Early management of acute respiratory failure. Part 1: Primary assessment and initial therapy

RICK HODDER and PIERRE CARDINAL

Rick Hodder, MD, FRCPC, Professor of Medicine, Divisions of Pulmonary and Critical Care, and Pierre Cardinal, Associate Professor of Medicine, Division of Critical Care, University of Ottawa and The Ottawa Hospital, 1053 Carling Avenue, Ottawa, Ontario, Canada K1Y4E9

Acute respiratory failure (ARF) is the term used to signify any acute disturbance of either oxygen uptake or carbon dioxide elimination, or both, that is severe enough to pose a threat to life. It is important, therefore, to have an understanding of what levels of hypoxemia and hypercarbia are considered ‘dangerous’. ARF may be suspected in the patient who presents with breathing distress, but must also be considered in the patient with minimal or no pulmonary symptoms who presents with signs of neurologic or cardiac dysfunction. Both primary pulmonary disorders and extrapulmonary disorders may cause ARF. Because there is often an element of urgency when faced with a patient in ARF, the most prudent approach is the ‘primary assessment’, a coordinated combination of assessment and concurrent initial management based on simple principles and with rational therapeutic goals, even before a firm etiologic diagnosis is reached. Part 1 of this series discusses the definition and classification of respiratory failure, and outlines the primary assessment of such patients, using a case history as an example. Part 2 of this two-part series will deal with the basic principles and initial approach to mechanical ventilation for ARF, both non-invasively by mask and by endotracheal intubation.

What is respiratory failure?
Acute respiratory failure can be defined as ‘any impairment of oxygen uptake or carbon dioxide elimination, or both, that is severe enough to be a threat to life’. Although this definition describes a fundamental concept, it is not likely to be very helpful at the bedside, where we are concerned primarily with treating patients’ symptoms and ensuring their safety, rather than with labeling them as being either in or out of ‘respiratory failure’. Nevertheless, it is important to have an understanding of what constitutes a dangerous gas exchange abnormality so that it can be prevented or corrected.
What, then, is a dangerous level of hypoxia or hypoxemia, and what is a dangerous level of hypercapnia or acidemia? The answer, as is often the case, is not always clear, and so an albeit unhelpful but logical response is ‘It depends’ (see Box 1). While most people would not be in immediate risk of dying with an arterial partial pressure of oxygen (PaO2) of 6.7 kPa (50 mmHg), this level of hypoxemia might be very dangerous in an anemic, hypotensive man with acute coronary insufficiency and heart failure whose oxygen delivery is already compromised. Similarly, in a young person suffering from a sedative overdose, a partial pressure of carbon dioxide (PCO2) of 9.3 kPa     (70 mmHg) and a pH of 7.16 in and of themselves are unlikely to be life threatening.2 Other factors such as the attendant hypoxemia, hypotension and inability to protect the airway, rather than the hypercarbia, are more likely to threaten the life of such a patient, and it is these issues that would require prompt attention.
Although blood gases should never be interpreted in isolation and must always be assessed in the context of the patient’s clinical situation, there clearly are thresholds of blood gas derangement that should be considered to be warning signs and should not go unheeded. Hypoxemia involving an acute fall in PaO2 to less than 6.7-7.3 kPa (50-55 mmHg) or an arterial oxygen saturation (SaO2) less than 85-88% is important because it indicates a significant loss of homeostasis and that the patient has minimal oxygenation reserve. Similarly, an acute rise in PCO2 causing a fall in pH to less than 7.20-7.30 cannot be ignored because it also indicates a potentially dangerous loss of ventilatory reserve and capacity and, equally importantly, probably a fatiguing patient. Patients with these degrees of blood gas derangement need to be watched closely and managed aggressively, as they may suddenly lose their remaining reserves and plummet into cardiorespiratory collapse. The accompanying case history of Mr Puffer will help to illustrate some management principles and tips.

