Special Feature

Ventilatory Support in Patients with COPD: An Update

By Mark T. Gladwin, MD

In the rapidly changing field of mechanical ventilation, how best to manage patients with chronic obstructive pulmonary disease (COPD) or acute severe asthma remains one of the most active and important topics. In this brief review, using a question-and-answer format, I discuss several aspects of ventilatory support in patients with severe COPD, emphasizing recent developments.1

Noninvasive Ventilation for Acute Exacerbations

• Can noninvasive positive-pressure ventilation (NPPV) be used safely and effectively in COPD patients with severe respiratory acidosis (for example, arterial PCO2 > 70 mm Hg and pH < 7.20)?

• How much inspiratory positive pressure is needed to achieve adequate unloading of the ventilatory muscles?

• Is it necessary to apply positive end-expiratory pressure (PEEP) to counteract the effects of air trapping and auto-PEEP in order to achieve clinical success?

When initiating NPPV, inspiratory and expiratory pressures are typically set at low levels for patient comfort and acceptance, and gradually raised to levels between 8 and 20 cm H2O inspiratory pressure and between 0 and 6 cm H2O expiratory pressure. Two recent studies of NPPV in patients with severe exacerbations of COPD used different levels of pressure support with equally positive results. Kramer and colleagues2 used an average inspiratory pressure of 11 cm H2O delivered by nasal mask with an average expiratory pressure of 3 cm H2O, while Brochard and colleagues3 used an inspiratory pressure of 20 cm H2O but no PEEP, delivered via face mask. Thus, the application of PEEP (called expiratory positive airway pressure, EPAP, on some ventilators) may not be necessary in all cases. Gastric distention with air is considered unlikely with pressure support levels less than 25 cm H2O.4

The application of modest amounts of PEEP will reduce the inspiratory work of spontaneous breathing associated with dynamic hyperinflation and auto-PEEP. This is supported by a trial of continuous positive airway pressure (CPAP) without additional inspiratory pressure in patients with COPD and acute ventilatory failure, which demonstrated improvements in dyspnea, inspiratory effort, and arterial blood gas values.5

Numerous uncontrolled studies, as well as case series using historical controls, have consistently demonstrated an improvement in ventilatory failure with NPPV. Recent prospective randomized trials have demonstrated similar positive results. Brochard et al3 randomized 85 patients with severe COPD and acute ventilatory failure to noninvasive inspiratory positive airway pressure of 20 cm H2O or to standard therapy without NPPV. Intubation rates in the two patient groups were 26% and 74%, respectively, mortality 9% vs. 29%, complications 16% vs. 45%, and hospital length of stay 23 ± 17 days vs. 35 ± 33 days. Kramer et al2 studied 31 patients with ventilatory failure, and in the subgroup with COPD the rate of intubation was 9% with NPPV as compared to 67% in the standard treatment group.

With the advent of NPPV, mandatory criteria for endotracheal intubation are becoming more difficult to identify. While apnea or agonal respiration, uncontrolled agitation, uncorrectable life-threatening hypoxia, hemodynamic instability or serious dysrhythmia, and a high risk for aspiration remain relative exclusion criteria for NPPV and, thus, indications for endotracheal intubation, a falling arterial pH secondary to ventilatory failure and rising PaCO2 is less so. Thus, while a pH of 7.25 or less has historically been considered a reasonable "line in the sand" beyond which intubation was necessary, this may now be a manageable degree of ventilatory failure with properly applied NPPV. Indeed, in the Brochard study,3 the average arterial pH in patients in whom intubation was successfully avoided was as low as 7.28 ± 0.1, with average PaCO2 70 ± 12 mmHg.

The level of consciousness acceptable for noninvasive ventilation is likewise controversial. While patients with severe obtundation have frequently been excluded from studies of NPPV, other authors have documented success even in patients with this finding.6

Dynamic Hyperinflation and Auto-PEEP

• How serious a problem is auto-PEEP in hospitalized patients with COPD?

• What is the difference between static and dynamic auto-PEEP?

• What is the best way to quantitate dynamic hyperinflation?

• Does externally applied PEEP help to maintain the patency of collapsible airways in patients with severe COPD?

