Abstract and Introduction

Abstract

Objective: Although noninvasive positive pressure ventilation (NPPV) is a widely accepted treatment for some patients with acute respiratory failure, the use of NPPV in patients who have decided to forego endotracheal intubation is controversial. Therefore, the Society of Critical Care Medicine charged this Task Force with developing an approach for considering use of NPPV for patients who choose to forego endotracheal intubation.
Data Sources and Methods: The Task Force met in person once, by conference call twice, and wrote this document during six subsequent months. We reviewed English-language literature on NPPV for acute respiratory failure.
Synthesis and Overview: The use of NPPV for patients with acute respiratory failure can be classified into three categories: 1) NPPV as life support with no preset limitations on life-sustaining treatments, 2) NPPV as life support when patients and families have decided to forego endotracheal intubation, and 3) NPPV as a palliative measure when patients and families have chosen to forego all life support, receiving comfort measures only. For each category, we reviewed the rationale and evidence for NPPV, key points to communicate to patients and families, determinants of success and failure, appropriate healthcare settings, and alternative approaches if NPPV fails to achieve the original goals.
Conclusions: This Task Force suggests an approach to use of NPPV for patients and families who choose to forego endotracheal intubation. NPPV should be applied after careful discussion of the goals of care, with explicit parameters for success and failure, by experienced personnel, and in appropriate healthcare settings. Future studies are needed to evaluate the clinical outcomes of using NPPV for patients who choose to forego endotracheal intubation and to examine the perspectives of patients, families, and clinicians on use of NPPV in these contexts.

Introduction

Noninvasive positive pressure ventilation (NPPV) has become a standard therapy for the treatment of acute respiratory failure (ARF) in select populations and is increasingly being used in the critical and acute care setting.^[1] This increased utilization has been driven in large part by the desire to avoid complications of invasive mechanical ventilation, such as ventilator-induced lung injury, ventilator-associated pneumonia and other nosocomial infections, patient discomfort, and increased sedation needed with invasive mechanical ventilation. In addition, NPPV has been shown to reduce the rates of endotracheal intubation, mortality, and hospital length of stay in specific conditions, particularly hypercarbic respiratory failure due to exacerbations of chronic obstructive pulmonary disease (COPD)^[2-5] and hypoxic respiratory failure in immunocompromised hosts.^[6,7] In addition, NPPV is beneficial in many patients with cardiogenic pulmonary edema in the absence of acute coronary ischemia.^[6,8,9] Technical aspects of NPPV are beyond the scope of this review and have been discussed in detail elsewhere.^[1]

Although NPPV is widely accepted for patients with ARF who desire maximum life-prolonging treatment without preset limits on advanced life support measures, the emerging use of NPPV in patients who have decided to forego endotracheal intubation has raised controversy. Some authors have suggested that NPPV should be used in these patients on a trial basis, even for those with terminal diseases, to help alleviate respiratory distress and attempt to provide some additional time to finalize personal affairs.^[10] Others have suggested that it is inappropriate to use NPPV for patients who prefer to forgo life support in the form of endotracheal intubation and mechanical ventilation because NPPV is still a form of life support delivered by mask and may cause discomfort while prolonging the dying process.^[11] Clinicians themselves may be unclear about the goals of care when NPPV is used for patients and families who have decided to forego endotracheal intubation. This lack of clarity can lead to inadequate articulation of the rationale for NPPV; ambiguous dialogue between clinicians, patients, and families; and misleading recommendations. Adverse consequences may be serious, including unmet or conflicting expectations, avoidable adverse outcomes associated with NPPV, inappropriate use of medical resources, inadvertent prolongation of the dying process, and intensification of patient suffering and family duress.

