25 Non-invasive ventilation

During the last 25 years, the use of non-invasive ventilation has grown substantially. Non-invasive ventilation refers to the delivery of positive pressure to the lungs without endotracheal intubation and plays a significant role in the treatment of patients with acute respiratory failure and in the domiciliary management of some chronic respiratory and sleep disorders. In the intensive and acute care setting, the primary aim of non-invasive ventilation is to avoid intubation, and it is mainly used in patients with chronic obstructive pulmonary disease exacerbations, acute cardiogenic pulmonary oedema, immunocompromised or in the context of weaning, situations in which a reduction in mortality has been demonstrated. The principal techniques are continuous positive airway pressure, bilevel pressure support ventilation and more recently, high flow nasal cannula. Whereas non-invasive pressure support ventilation requires a ventilator, the other two techniques are simpler and can be easily used in non-equipped areas by less experienced teams, including the pre-hospital setting. The success of non-invasive ventilation is related to an adequate timing, proper selection of patients and interfaces, close monitoring as well as the achievement of a good adaptation to patients’ demand.

Non-invasive ventilation (NIV) started in the 1930s when continuous negative extrathoracic pressure was applied (iron lung, chest shell, rocking bed, pneumobelt) in the treatment of epidemic poliomyelitis [1]. In the late 1950s, mechanical ventilation through endotracheal intubation led to a dramatic decrease in the use of NIV. However, after the introduction of pressure support ventilation in the late 1980s, NIV began to be used again in patients with acute respiratory failure (ARF) as an alternative to endotracheal intubation in critical care settings [2, 3] as well as in chronic domiciliary ventilation and sleep disorders. Currently, the rate of NIV in European intensive care units (ICUs) is 12–25% [4, 5], and >50% of all patients with respiratory failure without intubation at ICU admission will end up with NIV. This rate increases to over 90% in patients with chronic obstructive pulmonary disease or acute cardiogenic pulmonary oedema (ACPO). NIV is now a first-line therapy in the emergency department (ED) [6], regular hospital wards [7], palliative [8] or paediatric [9] care units, and even for patients out of hospital [10, 11].

NIV refers to the delivery of ventilatory support or positive pressure into the lungs without an endotracheal tube (ETT) [12, 13]. There are several advantages over ‘invasive’ mechanical ventilation. NIV techniques leave the upper airway intact and may avoid the complications and drawbacks that might occur during the procedure of endotracheal intubation (i.e. upper airway trauma, laryngeal swelling, cardiovascular complications in difficult cases), during mechanical ventilation (i.e. oversedation, infections, weaning failure, tracheal granuloma), and at the time of extubation (i.e. reintubation, vocal cord dysfunction). The most relevant difference with intubated patients is the reduced incidence of infections, especially ventilator-acquired pneumonia [14, 15], due to the preservation of the natural upper airway defense mechanisms. By avoiding intubation and its inherent consequences, the length of stay and mortality in the ICU may be reduced, decreasing financial costs [16]. Furthermore, patients do not need sedation, allowing for oral nourishment and communication.

Besides all these benefits, there are some disadvantages. NIV does not allow the same ventilation performance as invasive ventilation and often requires increasing nursing attention, especially in cases of respiratory impairment due to device uncoupling. In addition, NIV may have some complications [17] (see Table 25.1), which are usually mild and related to the mask. Major complications are very uncommon.

There are three major modes of NIV: continuous positive airway pressure (CPAP), non-invasive pressure support ventilation (NIPSV) and high flow nasal cannula (HFNC).

Although it was introduced earlier, it is not essentially a ‘true’ ventilation mode, because it does not provide any inspiratory support [18]. CPAP can be generated with a simple oxygen source through a hermetical mask with a positive end-expiratory pressure (PEEP) valve or a Boussignac mask, which holds a quantity of air in the lungs on expiration (see Figure 25.1). The continuous positive intrathoracic pressure recruits collapsed alveolar units and increases functional residual capacity (FRC) and lung compliance, improving oxygenation and the work of breathing. However, the control of the fraction of inspired oxygen (FiO₂) can be difficult, unless a mixer or a ventilator is used. CPAP is currently applied to hypoxaemic patients, particularly in those with atelectasis or ACPO.

