The first objective in the management of respiratory failure is to reverse and/or prevent tissue hypoxia. Hypercapnia unaccompanied by hypoxemia generally is well tolerated and probably is not a threat to organ function unless accompanied by severe acidosis. Many experts believe that hypercapnia should be tolerated until the arterial blood pH falls below 7.2. Appropriate management of the underlying disease obviously is an important component in the management of respiratory failure.

A patient with acute respiratory failure generally should be admitted to a respiratory care unit or intensive care unit (ICU). Most patients with chronic respiratory failure can be treated at home with oxygen supplementation and/or ventilatory assist devices along with therapy for their underlying disease.

Extracorporeal membrane oxygenation (ECMO) may be more effective than conventional management for patients with severe but potentially reversible respiratory failure.

Peek et al found that survival without severe disability was significantly higher in patients who were transferred to a single specialized center for consideration of ECMO.[7] In a randomized, controlled trial in 180 patients either with a Murray lung injury score of 3.0 or higher or with uncompensated hypercapnia and a pH lower than 7.20 despite optimal conventional treatment, 36.7% of patients in the ECMO arm had died or were severely disabled 6 months after randomization, compared with 52.9% of patients in the conventional treatment arm.

Although average total costs were more than twice as high for ECMO than for conventional care in this study, lifetime quality-adjusted life-years (QALYs) gained were 10.75 for the ECMO group and 7.31 for the conventional group.^[7]

Assurance of an adequate airway is vital in a patient with acute respiratory distress. The most common indication for endotracheal intubation is respiratory failure. Endotracheal intubation serves as an interface between the patient and the ventilator. Another indication is airway protection in patients with altered mental status.

Once the airway is secured, attention is turned toward correcting the underlying hypoxemia, the most life-threatening facet of acute respiratory failure. The goal is to assure adequate oxygen delivery to tissues, generally achieved with an arterial oxygen tension (P_a O₂) of 60 mm Hg or an arterial oxygen saturation (S_a O₂) greater than 90%. Supplemental oxygen is administered via nasal prongs or face mask; however, in patients with severe hypoxemia, intubation and mechanical ventilation are often required.

Coexistent hypercapnia and respiratory acidosis may have to be addressed. This is done by correcting the underlying cause or providing ventilatory assistance.

Mechanical ventilation is used for 2 essential reasons: (1) to increase P_a O₂ and (2) to lower P_a CO₂. Mechanical ventilation also rests the respiratory muscles and is an appropriate therapy for respiratory muscle fatigue.

The use of mechanical ventilation during the polio epidemics of the 1950s was the impetus that led to the development of the discipline of critical care medicine. Before the mid-1950s, negative-pressure ventilation with the use of iron lungs was the predominant method of ventilatory support. Currently, virtually all mechanical ventilatory support for acute respiratory failure is provided by positive-pressure ventilation. Nevertheless, negative-pressure ventilation still is used occasionally in patients with chronic respiratory failure.

Types of mechanical ventilation

Over the years, mechanical ventilators have evolved from simple pressure-cycled machines to sophisticated microprocessor-controlled systems. The following is a brief overview of the basic principles of their use.

Positive-pressure versus negative-pressure ventilation

For air to enter the lungs, a pressure gradient must exist between the airway and the alveoli. This can be accomplished either by raising pressure at the airway (positive-pressure ventilation) or by lowering pressure at the level of the alveolus (negative-pressure ventilation).

The iron lung or tank ventilator is the most common type of negative-pressure ventilator used in the past. These ventilators work by creating subatmospheric pressure around the chest, thereby lowering pleural and alveolar pressure and facilitating flow of air into the patient’s lungs. These ventilators are bulky and poorly tolerated and are not suitable for use in modern critical care units. Positive-pressure ventilation can be achieved via an endotracheal or tracheostomy tube or noninvasively through a nasal mask or face mask.

Controlled versus patient-initiated ventilation

Ventilatory assistance can be controlled or patient-initiated. In controlled ventilation, the ventilator delivers assistance independent of the patient’s own spontaneous inspiratory efforts. In contrast, during patient-initiated ventilation, the ventilator delivers assistance in response to the patient’s own inspiratory efforts. The patient’s inspiratory efforts can be sensed either by pressure or flow-triggering mechanisms.

