most common approach used was to increase expiratory positive airway pressure (EPAP; 42 (79.2%) out of 53 occasions), which was effective in 25 (59.5%) out of 42 cases. For continuous desaturation, the effective modifications were to increase IPAP (10 (55.55%) out of 18 cases) or add oxygen (eight (44.4%) out of 18 cases). Only one modification was required in 81 cases, two in 33 patients and three or more in 30 patients. In 35 (15.7%) cases, it was judged necessary to perform PG/PSG under ventilation (see online supplement). The need for PG was significantly more common in ARF (19 (27.5%) out of 69) than in LTMV (seven (7%) out of 100) (p<0.01). The abnormality that most frequently indicated a PG/PSG was desaturation dips (21 (60%) out of 35 cases).

DISCUSSION

NIV is a non-hermetic system compared to invasive ventilation, and the ventilator-lung assembly cannot be considered as a single-compartment model because of the presence of variable resistance in the upper airway (UA) 7. Therefore, increasing the volume or the delivered inspiratory pressure does not necessarily result in increased effective ventilation reaching the lungs 8–10. It is therefore essential to assess the clinical efficacy of NIV in all patients and this evaluation should include at least a clinical assessment and measurement of ABG and nocturnal Sp,O2 under NIV 11, 12. However, this may not be sufficient. For example, while an abnormal Sp,O2 reflects, on the whole, inoptimal ventilation, it does not determine the underlying mechanism 11, 12. Moreover, even a “normal” Sp,O2 does not rule out the possibility of nocturnal hypoventilation, particularly if the patient is receiving supplemental oxygen 12–14. Consequently, more specific monitoring is necessary. Some authors have suggested that all patients receiving NIV should be tested by PSG to verify its efficacy 1, 15, 16 but, in reality, this practice is infrequent 3. This is on account of the workload and the lack of available PSG devices, but also due to the difficulty in interpreting the ventilatory signals under NIV, which needs a high level of experience. In addition, in order to understand patient–ventilator interactions, the PSG must include at least one signal to detect inspiratory activity (oesophageal pressure or diaphragmatic electromyography) and monitoring of this type may itself change sleep quality. Therefore, “interpreting polysomnography under NIV is a very difficult task” 17. Finally, it is impossible to perform polysomnography after every therapeutic modification. Thus, the systematic use of PSG would be very difficult. In addition there has been an exponential growth in the number of patients receiving NIV in recent years and, therefore, performing polysomnography on all ventilated patients becomes increasingly difficult.

Some portable ventilators include technologies that allow monitoring of ventilatory parameters. In this study we assessed the efficacy of NIV using the VPAPTMIII-ResLinkTM. We first wanted to evaluate the accuracy of the parameters measured in a bench model. We tested the system at different settings “prompting” various leak ranges, at different compliance and resistance levels, to reproduce situations observed in clinical practice. These tests confirmed the accuracy of this system in terms of leak detection and MV. The correlation between the bench test and ResLinkTM was less reliable with low leak levels, but, since such leak levels would not degrade the quality of ventilation they are probably not clinically relevant.

We have assessed the usefulness of the ResLinkTM device in monitoring NIV in clinical practice. This approach is pertinent since, to date, only a limited number of studies have specifically addressed the question of NIV evaluation during the night. Moreover, the validity of several parameters estimated by NIV devices is questionable and must be validated by independent clinical and/or bench test studies 17.

