Clinical Anesthesia

Chapter 11

Respiratory Function

Michael L. Ault

M. Christine Stock

Key Points

1. In a person with normal lungs, both breathing and coughing can be performed exclusively by the diaphragm.

2. In the adult, the tip of an orotracheal tube moves an average of 3.8 cm with flexion and extension of the neck, but can travel as much as 6.4 cm. In infants and children, displacement of even 1 cm can move the tube above the vocal cords or below the carina.

3. The following anatomy should be considered when contemplating the use of a double-lumen tube. The adult right main stem bronchus is ~2.5 cm long before it branches into lobar bronchi. In 10% of adults, the right upper lobe bronchus departs from the right main stem bronchus <2.5 cm below the carina. In 2 to 3% of adults, the right upper lobe bronchus opens into the trachea, above the carina.

4. When lung compliance is reduced, larger changes in pleural pressure are needed to create the same tidal volume (VT). Patients with low lung compliance breathe with smaller VT and more rapidly, making spontaneous respiratory rate the most sensitive clinical index of lung compliance.

5. Carotid and aortic bodies are stimulated by Pao2 values less than 60 to 65 mm Hg. Thus, patients who depend on hypoxic ventilatory drive do not have Pao2 values >65 mm Hg. The peripheral receptors' response will not reliably increase ventilatory rate or minute ventilation to herald the onset of hypoxemia during general anesthesia or recovery.

6. There are three causes of hyperventilation: arterial hypoxemia, metabolic acidemia, and central etiologies (e.g., intracranial hypertension, hepatic cirrhosis, anxiety, pharmacologic agents).

7. Increases in dead space ventilation primarily affect CO2 elimination (with minimal influence on arterial oxygenation), and physiologic shunt increase primarily affects arterial oxygenation (with minimal influence on CO2 elimination).

8. During spontaneous ventilation, the ratio of alveolar ventilation to dead space ventilation is 2:1. The alveolar-to-dead space ventilation ratio during positive-pressure ventilation is 1:1. Thus, minute ventilation during mechanical ventilatory support must be greater than that during spontaneous ventilation to achieve the same Paco2.

9. Paco2 ≥ PETCO2 unless the patient inspires or receives exogenous CO2. The difference between Paco2 and PETCO2 is because of dead space ventilation. The most common reason for an acute increase in dead space ventilation is decreased cardiac output.

10.     The best evaluation of the efficiency with which the lungs oxygenate the arterial blood is the calculation of shunt fraction. It is the only index of oxygenation that takes into account the contribution of mixed venous blood to arterial oxygenation.

11.     When functional residual capacity is reduced, lung compliance falls and results in tachypnea, and venous admixture increases, creating arterial hypoxemia.

12.     There is no compelling evidence that defines rules or pa-rameters for ordering preoperative pulmonary function tests. Rather, they should be obtained to ascertain the presence of the reversible pulmonary dysfunction (bronchospasm) or to define the severity of advanced pulmonary disease.


13.     Smoking patients should be advised to stop smoking at least 2 months prior to an elective operation to decrease the risk of postoperative pulmonary complications (PPCs).

14.     The operative site is one of the most important determinants of the risk of PPC. The highest risk for PPC is associated with nonlaparoscopic upper abdominal operations, followed by lower abdominal and intrathoracic operations.

15.     The single most important aspect of postoperative pulmonary care and prevention of PPC is getting the patient out of bed, preferably walking.

Anesthesiologists directly manipulate pulmonary function. Thus, a sound and thorough working knowledge of applied pulmonary physiology is essential to the safe conduct of anesthesia. This chapter discusses pulmonary anatomy, the control of ventilation, oxygen and carbon dioxide transport, ventilation–perfusion relationships, lung volumes and pulmonary function testing, abnormal physiology and anesthesia, the effect of smoking on pulmonary function, and assessing risk for postoperative pulmonary complications (PPCs).

Functional Anatomy of the Lungs

This section emphasizes functional lung anatomy, with structure described as it applies to the mechanical and physiologic function of the lungs.


The thoracic cage is shaped like a truncated cone, with a small superior aperture and a larger inferior opening to which the diaphragm is attached. The sternal angle is located in the horizontal plane that passes through the vertebral column at the T4 or T5 level. This plane separates the superior from the inferior mediastinum. During ventilation, the predominant changes in thoracic diameter occur in the anteroposterior direction in the upper thoracic region and in the lateral or transverse direction in the lower thorax.

Muscles of Ventilation

Work of breathing is the energy expenditure of ventilatory muscles. Like other skeletal muscles, the ventilatory muscles are endurance muscles that are subject to fatigue from inadequate oxygen delivery, poor nutrition, increased work secondary to chronic obstructive pulmonary disease (COPD) with gas trapping, or increased airway resistance. The ventilatory muscles include the diaphragm, intercostal muscles, abdominal muscles, cervical strap muscles, sternocleidomastoid muscles, and the large back and intervertebral muscles of the shoulder girdle. During breathing the diaphragm performs most of the muscle work. Work contribution from the intercostal muscles is minor. Normally, at rest, inspiration requires work and exhalation is passive. As work of breathing increases, abdominal muscles assist with rib depression and increase intra-abdominal pressure to facilitate forced exhalation causing the “stitch” athletes experience when they actively exhale. With further increases in work, the cervical strap muscles help elevate the sternum and upper portions of the chest. Finally, during periods of maximal work, the large back and paravertebral muscles of the shoulder girdle contribute to ventilatory effort. The muscles of the abdominal wall, the most powerful muscles of expiration, are important for expulsive efforts such as coughing.1 With normal lungs, both breathing and coughing can be performed solely by the diaphragm.

Breathing is an endurance phenomenon involving fatigue-resistant muscle fibers, characterized by a slow-twitch response to electrical stimulation that must create sufficient force to lift the ribs and generate subatmospheric pressure in the intrapleural space. These fatigue-resistant fibers comprise approximately 50% of the total diaphragmatic muscle fibers. The high oxidative capacity of theses fibers creates endurance units.2 Fast-twitch muscle fibers, more susceptible to fatigue, have rapid responses to electrical stimulation imparting strength and allowing greater force over less time. The combination of fast-twitch fibers useful during brief maximal ventilatory effort periods (coughing, sneezing) and slow-twitch fibers providing endurance (breathing without rest) belie its unique duplicitous function as a muscle.3

A working muscle like the diaphragm must be firmly anchored at both its origin and insertion. However, its unique insertion is mobile—a central tendon originates from fibers attached to the vertebral bodies as well as the lower ribs and sternum. Diaphragmatic contraction results in descent of the diaphragmatic dome and expansion of the thoracic base creating decreases in intrathoracic and intrapleural pressure and an increase in intra-abdominal pressure.

The cervical strap muscles, active even during breathing at rest, are the most important inspiratory accessory muscles. When diaphragm function is impaired, as in patients with cervical spinal cord transection, they can become the primary inspiratory muscles.

Lung Structures

In an intact respiratory system, the expandable lung tissue fills the pleural cavity. The visceral and parietal pleurae oppose each other, creating a potential intrapleural space where pressure decreases when the diaphragm descends and the rib cage expands. At the end of inspiration, the resultant subatmospheric intrapleural pressure is a reflection of the opposing and equal forces between the tendency of the lung to collapse and the chest wall musculature to remain expanded to create subatmospheric pleural pressure. These equal and opposing forces at end inspiration result in the functional residual capacity (FRC), the volume of gas in the lungs at passive end expiration. At FRC the intrapleural space normally has a slightly subambient pressure (-2 to -3 mm Hg). Major divisions of the right and left lung are listed in Table 11-1. Knowledge of the bronchopulmonary segments is important for localizing lung pathology, interpreting lung radiographs, identifying lung regions during bronchoscopy, and operating on the lung. Each bronchopulmonary segment is separated from its adjacent segments by well-defined connective tissue planes that often anatomically confine initial primary lung pathologies.

The lung parenchyma can be subdivided into three airway categories based on functional lung anatomy (Table 11-2). The conductive airways allow or conduct basic gas transport without gas exchange. The next group of airways, which have smaller diameters, are transitional airways. Transitional airways are not only conduits for gas movement, but also allow limited gas diffusion and exchange. Finally, the primary function of the smallest respiratory airways is gas exchange.

Conventionally, large airways with diameters of >2 mm create 90% of total airway resistance. The number of alveoli increases progressively from approximately 24 million at birth to the final adult count of 300 million at the age of 8 or 9 years. These alveoli are associated with about 250 million


precapillaries and 280 billion capillary segments, resulting in a surface area of ~70 m2 for gas exchange.

Table 11-1 Major Divisions of the Lung









Medial basal
Lateral basal
Anterior basal
Posterior basal



Apical posterior




Posterior basal
Anteromedial basal
Lateral basal

Conductive Airways

In the adult, the trachea is a fibromuscular tube ~10 to 12 cm long with an outside diameter of ~20 mm. Structural support is provided by 20 U-shaped hyaline cartilages, with the open part of the U facing posteriorly. The cricoid membrane tethers the trachea to the cricoid cartilage at the level of the sixth cervical vertebral body. The trachea enters the superior mediastinum and bifurcates at the sternal angle (the lower border of the fourth thoracic vertebral body). Normally, half of the trachea is intrathoracic and half is extrathoracic. Because both ends of the trachea are attached to mobile structures, the adult carina can move superiorly as much as 5 cm from its normal resting position. Awareness of airway “motion” is essential to proper care of the intubated patient. In the adult, the tip of an orotracheal tube moves an average of 3.8 cm with flexion and extension of the neck but can travel as far as 6.4 cm.4 In infants and children, tracheal tube movement with respect to the trachea is even more critical: displacement of even 1 cm can result in unintentional extubation or bronchial intubation.

The next airway generation is composed of the right and left main stem bronchi. The diameter of the right bronchus is generally greater than that of the left. In the adult, the right bronchus leaves the trachea at ~25 degrees from the vertical tracheal axis, whereas the angle of the left bronchus is ~45 degrees. Thus, inadvertent endobronchial intubation or aspiration of foreign material is more likely to occur on the right than the left. Furthermore, the right upper lobe bronchus dives almost directly posterior at ~90 degrees from the right main bronchus facilitating aspiration of foreign bodies and fluid into the right upper lobe in the supine patient. In children younger than 3 years of age, the angles created by the right and left main stem bronchi are approximately equal, with takeoff angles of about 55 degrees.

The adult right main bronchus is ~2.5 cm long before it initially branches into lobar bronchi. However, in 10% of adults, the right upper lobe bronchus departs from the right main stem bronchus <2.5 cm from the carina. Furthermore, in ~2 to 3% of adults, the right upper lobe bronchus opens into the trachea, superior to the carina. Patients with these anomalies require special consideration when placing double-lumen tracheal tubes, especially if one contemplates inserting a right-sided endobronchial tube. After the right upper and middle lobe bronchi divide from the right main bronchus, the main channel becomes the right lower lobe bronchus.

The left main bronchus is ~5 cm long before its initial branching point to the left upper lobe and the lingula; it then continues as the left lower lobe bronchus.

The bronchioles, typically 1 mm in diameter, are devoid of cartilaginous support and have the highest proportion of smooth muscle in their walls. Of the three to four bronchiolar generations, the final generation is the terminal bronchiole, which is the last airway component incapable of gas exchange.

Transitional Airways

The respiratory bronchiole, which follows the terminal bronchiole, is the first site in the tracheobronchial tree where gas exchange occurs. In adults, two or three generations of respiratory bronchioles lead to alveolar ducts, of which there are four to five generations, each with multiple openings into alveolar sacs. The final divisions of alveolar ducts terminate in alveolar sacs that open into alveolar clusters.

Respiratory Airways and the Alveolar–Capillary Membrane

The alveolar-capillary membrane has two primary functions: transport of respiratory gases (oxygen and carbon dioxide) and the production of a wide variety of local and humoral substances. Gas transport is facilitated by the pulmonary capillary beds that logically are the densest capillary networks in the body. This extensive vascular branching system starts with pulmonary arterioles in the region of the respiratory bronchioles. Each alveolus is closely associated with ~1,000 short capillary segments.

The alveolar-capillary interface is complicated but well designed to facilitate gas exchange. Viewed with electron microscopy, the alveolar wall consists of a thin capillary epithelial cell, a basement membrane, a pulmonary capillary endothelial cell, and a surfactant lining layer. The flattened, squamous type I alveolar cells cover ~80% of the alveolar surface. Type I cells contain flattened nuclei and extremely thin


cytoplasmic extensions that provide the surface for gas exchange. Type I cells are highly differentiated and metabolically limited, which makes them highly susceptible to injury. When type I cells are damaged severely (during acute lung injury or adult respiratory distress syndrome), type II cells replicate and modify to form new type I cells.5

Table 11-2 Functional Airway Divisions





Bulk gas movement

Trachea to terminal bronchioles


Bulk gas movement
Limited gas exchange

Respiratory bronchioles
Alveolar ducts


Gas exchange

Alveolar sacs

Type II alveolar cells are interspersed among type I cells, primarily at alveolar–septal junctions. These polygonal cells have vast metabolic and enzymatic activity and manufacture surfactant. The enzymatic activity required to produce surfactant is only 50% of the total enzymatic activity present in type II alveolar cells.6 The remaining enzymatic activity modulates local electrolyte balance, as well as endothelial and lymphatic cell functions. Both type I and type II alveolar cells have tight intracellular junctions, providing a relatively impermeable barrier to fluids.

Type III alveolar cells, alveolar macrophages, are an important element of immunologic lung defense. Their migratory and phagocytic activities permit ingestion of foreign materials within alveolar spaces.7 Although functional pulmonary macrophages reduce the incidence of lung infection,8 they are also an integral part of the organwide pulmonary inflammatory response. Thus, it is highly controversial whether the presence of these cells is beneficial (reducing the sequelae of infection) or harmful (contributing to the inflammatory response).9

Pulmonary Vascular Systems

Two major circulatory systems supply blood to the lungs: the pulmonary and bronchial vascular networks. The pulmonary vascular system delivers mixed-venous blood from the right ventricle to the pulmonary capillary bed via two pulmonary arteries. After gas exchange occurs in the pulmonary capillary bed, blood is returned to the left atrium via four pulmonary veins. The pulmonary veins run independently along the intralobar connective tissue planes. The pulmonary capillary system adequately provides for the metabolic and oxygen needs of the alveolar parenchyma. The bronchial arterial system provides oxygen to the conductive airways and pulmonary vessels. Anatomic connections between the bronchial and pulmonary venous circulations create an absolute shunt of ~2 to 5% of the total cardiac output, and represents “normal” shunt.

Lung Mechanics

Lung movement occurs secondary to forces external to the lungs. During spontaneous ventilation, the external forces are produced by ventilatory muscles. The lungs' response is governed by two main categories: ease of elastic recoil of the chest wall and by resistance to gas flow within airways.

Elastic Work

The natural tendency of the lungs is to collapse because of elastic recoil; thus, expiration at rest is normally passive as gas flows out of the lungs. The thoracic cage exerts an outward-directed force and the lungs exert an inward-directed force, and because the outward force of the thoracic cage exceeds the inward force of the lung, the overall tendency of the lung within the thoracic cage is to remain inflated. FRC represents the gas volume in the lungs when the outward and inward forces on the lung are equal. Gravitational forces create a more subatmospheric pressure in nondependent areas of the lung than in dependent areas. In the upright adult, the difference in intrapleural pressure from the top to the bottom of the lung is ~7 cm H2O.