Unlike CPAP, this modality requires a ventilator and constitutes the essence of NIV. It is usually programmed with two levels of pressure: expiratory positive airway pressure (EPAP) (similar to CPAP) and inspiratory positive airway pressure (IPAP) (see Figure 25.1). When the patient starts the inspiratory effort, the ventilator delivers inspiratory assistance (pressure support) using a decelerated flow, which keeps the prefixed IPAP constant. When the inspiratory flow descends below a preset percentage of its maximum value (usually 25–30%) or the patient ends the inspiratory drive, the ventilator assistance is discontinued and the pressure drops down to the predetermined EPAP (see Figure 25.2).

Whatever NIV technique used, an interface is needed to connect the patient to a ventilator or air/oxygen source (see Figure 25.3). That interface is the component that most defines NIV, and it is crucial for the success of the treatment. In order to avoid leaks, a tight seal between the patient’s face and the device is essential, but it is often difficult to obtain. There are different types of interfaces [36].

This has been proven useful in chronic patients or those with sleep apnoea disorders. Although nasal masks can be better tolerated [36], they are less useful in critical situations, generating more resistance [37, 38] and massive leakage through the mouth, often requiring mask change [39]. Conversely, they permit speech, feeding, coughing, and secretion expectoration, reducing the risk of vomiting [40]. A variant is the nasal pillow inserted into the nostrils, which are commonly used in paediatrics. In neonatal units, heated and humidified HFNC delivering oxygen is widely used to support breathing, as an alternative to nasal CPAP and for weaning off CPAP, despite a lack of supportive evidence [41].

This is the most used interface in clinical practice, in over 70% of all patients requiring NIV [42]. Disadvantages include a lack of protection from vomiting, nasal skin injuries, nasal congestion, mouth dryness, eye irritation, speaking difficulty, and possible claustrophobia [39].

There are three different types of face masks:

The helmet covers the whole head and part of the neck, without contact with the patient’s face. It is very useful when anticipating prolonged NIV treatment. It can be attached to the patient’s axillae, to the abdomen, or even to the bed. It allows more patient autonomy (speaking and eating), but its noise may occasionally be uncomfortable [49]. It is not recommended with traditional ventilators, as a fresh gas flow, high enough to minimize rebreathing, is necessary [50]. It is more useful for CPAP, because the increased dead space may generate asynchrony when NIPSV is applied [51, 52].

Mouthpieces placed between the lips and held in place by lip seals are less effective due to higher leak and asynchrony rates and greater patient discomfort [53, 54]. Laryngeal masks with volume control may be useful for patients in coma.

All the ventilators have particular settings for CPAP. Indeed, there are three types of ventilators for NIPSV: portable ventilators designed specifically for NIV, transport ventilators, and ICU ventilators. Classical ICU ventilators and transport ventilators were primarily configured to be used with endotracheal intubation and provided different levels of monitoring and security alarm systems, but they often failed during NIPSV when leaks were present. Modern ICU ventilators and some transport ventilators have solved this drawback by incorporating NIV algorithms.

NIV ventilators are more economical, have a higher mobility, and do not need an airflow source. A wide range of portable ventilators is currently on the market from the most simple (only pressure is modifiable) to the latest generation high-tech ventilators (monitoring, alarm setting, leakage compensation, different triggers, cycling, and flow ramp control, etc.) [55]. These NIV ventilators usually allow better synchrony than ICU and transport ones, even those that have NIV algorithms [55].

The most important attribute of the equipment is leakage compensation through an increase of airflow (up to 120–180 L/min), which maintains the tidal volume, thus producing better patient–ventilator synchrony and higher efficacy of the system. Since pressure cycling can increase auto-PEEP, a trigger is usually activated with airflow [56]. Auto-PEEP is defined by air trapped at end-expiration due to inadequate time for expiration (too short), bronchoconstriction, or mucus plugging. In this case, flow and pressure curves do not arrive to zero.

