Asynchrony and cyclic variability in pressure support noninvasive ventilation

https://doi.org/10.1016/j.compbiomed.2006.12.005Get rights and content

Abstract

Noninvasive mechanical ventilation is an effective procedure to manage patients with acute or chronic respiratory failure. Most ventilators act as flow generators that assist spontaneous respiratory cycles by delivering inspiratory and expiratory pressures. This allows the patient to improve alveolar ventilation and subsequent pulmonary gas exchanges. The interaction between the patient and his ventilator are therefore crucial for tolerance and acceptability and part of this interaction is the facility to trigger the ventilator at the beginning of the inspiration. This is directly related to patients’ discomfort which is not quantified today. Phase portraits reconstructed from the airflow and first-return maps built on the total breath duration were used to investigate the quality of the patient–ventilator interaction. Phase synchronization can be identified from phase portrait and the breath-to-breath variability is well characterized by return maps. This paper is a first step in the direction of automatically estimating the comfort from measurements and not from a necessarily subjective answer given by the patient. These tools could be helpful for the physicians to set the ventilator parameters.

Introduction

The aim of mechanical ventilation is to assist or to replace spontaneous breathing in order to allow the physiological gas exchanges to be performed (oxygen capture and carbon dioxide release). In order to exchange oxygen and carbon dioxide through the alveolo-capillary membrane, air should move into and out of the lungs by the action of inspiratory muscles, especially the diaphragm. During a spontaneous inspiration, a negative pressure swing is generated by the inspiratory muscles; the resulting gradient between the mouth (atmospheric pressure) and the alveoli generates an airflow and the pulmonary volume then increases. During normal exhalation, the inspiratory muscles relax and the pulmonary volume passively decreases due to the lung elastic recoil. Pressure at the mouth is thus lower than inside the alveoli, and air is forced out the alveoli. When the pressure in the alveoli and the mouth equalizes, expiration ends.

Two methods have been developed to assist or replace the normal lung mechanics of breathing. First, negative pressure ventilation using ventilators like the “iron lung” [1] tends to mimic the normal lungs mechanics and was developed during the polio epidemics [2], [3]. When the patient's airways are opened, the underpressure in the chamber enclosing the body and sealed at the neck, allows the lungs to inflate. Expiration occurs when the negative pressure around the chest wall is removed and the lung elastic recoil causes airflow out of the lungs passively [4]. Second, positive pressure ventilation is a technique where air is driven into the patient's lung by a positive airflow generated through an endotracheal tube (invasive ventilation) or a mask (noninvasive ventilation) [4]. During inspiration the ventilator increases the pressure at the mouth up to a value greater than alveolar pressure. This forces gas into the lungs. During expiration, the applied pressure is reduced and, due to lung and chest wall elastic recoil, gas is passively forced out of the lungs since the alveolar pressure is thus greater than the mouth pressure. Noninvasive ventilation gained popularity over 20 years ago for therapy of both acute and chronic respiratory failure. Advantages of noninvasive ventilation over invasive ventilation include avoidance of airway trauma, greater comfort, preservation of airway defenses, speech and swallowing functions.

The application of nonlinear dynamics to the study of the respiratory rhythm during mechanical ventilation can be traced back to the analysis of different coupling patterns (between a mechanical ventilator and the respiratory rhythm) in the works by Petrillo and Glass [5], [6] where the concept of phase locking was first proposed. Experiments were performed on anesthetized cats which were paralyzed by neuromuscular blockade and invasively ventilated. The rhythm of central respiratory activity was monitored by recording inspiratory-promoting activity from the phrenic nerve, which innervates the diaphragm. As the ventilator volume and frequency are varied, a number of different rhythms were identified between the ventilator and phrenic activity [6].

Tools borrowed from nonlinear dynamical systems theory have been widely used in biomedicine, in particular for investigating cardiac variability, see [7], [8], [9], [10] among others. Breathing is a phenomenon naturally related to the cardiac activity. Thus global modeling and dynamical analysis of physiological data—blood oxygen saturation, heart rate and respiration—have been performed by Aguirre and co-workers [11]. They were able to predict the blood oxygen saturation from the knowledge of the heart rate and the respiration. Also in normal subjects, breathing patterns display a certain variability which is maintained by a central neural mechanism and the feedback loop of arterial chemoreceptors and lung vagal sensory receptors [12]. Peripheral factors, such as mechanical and chemical changes within the respiratory system, may modify the breathing pattern variability from the normal level in individuals under pathological conditions. Quantitative methods, including calculations of coefficients of variations and first-return maps—also called Poincaré maps—have been applied to the analysis of breathing pattern variability to serve as indicators of pathological conditions in patients with respiratory diseases [13], [14], [15].

