Update: Non-Invasive Positive Pressure Ventilation in Chronic Respiratory Failure Due to COPD

Abstract

Long-term non-invasive positive pressure ventilation (NPPV) has widely been accepted to treat chronic hypercapnic respiratory failure arising from different etiologies. Although the survival benefits provided by long-term NPPV in individuals with restrictive thoracic disorders or stable, slowly-progressing neuromuscular disorders are overwhelming, the benefits provided by long-term NPPV in patients with chronic obstructive pulmonary disease (COPD) remain under question, due to a lack of convincing evidence in the literature. In addition, long-term NPPV reportedly failed in the classic trials to improve important physiological parameters such as arterial blood gases, which might serve as an explanation as to why long-term NPPV has not been shown to substantially impact on survival. However, high intensity NPPV (HI-NPPV) using controlled NPPV with the highest possible inspiratory pressures tolerated by the patient has recently been described as a new and promising approach that is well-tolerated and is also capable of improving important physiological parameters such as arterial blood gases and lung function. This clearly contrasts with the conventional approach of low-intensity NPPV (LI-NPPV) that uses considerably lower inspiratory pressures with assisted forms of NPPV. Importantly, HI-NPPV was very recently shown to be superior to LI-NPPV in terms of improved overnight blood gases, and was also better tolerated than LI-NPPV. Furthermore, HI-NPPV, but not LI-NPPV, improved dyspnea, lung function and disease-specific aspects of health-related quality of life. A recent study showed that long-term treatment with NPPV with increased ventilatory pressures that reduced hypercapnia was associated with significant and sustained improvements in overall mortality. Thus, long-term NPPV seems to offer important benefits in this patient group, but the treatment success might be dependent on effective ventilatory strategies.

Keywords:

• chronic obstructive pulmonary disease
• chronic respiratory failure
• hypercapnia
• non-invasive positive pressure ventilation
• systematic review

Introduction

Chronic obstructive pulmonary disease (COPD) is a major cause of morbidity and mortality and represents a substantial economic and social burden throughout the world (1). It is expected to become the third-leading cause of death by 2020 (2). COPD is characterized by progressive airflow obstruction and the destruction of lung parenchyma, which in turn leads to reduced alveolar ventilation, nocturnal and daytime gas exchange abnormalities, dyspnea, increased work of breathing and sleep-disordered breathing. The eventual development of chronic respiratory failure is characterized by varying degrees of ventilation perfusion mismatch, hypoxia, and hypercapnia (3). Once hypercapnia develops, the 2-year mortality is approximately 30–40% (4).

The utilization of non-invasive positive pressure ventilation (NPPV) has now been reported for a variety of clinical indications (5). NPPV has become widely accepted as the standard method of ventilation used in patients with chronic hypercapnic respiratory failure caused by chest wall deformity, neuromuscular disease, or impaired central respiratory drive (6). Additionally, current evidence supports the use of NPPV in acute respiratory failure due to chronic obstructive pulmonary disease (7). Use of NPPV in acute-on-chronic respiratory failure decreases work of breathing, improves ventilation/perfusion (V/Q) matching, decreases fatigue, and increases minute ventilation (5,8,9). As we will discuss, accumulating evidence suggests some of these physiological benefits may be sustained, thereby conferring long-term benefits to chronic therapy. Compared to endotracheal intubation and/or tracheostomy, there are also the added advantages of simplicity, patient comfort, the ability of the patient to talk and swallow, and relative ease of implementation and discontinuation.

Although NPPV has become an accepted management approach for patients with acute-on-chronic respiratory failure due to COPD, it remains unclear whether it can also be beneficial in patients with CRF (10). Theoretically, NPPV could be helpful in patients with stable COPD, bronchoconstriction as well as increased respiratory frequency result in dynamic hyperinflation (DHI), which is progressive (dynamic) because of air accumulation in the lung with each breath as a result of a failure to achieve complete exhalation before the onset of the next breath. DHI creates elevated levels of intrinsic positive end-expiratory pressure (PEEPi or “auto-PEEP”) that can lead to increased work of breathing (WOB) (11). The treatment of DHI is primarily achieved through a reduction in respiratory rate.

It has been suggested that NPPV decreases respiratory rate by augmenting tidal volume, and external-PEEP delivered with NPPV can counterbalance auto-PEEP, which in turn decreases WOB and improves respiratory muscle function by resting the respiratory muscles (12). In addition, NPPV might lower upper-airway resistance, improves gas exchange and the response of the respiratory center to CO₂ thereby favorably impacting both daytime ventilation, nocturnal ventilation and the quality of sleep (11,13). Furthermore, a Cochrane review of 14 randomized trials suggested that NPPV decreased mortality, the need for mechanical ventilation, and complications associated with treatment and length of hospital stay were decreased with NPPV (14). Based on these theoretical benefits investigators have evaluated two strategies of long-term NPPV termed “low intensity” NPPV (LI-NPPV) and “high intensity” NPPV (HI-NPPV). Mounting evidence of two treatment strategies for delivery of NPPV in patients with chronic obstructive pulmonary disease is reported in the literature; however the superiority of one strategy over the other remains controversial.