Surface tension at an air–fluid interface produces forces that tend to further reduce the area of interface. For a bubble to remain inflated the gas pressure within a bubble that is contained by surface tension must be higher than the surrounding gas pressure. Alveoli resemble bubbles in this respect, but unlike a bubble, alveolar gas communicates with the atmosphere via the airways. The Laplace equation describes this phenomenon: P = 2T/R, where P is the pressure within the bubble (dyne · cm-2), T is the surface tension of the liquid (dyne · cm-1), and R is the radius of the bubble (in centimeters).

During inspiration, the surface tension of the liquid in the lung increases to 40 mN/m, a value close to that of plasma. During expiration, this surface tension falls to 19 mN/m, a value lower than that of most other fluids. This change in surface tension creates hysteresis of the alveoli, different pressure-volume relationships of the alveoli during inspiration and expiration. Unlike a bubble, the pressure within an alveolus decreases as the radius of curvature decreases, creating gas flow from larger to smaller alveoli, which maintains structural stability and prevents lung collapse.

The alveolar transmural pressure gradient, or transpulmonary pressure, is the difference between intrapleural and alveolar pressure and is directly proportional to lung volume. Intrapleural pressure can be safely measured with a percutaneously inserted catheter10; however, clinicians rarely perform this technique. When measured with an esophageal balloon in the midesophagus, esophageal pressure can be used as a reflection of intrapleural pressure.11 Commercially available esophageal pressure monitors increase the ease and accuracy of measuring esophageal pressure as a reflection of intrapleural pressure.12 These monitors are useful for estimating the elastic work performed by the patient during spontaneous ventilation, mechanical ventilation, or a combination of spontaneous and mechanical ventilation. By estimating intrapleural pressure on a real-time basis, it is possible to quantitate the patient's work of breathing and changes secondary to intervention. Low levels of inspiratory pressure support can compensate for the work of breathing imposed by the endotracheal tube.13

Physiologic work of breathing includes elastic work (inspiratory work required to overcome the elastic recoil of the pulmonary system) and resistive work (work to overcome resistance to gas flow in the airway). For a patient in whom breathing apparatus is employed, the concept of total work of breathing encompasses physiologic work plus equipment-imposed ventilatory work to overcome the resistance imposed by the breathing apparatus; such as an endotracheal tube or a ventilator demand valve.

If the lungs are slowly inflated and deflated, the pressure– volume curve during inflation differs from that obtained during deflation. The two curves form a hysteresis loop that becomes progressively broader as the tidal volume is increased (Fig. 11-1). To inflate the lungs, pressure greater than the recoil pressure of deflation is needed, which means that the lung accepts deformation poorly and, once deformed, reforms to its original shape slowly. Elastic hysteresis is important for the maintenance of normal lung compliance but is not clinically significant.

The sum of the pressure-volume relationships of the thorax and lung results in a sigmoidal curve (Fig. 11-2). The vertical line drawn at end expiration coincides with FRC. Normally, humans breathe on the steepest part of the sigmoidal curve, where compliance (ΔVP) or slope is highest. In restrictive pulmonary diseases, the compliance curve shifts to the right, has decreased slope, or both. This decreased lung compliance results in smaller FRCs. When lung compliance is reduced, larger changes in intrapleural pressure are needed to create the same tidal volume; that is, the thorax has to work harder to get the same volume of gas into the lungs. The body, being an energy-conserving organism, prefers to move less gas with each breath rather than working harder to achieve the same tidal volume. Thus, patients with restrictive lung disease typically breathe with smaller tidal volumes at more rapid rates, making the spontaneous ventilatory rate one of the most sensitive indices of


lung compliance. When lung compliance is decreased, continuous positive airway pressure (CPAP) will shift the vertical line to the right, allowing the patient to breathe on a steeper, more efficient portion of the volume–pressure curve, resulting in a slower ventilatory rate with a larger tidal volume.

Figure 11-1. Dynamic pressure–volume loop of resting tidal volume. Quiet, normal breathing is characterized by hysteresis of the pressure–volume loop. The lung is more resistant to deformation than expected and returns to its original configuration less easily than expected. The slope of the line connecting the zenith and nadir lung volumes is lung compliance, ~500 mL/3 cm H2O = 167 mL/cm H2O.

At the other end of the spectrum, patients with diseases that increase lung compliance expend less elastic work to inspire but have decreased elastic recoil creating larger than normal FRC (gas trapping), and their pressure–volume curves shift to the left and steepen. Chronic obstructive lung disease and acute asthma are the most common examples of diseases with high lung compliance. If lung compliance and FRC are sufficiently high that elastic recoil is minimal, the patient must use ventilatory muscles to actively exhale. The difficulty these patients experience in emptying the lungs is compounded by the increased airway resistance.

Figure 11-2. Pulmonary pressure–volume relationships at different values of total lung capacity (TLC), ignoring hysteresis. The solid line depicts the normal pulmonary pressure–volume relationships. Humans normally breathe on the linear, steep part of this sigmoidal curve, where the slope, which is equal to compliance, is greatest. The vertical line at zero defines functional residual capacity (FRC), regardless of the position of the curve on the graph. Mild restrictive lung disease, indicated by the dashed line, shifts the curve to the right with little change in slope. However, with restrictive disease, the patient breathes on a lower FRC, at a point on the curve where the slope is less. Severe restrictive pulmonary disease profoundly depresses the FRC and diminishes the slope of the entire curve (dashed-dotted line). Obstructive disease (dotted line) elevates both FRC and compliance.

Both compliance and inspiratory elastic work can be measured for a single breath by measuring airway (Paw), intrapleural (Ppl) pressures, and tidal volume. If esophageal pressure is measured correctly, the esophageal pressure values can be substituted for Ppl values. Lung compliance, CL, the slope of the volume–pressure curve, is given by the equation:

where PL is transpulmonary pressure, PLi and PLe are transpulmonary pressure at end-inspiratory and end-expiratory, VT is tidal volume, Pawe and Pawi are expiratory and inspiratory airway pressures, and Pple and Ppli are expiratory and inspiratory intrapleural pressures.

Elastic work (Wel) is performed during inspiration only because expiration is passive during normal breathing. The area within the triangle in Figure 11-2 describes the work required to inspire. The equation that yields elastic work (and the area of the triangle) is:

Resistance to Gas Flow

Both laminar and turbulent flows exist within the respiratory tract, usually in mixed patterns. The physics of each, however, is significantly different and worth consideration.

Laminar Flow

Below critical flow rates that create turbulent flow, gas proceeds through a straight tube as a series of concentric cylinders that slide over one another. Fully developed flow has a parabolic profile with a velocity of zero at the cylinder wall and a maximum velocity at the center of the advancing “cone.” This type of streamlined flow is usually inaudible. The advancing conical front means that some fresh gas reaches the end of the tube before the tube has been completely filled with fresh gas. Thus, laminar flow in the airways results in alveolar ventilation which can occur even when the tidal volume (VT) is less than anatomic dead space. This phenomenon certainly has significant clinical implications, and as noted by Rohrer14 in 1915 it allows high-frequency ventilation to achieve adequate alveolar ventilation.

Resistance to laminar gas flows in a straight, unbranched tube can be calculated by the following equation:

where PB and PA are barometric and alveolar pressures. It is essential to note that as radius decreases in narrowed airways, resistance will increase by a power of four. Viscosity is the only physical gas property that is relevant under conditions of laminar flow. Helium has a low density, but its viscosity is close to that of air. Therefore, helium will not improve gas flow if the flow is laminar. Flow is usually turbulent when there is critical airway narrowing or abnormally high airway resistance, making low-density helium useful therapy (see next section).

Turbulent Flow

High flow rates, particularly through branched or irregularly shaped tubes, disrupt the orderly flow of laminar gas. When


resistance to gas flow is significant, turbulent flow occurs and is usually audible. Turbulent flow usually presents with a square front so fresh gas will not reach the end of the tube until the amount of gas entering the tube is almost equal to the volume of the tube. Thus, turbulent flow effectively purges the contents of a tube. Four conditions that will change laminar flow to turbulent flow are high gas flows, sharp angles within the tube, branching in the tube, and a decrease in the tube's diameter. During laminar flow, resistance is inversely proportional to gas flow rate. Conversely, during turbulent flow, resistance increases significantly in proportion to the flow rate. A detailed description of these phenomena is beyond the scope of this chapter, but the reader is referred to descriptions by Nunn.15

Increased Airway Resistance

Bronchiolar smooth muscle hyperreactivity (true bronchospasm), mucosal edema, mucous plugging, epithelial desquamation, tumors, and foreign bodies all increase airway resistance. The conscious subject can detect small increases in inspiratory resistance.16 The normal response to increased inspiratory resistance is increased inspiratory muscle effort, with little change in FRC.17 Emphysematous patients retain remarkable ability to preserve an adequate alveolar ventilation, even with gross airway obstruction. In patients with preoperative values of forced expiratory volume in the first second of expiration (FEV1) that are <1 L, Paco2 is normal in most patients. Furthermore, asthmatic patients compensate well for increased airway resistance and also keep the mean Paco2 in the lower end of normal range.18 Thus, an increased Paco2 in the setting of increased airway resistance warrants serious attention as it may signal that the patient's compensatory mechanisms are nearly exhausted. Mild expiratory resistance does not result in muscle use for active exhalation in conscious or anesthetized subjects. Instead, the initial work to overcome expiratory resistance is performed by augmenting inspiratory force until a sufficiently high lung volume is achieved that allows elastic recoil to overcome expiratory resistance.19 When expiratory resistance becomes excessive, accessory muscles are used to force gas from the lungs. During acute increases in expiratory resistance, this response can be well tolerated by most patients. However, chronic use of accessory muscles to exhale significantly increases the risk of ventilatory failure if work of breathing is further increased. When work of breathing exceeds physiologic reserves, work of breathing becomes detrimental to physiologic homeostasis and impending ventilatory failure secondary to ventilatory muscle fatigue becomes acute ventilatory failure evidenced by an acute increase in arterial carbon dioxide. Commonly, this is precipitated by pneumonia or heart failure.

Physiologic Changes in Respiratory Function Associated With Aging

Physiologic aging of the lung is associated with dilation of the alveoli, enlargement of the airspaces, decrease in exchange surface area, and loss of supporting tissue.20 Changes in the aging lung and chest wall result in decreased lung recoil (elastance), creating an increased residual volume and FRC. Additionally, compliance of the chest wall diminishes, thereby increasing the work of breathing compared with younger subjects. Respiratory muscle strength decreases with aging and is strongly correlated with nutritional status and cardiac index. Expiratory flow rates decrease with a flow–volume curve suggestive of small airway resistance. Despite these changes, the respiratory system is normally able to maintain adequate gas exchange at rest and during exertion throughout life, with only modest decrements in Pao2 and no change in Paco2. With aging, respiratory centers in the nervous system demonstrate decreased sensitivity to hypoxemia and hypercapnia, resulting in a blunted ventilatory response when challenged by heart failure, airway obstruction, or pneumonia.

Control of Ventilation

Mechanisms that control ventilation are extremely complex, requiring integration with many parts of the central and peripheral nervous systems (Fig. 11-3). LeGallois,21 who localized the respiratory centers in the brainstem in 1812, demonstrated that breathing does not depend on an intact cerebrum. Rather, breathing depends on a small region of the medulla near the origin of the vagus nerves. Countless studies in the past two centuries have greatly increased our knowledge and understanding of the anatomic components of ventilatory control. However, experimental work performed in animals is difficult to apply to humans because of interspecies variation.


Breathing, ventilation, and respiration are often used interchangeably. However, it is important to realize that these terms have distinct meanings. The term breathing refers to the act of inspiring and exhaling, which requires energy utilization for muscle work and thus is limited by energy reserves. Ventilation, on the other hand, is the movement of gas in and out of the lungs. When spontaneous, ventilation requires energy for muscle work and is, thus, breathing. Respiration occurs when energy is released from organic molecules. Such energy release depends on the movement of gas molecules such as carbon dioxide and oxygen across membranes, whether alveolar or mitochondrial. Thus, humans breathe to ventilate and ventilate to respire. Despite what appears to be clear distinctions in terminology, vernacular use of these terms are often confused in daily dialog. For example:respirators are used to treat those who have succumbed to respiratory arrest and do not have a respiratory rate, and residents are sometimes advised to breathe down a patient using potent anesthetic agent.

Generation of Ventilatory Pattern

Refer to Table 11-3 for definitions of terms used in this section. A respiratory center is a specific area in the brain that integrates neural traffic, resulting in spontaneous ventilation. Within the pontine and medullary reticular formations, there are several discrete respiratory centers that function as the control system (see Fig. 11-3).

Initial descriptions of brainstem respiratory functions are based on classic ablation and electrical stimulation studies. Another method for localizing respiratory centers entails recording action potentials from different areas of the brainstem with microelectrodes. This method is based on the assumption that local brain activity that occurs in phase with respiratory activity is evidence that the area under study has “respiratory neurons.”22 These techniques are imperfect for precisely localizing discrete respiratory centers.

Medullary Centers

The medulla oblongata contains the most basic ventilatory control centers in the brain. Specific medullary areas are primarily active during inspiration or during expiration, with many neural inspiratory or expiratory interconnections. The inspiratory centers that reside in the dorsal respiratory group


(DRG) are located in the dorsal medullary reticular formation. The DRG is the source of elementary ventilatory rhythmicity23,24 and serves as the “pacemaker” for the respiratory system.25 Whereas resting lung volume occurs at end expiration, the electrical activity of the ventilatory centers is at rest at end inspiration. The rhythmic activity of the DRG persists even when all incoming peripheral and interconnecting nerves are sectioned or blocked completely. Isolating the DRG in this manner results in ataxic, gasping ventilation with frequent maximum inspiratory efforts: apneustic breathing.

Figure 11-3. Classic central nervous system (CNS) respiratory centers. Diagram illustrates major respiratory centers, neurofeedback circuits, primary neurohumoral sensory inputs, and mechanical outputs.

The ventral respiratory group (VRG), which is located in the ventral medullary reticular formation, serves as the expiratory coordinating center. The inspiratory and expiratory neurons function by a system of reciprocal innervation, or negative feedback.22 When the DRG creates an impulse to inspire, inspiration occurs and the DRG impulse is quenched by a reciprocating VRG impulse. This VRG transmission prohibits further use of the inspiratory muscles, thus allowing passive expiration to occur.

Pontine Centers

The pontine centers process information that originates in the medulla. The apneustic center is located in the middle or lower pons. With activation, this center sends impulses to inspiratory DRG neurons and is designed to sustain inspiration. Electrical stimulation of this area results in inspiratory spasm.26 The middle and lower pons contain specific areas for phase-spanning neurons.27 These neurons assist with the transition between inspiration and expiration, and do not exert direct control over ventilatory muscles.

The pneumotaxic respiratory center is in the rostral pons. A simple transection through the brainstem that isolates this portion of the pons from the upper brainstem reduces ventilatory


rate and increases tidal volume. If both vagus nerves are additionally transected, apneusis results.28 Thus, the primary function of the pneumotaxic center is to limit the depth of inspiration. When maximally activated, the pneumotaxic center secondarily increases ventilatory frequency. However, the pneumotaxic center performs no pacemaking function and has no intrinsic rhythmicity.

Table 11-3 Definition of Respiratory Pattern Terminology




“Good breathing”: continuous inspiratory and expiratory movement without interruption


“No breathing”: cessation of ventilatory effort at passive end-expiration (lung volume = FRC)


Cessation of ventilatory effort with lungs filled at TLC

Apneustic ventilation

Apneusis with periodic expiratory spasms


Ventilatory gasps interposed between periods of ventilation apnea; also “agonal ventilation”

FRC, functional residual capacity; TLC, total lung capacity.