Pressure-targeted versus volume-targeted ventilation

During positive-pressure ventilation, either pressure or volume may be set as the independent variable.

In volume-targeted (or volume preset) ventilation, tidal volume is the independent variable set by the physician or respiratory therapist, and airway pressure is the dependent variable. In this type of ventilation, airway pressure is a function of the set tidal volume and inspiratory flow rate, the patient’s respiratory mechanics (compliance and resistance), and the patient’s respiratory muscle activity.

In pressure-targeted (or pressure preset) ventilation, airway pressure is the independent variable, and tidal volume is the dependent variable. The tidal volume during pressure-targeted ventilation is a complex function of inspiratory time, the patient’s respiratory mechanics, and the patient’s own respiratory muscle activity.

Endotracheal intubation

Mechanical ventilation requires an interface between the patient and the ventilator. In the past, this invariably occurred through an endotracheal or tracheostomy tube, but there is a growing trend toward noninvasive ventilation, which can be accomplished by the use of either a full face mask (covering both the nose and mouth) or a nasal mask (see Noninvasive Ventilatory Support).^[8] Care of an endotracheal tube includes correct placement of the tube, maintenance of proper cuff pressure, and suctioning to maintain a patent airway.

After intubation, the position of the tube in the airway (rather than the esophagus) should be confirmed by auscultation of the chest and, ideally, by a carbon dioxide detector. As a general rule, the endotracheal tube should be inserted to an average depth of 23 cm in men and 21 cm in women (measured at the incisor). Confirming proper placement of the endotracheal tube with a chest radiograph is recommended.

The tube should be secured to prevent accidental extubation or migration into the mainstem bronchus, and the endotracheal tube cuff pressure should be monitored periodically. The pressure in the cuff generally should not exceed 25 mm Hg.

Endotracheal suctioning can be accomplished via either open-circuit or closed-circuit suction catheters. Routine suctioning is not recommended, because suctioning may be associated with a variety of complications, including desaturation, arrhythmias, bronchospasm, severe coughing, and introduction of secretions into the lower respiratory tract.

Ventilator modes

Pressure support ventilation (PSV) can be categorized as patient-initiated, pressure-targeted ventilation. With PSV, ventilatory assistance occurs only in response to the patient’s spontaneous inspiratory efforts. With each inspiratory effort, the ventilator raises airway pressure by a preset amount. When the inspiratory flow rate decays to a minimal level or to a percentage of initial inspiratory flow (eg, 25% of peak flow), inspiration is terminated.

During PSV, patients are free to choose their own respiratory rate; inspiratory time, inspiratory flow rate, and tidal volume are determined, in part, by the patient’s respiratory efforts. This mode of ventilation should not be used in patients with unstable ventilatory drive, and care must be exercised when the patient’s respiratory mechanics are changing because of bronchospasm, secretions, or varying levels of auto–positive end-expiratory pressure (auto-PEEP).

Intermittent mandatory ventilation (IMV) is a mode whereby mandatory breaths are delivered at a set frequency, tidal volume, and inspiratory flow rate. However, the patient can breathe spontaneously between the machine-delivered breaths.

Most modern ventilators are capable of synchronized IMV (SIMV), whereby the ventilator attempts to deliver the mandatory breaths in synchrony with the patient’s own inspiratory efforts. In essence, the ventilator allows the patient an opportunity to breathe. If the patient makes an inspiratory effort during a window of time determined by the IMV rate, the ventilator delivers a mandatory breath in response to the patient’s inspiratory effort. However, if no inspiratory effort is detected by the ventilator, a time-triggered breath is delivered.

Compared with standard IMV, SIMV may improve patient comfort and may limit dynamic hyperinflation, which may occur when a preset breath is delivered immediately after the patient’s spontaneous inspiratory effort (ie, before exhalation).

In assist-control ventilation, patients receive a fixed tidal volume and inspiratory flow rate with each inspiratory effort, regardless of their respiratory rate. However, a backup rate is selected that guarantees that the patient receives a minimum number of breaths per minute. If the patient’s respiratory rate falls below the backup rate, the ventilator delivers the number of breaths necessary to reach that rate; such breaths are delivered independent of any inspiratory effort by the patient.

In volume-control mode, respiratory rate, tidal volume, and inspiratory flow rate (or inspiratory time) are fixed. This mode is used most often in heavily sedated or paralyzed patients.