Classifying respiratory failure
Traditionally, respiratory failure is divided into primary failure to oxygenate (hypoxemic failure) or primary failure to ventilate (hypercapnic or hypoventilatory failure) (Figure 1).
Examples of primary hypoxemic failure include severe pneumonia, pulmonary edema and acute respiratory distress syndrome (ARDS). Early on in the course of hypoxemic failure, there is a high drive to breathe (stimulated by hypoxemia, various lung and airway receptors and the disease process itself), so that the spared and relatively healthy lung regions can actually support alveolar hyperventilation and produce an initial hypocapnic respiratory alkalosis (reduced PaCO2), despite significant hypoxemia. However, as the disease process advances and less normal lung remains, central nervous system depression occurs, or the breathing muscles lose their ability to cope with increasing ventilatory loads, alveolar hypoventilation may supervene and the PaCO2 begins to rise, culminating in secondary hypercapnia. In the light of this, a normal or rising PaCO2 is a cause for concern, as it indicates a loss of patient ventilatory reserve that may require intervention. In primary hypoxemic failure, the alveolar to arterial oxygen tension difference [D(A–a)O2] and the ratio of the PaO2 to the inspired oxygen fraction (FIO2) are abnormal, indicating that the observed hypoxemia is secondary to sick lungs.
Primary hypercapnic respiratory failure can be further subdivided into failure of the respiratory centers (central failure) or failure of the ventilatory pump (peripheral failure). The ventilatory pump consists of the chest wall, diaphragm and accessory muscles of breathing. Patients with peripheral failure of the ventilatory pump often have to work very hard to breathe (e.g. those with kyphoscoliosis, myopathy or acute exacerbations of chronic obstructive pulmonary disease [COPD]), but cannot breathe effectively and so develop hypercapnia. Ventilatory pump failure also commonly occurs if primary hypoxemic failure is sustained and the diaphragm begins to fail due to high breathing loads. Patients with central respiratory failure usually show no signs of distress and have a decreased respiratory rate (e.g. those with narcotic or sedative overdoses, or brain tumor). These patients may have healthy lungs and the hypoxemia is secondary to the hypercapnia, as evidenced by a normal D(A–a)O2 and PaO2/FIO2 ratio.
The classification shown in  Figure 1, although artificial, is nonetheless useful in that it allows patients to be divided into well-defined categories, thus narrowing the differential diagnosis and facilitating further investigation and therapy. There are, however, limitations to this approach. Patients often present with a mixed picture characterized by both hypoxemic and hypocapnic respiratory failure, such as acute COPD exacerbation in a patient with chronic hypercapnia. In addition, this classification is based on arterial blood gases but, as discussed below, measuring the arterial blood gases should not be a primary preoccupation in the early management of patients with possible respiratory failure. A more rational and efficient approach to suspected respiratory failure relies mostly on clinical examination and tests such as oximetry that are readily available at the bedside. Once the patient has been stabilized, therapy should be refined to treat the underlying cause and prevent or treat the potential complications of respiratory failure.

Recognizing respiratory failure
If acute respiratory failure is defined in terms of blood gas and SaO2 criteria, in order to confirm the presence of respiratory failure these data must be obtained. However, it does not follow that an arterial puncture is immediately necessary for this purpose. An analysis of the oxygenation and acid-base status of any patient can be quickly obtained from the oxygen saturation reading of a pulse oximeter (SpO2) and from a venous blood gas sample (see Box 2).
While it may be easy to establish a diagnosis once all investigations are available or if the situation is extreme, testing takes time and therapy cannot be unduly delayed. Even without gas exchange data, a thoughtful clinical examination should raise the possibility of respiratory failure. The signs and symptoms of respiratory failure (i.e. dangerous gas exchange abnormalities) are non-specific (and often non-respiratory) and mainly reflect end-organ dysfunction of the neurologic and cardiovascular systems (Table 1).
Signs and symptoms specific to the respiratory system (e.g. wheeze, breathlessness, cough), while indicative of significant lung disease, do not necessarily define the presence of dangerous gas exchange. Not­withstanding these observations, there are some important respiratory warning signs that should not be ignored. Although non-specific, a high respiratory rate is one of the most sensitive signs of respiratory failure. For example, in the setting of community-acquired pneumonia, a breathing rate above 30 breaths/ minute correlates with an increased risk of dying from the pneumonia.12 Similarly, an increasing respiratory rate compared with baseline is of greater significance and urgency than an elevated but stable rate in a patient with COPD. Trending heart rate can be helpful. For example, patients with status asthmaticus are tachycardic due to the distress of the disease, and a falling heart rate (despite large doses of bronchodilators) is usually a sign that therapy is working and that the patient is less distressed (although bradycardia in the setting of progressive hypoxemia is an ominous sign). Assessment of the breathing muscles (the ‘vital pump’) can be quite helpful in prognosticating the expected course of acute respiratory failure. Use of accessory breathing muscles, in particular use of the inspiratory sternocleidomastoid muscles and of the abdominal muscles on expiration, indicates excessive loading of the breathing muscles and strongly suggests a serious underlying respiratory problem that, if not remedied, may progress to respiratory muscle fatigue and failure with collapse of the patient from breathing exhaustion. In acute asthma or COPD, for example, use of inspiratory accessory muscles is usually a sign of worsening hyperinflation, and the use of abdominal muscles to assist exhalation signifies a high degree of airflow obstruction, even in the absence of audible wheezing. As the breathing muscles become tired, the usual sequence of events is dyspnea, followed by tachypnea, and then use of accessory breathing muscles, at which time the PaCO2 may still be normal or even low. As diaphragm fatigue and failure supervene, only then does the PaCO2 begin to rise as a relatively late manifestation. Thus, waiting for the PaCO2 to rise to warning thresholds before intervening may delay the diagnosis and put patients at risk unnecessarily.
Patients in respiratory failure may present either in obvious respiratory distress, or they may appear somewhat calm if their ventilatory drive is depressed. For example, patients with primary central hypoventilatory failure usually present with a depressed level of consciousness and a decreased respiratory rate. The most common cause of central hypoventilatory failure is sedative or narcotic overdose (either intentional or not) and all too often the diagnosis is delayed, with the patient being found unresponsive in the middle of the night, having been presumed to be sleeping. Recognizing a patient with primary hypoxemic respiratory failure is usually more straightforward, particularly if the patient presents with a high respiratory rate, use of accessory muscles, intercostal indrawing and marked anxiety. However, the diagnosis can often be much more subtle. For example, older patients, often with a co-morbid illness such as COPD, may develop primary hypoxemic respiratory failure (e.g. due to pneumonia) but, due to weak respiratory muscles that quickly become fatigued, may paradoxically present quietly, with very few signs of overt breathing distress, as they succumb to overwhelming fatigue. Clinically, they may seem merely to be confused or drowsy, but blood gas analysis reveals a combination of hypoxemic and hypoventilatory failure. Similarly, patients with Guillain-Barré syndrome may be in impending or even established respiratory failure, yet only present with a high respiratory rate, with few additional signs other than paralysis and a sense of air hunger or dyspnea (which should never be ascribed to ‘anxiety’).
These points can be applied to our case history: does Mr Puffer have respiratory failure?

