Why High-Intensity NPPV is Favourable to Low-Intensity NPPV: Clinical and Physiological Reasons

ABSTRACT

High-intensity non-invasive positive pressure ventilation (NPPV) was originally described for chronic hypercapnic chronic obstructive pulmonary disease (COPD) patients in 2009, and refers to a specific ventilatory approach whereby NPPV settings are aimed at achieving the lowest arterial partial pressure of carbon dioxide (PaCO₂) values possible. Thus, high-intensity NPPV requires ventilator settings to be increased in a stepwise approach to either an individually tolerated maximum, or to the levels necessary to achieve normocapnia. This differs from the classic approach to low-intensity NPPV, which comprises considerably lower ventilator settings and typically fails to lower elevated PaCO₂ values. The ongoing discussion about whether or not long-term NPPV should be used in chronic hypercapnic COPD patients is based on the observation that many studies in the last two decades have failed to provide evidence for this particular patient cohort. In addition, these trials preferably used low-intensity NPPV. There is now, however, increasing evidence to suggest that high-intensity NPPV is capable of improving important physiological parameters such as blood gases and lung function, as well as health-related quality of life. Moreover, this approach also produced positive outcomes following two recent randomized controlled trials, e.g., improved survival rates in stable COPD patients, and admission-free survival in patients with persisting hypercapnia following acute in-hospital NPPV to treat acute acidotic respiratory failure. As a consequence, the time has now come to evaluate the impact of long-term NPPV on both the physiological and clinical outcomes, with emphasis on the different approaches to NPPV. Therefore, the aim of the current review article is to elaborate on the clinical and physiological reasons for why high-intensity NPPV is favourable to low-intensity NPPV.

KEYWORDS:

• COPD
• heart failure
• hypercapnia
• non-invasive positive pressure ventilation
• quality of life
• respiratory muscles

Introduction

There is an ongoing debate about whether long-term non-invasive positive pressure ventilation (NPPV), delivered by a nasal or a full-face mask, is useful for patients with chronic hypercapnic chronic obstructive pulmonary disease (COPD) (1–3). The most recent update of the GOLD guidelines (November 17th, 2016) emphasized that because various controlled trials had reported contradictory results about the clinical benefits of long-term NPPV in these patients, the evidence needed to formulate recommendations is therefore insufficient. Nevertheless, the, GOLD report also indicated that long-term NPPV may improve hospitalization-free survival in selected patients (4). Similarly, a recent meta-analysis that covered seven trials (245 patients) and was published in 2014 concluded that there is still insufficient evidence to support the routine application of NPPV in patients with stable COPD. However, the authors of the latter report also pointed out that higher levels of inspiratory positive airway pressure (IPAP), better compliance and higher baseline arterial partial pressure of carbon dioxide (PaCO₂) values were associated with improved hypercapnia. Thereby, it has been clearly acknowledged that both the settings and the techniques used for long-term NPPV application substantially influence clinical outcomes (3).

More recently, two randomized controlled trials (RCTs) using relatively different approaches to those in previous studies also produced positive outcomes, which were attributed to the ability of long-term NPPV to improve PaCO₂(1,2). Therefore, the time has now come to change the direction of questioning. In this regard, the question is no longer whether long-term NPPV is beneficial to chronic hypercapnic COPD patients. It is rather: What are the best settings and techniques for long-term NPPV for COPD patients?

In recent years, these more aggressive forms of NPPV aimed at normalizing elevated PaCO₂ values have been described as high-intensity NPPV (5). This is in contrast to the traditional approach to NPPV, which comprises less aggressive ventilator settings and is referred to as low-intensity NPPV. The current article describes the differences between these two approaches, with specific emphasis on both the physiological and clinical impact. The evolution of NPPV techniques used over the last two decades is also discussed, and the most recent evidence for the current long-term NPPV techniques used in COPD is summarized. A synopsis of the two approaches is provided in Table 1.

Table 1. Synopsis of physiological and clinical differences between low- and high-intensity long-term non-invasive positive pressure ventilation (NPPV), as used for chronic hypercapnic COPD patients.

