It seems surprising now, but the concept of respiratory failure is quite a recent development in respiratory medicine. Up until the middle of the 20th century, when chest physicians were dealing primarily with tuberculosis, it was virtually overlooked.1 The principles of the biochemical control of respiration had only recently been demonstrated, arterial blood analysis was difficult and time-consuming, and the understanding of acid-base metabolism was primitive by modern standards. The work of Moran Campbell and his colleagues,2 however, clarified the physiological basis of the clinical presentations of respiratory failure. The “blue bloater” who hypoventilated and developed right heart failure was contrasted with the “pink puffer” who suffered breathlessness but maintained a normal carbon dioxide tension and near normal oxygen tension. The adverse effect of supplemental oxygen of removing the hypoxic drive and exacerbating respiratory failure was recognised and controlled oxygen therapy using Venturi masks became established.

At the same time as these ideas about respiratory failure due to chronic obstructive pulmonary disease (COPD) were developing, a separate line of thought was evolving. Poliomyelitis—which until the early 1900s had caused only sporadic infections—began to appear in epidemics in the USA and many northern European countries and left thousands of children and some adults paralysed. These epidemics reached their peak in the 1950s until the advent of the Salk and Sabin immunisations. Acute hypercapnic respiratory failure due to respiratory muscle weakness was often fatal.

The first effective treatment was the tank ventilator or iron lung which was developed in 1928 by Drinker, an engineer with the Consolidated Gas Company of New York City.3 This comprised a negative pressure chamber which surrounded the body up to the neck and which was connected to a powerful pump capable of evacuating the chamber and thereby expanding the chest and abdomen. When the negative pressure was released, the elastic recoil of the chest wall and the lungs led to expiration before the next cycle of negative pressure was generated. This treatment was widely used but led to problems when bulbar function was compromised, since it did not protect the airway. The importance of this was demonstrated in the massive 1952 Copenhagen poliomyelitis epidemic. Many patients were intubated and then ventilated by hand by medical students because of a lack of mechanical ventilators. The survival rate was greater than in those treated in iron lungs.4 This important observation led to the widespread use of positive pressure invasive ventilation for ventilatory failure and to the development of modern intensive care units.

It was soon recognised, however, that survival in patients with COPD was poor if they were invasively ventilated during an exacerbation, and a generation of physicians and anaesthetists came to believe that respiratory support was usually inappropriate for these patients. Ventilatory support for hypercapnic respiratory failure from the 1950s until the mid 1980s remained largely confined to patients with neuromuscular and skeletal disorders. Compact variants of the tank ventilators such as jacket and cuirass ventilators were developed for long-term use, and they prolonged the life expectancy, normalised blood gases and improved the quality of life.5 Other effective appliances were developed, including those that generated a positive expiratory pressure around the chest or abdomen, high-frequency chest wall oscillators, rocking beds which used gravitational force to compensate for a paralysed diaphragm, and phrenic nerve stimulators to make up for a defective medullary respiratory drive or damage to the respiratory pathways in the spinal cord.6 However, none of these treatments was widely applied, probably because of their cost, the difficulty in selecting appropriate patients and the considerable expertise required to make them work.

Even before the identification of rapid eye movement (REM) sleep in 1953 it was recognised that mechanical ventilatory support was effective if it was used only at night. Sleep was thought to be a vulnerable period for respiration during which voluntary control of breathing was absent and respiration relied on reflex pathways. Ideas about respiratory control during sleep and wakefulness have developed considerably since then,7 but ventilatory support at night usually remains sufficient to maintain normal arterial blood gases and to relieve symptoms of respiratory failure throughout the day.

The breakthrough in the mid 1980s was the development of comfortable and effective nasal and oronasal masks, with the confidence to apply positive pressure ventilation to the upper airway rather than through a tracheostomy.8 This relied on bulbar function to protect the tracheobronchial tree from aspiration. The ventilators for non-invasive positive pressure ventilation became simpler, smaller, had better triggering systems and were at least as reliable as the ventilators previously used for negative pressure ventilation. Mask ventilation was much more readily applied and more convenient for patients to use in the long term in their homes.

The evidence for the value of non-invasive positive pressure ventilation has unfortunately come mainly from case reports and retrospective series rather than randomised controlled trials. The studies have nevertheless shown an improvement in life expectancy, quality of life and blood gases, both during the day and at night in neuromuscular and skeletal disorders.9 10 The situation is not so clear for COPD. A recent meta-analysis11 of non-invasive ventilation in COPD has shown promising but not decisive results, although they are certainly better than with long-term invasive ventilation.12 In a cash-limited National Health Service where evidence is increasingly required to justify expenditure on treatments, it is particularly unfortunate that the evidence for the effectiveness of non-invasive ventilation in hypercapnic respiratory failure is so limited.

The use of non-invasive ventilation is spreading to patient groups with only mild respiratory failure or those who are at high risk of developing it. These include patients with Duchenne’s muscular dystrophy,13 motor neurone disease and those with previous poliomyelitis in whom it is important to avoid excessive respiratory muscle activity because of the risk of developing the post poliomyelitis syndrome. Early intervention in some of these groups may well be effective, but the lessons from the recent past of not carrying out randomised controlled trials need to be learnt. Non-invasive ventilation is also becoming a standard technique in critical care units, both to avoid intubation and to assist in the weaning process.14 It is now widely used for acute exacerbations of COPD15 and in specialist units to wean these patients from invasive ventilation.16 There is little doubt that non-invasive ventilation will soon also become standard practice in treating the obesity hypoventilation syndrome as the obesity epidemic spreads through urbanised societies.

Despite these wide applications of ventilatory support for hypercapnic respiratory failure, the funding and resources—particularly of skilled medical and non-medical staff—are limited. While the provision of ventilatory support for acute respiratory failure should be provided in all acute district general hospitals, it is often difficult to provide continuously available high quality care. The issue of how much centralisation is best for long-term care also needs to be addressed. When the British Thoracic Society was established in 1982 virtually all non-invasive ventilation was provided in three units: the South Western Hospital in London (later transferred to St Thomas’ Hospital) under the supervision of Geoffrey Spencer, Newmarket General Hospital (now at Papworth Hospital) under myself, and the Brompton Hospital under Margaret Branthwaite. Since then the provision of care has become much more widespread, although these three large units continue to care for many patients. Evidence regarding the value for money, quality of care and accessibility are issues that commissioners of health care are increasingly concerned about and which will determine the future pattern of provision of respiratory support in the UK.

The benefits from non-invasive ventilation show how the commonly held widespread negative attitudes about the outlook in respiratory failure need to be re-fashioned. New techniques will need to be assessed and the available ones better evaluated so that they can be targeted more precisely to those patients who are likely to benefit and so that home ventilation can be successfully integrated into the lives of both patients and carers.