Higher Respiratory Centers

Many higher brain structures clearly affect ventilatory control processes. In the midbrain, stimulation of the reticular activating system increases the rate and amplitude of ventilation.29 The cerebral cortex also affects breathing pattern, although precise neural pathways are not known. Occasionally, the ventilatory control process becomes subservient to other regulatory centers. For example, the respiratory system plays an important role in the control of body temperature because it supplies a large surface area for heat exchange. This is especially important in animals in which panting is a primary means of dissipating heat. Thus, the, ventilatory pattern is influenced by neural input from descending pathways from the anterior and posterior hypothalamus to the pneumotaxic center of the upper pons.

Vasomotor control and certain respiratory responses are closely linked. Stimulation of the carotid sinus not only decreases vasomotor tone, but also inhibits ventilation. Alternatively, stimulation of the carotid body chemoreceptors (see “Chemical Control of Ventilation”) results in an increase in both ventilatory activity and vasomotor tone.

Reflex Control of Ventilation

Reflexes that directly influence ventilatory pattern usually do so to prevent airway obstruction. Deglutition, or swallowing, involves the glossopharyngeal and vagus nerves. Stimulation of the anterior and posterior pharyngeal pillars of the posterior pharynx induces swallowing. During swallowing, inspiration ceases momentarily, is usually followed by a single large breath, and briefly increases ventilation.

Vomiting significantly modifies normal ventilatory activity.30 Swallowing, salivation, gastrointestinal reflexes, rhythmic spasmodic ventilatory movements, and significant diaphragmatic and abdominal muscular activity must be coordinated over a very brief interval. Because of the obvious risk of aspirating gastric contents, it is advantageous to inhibit inspiration during vomiting. Input into the respiratory centers occurs from both cranial and spinal cord nerves.

Coughing results from stimulation of the tracheal subepithelium, especially along the posterior tracheal wall and carina.31 Coughing also requires coordination of both airway and ventilatory muscle activity. An effective cough requires deep inspiration and then forced exhalation against a momentarily closed glottis to increase intrathoracic pressure, thus allowing an expulsive expiratory maneuver.

Proprioception in the pulmonary system, the qualitative knowledge of the gas volume within the lungs, probably arises from smooth muscle spindle receptors. These proprioceptors, which are located within the smooth muscle of all airways, are sensitive to pressure changes. Airway stretch reflexes can be demonstrated during distention of isolated airways so airway pressure, rather than volume distention, appears to be the primary stimulation.32 Clinical conditions in which pulmonary airway stretch receptors are stimulated include pulmonary edema and atelectasis.

Golgi tendon organs (tendon spindles), which occur in series arrangements within ventilatory muscles, facilitate proprioception. The intercostal muscles are rich in tendon spindles, whereas the diaphragm has a limited number. Thus, the pulmonary stretch reflex primarily involves the intercostal muscles but not the diaphragm. When the lungs are full and the chest wall is stretched, these receptors send signals to the brainstem that inhibit further inspiration.

In 1868, Hering and Breuer33 reported that lightly anesthetized, spontaneously breathing animals would cease or decrease ventilatory effort during sustained lung distention. This response was blocked by bilateral vagotomy. The HeringBreuer reflex is prominent in lower-order mammals and is sufficiently active in lower mammals that even 5 cm H2O CPAP will induce apnea. In humans, however, the reflex is only weakly present, as evidenced by the fact that humans will continue to breathe spontaneously with CPAP in excess of 40 cm H2O.

Chemical Control of Ventilation

Peripheral Chemoreceptors

In a simplistic view of chemical ventilatory control, the peripheral chemoreceptors respond primarily to lack of oxygen, and the central nervous system (CNS) receptors respond primarily to changes in PCO2, pH, and acid-base disturbances.

The peripheral chemoreceptors are composed of the carotid and aortic bodies. The carotid bodies, located at the bifurcation of the common carotid artery, have predominantly ventilatory effects. The aortic bodies, which are scattered about the aortic arch and its branches, have predominantly circulatory effects. The neural output from the carotid body reaches the central respiratory centers via the afferent glossopharyngeal nerves. Output from the aortic bodies travels to the medullary centers via the vagus nerve. Both carotid and aortic bodies are stimulated by decreased Pao2, but not by decreased Sao2 or Cao2. When Pao2 falls to <100 mm Hg, neural activity from these receptors begins to increase. However, it is not until the Pao2 reaches 60 to 65 mm Hg that neural activity increases sufficiently to substantially augment minute ventilation. Thus, patients who depend on hypoxic ventilatory drive have Pao2 values in the middle 60s. Once these patients' Pao2 values exceed 60 to 65 mm Hg, ventilatory drive diminishes and Pao2 falls until ventilation is again stimulated by arterial hypoxemia. Thus, during withdrawal of mechanical ventilatory support in the patient who depends on hypoxic ventilatory drive, the Pao2 must fall to <65 mm Hg for spontaneous ventilation to resume (see Chapter 56.)

The carotid bodies are also sensitive to decreased pHa, but this response is minor. Similarly, changes in Paco2 do not stimulate these receptors sufficiently to alter minute ventilation. Increases in blood temperature, hypoperfusion of the carotid bodies themselves, and some chemicals will stimulate these receptors. Sympathetic ganglion stimulation by nicotine or acetylcholine will stimulate the carotid and aortic bodies; this effect is blocked by hexamethonium. Blockade of the cytochrome electron transport system by cyanide will prevent oxidative metabolism and will also stimulate these receptors.

Ventilatory effects resulting from stimulation of these receptors are increased ventilatory rate and tidal volume. Hemodynamic changes resulting from stimulation of these receptors include bradycardia, hypertension, increases in bronchiolar tone, and increases in adrenal secretion. The carotid body chemical receptors have been termed ultimum moriens (“last to die”). Although the response of peripheral receptors to hypoxemia was formerly believed to be resistant to the influences of anesthesia, potent inhaled anesthetics appear to depress hypoxic ventilatory response by depressing carotid body response to hypoxemia.34 The response of the peripheral receptors is not sufficiently robust to reliably increase ventilatory rate or minute ventilation to herald the onset of arterial hypoxemia during general anesthesia or


recovery from anesthesia. Furthermore, flumazenil, in a 1-mg intravenous dose, only partially reversed the diazepam-induced depression of hypoxic ventilatory drive.35 The data of Mora et al.35 further suggest that humans may develop tolerance to respiratory depressant effects of diazepam.

Central Chemoreceptors

Approximately 80% of the ventilatory response to inhaled carbon dioxide originates in the central medullary centers. Acid-base regulation involving carbon dioxide, H+, and bicarbonate is related primarily to chemosensitive receptors located in the medulla close to or in contact with the cerebrospinal fluid (CSF). The chemosensitive areas of the brainstem are in the inferolateral aspects of the medulla near the origin of cranial nerves IX and X. The area just beneath the surface of the ventral medulla is exquisitely sensitive to the extracellular fluid H+ concentration.36 Although the central response is the major factor in the regulation of breathing by carbon dioxide, carbon dioxide has little direct stimulating effect on these chemosensitive areas. These receptors are primarily sensitive to changes in H+ concentration. Carbon dioxide has a potent but indirect effect by reacting with water to form carbonic acid, which dissociates into hydrogen and bicarbonate ions.37

An acute increase in Paco2 is a more potent ventilatory stimulus than an acute increase in arterial protons concentration from a metabolic source. Carbon dioxide, but not H+, passes readily through the blood–brain and blood–CSF barriers. Local buffering systems immediately neutralize H+ in arterial blood and body fluids. In contrast, the CSF has minimal buffering capacity. Thus, once carbon dioxide crosses into the CSF, H+ are created and trapped in the CSF, resulting in a CSF H+ concentration considerably greater than that found in the blood. Because carbon dioxide crosses the blood–brain barrier readily, the Paco2 values in the CSF, cerebral tissue, and jugular venous blood rise quickly and to the same degree as the Paco2, although the central values are ~10 mm Hg higher than those measured in arterial blood.

The ventilatory response to changes in Paco2 (increased VT, increased respiratory rate) is rapid and peaks within 1 to 2 minutes after an acute change in Paco2. With the same, persistent level of carbon dioxide stimulation, the resultant increase in ventilation declines over a period of several hours, probably as a result of bicarbonate ions that are actively transported from the blood into the CSF through the arachnoid villi.38 This phenomenon explains the differing effects of acute hypercapnia versus chronic hypercapnia on the CNS-mediated ventilatory response. Finally, central medullary chemoreceptors also respond to temperature change. Cold CSF (with normal pH) or local anesthetic applied to the medullary surface will depress ventilation.

Ventilatory Response to Altitude

Ventilatory response and adaptation to high altitude are good examples of the integration of peripheral and central chemoreceptor control of ventilation. The following mechanism of acclimatization was proposed by Severinghaus et al.39 in 1963 and has since been confirmed.

Following ascent from sea level to 4,000 m, acute exposure to high altitude and low PIO2 results in arterial hypoxemia. This decrease in Pao2 activates the peripheral hypoxemic ventilatory drive by stimulating the carotid and aortic bodies, and causes increased minute ventilation. As minute ventilation increases, Paco2 and CSF Pco2 decrease, causing concomitant increases in pHa and CSF pH. The alkaline shift of the CSF decreases ventilatory drive via medullary chemoreceptors, partially offsetting hypoxemic drive. A temporary equilibrium is attained within minutes, with Paco2 only 2 to 5 mm Hg less than normal and Pao2 approximately 45 mm Hg. This initially profound hypoxemia probably causes the acute respiratory distress and other associated symptoms (headache, diarrhea) associated with rapid ascent. However, the CNS is able to restore CSF pH to normal (7.326) by pumping bicarbonate ions out of the CSF over 2 to 3 days. In 2 to 3 days, CSF bicarbonate concentration decreases approximately 5 mEq/L and restores CSF pH to within 0.01 pH unit of values at sea level. Then, centrally mediated ventilatory drive returns to normal, and hypoxic drive and stimulation of peripheral receptors can proceed unopposed. Thus, after 3 days' exposure to 4,000 m altitude, ventilatory adaptation would result in a new equilibrium, with Paco2 approximately 30 mm Hg and Pao2 approximately 55 mm Hg. Following descent to sea level, the low CSF bicarbonate concentration persists for several days, and the climber “overbreathes” until CSF bicarbonate and pH values return to normal.


Most adults with normal lungs and gas exchange can hold their breath for ~1 minute when breathing room air without previously hyperventilating. After 1 minute of breath-holding under these circumstances, Pao2 decreases to ~65 to 70 mm Hg and Paco2 increases by ~12 mm Hg. In the absence of supplemental oxygen and hyperventilation, the “breakpoint” at which normal people are compelled to breathe is remarkably constant at a Paco2 of 50 mm Hg.40,41 However, if the individual breathes 100% oxygen prior to breath-holding, he or she should be able to hold his or her breath for 2 to 3 minutes, or until Paco2 rises to 60 mm Hg. Hyperventilation sufficient to reduce Paco2 to 20 mm Hg can lengthen the period of breath-holding to 3 to 4 minutes.42 Hyperventilation with 100% oxygen prior to breath-holding should extend the apneic period to 6 to 10 minutes. The Paco2 rate of rise in awake, preoxygenated adults with normal lungs who hold their breath without previous hyperventilation is 7 mm Hg/min in the first 10 seconds, 2 mm Hg/min in the next 10 seconds, and 6 mm Hg/min thereafter.41

The duration of voluntary breath-holding is directly proportional to lung volume at onset and is probably related both to oxygen stores in the alveoli and to the rate at which Paco2rises. With smaller lung volumes, the same amount of carbon dioxide is emptied into a smaller volume during the apneic period, thus increasing the carbon dioxide concentration more rapidly than occurs with larger lung volumes. Of note, apneic patients during general anesthesia actually “breath-hold” at FRC rather than at vital capacity, which would tend to accelerate the rate of rise of carbon dioxide. Despite this difference in lung volume, the rate of rise of Paco2 in apneic anesthetized patients is 12 mm Hg during the first minute and 3.5 mm Hg/min thereafter, significantly lower than in the awake state.42,43 During anesthesia, metabolic rate and carbon dioxide production are significantly less than during ambulatory wakefulness, which probably accounts for the different rates of rise in carbon dioxide levels.

Hyperventilation with room air prior to prolonged breath-holding during exercise is inadvisable. During underwater swimming after poolside hyperventilation, the urge to breathe is first stimulated by a rising Paco2. Because an increased arterial carbon dioxide tension provides the stimulus to inspire, swimmers who hyperventilate with room air prior to swimming long distances frequently lose consciousness from arterial hypoxemia before the Paco2 is sufficiently increased to stimulate the “need” to breathe.

Hyperventilation is rarely followed by an apneic period in awake humans, despite a markedly depressed Paco2. However, minute ventilation may decrease significantly. Aggressive intermittent positive-pressure breathing treatments for patients with COPD who continue to have a carbon dioxide-based ventilatory drive can depress minute ventilation sufficiently to create arterial hypoxemia if they breathe room air after cessation of therapy.44 In contrast,


even mild hyperventilation during general anesthesia will produce prolonged apneic periods.45

Quantitative Aspects of Chemical Control of Breathing

The ventilatory responses to oxygen and carbon dioxide can be assessed quantitatively. Unfortunately, the quantitative indices of hypoxemic sensitivity are not clinically useful because the normal range is wide and confounded by many environmental factors. The reader is referred to a classic discussion of the quantitative indices of hypoxemic sensitivity.46

Ventilatory responses to Paco2 changes are measured in several ways, provided that carbon dioxide production remains constant. When subjects voluntarily increase minute ventilation to a prescribed level, the Paco2 decreases hyperbolically. The plot of minute ventilation (independent variable) and Paco2 (dependent variable) is the metabolic hyperbola (Fig. 11-4). The metabolic hyperbola is cumbersome to evaluate and difficult to use clinically.

The curve more commonly used is the Paco2 ventilatory response curve (see Fig. 11-4). It describes the effect of changing Paco2 on the resultant minute ventilation. Usually, subjects inspire carbon dioxide to raise Paco2, and the effect on minute ventilation is measured. Creating these curves and observing how they change in various circumstances allow quantitative study of factors that affect the chemical carbon dioxide control of ventilation. The carbon dioxide response curve approaches linearity in the range most often encountered in life: at Paco2 values between 20 and 80 mm Hg. Once the Paco2 exceeds 80 mm Hg, the curve becomes parabolic, with its peak ventilatory response at a Paco2between 100 and 120 mm Hg. Increasing the Paco2 to higher than 100 mm Hg allows carbon dioxide to act as a ventilatory and CNS depressant, the origin of the term carbon dioxide narcosis, with 1 minimum alveolar concentration being approximately 200 mm Hg.

Figure 11-4. Carbon dioxide–ventilatory response curve. The metabolic hyperbola, curve A, is generated by varying [V with dot above]A/[Q with dot above]E and measuring changes in carbon dioxide concentration. The hyperbolic configuration makes it cumbersome for clinical use. The carbon dioxide–ventilatory response curve, B, is linear between approximately 20 and 80 mm Hg. It is generated by varying Paco2 (usually by controlling inspired carbon dioxide concentration) and measuring the resultant [V with dot above]A/[Q with dot above]E. This is the most commonly used test of ventilatory response. The slope defines “sensitivity”; the set point, or resting Paco2, occurs at the intersection of the metabolic hyperbola and the carbon dioxide–ventilatory response curve; the apneic threshold can be obtained by extrapolating the carbon dioxide–ventilatory response curve to the x-intercept. In the absence of surgical stimulation, increasing doses of potent inhaled anesthesia or opioids will shift the curve to the right and eventually depress the slope (dashed lines). Painful stimulation will reverse these changes to varying and unpredictable degrees.