No concept in the management of acute exacerbations of COPD is more important than the concept of dynamic hyperinflation and auto-PEEP.1,7,8 Whenever there is insufficient emptying of alveoli during exhalation, as occurs with rapid respiratory rates in persons with normal airways or in persons with airway obstruction due to emphysema or reactive airways disease, the alveoli and the lung will be overdistended at end-exhalation. This elevation of end-expiratory resting lung volume (functional residual capacity, FRC) above normal is termed "dynamic hyperinflation" and is accompanied by an elevated net static recoil pressure of the respiratory system. This pressure has been called auto-PEEP, occult-PEEP, or intrinsic-PEEP because it cannot be measured on the ventilator manometer (because the manometer communicates with atmospheric pressure during exhalation) and is produced intrinsically by the patient and not set by the ventilator.

Methods for quantitating dynamic hyperinflation and auto-PEEP include the following:

• Directly measure the excess gas trapped at end-exhalation.9 Measure expired volume during a 30- to 50-second period of apnea following a mechanically delivered breath. Subtraction of the delivered tidal volume from the total expired volume will give the volume of trapped air. This technique requires a completely relaxed patient without inspiratory effort.

• Apply the end-expiratory occlusion technique. Manually occlude the ventilator’s expiratory port at end-expiration, or with some ventilators activate an end-expiratory pause button on the ventilator control panel. An occlusion of about one second allows time for the trapped air to equilibrate across most obstructed airways. The pressure measured on the ventilator’s pressure manometer during this maneuver is the auto-PEEP. This technique is only applicable in patients who are not actively attempting to breathe.

• Directly measure esophageal pressure (which is equivalent to pleural pressure) using an esophageal balloon and pressure transducer.

• Monitor inspiratory flow, peak inspiratory pressure, or end-inspiratory plateau pressure while increasing dialed-in PEEP by increments of 3 cm H2O. As long as flow, peak inspiratory pressure, or plateau pressure does not change, the set (external) PEEP is below the level of auto-PEEP. If flow decreases or peak inspiratory or plateau pressure goes up with the addition of external PEEP, the auto-PEEP level has been exceeded. Only peak inspiratory pressure can be measured if patients are actively attempting to breathe.

Static PEEP refers to the measurement of PEEP under static conditions and dynamic PEEP under dynamic conditions. A static condition occurs when there is no inspiratory effort and measures the pressure or volume of gas emptying from the lungs. During the end-expiratory occlusion method, a prolonged occlusion allows for emptying of even the most obstructed regions of the lung (that is, those with the longest time constants). This also occurs with the direct measurement of total trapped gas. These two static measurements will more accurately reflect the entire volume or pressure of trapped gas (as there is time for almost complete emptying), and will thus more accurately represent the detrimental effects of auto-PEEP on intrathoracic pressure and venous return than will more rapid approximations.

Dynamic measurements can be made with an esophageal balloon measuring intrathoracic pressure and the simultaneous measurement of airflow. When auto-PEEP is present and the patient attempts to breathe, the intrathoracic (esophageal) pressure will fall while initially no flow will occur. Eventually the pressure drops below the level of auto-PEEP and flow begins. The amount of negative intrathoracic pressue required to initiate flow is the dynamic auto-PEEP.

Understanding this mechanism leads to further observations. First, airflow should occur when the intrathoracic pressure exceeds the auto-PEEP of the lung units with the least auto-PEEP. This value may be substantially less than the level of auto-PEEP measured under static conditions and can be measured in a spontaneously breathing patient. Furthermore, this measurement more accurately reflects the elevated work of breathing associated with auto-PEEP because this is the minimal threshold (hurdle) that must be overcome to initiate inspiratory flow. Titrating PEEP in increments of 3 cm H2O until flow occurs or peak or plateau inspiratory pressures increase will also reflect the lowest auto-PEEP of any lung unit and is a dynamic measurement.

Auto-PEEP likely contributes to deaths from pulseless electrical activity (PEA) in hospitalized patients. The patient with COPD is at greatest risk for this complication immediately after intubation. The urgency of the situation often leads to overzealous manual ventilation prior to connection to the ventilator. Large tidal volumes and rapid respiratory rates increase lung volume and shorten expiratory time, preventing adequate exhalation and potentiating dynamic hyperinflation. In the setting of non-elective emergency intubation, the reduction in cardiac output associated with dynamic hyperinflation is often compounded by volume depletion and sedation.

