The main findings of this observational study performed in patients with stable respiratory disease ready to be enrolled in a home ventilatory program can be summarized as follows: (i) there is an high prevalence of asynchrony in non-invasively ventilated patients; (ii) the occurrence of asynchrony is not correlated to any variable of respiratory mechanics recorded during spontaneous breathing and does not differ between patients with obstructive or restrictive disease; (iii) the frequency of ineffective efforts is associated with a higher level of pressure support; (iv) an high incidence of asynchronies, above all of ineffective efforts, is associated to a poorer tolerance of NIV.
Incidence of asynchronies
Several studies have investigated the effect of different ventilator settings on the development of asynchronies, mostly in acutely ill patients. Thille
et al. [
3] showed that approximately one-fourth of invasively ventilated patients had a high incidence of asynchrony during assisted breathing. A high level of pressure support and a large tidal volume were associated with an increased incidence of asynchronies [
13]. On the other hand, Vignaux
et al. [
2] recently found an AI >10% in more than 40% of patients during NIV. The level of pressure support and the magnitude of leaks were independent predictive factors of severe asynchronies. The incidence of severe asynchronies was lower in our study (30%) than in the latter studies [
2,
3], which could be explained by the different ventilators used for NIV. The study by Vignaux
et al. [
2], in fact, was performed in an ICU in which intensive care ventilators were used for NIV with and without the use of an algorithm for compensation for leaks. But it is known, as was recently shown [
14], that the asynchrony index is significantly lower with a dedicated NIV ventilator than with ICU ventilators even when the latter are used with their NIV algorithm. In our study, we used a ventilator exclusively designed for NIV.
Concerning the types of asynchronies, we have also confirmed the finding of most of the other studies [
2,
3,
13], that IEs are much more common than either auto-triggering and double triggering.
Effects of respiratory mechanics
The most important finding of our study was that AI was not different between patients with obstructive or restrictive disease and that none of the parameters of respiratory mechanics was significantly different in those patients showing or not a high frequency of asynchronies. Until now, observational studies have been performed only in a single group of patients (that is, those with COPD) [
15] or in a heterogeneous group of patients without subgroup analysis [
2,
3,
13,
16]. The relationship between a high level of pressure support and rate of IE was emphasized in COPD patients, due to high lung compliance, which could be responsible for large tidal volumes [
13]. It has been shown that in these patients with high compliance, the ventilator continues to inflate the respiratory system long after the inspiratory muscles have ceased to contract and the next inspiratory attempt is likely to occur at a high lung volume, when airway pressure is still markedly positive; the inspiratory effort will not, therefore, always be sufficient to create a pressure gradient capable of being sensed by the ventilator [
15,
17]. Rather surprisingly, in our study, despite the average lung compliance being higher in COPD patients than in patients with restrictive disease, particularly in those in whom asynchronous events occurred, no correlation was found between lung compliance and the onset of asynchrony. Moreover, we found the same incidence of a high rate of IE both in obstructive and restrictive patients. This could mean that the mechanism rather than the compliance (that is, leaks) could be involved.
PEEP
e has been shown to decrease ineffective triggering in patients with a high PEEP
dyn by reducing the portion of Pdi spent to overcome the amount of PEEP
dyn and needed to trigger the ventilator [
18]. On the contrary, other studies [
13,
16] did not find any influence on the amount of ineffective effort when the PEEP
e was applied as a "fixed" value (that is, 5 cmH
2O). In our study, the level of PEEP
dyn, despite being significantly higher in patients with obstructive respiratory disease, was overall still too small to induce a potential problem of triggering and, therefore, to be the major determinant of the asynchronies described in our patients. Indeed, the fact that overall the patients affected by different underlying diseases showed similar amounts of asynchronies, led us to conclude that there are other potential mechanisms, at least in patients with stable disease. One possible explanation for the relatively high occurrence of asynchrony may be the setting of the ventilator. Nowadays, ever more parameters can be adjusted on the machine, and each of them could be responsible for asynchronies. For example, Haynes
et al. [
19] showed how increasing the rise time while keeping the flow-cycling threshold constant, could significantly reduce tidal volume and, consequently, lower the incidence of IE. The importance of settings was also demonstrated in intubated COPD patients, in whom increasing the expiratory threshold to 70% of the peak inspiratory flow improved patient-ventilator interaction and decreased ineffective efforts without changing inspiratory muscle effort or alveolar ventilation [
7].