The slope of the carbon dioxide response curve is considered to represent carbon dioxide sensitivity. When Paco2 reaches 100 mm Hg, carbon dioxide sensitivity is at its peak. The set point, the point of intersection of the carbon dioxide response curve and the metabolic hyperbola, defines normal resting Paco2. Extrapolation of the carbon dioxide response curve to the x-intercept (where minute ventilation is 0) defines the apneic threshold. In awake, normal adults, the apneic threshold normally occurs at a Paco2 of ~32 mm Hg, although awake adults usually continue to breathe when they achieve the apneic threshold because the sensation of apnea is disturbing. The slope of the curve is a measure of the response of the entire ventilatory mechanism to carbon dioxide stimulation.

Once Pao2 exceeds 100 mm Hg, it no longer influences the carbon dioxide response curve. When the Pao2 is between 65 and 100 mm Hg, its effect on the carbon dioxide response curve is small. However, when Pao2 falls to <65 mm Hg, the carbon dioxide response curve shifts to the left and its slope increases, probably as a result of increased ventilatory drive stimulated by the peripheral chemoreceptors. Thus, during measurements of carbon dioxide ventilatory response, the subject should breathe supplemental oxygen to prevent hypoxic ventilatory drive interference.

The carbon dioxide response curve can be generated rapidly by increasing the fraction of inspired carbon dioxide (FICO2) by requiring the subject to rebreathe exhaled gas. The results obtained with this technique are less pure because the FICO2 is not controlled.

Three clinical states result in a left shift and/or a steepened slope of the carbon dioxide response curve. These same three situations are the only causes of true hyperventilation; that is, an increase in minute ventilation such that the decreased Paco2 creates respiratory alkalemia (either primary or compensatory). The three causes of hyperventilation (enhanced carbon dioxide response) are arterial hypoxemia, metabolic acidemia, and CNS etiologies. Examples of central causes that cause hyperventilation include drug administration, intracranial hypertension, hepatic cirrhosis, and nonspecific arousal states such as anxiety and fear. Aminophylline, doxapram, salicylates, and norepinephrine stimulate ventilation independent of peripheral chemoreceptors. Opioid antagonists, given in the absence of opioids, do not stimulate ventilation. However, when given after opioid administration, they do reverse the effects of opioids on the carbon dioxide response curve.

Ventilatory depressants displace the carbon dioxide response curve to the right or decrease its slope or both. Changes in physiology that depress ventilation include metabolic alkalemia, denervation of peripheral chemoreceptors, normal sleep, and drugs. During normal sleep, the carbon dioxide response curve is displaced to the right, with the degree of displacement depending on the depth of sleep. Usually, Paco2 increases up to 10 mm Hg during deep sleep. Hypoxemic responses are not impaired by sleep, which is convenient for continued survival at high altitude while sleeping.

Opioids displace the carbon dioxide response curve to the right with little change in slope at sedative doses (see Chapter 19.) With higher, “anesthetic” doses, the curve shifts farther to the right and its slope is depressed, simulating the effect of potent inhalation agents on the carbon dioxide response curve (see Fig. 11-4). In the absence of other ventilatory-depressant drugs, opioids induce pathognomonic changes in ventilatory patterns: a decreased ventilatory rate with an increased tidal volume. Not until opioids nearly induce apnea is tidal volume decreased. Large narcotic doses usually result in apnea responsive to verbal encouragement before consciousness is lost.

Barbiturates in sedative or light hypnotic doses have little effect on the carbon dioxide response curve. However, in doses adequate to allow skin incision, barbiturates shift the carbon dioxide response curve to the right. The ventilatory pattern


resulting from barbiturate administration is characterized by decreased tidal volume and increased ventilatory rate. Potent inhaled anesthetics displace the carbon dioxide response curve to the right and decrease the slope, the degree of which depends on the anesthetic dose and the level of surgical stimulation. Like barbiturates, the ventilatory pattern following administration of potent inhaled anesthetics is initially represented by a decreased tidal volume and increased ventilatory rate. As more potent anesthetic agent is administered, however, ventilatory rate decreases toward an apneic end point. This clinical response occurs when the carbon dioxide response curve eventually becomes horizontal (slope = 0), resulting in essentially no ventilatory response to Paco2 changes.

Potent inhaled anesthetics and opioids displace the setpoint to the right, implying that the resting, steady-state Paco2 is higher and minute ventilation lower. Furthermore, when the carbon dioxide response curve shifts to the right, the apneic threshold also increases (see Fig. 11-4). Surgical stimulation reverses the ventilatory response changes induced by inhaled anesthetics and opioids, but the degree of reversal is not predictable.

Oxygen and Carbon Dioxide Transport

This chapter discusses only external respiration, in which oxygen moves from the ambient environment into the pulmonary capillaries and carbon dioxide leaves the pulmonary capillaries to enter the atmosphere. The movement of gas across the alveolar-capillary membrane depends on the integrity of the pulmonary and cardiac systems. Unless it is otherwise stated, the reader should assume the ventilation and perfusion of alveolar-capillary units are normal. Abnormal distribution of ventilation or perfusion of the lungs is discussed later (see “Ventilation-Perfusion Relationships”).

Bulk Flow of Gas (Convection)

Convection, in which all gas molecules move in the same direction, is the primary mechanism responsible for gas flow in large and most small airways, down to the bronchi and bronchiolar airways of the 14th or 15th generation. Because the cross-sectional area of the airways progressively increases as gas moves toward the lung periphery, the average velocity of gas particles decreases as they travel toward the alveoli. Because resistance depends on flow, the greatest part of airway resistance occurs in the larger airways, where gas molecules travel more quickly. During normal quiet ventilation, gas flow within convective airways is mainly laminar.

Figure 11-5. Distribution of blood flow in the isolated lung. In zone 1, alveolar pressure (PA) exceeds pulmonary artery pressure (Ppa), and no flow occurs because the vessels are collapsed. In zone 2, arterial pressure exceeds alveolar pressure, but alveolar pressure exceeds pulmonary venous pressure (Ppv). Flow in zone 2 is determined by the arterial–alveolar pressure difference (Ppa – PA), which steadily increases down the zone. In zone 3, pulmonary venous pressure exceeds alveolar pressure and flow is determined by the arterial–venous pressure difference (Ppa - Ppv), which is constant down this pulmonary zone. However, the pressure across the vessel walls increases down the zone so their caliber increases, as does flow. (From West JB, Dollery CT, Naimark A: Distribution of blood flow in isolated lung: Relation to vascular and alveolar pressures. J Appl Physiol 1964;19: 713, with permission.)

Gas Diffusion

Diffusion within a gas-filled space is random molecular motion that results in complete mixing of all gases. In the distal airways of the lung beginning with the terminal bronchioles (16th airway generation), diffusion becomes the predominant mode of gas transport. Once gas reaches the small alveolar ducts, alveolar sacs, and alveoli, both diffusion and regional ([V with dot above]A/[Q with dot above]) relationships influence gas transport. Historically, clinicians assumed defects in gas diffusion were responsible for arterial hypoxemia. However, the most frequent cause of arterial hypoxemia is physiologic shunt (see “Ventilation-Perfusion Relationships”).47

The other usage of “diffusion” refers to the passive movement of molecules across a membrane that is governed primarily by concentration gradient. In this sense, carbon dioxide is 20 times more diffusible across human membranes than is oxygen; therefore, carbon dioxide crosses alveoli easily. As a result, hypercapnia is never the result of defective diffusion; rather, it is the result of inadequate alveolar ventilation with respect to carbon dioxide production.

True diffusion defects that create arterial hypoxemia are rare. The most common reason for a measured decrease in diffusing capacity (see “Pulmonary Function Tests”) is mismatched ventilation and perfusion, which functionally results in a decreased surface area available for diffusion.

Distribution of Ventilation and Perfusion

The efficiency with which oxygen and carbon dioxide exchange at the alveolar–capillary level highly depends on the matching of capillary perfusion and alveolar ventilation. At this level, the marriage between the lung and the circulatory system must be well matched and intimate.

Distribution of Blood Flow

Blood flow within the lung is mainly gravity-dependent. Because the alveolar–capillary beds are not composed of rigid vessels, the pressure of the surrounding tissues can influence the resistance to flow through the individual capillaries. Thus, blood flow depends on the relationship between pulmonary artery pressure (Ppa), alveolar pressure (PA), and pulmonary venous pressure (Ppv; Fig. 11-5). West et al.47 and West and Dollery48 created a lung model that divides the lung into three zones. Zone 1 conditions occur in the most gravity-independent part of the lung. Because alveolar pressure is approximately equal to atmospheric pressure; and pulmonary artery pressure,


which is always in excess of pulmonary venous pressure, is subatmospheric in zone 1, then zone 1 is described by the following relationship: PA > Ppa > Ppv. In zone 1 alveolar pressure that is transmitted to the pulmonary capillaries promotes their collapse, with a consequent theoretical blood flow of zero to this lung region. Thus, zone 1 receives ventilation in the absence of perfusion. This relationship is alveolar dead space ventilation. Normally, zone 1 areas exist only to a limited extent. However, in conditions of decreased pulmonary artery pressure such as hypovolemic shock, zone 1 enlarges.

Zone 2 occurs from the lower limit of zone 1 to the upper limit of zone 3, where Ppa > Pa > Ppv. The pressure difference between pulmonary artery and alveolar pressure determines blood flow in zone 2. Pulmonary venous pressure has little influence. Well-matched ventilation and perfusion occur in zone 2, which contains the majority of alveoli.

Finally, zone 3 occurs in the most gravity-dependent areas of the lung, where Ppa > PpV > PA and blood flow is primarily governed by the pulmonary arterial to venous pressure difference. Because gravity also increases pulmonary venous pressure, the pulmonary capillaries become distended. Thus, perfusion in zone 3 is lush, resulting in capillary perfusion in excess of ventilation, or physiologic shunt.

Distribution of Ventilation

Alveolar pressure is the same throughout the lung; therefore, the more negative intrapleural pressure at the apex (or the least gravity-dependent area) results in larger, more distended apical alveoli than in other areas of the lung. The transpulmonary pressure (Paw – Ppl), or distending pressure of the lung, is greater at the top and lower at the bottom, where intrapleural pressure is less negative. Despite the smaller alveolar size, more ventilation is delivered to dependent pulmonary areas. The decrease in intrapleural pressure at the base of the lungs during inspiration is greater than at the apex because of diaphragmatic proximity. Thus, because the dependent area of the lung generates the greatest change in transpulmonary pressure, more gas is sucked into dependent areas of the lung.

Ventilation–Perfusion Relationships

As discussed previously, the majority of blood flow is distributed to the gravity-dependent part of the lung. Also, during a spontaneous breath, the largest portion of the tidal volume also reaches the gravity-dependent part of the lung. Thus, the nondependent area of the lung receives a lower proportion of both ventilation and perfusion, and dependent lung receives greater proportions of ventilation and perfusion. Nevertheless, ventilation and perfusion are not matched perfectly, and various [V with dot above]A/[Q with dot above] ratios result throughout the lung. The ideal [V with dot above]A/[Q with dot above] ratio of 1 is believed to occur at approximately the level of the third rib. Above this level, ventilation occurs slightly in excess of perfusion, whereas below the third rib the [V with dot above]A/[Q with dot above] ratio becomes less than 1 (Fig. 11-6).

In a simplified model, gas exchange units can be divided into normal ([V with dot above]A/[Q with dot above] = 1:1), dead space ([V with dot above]A/[Q with dot above] = 1:0), shunt ([V with dot above]A/[Q with dot above] = 0:1), or a silent unit ([V with dot above]A/[Q with dot above] = 0:0; Fig. 11-7). Although this model is helpful in understanding [V with dot above]A/[Q with dot above] relationships and their influences on gas exchange, [V with dot above]A/[Q with dot above] really occurs as a continuum. In the lungs of a healthy, upright, spontaneously breathing individual, the majority of alveolar-capillary units are normal gas exchange units. The [V with dot above]A/[Q with dot above] ratio varies between absolute shunt (in which [V with dot above]A/[Q with dot above] = 0) to absolute dead space (in which [V with dot above]A/[Q with dot above] = ∞). Rather than absolute shunt, most units with low [V with dot above]A/[Q with dot above] mismatch receive a small amount of ventilation relative to blood flow. Similarly, most dead space units are not absolute, but rather are characterized by low blood flow relative to ventilation. During acute lung injury and adult respiratory distress syndrome, areas of low [V with dot above]A/[Q with dot above] matching commonly lie adjacent to areas of high [V with dot above]A/[Q with dot above] matching.49 Thus, the lung zone model proposed by West and coworkers47,48 should be used to aid the understanding of pulmonary physiology and not be regarded as an incontrovertible anatomic truism.

Figure 11-6. Distribution of ventilation, blood flow, and ventilation– perfusion ratio in the normal, upright lung. Straight lines have been drawn through the ventilation and blood flow data. Because blood flow falls more rapidly than ventilation with distance up the lung, ventilation–perfusion ratio rises, slowly at first, then rapidly. (From West JB: Ventilation/Blood Flow and Gas Exchange, 4th ed. Oxford, England, Blackwell Scientific, 1985, with permission.)

Hypoxic pulmonary vasoconstriction and bronchoconstriction allow the lungs to maintain optimal [V with dot above]A/[Q with dot above] matching (see Chapter 40.) Hypoxic pulmonary vasoconstriction,


stimulated by alveolar hypoxia, severely decreases blood flow. Thus, poorly ventilated alveoli also receive minuscule blood flow. Furthermore, decreased regional pulmonary blood flow results in bronchiolar constriction and diminishes the degree of dead space ventilation.50,51 When either phenomena occurs, the shunt or dead space units effectively become silent units in which little ventilation or perfusion occurs.

Figure 11-7. Continuum of ventilation–perfusion relationships. Gas exchange is maximally effective in normal lung units and only partially effective in shunt and dead space effect units. Gas exchange is totally absent in silent units, absolute shunt, and dead space units.

Many pulmonary diseases result in both physiologic shunt and dead space abnormalities. However, most disease processes can be characterized as producing either primarily shunt or dead space in their early stages. Increases in dead space ventilation primarily affect carbon dioxide elimination and have little influence on arterial oxygenation until dead space ventilation exceeds 80 to 90% of minute ventilation ([V with dot above]E). Similarly, physiologic shunt primarily affects arterial oxygenation with little effect on carbon dioxide elimination until the physiologic shunt fraction exceeds 75 to 80% of the cardiac output. Defective to absent gas exchange can be the net effect of either abnormality in the extreme.

Physiologic Dead Space

Each inspired breath is composed of gas that contributes to alveolar ventilation (VA) and gas that becomes dead space ventilation (VD). Thus, tidal volume (VT) = VA + VD. In the normal, spontaneously breathing person, the ratio of alveolar-to-dead space ventilation for each breath is 2:1. Conveniently, the rule of “1, 2, 3” applies to normal, spontaneously breathing persons. For each breath, 1 mL/lb (lean body weight) becomes VD, 2 mL · lb-1 becomes Va, and 3 mL · lb-1 constitutes the VT.

Physiologic dead space consists of anatomic and alveolar dead space. Anatomic dead space ventilation, approximately 2 mL/kg ideal body weight, accounts for the majority of physiologic dead space. It arises from ventilation of structures that do not exchange respiratory gases: the oronasopharynx to the terminal and respiratory bronchioles. Clinical conditions that modify anatomic dead space include tracheal intubation, tracheostomy, and large lengths of ventilator tubing between the tracheal tube and the ventilator Y-piece. It is important to note that ventilation occurs because gas flows into and out of the alveoli. In contrast, the inspiratory or expiratory limb of anesthesia circle system has unidirectional flow, and therefore is not a component of anatomic dead space ventilation.