Noninvasive mechanical ventilation is a clinical procedure which can be also successfully investigated in terms of phase locking [16], [17]. This results directly from the fact that mechanical ventilation is effective when the patient does not “struggle” with his ventilator. The interplay between these two pumps is thus directly related to the general concept of synchronization [18]. More recently, breathing pattern variability has been investigated using first-return map built on the total breath duration. It was shown that first-return maps can serve as a potential weaning predictor for postoperative patients recovering from systemic inflammatory response syndrome [19].

One relevant factor for the tolerance of assisted mechanical ventilation could be the patient's comfort. Unfortunately, quantifying ventilatory comfort is still an open problem since it is based on subjective answers to questions asked to the patient. Our objective is therefore to investigate how tools borrowed from the nonlinear dynamical systems theory can improve this rough estimation of the patients’ discomfort. We will therefore investigate how phase portrait may detect lack of phase synchronization and how first-return maps—usually used as indicators for the level of variability in a time series—can provide information on the patients’ ability to manage their ventilators. The use of these tools required the development of an algorithm to automatically detect the main phases of the respiratory cycles, that is, inspiratory and expiratory phases, from noninvasive measurements (flow and airway pressure).

The subsequent part of this paper is organized as follows. Section 2 is devoted to the experimental device used for the measurements and to some characteristics of the subjects who participated in this study. In Section 3, phase portraits are reconstructed from the airflow and it is shown how the ventilator parameters may affect its shape and how lack of ventilator triggering may change its structure. In Section 4, it is explained how trigger asynchrony can be identified from airflow and airway pressure. In Section 5 the effect of ineffective triggerings on the patient–ventilator dynamics is investigated for the 12 subjects. Section 6 gives a conclusion.

Section snippets

The respiratory circuit and the ventilator

The main physiological aim for instituting mechanical ventilation is to decrease patients’ work of breathing. The most effective unloading of the inspiratory muscles is obtained when the ventilator cycles in synchrony with the activity of the patient's own respiratory rhythm [20]. The interplay between these two pumps is complex, and asynchronies may arise at several points during the respiratory cycle. In order to investigate this interplay, the experimental device is built as follows (Fig. 1

Spontaneous breathing

Let us start with the phase portrait reconstructed from the airflow measured in a circuit where the subject spontaneously breathes, that is, when the ventilator is removed from the circuit. Typically, each respiratory cycle is associated with a loop around the origin of the reconstructed phase space (Fig. 3). The respiratory cycle is subdivided into the inspiratory phase, when diaphragmatic activity is sustained and the lungs are inflating, and the expiratory phase, when diaphragmatic activity

Trigger asynchrony

Ideally, during mechanical ventilation the triggering of the ventilator should result from inspiratory muscles contraction. In some circumstances, however, a mechanical cycle may be triggered without any inspiratory effort (auto-triggering). It may be caused by random noise in the circuit, water in the circuit, leaks, cardiac oscillations, etc. Auto-triggering most often occurs with low respiratory drive and breathing frequency and when dynamic hyperinflation is absent. In such a situation,

Intersubject dependence

For the 12 subjects involved in the protocol, we computed the rate of ineffective triggerings using the procedure described in the previous section. Depending on the subject, this rate varies between 0% and 64.4% (Table 2). In each of the three groups of subjects—patients with COPD or OHS, or healthy subjects—the rate of ineffective triggerings can be near zero or greater than 50%. Even healthy subjects can have a significant rate of ineffective triggerings. Such a feature means that this form

Conclusions

It is well known that the rate of ineffective triggerings is related to the level of ventilatory assistance (IPAP value). A decrease in the magnitude of inspiratory effort at a given level of assistance is not necessarily the cause of ineffective triggerings since Leung and co-workers [37] showed that inspiratory efforts followed by an ineffective triggering of the ventilator were also more than 30% higher than efforts successfully triggering the ventilator. Significant differences, however,

Acknowledgments

L.A. is supported by Association d’Aide à Domicile aux Insuffisants Repiratoires (ADIR), member of the ANTADIR federation (Paris, France). We wish to thank J. Kurth for his encouraging comments on this work during the Eighth Experimental Chaos Conference.