In pressure-control mode, as contrasted with volume-control mode, airway pressure is raised by a set amount at a fixed number of times per minute. The physician or respiratory therapist also sets the inspiratory-to-expiratory (I:E) ratio or the inspiratory time. This mode is used most often in heavily sedated or paralyzed patients.

Pressure-control inverse-ratio ventilation (PCIRV) is a variation of simple pressure-control ventilation. In this mode, inspiration is set to be longer than expiration. The I:E ratio should rarely, if ever, exceed 3:1.

Triggering mechanisms

In patient-initiated (assisted) ventilation, the ventilator must sense the patient’s inspiratory effort in order to deliver assistance. Ventilator triggering may be based on a change in either pressure or flow.

With pressure triggering, the ventilator is set to detect a certain change in pressure. The ventilator is triggered whenever airway pressure drops by the set amount. For example, in a patient on no positive end-expiratory pressure (PEEP) with a trigger sensitivity set at 1 cm H₂ O, a breath is triggered whenever airway pressure falls below –1 cm H₂ O. In a patient on 5 cm H₂ O PEEP with the same trigger sensitivity, a breath is triggered whenever airway pressure falls below +4 cm H₂ O.

With flow triggering, a continuous flow of gas is sent through the ventilator circuit. In some ventilators, this continuous flow rate may be set by the physician or respiratory therapist, whereas in other ventilators, the continuous flow rate is fixed. A flow sensitivity is selected, and the ventilator senses the patient’s inspiratory efforts by detecting a change in flow.

When the patient makes an inspiratory effort, some of the gas that was previously flowing continuously through the circuit is diverted to the patient. The ventilator senses the decrease in flow returning through the circuit, and a breath is triggered. One problem with flow triggering is that automatic triggering sometimes results from leaks in the ventilator circuit.

Positive end-expiratory pressure

By maintaining airway (and hence alveolar) pressure greater than zero, PEEP may recruit atelectatic alveoli and prevent their collapse during the succeeding expiration. PEEP also shifts lung water from the alveoli into the perivascular interstitial space and helps with recruitment of alveoli. However, it does not decrease the total amount of extravascular lung water.

In patients with disorders such as acute respiratory distress syndrome (ARDS) or acute lung injury (ALI), PEEP is applied to recruit atelectatic alveoli, thereby improving oxygenation and allowing a reduction in F_I O₂ to nontoxic levels (< 0.6). Applying PEEP of 3-5 cm H₂ O to prevent a decrease in functional residual capacity in patients with normal lungs is a common practice.

In an ARDS Network trial, higher PEEP produced better oxygenation and lung compliance but no benefit to survival, time on ventilator, or nonpulmonary organ dysfunction.^[3] Although sufficient PEEP is essential in ventilator management of patients with ARDS, this level varies from patient to patient. Ideal PEEP helps to achieve adequate oxygenation and decrease the requirement for high fractions of inspiratory oxygen without causing any of the harmful effects of PEEP.

Current evidence does not support routine application of high PEEP strategy in people with ALI or ARDS; however, a study by Briel et al found higher PEEP levels have been associated with improved survival among patients with ARDS.^[9]

PEEP causes an increase in intrathoracic pressure, which may decrease venous return and cardiac output, particularly in patients with hypovolemia.

Inspiratory flow

In volume-targeted ventilation, inspiratory flow is a variable that is set by the physician or respiratory therapist. The inspiratory flow rate is selected on the basis of a number of factors, including the patient’s inspiratory drive and the underlying disease.

Two flow patterns are used commonly: (1) a constant-flow (ie, square-wave) pattern (see the image below) and (2) a decelerating-flow pattern. With a constant-flow pattern, inspiratory flow is held constant throughout the breath, whereas with a decelerating-flow pattern, flow rises quickly to a maximal value and then decreases progressively throughout the breath.

In pressure-targeted ventilation, the inspiratory flow rate is a dependent variable that varies as a function of the preset pressure and the patient’s own inspiratory effort. Because airway pressure is held constant while alveolar pressure rises during inspiration, the pressure difference between airway and alveoli decreases, leading to a decelerating pattern of inspiratory flow.

Ventilator-associated lung injury

Mechanical ventilation is associated with a variety of insults to the lung.