Since the importance of respiratory mechanics and elevated tidal volume only partially explain the occurrence of asynchronies, we believe that other factors need to be taken into account during NIV. Despite our attention and care to minimize the problem of air leaks, these events may have deeply influenced our results, as already shown [
2]. Indeed, even in centers with NIV expertise, it has been shown that the occurrence of leaks is unavoidable [
2].
Last, in order to avoid any confounders, such as different settings, we standardized the ventilator adjustments by protocol. Concerning the fixed rise time and expiratory trigger, the latter was automatically set by the ventilator we have used, and has been previously shown to be effective in reducing the effort and improving the ventilatory phase [
20]. Moreover, we chose the rise time that in a previous study showed a good balance between the amount of air-leaks and patient tolerance [
21]
Asynchrony and tolerance to NIV
As a secondary outcome, this study showed that those patients with a high incidence of ineffective efforts and asynchronies had also a poorer tolerance of NIV. Until now the association between NIV tolerance and asynchrony was studied only in an acute setting [
2].
In this study, performed in patients receiving NIV for acute respiratory failure, the tolerance score was higher in those who showed an AI >10% [
2]. However, the impact of a single different type of asynchrony was not studied.
Ineffective effort is a sort of inspiratory muscle effort in a closed system because it is not followed by an activation of the ventilator. This could explain the perception of discomfort with a ventilator when the incidence of this kind of asynchrony is more frequent.
Clinical implications
As for any physiological study, the main aim of this study was to assess the mechanism(s) of an event, rather than suggesting a clinical practice. The extrapolation of the results obtained in this study for clinical practice is, therefore, questionable, especially because the recordings were performed for a short period during the daytime, while most of the patients requiring home NIV are also ventilated during the night. Having said this, we have shown that the occurrence of asynchronies is very frequent even when the settings are performed in expert centers [
2]. A close look at some indirect indices of mismatching (that is, the flow waveform), may help the clinician to detect the most frequent asynchrony between the patient and the machine, that is, ineffective effort [
22]. Indeed, it is of clinical interest that, despite the different underlying pathologies, the settings decided by the operators to achieve the same aims (that is, gas exchange amelioration and tidal volume increase) are very similar, and also induced comparable amounts of asynchronies.
Limitation of the study
As already stated, the gold standard for measuring the interaction between the patient and a ventilator is electromyography of the diaphragm because indirect estimates of the onset and duration of neural inspiratory time based on esophageal pressure and flow could lead to errors, compared to neural inspiratory time measurement. For the same reasons, we were unable to assess the inspiratory trigger delay and the expiratory delay because of the use of the catheter-balloon technique. On the other hand, a good correlation among IE, AT and DT detected from flow/airway pressure and from esophageal pressure signals has been already shown [
3,
23].
Lastly, we used "only" one type of ventilator and an
a priori "fixed" setting. However, the ventilator chosen, the Vision, used the same algorithms (above all the "autotrack") present on all the bilevel ventilators produced by Philips-Respironics and widely used for NIV at home (BiPAP Synchrony, BiPAP S/T, BiPAP A30, Trilogy). Moreover, we are aware of the existence of an updated software in the V60. However, a recent bench and clinical study was not able to find any difference between the two generations of ventilator in terms of ineffective trigger, trigger delay and delayed cycling [
14]. Of course, the results of the study may not be generalizable and may not represent the "real life" situation in which other different ventilator models are employed, and the clinicians are free to set the ventilatory parameters as they wish.