One can estimate that by permitting evaluation of some “critical signals” (for instance arterial oxygen saturation, leaks and MV) this system could provide useful information about the global quality of ventilation. The occurrence of oxygen desaturations (both desaturation dips and continuous desaturation) is highly sensitive at detecting breathing abnormalities in NIV users. Moreover, unintentional leaks are probably the main phenomena that impair ventilatory effectiveness 17 and are extremely frequent in patients receiving NIV 18, 19. Finally, whereas the main goal of ventilatory assistance is to ensure physiological MV, it allows us to estimate this parameter. Therefore, an analysis combining the three signals provided by this device may offer a global estimation of NIV quality. Using this procedure enabled us to observe abnormalities liable to change the efficacy of the NIV in 66% of cases. This is higher than reported in the literature (5–60%) 20, 21, perhaps because the efficacy criteria selected were very rigorous, but also because a more subtle methodology of analysis was used in order to rule out false negative results. We defined ventilation as ineffective when we observed one of the following abnormalities: leak, episodes of continuous desaturation or desaturation dips. Suboptimal ventilation was more common at the start of ventilation treatment; therefore, the number of traces and PG/PSG required, and consequently the number of modifications made, was higher in that period. The better ventilation performance observed in patients assessed at a later period may reflect greater clinical stability or a “training effect” 22, but may also reflect that many of these patients had undergone monitoring at the start of the NIV, making early intervention possible. The abnormalities were less common when an FM was used, but this difference was only significant in the acute phase. The better results with FM in the acute phase is found in clinical practice, given that in acute NIV FM are most commonly used, while for chronic NIV NM use predominates 4, 23. Finally, the proportion of patients with suboptimal ventilation was unrelated to the type of underlying disease.

In the acute setting, greater hypercapnia at the start of the ventilation correlated with a greater probability of inefficacy of NIV. This can be explained by the need, in these patients, to modify the Pa,CO2 to a greater extent, which could cause “resistance” to the ventilation, as has been described previously 24. For stable patients, there was a significant correlation between the presence of abnormalities in ResLinkTM data and a reduced efficacy of NIV, particularly in terms of Pa,O2, pH and HCO3-. As well as Pa,CO2, serum bicarbonate levels reflect the effectiveness of nocturnal ventilation 25, 26. Given that compliance in both groups (normal and abnormal) was similar, these differences in blood gases cannot be explained by a difference in the use of the ventilator. If one assumes that the primary objective of ventilation is to correct the ABG 2, these findings call attention to the importance of monitoring results as markers of ineffective ventilation.

Leak was the most commonly found abnormality as is expected in a non-hermetic system such as NIV 21, 26, and can compromise the efficacy of the method, reduce compliance and disturb sleep quality 18, 19, 27. Generally, a leak of <24 L·min−1 is considered clinically tolerable 18, 20, 25, 27. Different portable ventilators allow leak to be estimated, but the performance of their algorithms is variable 26. In addition, for most of these algorithms, intentional leaks are included in the leak estimations, which are variable and proportional to the level of pressure and different for each interface. The ResLinkTM has the particular feature of taking into account the intentional leak and subtracting it, thereby estimating the “undesired” leak level. Solutions for reducing leak include chin straps and using an FM or mouth interfaces 18. Although the use of chin straps has been reported to reduce leak volume by <50% 25, in our study and in other studies 28 this was effective in a large number of patients.

Desaturation dips reflect intermittent obstruction of the UA under NIV and are related to two mechanisms. The first corresponds to obstructive events at the oropharyngeal level because of UA collapse, as a result of insufficient EPAP. This mechanism may be present in patients with an unstable UA, such as patients with obstructive sleep apnoea 29, 30. Another mechanism corresponds to episodes of intermittent obstruction at the glottal level reflecting cyclic glottal closure induced by hyperventilation, a type of “ventilation resistance” reflex 9, 31–33. Interestingly, desaturation dips were more frequent when an FM was used, particularly in patients in an acute situation. There are several possible explanations for this. One is that greater modification of Pa,CO2 as a result of more hermetic ventilation could promote the occurrence of glottal closure episodes 24. Another is that the use of FM predominantly induces mouth breathing, which may promote significant narrowing of the retropalatal distance predisposing to oropharyngeal obstruction 34. The application of positive pressure in the mouth could worsen this situation even further 34.

The fact that it is impossible to identify the underlying mechanism of these desaturation dips explains that this abnormality was the one that led most frequently to performing a PG/PSG. In these situations, the purpose of PG/PSG was to identify the mechanism of the UA obstruction 21, 35. If obstruction occurs at the oropharyngeal level it will be accompanied by progressively increased inspiratory activity, indicating a struggle against UA collapse. Conversely, if the mechanism is glottal closure it will be accompanied by reduced inspiratory effort, the result of an excessive level of ventilation promoting respiratory pauses 31, 36. In the case of oropharyngeal apnoea, the approach is to increase the level of EPAP to stabilise the UA. In the case of glottal apnoea, however, it is to reduce the MV 31.

