In any discussion about location of a noninvasive ventilation (NIV) service, it is important to note that the model of hospital care differs between countries and that there may be significant differences even between hospitals within the same country. There will be variations in staffing levels; the skills of doctors, nurses, and paramedical staff; and the sophistication of monitoring. The terms intensive care unit (ICU), high-dependency unit (HDU), and general ward have a different meaning to different people. Care therefore must be taken when extrapolating experience and results obtained in one environment to other hospitals and countries.

The United Kingdom’s King’s Fund panel1 defines intensive care as “a service for patients with potentially recoverable diseases who can benefit from more detailed observation and treatment than is generally available in the standard wards and departments.” The definition of HDU is less clear, with some HDUs allowing invasive monitoring, whereas in others only noninvasive monitoring is performed. In some countries, specific respiratory ICUs and intermediate ICUs have been developed.2,3 Specifically, within the King’s Fund definition is the consideration of intensive care as a service rather than a place; critical care is provided within a continuum of primary, secondary, and tertiary care, and patients are categorized on the basis of their needs4 (Table 34-1). Movement through the different levels usually means transfer from one location to another. Critical care outreach teams can advise on care as patients cross organizational boundaries and also facilitate transfer when this is needed.5,6 Although most acute NIV services are situated in a specific clinical area, a peripatetic model has been described and has some advantages.7,8

For the purposes of this chapter, the following definitions are used:

• Intensive care. High ratio of staff to patients, facility for invasive ventilation and sophisticated monitoring.
  
• Intermediate respiratory ICU or HDU. Continuous monitoring of vital signs, with a staffing ratio intermediate between an ICU and a general ward, in a specified clinical area. Intubated patients (unless with tracheostomy) usually are not cared for in this environment.
  
• General ward. Takes unselected emergency admission, and although most wards will have a particular speciality interest, it is likely that because of the unpredictability of demand, patients with a variety of conditions and degrees of severity will be cared for in the same clinical area. Nurse staffing levels vary, but the intensity of nursing input available in HDUs and ICUs is not possible. Only basic monitoring is available.

Another important issue when considering NIV in different locations is the severity and acuteness of the insult leading to ventilatory failure. Ventilatory failure can be considered acute when it occurs on a background of normal function, acute-on-chronic when there is a sudden deterioration on a background of impaired function, or chronic when there is ventilatory failure but with no precipitating acute event. Assisted ventilation can be considered necessary when without it death will ensue over a few hours or desirable when the primary aim is to improve quality of life and also to improve survival over the longer term.

For additional information about noninvasive ventilation, see Chapter 18.

Acute Noninvasive Ventilation

Chronic Obstructive Pulmonary Disease

NIV first became established as a viable technique for patients with acute respiratory failure secondary to an exacerbation of chronic obstructive pulmonary disease (COPD) in the ICU. The most striking finding from the early randomized, controlled trials (RCTs) comparing NIV with conventional therapy was a reduction in the need for intubation,9,10 which in the largest study translated into improved survival and reduced length of both ICU and hospital stays.9 Complications, particularly pneumonia and other infectious complications, were reduced markedly.9,11–15 It is striking that NIV was administered for only a relatively small proportion (mean: 6 hours) of each day9 or at modest levels for a longer period.10 With NIV, paralysis and sedation are not needed, and ventilation outside the ICU is an option. Given the considerable pressure on ICU beds in some countries, the high costs, and that for some patients admission to ICU is a distressing experience,16 this is an attractive option.

