Sleep and Mechanical Ventilation in Stable COPD Patients

ABSTRACT

Respiratory work is physiologically increased during sleep and leads to severe alterations in COPD patients, especially by raising sleep hypoventilation. The diurnal impact of these nocturnal events may have been underestimated in COPD patients. Impaired sleep and the increase of respiratory work may be one of the major trigger of diurnal events like hypoventilation, exacerbation and even mortality. One of the most commonly used nocturnal treatments at the present time is noninvasive ventilation (NIV). However, there is an on-going debate concerning the indications and objectives of NIV in COPD patients. In most studies, NIV initiation and monitoring depend on diurnal tools like PaCO2, and the nocturnal efficacy of this treatment has not yet been adequately determined. In other respiratory diseases, sleep events have a predominant role in NIV therapy. Such nocturnal events drive NIV initiation and setting adaptation. Monitoring of sleep events is associated with an increase in health related to quality of life and a decrease in mortality. The monitoring may be the solution to solve the debate of NIV in COPD patients. This article reviews the impact of sleep in COPD patients and the value of long-term NIV.

KEYWORDS:

• Accessory respiratory muscles
• COPD
• hypoventilation
• noninvasive ventilation
• sleep

Introduction

Chronic obstructive pulmonary disease (COPD) is a lung disease characterized by bronchial obstruction (related to chronic airway inflammation and bronchial remodelling), and lung hyperinflation secondary to destruction of the lung parenchyma. These various abnormalities, present in varying degrees in patients with COPD, are responsible for marked heterogeneity of ventilation/perfusion ratios and increased respiratory work.

These abnormalities result in a fragile clinical equilibrium, in which each added constraint such as an effort or superinfection is responsible for respiratory symptoms, possibly leading to therapeutic intervention, hospitalization or even death.

Although the great majority of interventions are designed to prevent (vaccination, smoking cessation, etc.) or reduce (bronchodilators, antibiotics, corticosteroids and oxygen therapy) these added constraints throughout the day and night, all patients are exposed to a period during which respiratory work is physiologically increased: sleep.

Sleep has multiple physiological effects on ventilation, such as decreased chemosensitivity, increased upper airway resistance, decreased muscle activity, a change of ventilation/perfusion ratios and redistribution of body fluids (Table 1). Paradoxically, sleep has been poorly described in COPD patients and has been the subject of few specific treatment proposals. However, several publications have studied sleep in COPD and long-term noninvasive ventilation (NIV) has been recently shown to be particularly useful during sleep (1). A better understanding of the pathophysiology of sleep in COPD therefore appears to be essential in order to optimally adapt the treatments that are mainly used at night.

Table 1. Major physiological change during sleep in COPD patients.

	Upper airway resistance	Respiratory muscle activity	Vt	FRC	PtCO2
Physiological change	Increase But no more than control patients	Decrease. More profond during REM sleep	Decrease. More profond during REM sleep	No change	Increase. More profond during REM sleep
Ref.	(2–4)	(25,26)	(2,3,10)	(2)	(10,11)

Vt, tidal volume; FRC, functional residual capacity; PtCO₂, transcutaneous CO₂.

Part 1: Sleep and COPD

Upper airway resistance during sleep in COPD: contradictory results

Contradictory results have been reported concerning upper airway obstruction in COPD. In 1995, Ballard et al. (2) first proposed the hypothesis of increased upper airway resistance based on nocturnal plethysmography measurements in five patients. In 2004, O'Donoghue et al. (3), in a series of 17 hypercapnic, nonobese and nonapnoeic COPD patients, reported a marked increase of airway resistance, which increased from 3 to 10 cm H₂O l⁻¹ s. during sleep. However, Meurice et al. (4) although they confirmed this increased airway resistance on nocturnal polysomnography recordings, showed that it was not more marked than that observed in healthy subjects. Unfortunately, these three studies used more or less invasive nocturnal investigation tools, which suppressed REM sleep in some of these studies (3), and the very diverse methods of measurement preclude any formal conclusions at the present time. More conclusive evidence may be provided by recent large-scale clinical trials of nocturnal NIV in COPD, which did not show any episodes of upper airway obstructive apnoea (OA) on NIV, although it is unclear how the quality of nocturnal NIV was verified (5). Furthermore, none of the studies required high levels of positive end-expiratory pressure (PEEP) (1,6). Finally, in a study based on nocturnal recordings in a large cohort of patients, an apnoea-hypopnoea index (AHI) >15/h was observed in only 14% of patients with COPD (7). In on-going studies by our team, only 19% of patients presented an AHI >15/h (8). At the present time, it therefore appears reasonable to conclude that COPD patients do not present more episodes of nocturnal OA than the general population. Moreover, the rarity of OA in COPD during sleep could be explained by the fact that these patients present high volume breathing, which is known to induce an upper airway conformation that tends to prevent obstruction (9).