Initial management of respiratory failure: primary assessment
When confronted with a patient with possible respiratory failure, the experienced clinician begins an instinctive, coordinated combination of assessment and concurrent initial management based on fundamental principles and past experience. This primary assessment involves a rapid initial evaluation for danger signs and the simultaneous institution of therapy, even before a firm diagnosis is reached. The conventional approach consisting of a thorough history, followed by a complete physical examination and then investigations and treatment is often inappropriate in critically ill patients, when things are moving quickly and time is limited. Initial evaluation should involve assessing not only the patient, but also the overall situation including the resources available and any initial management that has already been started. An example of a general approach to primary assessment in acute respiratory failure is outlined in Box 3.
Simultaneously with this brief initial examination, which should always include an assessment of the ‘vital pump’ for signs of excess ventilatory loads that might provoke respiratory muscle failure (see Box 4), an oximeter should be positioned for the measurement of oxygen saturation. If an assistant is present, blood should also be drawn for initial biochemistry, hematology and venous blood gas assessment. Very few respiratory conditions necessitate immediate intervention at this stage, except possibly significant upper airway obstruction or an obvious tension pneumothorax. A tension pneumothorax is a rare occurrence unless a patient has undergone a procedure such as line insertion or thoracentesis or is a trauma victim. Even patients who are mechanically ventilated rarely develop a pneumothorax unless they are being ventilated with excessive volumes and pressures. If the clinical suspicion is high, or if the clinical examination leaves absolutely no doubt as to this diagnosis, a needle should be immediately inserted into the second intercostal space on the affected side, followed by a chest tube. However, for patients at lower risk of developing a pneumothorax, there is often sufficient time to confirm the diagnosis with a chest radiograph.