NIV is often applied without humidifying devices. However, dry gas provokes dryness of the mouth, nose, and respiratory tract, resulting in nasal congestion and an increase of airway resistance. Consensus statements and guidelines contain conflicting recommendations concerning humidification and interactions with the ventilator system [57]. Heat humidification was recommended, because it seemed to facilitate NIV [58, 59] by reducing nasal resistance, helping expectoration, and improving adherence and comfort [60], especially in patients with secretions. Heat and moisture exchangers were not indicated when using NIV, since they might increase circuit dead space (increased PaCO₂) and the work of breathing [61, 62]. However, in a recent randomized trial, there were no differences in the intubation rate between both systems [63]. When using HFNC it is mandatory to use humidification. On the other hand, nebulizers can be used safely without interrupting NIV therapy. The association of oxygen and helium has been tested, with no clear advantages [64]. It is important to use skin protectors to avoid face mask skin injuries.

NIV can be used in a wide range of disorders that may lead to ARF. There are different levels of evidence to support its use in most of them. This is a key point for the appropriate selection of patients that may benefit from this therapy (see Table 25.3) [36].

Many randomized trials have compared the efficacy of NIPSV with conventional oxygen therapy in patients with exacerbation of COPD, showing NIPSV significantly better in improving gas exchange and symptoms [63,64,65,66,67,68,69]. Several meta-analyses and systematic reviews confirmed the superiority of NIPSV over conventional oxygen therapy by reducing the intubation rate, ICU or hospital length of stay, and mortality [70,71,72,73,74,75]. Therefore, NIPSV should be considered the first-line treatment for decompensated COPD patients [76,77,78] in the ED, ICU, and even on general wards, although the latter is suggested in less severe cases (pH >7.30) [68, 78]. Recent guidelines [12] recommended a trial of bilevel NIV even in COPD patients considered to require endotracheal intubation (unless in cases of immediately deterioration), but not in COPD exacerbation without acidosis.

Although , some studies [81,82,83,84,85] suggest that NIPSV may be effective in improving airflow, correcting gas exchange abnormalities, avoiding intubation, and reducing the need for hospitalization, The evidence is low and it was not recommended in recent guidelines [79, 80, 12].

ACPO and COPD exacerbations are the most frequent indications for NIV in acute care settings [4, 5, 86, 87]. Either CPAP or NIPSV are used in ACPO.

The utilization of CPAP in patients with cardiac failure has a physiological basis. In addition to previously mentioned effects on oxygenation, positive intrathoracic pressure reduces venous return and LV transmural pressure (systolic wall stress) [88,89,90,91]. In patients with normal cardiac function, this may produce a mild decrease in blood pressure and cardiac output. Conversely, in patients with decompensated heart failure with hypervolaemia and elevated preload, it may decrease pulmonary congestion and increase cardiac output [91, 92]. Since 1985, numerous studies have proved the superiority of CPAP over standard oxygen therapy in patients with ACPO, improving gas exchange and symptoms [93,94,95] and reducing the endotracheal intubation rate [90, 96,97,98,99,100]. The technique has also been experienced in the pre-hospital setting in recent French studies, two of them with positive results [11, 101] and one neutral [102]. CPAP has also successfully used with helmet in a recent trial in the pre-hospital area, even as a sole therapy. In most of the studies CPAP is generally et at 10 cmH₂O.

On the other hand, the first randomized trial with NIPSV was published in 2000 [103]. Although a contemporary study carried out in Israel with very low level pressure support showed worse outcomes with NIPSV [104], several trials [100, 103, 105] showed a reduction in the intubation rate, compared to standard therapy, especially in hypercapnic patients [106].

Several meta-analyses were concordant in demonstrating that CPAP and NIPSV reduced nearly to half the risk of endotracheal intubation when compared with standard therapy [107,108,109,110,111]. In addition, both techniques reduced mortality by nearly 40%, although only CPAP reached statistical significance. No superiority of one technique over the other was shown in clinical trials designed to compare both techniques (and sometimes PAV) [112,113,114,115,116,117,118,119,120] or in meta-analyses [107,108,109,110,111], although NIPSV tended to show a faster improvement in some respiratory parameters in some series and PAV was better tolerated.