To give an update of the current literature; the present systematic review was undertaken to evaluate and summarise studies examining NPPV in the management of chronic respiratory failure (CRF) in patients with stable-COPD. Furthermore the effectiveness of two NPPV treatment strategies were compared in regard to lung function (forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC)), arterial blood gas tensions (PaCO₂, PaO₂), exercise tolerance, health status (health-related quality-of-life measurements), sleep efficiency, COPD exacerbations, hospitalisations and, finally, survival.

Methods

Search and Selection of Studies

Search terms employed were “bilevel,” “bilevel airway pressure OR bilevel CPAP OR biphasic positive airway pressure,” as well as “non-invasive OR nasal ventilation OR intermittent OR positive pressure ventilation OR NPPV.” Electronic databases searched included MEDLINE, preMEDLINE, EMBASE, CINAHL, Conference Papers Index, Online Computer Library Centre, Inc., Papers First (Conference Papers), Cochrane Library (including Cochrane Database of Systematic Reviews, DARE and Cochrane Controlled Trials), American College of Physicians Journal Club, PubMed, Biological Abstracts, and Dissertation Abstracts for the years 2000–2015. The following journals were hand searched for the years 2000–2015: American Journal of Respiratory and Critical Care Medicine, Chest, European Respiratory Journal, Lung, The New England Journal of Medicine and Thorax. Reference lists of all articles identified for inclusion were manually screened to identify any additional studies. Only studies reported in English were included.

Quality Assessment of Studies

The quality of the eligible studies was assessed by criteria for assessment of risk of bias provided in the Cochrane Handbook for Systematic Reviews of Interventions (15), which included design allocation, recruitment, inclusion and exclusion criteria, follow-up, control of confounders, description of intervention, data collection, outcome measurement and statistical analysis. The methodological quality of studies was then estimated as low, medium or high.

Results

Low Intensity NPPV (LI-NPPV)

The terminology LI-NPPV been developed to differentiate the strategy used in early studies (mean inspiratory pressure < 18 cm H₂O) of NPPV in chronic COPD from the more recently studied HI-NPPV (mean inspiratory pressure ∼28 cm H₂O). The evidence to support the use of LI-NPPV in the setting of severe, stable COPD has been inconsistent. Although some studies have found LI-NPPV to be beneficial in patients with CRF due to stable-COPD (Table 1) (16–21), most of the reports have been predominately unfavorable (Table 2) (22–30). For instance, a two-year prospective multicenter randomized controlled trial evaluated the effect of NPPV plus long-term oxygen therapy (LTOT) versus LTOT alone in 90 patients with stable COPD.

Table 1. Studies employing LI-NPPV; showing benefit of NPPV

Prospective randomized studies
First author year (reference)	Mean IPAP cmH2O	Mean EPAP cmH2O	Study design	Study Length	Number of patients	Outcomes
Garrod et al. 2000 (16)	16	4	NPPV+ E.T vs. E.T alone	8 weeks	55	Significant Improvement in mean shuttle walk test, chronic respiratory disease questionnaire, and PaO2 in NPPV+E.T group.
Diaz et al. 2002 (17)	18	2	NPPV vs. Sham NPPV	3 weeks	36	Day time PaCO2 and PaO2 and also PFT improved in NPPV group
Sin et al. 2007 (18)	20	4	NPPV vs. Sham NPPV	3 months	23	Blood gasses, 6 MWD, HRV were improved and natriuretic peptide was reduced in NPPV group however FEV1 was not different between groups.
Duiverman et al. 2008 (19)	20	6	NPPV + rehabilitation vs. rehabilitation	3 months	72	Day time PaCO2, cognition domain, fatigue domain and SRI summary score, minute ventilation, tidal volume were improved in NPPV group.
Prospective crossover studies
Ambrosino 1992 (20)	22	0	NPPV vs. standard treatment sequentially	<1 week	8	Significant improvement in pH, PaCO2, PaO2, minute ventilation and tidal volume in NPPV group.
Meecham et al. 1995 (21)	18	2	NPPV+LTOT vs. LTOT sequentially	3 months	14	Day time PaO2 and PaCO2, overnight PaCO2 total sleep time, sleep efficiency, HRQL were improved in favor of NPPV group.

LI-NPPV, Low intensity non-invasive positive pressure ventilation; IPAP, inspiratory positive airway pressure; EPAP, expiratory positive airway pressure; E.T, exercise training; vs,versus; LTOT, long-term oxygen therapy; PFT, pulmonary function test; HRQL, health-related quality of life; SRI, The severe respiratory insufficiency questionnaire; HRV, heart rate variability; 6 MWD, 6-minute walk distance.