In the past, physicians focused on barotrauma, including pneumothorax, pneumomediastinum, and subcutaneous and pulmonary interstitial emphysema. The manifestations of barotrauma probably result from excessive alveolar wall stress; excessive airway pressure by itself does not appear to cause barotrauma. In critically ill patients, the manifestations of barotrauma can be subtle. For example, the earliest sign of pneumothorax in supine patients may be the deep-sulcus sign or a collection of air anteriorly along cardiophrenic angle.

It is now recognized that lung damage indistinguishable from ARDS may be caused by certain patterns of ventilatory support. Early animal experiments showed that mechanical ventilation employing high peak airway pressures and high tidal volume led to pulmonary edema, possibly as a result of direct parenchymal injury and altered microvascular permeability secondary to high peak alveolar pressures. Subsequent work indicates that excessive tidal volumes resulting in alveolar overdistention are the most important factor in ventilator-associated lung injury.

A strategy of using low tidal volumes in patients with ARDS who are on mechanical ventilation has led to a reduced incidence of barotrauma and improved survival rates in clinical trials.

The mode of ventilation should be suited to the needs of the patient. After the initiation of mechanical ventilation, ventilator settings should be adjusted on the basis of the patient’s lung mechanics, underlying disease process, gas exchange, and response to mechanical ventilation. SIMV and assist-control ventilation are often used for the initiation of mechanical ventilation. In patients with intact respiratory drive and mild-to-moderate respiratory failure, PSV may be a good initial choice.

The lowest F_I O₂ that produces an S_a O₂ greater than 90% and a P_a O₂ greater than 60 mm Hg generally is recommended. The prolonged use of an F_I O₂ lower than 0.6 is unlikely to cause pulmonary oxygen toxicity.

Acute respiratory distress syndrome

In ARDS, the primary objective of mechanical ventilation is to accomplish adequate gas exchange while avoiding excessive inspired oxygen concentrations and alveolar overdistention.

The traditional ventilatory strategy of delivering high tidal volumes leads to high end-inspiratory alveolar pressures (ie, plateau pressure). Many investigators now believe that repeated cycles of opening and collapsing of inflamed and atelectatic alveoli are detrimental to the lung. Failure to maintain a certain minimum alveolar volume may further accentuate the lung damage. Furthermore, transalveolar pressure (reflected by plateau pressure) exceeding 25-30 cm H₂ O is considered to be an important risk factor for stretch injury to the lungs.

Patients with ARDS should be targeted to receive a tidal volume of 6 mL/kg. It is important to remember that the set tidal volume should be based on ideal rather than actual body weight. If the plateau pressure remains excessive (>30 cm H₂ O), further reductions in tidal volume may be necessary.

ARDSNet, a prospective randomized clinical trial, demonstrated a striking reduction in hospital mortality in ARDS patients who were ventilated with 6 mL/kg predicted body weight rather than with 12 mL/kg.^[3] Patients who received the lower tidal volume strategy also had more ventilator-free and organ failure-free days. This strategy may lead to respiratory acidosis, which requires either high respiratory rates and or sodium bicarbonate infusion.

Application of PEEP sufficient to raise the tidal volume above the lower inflection point (P_flex) on the pressure-volume curve may minimize alveolar wall stress and improve oxygenation. A pressure-volume curve can be constructed for an individual patient by measuring plateau pressures at different lung volumes (see the image below). P_flex is the point where the slope of the curve changes, indicating that the lung is operating at the most compliant part of the curve.

A lung-protective strategy in which the P_a CO₂ is allowed to rise (permissive hypercapnia) may reduce barotrauma and enhance survival.

In some patients with ARDS, the prone position may lead to significant improvements in oxygenation; whether this translates to improved outcome is unknown.

Obstructive airway diseases

In patients with chronic obstructive pulmonary disease (COPD) or asthma, initiation of mechanical ventilation may worsen dynamic hyperinflation (auto-PEEP or intrinsic PEEP [PEEPi]). The dangers of auto-PEEP include reduction in cardiac output and hypotension (because of decreased venous return), as well as barotrauma.

The goals of mechanical ventilation in obstructive airway diseases are to unload the respiratory muscles, achieve adequate oxygenation, and minimize the development of dynamic hyperinflation and its associated adverse consequences.