	Low-intensity NPPV	High-intensity NPPV
Description/definition	Classic mode of NPPV using assisted ventilation with an IPAP typically < 18 cm H2O	NPPV in the assist/control mode aimed at maximally improving PaCO2, with IPAP typically 20–30 cm H2O
Hypercapnia	Not capable of improving PaCO2	Capable of improving PaCO2, both during NPPV and subsequent spontaneous breathing (as result of an improved breathing pattern)
Acclimatization to NPPV	High-intensity NPPV requires more time in hospital for sufficient acclimatization
Tolerance/adherence to therapy	Once NPPV is carefully established in hospital, adherence to therapy is higher for high- compared to low-intensity NPPV
Lung function	Does not improve lung function	Improves lung function
Inspiratory effort	High-intensity NPPV is more effective than low-intensity NPPV in reducing inspiratory effort
Inspiratory muscle strength	Improvements shown neither in low- nor high-intensity NPPV
Cardiac function	Physiologically, high-intensity NPPV is more likely to reduce cardiac output compared to low-intensity NPPV. However, clinical data show that there are no contraindications in heart failure patients for either mode, while a long-term reduction in proBNP has been reported for high-intensity NPPV
Sleep quality	Sleep quality is similarly preserved during each mode of NPPV, even though leak volumes during nocturnal ventilation are higher for high- compared to low-intensity NPPV
HRQL	Conflicting results: generic aspects of HRQL are not improved and can even deteriorate, while there are some positive effects on specific aspects of HRQL	Substantially improves both generic and specific aspects of HRQL
Outcome for patients started on NPPV during stable disease	Conflicting results: Most studies show no effect on long-term survival, while one study shows a small survival benefit	Long-term survival has been shown to be dramatically improved
Outcome for patients started on NPPV following exacerbation with acute NPPV	No clear evidence	Admission-free survival has been shown to be improved, but only if NPPV is started in patients with PaCO2 > 7 kPa at least 2 weeks after acute NPPV has been stopped.

HRQL, Health-related quality of life; IPAP, inspiratory positive airway pressure; PaCO₂, arterial partial pressure of carbon dioxide; proBNP, pro-Brain natriuretic peptide.

What is high-intensity NPPV?

High-intensity NPPV was originally described in 2009 (5), even though earlier studies had already reported the use of this technique in COPD patients (6–9). High-intensity NPPV refers to a specific ventilatory approach in which NPPV settings are aimed at achieving either normocapnia (when feasible), or the lowest PaCO₂ values possible. For this purpose, ventilator settings are increased in a stepwise manner until normocapnia is achieved, or at least until the maximum level tolerated by the individual patient is reached (10).

Although IPAP levels are typically high, it is important to note that high-intensity NPPV is not defined by specific IPAP cut-off values; rather, it is driven by the physiological aim of maximally improving alveolar ventilation, as evidenced by normalized or improved PaCO₂ levels (at the least) (10). Furthermore, despite IPAP levels being generally high, it should not be increased any further if it is not tolerated by the patient. As a consequence, IPAP levels turn out to be heterogeneous across individuals, given that they depend on the patient's baseline physiology, as well as subjective conditions such as symptoms, side effects, tolerance and adherence to treatment (5). IPAP levels that typically range between 20 and 30 cm H₂O are chosen for the purpose of high-intensity NPPV, which is first established in hospital and subsequently used as home mechanical ventilation. It should also be noted that the term high-intensity NPPV does not refer to a specific ventilatory mode, even though the assist/control mode is most often chosen for high-intensity NPPV with the aim of achieving controlled ventilation (5). Furthermore, caution is needed to discriminate between the level of pressure support typically added to the expiratory positive airway pressure (EPAP) and the absolute level of IPAP when comparing different studies in regard to ventilator settings.

In contrast to high-intensity NPPV, low-intensity NPPV typically uses lower IPAP settings of < 18 cm H₂O during assisted ventilation. Here, ventilator settings are not increased for the purpose of reducing PaCO₂. Low-intensity NPPV has served as the primary approach to all of the outcome studies published prior to 2014 (11–13). Therefore, the low-intensity version is deemed to be the classic approach to NPPV in patients with COPD.