Table 2. Studies employing LI-NPPV showing non-benefit of NPPV

Prospective randomized studies
First author Date (reference)	Mean IPAP cmH2O	Mean EPAP cmH2O	Study design	Study length	Number of patients	Outcomes
Renston et al. 1994 (22)	15-20	2	NPPV vs. Sham NPPV	<1 week	9	No difference on PaCO2 and PFT in both groups however improved exercise capacity and Borg dyspnea score in NPPV group.
Gay et al. 1996 (23)	10	2	NPPV vs. Sham NPPV	3 months	13	Lung function, nocturnal oxygen saturation, and sleep efficiency remained unchanged in both groups.
Casanova 2000 (24)	13	4	NPPV+LTOT vs. LTOT	1 year	52	Acute COPD exacerbations, hospital admissions, intubations, mortality rates, gas exchange, pulmonary function, cardiac function did not change in either group. However Borg dyspnea score dropped from 6 to 5 and one of the neuropsychological performance was improved in NPPV group.
Clini et al. 2002 (25)	14	2	NPPV+LTOT vs. LTOT	2 years	90	Lung function, inspiratory muscle function, exercise tolerance, sleep quality score, hospital admissions and survival rates did not change in either group. However dyspnea and HRQL, day time PaCO2 were improved in NPPV group.
McEvoy et al. 2009 (26)	13	5	NPPV+LTOT vs LTOT	28 months	144	FEV1 and PaCO2 measured at 6 and 12 months were not different between groups. However nocturnal NPPV improved survival in stable hypercapnic COPD, but this was at the cost of worsening quality of life.
Prospective crossover studies
Strumpf et al. 1991 (27)	15	2	NPPV vs. standard care sequentially.	6 months	7	NPPV revealed no improvements in pulmonary function, respiratory muscle strength, gas exchange, exercise endurance, sleep efficiency, quality of oxygenation or dyspnea ratings.
Lin 1996 (28)	12	2	O2 monotherapy vs. NPPV sequentially.	6 week	12	NPPV did not improve nocturnal oxygenation and also sleep efficiency was poorer in NPPV group.
Krachman et al. 1997 (29)	22	3	NPPV vs. Sham NPPV sequentially.	< 1 week	6	PaCO2 was not different during NREM and REM sleep in both groups. Sleep efficiency increased by 14% with NPPV. However Sleep architecture was not changed.
Highcock et al. 2003 (30)	12–15	-	Walking with 3 different types of NPPVs vs. unencumbered walking sequentially.	<1 week	8	NPPV did not improve exercise capacity however it increased ventilation.

HI-NPPV, High intensity non-invasive positive pressure ventilation; IPAP, inspiratory positive airway pressure; EPAP, expiratory positive airway pressure; E.T., exercise training; vs, versus; LTOT, long-term oxygen therapy; PFT, pulmonary function test; HRQL, health-related quality of life; COPD, chronic obstructive pulmonary disease.

Researchers implemented NPPV by means of pressure support ventilation (PSV) with a back-up respiratory rate of 8 breaths per minute, a mean inspiratory positive airway pressure (IPAP) of < 14 cmH₂O, and a mean expiratory positive airway pressure (EPAP) of 2 cmH₂O (25). The study demonstrated that lung functions, inspiratory muscle function, exercise tolerance and sleep quality score did not change over time in either group. By contrast the carbon dioxide tension in arterial blood on usual oxygen, resting dyspnea and health-related quality of life (HRQL) as assessed by the Maugeri Foundation Respiratory Failure Questionnaire (31) improved over time in favor of NPPV plus LTOT group. Hospital admissions were not different between groups during the follow-up. Nevertheless, overall hospital admissions showed a trend in favor of the NPPV plus LTOT group (decreased by 45%) as compared with the LTOT group (increased by 27%). Intensive care unit (ICU) stay decreased over time by 75% and 20% in the NPPV+LTOT and LTOT groups, respectively. Finally survival was similar between groups.

The Australian Trial of Non-invasive Ventilation in Chronic Airflow Limitation Study is the largest RCT to date and utilized an IPAP of 12.9 cmH₂O and a mean EPAP of 5.1 cmH₂O (26). Patients with stable hypercapnic COPD (n=144) were randomly assigned to either receive NPPV plus LTOT or LTOT alone. Mean adherence to NPPV was 4.5 h/night. Median follow-up time was 28.5 months in the NPPV+LTOT group and 20.5 months in the LTOT group. No differences were observed between the two groups regarding PaO₂, PaCO₂ or FEV₁ in the first 12 months of follow-up. There was no significant change in other parameters of lung function (DLco, static lung volumes). Hospitalization rates were not different between the two groups. However, there was a survival advantage in favor of the NPPV + LTOT. Specific HRQL as measured by the St George's Respiratory Questionnaire (SGRQ) (32) did not change in the NPPV group, while HRQL as measured by the generic instrument MOS 36-Item Short-Form Health Status Survey (SF-36) (33) deteriorated in two of the eight sub-dimensions (General Health and Mental Health).

Tables 1 and 2 depict the outcomes and study protocols of the prospective, randomized controlled trials and crossover designed studies involving COPD patients who received NPPV via nasal, oronasal and/or total face mask interfaces.