Hypotension occurs in 25% of all emergency intubations, and more frequently in patients with COPD and baseline hypercapnia. The transition from manual to mechanical ventilation carries the iatrogenic hazards of attempting to apply physiological tidal volumes and respiratory rates and ill-advised attempts to drive PaCO2 to the normal range—interventions that critically reduce expiratory time.

It is instructive to consider case reports of patients with COPD who develop PEA and do not respond to aggressive resuscitation.10,11 In several cases, after resuscitative efforts had been discontinued, the arterial line began to pick up a pulse. Discontinuation of mechanical ventilation allowed exhalation to occur and intrathoracic pressure to decrease, resulting in enhanced venous return and restoration of cardiac output. A recent review of 89 in-hospital cardiac arrests,12 35 of which involved PEA, revealed no discernible etiology in 18, and 13 (74%) of these patients were subsequently found to have COPD by history, pulmonary function testing, or at autopsy. By contrast, only 11% of the remaining patients had COPD. This suggests that unrecognized auto-PEEP may be a common cause of PEA.

Application of external PEEP in the presence of auto-PEEP will reduce the patient’s work of spontaneous breathing13 but can also increase dynamic hyperinflation. External PEEP should only be applied to the spontaneously breathing patient to decrease work of breathing. The impact of auto-PEEP on mechanical work of breathing, both on and off the ventilator, is significant. With dynamic hyperinflation, at end-expiration there will be a significant pressure gradient between the (positive) alveolar pressure and (zero) pressure in the airway distal to the critical closure point. The process of inspiration involves reversing this pressure gradient, such that alveolar pressure is less than airway pressure and flow reverses. If end-expiratory alveolar pressure (that is, auto-PEEP) remains, for example, at 10 cm H2O, intrapleural pressures must exceed -10 cm H2O in order to drop alveolar pressure to less than atmospheric and initiate inspiratory flow. Therefore, auto-PEEP represents an inspiratory threshold load that must be overcome. Furthermore, in the presence of dynamic hyperinflation, breathing occurs at higher lung volumes, such that inspiration occurs at a higher portion of the pressure-volume curve, where there is a greater inward elastic recoil of the overexpanded chest wall.

The mechanically ventilated patient with dynamic hyperinflation must first overcome this inspiratory threshold load, and, depending on the triggering mechanism of the ventilator, further reverse flow (flow-triggered) or lower the airway pressure (pressure triggered) to initiate the delivery of a breath. This can be detected at the bedside by observing chest wall expansion and accessory muscle activity that do not trigger a ventilated breath. Airway and esophageal pressure tracings reveal negative inspiratory excursions that do not trigger a breath, alternating with those that do. Carefully applying extrinsic PEEP increases airway pressure and reduces the pressure gradient required to reverse flow on inspiration. Applied PEEP can be titrated to a level at which every inspiratory effort triggers a breath, resulting in sharp reductions in the patient’s work of breathing.

Patients with emphysema, characterized by compliant or floppy airways, poorly tethered by damaged elastic fibers, develop an early "equal pressure point" as the positive extramural pressure exceeds the elastic recoil forces of the airway and the positive intramural airway pressure. Theoretically, the application of extrinsic PEEP less than the original level of auto-PEEP should serve to prevent this early dynamic airway closure at the equal pressure point by maintaining a positive airway pressure that counterbalances the positive extramural pressure that surrounds the airways.

Unfortunately, this concept that applied PEEP thus "stents" open airways and facilitates lung emptying is not supported by clinical studies. Application of external PEEP rarely leads to actual reductions in lung volume. As the level of applied PEEP approaches that of auto-PEEP, lung volume actually increases. This is in keeping with the waterfall analogy of auto-PEEP,1 where a waterfall represents the critical airways narrowing, the water above the behind the dam represents the auto-PEEP, and the water beneath the waterfall represents the pressure at the airway. As the downstream water level rises above that of the waterfall, the water backs up behind the waterfall. In fact, the greatest risk of applied PEEP in the COPD patient is inducing further hyperinflation with cardiovascular compromise.