Alveolar dead space ventilation arises from ventilation of alveoli where there is little or no perfusion. Because disease produces little change in anatomic dead space, physiologic dead space is primarily influenced by changes in alveolar dead space. Rapid changes in physiologic dead space ventilation most often arise from changes in pulmonary blood flow, resulting in decreased perfusion to ventilated alveoli. The most common cause of acutely increased physiologic dead space is an abrupt decrease in cardiac output. Another pathologic condition that interferes with pulmonary blood flow, and thereby creates dead space, is pulmonary embolism, whether due to thrombus or to fat, air, or amniotic fluid. Although there may be obstruction to blood flow with some types of pulmonary emboli, the greatest decrease in pulmonary blood flow is due to vasoconstriction induced by locally released vasoactive substances such as leukotrienes.

Chronic pulmonary diseases create dead space ventilation by irreversibly changing the relationship between alveolar ventilation and blood flow; this alteration is especially prominent in patients with COPD. Furthermore, acute diseases such as adult respiratory distress syndrome can cause an increase in dead space ventilation owing to intense pulmonary vasoconstriction. Finally, therapeutic or supportive manipulations such as positive-pressure ventilation or positive airway pressure therapy can increase alveolar dead space because depressed venous return to the right heart will decrease cardiac output, which can usually be overcome by intravenous fluid administration. Occasionally, therapeutics that create intrapulmonary positive pressure may increase physiologic shunt when blood flow to a previously silent area of [V with dot above]A/[Q with dot above] matching now receives blood redistributed by positive pressure from more compliant areas of the lung.

Assessment of Physiologic Dead Space

Because the lung receives nearly 100% of the cardiac output, assessment of physiologic dead space ventilation in the acute setting yields valuable information about pulmonary blood flow and, ultimately, about cardiac output. If pulmonary blood flow decreases, the most likely cause is a decreased cardiac output. Thus, it is clinically useful to be able to readily assess the degree of physiologic dead space ventilation.

There are two easy and several difficult ways to assess dead space ventilation. A comparison of minute ventilation and Paco2 allows a gross qualitative assessment of physiologic dead space ventilation. The Paco2 is determined only by alveolar ventilation and [V with dot above]co2. If [V with dot above]co2 remains constant, Paco2 also will remain constant as long as minute ventilation supplies the same degree of alveolar ventilation. If the spontaneously breathing individual must increase minute ventilation to maintain the same Paco2, he or she has experienced an increase in dead space ventilation because less of the minute ventilation is contributing to alveolar ventilation. Alternatively, a mechanically ventilated patient with a fixed minute ventilation and no increase in [V with dot above]co2 also experiences an increased dead space ventilation if the Paco2 rises. Hence, when Paco2 in a mechanically ventilated patient increases, it is necessary to determine if the cause is increased dead space ventilation or an increased [V with dot above]co2.

Because positive pressure ventilation increases alveolar pressure, the mechanically ventilated patient with normal lungs has a dead space to alveolar ventilation ratio (VD/Va) of 1:1 (more West zone 1) rather than 1:2, as during spontaneous ventilation. If mechanical VT is 1,000 mL, 500 mL contributes to VA, and 500 mL contributes to VD. At rest, the required [V with dot above]A with normal [V with dot above]co2 is approximately 60 mL/kg/min. A 70-kg man would then require a [V with dot above]A of 4,200 mL/min. During spontaneous breathing, the required [V with dot above] E would be 6,300 mL/min, but during mechanical ventilation [V with dot above] E would have to be 8,400 mL/min. Using this calculation, if a 70-kg resting patient requires [V with dot above] E much in excess of 8,400 mL/min, either [V with dot above] D or [V with dot above] co2 is increased. A rule of thumb for mechanically ventilated patients is that doubling baseline minute ventilation decreases Paco2 from 40 to 30 mm Hg, and quadrupling minute ventilation decreases Paco2 from 40 to 20 mm Hg.

The Paco2 will be greater than or equal to end-tidal Paco2 (PETCO2) unless the patient inspires or receives exogenous carbon dioxide (e.g., from peritoneal insufflation). The difference between PETCO2 and Paco2 is because of dead space ventilation. The most common reason for an acute increase in dead space ventilation is decreased cardiac output. Measurement of this difference—which is simple, readily obtainable, and fairly inexpensive—yields reliable information relative to the degree of dead space ventilation. Clinical situations that change pulmonary blood flow sufficiently to increase dead space ventilation can be detected by comparing PETCO2 with temperature-corrected Paco2. Yamanaka and Sue52 found that the PETCO2 in ventilated patients varied linearly with the dead space to tidal volume ratio (VD/VT) and that PETCO2 correlated poorly with Paco2. Thus, in the critically ill, mechanically ventilated patient, and in anesthetized patients, monitoring PETCO2 gives far more information about ventilatory efficiency or dead space ventilation than it does about the absolute value of Paco2.


Anesthesiologists commonly measure PETCO2 to detect venous air embolism during anesthesia. A lowered cardiac output alone, in the absence of venous air embolism, may sufficiently decrease pulmonary perfusion so dead space ventilation increases and PETCO2 falls. Thus, a depressed PETCO2 is sensitive for decreased cardiac output but nonspecific pulmonary embolism. Air in the pulmonary arteries mechanically interferes with blood flow and also causes pulmonary arterial constriction, further decreasing pulmonary blood flow. A decreased PETCO2 suggests that a physiologically significant air embolism has occurred. The same physiologic considerations apply to detecting pulmonary thromboembolism.

Some clinicians use the divergence of PETCO2 from Paco2 as a reflection of pulmonary blood flow for other applications. During intentional pharmacologic or surgical manipulation of pulmonary blood flow, the difference between Paco2 and PETCO2 serves as a useful physiologic monitor of the effectiveness of these interventions. Furthermore, PETCO2 as a reflection of pulmonary perfusion is a useful tool for studying and monitoring the effectiveness of resuscitation efforts and may provide a marker for survival after resuscitation.53

The most quantitative technique used to measure physiologic dead space uses a modification of the Bohr equation:

where PĒco2 is the Pco2 from the mixture of all expired gases over the period of time during which measurements are made. This calculation estimates the fraction of each breath that does not contribute to gas exchange. In spontaneously breathing patients, normal VD/VT is between 0.2 and 0.4, or ~0.33. In patients receiving positive-pressure ventilation, VD/VT becomes ~0.5. The major limitation of performing this calculation is the difficulty in collecting exhaled gas for PĒco2 measurement. Exhaled gases, collected in cumbersome 50 L bags, can easily be contaminated with inspired air or supplemental oxygen. The measurement will also be inaccurate if the patient does not maintain a steady ventilatory pattern. Therefore, extreme care must be taken to ensure all measurements are performed accurately. In practice, this measurement is rarely performed.


Physiologic Shunt

Whereas physiologic dead space ventilation applies to areas of the lung that are ventilated but poorly perfused, physiologic shunt occurs in lung that is perfused but poorly ventilated. The physiologic shunt ([Q with dot above]SP) is that portion of the total cardiac output ([Q with dot above]T) that returns to the left heart and systemic circulation without receiving oxygen in the lung. When pulmonary blood is not exposed to alveoli or when those alveoli are devoid of ventilation, the result is absolute or true shunt, in which [V with dot above]A/[Q with dot above] = 0. Shunt effect, or venous admixture, is the more common clinical phenomenon and occurs in areas where alveolar ventilation is deficient compared with the degree of perfusion: 0 < [V with dot above]A/[Q with dot above] < 1.

Because blood passing through areas of absolute shunt receives no oxygen, arterial hypoxemia resulting from absolute shunt is minimally reversed with supplemental oxygen. Alternatively, supplemental oxygen supplied to patients with arterial hypoxemia due to venous admixture will increase the Pao2. Although ventilation to these alveoli is deficient, they do carry a small amount of oxygen to the capillary bed. Thus, assessment of arterial oxygen responsiveness to supplemental oxygen administration is a helpful diagnostic tool.

A small percentage of venous blood normally bypasses the right ventricle and empties directly into the left atrium. This anatomic, absolute, or true shunt arises from the venous return from the pleural, bronchiolar, and thebesian veins. This venous drainage accounts for 2 to 5% of total cardiac output and explains the small shunt that normally occurs. Anatomic shunts of greatest magnitude are usually associated with congenital heart disease that causes right-to-left shunt. Intrapulmonary anatomic shunts can also cause anatomic shunt. For example, the arterial hypoxemia associated with advanced hepatic failure (hepatopulmonary syndrome) is partly due to arteriovenous malformations.54,55 Diseases that may cause absolute or true shunt include acute lobar atelectasis, extensive acute lung injury, advanced pulmonary edema, and consolidated pneumonia. Disease entities that tend to produce venous admixture include mild pulmonary edema, postoperative atelectasis, and COPD.

Assessment of Arterial Oxygenation and Physiologic Shunt

The simplest assessment of oxygenation is qualitative comparison of the patient's FIO2 and Pao2. The highest possible Pao2 for any given FIO2 (and Paco2) can be calculated from the alveolar gas equation:

where PAO2 and PACO2 are alveolar Po2 and Pco2, PH2o is water vapor pressure at 100% saturation and 37°C, Pb is barometric pressure, and R is respiratory quotient. Assuming one makes the calculation for a well-perfused alveolus, the alveolar and arterial Pco2 are equal. Therefore, Paco2 can be substituted for PACO2. Respiratory quotient (R) is the ratio of O2consumed ([V with dot above] o2) to CO2 produced ([V with dot above]co2):

Oxygen tension–based indices do not reflect mixed venous contribution to arterial oxygenation and can be misleading.56 Even if venous admixture is small, mixed venous blood with very low oxygen content will magnify the effect of a small shunt. Oxygen tension–based indices, for example, Pao2/FIO2, alveolar to arterial Po2 difference (P(A-a)O2), and ratio Pao2/PAO2, do not take into account the influence of C·v_o2 on arterial oxygenation. Therefore, in critically ill patients who are hypoxemic, the insertion of a pulmonary artery catheter to assess shunt and to measure cardiac output may be essential to understanding the influence of cardiac function on arterial oxygenation.

P(A-a)O2 is a useful quantitative assessment of arterial oxygenation mainly when arterial hemoglobin is well saturated when normal DA-ao2 is <5 mm Hg. When Pao2 is <150 mm Hg (and certainly when it is <100 mm Hg), the relationship between oxygen content and oxygen tension is nonlinear, thus making DA-ao2 more difficult to interpret.

The assessment of arterial oxygenation requires, at least, knowledge of FIO2 and either Pao2 or Sao2. Oxygen tension–based indices of oxygenation are useful, but they do not take into account the contribution of mixed venous blood to arterial oxygenation. Mixed venous blood can become extremely desaturated in the critically ill patient owing to inadequate cardiac output, anemia, arterial hypoxemia, or increased [V with dot above]o2. The best knowledge of the efficiency with which the lungs oxygenate the arterial blood can be obtained only by calculating shunt fraction or ventilation–perfusion index (VQI).

Physiologic Shunt Calculation

The clinical reference standard for the calculation of physiologic shunt fraction is derived from a two-compartment pulmonary blood flow model where one compartment performs ideal gas exchange and contains perfectly married alveolar–capillary units. The other compartment is the shunt compartment and contains pulmonary capillaries that have no exposure to ventilated alveoli. Using the Fick relationship, the following equation can be derived:

where [Q with dot above] SP/[Q with dot above] T is the shunt fraction, [Q with dot above] SP is blood flow through the physiologic shunt compartment, [Q with dot above]T is total cardiac output, and Cc′o2 and Cv-O2 are end-capillary and mixed-venous oxygen contents, respectively. Normal intrapulmonary shunt is approximately 5%. Because this equation is based on an artificial two-compartment model, the absolute value is physically meaningless. A calculated [Q with dot above]SP/[Q with dot above] T of 25% means that if the lung existed in two compartments, 25% of the cardiac output would travel through the shunt compartment. Because the lung does not exist in two compartments, this equation only grossly estimates pulmonary oxygen exchange defects. Nevertheless, it remains our best tool for clinically evaluating the efficiency with which the lungs oxygenate arterial blood. Observing shunt fraction change with therapeutic intervention or with the progress of disease is more valuable than knowing the absolute value per se.

Because hemoglobin concentration is uniform throughout the vascular system, the oxygen contents in the shunt equation are determined primarily by oxyhemoglobin saturation. Thus, the shunt equation can be approximated by substituting saturation values for each term; the new value, called ventilationperfusion ratio (VQI),55 is determined as follows:

If the patient is neither breathing a hypoxic gas mixture nor has a methemoglobin or carboxyhemoglobin value in excess of 5 to 6%, Sc′o2 must equal 1 because the model requires a perfect alveolar–capillary interface. This substitution results in the final expression in the previous equation. The absolute values of VQI are meaningless, although “normal” should be 0 to 4%. Like [Q with dot above] SP/[Q with dot above] T, the importance of these values lies in their trend as disease and treatment progress.

Sao2 and S·v–o2 can be estimated continuously with pulse oximetry and by using a pulmonary artery catheter with oximetry capability. By interfacing the outputs of these two devices with a computer, VQI can be calculated continuously. The greatest advantage of calculating [Q with dot above]SP/[Q with dot above]T or VQI to assess arterial oxygenation efficiency is that these values include the contribution of mixed venous blood.

Pulmonary Function Testing

Anesthesiologists frequently care for patients with significant pulmonary dysfunction (see Chapter 23). It is important for the anesthesiologist to be able to interpret tests of pulmonary function intelligently and to know which tests will help define dysfunction if the patient's history and physical are suggestive of disease. This section discusses lung volumes, tests of pulmonary mechanics, and diffusing capacity.

Lung Volumes and Capacities

Known, reproducible pulmonary gas volumes and capacities provide a reliable basis for comparison between normal and abnormal measurements.57 Because normal measurements vary with size, height is most frequently used to define “normal.” Lung capacities are composed of two or more lung volumes. Lung volumes and capacities are schematically illustrated in Figure 11-8.

Tidal volume is the volume of gas that moves in and out of the lungs during quiet breathing and is ~6 to 8 mL/kg. Tidal volume falls with decreased lung compliance or when the patient has reduced ventilatory muscle strength.

Vital capacity is usually ~60 mL/kg but may vary as much as 20% from normal in healthy individuals. Vital capacity correlates well with the capability for deep breathing and effective coughing. It is decreased by restrictive pulmonary disease such as pulmonary edema or atelectasis. Vital capacity may also be reduced by the mechanically induced extrapulmonary restriction seen in pleural effusion, pneumothorax, pregnancy, large ascites, or ventilatory muscle weakness.

The inspiratory capacity is the largest volume of gas that can be inspired from the resting expiratory level and is frequently decreased in the presence of significant extrathoracic airway obstruction. This measurement is one of the few simple tests that can detect extrathoracic airway obstruction. Most routine pulmonary function tests measure only exhaled flows and volumes, which may be relatively unaffected by extrathoracic obstruction until it is severe. Changes in the absolute volume of inspiratory capacity usually parallel changes in vital capacity. Expiratory reserve volume is not of great diagnostic value.

Functional residual capacity (FRC) is the volume of gas remaining in the lungs at passive end expiration. Residual volume is that gas remaining within the lungs at the end of forced maximal expiration. The FRC serves two primary physiologic functions. It determines the point on the


pulmonary volume–pressure curve for resting ventilation (see Fig. 11-2). The tangent defined by the midportion pulmonary volume–pressure curve at FRC defines lung compliance. Thus, FRC determines the elastic pressure–volume relationships within the lung. Furthermore, FRC is the resting expiratory volume of the lung and is the primary determinant of oxygen reserve in humans when apnea occurs. As such, it greatly influences ventilation–perfusion relationships within the lung. When FRC is reduced, venous admixture (low [V with dot above]A/[Q with dot above]) increases and results in arterial hypoxemia (see “Oxygen and Carbon Dioxide Transport” and “Lung Mechanics”).