High Intensity NPPV (HI-NPPV)

Windisch et al. have suggested that the variable effects reported above may be related to the relatively low levels of ventilatory support (i.e., LI- NPPV) utilized in these trials where IPAP ranged from 10 to 18 cmH₂O without a back-up rate in most cases (see Tables 1 and 2). This belief led to research into and a small but growing pool of data assessing the effect of increased ventilatory support in chronic stable COPD (Tables 3a and 3b). In what these researchers term HI-NPPV, pressure controlled ventilation is utilized with stepwise titration of IPAP up to 40 cmH₂O, depending on the tolerance and the ability to achieve normocapnia (34,35). The ventilator is initially set with a low back-up rate and a relatively low trigger threshold, so to avoid auto triggering. These initial settings are used in conjunction with low IPAP levels, typically ranging between 12 and 16 cmH₂O, and low EPAP levels. IPAP is carefully increased, in a stepwise fashion up to the point where it is no longer tolerated. The respiratory rate is increased beyond the spontaneous rate attempting to establish controlled ventilation, while EPAP is increased in order to rest the inspiratory muscle by decreasing the work of breathing due to auto-PEEP. The EPAP is usually set between 3 and 6 cmH₂O, depending on individual tolerance, while maintaining an inspiration : expiration (I:E) ratio of approximately 1:2 (36). Although most of HI-NPPV ventilation data has been produced in Germany, a more recent study produced compelling results in the United Kingdom (37).

Table 3a. Studies employing HI-NPPV prospective randomized studies

First author date (reference)	Mean IPAP cmH2O	Backup rate	Study design	Study length	Number of patients	Outcomes
Windisch et al.2006 (40)	31	21	Randomized controlled Trial Half of the patients with COPD; received HI-NPPV other half received symptomatic treatment and physiologic parameters were compared.	2 months	24**	PaCO2 decreased by 13 mm Hg, PO2 increased by 18 mm Hg FEV1 increased by 0.3 L in patients with COPD after 2 months of HI-NPPV. And there was a significant stepwise decrease of PaCO2 during the following 3 h after cessation of nocturnal NPPV in the early morning in HI-NPPV group whereas no change was observed in control group.
Dreher et al. 2007 (34)	29	20	Randomized controlled Trial Patients with COPD (GOLD stage 4), who used nocturnal HMV for 35 months, underwent two 6 MWT with oxygen or oxygen plus NPPV with help of walking rollator. Physiologic parameters and walking distances were measured.	35 months	20	PaCO2 increased by 0.4 kPa during the 6MWT when supplemental oxygen was used, but did not significantly change when NPPV was used in addition to supplemental oxygen. PaO2 significantly increased by after walking with NPPV, but significantly decreased by without NPPV. Dyspnea, as assessed by the BDS, significantly decreased from 6 to 4 and walking distance significantly increased from 209 m to 252 m when walking was NPPV-aided.
Dreher et al. 2009 (52)	29	19	Randomized crossover study. Patients with COPD (GOLD stage 4) used nocturnal HMV for mean of 27 months, underwent 12 MWT with oxygen or oxygen plus NPPV without help of walking rollator. NPPV replaced backpack of patients.	27 months	11	After 12MWT; PaO2 decreased by 8 mm Hg in oxygen group whereas 6 mm Hg increased in NPPV group. PCO2 increased in both study arms. BDS increased in both groups. FEV1 increased 0.1 L in NPPV group whereas 1.4 L decreased in O2 group. However, walking distance was 64 m shorter in the NPPV group than oxygen group. Authors think it is due to the burden of carrying the NPPV in a backpack.
Dreher et al. 2010 (41)	29 vs 15	18 vs assist controlled	Randomized controlled crossover study. Compared 6 weeks of HI-NPPV using controlled ventilation with LI-NPPV in patients with stable COPD.	12 weeks	17	Nighttime PCO2, HCO3 decreased in Hi-NPPV group. After the walking test, BDS score was improved in HI-NPPV group. There were no significant treatment effects between LI-NPPV and HI-NPPV in terms of SRI subscales.
Dreher et al. 2011 (61)	29 vs 14	19 vs assist controlled	Randomized crossover study. Assessed the difference in sleep quality in HI-NPPV and LI-NPPV.	42 months	17	In both group there were no changes in terms of slow wave sleep, sleep quality, sleep stages and arousal rates.
Murphy et al. 2012 (37)	29	16 vs 6	Crossover randomized controlled study. Compared high-intensity (high IPAP pressure with back up rate) and high-pressure (high IPAP pressure without back-up rate) ventilation groups.	12 week	12	There were no between-group differences in health-related quality of life, mean nocturnal ventilator usage, gas exchange, objective sleep quality and subjective sleep quality. However the respiratory domain of the Severe Respiratory Insufficiency questionnaire was higher in the high-intensity group.
Ekkernkamp et al. 2014 (42)	25	18	Minute ventilation measured in patients who received for a median 22 months HI-NPPV compared with spontaneous breathing.	22 months	27	Long-term HI-NPPV increased minute ventilation by an average of 26% compared with spontaneous breathing in stable hypercapnic COPD patients.
Köhnlein et al. 2014 (77)	22	16	COPD GOLD stage IV patients were randomly assigned either to optimized standard COPD treatment (control group) or to standard COPD treatment plus NPPV (intervention group) for at least 12 months.	Approximately 7 years	195	31 (33%) of 93 patients in the control group, and 12 (12%) of 102 patients in the intervention group died within 1 year after randomization (log rank p=0.0004). PaCO2, pH, SaO, HCO–, and FEV were significantly increased in intervention group. HRQL significantly improved with NPPV treatment. However PaO2, FVC, 6-minute walk distance and residual volume/total lung capacity were not statistically different between groups.