Empiric oxygen therapy
When respiratory failure is highly likely, administration of empiric oxygen is usually the most prudent course. The various methods of delivering initial oxygen therapy in respiratory failure are outlined in Box 5. Regardless of the delivery method used, the most important principle is to assess the adequacy of the oxygen therapy often; this is most conveniently done using a pulse oximeter.
Particularly in the setting of COPD, concern is frequently expressed that empiric administration of high concentrations of oxygen may result in a dangerous rise in PaCO2, and carbon dioxide narcosis secondary to elimination of the so-called hypoxic drive to breathe. For the majority of patients, even those with COPD, this is probably an unwarranted fear, as the existence of carbon dioxide narcosis is controversial.4 More importantly, untreated hypoxia is much more dangerous than hypercapnia. It appears that when supplemental oxygen is administered to hypoxemic COPD patients, the subsequent rise in PCO2 is limited, as the drive to breathe from increasing acidemia increases and matches the patient’s former hypoxic drive to breathe, so that the PCO2 eventually plateaus (provided that the patient has not been overly sedated, or is extremely fatigued and cannot increase his minute ventilation). In one study, when 100% oxygen was given to hypoxemic COPD patients in respiratory failure, even though PCO2 levels rose on average 3.1 kPa (23 mmHg) over 15 minutes, no changes in the patients’ level of consciousness were noted.7 Studies in patients with COPD have shown that when the PCO2 rises in response to oxygen therapy, the causes are multifactorial and not solely due to suppression of the hypoxic drive to breathe.7,14 The rise in PaCO2 is also related to a release of carbon dioxide from the newly oxygenated hemoglobin (Haldane effect) and possibly also an increase in ventilation/perfusion mismatching and an increase in dead space ventilation.
In summary, although PCO2 levels may rise somewhat in patients with hypoxic COPD in response to oxygen therapy, the magnitude of hypercarbia is small and not usually clinically relevant. A reduced minute ventilation in this setting is more commonly due to fatigue from the increased work of breathing due to the underlying pulmonary disease than to the necessary oxygen therapy.
In patients who are severely hypoxemic, it is probably best to start with an inspired oxygen concentration of as close to 100% as possible, because the greatest danger is uncorrected hypoxemia. Rather than exposing the patient to a prolonged period of potentially dangerous hypoxemia by starting therapy with low concentrations of oxygen, it is preferable to secure an adequate SpO2 using higher concentration of oxygen and then to titrate the concentration down later. On the other hand, in patients with only modest hypoxemia, in particular those with an acute COPD exacerbation, it is probably prudent to begin with low flow oxygen, or a low FIO2 Venturi mask, and to follow the SpO2 response closely and often. The oxygen concentration need only be titrated to achieve a modest goal of arterial oxygenation, such as an SpO2 of 85-88% measured using pulse oximetry. As discussed earlier, this modest oxygenation goal is usually safe and adequate for the patient and is therefore acceptable in those patients who are very difficult to oxygenate. Of course, it is more comforting to achieve an SpO2 of over 90%, but in difficult patients this my only be achieved with positive pressure ventilators.
The response to oxygen therapy may provide important clues to the etiology of the respiratory failure. The hypoxemia of patients with central hypoventilatory failure or with an exacerbation of COPD and often asthma, will usually easily correct with a relatively low concentration of oxygen. For example, a patient with a narcotic overdose and a PaCO2 as high as 10.7 kPa (80 mmHg) could have an SaO2 of 100% on an FIO2 as low as 0.30 if the lungs are otherwise normal. On the other hand, failure to rapidly correct hypoxemia suggests diseases that produce a severe intrapulmonary shunt (i.e. alveoli that are perfused but not ventilated) such as pneumonia, ARDS, or a right to left intracardiac shunt, as is occasionally seen in patients with a patent foramen ovale and pulmonary hypertension. Failure to correct the hypoxemia despite a high inspired concentration of oxygen should be recognized early and should prompt the physician to consider alternative therapies such as assisted ventilation, either invasive or non-invasive (see Part 2).
The application of these principles can be seen in our case history.

Acute exacerbations in COPD
Several recent reviews on the management of acute COPD exacerbations are available.15-17 Despite this, there are wide variations in practice amongst physicians regarding the management of COPD exacerbations, and the length of stay in hospital for this condition is frequently unnecessarily long.18-20 This has led to the concept that the initiation of standing orders, or a care map, might help to make the management of acute COPD exacerbations more efficient and effective.
Standing orders or ‘critical pathways’ are predefined, evidence-based management strategies that aim to improve the quality of care and reduce deviation from best practice. A good example of a successful critical pathway was that used in the Canadian ‘Capital Trial’; this study confirmed the value of a critical pathway over standard care for the treatment of community-acquired pneumonia.21 Application of the critical pathway was associated with fewer days in hospital and a reduction in the number of unnecessary hospital admissions. In the case of acute exacerbations of COPD, standing orders have been shown to lead to more appropriate use of antibiotics and corticosteroids.20 Examples of standing orders for acute COPD exacerbations can be found in the study of Goddard et al.20
In our case history, Mr Puffer required admission to hospital and treatment was begun according to an acute COPD exacerbation critical pathway, which included targeted oxygen therapy, initial aggressive bronchodilator therapy with both b2-agonist and anticholinergic agents, systemic corticosteroids and appropriate antibiotics. In Mr Puffer’s case, because he was considered to have ‘complicated’ COPD (advanced airflow obstruction, frequent exacerbations and recent need for oral corticosteroids), antibiotic coverage against Pseudomonas species was chosen.15 The critical pathway used in Mr Puffer’s case did not include routine use of non-invasive mechanical ventilation, but such an approach has been suggested for this type of patient with advanced COPD.22 This will be discussed in greater detail in Part 2.