Linda Achour was born in 1978. She received a Bachelor degree in 1996. She received a Master degree in Biomedical Sciences from the University of Lyon. In September 2000, she was involved in the DEA in Medical and Biological Engineering with medical imaging speciality. In November 2005, she received a Ph.D. degree in bio-engineering on asynchronisms in noninvasive ventilation. Actually, she is a Research Engineer in Biomedicine for ADIR Association in Rouen (Association d’aide á Domicile des

References (37)

  • G.A. Petrillo et al.

    A theory for phase locking respiration in cats to a mechanical ventilator

    Am. J. Physiol.

    (1984)
  • A.L. Goldberger et al.

    Applications of nonlinear dynamics to clinical cardiology

    Ann. N. Y. Acad. Sci.

    (1987)
  • A. Babylogantz et al.

    Is the normal heart a periodic oscillator?

    Biol. Cybern.

    (1988)
  • J. Kurth et al.

    Quantitative analysis of heart rate variability

    Chaos

    (1995)
  • M.E.D. Gomes et al.

    Investigation of determinism in heart rate variability

    Chaos

    (2000)
  • E.N. Bruce et al.

    Mechanisms and analysis of ventilatory stability

  • T. Brack et al.

    Dyspnea and decreased variability of breathing in patients with restrictive lung disease

    Am. J. Respir. Crit. Care Med.

    (2002)
  • B. Loveridge et al.

    Breathing patterns in patients with chronic obstructive pulmonary disease

    Am. Rev. Respir. Dis.

    (1984)
  • Cited by (0)

    Linda Achour was born in 1978. She received a Bachelor degree in 1996. She received a Master degree in Biomedical Sciences from the University of Lyon. In September 2000, she was involved in the DEA in Medical and Biological Engineering with medical imaging speciality. In November 2005, she received a Ph.D. degree in bio-engineering on asynchronisms in noninvasive ventilation. Actually, she is a Research Engineer in Biomedicine for ADIR Association in Rouen (Association d’aide á Domicile des Insuffisants Respiratoires) and her main research topics are patient–ventilator interactions in adults or pediatrics patients with chronic respiratory failure, and bench test study for home ventilators.

    Christophe Letellier received a Master degree in Particule Physics and a Ph.D. degree in the nonlinear dynamical system theory from the University of Paris VII in 1991 and 1994, respectively. In 1996 he was offered a permanent position at University of Rouen where he supervises undergraduate and graduate research on modeling and analysis of nonlinear systems. He manages the research group ATOMOSYD from CORIA CNRS UMR 6614 and his main research topics include: global modeling, topological characterization of nonlinear time series, symmetry properties of nonlinear systems, observability and controllability of chaotic systems.

    Antoine Cuvelier is a Medical Doctor at the Department of Pulmonary and Intensive Care Medicine at Rouen University (France, EU). He received a Master Degree in Respiratory Physiology from Paris V René Descartes University and a Ph.D. in Pulmonary Medicine from Rouen University about chronic obstructive respiratory failure. His research interests are focused on the clinical and physiological aspects of the management of patients with respiratory failure, either in the chronic or the acute setting. He has studied the indications, the modalities and the monitoring of noninvasive ventilation, especially in patients with COPD and/or obesity. Antoine Cuvelier manages the Noninvasive Ventilation Group at the Société de Pneumologie de Langue Française. He serves as a member of editorial boards and scientific committees in the field of chronic respiratory failure.

    Eric Verin received his M.Sc. in Medical Biology in 1996, his Medical Degree and chest specialization in 1997, Master's Degree in 1998 and Ph.D. in life sciences in 2002. He is a member of French Physiology Society, Respiratory Muscles Group, Societé de Pneumologie de Langue Française (SPLF) and of European Respiratory Society. He has a permanent position in Rouen University Hospital in 2003, and is working in the research group of GRHV, UPRES EA 3830, IFRMP 23.

    Jean-François Muir is head of the Respiratory Department and Respiratory Intensive Care Unit of Rouen University Hospital and Director of the EA3830 Research Unit devoted to Respiratory Handicap. He is involved for many years in research in respiratory assistance and mechanical ventilation.

    View full text