However, the results of the 3CPO study [121], the largest clinical trial on NIV carried out to date and designed to assess mortality, contradicted the results of the previous meta-analyses. Although NIV patients showed faster improvement in acidosis and respiratory distress, short-term mortality was not different when comparing both techniques to conventional therapy. The fact that patients were not hypoxaemic at study entry (average PaO₂ was 100 mmHg in all groups) and showed a very low intubation rate (<3%) suggested that they were less severe than those included in the meta-analyses [122]. In addition, the design of the trial not restricting the use of rescue NIV would probably have induced a bias against NIV, because there was a significantly higher crossover rate to NIV due to respiratory distress. However, recent meta-analyses, including the 3CPO trial, continued to reflect a significant reduction of mortality rate with CPAP [123, 124] (RR 0.75 [0.61–0.92]) [111], showing maximal benefit in patients with acute myocardial infarction (AMI) or myocardial ischaemia [111]. This finding reinforces the role of NIV in AMI [98, 125], which has also been described recently to be successful during primary angioplasty [126]. However, prognosis in this setting does depend on the severity of myocardial injury, rather than the degree of respiratory failure [125, 127], and a strong impact on mortality may not be expected with the use of NIV. Further research will be necessary to define which patients are most likely to benefit from NIV in terms of mortality [128]. Probably those with a high risk for intubation, defined by severe acidosis, hypercapnia, being non-hypertensive, or having ACS, would be the target population [129]. Recent ESC consensus paper have recommended the use of NIV in patients with AHF who show an increased work of breathing with tachypnoea and oxygen saturation < 90% [130], have defined criteria for diagnosis of ACPO and have addressed the role of different modalities of NIV in several AHF scenarios (Masip et al, 2017; Harjola et al, 2017)

The latest update of the ESC guidelines for the diagnosis and treatment of acute and chronic heart failure, published in 2016, has considered NIV to be a class IIa recommendation [131]. NICE guidelines recommended NIV in patients with APO showing severe dyspnoea and acidemia (National Clinical Guideline Centre, 2014).

In conclusion, NIV is indicated as the first-line therapy in ACPO, since it improves ARF more rapidly and may reduce the intubation rate and mortality when compared with standard therapy. CPAP may certainly be the technique of choice for its low cost and simplicity, being easily implemented in the pre-hospital setting. NIPSV may be equally effective but should be performed by experienced personnel and may be specially indicated in the most severe patients (hypercapnia or respiratory fatigue). Although a recent randomized trial comparing HFNC to conventional oxygen therapy in ACPO have demonstrated its feasibility in this setting (Makdee et al, 2017), further research is necessary to stablish the real indications of this technique.

NIV has been proven to be useful in patients with persistent weaning failure [132], showing shorter time with mechanical ventilation, shorter length of stay, lower incidence of complications (ventilator-acquired pneumonia or septic shock), and better survival, mainly in COPD patients [133,134,135,136,137,138,139,140]. On the other hand, because of the high rate of complications observed in patients that need reintubation (around 15% of the cases) [141], NIPSV has been used to prevent ARF after extubation. Some studies [134, 135] demonstrated that the use of NIV was more effective than standard therapy in preventing post-extubation ARF and reintubation in high-risk patients, but not in unselected patients [142], with a reduction in mortality in hypercapnic patients and chronic respiratory disorders [135, 136].

Another issue is the effect of NIV to treat the established ARF after extubation where no trial has reported benefits [143, 144]. A multicentre study found an even slightly higher mortality in the group of NIPSV, which was attributed to delayed reintubation (12 hours vs 2.5 hours). Although this trial included only 8–12% of COPD patients and could be biased by a high crossover rate, the authors concluded that NIPSV was not effective in averting the need for reintubation in unselected patients [144]. More recently, in the post-extubation period HFNC has shown to be superior to conventional oxygen therapy in low-risk patients and equivalent to NIPSV in high risk of reintubations [26, 27].

In conclusion, it seems reasonable to use NIPSV as an alternative to traditional weaning and in the prevention of post-extubation ARF in patients at risk, especially in those with COPD or hypercapnia, but should not generally be used in the treatment of post-extubation ARF, since a reintubation delay may be hazardous. HFNC is a reliable alternative in the prevention of ARF in this setting.