**Half of the subjects were patients with restrictive lung disease.

GOLD, Global initiative for chronic obstructive lung disease; 6 MWT, 6-minute walking test; HMV, home mechanical ventilation; SF-36, Medical Outcome Survey 36-Item Short-form Health Status Survey; BDS, Borg Dyspnea Scale; SRI, the Severe Respiratory Insufficiency Questionnaire; CRF, chronic respiratory failure; IPAP, inspiratory positive airway pressure; EPAP, expiratory positive airway pressure; NPPV, non-invasive positive pressure ventilation; ET, exercise training; vs, versus; LTOT, long-term oxygen therapy; PFT, pulmonary function test; HRQL, health-related quality of life.

Table 3b. Studies employing HI-NPPV non-randomized cohort studies

First author Date (reference)	Mean IPAP cmH2O	Mean EPAP cmH2O	Study design	Study length	Number of patients	Outcomes
Windisch et al. 2002 (67)	30	23	Non-RCT; HI-NPPV was given in GOLD stage 4 COPD patients and physiologic parameters were measured	6 months	14	Normocapnia was achieved in mean of 9 days. PCO2 decreased by 20 mm Hg, HCO3 decreased by 7 mmol/l, PO2 increased by 25 mm Hg.
Windisch 2005 (39)	28	21	Retrospective non-RCT, Nocturnal HI-NPPV was given in GOLD stage 4 patients and physiologic parameters were measured	6 years	34	In a 2 month-period; PaCO2 decreased by 7 mm Hg,HCO3 decreased by 3 mmol/l,PO2 increased by 6 mm Hg, FEV1 increased by 0.14 L.
Windisch et al. 2008 (65)	25	18	Non-RCT; patients received HI-NPPV for 12 months and HRQL and physiologic parameters were measured in the 1^st and 12^th month.	12months	27	In the first month; PaCO2 decreased by 13 mmHg, PaO2 increased by 13 mmHg,HCO3 decreased by 4 mmHg In the 12^th month; PaCO2 decreased by 14 mmHg, PaO2 increased by 12 mmHg, HCO3 decreased by 5 mmHg. General and condition-specific aspects of HRQL which was measured by SF-36 and SRI increased following NPPV establishment.
Windisch et al. 2009 (36)	29	21	Retrospective non-RCT Physiologic parameters, exacerbation rates, long-term survival rates were evaluated in patients with COPD	9 years	73	After one year, PCO2 decresed by 7 mm Hg, PO2 increased by 12 mmHg, Hematocrit levels were improved and hospitalization rates were decreased. Two- and 5-year survival rates of all patients were 82% and 58%, respectively.

GOLD, Global initiative for chronic obstructive lung disease; 6 MWT, 6-minute walking test; HMV, home mechanical ventilation; SF-36, Medical Outcome Survey 36-Item Short-form Health Status Survey; SRI, the Severe Respiratory Insufficiency Questionnaire; CRF, chronic respiratory failure; IPAP, inspiratory positive airway pressure; EPAP, expiratory positive airway pressure; NPPV, non-invasive positive pressure ventilation; ET, exercise training; vs, versus; LTOT, long-term oxygen therapy; PFT, pulmonary function test; HRQL, health-related quality of life.

Impact of NPPV

Physiology

The physiologic response of patients with COPD to both ventilation modalities has been assessed with an overall benefit found in most instances. Whether these physiologic benefits lead to clinically meaningful outcomes remains uncertain.

As shown in Table 1, an end point in these trials of LI-NPPV was gas exchange; all of these trials revealed a significant improvement in PaO₂, PCO₂, or both. Even in the studies employing LI-NPPV without conclusive benefit (Table 2), the two largest studies revealed an improvement in day time PaCO₂ (25) and sleep-related hypercapnia acutely (26). However a recent Cochrane meta-analysis showed that there was no significant difference between NPPV and control groups when looking at PaO₂ and PaCO₂. Interestingly, in the same meta-analysis, patients who used ventilation on average for more than 5 hours per night compared to those who used it for less than 5 hours per night, and patients started NPPV with baseline hypercapnia levels of over 55 mm Hg, than those who with PaCO₂ levels below 55 mm Hg showed a significantly bigger improvement in PaCO₂ after 3 months (38). Additionally patients who received IPAP levels of 18 cm H₂O or higher, which is in line with the Meecham Jones et al. study (21), showed a significant improvement. This raises the question as to whether Hi-NIPPV pressures of around 29 cm H₂O and at least 5 hours of NPPV usage are required to improve gas exchange during the day. However data on HI-NPPV is less aggregated, and the majority of this data comes from one region of Europe. Nonetheless, Tables 3a and 3b illustrate the profound impact HI-NPPV has on gas exchange.