High- versus low-intensity NPPV: Blood gases

Many studies have shown that low-intensity NPPV^{Figure 1} is not capable of improving elevated PaCO₂ levels (14). In 1995, Meecham Jones et al. were the first to demonstrate in a RCT that NPPV reduced PaCO₂ when mean IPAP levels of 18 cm H₂O were used. In this study, PaCO₂ during nocturnal NPPV correlated with daytime PaCO₂ when NPPV was paused (15).

In 2005, Windisch et al. were the first to apply very aggressive ventilator settings in severely hypercapnic COPD patients. The mean IPAP level and respiratory rate in this experimental clinical study were 30 cm H₂O and 23 breaths per minute, respectively, thus establishing controlled NPPV. Consequently, PaCO₂ decreased by a mean of nearly 20 mm Hg during NPPV application (p < 0.001). In addition, there was a high correlation between PaCO₂ during NPPV, and PaCO₂ during spontaneous breathing four hours after NPPV cessation (r = 0.77; p < 0.005) (7). Accordingly, many studies have provided evidence for NPPV being capable of significantly reducing elevated PaCO₂ levels if the high-intensity version is used (6–9,16–17).

Of note, in a randomized cross-over study that directly compared high- and low-intensity NPPV for home treatment, mean IPAP for high- vs. low-intensity NPPV was 29 vs. 15 cm H₂O, respectively (16). Low-intensity NPPV was applied using assisted ventilation, while the assist-control mode with mean respiratory rate settings of 18 per minute was chosen for high-intensity NPPV. The main result of the study showed a mean treatment effect of 9.2 mm Hg (95% CI −13.7/−4.6 mm Hg; p < 0.001) in favour of high-intensity NPPV for the reduction in PaCO₂ (primary outcome) (Figure 1). Of note, high-intensity NPPV in this trial was associated with better compliance rates, health-related quality of life (HRQL), lung function, and exercise-related dyspnea (16).

Figure 1. Effects of high- vs. low-intensity NPPV on nocturnal arterial carbon dioxide tension (PaCO₂) following a 6-week randomized cross-over study (published with the kind permission of xxx (16)): Nocturnal mean ± SD PaCO₂ at baseline and follow-up visits.

Finally, one study was able to show that supplementing high IPAP levels with high back-up rates did not further improve physiology (18). Based on this finding, it was suggested that it is the high-pressure component of high-intensity NPPV that plays a key therapeutic role in the management of hypercapnic respiratory failure in COPD patients. Notably, the most recent outcome studies in which NPPV was shown to improve PaCO₂ reported positive results using higher IPAP levels compared to previous RCTs; however, mean respiratory rates were lower compared to those in the original description of high-intensity NPPV (1,2). This observation therefore supports the notion that IPAP levels are the most important setting for successful NPPV.

High- versus low-intensity NPPV: Breathing pattern, respiratory muscles, lung and cardiac function

As mentioned above, using the high-intensity approach to long-term NPPV can improve blood gases not only during NPPV application, but also during periods of subsequent spontaneous breathing (5–7,9). The mechanisms by which NPPV can contribute to an improvement in spontaneous breathing are not fully understood, however, three basic underlying theories exist (19–21): (1) respiratory muscle rest; (2) re-setting CO₂ sensitivity in the central breathing centre; and (3) changes in pulmonary mechanics. These mechanisms are more likely to be complementary rather than contradictory, but have been studied in various conditions and different patient groups. Thus, the most important mechanisms for improved spontaneous breathing following a period of NPPV in COPD patients still need to be elucidated.

The question of^{Figure 2} whether NPPV influences respiratory muscle function has been targeted by several trials. However, volitional tests such as the widely-used maximal inspiratory pressure method are highly dependent on the patient making a truly maximal effort (22), and normal values appear to be greatly reliant both on the study cohort and on several methodological aspects. This has led to the formulation of contradictory regression equations for calculating normal values for maximal inspiratory pressure (23), thus rendering volitional tests unsuitable for physiological studies. Therefore, the studies that have used this technique will not be discussed in this article, due to conflicting results. Alternatively, the assessment of non-volitional transdiaphragmatic twitch pressures following magnetic phrenic nerve stimulation and the placement of enteral balloon catheters are regarded as gold standard techniques (22). Here, there is one study in COPD patients using these techniques that showed no changes in diaphragmatic muscle strength following NPPV (24).