Clinician-family communication in the intensive care unit (ICU) is an area in need of improvement.^[12-14] Specific concerns have been raised about whether patients and their families have an adequate discussion and clear understanding about the goals of care when NPPV is used in different circumstances.^[15] If clinicians are not certain about the goals of care for patients with ARF, this precludes clear discussion about the role of NPPV. In response to this potential problem, the Society of Critical Care Medicine formed a Task Force on the Palliative Use of NPPV to develop a framework for using NPPV in patients with ARF, especially for those patients who decline endotracheal intubation or who are receiving palliative care. This Task Force Report describes one approach to categorizing the use of NPPV, which uses an ethical framework based on explicating the goals of care and emphasizes the importance of informed patient preferences and communication among patients, family members, and clinicians. Herein, we review a) three potential goals of NPPV, b) the key points to communicate to patients and family members, c) the determinants of success and failure, d) suitable healthcare settings for use of NPPV, and e) appropriate alternative clinical measures if NPPV fails to achieve the goals of care. These issues are discussed within the context of a case scenario, modified to underscore the three different goals of care and roles of NPPV.

Proposed Three-category Approach

The use of NPPV for patients with ARF or acute on chronic respiratory failure can be conceptualized in three categories, as outlined in Table 1 and as described in more detail in the subsequent sections. Although the specific characteristics and nuances of every patient and family must be incorporated into good quality care, especially palliative care, this schema provides clinicians with a conceptual framework with which to approach NPPV in these different scenarios. Furthermore, although we present three distinct categories, it is important to acknowledge that individual patients may transition from one category to another as the goals of care or the risks and benefits of NPPV change.

Category 1: NPPV Without Preset Limits on the Provision of Advanced Life Support

Scenario 1. A 68-yr-old man with severe oxygen-dependent COPD presents to the emergency department with an acute exacerbation of his COPD. The clinicians evaluating the patient determine that before this exacerbation, his quality of life was acceptable to him, although he was only able to walk a half a block before stopping because of dyspnea. He has been admitted to the ICU for two previous COPD exacerbations, requiring endotracheal intubation and mechanical ventilation on both occasions. This time, he presents again with acute on chronic respiratory failure due to an upper respiratory tract infection, which requires ventilatory support. Arterial blood gas analysis shows a moderately severe (7.20<7.30) acute on chronic respiratory failure with hypercapnia. He is started on intravenous corticosteroids and antibiotics. Careful discussion with the patient and his partner reveals that the patient would want endotracheal intubation and invasive mechanical ventilation if NPPV fails, provided there is a reasonable chance of return to his baseline function. The patient is transferred to the ICU, and NPPV is started as the therapeutic ventilation strategy of choice.

Discussion. The first category for NPPV use is full life support without preset limits in which the primary goal is to restore the patient's health back to baseline. In this category, NPPV may be the preferred form of ventilatory support with advantages over endotracheal intubation and mechanical ventilation. Furthermore, in this scenario, the patient agrees to undergo endotracheal intubation and mechanical ventilation in the event that NPPV fails. Therefore, NPPV is considered successful if 1) it improves ventilation or oxygenation, 2) is reasonably well tolerated, and 3) provides support for the patient while the underlying cause of the ARF is treated and resolves. It is also important for the patient to be comfortable on NPPV, although the patient may be encouraged to tolerate some discomfort if the NPPV is improving ventilation or oxygenation and it seems that endotracheal intubation can be avoided by using NPPV. This mode of treatment is discontinued when the patient is able to support ventilation and oxygenation without assistance or when the patient is not responding to or tolerating NPPV, such that endotracheal intubation and invasive mechanical ventilation are indicated. Decreased level of consciousness is a relative contraindication to NPPV; however, there are occasions when NPPV is initiated in patients who have a decreased level of consciousness if 1) the potential benefit is perceived to outweigh the potential risk, 2) NPPV is being used in a monitored setting with a skilled, experienced clinical team and an adequate nurse-to-patient ratio, and 3) the patient's level of consciousness is closely monitored and seen to improve soon after NPPV is initiated.^[1,16] If the patient's level of consciousness does not improve sufficiently with the application of NPPV, this form of ventilatory support should be discontinued in favor of endotracheal intubation and invasive mechanical ventilation.