There have been seven prospective, randomized, controlled studies of NIV outside the ICU either on general wards, in an intermediate unit, or in the emergency department.17–23 A more rapid improvement in abnormal physiology is a consistent finding, but it was only in the largest,22 adequately powered, study that a benefit in terms of outcome was seen. Plant et al22 recruited 236 patients with an acute exacerbation of COPD, who were still hypercapnic, with a pH less than 7.35, and respiratory rate greater than 23 breaths/min on arrival on the ward. A proportion of patients will improve just with medical therapy. In a 1-year-period prevalence study24 of patients with acute exacerbations of COPD, 20% of 954 patients were acidotic on arrival in the emergency department; of these, 25% had completely corrected their pH by the time of arrival on the ward. There was a weak relationship between partial pressure of arterial oxygen (PaO2) on arrival at hospital and the presence of acidosis, suggesting that, in at least some patients, respiratory acidosis had been precipitated by high-flow oxygen therapy administered on the way to hospital.

The study was performed on general respiratory wards in thirteen centers. NIV was applied, by the usual ward staff, using a bilevel device in spontaneous mode according to a simple protocol. “Treatment failure,” a surrogate for the need for intubation, defined by a priori criteria, was reduced from 27% to 15% by NIV. In-hospital mortality was reduced from 20% to 10%. This study suggests that with adequate staff training, NIV can be applied with benefit outside the ICU by the usual ward staff and that early introduction of NIV in a general ward results in better outcomes than providing no ventilator support for acidotic patients outside the ICU. A recent national audit in the United Kingdom,25 however, raised significant concerns about the provision of NIV in the “real” world. Although it was not recorded in the audit, it is likely that the majority of patients received NIV outside of ICUs, mostly on general wards. Two hundred and thirty-two hospital units collected data on 9716 patients of whom 1077 received NIV. Of concern 30% of patients with persisting respiratory acidosis did not receive NIV. The mortality was higher in all acidotic groups receiving NIV than in those treated without. Patients who had late-onset acidosis had a particularly poor prognosis confirming the results of an earlier case series.26 Interestingly, 11% of acidotic admissions had a pure metabolic acidosis. There is a challenge in translating the results of RCTs into everyday clinical practice, especially when the particular technique involves significant technical expertise. It reinforces the need for ongoing audit to ensure that standards are maintained.

Conditions Other Than Chronic Obstructive Pulmonary Disease

Trials in acute exacerbations of COPD provide the biggest body of evidence on NIV. NIV, however, is also used in other conditions, often on the basis of what has been learned in COPD.

Hypoxemic and Hypercapnic Respiratory Failure.

Obesity.

Obese patients may present with acute or acute-on-chronic respiratory failure. In numerical terms this patient group is increasing; the number of patients requiring home ventilation because of obesity-hypoventilation syndrome is increasing year on year, and in one study, patients with obesity now comprise the largest single group.27 For obese patients requiring ventilator support acutely, the outcome from invasive ventilation is generally poor.28 There are major practical problems associated with nursing critically ill obese patients, often requiring many pairs of hands and specialized lifting equipment for basic tasks. There are no RCTs of the use of NIV in patients with ventilatory failure secondary to obesity. A case series in which patients who received NIV were compared with those who refused it showed a survival advantage for those receiving NIV (97% vs. 42%)29; this was not controlled and there may have been other reasons for the difference. Very obese patients may have upper airway obstruction during sleep and because the impedance to inflation may be very high may require different ventilator modes.30

Neuromuscular Disease and Chest Wall Deformity.

Patients with acute respiratory failure secondary to neuromuscular disease and chest wall deformity are not widely studied because they are small in number. Because of markedly reduced respiratory reserve, however, these patients are often challenging to wean from invasive ventilation and endotracheal intubation is best avoided if possible. Ideally, at-risk patients should already be under follow-up in a specialist unit and have been warned of the symptoms of evolving respiratory failure and of the necessity to present to hospital early in case of changes in their condition. Some patients will already have experienced a trial of domiciliary NIV. As such they represent good candidates for NIV outside the ICU. In addition to staff skilled in the delivery of NIV, therapists with expertise in secretion clearance techniques, including the use of mechanical insufflators or exsufflators,31,32 are vital in the management of these patients.

Cardiogenic Pulmonary Edema.