Progressive increments of PEEP compared with 0 PEEP induce hemodynamic impairment when levels of applied PEEP exceed 80-85% of measured auto-PEEP. It is therefore reasonable to judiciously apply PEEP to about 80% of the measured auto-PEEP level, to reduce work of breathing associated with patient-initiated mechanical ventilation. However, there is no rationale for its use during controlled mechanical ventilation when there is no patient inspiratory effort (such as in the paralyzed patient), or in an attempt to "stent" open airways to reduce lung volume.

Identifying clinical assessments that actually reflect alveolar distention and limit the magnitude of this distention should prevent complications of dynamic hyperinflation. Plateau pressure is measured by occluding the proximal airway at end inspiration, allowing the peak proximal airway pressure to equilibrate across airways with high resistance. Peak inspiratory pressure falls to a lower post-occlusion pressure that more accurately reflects the elastic recoil of the respiratory system and thus alveolar distention. In patients with COPD and asthma, peak airway pressure is highly dependent on inspiratory airflow resistance and peak inspiratory flow, while plateau pressure is independent of flow-resistive properties unless the pause time is inadequate to allow complete equilibration or the airways are completely obstructed.

It has been recommended that plateau pressures be kept below 35 cm of water, although this remains theoretical. Alternatively, limiting end-expiratory volume (the excess volume exhaled in addition to tidal volume during a period of apnea) to less than 1.4 liters14 may be relevant to patients with asthma and COPD, in that hypotension and barotrauma are uncommon in patients with severe asthma, when VEI is less than this. Unfortunately, the measurement of VEI requires a 30- to 60-second complete exhalation, which clinically can only be achieved with deep sedation, usually with neuromuscular blockade.

Administration of Bronchodilators to Mechanically Ventilated Patients

• Are nebulizers more effective than metered-dose inhalers (MDIs)?

• Where in the ventilator circuit should the aerosol generator be placed for optimal delivery of drugs to the patient?

• What steps can the clinician take to maximize bronchodilator delivery during mechanical ventilation?

Beta-adrenergic and anticholinergic bronchodilators can be effectively delivered to the mechanically ventilated patient via small-volume nebulizers (SVN) or MDIs. These medications, when dosed and delivered effectively, will reduce airway resistance (measured by peak and plateau inspiratory pressures) and dynamic hyperinflation.

Delivery of bronchodilator aerosols can be improved by a number of measures under control of the clinician.1,15,16 First, medication should be delivered at a distance from the endotracheal tube in the inspiratory limb of the ventilator circuit to avoid impaction and deposition of larger aerosol particles. Second, humidification of the ventilator circuit reduces aerosol delivery by one-third to one-half, likely by increasing particle size, resulting in particle impaction in the circuit. Finally, the ventilator settings will affect delivery. In a lung model, delivery was enhanced by spontaneous breathing (CPAP mode), larger tidal volumes, and an increase in duty cycle (total inspiratory time/duration of total breathing cycle). Some authors recommend a breath hold after an MDI actuation to increase duty cycle. This recommendation is reasonable provided dynamic hyperinflation is modest and the clinician is cognizant that increases in duty cycle will increase dynamic hyperinflation.

There has been considerable debate about whether nebulizers are more effective than metered-dose inhalers in ventilated patients. Numerous studies support the efficacy of MDIs provided spacer devices are used, the MDI is actuated immediately prior to inspiratory flow with a 30-second to 1-minute pause between actuations, and the above general recommendations are followed. Using these techniques, four puffs (90 mcg/puff) of albuterol reduced airway resistance to the same degree as eight and 16 puffs, without the rise in heart rate that occurred with the higher doses.17 Advantages of MDIs over nebulizers are lower cost, freedom from contamination, and ease of dosing.

Inspiratory Pressure and Work of Breathing During Weaning

• What is the effect of the endotracheal tube on work of breathing?

• Is it necessary to provide pressure support during trials of spontaneous breathing?