Figure 11-8. Lung volumes and capacities. The darkest bar on the far right depicts the four basic lung volumes that sum to create total lung capacity (TLC). Other lung capacities are composed of two or more lung volumes. The overlying spirographic tracing orients the reader to the relationship between the lung volumes and capacities and the spirogram. ERV, expiratory reserve volume; FRC, functional residual capacity; IC, inspiratory capacity; IRV, inspiratory reserve volume; RV, residual volume; VC, vital capacity; TV, tidal volume.

Further, the FRC may be used to quantify the degree of pulmonary restriction. Disease processes that reduce FRC and lung compliance include acute lung injury, pulmonary edema, pulmonary fibrotic processes, and atelectasis. Mechanical factors also reduce FRC; examples include pregnancy, obesity, pleural effusion, and posture. The FRC decreases 10% when a healthy subject lies down. Ventilatory muscle weakness or paralysis will also decrease FRC. In contrast, patients with COPD have excessively compliant lungs that recoil less forcibly. Their lungs retain an abnormally large volume at the end of passive expiration, a phenomenon called gas trapping.

Functional Residual Capacity Measurement

The FRC and residual volume must be measured indirectly because residual volume cannot be removed from the lung. The multiple-breath nitrogen washout test is performed by having the subject breathe 100% oxygen for several minutes so alveolar nitrogen is gradually “washed out.” With each breath, the volume of gas and the concentration of nitrogen in the exhaled gas are measured. A rapid nitrogen analyzer coupled to a spirometer or pneumotachometer provides a breath-by-breath analysis of nitrogen washout. Electronic signals proportional to nitrogen concentrations and exhaled volumes (or flow, if a pneumotachometer is used) are integrated to derive the exhaled volume of nitrogen for each breath. Then the values for all breaths are summed to provide a total volume of nitrogen washed out of the lungs. The test proceeds until the alveolar nitrogen concentration is reduced to <7%, usually requiring 7 to 10 minutes. FRC is calculated using the equation:

where [N2]i and [N2]f are the fractional concentrations of alveolar nitrogen at the beginning and end of the test, respectively.

Pulmonary Function Tests

Forced Vital Capacity

The forced vital capacity (FVC) is the volume of gas that can be expired as forcefully and rapidly as possible after maximal inspiration. Normally, FVC is equal to vital capacity. Because forced expiration significantly increases intrapleural pressures but changes airway pressure little, bronchiolar collapse, obstructive lesions, and gas trapping are exaggerated. Thus, FVC may be reduced in chronic obstructive diseases even when the vital capacity appears near normal. FVC is nearly always decreased by restrictive diseases. FVC values <15 mL/kg are associated with an increased incidence of PPCs, probably because patients in this condition cough ineffectively.58 FVC reduced to this level represents a profound defect, most commonly seen in quadriplegic patients or patients with severe neuromuscular disease. Finally, FVC is largely dependent on patient effort and cooperation.

Forced Expiratory Volume

FEVT is the forced expiratory volume of gas over a given time interval during the FVC maneuver. The interval, described by the subscript T, is the time elapsed in seconds from the onset of expiration. Because FEVT records a volume of gas expired over time, it is actually a measure of flow. By measuring expiratory flow at specific intervals, the severity of airway obstruction can be ascertained. Decreased FEVT values are common in both obstructive and restrictive disease patterns. The most important application of FEVT is its comparison with the patient's FVC. Normal subjects can expire at least three fourths of FVC within the first second of the forced expiratory maneuver. The FEV1, the most frequently employed value, is normally ≥75% of the FVC, or FEV1/FVC ≥ 0.75.

Normally, an individual can expire 50 to 60% of FVC in 0.5 second, 75 to 85% in 1 second, 94% in 2 seconds, and 97% in 3 seconds. Cooperative patients with obstructive disease will exhibit a reduced FEV1/FVC in most cases. However, patients with restrictive disease usually have normal FEV1/FVC ratios. The validity of the evaluation of the FEV1/FVC is highly dependent on patient cooperation and effort. It is possible to deliberately produce an artificially low FEV1/FVC.

Forced Expiratory Flow

FEF25–75% is the average forced expiratory flow during the middle half of the FEV maneuver. This test is also called maximum midexpiratory flow rate. The length of time required for a subject to expire the middle half of the FVC is divided into 50% of the FVC. The spirogram in Figure 11-9 marks the place from 25 to 75% of FVC, constituting the middle 50% of FVC. The straight line connecting the 25% and 75% volumes has a slope approximately equal to average flow. A normal value for a healthy 70-kg man is approximately 4.7 L/sec (or 280 mL/min). Normally, both the absolute value and the percentage of predicted value for the individual being studied are recorded. A normal value is 100 ± 25% of predicted value. Decreased flow rates from this middle 50% of FVC anatomically represent flow in medium-sized airways, and when decreased, there is obstructive disease of medium-sized airways. This value is typically normal in restrictive diseases. This test is fairly sensitive in the early stages of obstructive airway disease. Decreased FEV25–75% frequently will be observed


before other obstructive manifestations occur. Although somewhat effort-dependent, the test is much more reliable and reproducible than FEV1/FVC.

Figure 11-9. Forced expiratory flow, 25 to 75% (FEF25–75%). The spirogram depicts a 4-L forced vital capacity (FVC) on which the points representing 25% and 75% FVC are marked. The slope of the line connecting these points is the FEF25–75%.

Maximum Voluntary Ventilation

Maximum voluntary ventilation (MVV) is the largest volume of gas that can be breathed in 1 minute by voluntary effort. The MVV is measured by having the subject breathe as deeply and as rapidly as possible for 10, 12, or 15 seconds. The results are extrapolated to 1 minute. The subject is instructed to set his or her own ventilatory rate and move more than tidal volume but less than vital capacity in each breath.

MVV measures the endurance of the ventilatory muscles and indirectly reflects lung–thorax compliance and airway resistance. MVV is the best ventilatory endurance test that can be performed in the laboratory. Values that vary by as much as 30% from predicted values may be normal, so only large reductions in MVV are significant. Healthy, young adults average ~170 L/min. Values are lower in women and decrease with age in both sexes. Because this maneuver exaggerates air trapping and exerts the ventilatory muscles, MVV is decreased greatly in patients with moderate-to-severe obstructive disease. MVV is usually normal in patients with restrictive disease.

Flow–Volume Loops

The flow–volume loop graphically demonstrates the flow generated during a forced expiratory maneuver followed by a forced inspiratory maneuver, plotted against the volume of gas expired (Fig. 11-10; see Chapter 40). The subject forcefully exhales completely, then immediately and forcefully inhales to vital capacity. The expired and inspired volumes are plotted on the abscissa and flow is plotted on the ordinate. Although various numbers can be generated from the flow–volume loop, the configuration of the loop itself is probably the most informative part of the test.

Flow–volume loops were formerly useful in the diagnosis of large airway and extrathoracic airway obstruction prior to the availability of precise imaging techniques. Imaging techniques such as magnetic resonance imaging give more precise and useful information in the diagnosis of upper airway and extrathoracic obstruction and superseded the use of flow–volume loops for diagnosis of these conditions. Therefore, it is rare that flow–volume loops are useful for preoperative pulmonary evaluation in the modern era of imaging.

Figure 11-10. Flow–volume loop. The figure depicts a normally configured adult flow–volume loop. The slope of the loop after the subject reaches peak expiratory flow is nearly linear.

Carbon Monoxide Diffusing Capacity

Because PO2 in the pulmonary capillary blood varies with time as it moves through the pulmonary capillary bed, oxygen cannot be used to assess diffusing capacity. A gas mixture containing carbon monoxide is the traditional diagnostic gas used to measure diffusing capacity. Its partial pressure in the blood is nearly zero, and its affinity for hemoglobin is 200 times that of oxygen.59 Carbon monoxide diffusing capacity (DLCO) collectively measures all the factors that affect the diffusion of gas across the alveolar–capillary membrane. The DLCO is recorded in mL CO/min/mm Hg at STPD (standard temperature and pressure, dry). In persons with normal hemoglobin concentrations and normal [V with dot above]A/[Q with dot above] matching, the main factor limiting diffusion is the alveolar–capillary membrane. Small amounts of carbon dioxide and inspired gas can produce measurable changes in the concentration of inspired gas compared with expired gas. There are several methods for determining DLCO, but all methods measure diffusing capacity according to the equation:

The average value for resting subjects when the single-breath method is used is 25 mL CO/min/mm Hg. DLCO values can increase to 2 or 3 times normal during exercise.

The DLO2 may be estimated from the DLCO by multiplying DLCO by 1.23, although the DLCO is usually the reported value. DLCO can be divided by the lung volume at which the measurement was made to obtain an expression of diffusing capacity per unit lung volume.

Some of the other factors that can influence DLCO are as follows:

1.   Hemoglobin concentration: decreased hemoglobin concentration decreases the DLCO.

2.   Alveolar Pco2: an increased PACO2 raises DLCO.

3.   Body position: the supine position increases DLCO.

4.   Pulmonary capillary blood volume.

Diffusing capacity is decreased in alveolar fibrosis associated with sarcoidosis, asbestosis, berylliosis, oxygen toxicity, and pulmonary edema. These states are frequently categorized as diffusion defects, but low DLCO is probably more closely related to loss of lung volume or capillary bed perfusion. DLCO is decreased in obstructive disease because of the decreased alveolar surface area, loss of capillary bed, the increased distance from the terminal bronchiole to the alveolar–capillary membrane, and [V with dot above]A/[Q with dot above] mismatching. In short, few disease states truly inhibit oxygen diffusion across the alveolar–capillary membrane.

Practical Application of Pulmonary Function Tests

Although we have a host of pulmonary function tests from which to choose, spirometry is the most useful, cost-effective, and commonly used test.60 Screening spirometry yields vital capacity (VC), FVC, and FEV1. From these values, two basic types of pulmonary dysfunction can be identified and quantitated: obstructive defects and restrictive defects. The primary criterion for airflow obstruction is decreased FEV1/FCV ratio. Other measurements such as FEF25–75% can be used to support


the diagnosis of an obstructive defect or to assist in making decisions (e.g., whether to institute bronchodilation). A restrictive defect is a proportional decrease in all lung volumes; thus, VC, FVC, and FEV1 all are reduced, but FEV1/FVC remains normal. When there is a question about whether a decreased VC is due to restriction, total lung capacity should be measured. Reduced total lung capacity defines a restrictive defect but is not necessary unless VC on screening spirometry is reduced. The American Thoracic Society published an experts' consensus concerning interpretation of lung function tests.61 Table 11-4 summarizes the distinction between pulmonary function results obtained from those with restrictive and obstructive defects. Refer to “Pulmonary Function Postoperatively” for a discussion of the use of pulmonary testing.

Table 11-4 Pulmonary Function Tests in Restrictive and Obstructive Lung Disease


Restrictive Disease

Obstructive Disease


Proportional decreases in all lung volumes

Small airway obstruction to expiratory flow



Normal or slightly ↑



Normal or slightly ↓









Normal or ↑ if gas trapping



Normal or ↑ if gas trapping

FVC, forced vital capacity; ↓↓↓, ↑↑↑ = large decrease or increase, respectively; ↓, ↑ = small/moderate decrease or increase, respectively; FEV, forced expiratory volume; FRC, functional residual capacity; TLC, total lung capacity.

Preoperative Pulmonary Assessment

Markedly impaired pulmonary function is likely in patients who have the following:

1.   Any chronic disease that involves the lung

2.   Smoking history, persistent cough, and/or wheezing

3.   Chest wall and spinal deformities

4.   Morbid obesity

5.   Requirement for single-lung anesthesia or lung resection

6.   Severe neuromuscular disease

Preoperative pulmonary evaluation must include history and physical examination and may include chest radiograph, arterial blood gas analysis, and screening spirometry, depending on the patient's history. A history of sputum production, wheezing or dyspnea, exercise intolerance, or limited daily activities may yield more practical information than does formal testing. Arterial blood analysis, which should be sampled while the patient breathes room air, adds information regarding gas exchange and acid-base balance. Arterial blood gas sampling is primarily useful if the patient's history suggests that he or she may be chronically hypoxemic or may “retain” CO2 (i.e., a patient with a chronic, compensated arterial acidemia) and be used to guide ventilatory management goals.

The goals one might hope to achieve through preoperative pulmonary function would be to predict the likelihood of pulmonary complications, obtain quantitative baseline information concerning pulmonary function that guides decision making, and identify patients who may benefit from therapy to improve pulmonary function preoperatively. For patients who will have lung resections, pulmonary function testing does provide some predictive benefit.62 For all other patients, however, overwhelming evidence suggests that preoperative pulmonary function testing does not predict or assign risk for PPCs.63,64

In 2002, the American Society of Anesthesiologists' Task Force on Preanesthetic Evaluation published a practice advisory65 wherein they recommended that “there is insufficient evidence to identify explicit decision parameters or rules for ordering preoperative tests on the basis of specific clinical characteristics.” Review of the literature66 also reveals that specific measurements of lung function do not predict PPCs. Rather, they should be obtained to ascertain the presence of reversible pulmonary disease (bronchospasm) or to define the severity of advanced pulmonary disease. Instead, the clinician obtains more information from the patient's history. In a series of 272 adults undergoing nonthoracic surgery, McAlister et al.67 found that the following historical factors independently increased the risk of PPC: age >65 years, smoking >40 pack-years, COPD, asthma, productive cough, and exercise tolerance of less than one flight of stairs.

The need to obtain baseline pulmonary function data should be reserved for those patients with severely impaired preoperative pulmonary function, such as tetraplegics or myasthenics, so assessment for liberation from mechanical ventilation and/or tracheal extubation might be based on the patient's baseline pulmonary function.

Arterial blood gases are not indicated unless the patient's history suggests arterial hypoxemia or severe enough COPD that one suspects CO2 retention. Then the arterial blood gas finding should be used in essentially the same manner as one might use preoperative pulmonary function tests: to look for reversible disease or to define the severity of the disease at its baseline. Defining baseline Pao2 and Paco2 is particularly important if one anticipates postoperatively ventilating a patient who has severe COPD. Table 11-5 summarizes the respiratory physiology formulas discussed in this chapter.

Anesthesia and Obstructive Pulmonary Disease

Patients with marked obstructive pulmonary disease are at increased risk for both intraoperative and PPCs. For example, patients with reduced FEV1/FVC or reduced midexpiratory flow not only suffer airway obstruction, but also usually exhibit increased airway reactivity. Because of the hazard of provoking reflex bronchoconstriction during laryngoscopy and tracheal intubation, patients with COPD or asthma should receive aggressive bronchodilator therapy preoperatively.


High alveolar concentrations of most potent inhalational anesthetics will blunt airway reflexes and reflex bronchoconstriction, but require a fairly robust cardiovascular system. Adjunctive intravenous administration of opioids and lidocaine prior to airway instrumentation will decrease airway reactivity by deepening anesthesia. Furthermore, a single dose of corticosteroids may help prevent postoperative increases in airway resistance.