Lung volume changes: current therapeutic target

During sleep, COPD patients present a marked reduction of their tidal volume ranging from 26% (3) to 37% (2). In 54 hypercapnic hyperinflated patients, this reduction of tidal volume (VT) was associated with hypoventilation in 43% of cases (10). In a study comprising 100 patients, Holmedahl et al. (11) showed an association between nocturnal hypoventilation and diurnal hypercapnia. Nocturnal hypoventilation has become the target of treatment by NIV, with very intense ventilator settings right from initiation of NIV (12), also very recently recommended at night in order to correct or decrease hypoventilation by 20% (1). However, the proposed objectives at the present time are generally based on waking measurements, as initially described (12).

Sleep alterations are underestimated

In addition to these nocturnal consequences on ventilatory parameters, sleep itself can be altered. During simple clinical interviews with 146 patients, the quality of sleep represented the third leading complaint after dyspnoea and asthenia (13). Valipour et al. (14), in a series of 52 patients, confirmed a reduction of total sleep time and alteration of sleep structure evaluated by polysomnography, but unfortunately without specifying the cause. However, the symptoms of poor quality sleep, very largely underestimated by clinicians, have a major clinical impact on daytime activities with impaired quality of life (15) and overconsumption of emergency care (16). Apart from the ventilatory changes described above, the causes of altered sleep in these patients have not been fully elucidated: nocturnal dyspnoea? pain? cough? anxiety? sedatives? discomfort related to ventilator equipment (oxygen tubes, extractor noise, nasal air flow). Numerous causes would appear to be involved and our team is currently investigating the hypothesis of an impact of diaphragmatic dysfunction, as in neuromuscular diseases (17).

Respiratory muscles and sleep: one of the keys to the nocturnal problem?

COPD patients present a dual impairment of the ventilatory pump: the work imposed on respiratory muscles is increased compared to healthy subjects due to several factors, including the presence of airway obstruction and dynamic hyperinflation related to the intrinsic PEEP; and the decreased capacity of inspiratory muscles in these patients. Lung hyperinflation induces a reduction of diaphragmatic muscle strength (18,19). Lung hyperinflation induces also a modification of the muscle phenotype (20) with an increased proportion of type 1 fibres (21), decreased myosin concentrations, and decreased maximum contractility of these fibres (22). Finally, malnutrition or treatments (corticosteroids, for example) can also be responsible for decreased muscle capacity.

COPD patients are now widely considered to present diaphragm dysfunction. The mechanisms of diurnal compensation [especially the use of neck muscles (23,24)] have already been investigated, but few authors have studied nocturnal mechanisms of compensation and their consequences, which could be one of the keys to the nocturnal problem, explaining the major value of nocturnal NIV.

Old studies based on small series have reported nocturnal electromyographic activation of respiratory accessory muscles and a reduction of this activity during periods of REM sleep compared to periods of Non-REM (NREM) sleep on scalene, sternocleidomastoid (25) and intercostal muscles (26) in 6 and 10 hyperinflated COPD patients, respectively. Similarly, phenotypic adaptations of respiratory accessory muscles have been reported in neuromuscular diseases (17,27) with nonphysiological use of these muscles during REM sleep in some patients. Several studies have demonstrated histological phenotypic changes of respiratory accessory muscles in COPD patients (28). Our team is studying the hypothesis that respiratory accessory muscles could play a role in sleep of COPD patients, and the preliminary results appear to confirm this hypothesis (8).