Category 2: NPPV for Patients Who Decline Endotracheal Intubation and Invasive Mechanical Ventilation

Scenario 2. A similar 68-yr-old man with severe oxygen-dependent COPD presents to the emergency department with an acute exacerbation of his COPD. In contrast to the patient in scenario 1, the clinicians caring for this patient determine that he does not want endotracheal intubation and invasive mechanical ventilation. He understands the potential consequences of this decision. He is willing to receive treatment with NPPV if he has a reversible cause of his COPD exacerbation, provided there is a reasonable chance that he can return home to his previous baseline health. His partner concurs with his decision. NPPV is started in the ICU with the understanding that if it is not successful or the patient opts not to continue on NPPV, it will be discontinued and comfort measures will be intensified.

Discussion. In the second category, the patient is willing to undergo NPPV but not endotracheal intubation and invasive mechanical ventilation because of the additional burden and risks that invasive mechanical ventilation entails, which, according to his values and preferences, outweigh the potential benefits. In this scenario, NPPV is being used as a form of ventilatory support. Therefore, NPPV is considered successful if it improves ventilation or oxygenation and provides support for the patient while the underlying cause of the respiratory failure is treated. It is important that this patient's dyspnea is well managed so that he is comfortable on NPPV; however, this patient may also be encouraged to tolerate some discomfort if the NPPV is improving ventilation or oxygenation. NPPV is discontinued when the patient is able to support ventilation and oxygenation without it, when NPPV is determined not to be working, or when the patient is not tolerating NPPV. In the latter two circumstances, the healthcare team, patient, and family decide together that NPPV is not producing the desired response and that the burdens outweigh the benefits; it is at that point that the patient's preference for comfort measures without life-sustaining therapies is called forth.

Category 3: NPPV as a Comfort Measure for Patients Who Decline Endotracheal Intubation

Scenario 3. A similar 68-yr-old man with severe oxygen-dependent COPD presents with an acute exacerbation of his COPD. Careful discussion with him reveals that this patient has a baseline quality of life that he finds unacceptable; before this exacerbation, he was unable to walk across his living room without stopping because of dyspnea. He reports being very distressed at the slow progressive decline in his functional capacity and his severe, limiting symptoms, despite maximal outpatient therapy. The patient wishes to be kept comfortable. His preferences are clear; he does not want any form of life-prolonging therapy. He and his partner understand the consequences of this decision. His partner is sad at the prospect of losing her husband but wants to ensure that his treatment preferences are honored. The patient has been on NPPV in the past, and she asks whether NPPV might be attempted to help relieve his breathlessness so that he and his partner can communicate with one another.

Discussion. Patients in the third category, especially those with a terminal disease (such as an underlying malignancy) or at the final stage of a chronically progressive disease (such as COPD, chronic heart failure, or neuromuscular disease), may decide that they do not wish life-prolonging therapy and choose comfort measures only. In this setting, it is conceivable that NPPV might be used as a form of palliative care, to attempt to reduce the dyspnea. Furthermore, although most patients and families are interested in ensuring comfort while dying, some are also interested in maintaining cognition and the ability to communicate.^[17] NPPV in this circumstance would only be considered successful if it improves the patient's symptoms of dyspnea or other distress without causing other troubling consequences. Patients in this category should not be encouraged to tolerate discomfort associated with NPPV because the primary goal of the therapy is patient comfort. NPPV should be discontinued if patients feel that the NPPV is not making them more comfortable. Furthermore, when used in this context, there is no justification for providing NPPV to patients who are unable to communicate about whether NPPV has improved their symptoms.

It is important to acknowledge that some patients or family members who choose comfort measures only may still maintain a desire or hope for a cure or a miracle and that this may reflect ambivalence toward life-sustaining treatments. Discussing such ambivalence can provide an opportunity to explore patients' and families' concerns and fears about end-of-life care, but the existence of this ambivalence does not necessarily require life-sustaining therapies that are not indicated or for which the burden outweighs the benefit.^[18] However, a patient who has chosen not to undergo endotracheal intubation and is using NPPV, hoping it will even briefly prolong life, should be considered as category 2 rather than 3. It is also important that the goals of care be continually reassessed, as risks, benefits, and treatment preferences can change over time, and therefore, the goals of care and the category of NPPV may change over time.