Data supporting changes in lung function have been less robust. Table 2 illustrates studies in which LI-NPPV was shown to have no impact on pulmonary function. A recent Cochrane meta-analysis showed that NPPV has none or a small negative effect on FVC and FEV₁ after 3 and 12 months (38). However a few studies employing LI-NPPV showed statistically significant improvement in pulmonary function (17,20). Whether these improvements have a clinically meaningful impact is debatable. Work by Windisch and Dreher have suggested a statistically significant improvement in FEV₁ in COPD patients utilizing HI-NIPPV; however, the improvements reported were 0.14 L and 0.1 L, respectively (39–41). Recently Ekkernkamp et al. showed that long-term (median 22 months) HI-NPPV increased minute ventilation by an average of 26% compared with spontaneous breathing (42). In a randomized crossover study involving 15 patients, Lukácsovits et al. assessed the physiologic impact of LI-NPPV versus HI-NPPV(43). In their study, HI-NPPV induced a greater reduction in the pressure-time product of the diaphragm per minute suggesting a decreased work of breathing. In 9 out 15 patients the transdiaphragmatic pressure was flat and the pleural pressure became positive in HI-NPPV suggesting that high intensity led to completely controlled ventilation.

Exercise

Hallmarks of advanced COPD include dyspnea, reduction in exercise capacity, and hypoxemia (44,45). Patients with COPD are at increased risk of dying from cardiovascular complications following exercise-induced hypoxemia; thus interventions to reduce dyspnea, the hypoxemic and cardiac output, burden may be of clinical importance (46). Use of NPPV has been illustrated to improve exercise tolerance (47) and cardiac output in patients with COPD (43); this improvement was more prominent in Hi-NPPV strategy (48), even though this is speculative, as a reduction in cardiac output may simply reflect that the heart is more “rested” because of reduced oxygen consumption that results from rested respiratory muscles.

NPPV has been used as an aid to exercise training in COPD patients. Although to date, it is difficult to arrive at any definitive conclusion on the role NPPV in exercise/pulmonary rehabilitation, a few studies have suggested benefit. An early study utilized non-invasive pressure support ventilation (∼10 cm H₂O), while patients performed constant load bicycle exercise (49). Participants were found to have a marked objective reduction in respiratory effort and reported a significantly improved sensation of breathlessness. Subsequent studies utilizing pressure support during exercise concluded that use of NPPV leads to unloading of the respiratory muscles and a significant delay in the emergence of exercise-induced lactic acidosis (50,51). Supporting previous studies, a recent meta-analysis result for 6-minute walk distance (MWD) with upper limit of 66 m could be promising in patients with COPD. Articulating the meta-analysis “NIPPV probably has a beneficial effect on walking distance at least in a subgroup of patients” (38).

In hopes of assessing the impact of HI-NPPV on exercise, Dreher and his co-investigators subsequently performed two studies. In one study of crossover design, 20 patients with advanced COPD underwent a 6-minute walk test with a rollator, supplemental oxygen, with and without HI-NPPV (Figure 1) (34). Average IPAP for this study was 29 H₂O; average EPAP 4 H₂O; and average respiratory rate was 20 bpm. Implementation of HI-NPPV resulted in improved oxygenation, decreased dyspnea, and increased walking distance, without significant change in PaCO₂ (Figure 2). A subsequent study aimed to identify the most effective means of preserving oxygenation during activity in patients with advanced COPD (52). Again HI-NPPV produced a superior improvement in PaO₂ during walking however walking distance and dyspnea were not improved. This observation was attributed to the burden of carrying the heavy ventilator equipment in a backpack (Figure 3).

Figure 1. A chronic obstructive pulmonary disease patient during a 6-minute walking test while walking with a rollator [Reproduced by permission of Dr. Wolfram Windisch from Dreher et al. (34)).

Figure 2. Changes in arterial oxygen tension (PaO₂) before and after 6-minute walking test while on a) supplemental oxygen and b) non-invasive positive-pressure ventilation in addition to supplemental oxygen. [Reproduced by permission of Dr. Wolfram Windisch from Dreher et al. (34)].

Figure 3. A chronic obstructive pulmonary disease patient during a 6-minute walking test while walking with a backpack [Reproduced by permission of Dr. Wolfram Windisch from Dreher et al. (52)].

Nocturnal Hypoventilation and Sleep

During sleep, minute ventilation, ventilatory responsiveness, and chemoresponsiveness to CO₂ progressively decrease as the depth of sleep increases, with PaCO₂ rising to maximum during rapid eye movement sleep (53). This decrease in ventilation does not generally result in significant hypoxemia in normal individuals (54). However, nocturnal ventilatory drive is affected to a larger extent in patients with COPD compared to normal individuals, which can result in clinically significant alveolar hypoventilation with resulting hypoxemia (55).

Controlling nocturnal hypoventilation with NPPV has not been objectively assessed in the majority of studies. In an attempt to examine the impact of NPPV on nocturnal hypoventilation, GOLD Stage IV COPD patients were randomized and crossed over to receive either HI-NPPV followed by LI-HIPPV or LI-NPPV followed by HI-NPPV for 6 weeks (41). This crossover comparison demonstrated that HI-NPPV compared to LI-NPPV led to superior improvements in nocturnal PaCO₂, lung function, and HRQL as measured by the Severe Respiratory Insufficiency Questionnaire (SRI) (56). No significant differences were observed in daytime PaCO₂, lung function, 6MWT, maximum inspiratory pressure or the summary scale of SRI.