Cardiogenic pulmonary edema (CPE) represents a special case because the onset and recovery are usually both rapid. Most patients present to the emergency room, but some develop CPE in the ward. There have been seven systematic reviews (meta-analyses) on noninvasive ventilator assistance in CPE published since 2005.33–40 Overall, there was a significant reduction in mortality for those patients treated with continuous positive airway pressure (CPAP) and a trend toward improved survival with NIV.34 Both CPAP and NIV showed benefit when intubation was an outcome. There was no difference in any outcome when CPAP and NIV were compared. There was a trend toward an increase in myocardial infarction rate with NIV, but this was largely caused by the weighting of one study.41 Two recent trials may result in the reappraisal of the role of NIV in acute CPE.42,43

In the 3CPO trial,42 a multicenter, open, prospective RCT, patients were randomized to standard oxygen therapy, CPAP, or bilevel ventilation. There was no difference between 7-day mortality for standard oxygen therapy (9.8%) and NIV (CPAP and bilevel ventilation, 9.5%; P = 0.87). The combined end point of 7-day death or intubation rate was similar irrespective of NIV modality (11.7% vs. 11.1%, CPAP vs. bilevel ventilation respectively; P = 0.81). In comparison to standard oxygen therapy, NIV was associated with greater reductions in breathlessness scores, heart rate, acidosis, and hypercapnia at 1 hour. There were no treatment-related adverse events. There were no differences in other secondary outcomes, such as myocardial infarction rate, intubation, length of hospital stay, or ICU admission rate.

In another trial,43 120 patients were enrolled in three French emergency departments to either CPAP or NIV. There was no difference between interventions for any outcome. Respiratory distress and physiology improved in both arms. Only 3% of patients required intubation and one died within the first 24 hours.

These outcomes are different from the outcomes in the above meta-analyses, despite similar improvements in physiologic and gas exchange variables. The 3CPO trial was adequately powered and recruited more patients than the total of all the studies included in the meta-analyses. The discrepancy between results from one large, multicenter RCT and previous pooled data are not unique and the limitations of meta-analysis are well known.44 Individual trials were composed of small treatment group sizes that varied between nine and sixty-five patients with recruitment rates of only 10% to 30% (compared to 62% randomized in the 3CPO trial). In the meta-analyses, the small total number of outcome events was well below the recommended threshold of 200,45 limiting the generalizability of the findings.

The 3CPO trial may have failed to reveal a difference because the intervention was ineffectively delivered. Mean pressures for both CPAP (10 cm H2O) and noninvasive positive pressure ventilation (IPAP 14/EPAP 7 cm H2O) are comparable with previous studies, and improvements in physiologic variables are similar. There was crossover between interventions in all three arms of the 3CPO trial and these were analyzed on an intention-to-treat basis. There were differing reasons with respiratory distress and hypoxia being more likely in the control arm and lack of patient tolerance in the two intervention arms. After these patients were removed from primary outcome analysis, there remained no significant difference between groups, although mortality rates were lower.

Previous trials have indicated that the physiologic improvement seen with NIV is translated into a reduction in tracheal intubation rates.33,34 In contrast, the 3CPO trial found no benefit in reducing intubation rates by NIV. Reasons for this are unclear but may reflect the differing patient populations, concomitant therapies, and thresholds for intubation and mechanical ventilation. Intubations rates in the standard therapy arms vary from 35% to 65% in early trials to 5% to 7% for recent trials in emergency department settings, despite similar severity of illness. Intubation rate in the intervention arms have fallen considerably over time, with some initial trials reporting intubation rates of up to 35% whereas recent reports have consistently suggested rates of around 5%. The recent trial43 from France reported a 3% intubation rate, almost identical to that in the 3CPO trial. It is difficult to make direct comparisons because studies differ in the time at which mortality is recorded, but, if anything, survival has probably improved as intubation rates have fallen, suggesting better overall management.