Table 11-5 Respiratory Formulas


Normal Values (70 kg)

Alveolar oxygen tension

110 mm Hg

  PAO2 = (PB - 47) FIO2 –; (PACO2/R)

(FIO2 = 0.21)

Alveolar-arterial oxygen gradient

<10 mm Hg

  (A-aO2) = PAO2 – PaO2

(FIO2 = 0.21)

Arterial-to-alveolar oxygen ratio, PaO2/PAO2 ratio


Arterial oxygen content

20 mL/100 mL blood

  Cao2 = (Sao2) (Hb × 1.34) + Pao2 (0.0031)


Mixed venous oxygen content

15 mL/100 mL blood

  C[v with bar above]O2 = (S [v with bar above] o2) (Hb × 1.34) + P[v with bar above]o2 (0.0031)


Arterial-venous oxygen content difference

4–6 mL/100 mL blood

  C(a-[v with bar above])O2 = Cao2 – C[v with bar above]o2


Intrapulmonary shunt


  [Q with dot above]sp/[Q with dot above]T = (Cc′o2 – Cao2)/(Cc′o2 – C[v with bar above]o2)
    where Cc′o2 = (Hb × 1.34) + (PAO2 × 0.0031)


Physiologic dead space


  VD/VT = (PaCO2 – PĒco2)/PaCO2


Oxygen consumption

250 mL/min

  [V with dot above]O2 = CO (Cao2 – C[v with bar above]o2)


Oxygen transport

1,000 mL/min

  Do2 = CO (Cao2)


Respiratory quotient


  [V with dot above]CO2/[V with dot above]o2 = R


PAO2, alveolar oxygen tension; PB, barometric pressure; FIO2, fraction inspired oxygen; PACO2, alveolar carbon dioxide tension; R, respiratory quotient; Pao2, arterial oxygen tension; Cao2, arterial oxygen content; Sao2, arterial oxygen saturation; Hb, hemoglobin concentration; Cv–O2, mixed venous oxygen content; Sv–o2, mixed venous oxygen saturation; Pv–o2, mixed venous oxygen tension;[Q with dot above]SP/[Q with dot above]T, intrapulmonary shunt; Cc′o2, end-pulmonary capillary oxygen content; VD, dead space gas volume; VT, tidal volume; PaCO2, arterial carbon dioxide tension; PĒCO2, mixed expired carbon dioxide tension; [V with dot above]O2, oxygen consumption (mL/min); CO, cardiac output; [V with dot above] CO2, carbon dioxide production (mL/min); Do2, oxygen transport.

Spontaneous ventilation during general anesthesia in patients with severe obstructive disease is more likely to result in hypercapnia than in patients with normal pulmonary function.68 Preoperative FEV1 reduction correlates with the Paco2 increase during anesthesia. Slower rates of mechanical ventilation (8 to 10 breaths · min-1) should be used to allow time for exhalation. Low ventilatory rates necessitate larger tidal volume if one desires a normal Paco2, but larger VT and resultant higher peak airway pressure may predispose the patient to pulmonary barotrauma. Tidal volume and inspiratory flows should be adjusted to keep peak airway pressure less than 40 cm H2O,69,70 if possible. Higher inspiratory flows produce a shorter inspiratory time and, usually, a high peak airway pressure. Thus, a balance that avoids high peak airway pressure and excessively large VT that allows the longest possible expiratory time should be sought.

Ideally, depending on the procedure and the duration of anesthesia, one would extubate the patient's trachea at the end of the operation. The irritating tracheal tube increases both airway resistance and reflex bronchoconstriction, limits the ability of the patient to clear secretions effectively, and increases the risk of iatrogenic infection. For some patients with obstructive disease (e.g., the young asthmatic patient), many advocate tracheal extubation during deep anesthesia at the conclusion of the operation.

Anesthesia and Restrictive Pulmonary Disease

Restrictive disease is characterized by proportional decreases in all lung volumes. The decreased FRC produces low lung compliance and also results in arterial hypoxemia because of low [V with dot above]A/[Q with dot above] mismatching. Patients with this disease typically breathe rapidly and shallowly.

Positive-pressure ventilation of patients with restrictive disease is fraught with high peak airway pressures because more pressure is required to expand stiff lungs. Lower mechanical tidal volumes at more rapid rates reduce the risk of barotrauma but augment ventilation-induced cardiovascular depression and increase the chances of developing atelectasis. Larger tidal volumes should be avoided because of the increased risk of both barotrauma71 and volutrauma.72 Various lung-protective strategies have been developed to ventilate patients with profound restrictive lung disease (see Chapter 56).

Because the FRC is reduced, a lower oxygen store is available during apneic periods. Even preoxygenation with an FIO2 of 1.0 can result in arterial hypoxemia seconds after the cessation of breathing or disconnection from a ventilator circuit. Patients with severe restrictive diseases tolerate apnea poorly. Because arterial hypoxemia develops so rapidly, transportation of these patients within the hospital should be performed with a pulse oximeter.


Even healthy individuals develop mild restrictive defects during anesthesia. FRC decreases 10 to 15% when healthy, spontaneously breathing individuals lie supine. Tracheal intubation further reduces FRC only slightly. General anesthesia consistently decreases FRC by a further 5 to 10%,73 which usually results in decreased lung compliance.74 The FRC reaches its nadir within the first 10 minutes of anesthesia73,75,76 and is independent of whether ventilation is spontaneous or controlled. The diminished FRC persists in the postoperative period but may be restored postoperatively by the use of positive end-expiratory pressure or CPAP.73,77,78 However, once positive airway pressure is removed, FRC plummets to previously diminished levels, which reach a postoperative nadir 12 hours after operation.79

Effects of Cigarette Smoking on Pulmonary Function

Smoking affects pulmonary function in many ways (see Chapter 23). The irritant smoke decreases ciliary motility and increases sputum production. Thus, these patients have a high volume of sputum and decreased ability to clear it effectively. As smoking habits persist, airway reactivity and the development of obstructive disease become problematic. Studies of the pathogenesis of COPD suggest that smoking results in an excess of pulmonary proteolytic enzymes, which directly cause damage to the lung parenchyma.80 Exposure to smoke increases synthesis and release of elastolytic enzymes from alveolar macrophages—cells instrumental in the genesis of COPD from smoking. Further damage to the lung tissue is probably caused by reactive metabolites of oxygen, such as hydroxyl radicals and hydrogen peroxide, which are usually used by the macrophages to kill micro-organisms. The immunoregulatory function of the macrophages is also changed by cigarette smoking, with changes occurring in the presentation of antigens and interaction with T lymphocytes.81Other direct effects on lung tissue caused by smoking include increased epithelial permeability82 and changed pulmonary surfactant.83 The airway irritation or small airway reactivity evoked by inhaling cigarette smoke is the result of activation of sensory endings located in the central airways, which is primarily caused by nicotine.84

Early in the disease, mild [V with dot above]A/[Q with dot above] mismatch, bronchitic disease, and airway hyperreactivity are primary problems. Later, these problems are accompanied by the hallmarks of COPD: gas trapping, flattened diaphragmatic configuration (which decreases the efficiency with which the diaphragm functions), and barrel-chest deformity. Lung compliance increases significantly so limited elastic recoil prevents complete passive emptying. As a result, many patients exhale forcibly to reduce gas trapping.

With gas trapping, ventilation and perfusion become increasingly mismatched. Large areas of dead space ventilation and venous admixture occur. Carbon dioxide elimination is inefficient because of dead space ventilation. The typical minute ventilation for patients with advanced obstructive lung disease can be 1.5 to 2 times normal. In addition, venous admixture produces arterial hypoxemia that is exquisitely sensitive to low concentrations of supplemental oxygen. Gas exchange is further impaired by the increased carboxyhemoglobin concentration that results from inspiring smoke. Normal carboxyhemoglobin concentration in nonsmokers is approximately 1%; in smokers, however, it can be as high as 8 to 10%. Cessation of smoking, even for 12 to 24 hours preoperatively, can decrease CO concentration to near normal.

Smoking is one of the main and most prevalent risk factors associated with postoperative morbidity.85 COPD patients who smoke have a two- to sixfold86 risk of developing postoperative pneumonia compared with nonsmokers. Further, smokers' relative risk of PPC is doubled, even if they do not have evidence of clinical pulmonary disease or abnormal pulmonary function.87 The incidence of PPC in smokers can be reduced by abstinence from smoking, although there is no consensus on the minimal or optimal duration of preoperative smoking abstinence.88,89,90 Warner et al.85 studied 200 patients undergoing coronary artery bypass grafting and found that patients who continued to smoke or stopped <8 weeks before the operation had a complication rate nearly 4 times that of patients who had quit smoking more than 8 weeks preoperatively. These data further demonstrated that those who quit smoking <8 weeks preoperatively had a higher rate of complication than those who continued to smoke. Normalization of mucociliary function requires 2 to 3 weeks of abstinence from smoking, during which time sputum increases. Several months of smoking abstinence is required to return sputum clearance to normal.91 In a study of bupropion-assisted smoking cessation, Hurt et al.92 demonstrated decreased risk of postoperative complications even after 4 weeks of abstinence from smoking.

Nonetheless, Public Health Service guidelines published in 2000 emphasize the responsibility of health care facilities to coordinate interventions aimed at tobacco-dependence treatment. In addition to the guidelines noting that tobacco dependence often necessitates repeated interventions, “every patient who uses tobacco should be offered at least brief treatment” as brief tobacco-dependence therapy has been shown to be effective. These guidelines recognize five first-line pharmacologic adjuncts that increase smoking cessation success: bupropion SR, nicotine gum, nicotine inhaler, nicotine nasal spray and nicotine patch. Additionally, clonidine and nortriptyline were identified as second-line pharmacologic adjuncts.93

Following publication of these 2000 guidelines, a randomized controlled trial using the partial nicotinic acetylcholine agonist, varenicline, showed improved smoking abstinence rates at all times evaluated during the study when compared with bupropion SR treatment.94 Based on this information, the utilization of varenicline in a smoking-cessation program should be considered.

Smokers who decrease, but do not stop, cigarette consumption without the aid of nicotine-replacement therapy continue to acquire equal amounts of nicotine from fewer cigarettes by changing their technique of smoking to maximize nicotine intake.95 Levels of serum nicotine and cotinine and urinary mutagenesis levels remain unchanged. Thus, reduction in the number of cigarettes smoked will likely have little effect on the risk of developing PPCs.86 Smoking patients should be advised to stop smoking 2 months prior to elective operations to maximize the effect of smoking cessation,85 or for at least 4 weeks to benefit from improved mucociliary function and some reduction in PPC rate. If patients cannot stop smoking for 4 to 8 weeks preoperatively, it is controversial whether they should be advised to stop smoking 24 hours preoperatively. A 24-hour smoking abstinence would allow carboxyhemoglobin levels to fall to normal but may increase the risk of PPC.

Pulmonary Function Postoperatively

Risk of Postoperative Pulmonary Complications

Postoperative Pulmonary Function

The changes in pulmonary function that occur postoperatively are primarily restrictive, with proportional decreases in all lung volumes and no change in airway resistance. The decrease in FRC, however, is the yardstick by which the severity of the


restrictive defect is gauged. This defect is generated by abdominal contents that impinge on and prevent normal movement of the diaphragm and by an abnormal respiratory pattern devoid of sighs and characterized by shallow, rapid respirations. The normal resting respiratory rate for adults is 12 breaths per minute, whereas the postoperative patient usually breathes approximately 20 breaths per minute. Furthermore, most (but not all) factors that tend to make the restrictive defect worse are also those associated with a higher risk of PPCs.

The operative site is one of the single most important determinants of the degree of pulmonary restriction and the risk of PPCs. Nonlaparoscopic upper abdominal operations cause the most profound restrictive defect, precipitating a 40 to 50% decrease in FRC compared with preoperative levels, when conventional postoperative analgesia is employed. Lower abdominal and thoracic operations cause the next most severe change in pulmonary function, with decreases in FRC to 30% of preoperative levels. Most other operative sites—intracranial, peripheral vascular, otolaryngologic—have approximately the same effect on FRC, with reductions to 15 to 20% of preoperative levels.

Postoperative Pulmonary Complications

Two problems confound interpretation of the literature examining PPCs (see Chapter 65). First, there is no clear definition of what constitutes a PPC. For example, some clinical studies include only pneumonia, whereas others add atelectasis and/or ventilatory failure. Thus, to interpret data concerning rates of PPCs, it is important to discern what complications are specifically being addressed. Second, the criteria by which the diagnosis of postoperative pneumonia or atelectasis is made vary from study to study. For this discussion, PPCs include atelectasis and pneumonia only. Reasonable, well-accepted diagnostic criteria for these diagnoses include change in the color and quantity of sputum, oral temperature exceeding 38.5°C, and a new infiltrate on chest radiograph.

The operative site is an important risk factor for the development of PPCs. Nonlaparoscopic upper abdominal operations increase risk for PPC at least twofold,89 with rates of occurrence varying from 20 to 70%.95. Lower abdominal and intrathoracic operations are associated with slightly less risk, but still higher risk than extremity, intracranial, and head/neck operations.

Patients with COPD are at risk for PPC. Their risks can be minimized by ensuring they do not have an active pulmonary infection and any increased resistance associated with reactive airways disease is minimized by the use of bronchodilator therapy. Interestingly, those with asthma are not at increased risk for atelectasis or pneumonia. However, exacerbation of asthma in the postoperative period can be problematic. Careful attention must be given to ensuring the continuation of bronchodilating regimens and steroid administration (either inhaled or systemic) through the perioperative period.

There are several strategies by which it is possible to reduce risk of PPC: the use of lung-expanding therapies postoperatively, choice of analgesia,96 and cessation of smoking. After upper abdominal operations, which are associated with the highest incidence of PPCs, FRC recovers over 3 to 7 days. With the use of intermittent CPAP by mask, FRC will recover within 72 hours.97 Patients correctly use incentive spirometers only 10% of the time unless therapy is supervised.98 Stir-up regimens are as effective as incentive spirometry at preventing PPCs99 and they are less expensive than supervised incentive spirometry; thus, they are preferred over incentive spirometry therapy.

After median sternotomy for cardiac operations, FRC does not return to normal for several weeks, regardless of postoperative pulmonary therapy.100 The persistently low FRC in this population is probably due to mechanical factors such as a widened mediastinum, intrapleural fluid, and altered chest wall compliance. The single most important aspect of postoperative pulmonary care is getting the patient out of bed, preferably walking.

The choice of anesthetic technique for intraoperative anesthesia does not change the risk for PPC independent of the operative site or duration of the operation. Operations exceeding 3 hours are associated with a higher rate of PPC. Choice of postoperative analgesia strongly influences the risk of PPC.89 The use of postoperative epidural analgesia, particularly for abdominal and thoracic operations, markedly decreases the risk of PPC and appears to decrease length of stay in the hospital.

Although obesity is associated with marked restrictive defects, some studies demonstrate that obesity does not independently increase the risk of PPC, whereas others do demonstrate increased independent risk for PPCs in the obese population.101 However, there are data to support101 advanced age as an independent risk factor for PPCs.

Several authors have attempted to assess the influence of overall health on PPC risk. The use of indices that weight and score various aspects of physiology and health shows that patients who are in a poor state of health preoperatively tend to be at higher risk of PPC.90

Patients with obstructive airway disease and decreased expiratory flows may benefit from preoperative bronchodilator therapy and formal pulmonary toilet.102 High-risk patients with COPD who receive bronchodilation, chest physical therapy, deep breathing, forced oral fluids (>3 L/day), and preoperative instruction in postoperative respiratory techniques, as well as those who stop smoking for more than 2 months preoperatively, experience a PPC rate approximately equal to that observed in normal patients.103 Interestingly, although a regimen of this nature significantly reduces the incidence of PPCs,104 airway obstruction and arterial hypoxemia are not measurably reversed during the 48 to 72 hours of preoperative therapy.105 It is possible that the reduced complication rate results from the additional attention that these patients receive rather than from the specific regimen employed.