Before starting NIV, it is crucial to identify if the patient is a good candidate [166, 167]. This should be done considering the cause and the severity of ARF, which should be moderate to severe. There is a therapeutic window when NIV should be used, avoiding those patients with mild ARF that would easily respond to conventional oxygen therapy or conversely those who present very severe ARF needing intubation [59]. It is necessary to check contraindications for the technique (see Table 25.4) and consider predictors of failure (see Table 25.5) that warrant closer monitoring, paying attention to possible complications (see Table 25.1). Intubation may be preferred if the likelihood of NIV failure is very high. Subjects who have a pH <7.25, an APACHE II score >29, and a GCS <11 have failure rates ranging from 64% to 82% [58, 59, 168]. Patients with excessive secretions or without improvement after 60 min of NIV may be at high risk of failure [169,170,171,172]. Clinical signs that are only equivocal on presentation become more definitively predictive of failure if they persist after 2 hours of NIV [168].

In our experience, there are three levels that may influence NIV success (see Figure 25.4): the patient (severity, adaptation, mental status), the physician (concomitant therapy, experience, team attitude), and the device (incorrect adjustment, inadequate interface, excessive leakage).

Clear explanations and instructions should be given to the patients about the technique. Select the mask of appropriate size and shape to improve patient comfort, minimizing the risk of ulcers [173]. Prepare the straps, skin protections, and connections, and set the ventilator in order to minimize the time wasted in changing the mode. Hold the mask manually at the beginning, even having down the patient to do it by himself. Spend the initial time stimulating the respiration and teaching the patient the clues for synchronization.

The technique is very simple and does not require special training. However, it is important to use appropriate masks, adjusted with head straps, and to monitor, when possible, the resultant FiO₂. The initial CPAP may be 5 cmH₂O. If the patient is correctly adapted, the CPAP may be increased to 10 cmH₂O, the most common level of pressure applied in the majority of trials.

Out of sleep disorders and treatment of atelectasis, CPAP is mainly restricted to patients with ACPO. The improvement in these cases is mostly due to changes in the haemodynamic state and oxygenation, rather than the ventilatory effect of the technique. In the majority of cases, there is a progressive improvement in physiologic parameters and blood gases in minutes, and CPAP may be stopped when a negative balance has been obtained and the patients have no dyspnoea. CPAP does not improve outcomes in patients with hypoxaemic ARF from different aetiologies. When the response to CPAP is poor, a trial with NIPSV may be attempted in patients who show ventilatory impairment, reflected in CO₂, although there is no evidence to support this recommendation.

In critically ill patients HFNC is often started with a F_IO₂ of 100% and the maximum tolerated flow. Later, F_IO₂ and flow rate can be decreased according to SpO₂⁴¹ and patient’s demand. In less severe cases it is usually started with lower flow and F_IO₂.

This modality requires more experience. The selection or preferred use of specific interfaces, the main modalities, and the type of ventilators have been presented previously.

It is recommendable to start with low levels of pressure (IPAP 8–10 cmH₂O/EPAP 3–4 cmH₂O), increasing the pressure support progressively, according to patient adaptation, ensuring expired tidal volumes >4–6 mL/kg (it can be lower in COPD patients). If there is no control display of the expired volume, pressure support must be increased while tolerated. Normally, with a pressure support of 12–18 cmH₂O above PEEP, a tidal volume of 400–500 mL is reached. Elevated pressures may cause excessive air leakage, asynchrony (especially when the patient is tachypnoeic) and discomfort. On the other hand, a PEEP over 4 cmH₂O is necessary to avoid rebreathing when using portable ventilators, which may not include an expiratory valve or a double inspiratory–expiratory circuit [55, 168]. F_IO₂ should be titrated to achieve an SpO₂ >95%. Visualization of flow and pressure waveforms on display is essential. In a recent study, physicians obtained better results analysing the flow and pressure waveforms than just controlling numerical variables [174].

To ensure the success of NIV, close monitoring is necessary (see Table 25.6), especially patients’ oxygen saturation (to adjust FiO₂), PaCO₂ (to assess efficacy), and synchrony. Overall reassessments are usually performed at 60 and 90–120 min.