Sleep therefore appears to be an essential but poorly understood period in COPD patients, which probably explains the marked efficacy of treatments applied at night, such as O₂ or NIV.

Rationale for nocturnal NIV in COPD patients

Curative treatment is not a realistic goal in COPD patients at the present time. The aim of treatment is therefore to limit the consequences of the disease, especially the main consequence: mortality. However, as living is more than simply surviving, this treatment must also, and above all, improve the patient's overall quality of life. These objectives raise the question of the role of NIV in the treatment of COPD patients. In order to answer this question, we need to investigate the general mechanics underlying this treatment.

The theoretical effect of NIV during sleep is threefold: 1) it limits upper airway resistance (by the addition of PEEP); 2) it limits nocturnal hypoventilation; and 3) it allows resting of all respiratory muscles.

NIV therefore appears to be a perfectly adapted nighttime treatment, being able to largely compensate for the harmful effects of sleep on the ventilation of COPD patients (Figure 1).

Figure 1. Impact of sleep and nocturnal noninvasive ventilation on hypoventilation in COPD patients. UAR, upper airway resistance; IPAP, inspiratory positive pressure; EPAP, expiratory positive pressure; Vt, tidal volume.

The most extensively studied beneficial effect of nocturnal NIV at the present time is a reduction of alveolar hypoventilation, a cornerstone of the most recently published studies (1,5,6). However, this is not the only possible consequence of this treatment, as, on the basis of the effects described in patients with neuromuscular diseases, nocturnal noninvasive ventilation could also allow a reduction of diurnal and nocturnal alveolar hypoventilation, as well as improvement of quality of sleep in stable patient (29), [but also in intensive care unit (30)] reduction of cortical work (31), improvement of quality of life (32), reduction of mortality (32) and possibly dyspnoea [as suggested by recent preliminary studies (33)]; these objectives could probably also be achieved in COPD patients.

Part 2: COPD and long-term nocturnal NIV

The results of^{Figure 2} the main prospective randomized controlled trials (1,5,6,34–37) (Table 2) published up until 2014 appear to be contradictory. The studies published between 2002 and 2014 (34–37) did not demonstrate any effect of NIV therapy on mortality and reported few objective positive results (Table 2), while a growing number of patients were ventilated for COPD (38). In 2009, McEvoy et al. (5) was the first team to demonstrate an improvement of survival, but at the price of deterioration of certain quality of life items in the NIV arm. More recent studies finally reported different results, especially on improvement of survival (1) or hospital-free survival (39), by more precisely selecting patients likely to benefit from NIV. It is now generally accepted that patients more likely to benefit from long-term NIV are those with persistent hypoventilation (>52 mmHg) more than 2 weeks after an acute exacerbation. Also in 2009, Struik et al. (6) showed that the majority of patients with persistent hypercapnia following an acute exacerbation spontaneously corrected their hypoventilation and therefore do not all require home NIV based on the criterion of exacerbation-free survival.

Table 2. Main outcomes of randomized controlled trials about COPD patients following NPPV treatment

Study (year)	PaCO2 at baseline	IPAP/EPAP (cmH2O)	Effect on PaCO2 (between group)	Effect on HRQL (beetwen group)	Effect on mortality	Ref.
Gay et al. (1996)	Control: 48.5	2/10	None	x	x	[32]
(n = 35)	NPPV: 54.7 (mmHg)
Casanova et al. (2000)	Control: 53.2	4/12	None	x	12 months:	[34]
(n = 52)	NPPV: 50.7 (mmHg)				No difference
Clini et al. (2002)(n = 90)	Control: 7.4NPPV 7.2 (kPa)	2/14	None	24 months: SGRQ: no differenceMRF: improvment by NPPV	24 months: No difference	[33]
McEvoy et al. (2009)(n= 144)	Control: 54.4NPPV: 52.6 (mmHg)	5/13	None	12 months: SGRQ: no differenceSF 36 and POMS: deterioration in2 subscales each	24 months: Reduced by NPPV	[5]
Bhatt et al. (2013)(n = 30)	Control: 41.9NPPV 42.4 (mmHg)	5/15	None	6 months: CRQ: no difference	x	[35]
Struik et al. (2014)(n = 201)	Control: 7.7NPPV: 7.9 (kPa)	4.8/19.2	None	12 months: SRI: no difference	12 months: No difference	[6]
Köhlein et al. (2014)(n = 194)	Control: 7.7NPPV: 7.8 (kPa)	4.8/21.6	12 months: Reduced by NPPV	12 months: SGRQ and SRI: improvment by NPPVSF 36: no difference	12 months: Reduced by NPPV	[1]