The improvement in PaCO₂ during daytime spontaneous breathing following nocturnal high-intensity NPPV has been attributed by one physiological study to an improved breathing pattern with increased tidal volume at an unchanged respiratory rate (9). In line with this, another interesting theory for improved spontaneous breathing stems from recent observations that lung function and, importantly, FEV₁, are reportedly improved in COPD patients following high-intensity NPPV (5,7,16). This raises the possibility that long-term NPPV has an effect on the airways themselves. Whether this is due to an anti-inflammatory effect of NPPV, or is the result of chronically fibrosed airways being stretched open, remains speculative (25).

Alternatively, a possible mechanism to explain improved FEV₁ following high-intensity NPPV is a reduction in airway oedema. Importantly, oedema can result from CO₂-associated vasodilatation and is therefore a common finding in hypercapnic COPD patients (14,26). Since oedema is also likely to be present in the airways, reducing elevated PaCO₂ values post high-intensity NPPV would then reverse dilation of the precapillary sphincters, thereby impacting positively on the additional oedema in the airways (25). This, in turn, would eventually improve respiratory mechanics as well as FEV₁. As pointed out above, however, this remains to be investigated.

In another randomized cross-over study, the short-term effects on physiology of high-intensity NPPV (mean IPAP 28 cm H₂O, mean EPAP 4 cm H₂O, mean respiratory rate 22 breath per minute) vs. low-intensity NPPV (mean IPAP 18 cm H₂O, mean EPAP 4 cm H₂O, mean respiratory rate 12 breath per minute) were compared (27). This elegant confirmed that high-intensity NPPV more effectively reduces PaCO₂. In addition, this study also showed that high-intensity NPPV established a greater reduction in the pressure–time product of the diaphragm, as assessed by oesophageal and gastric balloon catheters, and completely abolished spontaneous breathing activity in 9 out of 15 patients. From these data, it was concluded that high-intensity NPPV is more effective than low-intensity NPPV, both in improving gas exchange and reducing inspiratory effort (Figure 2) (27). Therefore, resting the diaphragm could be another mechanism by which high- rather than low-intensity NPPV improves respiratory function.

Figure 2. Respiratory mechanical parameters during spontaneous breathing (SB), low-intensity (Li) non-invasive positive pressure ventilation (NPPV) and high-intensity (Hi)-NPPV in a characteristic subject. P_aw: airway pressure; P_pl: pleural pressure; P_di: transdiaphragmatic pressure (published with the kind permission of xxx (27)).

Another important novel finding from the aforementioned physiological study was that high-intensity NPPV was more likely to significantly reduce non-invasively measured cardiac output than low-intensity NPPV (27). The authors accordingly emphasized that this effect needs to be taken into account when high-intensity NPPV is prescribed to patients with pre-existing cardiac disease, and the interaction between positive pressure application and cardiac function has been well reflected. Nevertheless, clinical data have revealed a reduction in pro-Brain natriuretic peptide values following either high- or low-intensity NPPV (28,29). In addition, a very recent randomized cross-over clinical RCT comparing high- and low-intensity NPPV for home treatment did not show an overall adverse effect of high-intensity NPPV on cardiac performance, although it could reduce cardiac output in patients with pre-existing heart failure (30,31). Therefore, there is currently no reason to withhold high-intensity NPPV from COPD patients due to fear of adverse cardiac outcomes, although heart function should be checked regularly in those patients with pre-existing heart disease.

High- versus low-intensity NPPV: Sleep quality and health-related quality of life

The above-mentioned Dreher et al. study in which high-intensity NPPV was found to be physiologically and clinically superior to low-intensity NPPV, used external pneumotachygraphy to measure inspiratory and expiratory volumes, as well as leak volumes (16). Therefore, considerably higher leak volumes were observed during high-intensity NPPV compared to the low-intensity approach. This raised the question of whether high-intensity NPPV could cause sleep disturbances, and whether the improvement in hypercapnia is achieved at the cost of reduced sleep quality. A subsequent randomized cross-over study from the same group revealed that sleep quality is well preserved to a similar degree during both forms of NPPV, even though this study again confirmed the superiority of high-intensity NPPV in improving gas exchange (17).