The third pathological ResLinkTM pattern was represented by regular desaturation episodes, whether or not accompanied by a simultaneous reduction of MV. Two mechanisms could explain this abnormality: 1) the inefficacy of the system to ensure adequate ventilation, which is generally accompanied by diurnal hypercapnia; and 2) a worsening of the ventilation perfusion imbalance, known as a cause of nocturnal hypoxaemia in patients with respiratory failure 29. Monitoring transcutaneous carbon dioxide tension is a useful tool for the differential diagnosis between these mechanisms, although this assessment is under utilised 29. In our bench study, the ResLinkTM device gave a good estimate of MV and, thus, may be useful for differentiation between the mechanisms of regular desaturations, but this needs to be confirmed by other studies.

A number of criticisms may be made with regard to our study. Concerning the bench study, the tests were performed using a passive lung model, which may not accurately reproduce physiological conditions. The estimate of leak by the ResLinkTM may be less accurate with assisted breathing, given that the velocity of the turbine (and accordingly the estimate of leak level) can vary in the presence of inspiratory effort.

Regarding the clinical study, first, the study includes patients with various pathologies and heterogeneous clinical situations and, additionally, the parameters defined as a threshold of normality are arbitrary. However, our objective was to propose a screening method that can be used in daily practice. Thus it was necessary to include a representative sample of practice and establish criteria that reflect the quality of ventilation and which have already been suggested as markers of NIV inefficacy 12, 25. Secondly, our definition of ineffective ventilation was based only on the results of the traces without taking into account ABG or clinical response. Future research should assess the importance of combining the results of this monitoring with other variables. Thirdly, since ResLinkTM provides only trend data and not raw data of ventilation, this system does not allow an analysis of patient–ventilator interaction to detect asynchrony. Nevertheless, it has been well established that, in patients receiving NIV, major leak is probably the most important contributor to these events 12, and our evaluation includes its detection. Finally, as with all oximeters, oxygen supplementation may affect the reliability of arterial oxygen saturation measured by this device.

In summary, monitoring by this system makes it possible, with accessible technology, to assess the quality of the ventilation and interpret the results easily, which facilitates its application in screening patients in a real-life situation. If the results prove to be less than expected, this system makes it possible to first clarify the pathophysiological mechanism responsible and, if changes are required, to subsequently check the results of these changes. The systematic use of this system enables NIV to be optimised, limiting the indication of PG/PSG to complex cases. The importance of including the results of this simplified monitoring with other assessment variables of ventilation quality should be the objective of future research.

Statement of interest

None declared.

Acknowledgments

We would like to thank C. Duvivier (Unité de Physiopathologie Respiratoire, Université Poincaré, Nancy, France), D. Veale (ANTADIR, Paris, France), P. Leger (Laboratoire du sommeil, Service de Pneumologie, CHU Lyon Sud, Lyon, France) and S. Sortor-Leger (ResMed, France) for their kind and helpful collaboration and for correcting our English, and the technician team of Alize de Bourgogne (Dijon, France) for their excellent technical assistance.

Footnotes

  • Received November 10, 2008.
  • Accepted March 16, 2009.

