In defence of ward-based non-invasive ventilation

The National Confidential Enquiry into Patient Outcome and Death (NCEPOD) have just published a report reviewing acute non-invasive ventilation (NIV) in the UK. ¹ Historically, NIV has had an unusual place on the fringe of critical care medicine. Acute mechanical ventilation is and has always been a standard intensive care therapy. However, with intensive care unit (ICU) beds scarce in the late 1990s, a number of respiratory physicians recognised that their chronic obstructive pulmonary disease (COPD) patients could not easily access such respiratory support. As a response they experimented with NIV outside of the ICUs. The “Yoniv” study ² published in 2000 showed that NIV could safely be delivered on general respiratory wards, albeit with carefully selected patients and under the care of enthusiasts in this technique. This encouraged other clinicians to set up ward-based NIV services, and over the next 15 years, such services and the demand for them grew explosively. By 2014, the vast majority of hospitals admitting patients with acute exacerbations of COPD had the ability to provide NIV outside of the critical care unit, most often on respiratory wards and/or medical assessment units. ³

In many ways, ward-based NIV is a success, widening access to a potentially lifesaving treatment. It is however somewhat of a victim of its own success. Increasingly, COPD patients are surviving for longer with established type 2 respiratory failure as NIV allows them to survive acute episodes that would have previously been fatal. Judging by the pH, many patients currently treated with NIV are substantially sicker than was the norm in the Yoniv study. ⁴ In my own organisation, of the patients treated with acute NIV, nearly 25% have one or more further episodes requiring NIV in the subsequent year (unpublished data). The complexity of the patients treated with ward NIV has increased over the years as services expand from “simple” COPD to obesity hypoventilation to neuromuscular disorders and kyphoscoliosis.

No one in the critical care community would dispute that ventilation is a multifaceted therapy which requires specialist clinical knowledge and a decent working understanding of some of the finer points of respiratory physiology to do well. Medical cover out of hours for ward-based services is of necessity usually provided by junior- and middle-grade doctors from a variety of general medical backgrounds whose particular interests lie elsewhere. Without specialist gatekeeping, the risk is that NIV gets used inappropriately, for example in those with type 1 respiratory failure or metabolic acidosis or in patients who are unequivocally reaching the end of their lives and would be better served by palliation.

An unexpected challenge for NCEPOD was that simply identifying NIV patients was less than straightforward. The procedure codes used nationally do not differentiate between NIV and Continuous Positive Airways Pressure (CPAP), clearly a different treatment used for different indications. It is reasonable to speculate that this coding issue is also hampering local audits and service reviews. NCEPOD found several problems with the delivery of NIV by teams on the ground. Overall, NCEPOD reviewers felt that there was room for improvement in the clinical management of patients in over half of cases. Over oxygenation, inadequate ventilator management (particularly insufficient inspiratory pressures) and poor patient selection were some of the emerging themes. Mortality was high, 25% in patients with COPD (compared to 10% in the treatment arm of Plant et al. ² ) and up to 60% in those with “other” indications perhaps reflecting the inappropriateness of NIV for these “other” “indications”.

Forty percent of patients treated with NIV had evidence of consolidation on the chest X-ray. National NIV guidelines caution against the use of NIV in pneumonia. ⁵ Whilst these guidelines were published just after the NCEPOD study period, it would appear that they represent a substantial change from what was contemporary practice. Whilst NIV is a poor choice in pure pneumonia without significant underlying lung disease, it is difficult sometimes to know the best primary label to give to the patient with severe COPD who exacerbates and presents febrile with minor pneumonic chest radiograph changes. This group may be said to have evidence of pneumonia but NIV may be a reasonable treatment option.

Nearly half of the patients were referred to critical care at some point in their hospital stay and of those around 70% were accepted for admission. Half those referred to critical care were considered frail (Rockwood score of 6 or more), though this proportion was higher in the group not referred. The long-standing guideline recommendation to initiate NIV in an High Dependency Unit (HDU) or ICU environment in patients with a pH of 7.25 or below does not seem to be common practice; less than half of this group were referred to critical care.

NCEPOD, in their recommendations, suggest acute NIV is only done in areas equipped with oximetry and continuous electrocardiogram (ECG) monitoring and staffed, at a minimum, with a nurse:patient ratio of 1:2. This is effectively specifying that acute NIV is only done in an HDU or higher level environment. This is very much in keeping with the concept of ventilation as a complex specialist intervention, also promoted in the British Thoracic Society (BTS) guidelines. ⁵ Given the numbers receiving ward-based NIV, this is not going to be possible to implement immediately in many hospitals. Is it the ideal we should be striving for, however? Current guidelines suggest continuous ECG monitoring for patients with “pulse rate >120 bpm or if there is dysrhythmia or possible cardiomyopathy” as a good practice point and is no doubt sensible in the patient appropriate for escalation. The risk in patients at their ceiling of care is that such monitoring will increase the burden of treatment without any corresponding beneficial gain.

What should the critical care community be doing in response to this report? Firstly, we should clearly acknowledge that these patients are not someone else’s problem. These are often people in whom multi-organ support would be seen as likely to offer no overall benefit. That fact should not be allowed to belittle the life prolonging success of limited respiratory support. Intensivists are a repository for the physiological understanding required to optimise ventilatory support. That said, the critical care unit is not necessarily the best location for many such patients. Critical care staff may struggle with the differences in strategy between these and invasively ventilated patients. Six millilitres per kilogram ideal body weight, minimising driving pressure, etc. have no place in NIV in this group of patients. Air-driven ventilators are widely used in ward NIV but perhaps less familiar to those working in critical care. It can be uncomfortable tolerating lower oxygen saturation targets when such targets would be completely unacceptable in other areas of one’s work. The easy availability of other supportive therapies such as vasopressors can lead to “mission creep” and escalating futile care when palliation would be the most appropriate way forward.

The challenge is to offer rapid and reliable NIV to acute type 2 respiratory patients, 24/7, delivered by staff able to spot suitable patients and the confidence to adjust ventilator settings to get the most out of the machines. The full range of intensive care therapies needs to be accessible to those who would benefit from them, but NIV will be the definitive organ support for the majority and perhaps the majority need protecting from the excesses of critical care. Ultimately, hospitals will need to come up with local solutions that fit their situations and facilities. These will come (and in many places, have come) from conversations between critical care, acute and emergency medicine and the respiratory wards.