Another randomized cross-over study confirmed that the sleep quality is acceptable during high-intensity NPPV and that switching to target volume NPPV does not provide any further benefits, both in terms of physiology and sleep quality (31). Therefore, there is currently no data to suggest that more aggressive forms of NPPV such as high-intensity NPPV have a negative impact on sleep quality.

The challenge remains, however, to convince patients to use higher pressures during the acclimatization phase of NPPV. Experience is inevitably needed to help patients get used to high-intensity NPPV. Furthermore, more time in hospital was necessary to establish high-intensity NPPV in the Dreher cross-over trial (mean treatment effect: 2.5 days; 95% CI 1.3/3.7 days; p = 0.001) (16). This, however, was deemed to be justified, given the physiological and clinical advantages (see above). In addition, adherence to NPPV was higher with high- compared to low-intensity NPPV, once patients had been carefully adapted to NPPV (mean treatment effect: 3.6 hours; 95% CI 0.6/6.7 hours; p = 0.024) (16).

For the assessment of HRQL, it is fundamentally important that highly specific assessment tools are used when specific treatment interventions are longitudinally assessed in well-characterized patients. The Severe Respiratory Insufficiency (SRI) questionnaire is currently the most frequently used instrument for HRQL assessment in patients receiving NPPV (33). The SRI questionnaire has even been specifically validated for COPD patients under NPPV therapy (34), and has been shown to be capable of scoring the best in comparison to other established questionnaires in the specific subgroup of COPD patients receiving long-term NPPV (35). The SRI questionnaire was originally developed in German, but has also been professionally translated into a number of languages including English, French and Spanish, and has been re-validated to ensure international comparability. Information about the current status of the international adaptation can be found on the home page of the German Society of Pneumology and Mechanical Ventilation (36).

Here, all versions of the SRI questionnaire can be downloaded for non-profit purposes free of charge. There are many studies in which the SRI questionnaire reveals HRQL improvements in COPD patients following high-intensity NPPV commencement (14). In one large observational study, improvements in HRQL were detected both by the SRI questionnaire and the Short Form 36 (SF-36) following high-intensity NPPV in COPD patients, and these improvements were comparable to patients with restrictive thoracic and those with neuromuscular disorders (37). In contrast, McEvoy and co-workers used the SF-36 to show that generic aspects of HRQL were reduced in the NPPV group compared to controls, while the COPD-specific St. Georges Respiratory Questionnaire revealed a lack of difference between the two groups in terms of specific aspects of HRQL. Even though NPPV did provide a small survival benefit, this was reportedly at the expense of HRQL (13). However, HRQL was not assessed with tools specific to chronic respiratory failure. In addition, low-intensity NPPV which is not capable of improving PaCO₂ was used in this trial.

Therefore, NPPV techniques that do not have the capacity to augment alveolar ventilation are also incapable of improving HRQL. Furthermore, the generic aspects of HRQL can even deteriorate in association with low-intensity NPPV use, and this is possibly attributable to the side effects of NPPV, or at least to the burden caused by the need to wear a mask every night without any perceived benefit. Accordingly, mean adherence to NPPV was low in this study (4.5 ± 3.2 hours per night). Taken together, these findings suggest that low-intensity NPPV, in contrast to high-intensity NPPV, does not improve HRQL. This is supported by the Dreher trial showing that SRI scores improve after high-intensity, but not after low-intensity, NPPV. Finally, the two most recent RCTs that used higher IPAP levels to improve PaCO₂, observed not only an improvement in outcome (see below), but also improved specific aspects of SRI-assessed HRQL (16).

High- versus low-intensity NPPV: Outcome

Older outcome studies (published between 2000 and 2009) in which low-intensity NPPV was used have yielded conflicting results. The first two studies published in 2000 (11) and 2002 (12), respectively, showed no survival benefit, while the third study published in 2009 (13) showed a slight survival benefit for patients undergoing long-term NPPV, but this was reportedly at the expense of generic HRQL, which was reduced in comparison to controls (see comments above).