All study included stable patients except the study of Struik et al. [6] HRLQ (Health Related Quality of Life), n (Number of randomized patient), NPPV (Non-invasive Positive Pressure Ventilation), IPAP (Inspiratory Positive Airway Pressures), EPAP (Expiratory Positive Airway Pressures), SGRQ (St. George's Respiratory Questionnaire), MRF (Maugeri Foundation Respiratory Failure), SF-36 (MOS 36-Item short-form health survey), POMS (Profile Of Mood State), CRQ (Chronic Respiratory Disease Questionnaire), SRI (Severe Respiratory Insufficiency Questionnaire).

Correction of hypoventilation is therefore currently considered to be the major criterion for initiation of long-term NIV, but it also constitutes the treatment target. However this goal raises another question: when do we have to measure the CO₂? In the morning? In the middle of the day? Or during the night? To our knowledge, no studies have demonstrated a relationship between PtCO₂ with or without NIV and diurnal PaCO₂.

In six prospective trials (Table 2) conducted in stable patients, correction of alveolar hypoventilation was only obtained in the study by Köhlein et al. (1), which was also the only study to clearly demonstrate the clinical benefit of NIV, but by using much higher inspiratory pressures than the other studies (Table 2). In 2009, the German team proposed the hypothesis that High-Intensity Noninvasive Positive Pressure Ventilation (HI-NPPV) could be the preferred ventilatory mode in these patients, as shown by Sadoul et al. (12) in the acute setting in the first trial of NIV in COPD patients published in 1964. HI-NPPV is based on optimized ventilator settings to ensure correction or at least maximal reduction of hypercapnia (40). Ventilator settings are increased progressively to reach the pressures required to correct hypercapnia, or the highest pressures tolerated by the patient when normocapnia cannot be achieved. However, ventilator pressures must not be increased beyond the limit tolerated by the patient. Initiation and initial adjustment of ventilator settings must be performed in hospital under close monitoring. The main validated tool proposed for this monitoring is transcutaneous CO₂ (PtcCO₂) (41) [although this technique was not used by Köhlein et al. (1)].

HI-NPPV is therefore guided, not by target pressures, but by ventilation designed to reduce hypercapnia, so that the necessary pressures consequently vary from one patient to another depending on the baseline PaCO₂ and the patient's tolerance of NIV. However, the inspiratory pressures used are much higher than those used in conventional ventilation, for which the term low-intensity noninvasive positive pressure ventilation (LI-NPPV) has been proposed (42).

Dreher et al. (43) have recently shown that the use of HI-NPPV versus LI-NPPV allowed a reduction of daytime PaCO₂, but also improvement of quality of life (evaluated by the severe respiratory insufficiency, SRI, scale), while LI-NPPV failed to achieve either of these objectives. This study also reported better adhesion to treatment in the HI-NPPV arm, very probably due to the benefits perceived by the patients (mean difference in NIV use: 3.6 hours; 95% CI (0.6–6.7), p = 0.024). However, it should be noted that the need to use high pressures is associated with a longer hospital stay (length of stay difference: 2.5 days; 95% CI (1.3–3.7) p = 0.001). The leak rate is also higher during HI-NPPV, but these leaks do not appear to have any clinical impact, especially on the patient's sleep (44).