Some patients who do not want prolonged life-sustaining therapy may choose to undergo a time-limited trial of NPPV to achieve the goal of surviving until the arrival of family members or friends traveling to visit the patient. Although these patients may have a terminal illness and may be actively dying, this use of NPPV would be an example of a category 2 rationale. In this situation, NPPV would be used to provide life support, delaying death until the family member arrives to say goodbye, providing that the benefits of NPPV outweigh the burdens from the patient's perspective.

Evidence Supporting the Use of NPPV in Each Category

Clinical research evidence supporting the use of NPPV for patients in each of these categories is variable. Table 2 provides a narrative overview of data supporting the use of NPPV in the three categories. Table 3 outlines details on the number of studies, sample sizes, and specific clinical outcomes for three conditions about which research evidence is strongest. Evidence for each category of NPPV use is also described below.

Evidence Supporting NPPV in Category 1

Most of the randomized trials to date on use of NPPV have been conducted on patients in the first category: when NPPV is used as life support and the patient would undergo endotracheal intubation and mechanical ventilation if NPPV failed to support the ventilatory or oxygenation needs of the patient. These trials provide compelling evidence of the benefit of NPPV for several indications, including ARF in the setting of a COPD exacerbation,^[2-5] hypoxic respiratory failure in the setting of an immunocompromised state,^[6,7] and ARF in the setting of cardiogenic pulmonary edema.^[6,8,9,19]

Although the evidence is more limited, NPPV has also been used for carefully selected patients with acute exacerbations of asthma or cystic fibrosis, some postoperative patients with ARF, and to facilitate weaning and avoid extubation failure. Less commonly and more controversially, use of NPPV has been reported in carefully selected patients with community-acquired pneumonia or acute respiratory distress syndrome. For patients with neuromuscular disease, two small observational studies are available, one suggesting that NPPV may reduce intubation and mortality rates compared with historical controls^[20] and the other supporting the use of NPPV to avoid intubation in myasthenic crises.^[21] However, the evidence base for use of NPPV for these diseases is still being established; the need to avoid causing harm and to carefully select and monitor all acutely ill patients treated with NPPV cannot be overemphasized.

Evidence Supporting NPPV in Category 2

Three randomized trials conducted in the UK enrolled patients with acute exacerbations of COPD and did not explicitly exclude patients who declined endotracheal intubation and invasive mechanical ventilation.^[22-25] The largest of these trials (n = 236) showed a reduced rate of endotracheal intubation and mortality in patients with moderately severe exacerbations of COPD (which also included category 1 patients).^[22] However, these studies did not conduct a priori or post hoc subgroup analyses specifically on patients in category 2. Therefore, it is not possible to draw conclusions from these trials about the benefits of NPPV for patients in category 2.

Observational studies have examined the potential benefit of NPPV in patients with ARF who have do not intubate (DNI) orders. In a single-center study of 17 patients who were DNI because endotracheal intubation was contraindicated, NPPV was initially successful in 59% of patients (10 of 17), with a hospital mortality of 47% (9 of 17 patients).^[26] In a multiple-center observational study of 114 consecutive DNI patients with ARF treated with NPPV, 43% survived to hospital discharge, with approximately 50% of patients with COPD and 70% of patients with cardiogenic pulmonary edema surviving to hospital discharge.^[27] A higher baseline Paco₂, diagnosis of cardiogenic pulmonary edema or COPD, strong cough, and wakefulness were associated with greater hospital survival. A recent single-center prospective observational study of the use of NPPV in 131 consecutive DNI patients with ARF found a similar overall hospital mortality of 65%.^[28] Patients with advanced cancer had a much poorer prognosis, with a hospital mortality of 85% (p = .002). Of those in the nononcologic subgroups, 63% of patients with COPD and 60% of patients with cardiogenic pulmonary edema were discharged home. It is unlikely that these predictors are specific to patients in category 2. Furthermore, in none of these latter two studies were patients in category 2 (NPPV as ventilatory support in the patient with a DNI order) distinguishable from those in category 3 (NPPV as a comfort measure only); however, it is likely that the majority of these patients were in category 2 because most were discharged alive.