Patients with COPD are more likely to report regular use of hypnotic medications, difficulty falling or staying asleep, and daytime sleepiness (57). Nocturnal desaturation is associated with more sleep fragmentation and arousals, most notably during REM sleep, therefore nocturnal desaturation is suggested to be the predominant conditions impacting on sleep quality in patients with severe COPD (55,57). Studies with Li-NPPV a showed significant increase in sleep efficiency and total sleep time (TST). However sleep architecture, expressed as percentage TST was not changed with NPPV (21,29,58). Patients in these studies showed a significant increase in nocturnal PO₂ and decrease in PCO₂ that may have contributed to increased sleep efficiency. A few study with Li-NPPV showed no significant change in sleep efficiency; yet it should be noted that patients in these studies were not hypercapnic (27,59). In a recent Cochrane review, sleep efficiency was deteriorated after 3 months of NPPV; however, this effect was due to a small number of studies and, therefore, participants (38).

Although Li-NPPV has a positive effect on sleep efficiency, Hi-NPPV ventilation unearths two questions.  First, since air leak during NPPV is known to disrupt sleep architecture (60), does the greater air leak created by Hi-NPPV has a negative impact on sleep quality, when compared to the conventional approach of LI-NPPV? Second, what is the effect of higher pressures on sleep quality, and specifically, can it affect slow-wave sleep (NREM stages 3) and REM sleep?

Surprisingly, HI-NPPV does not appear to negatively impact sleep quality nor adherence to NPPV. An open label, crossover RCT was performed with the primary end point of sleep quality (61). Patients were randomized to two consecutive nights of HI-NPPV followed by LI-NPPV or LI-NPPV followed by HI-NPPV. Sleep was evaluated with attended polysomnography. Higher inspiratory pressures did not adversely affect sleep quality compared with lower pressure. Two additional trials utilizing pressure limited and volume limited NPPV, both suggested that HI-NPPV does not have a significant adverse impact on sleep quality (39,62).

Adherence to NPPV

NPPV therapy is notably intrusive and difficult for many patients to tolerate (63). The most common side effects that lead to poor adherence with NPPV are pressure intolerance (of particular concern in HI-NPPV) and difficulty exhaling against a positive pressure mask (64). Other adverse effects reported include skin breakdown, air leak, pressure related changes to the oral/nasal airway, and consequences of aerophagia (see Table 4) (65,66). Interestingly, in a recent crossover RCT revealed participants in HI-NPPV spent an average of 3.6 additional hours per day on NPPV compared with those treated with LI-NPPV, In addition, four patients could not tolerate LI-NPPV, whereas all patients tolerated HI-NPPV (41). Moreover, a few other small data sets also support this notion that more aggressive NPPV does not lead to worsened adherence in the COPD population (36,39,67).

Health-Related Quality of Life (HRQL)

Evaluation of HRQL is an important patient centered outcome measure that is increasingly being incorporated into outcomes research of critically ill patients (68,69). Unfortunately data regarding the impact of NPPV on HRQL is scant at best and trials that have been completed utilized different questionnaires, making pooled statistical analysis impossible. A few small trials suggested improved HRQL when patients with stable hypercapnic COPD were placed on HI-NPPV (41,65). However, this improvement was appreciated only utilizing assessments specifically validated for the COPD population, such as the Severe Respiratory Insufficiency (SRI) Questionnaire (56,70). The largest study to date involving LI-NPPV suggested a negative impact on HRQL despite an actual survival advantage (26). This study was supported by a meta-analysis that NPPV had a negative impact on HRQL after 12 months of treatment (38). However, outcome of meta-analysis was not considered appropriate due to the scant number of studies used in the statistics of study.

COPD Exacerbations and Hospitalizations

COPD exacerbations and hospitalizations related negatively to COPD have a tremendous negative impact on patient prognosis, HRQL, and on healthcare costs (71). Any modality reducing the frequency of exacerbations and hospitalizations for COPD would not only be of great benefit to the patients but also in the rationing of scarce medical resources. At this juncture the impact of either LI or HI-NPPV cannot be ascertained.

Data from randomized controlled trials have been conflicting. Meecham-Jones et al. found that LI-NPPV added to LTOT not only significantly improved gas exchange but also decreased hospital admissions at 3 months (21). However this decrease in hospital admissions was no longer present at 12 months. In a multicenter Italian study, LI-NPPV with LTOT was compared to LTOT alone (25). Hospital admissions and survival did not differ between the two groups during the 2-year follow-up period. In another study, >50% of COPD participants treated with LI-NPPV required hospitalization due to exacerbations during 1-year follow-up (72).

To our knowledge no RCTs assessing the impact of HI-NPPV on rates of exacerbations nor hospitalizations in patient with severe chronic COPD have been completed. One observational study employed HI-NPPV plus LTOT in CRF (36). Twenty-two percent of participants required hospitalization during the first year, with most being treated on the general ward. This small data set seems to suggest that HI-NPPV may decrease hospitalization rate or at least the need for ICU admission in the setting of an acute exacerbation.