References

  1. ↵
    Fanfulla F, Delmastro M, Berardinelli A, et al. Effects of different ventilator settings on sleep and inspiratory effort in patients with neuromuscular disease. Am J Respir Crit Care Med 2005;172:619–624.
    OpenUrlCrossRefPubMedWeb of Science
  2. ↵
    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.
    OpenUrlCrossRefPubMedWeb of Science
  3. ↵
    Gonzalez MM, Parreira VF, Rodenstein DO. Non-invasive ventilation and sleep. Sleep Med Rev 2002;6:29–44.
    OpenUrlCrossRefPubMedWeb of Science
  4. ↵
    Schonhofer B, Sortor-Leger S. Equipment needs for noninvasive mechanical ventilation. Eur Respir J 2002;20:1029–1036.
    OpenUrlAbstract/FREE Full Text
  5. ↵
    Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–310.
    OpenUrlCrossRefPubMedWeb of Science
  6. ↵
    Lloyd-Owen SJ, Donaldson GC, Ambrosino N, et al. Patterns of home mechanical ventilation use in Europe: results from the Eurovent survey. Eur Respir J 2005;25:1025–1031.
    OpenUrlAbstract/FREE Full Text
  7. ↵
    Jounieaux V, Parreira V, Delguste P, et al. Nasal mask pressure waveform and inspiratory muscle rest during nasal assisted ventilation. Am J Respir Crit Care Med 1997;155:2096–2101.
    OpenUrlPubMedWeb of Science
  8. ↵
    Parreira VF, Delguste P, Jounieaux V, et al. Glottic aperture and effective minute ventilation during nasal two-level positive pressure ventilation in spontaneous mode. Am J Respir Crit Care Med 1996;154:1857–1863.
    OpenUrlPubMedWeb of Science
  9. ↵
    Jounieaux V, Aubert G, Dury M, et al. Effects of nasal positive-pressure hyperventilation on the glottis in normal awake subjects. J Appl Physiol 1995;79:176–185.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    Parreira VF, Delguste P, Jounieaux V, et al. Effectiveness of controlled and spontaneous modes in nasal two-level positive pressure ventilation in awake and asleep normal subjects. Chest 1997;112:1267–1277.
    OpenUrlCrossRefPubMed
  11. ↵
    French Pulmonary Society. [Non-invasive positive pressure ventilation for neuromuscular diseases. Practice guidelines]. Rev Mal Respir 2006;23:14S15–14S40.
    OpenUrlPubMed
  12. ↵
    Guo YF, Sforza E, Janssens JP. Respiratory patterns during sleep in obesity-hypoventilation patients treated with nocturnal pressure support: a preliminary report. Chest 2007;131:1090–1099.
    OpenUrlCrossRefPubMedWeb of Science
  13. Freeman ML, Hennessy JT, Cass OW, et al. Carbon dioxide retention and oxygen desaturation during gastrointestinal endoscopy. Gastroenterology 1993;105:331–339.
    OpenUrlPubMedWeb of Science
  14. ↵
    Kopka A, Wallace E, Reilly G, et al. Observational study of perioperative PtcCO2 and SpO2 in non-ventilated patients receiving epidural infusion or patient-controlled analgesia using a single earlobe monitor (TOSCA). Br J Anaesth 2007;99:567–571.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    Jounieaux V, Rodenstein DO. [Home mechanical ventilation: indications and pathophysiological limitations]. Rev Mal Respir 2004;21:358–366.
    OpenUrlPubMedWeb of Science
  16. ↵
    Robert D, Argaud L. Non-invasive positive ventilation in the treatment of sleep-related breathing disorders. Sleep Medicine 2007;8:441–452.
    OpenUrlCrossRefPubMedWeb of Science
  17. ↵
    Pépin J-L, Borel JC, Janssens JP, et al. Sleep and NIV: monitoring of the patient under home ventilation. In: Muir J-F, Ambrosino N, Simonds AK, eds. Noninvasive Ventilation. Eur Respir Mon 2008;41:350–366.
    OpenUrl
  18. ↵
    Teschler H, Stampa J, Ragette R, et al. Effect of mouth leak on effectiveness of nasal bilevel ventilatory assistance and sleep architecture. Eur Respir J 1999;14:1251–1257.
    OpenUrlAbstract/FREE Full Text
  19. ↵
    Meyer T, Pressman M, Benditt J, et al. Air leaking through the mouth during nasal ventilation: effect on sleep quality. Sleep 1997;20:561–569.
    OpenUrlPubMedWeb of Science
  20. ↵