After the initiation of mechanical ventilation, patients with status asthmaticus frequently develop severe dynamic hyperinflation, which is often associated with adverse hemodynamic effects. The development of dynamic hyperinflation can be minimized by delivering the lowest possible minute ventilation in the least possible time. Therefore, the initial ventilatory strategy should involve the delivery of relatively low tidal volumes (eg, 8-10 mL/kg) and lower respiratory rates (eg, 8-12 breaths/min) with a high inspiratory flow rate.

In the absence of hypoxia, hypercapnia generally is well tolerated in most patients. Even marked levels of hypercapnia are preferable to attempts to normalize the carbon dioxide tension (PCO₂), which could lead to dangerous levels of hyperinflation.

Patients often require large amounts of sedation and, occasionally, paralysis until the bronchoconstriction and airway inflammation have improved.

If a decision is made to measure trapped-gas volume (ie, end-inspiratory volume [V_EI]), as recommended by some investigators, an attempt should be made to keep it below 20 mL/kg. Routine measurement of V_EI is not recommended, because measurement of plateau pressure and auto-PEEP provide similar information and are much easier to perform.

Patients with COPD have expiratory flow limitation and are prone to the development of dynamic hyperinflation. Here again, the goal of mechanical ventilation is to unload the respiratory muscles while minimizing the degree of hyperinflation. The use of extrinsic PEEP may be considered in spontaneously breathing patients in order to reduce the work of breathing and to facilitate triggering of the ventilator. Care must be exercised to avoid causing further hyperinflation, and the set level of PEEP should always be less than the level of auto-PEEP.

Patient-ventilator synchrony

During mechanical ventilation, many patients sometimes experience asynchrony between their own spontaneous respiratory efforts and the pattern of ventilation imposed by the ventilator. This can occur with both controlled and patient-initiated modes of ventilation.

To achieve synchrony, the ventilator not only must sense and respond quickly to the onset of the patient’s inspiratory efforts but also must terminate the inspiratory phase when the patient’s “respiratory clock” switches to expiration. Asynchronous interactions (“fighting the ventilator”) may occur when ventilator breaths and patient efforts are out of phase. This may lead to excessive work of breathing, increased respiratory muscle oxygen consumption, and decreased patient comfort.

There are several ways of minimizing patient-ventilator asynchrony. Modern ventilators are equipped with significantly better valve characteristics than older-generation ventilators had. In addition, flow triggering (with a continuous flow rate) appears to be more sensitive and more responsive to patient’s spontaneous inspiratory efforts.

Patient-ventilator asynchrony often occurs in the presence of auto-PEEP. Auto-PEEP creates an inspiratory threshold load and thereby decreases the effective trigger sensitivity. This may be partially offset by the application of external PEEP.

Sometimes, additional sedation may be necessary to achieve adequate patient-ventilator synchrony.

Ventilatory support via a nasal or full-face mask rather than via an endotracheal tube (see the images below) is increasingly being employed for patients with acute or chronic respiratory failure. Noninvasive ventilation should be considered in patients with mild-to-moderate acute respiratory failure. The patient should have an intact airway, airway-protective reflexes, and be alert enough to follow commands.

In clinical trials, noninvasive positive-pressure ventilation (NPPV) has proven beneficial in acute exacerbations of COPD and asthma, decompensated congestive heart failure (CHF) with mild-to-moderate pulmonary edema, and pulmonary edema from hypervolemia. Reports conflict regarding its efficacy in acute hypoxemia due to other causes (eg, pneumonia). A variety of methods and systems are available for delivering noninvasive ventilatory support.

The benefits of NPPV depend on the underlying cause of respiratory failure. In acute exacerbations of obstructive lung disease, NPPV decreases P_a CO₂ by unloading the respiratory muscles and supplementing alveolar ventilation. The results of several clinical trials support the use of NPPV in this setting.

In a large randomized trial comparing NPPV with a standard ICU approach, the use of NPPV was shown to reduce complications, duration of ICU stay, and mortality.^[10] In patients in whom NPPV failed, mortality rates were similar to the intubated group (25% vs 30%).