Survival

The clinical course of COPD is characterized by a high morbidity and mortality despite long-term oxygen therapy (73). The benefit of NPPV in acute exacerbations of COPD plus the evidence that nocturnal NPPV improves survival in restrictive chest wall disease and neuromuscular disease, has raised the possibility that NPPV may also improve survival in patients with end stage COPD (74). Prior reports of historical cohorts demonstrate that survival rates did not change with NPPV (24,25,75,76). However, a study with a large population shown that nocturnal LI-NPPV may improve survival in stable oxygen-dependent patients with hypercapnic COPD, but at the cost of possible worsening in quality-of-life (26). Utilizing HI-NPPV, the 2-year survival rate was 86% in one report (39). In another observational study, the 2-year survival was 82% and the 5-year survival rate was 58% with HI-NPPV (36).

Recently Köhnlein and colleagues showed a substantial improvement in survival, (1-year mortality in the NPPV group was 11.8% vs 33.3% in the control group) and additionally, continuous 1-year NPPV treatment was associated with significant improvements in PaCO₂, pH, bicarbonate, FEV₁, and HRQL (77). The key difference between this study and other studies is that the protocol required a substantial reduction in CO₂ (20% during spontaneous breathing during the initiation phase). This reduction was achieved by a combination of increasing inflation pressures and the back-up respiratory rate. A high back-up rate might be more important than increased inflation pressures since patient ventilator asynchrony during sleep is common in patients with COPD and reduces the tolerance and effectiveness of ventilation (78).

Opposing Views

Murphy et al. challenge the view that high-intensity NPPV is required to achieve both physiological and clinical improvement in stable hypercapnic COPD (37). As an opposing argument, they designed a crossover study and allocated stable hypercapnic COPD patients to either high-pressure with high backup rate named “high-intensity NPPV” or high pressure with low backup rate, named “high-pressure ventilation” by authors for a 6-week period. Their study showed that there was no difference in both treatment arms in terms of ventilator adherence, clinical and physiological parameters. They concluded that it is not the high back-up rate but rather the high-pressure component of high-intensity non-invasive ventilation that is the key factor in achieving its therapeutic role in the management of hypercapnic respiratory failure in COPD patients, (see Tables 3a and 3b).

Discussion

Researchers have theorized that NPPV may have a positive impact on chronic stable COPD via numerous mechanisms. Both LI-NPPV and the relatively newly studied HI-NPPV have shown some promise in correcting physiologic abnormalities. Yet to date, large studies powered to assess the impact of these modalities on rates of exacerbations, hospitalizations are lacking. Also, with regards to HI-NPPV the data has been dominated by German investigators, although researchers in other countries have begun to evaluate this modality (37,43). The authors of a recent Cochrane review nicely summarized the current state of NPPV in chronic COPD quoting, “the small sample size precludes a definitive statement regarding the clinical implications of NPPV, other than stating that at present there is insufficient evidence to support its widespread us” (38). However- as discussed above- the addition of at least 5 hours per day and IPAP of 18 cm H₂O or higher, long-term NPPV to standard treatment which is targeted to greatly reduce hypercapnia; improves overall survival, exercise capacity, arterial blood gases and HRQL in patients with hypercapnic, COPD. This assumption seems to be supported by findings of Köhnlein et al. and Cochrane review (38,77).

There is currently a lack of resources to support this approach to patients with advanced COPD. Pulmonary physicians and respiratory therapists have become very skilled in initiating NPPV in an ICU setting for acute-on-chronic respiratory failure due to COPD. This is not the case for stable chronic respiratory failure. To be successful, additional training of respiratory therapists and pulmonologists will certainly be required. It is likely that these efforts will need to be harmonized with sleep and pulmonary rehabilitation programs. Optimal mask fit, the provision of entrained supplemental oxygen, and the need for in line humidification will likely be key components of successful therapy. Incorporating an objective assessment of sleep quality may be vital to successful therapy.

Conclusions

Clinicians are beginning to realize it is naïve to assume that there is one COPD phenotype in which NPPV can be utilized. Delivering a comfortable assisted breath requires adjustment in the sensitivity of triggering, flow, and inspiratory to expiratory ratio to optimize ventilator synchrony, comfort and subsequent long-term adherence. Whether new variations of assisted ventilation with pressure support, such as average volume assured pressure, support will be more comfortable and / or effective remains to be determined.

NPPV for chronic respiratory failure due to COPD makes physiologic sense; however, should these modalities conclusively prove to be effective with large RCTs, implementation may prove difficult. As larger studies are completed, and, in the case of HI-NPPV, therapy is applied to more diverse populations, a clearer picture of the overall impact and the appropriate COPD phenotype for implementation of either LI-NPPV or HI-NPPV should emerge.

Declaration of Interest Statement

The author has reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article and is fully responsible for the content and writing of this paper.

Acknowledgment

The author is indebted to Prof. Dr. Patrick Strollo (Former President of American Academy of Sleep Medicine) for his excellent technical and editorial assistance and linguistic revision.