This remains an unclear and controversial area. Recent work by Straus and colleagues18 suggests that the work of breathing through a size 7- to 9-mm internal diameter endotracheal tube is the same during a T-piece trial as it is following extubation. Using the acoustic reflection method, they measured, in 14 successfully extubated patients, the resistance of the endotracheal tube and of the supraglottic airway during a two-hour weaning T-piece trial and again following extubation. While endotracheal tube resistance was greater than supraglottic airway resistance, work of breathing was the same with and without the endotracheal tube. Straus et al conclude that a two-hour trial of spontaneous breathing through an endotracheal tube, without any ventilator support, is similar to the work of breathing following extubation. Therefore, pressure support would likely provide excessive support. Application of continuous positive airway pressure (CPAP) at 5 cm H2O would increase functional residual capacity (assuming that auto-PEEP did not exceed this level) but would not provide excessive support. Flow triggering would reduce the work imposed by dead space in the ventilator and tubing.

The advantage of a trial of spontaneous breathing without any positive inspiratory pressure is further supported by the results of two much-discussed weaning studies,19,20 demonstrating that even difficult to wean patients may be successfully weaned using two-hour trials of spontaneous breathing via either a T-piece or CPAP of 5 cm H2O.


1. Gladwin MT, Pierson DJ. Mechanical ventilation of the patient with severe chronic obstructive pulmonary disease. Intensive Care Med 1998;24:898-910.

2. Kramer N, et al. Randomized, prospective trial of noninvasive positive pressure ventilation in acute respiratory failure. Am J Respir Crit Care Med 1995;151: 1799-1806.

3. Brochard L, et al. Noninvasive ventilation for acute exacerbations of chronic obstructive pulmonary disease. N Engl J Med 1995;333:817-822.

4. Brochard L, et al. Reversal of acute exacerbations of chronic obstructive lung disease by inspiratory assistance with a face mask. N Engl J Med 1990;323:1523-1530.

5. Goldberg P, et al. Efficacy of noninvasive CPAP in COPD with acute respiratory failure. Eur Respir J 1995;8:1894-1900.

6. Corrado A, et al. Intermittent negative-pressure ventilation in the treatment of hypoxic hypercapnic coma in chronic respiratory insufficiency. Thorax 1996;51: 1077-1082.

7. Ranieri VM, et al. Physiologic effects of positive end-expiratory pressure in COPD patients during acute ventilatory failure and controlled mechanical ventilation. Am Rev Respir Dis 1993;147:5-13.

8. Ranieri, VM, et al. Auto-positive end-expiratory pressure and dynamic hyperinflation. Clin Chest Med 1996;17;379-394.

9. Tuxen DV, Lane S. The effects of ventilatory pattern on hyperinflation, airway pressures, and circulation in mechanical ventilation of patients with severe air-flow obstruction. Am Rev Respir Dis 1987;136:872-879.

10. Myles PS, Madden H, Morgan EB. Intraoperative cardiac arrest after unrecognized dynamic hyperinflation. Br J Anaesth 1995;74:340-342.

11. Rogers PL, et al. Auto-PEEP during CPR. An "occult" cause of electromechanical dissociation? Chest 1991;99:492-493.

12. Lapinsky SE, Leung RS. Auto-PEEP and electromechanical dissociation (letter). N Engl J Med 1996;335:674.

13. Marini JJ. Should PEEP be used in airflow obstruction? Am Rev Respir Dis 1989;140:1-3.

14. Williams TJ, Tuxen DV, Scheinkestel CD. Risk factors for morbidity in mechanically ventilated patients with acute severe asthma. Am Rev Respir Dis 1992;146:607-615.

15. Dhand R, Tobin MJ. Bronchodilator delivery with metered-dose inhalers in mechanically-ventilated patients. Eur Respir J 1996;9:585-595.

16. Fink JB, et al. Aerosol delivery from a metered-dose inhaler during mechanical ventilation. Am J Respir Crit Care Med 1996;154:382-387.

17. Dhand R, et al. Dose response to bronchodilator delivered by metered-dose inhaler in ventilator-supported patients. Am Rev Respir Crit Care Med 1996;154:388-393.

18. Straus C, et al. Contribution of the endotracheal tube and the upper airway to breathing workload. Am J Respir Crit Care Med 1998;157(1):23-30.

19. Brochard LA, et al. Comparison of three methods of gradual withdrawal from ventilatory support during weaning from mechanical ventilation. Am J Respir Crit Care Med 1994;150(4):896-903.

20. Esteban AF, et al. A comparison of four methods of weaning patients from mechanical ventilation. Spanish Lung Failure Collaborative Group. N Engl J Med 1995;332(6):345-350.