Long-term follow-up for 80 consecutive patients presenting with an exacerbation of COPD and treated with NPPV was evaluated in one study; 37 COPD patients who had a DNI order were compared with 43 patients without DNI orders, and the 1-yr survival was 30% compared with 65%, respectively.^[29] Patients in the DNI group were significantly older, had worse dyspnea and APACHE II scores, and were more limited in their activities of daily living. These results may be useful when discussing the longer-term prognosis of a COPD exacerbation treated with NPPV among individuals electing to forego endotracheal intubation and invasive mechanical ventilation.

Evidence Supporting NPPV in Category 3

To date, no studies have fully assessed the effectiveness and safety of NPPV specifically in patients who are in category 3, even though NPPV may reduce symptoms of dyspnea in a broader group of patients with ARF from COPD,^[23,24] in patients with advanced cancer and respiratory failure,^[30] and in selected patients with stable chronic respiratory failure from COPD.^[31] However, further research is needed to determine whether patients in category 3 benefit from NPPV. It is important to highlight that not only do we have inadequate clinical research on benefit in this unique category 3 setting compared with standard therapy without NPPV, but also on the potential harm. For example, many dying patients might find NPPV undesirable because the potential for relief of dyspnea is outweighed by discomfort due to the need for a tight-fitting face mask. The potential theoretical advantages could also be countered by a stressful fixation on technology at the end of life and limited ability to communicate imposed by the face mask. Moreover, there are no studies comparing NPPV in this setting with pharmacologic therapies such as morphine. Further research is needed before this approach can be considered a standard of care.

Nonetheless, it is theoretically possible that NPPV could be used in the setting of comfort measures only as an adjunct to opiates and nonpharmacologic therapies in a way that might minimize some side effects of opiates, such as decreased level of consciousness. However, it is important to emphasize that there is no convincing evidence that NPPV will diminish the burden of symptoms in patients who choose comfort measures only. For occasional patients, however, the relief of dyspnea and the improved level of consciousness might allow for more effective communication, especially during periods when NPPV is temporarily stopped. Hence, NPPV may offer symptom palliation for patients who are in this category and offer families an opportunity to communicate with patients who are close to the end of life. Importantly, the patient can and must maintain control over the decision to continue support. If discomfort exceeds benefit, the patient can simply choose to discontinue NPPV, and comfort should be achieved quickly with pharmacologic therapies. This may require anticipatory dosing of opioids to achieve patient comfort, as is sometimes necessary during withdrawal of invasive mechanical ventilation.^[32] If the patient becomes unable to communicate, the benefits of NPPV in this setting will likely have ceased, and therefore, NPPV should be stopped.

Communication With the Patient and Family

For many reasons, talking about end-of-life care can be a source of stress for physicians and other clinicians.^[33] Discussing the pros and cons of NPPV administration with patients in categories 2 and 3 and their families is no exception. A recent International Consensus Conference for Critical Care identified shared decision-making between the patient (when possible), the family, and the clinicians as the preferred decision-making model for decisions about life-sustaining treatments.^[34,35] A shared approach exchanging clinical information and patient preferences may be especially important for decisions about the use of devices unfamiliar to patients and families, such as NPPV. In this model, physicians would provide a clear, understandable description of NPPV, including its potential risks and benefits, and they would also outline the goals of care associated with each of the three categories of NPPV use that are relevant for the given patient. Physicians would elicit from patients or families the patient's values and preferences, to understand, through dialogue, the treatment goals that are most consistent with the patient's values.^[36] Physicians would then summarize these points and outline the agreed-on plan of care, should NPPV succeed or fail in meeting the treatment goals. Table 4 provides some sample language that may be helpful to clinicians when summarizing why and how NPPV can be used. This table uses the example of communicating with the family, although if the patient is able to communicate, it is important to have these discussions with the patient.