In the largest prospective randomized study comparing NPPV with standard treatment in patients with COPD exacerbation, Plant et al found that treatment failed in significantly more patients in the control group (27% vs 15%) and that in-hospital mortality rates were significantly reduced by NPPV (20% to 10%). NPPV was administered on the ward; the nurses were trained for 8 hours in the preceding 3 months.^[11]

In addition, 3 Italian cohort studies with historical or matched control groups have suggested that long-term outcome of patients treated with NPPV is better than that of patients treated with medical therapy and/or endotracheal intubation.^{[12, 13, 14]}

In acute hypoxemic respiratory failure, NPPV also helps maintain an adequate P_a O₂ until the patient improves. In cardiogenic pulmonary edema, NPPV improves oxygenation, reduces work of breathing, and may increase cardiac output.

When applied continuously to patients with chronic ventilatory failure, NPPV provides sufficient oxygenation or carbon dioxide elimination to sustain life by reversing or preventing atelectasis or resting the respiratory muscles.

Patients with obesity-hypoventilation syndrome benefit from NPPV as a consequence of reversal of the alveolar hypoventilation and upper-airway obstruction.

Most studies have used NPPV as an intermittent rather than continuous mode of support. Most trials have used inspiratory pressures of 12-20 cm H₂ O and expiratory pressures of 0-6 cm H₂ O and have excluded patients with hemodynamic instability, uncontrolled arrhythmia, or a high risk of aspiration.

Weaning or liberation from mechanical ventilation is initiated when the underlying process that necessitated ventilatory support has improved. In some patients, such as those recovering from uncomplicated major surgery or a toxic ingestion, withdrawal of ventilator support may be done without weaning. In patients who required more prolonged respiratory therapy, the process of liberating the patient from ventilatory support may take much longer.

A patient who has stable underlying respiratory status, adequate oxygenation (eg, P_a O₂/ F_I O₂ >200 on PEEP < 10 cm H₂ O), intact respiratory drive, and stable cardiovascular status should be considered for discontinuance of mechanical ventilation.

Many criteria have been used to predict success in weaning, including a minute ventilation of less than 10 L/min, maximal inspiratory pressure more than –25 cm water, vital capacity more than 10 mL/kg, absence of dyspnea, absence of paradoxical respiratory muscle activity, and agitation or tachycardia during the weaning trial. However, the rapid-shallow breathing index—that is, the patient’s tidal volume (in liters) divided by the respiratory rate (in breaths/min) during a period of spontaneous breathing—may be a better predictor of successful extubation.

In one study, a daily trial of spontaneous breathing in patients with a rapid-shallow breathing index of less than 105 resulted in a shorter duration of mechanical ventilation. A spontaneous breathing trial of only 30 minutes appears adequate to identify patients in whom successful extubation is likely.

In patients who are not yet ready to be liberated from the ventilator, one should focus on the cause of ventilator dependency, such as excessive secretions, inadequate respiratory drive, impaired cardiac function, and ventilatory muscle weakness, rather than the type of ventilator or the mode of assistance.

The weaning protocol could be designed with assist-control ventilation, with gradually increasing time spent in trials of spontaneous breathing or by gradually reducing the level of PSV. SIMV appears to result in less rapid weaning than PSV or trials of spontaneous breathing. Patient-ventilator desynchrony is an important component in a carefully designed weaning protocol. Attention must be directed toward patient comfort, avoidance of fatigue, adequate nutrition, and prevention and treatment of medical complications during the weaning period.

Peak inspiratory and plateau pressures should be assessed frequently. Attempts should be made to limit the plateau pressure to less than 25 cm H₂ O. Expiratory volume is checked initially and periodically (continuously if ventilator-capable) to assure that the set tidal volume is delivered. In patients with severe airflow obstruction, auto-PEEP (PEEPi) should be monitored on a regular basis.

A patient with respiratory failure requires repeated assessments, which may range from bedside observations to the use of invasive monitoring. These patients should be admitted to a facility where close observation can be provided. Most patients who require mechanical ventilation are critically ill; therefore, constant monitoring in a critical care setting is a must.

Cardiac monitoring, blood pressure, pulse oximetry, S_a O₂, and capnometry are recommended. An arterial blood gas determination should be obtained 15-20 minutes after the institution of mechanical ventilation. The pulse oximetry readings direct efforts to reduce F_I O₂ to a value less than 0.6, and the P_a CO₂ guides adjustments of minute ventilation.

Medication Summary