    Heather S, Simmonds A, Ward S. Problem solving in acute NIPPV. In: Simmonds AK, ed. Non-Invasive Respiratory Support. London, Edward Arnold, 2001; pp. 146–159
  21. ↵
    Rabec CA, Reybet-Degat O, Bonniaud P, et al. Leak monitoring in noninvasive ventilation. Arch Bronconeumol 2004;40:508–517.
    OpenUrlPubMedWeb of Science
  22. ↵
    Achour L, Letellier C, Cuvelier A, et al. Asynchrony and cyclic variability in pressure support noninvasive ventilation. Comput Biol Med 2007;37:1308–1320.
    OpenUrlCrossRefPubMedWeb of Science
  23. ↵
    Elliott MW. The interface: crucial for successful noninvasive ventilation. Eur Respir J 2004;23:7–8.
    OpenUrlFREE Full Text
  24. ↵
    Parreira VF, Jounieaux V, Delguste P, et al. Determinants of effective ventilation during nasal intermittent positive pressure ventilation. Eur Respir J 1997;10:1975–1982.
    OpenUrlAbstract
  25. ↵
    Teschler H. Monitoring of the home mechanical ventilated patient. In: Muir J-F, Ambrosino N, Simonds AK, eds. Noninvasive Ventilation. Eur Respir Mon 2001;16:274–280.
    OpenUrl
  26. ↵
    Gonzalez J, Macey J, Cracco C, et al. Initiation and management of home mechanical ventilation. Rev Mal Respir 2004;21:783–790.
    OpenUrlPubMedWeb of Science
  27. ↵
    Bach J, Robert D, Leger P, et al. Sleep fragmentation in kiphoscoliotic individuals with alveolar hypoventilation treated by NIPPV. Chest 1995;107:1552–1558.
    OpenUrlCrossRefPubMedWeb of Science
  28. ↵
    Willson GN, Piper AJ, Norman M, et al. Nasal versus full face mask for noninvasive ventilation in chronic respiratory failure. Eur Respir J 2004;23:605–609.
    OpenUrlAbstract/FREE Full Text
  29. ↵
    Pepin JL, Chouri-Pontarollo N, Orliaguet O, et al. Alveolar hypoventilation during sleep and domiciliary assisted ventilation. Rev Mal Respir 2005;22:113–125.
    OpenUrlPubMedWeb of Science
  30. ↵
    Rabec C, Merati M, Baudouin N, et al. Management of obesity and respiratory insufficiency. The value of dual-level pressure nasal ventilation. Rev Mal Respir 1998;15:269–278.
    OpenUrlPubMedWeb of Science
  31. ↵
    Jounieaux V, Aubert G, Dury M, et al. Effects of nasal positive-pressure hyperventilation on the glottis in normal sleeping subjects. J Appl Physiol 1995;79:186–193.
    OpenUrlAbstract/FREE Full Text
  32. Parreira VF, Jounieaux V, Aubert G, et al. Nasal two-level positive-pressure ventilation in normal subjects. Effects of the glottis and ventilation. Am J Respir Crit Care Med 1996;153:1616–1623.
    OpenUrlPubMedWeb of Science
  33. ↵
    Delguste P, Aubert-Tulkens G, Rodenstein DO. Upper airway obstruction during nasal intermittent positive-pressure hyperventilation in sleep. Lancet 1991;338:1295–1297.
    OpenUrlCrossRefPubMedWeb of Science
  34. ↵
    Lee SH, Choi JH, Shin C, et al. How does open-mouth breathing influence upper airway anatomy?. Laryngoscope 2007;117:1102–1106.
    OpenUrlCrossRefPubMedWeb of Science
  35. ↵
    Fanfulla F, Taurino AE, Lupo NDA, et al. Effect of sleep on patient/ventilator asynchrony in patients undergoing chronic non-invasive mechanical ventilation. Respir Med 2007;101:1702–1707.
    OpenUrlCrossRefPubMedWeb of Science
  36. ↵
    Nakayama H, Smith CA, Rodman JR. et al. Effect of ventilatory drive on carbon dioxide sensitivity below eupnea during sleep. Am J Respir Crit Care Med 2002;165:1251–1260.
    OpenUrlCrossRefPubMedWeb of Science
View Abstract
Back to top
View this article with LENS
Email
Print
Citation Tools

Evaluating noninvasive ventilation using a monitoring system coupled to a ventilator: a bench-to-bedside study
C. Rabec, M. Georges, N. K. Kabeya, N. Baudouin, F. Massin, O. Reybet-Degat, P. Camus
European Respiratory Journal Oct 2009, 34 (4) 902-913; DOI: 10.1183/09031936.00170508
del.icio.us logo Digg logo Reddit logo Technorati logo Twitter logo CiteULike logo Connotea logo Facebook logo Google logo Mendeley logo
Full Text (PDF)

Jump To