For patients in all categories, clinicians need to continually reassess whether NPPV is succeeding or failing to meet the goals of care. If NPPV is failing, the task of communicating the next step in care is especially important. For patients in category 2, this will involve telling families that the patient is likely going to die of his disease and that NPPV has either failed to improve the patient's physiologic variables or that those improvements have been achieved at too great a cost of discomfort to the patient. When NPPV fails to alleviate suffering in category 3 patients, physicians should clearly communicate that NPPV is not working, and other treatments to achieve palliation of symptoms should be initiated or intensified.

Implications for Location of Therapy

The three-category approach to the use of NPPV has implications for the location of therapy. In many hospitals, the use of NPPV for ARF occurs only in a critical care unit (which we define here as an ICU or intermediate or step-down unit) because, if NPPV were to fail, an immediate response is facilitated in the critical care unit. National or regional differences in the organization and availability of critical care or the subspecialty of those physicians who first used NPPV in a region may explain some of the regional differences in hospital location for the use of NPPV.

A potential advantage of an explicit three-category approach to the use of NPPV for ARF is that it may help hospitals to develop policies that permit NPPV for ARF outside the critical care unit for either category 2 or category 3 patients because the response to failure of NPPV in these patients does not require rapid institution of more invasive life support. Instead, for category 2 or 3 patients, failure of NPPV often requires rapid intensification of comfort measures, which can often be addressed outside the critical care unit. For many patients in category 1, NPPV should be provided in the critical care unit. However, not all patients in category 1 require continuous NPPV, and as has been demonstrated in several studies, NPPV can be safely provided on some general medical-surgical wards with well-trained and experienced staff.^{[23,25,37-41]} The patients for whom this is most feasible are those who can easily sustain spontaneous ventilation independent of NPPV for some minimal period of time, perhaps 1 hr, depending on the monitoring capabilities of the ward. When NPPV is provided outside the critical care unit, the patient should be located close to the nursing station, and the ventilator should be appropriately configured to allow alarm notification to be heard immediately by clinicians. The ventilator alarms must include high-pressure and loss-of-pressure or disconnect alarms.

In many hospitals, patients in categories 2 and 3 are ideally cared for outside the ICU. The same NPPV monitoring and alarm standards as for category 1 patients should also be in place to ensure that patients are comfortable and to quickly identify patients for whom NPPV is no longer helpful so that other measures can be taken to ensure comfort. Such monitoring should include frequent direct observation of the patient, especially in the initial phase of NPPV therapy.

NPPV might also be considered for category 3 patients in the hospice setting to help palliate dyspnea while attempting to minimize adverse effects of opiates or when, in special cases, mechanical ventilation is being discontinued in the hospice setting. It is important to note that there are no data supporting this use of NPPV. Furthermore, in a setting in which clinicians are unfamiliar with the use of NPPV, the chance of benefit is low, and there is a distinct risk of increasing discomfort and stress at the end of life. If staff and family are appropriately trained and otherwise resourced, an inpatient or home hospice setting could be appropriate for some category 3 patients.

Conclusions and Recommendations for Clinicians

This Society of Critical Care Medicine Task Force suggests an approach to the use of NPPV for patients and families who have chosen to forego endotracheal intubation and mechanical ventilation. To clarify the goals of care, we recommend that clinicians consider which of these three categories apply to specific patients with ARF. These three categories underscore the need to elicit patient preferences and goals, communicate the rationale for NPPV, outline parameters for success and failure, and apply NPPV in correspondingly appropriate settings. Future studies are needed to identify the prevalence of the three approaches for use of NPPV and evaluate the outcomes associated with using NPPV, especially in the two categories of patients who have chosen to forego endotracheal intubation. This research should examine the perspectives of patients, families, and clinicians on issues such as symptoms, communication, decision making, satisfaction, other benefits, and harms. Research should also examine the perspective of the healthcare system on issues such as safety, life support resource utilization, and quality of end-of-life care. Such research is needed to help inform practice regarding the application of this emerging technology.