It has recently been shown that the combination of high respiratory rate and high pressures does not provide any improvement of clinical (adhesion to treatment, quality of life evaluated by SRI, quality of sleep) or laboratory endpoints (arterial blood gases) (45), suggesting the predominant role of high pressure in this ventilatory mode. It must also be remembered that hypercapnia and reduction of hypercapnia are not the only determinants of ventilatory parameters, as the patient's comfort, history and adhesion to treatment must also be taken into account. Nevertheless, there is an unresolved debate concerning the need to use HI-NPPV ventilation to obtain clinical improvement, and although reduction of PaCO₂ appears to be the target of noninvasive ventilation, the complications related to this treatment, such as dynamic hyperinflation or impaired sleep, must be avoided.

Dreher et al. (44) reported the absence of deterioration of quality of sleep of patients treated by HI-NPPV compared to those treated by LI-NPPV, some studies have demonstrated a beneficial effect of NIV on sleep by increasing total sleep time, and on sleep efficacy (46–48) by increasing the percentage of REM sleep (5). Sleep events, sleep quality and their impact in COPD patients with and without ventilation were largely underestimated and underevaluated in almost all of the recent studies. In intensive care unit, a failure of NIV treatment was associated with an altered quality of sleep (49). NIV parameters are set up on awake patients, occulting physiological and pathological events happening during sleep. Nocturnal tools designed to monitor sleep events and sleep quality under ventilation (i.e. NIV software, polygraphy and polysomnography) are never used for settings adaptation. Instead of that, awake criteria like blood gases and tolerance are used. Closer monitoring of nocturnal events during NIV is probably necessary to more clearly understand its effects. Adler et al. (50), using polysomnographic monitoring of NIV, showed that correction of nocturnal events allowed better clinical tolerability and decreased dyspnoea after stopping NIV, but also a significant reduction of PtcCO₂ and PaCO₂ with significantly lower inspiratory pressures (16.5 (±2.1) vs 19 cm H₂O (±2.1), p = 0.04). However, the main tool currently proposed by the various teams (PtcCO₂) to monitor NIV appears to be unable to detect the presence of these events (51). Use of a more precise monitoring tool therefore appears to be essential in order to optimize the ventilation of these patients.

Monitoring of ventilation

As described above, correction of respiratory events in COPD patients appears to allow improvement of several short-term parameters (PaCO₂, morning dyspnoea, etc.) (50). However, systematic use of sleep laboratory polysomnography in patients on NIV in order to correct these events appears to be logistically unfeasible. In order to overcome this difficulty, an international consensus group (SomnoNIV Group) has proposed, in a series of consensus articles and guidelines, the use of software and algorithms included in home ventilators and provided by manufacturers (52–54). A better knowledge of these tools and how to use them can allow correction of respiratory events inherent to the patient and to the disease, but also inherent to ventilators (55), as illustrated in Figure 2. In the longer term, routine use of this software to correct desaturations (56), or obstructive events (57) has been shown to have a beneficial effect on the survival of patients with amyotrophic lateral sclerosis. Although the use of these tools has not yet been studied in COPD patients, we can assume that they would also be beneficial in these patients.

Figure 2. Example of monitoring of noninvasive positive pressure ventilation (NPPV) in a patient with COPD. Monitoring performed by polygraphy integrated in ventilator software (Rescan®, Resmed, Sydney, Australia). (a) Pressure-flow curve at initiation of NPPV, with numerous asynchronies (not shown here), suggesting ‘air hunger’. (b) Pressure-flow curve after NPPV adaptation, without asynchronies. (c) Summarized course of pressure, respiratory rate and triggered cycle settings during NPPV.

Conclusion

Patients with COPD frequently present poorly understood alterations of quality of sleep, associated with greater physiological variations than in healthy subjects, with probably major daytime consequences. Nocturnal NIV is becoming an essential treatment in order to limit these consequences. At the present time, patients deriving a marked benefit from this treatment are patients with long-term persistent hypercapnia in the absence of acute exacerbation and receiving high pressure ventilation. In the future, the quality of sleep of COPD patients, with and without treatment, and nocturnal monitoring of the quality of NIV, will constitute promising fields of research.

Declaration of interest statement

All authors declare no conflicts of interest.