One danger of NIV is that other aspects of medical therapy may be forgotten because the focus is on the application of NIV. Nitrates are key and the total dose delivered has been shown to be an important predictor of outcome.46,47 Positive pressure is beneficial to the failing heart and has some similarities to the effects of nitrates (preload and afterload reduction); if medical management is suboptimal, ventilation will have a beneficial effect on the failing heart, which may be lost if these effects have already been achieved with medication.

Finally, the patients recruited may have been less unwell than those in other studies. There was no difference in survival between recruited and nonrecruited patients, and no interaction with disease severity making this unlikely. The physiologic disturbance in these patients put them at the sickest end of the spectrum of patients studied, and, indeed, in contrast to other studies, acidosis (mean pH: 7.22) and hypercapnia (mean partial pressure of arterial carbon dioxide [PaCO2]: 7.6 kPa) were invariable.

Despite these negative findings, a reduction in dyspnea, which was very intense, was a striking feature in patients receiving ventilator support, and this alone is sufficient reason to utilize ventilation in CPE. There is a trade-off between the beneficial effects of this reduction in dyspnea against discomfort from the mask and other factors.

Hypoxemic Respiratory Failure.

There are no RCTs of NIV outside the ICU in hypoxemic respiratory failure. An RCT in the ICU12 showed that patients receiving NIV had significantly lower rates of serious complications, and those treated successfully with NIV had shorter ICU stays. Post hoc analysis of patients grouped according to the Simplified Acute Physiology Score (SAPS) showed that NIV was superior to conventional mechanical ventilation in patients with a SAPS less than 16. In patients with a SAPS equal to or greater than 16, outcome was similar irrespective of the type of ventilation. Another study,48 in immunocompromised patients, introduced NIV at a much lower level of physiologic compromise than would be required for invasive ventilation, and the sequential strategy (predefined periods on and off NIV) suggests that these patients could manage periods of spontaneous breathing safely. Further data are needed but it is reasonable for selected patients to have a trial of NIV in an experienced noninvasive unit outside the ICU; rapid access to intubation and mechanical ventilation must be available.

Elective Ventilation for Chronic Ventilatory Failure

This subject is dealt with in more detail in Chapters 28 and 33. In summary, there is no prospective RCT evidence to support the chronic use of NIV in any patient group. Most practitioners, however, would consider it unethical not to offer NIV to patients with chest wall deformity and neuromuscular disease, and it is unlikely that there will ever be any RCTs of NIV in these conditions. Chronic NIV is not appropriate for most patients with COPD; RCTs are ongoing.

Acute Noninvasive Ventilation: ICU or General Ward?

There have been no direct comparisons of outcome with NIV delivered in the ICU, in intermediate units, and in a general ward. It should be appreciated that while there is some overlap, the skills needed for NIV are different from those required for invasive ventilation. Familiarity with and confidence in NIV by all members of the multidisciplinary team is the most important factor. Nurses, physiotherapists, or respiratory therapists may be the primary caregiver; this will depend on local availability, enthusiasm, and expertise. The outcome from NIV is likely to be better on a general ward where the staff has a lot of experience of NIV than in an ICU with high nurse-to-patient, therapist-to-patient, and doctor-to-patient ratios, and a high level of monitoring, but little experience of NIV. The less intensive atmosphere of a noninvasive unit may not be as distressing for patients and their relatives. NIV may be quite time-consuming in the early stages, and patients may benefit from extra attention, more likely in an ICU compared with less-well-staffed areas. Staffing is usually less at night, but a study in an ICU revealed no difference between patients failing NIV during the day and the night49; because of the greater number of patients under the care of an individual nurse, this may not be true on a general ward and should be evaluated further.

Assuming that the skills to deliver NIV are equal in the various possible locations, there are a number of other factors to be considered. These include whether or not intubation is considered appropriate should NIV fail, the presence of other system failure, comorbidity, severity of the respiratory failure, and likelihood of success with NIV (Fig. 34-1). Patients who cannot sustain ventilation for more than a few minutes when acutely unwell require continuous observation. This level of support is more likely to be available in the ICU than in other ward environments.