In contrast, a more recent German study from 2014 demonstrated both a substantial survival benefit, as well as improved HRQL following NPPV commencement (1). The key difference between the Köhnlein et al. study (1), and the three others mentioned above (11–13) is that the Köhnlein protocol required a 20% reduction in PaCO₂ during subsequent spontaneous breathing periods. This criterion was in line with the definition of high-intensity NPPV, even though settings were slightly lower than those originally described for high-intensity NPPV: mean IPAP/EPAP: 22/5 mbar; mean back-up respiratory rate 16 per minute. In this large RCT, 195 patients were randomly assigned to either the NPPV group (n = 102) or control group (n = 93). The one-year mortality rate was 12% (12 of 102 patients) in the intervention group, and 33% (31 of 93 patients) in the control group (hazard ratio 0.24; 95% CI 0.11–0.49; p = 0.0004).

In the most recent British trial (2), high-intensity NPPV using median IPAP/EPAP settings of 24/4 cm H₂O and a median back-up respiratory rate of 14 per minute was used in addition to LTOT following exacerbation that required acute NPPV in hospital; the effects of high-intensity NPPV plus LTOT were then compared to those of LTOT alone. Notably, patients were randomized in this trial if hypercapnia (PaCO₂ > 7 kPa) persisted for at least two weeks. This is in contrast to the Dutch study where NPPV was also used post exacerbation that required acute NPPV in hospital, since long-term NPPV was already started if hypercapnia had persisted for only two days after acute NPPV could have been stopped (38). The primary outcome in the British study was admission-free survival. Of note, the number of patients needed to be treated with NPPV was 6 for this outcome parameter (2). In contrast, both the control group and the treatment group in the Dutch trial experienced a significant reduction in PaCO₂, but there was reportedly no outcome benefit for the NPPV group.

These data therefore provide further evidence that long-term high-intensity NPPV is capable of improving: (i) survival in stable hypercapnic COPD patients (1) and (ii) admission-free survival in COPD patients with status post-acute exacerbation who required acute NPPV in hospital, and had persistent hypercapnia for at least 2 weeks after the cessation of acute NPPV in hospital.

Future considerations and unmet needs

Following the enduring discussion on the usefulness of long-term NPPV for chronic hypercapnic COPD patients, there is now increasing evidence to suggest that high-intensity NPPV confers both physiological and clinical benefits. However, there are still some important issues that need to be addressed:

Firstly, it remains unclear how to best select patients for high-intensity NPPV therapy. Even though there is growing evidence to suggest that patients with more severe hypercapnia benefit the most, it is not known how many of these patients are offered long-term NPPV, nor how many of them end up not receiving it. This is crucial, since high-intensity NPPV in particular has been shown to be advantageous in clinical trials performed by experienced research groups, however, its feasibility in real life still remains to be elucidated.

Secondly, the process of acclimatizing patients to NPPV in a hospital setting is expensive and dependent on the availability of a hospital bed. This is particularly true for high-intensity NPPV, for which more days in hospital are required to become established. This, however, is inevitably needed to guarantee clinical success. Therefore, outpatient commencement and control of NPPV aimed at reducing costs and overcoming hospital-bed shortages should be investigated in the future.

Thirdly, based on the observation that the establishment of high-intensity NPPV is complex, scientific research should also target alternative treatment strategies to treat chronic hypercapnic respiratory failure in COPD patients. Interestingly, two very recent short reports based on the findings in acute respiratory failure patients (39) also demonstrated a potential role for high-flow oxygen therapy as an alternative to NPPV in chronic hypercapnic COPD patients, since this approach improved PaCO₂, breathing pattern and inspiratory effort (40,41). These two studies were small, however, calling for more research in this direction to verify the findings.

Funding

WW received funds for research from the following companies: Weinmann, Germany; Vivisol, Germany; VitalAire, Germany and Heinen und Löwenstein, Germany.

Declaration of interest

The authors SBS, FSM and WW have accepted speaking fees and/or travel funding from companies involved in mechanical ventilation.

Acknowledgments

We acknowledge Dr. Sandra Dieni, Ph.D. for helpful comments on the manuscript before submission.