Pressure support ventilation unavoidably induces a certain degree of synchrony. In intubated patients, significant asynchrony has been found in 25% of the patients, especially at the beginning or after prolonged ventilation [175]. In patients with NIPSV, it may be nearly 50% [176, 177]. An asynchrony index (AI) >10% [AI (%) = number of events/(ineffective breaths + ventilator cycles) × 100] is considered as severe, leading to an increase in the work of breathing and patient discomfort [176]. Asynchrony is usually manifested under different forms that need specific approaches [174, 176, 177] (see Table 25.7 and Figure 25.5). Although several mechanisms may be responsible for asynchrony, air leakage is involved in many of them. In general, a leak of 0.4 L/s is well tolerated (<30 L/min) [178]. Higher levels of leakage may interfere with the patient’s efforts. In these cases, it is recommended to manually adjust the interface, reduce pressure support, and help the patient gain confidence with the technique.

Mild agitation and poor tolerance of the interface or the ventilator can provoke malfunction of, or patient refusal to, NIV. Although sedation can play a role in avoiding intolerance, it is also potentially dangerous because of the risk of oversedation [179]. In a large survey involving American and European physicians, the use of sedation in patients under NIV was infrequent and mainly determined by clinical experience [179]. The use of sedatives can improve tolerance and eventually avoid NIV failure. Morphine, remifentanil, dexmedetomidine, propofol, and midazolam-based regimens have been used with no serious complications in experienced units [177, 180, 181]. It is better the use of intermitent doses rather than continuous infusion [194]. The new α2 adrenoreceptor agonist dexmedetomidine showed better results than midazolam in patients with ACPO intolerant to NIV [180].

Sedation should only be considered when other maneuvers (reducing leakage, pressure support, etc) have failed.

NIV is usually stopped when a satisfactory recovery has been achieved or conversely there are signs of NIV failure. The identification of NIV failure has been presented previously in Adequate patient selection for non-invasive treatment section. Although there is some concern about the risk of delayed intubation [144], NIV is sustained while patients improve or at least maintain functional respiratory status. The approach would be different, depending on the cause and the duration of NIV. In mid or long-term use, a weaning period is often carried out by decreasing PEEP and ventilatory settings progressively. The application of a protocol-directed weaning has shown clear advantages in this context [182]. This approach does not seem to be necessary in short-term use. Typically, the interface will be removed, as requested by the patient, to provide facial hygiene or to administer oral medications. If the patient deteriorates when NIV is interrupted, the therapy is resumed, but otherwise NIV may be discontinued [168].

However, even in the best scenario and when applying NIPSV appropriately, a substantial number of patients will fail and need intubation. A prospective observational study reported a nearly 40% failure rate in a mixed population of patients with ARF [4, 168]. In patients with signs of NIV failure, endotracheal intubation is performed. Although some criteria for intubation may be proposed [174] (see Table 25.8), in everyday practice, clinicians typically decide to institute ventilation, based on their assessment of a patient’s signs and symptoms [89].

NIV developed in critical care areas and rapidly spread to ED, specialist areas, and finally to the ward [4, 5, 86, 183]. The main concern is that NIV needs device monitoring, expertise, and a higher rate of nursing staff than general wards. However, telemetry and some advanced alarm systems may facilitate its application. An attractive alternative is the control of peripheral NIV by medical emergency teams [184], who may be included in the organizational area of the ‘rapid response’ or ‘immediate vital support’ teams [185].

An optimal staffing and location for NIV therapy is needed. After connecting a patient to NIV, close attention from the therapeutic team is required. Several trials reported that patients with NIV may need more attention than intubated patients [186]. In our ICU and regarding workloads, NIV patients are considered as critical. When comparing costs, NIV seems to have a better cost–benefit relation than invasive ventilation [187]. Maximal attention is required within the first hours of starting NIV and especially in the first 20 min [174].

NIV plays an important role in the management of ARF in acute care areas (ICU, intensive cardiac care unit (ICCU), ED) and on general wards. It is indicated in patients with acute exacerbations of COPD and ACPO where it must be considered as a first-line treatment. NIV may be also indicated in difficult weaning, in ARF of immunocompromised patients, and in the prevention of post-extubation failure. Depending on the adequate selection of patients, it can be used in the post-operative period and in cases of pneumonia, asthma, or as a palliative treatment. Caregivers must develop skills in the technique, paying special attention to face mask fitting and patient–ventilator synchrony. Many other potential applications are undergoing further investigation, and additional studies are necessary to confirm all the benefits of NIV, extending its indications in the future.