1. Lieberman DA, Falkner JA, Craig AB Jr et al: Performance and histochemical composition of guinea pig and human diaphragm. J Appl Physiol 1973; 34: 233

2. Roussos C, Macklin PT: Diaphragmatic fatigue in man. J Appl Physiol 1977; 43: 189

3. Campbell EJM, Green JH: The behaviour of the abdominal muscles and intra-abdominal pressure during quiet breathing and increased pulmonary ventilation: A study in man. J Physiol (Lond) 1955; 127: 423

4. Conrardy PA, Goodman CR, Lainge F et al: Alteration of endotracheal tube position: Flexion and extension of the neck. Crit Care Med 1976; 4: 8

5. Bachoven M, Weibel ER: Basic pattern of tissue repair in human lungs following unspecific injury. Chest 1974; 65: 145

6. Fishman AP: Non-respiratory function of lung. Chest 1977; 72: 84

7. Hocking WG, Golden DW: The pulmonary-alveolar macrophage. N Engl J Med 1979; 301: 580

8. Whitehead TC, Zhang H, Mullen B, Slutsky AS: Effect of mechanical ventilation on cytokine response to intratracheal lipopolysaccharide. Anesthesiology 2004; 101: 1

9. Dreyfuss D, Rouby J-J: Mechanical ventilation-induced lung release of cytokines: A key for the future or Pandora's box? Anesthesiology 2004; 101: 1

10. Downs JB: A technique for direct measurement of intrapleural pressure. Crit Care Med 1976; 4: 207

11. Baydur A, Behrakis P, Zin WA: A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982; 126: 788

12. Blanch MJ, Kirby RR, Gabrielli A et al: Partially and totally unloading respiratory muscles based on real time measurements of work of breathing. A clinical approach. Chest 1994; 106: 1835

13. Brochard L, Rua F, Lorino H: Inspiratory pressure support compensates for the additional work of breathing caused by the endotracheal tube. Anesthesiology 1991; 75: 739

14. Rohrer F: Der Strömungswiderstand in den menschlichen Atemwegen. Pflugers Arch 1915; 162: 225


15. Nunn JF: Resistance to gas flow and airway closure. In: Applied Respiratory Physiology. Boston, Butterworths, 1987, pp 50

16. Campbell EJM, Freedman S, Smith PS, Taylor ME: The ability of man to detect added elastic loads to breathing. Clin Sci 1961; 20: 223

17. Fink BR, Ngai SH, Holiday DA: Effect of air flow resistance on ventilation and respiratory muscle activity. JAMA 1958; 168: 2245

18. Palmer KNV, Diament ML: Effect of aerosol isoprenaline on blood-gas tensions in severe bronchial asthma. Lancet 1967; 2: 1232

19. Campbell EJM: The effects of increased resistance to expiration on the respiratory behaviour of the abdominal muscles and intra-abdominal pressure. J Physiol 1957; 136: 556

20. Janssens JP, Pache JC, Nicod LP: Physiologic changes in respiratory function associated with aging. Eur Respir J 1999; 13: 107

21. LeGallois CJJ: Expériences sur le Principe de la Vie. Paris, D'Hautel, 1812, p 325

22. Salmoiraghi GC, Burns BD: Localization and patterns of discharge of respiratory neurones in brain-stem of cat. J Neurophysiol 1960; 23: 2

23. Cohen MI: Neurogenesis of respiratory rhythm in the mammal. Physiol Rev 1979; 59: 1105

24. Guz A: Regulation of respiration in man. Ann Respir Physiol 1975; 37: 303

25. Pitts RF, Magoun HW, Ranson SW: The origin of respiratory rhythmicity. Am J Physiol 1939; 127: 654

26. Lumsden TL: Observations on the respiratory centers in the cat. J Physiol (Lond) 1923; 57: 153

27. Cohen MI, Wang SC: Respiratory neuronal activity in the pons of the cat. J Neurophysiol 1959; 22: 33

28. Stella G: On the mechanism of production and the physiologic significance of “apneusis.” J Physiol (Lond) 1938; 93: 10

29. Kabat H: Electrical stimulation of points in the forebrain and mid-brain: The resultant alterations in respiration. J Comp Neurol 1936; 63: 211

30. Wang SC, Borison HL: The vomiting center: A critical experimental analysis. Arch Neurol Psychiatry 1950; 63: 928

31. Gaylor JB: The intrinsic nervous mechanisms of the human lung. Brain 1934; 57: 143

32. Davis HL, Fowler WS, Lambert EH: Effect of volume and rate of inflation and deflation on transpulmonary pressure and response of pulmonary stretch receptors. Am J Physiol 1956; 187: 558

33. Hering E, Breuer J: Die Sebsteuerung der Atmung durch den Nervus vagus. Stizber Akad Wiss Wien 1868; 57: 672

34. Ide T, Sakurai Y, Aono M, Nishino T: Contribution of peripheral chemoreception to the depression of the hypoxic ventilatory response during halothane anesthesia in cats. Anesthesiology 1998; 90: 1084

35. Mora CT, Torjman M, White PF: Effects of diazepam and flumazenil on sedation and hypoxic ventilatory response. Anesth Analg 1989; 68: 473

36. Leusen I: Regulation of cerebrospinal fluid composition with reference to breathing. Physiol Rev 1972; 52: 1

37. Cohen MI: Discharge patterns of brain-stem respiratory neurons in relation to carbon dioxide tension. J Neurophysiol 1968; 31: 142

38. Heinemann HO, Golaring RM: Bicarbonate and the regulation of ventilation. Am J Med 1974; 57: 361

39. Severinghaus JW, Mitchell RA, Richardson BW et al: Respiratory control at high altitude suggesting active transport regulation of CSF pH. J Appl Physiol 1963; 18: 1155

40. Ferris EB, Engel GL, Stevens CD, Webb J: Voluntary breath holding. J Clin Invest 1946; 25: 734

41. Stock MC, Downs JB, McDonald JS et al: The carbon dioxide rate of rise in awake apneic humans. J Clin Anesth 1988; 1: 96

42. Eger EI, Severinghaus JW: The rate of rise of Paco2 in the apneic anesthetized patient. Anesthesiology 1961; 22: 419

43. Stock MC, Schisler JQ, McSweeney TD: The Paco2 rate of rise in anesthetized patients with airway obstruction. J Clin Anesth 1989; 1: 328

44. Wright FG, Foley MF, Downs JB et al: Hypoxemia and hypocarbia following intermittent positive-pressure breathing. Anesth Analg 1976; 55: 555

45. Fink BR: The stimulant effect of wakefulness on respiration: Clinical aspects. Br J Anaesth 1961; 33: 97

46. Berger AJ, Mitchell RA, Severinghaus JW: Regulation of respiration: III. N Engl J Med 1977; 297: 194

47. West JB, Dollery CT, Naimark A: Distribution of blood flow in isolated lung: Relation to vascular and alveolar pressures. J Appl Physiol 1964; 19: 713

48. West JB, Dollery CT: Distribution of blood flow and the pressure-flow relations of the whole lung. J Appl Physiol 1965; 20: 175

49. Gattinoni L, Pesent A, Avalli L et al: Pressure-volume curve of total respiratory system in acute respiratory failure. Computed tomographic scan study. Am Rev Respir Dis 1987; 136: 730

50. Benumof JL, Pirla AF, Johanson I et al: Interaction of PVO2 with PAO2 on hypoxic pulmonary vasoconstriction. J Appl Physiol 1981; 51: 871

51. Swenson EW, Finley TN, Guzman SV: Unilateral hypoventilation in man during temporary occlusion of one pulmonary artery. J Clin Invest 1961; 40: 828

52. Yamanaka MK, Sue DY: Comparison of arterial-end-tidal PCO2 difference and deadspace/tidal volume ratio in respiratory failure. Chest 1987; 92: 832

53. Tyburski JG, Collinge JD, Wilson RF, Carlin AM, Albaran RG, Steffes CP: End-tidal CO2-derived values during emergency trauma surgery correlated with outcome: A prospective study. J Trauma 2002; 53: 738

54. Huffmyer JL, Nemergut EC: Respiratory dysfunction and pulmonary disease in cirrhosis and other hepatic disorders. Respir Care 2007; 52: 1030

55. Gaines DI, Fallon MB: Hepatopulmonary syndrome. Liver Int 2004; 24: 397

56. Räsänen J, Downs JB, Malec DJ, Oates K: Oxygen tensions and oxyhemoglobin saturations in the assessment of pulmonary gas exchange. Crit Care Med 1987; 15: 1058

57. Christi RV: Lung volume and its subdivisions. I. Methods of measurement. J Clin Invest 1932; 11: 1099

58. Tisi GM: Preoperative evaluation of pulmonary function. Validity, indications and benefits. [Review] Am Rev Respir Dis 1979; 119: 293

59. Apthorp GH, Marshall R: Pulmonary diffusing capacity: A comparison of breath-holding and steady-state methods using carbon monoxide. J Clin Invest 1961; 40: 1775

60. Crapo RO: Pulmonary function testing. N Engl J Med 1994; 331: 25

61. American Thoracic Society: Lung function testing: Selection of reference values and interpretive strategies. Am Rev Respir Dis 1991; 144: 1202

62. Kearney DJ, Lee TH, Reilly JJ et al: Assessment of operative risk in patients undergoing lung resection: Importance of predicted pulmonary function. Chest 1994; 105: 753

63. Ferguson MK: Preoperative assessment of pulmonary risk. Chest 1999; 115: 58S

64. Bapoje SR, Whitaker JF, Schulz T et al: Preoperative evaluation of the patient with pulmonary disease. Chest 2007; 132: 1637

65. Task Force on Preanesthetic Evaluation. Practice advisory for preanesthetia evaluation: A report by the American Society of Anesthesiologists. Anesthesiology 2002; 96: 485

66. Zollinger A, Hofer C, Pasch T: Preoperative pulmonary evaluation: Fact and myth. Curr Opin Anaesth 2001; 14: 59

67. McAlister FA, Khan NA, Strauss SE, Papaioakim M, Fisher BW, Majumdar SR, et al: Accuracy of the preoperative assessment in predicting pulmonary risk after nonthoracic surgery. Am J Respir Crit Care Med 2003; 167: 741

68. Pietak W, Weenig CS, Hickey RF et al: Anesthetic effects on ventilation in patients with chronic obstructive pulmonary disease. Anesthesiology 1975; 42: 160

69. Connors AF, McAferee D, Gray BA: Effect of inspiratory flow rate on gas exchange during mechanical ventilation. Am Rev Respir Dis 1981; 124: 537

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

71. Petersen GW, Baier H: Incidence of pulmonary barotrauma in a medical ICU. Crit Care Med 1983; 11: 67

72. Gattinoni L, Pesenti A: The concept of “baby lung.” Intensive Care Med 2005; 31: 776

73. Brisner B, Hedenstierna G, Lundquist H et al: Pulmonary densities during anesthesia with muscular relaxation: A proposal of atelectasis. Anesthesiology 1985; 62: 422

74. Don HF, Robson JG: The mechanics of the respiratory system during anesthesia. The effects of atropine and carbon dioxide. Anesthesiology 1965; 26: 168

75. Don HF, Wahba M, Cuadrado L et al: The effects of anesthesia and 100 per cent oxygen on the functional residual capacity of the lungs. Anesthesiology 1970; 32: 251

76. Westbrook PR, Stubbs SE, Sessler AD et al: Effects of anesthesia and muscle paralysis on respiratory mechanics in normal man. J Appl Physiol 1973; 34: 81

77. Wyche MQ, Teichner RL, Kallost T et al: Effects of continuous positive-pressure breathing on functional residual capacity and arterial oxygenation during intra-abdominal operation: studies in man during nitrous oxide and d-tubocurarine anesthesia. Anesthesiology 1973; 38: 68

78. Rose DM, Downs JB, Heenen TJ: Temporal responses of functional residual capacity and oxygen tension to changes in positive end-expiratory pressure. Crit Care Med 1981; 9: 79

79. Craig DB: Postoperative recovery of pulmonary function. Anesth Analg 1981; 60: 46

80. Diamond L, Lai YL: Augmentation of elastase-induced emphysema by cigarette smoke: effects of reducing tar and nicotine content. J Toxicol Environ Health 1987; 20: 287

81. deShazo RD, Banks DE, Diem JE, et al. Broncho-alveolar lavage cell–lymphocyte interactions in normal nonsmokers and smokers. Am Rev Respir Dis 1983; 127: 545

82. Hogg JC: The effect of smoking on airway permeability. Chest 1983; 83: 1

83. Clements JA: Smoking and pulmonary surfactant. N Engl J Med 1972; 286: 261

84. Lee L-Y, Gerhardstein DC, Wang AL, Burki NK: Nicotine is responsible for airway irritation evoked by cigarette smoke inhalation in men. J Appl Physiol 1993; 75: 1955

85. Warner MA, Divertie MB, Tinker JH: Preoperative cessation of smoking and pulmonary complications in coronary artery bypass patients. Anesthesiology 1984; 60: 380

86. Bluman LG, Mosca L, Newman N, Simon DG: Preoperative smoking habits and postoperative pulmonary complications. Chest 1998; 113: 883

87. Chalon J, Tayyab MA, Ramanathan S: Cytology of respiratory complications after operation. Chest 1975; 67: 32

88. Theadom A, Copley M: Effects of preoperative smoking cessation on the incidence and risk of intraoperative and postoperative complications in adult smokers: a systematic review. Tob Control 2006; 15: 352

89. Quraishi SA, Orkin FK, Roizen MF: The anesthesia preoperative assessment: an opportunity for smoking cessation intervention. J Clin Anesth 2006; 18: 635


90. Warner MA, Offord KP, Warner ME et al: Role of postoperative cessation of smoking and other factors in postoperative pulmonary complications: A blinded prospective study of coronary artery bypass patients. Mayo Clin Proc 1989; 64: 609

91. Beckers S, Camu F. The anesthetic risk of tobacco smoking. Acta Anaesthesiol Belg 1991; 42: 45

92. Hurt RD, Sachs DPL, Gover ED et al: A comparison of sustained-release bupropion and placebo of smoking cessation. N Engl J Med 1997; 337: 1195

93. Fiore MC; Bailey WC, Cohen SJ et al: US public health service clinical practice guideline: treating tobacco use and dependence. Respir Care 2000; 45: 1200

94. Gonzales D, Rennard SI, Nides M et al: Varenicline, an alpha4beta2 nicotinic acetylcholine receptor partial agonist, vs sustained-release bupropion and placebo for smoking cessation: a randomized controlled trial. JAMA 2006; 296: 47

95. Benowitz NL, Jacob P, Kozlowski LT et al: Influence of smoking fewer cigarettes on exposure to tar, nicotine and carbon monoxide. N Engl J Med 1986; 3115: 1310

96. Liu SS, Wu CL: Effect of postoperative analgesia on major postoperative complications: a systematic update of the evidence. Anesth Analg 2007; 104: 689

97. Gust R, Pecher S, Gust A et al: Effect of patient-controlled analgesia on pulmonary complications after coronary artery bypass grafting. Crit Care Med 1999; 27: 2218

98. Stock MC, Downs JB, Gauer PK et al: Prevention of postoperative pulmonary complications with CPAP, incentive spirometry and conservative therapy. Chest 1985; 87: 151

99. Lyager S, Wernberg M, Rajani N et al: Can postoperative pulmonary complications be improved by treatment with BartlettEdwards incentive spirometer after upper abdominal surgery? Acta Anaesthesiol Scand 1979; 23: 312

100. Stock MC, Downs JB, Cooper RB et al: Comparison of continuous positive airway pressure, incentive spirometry, and conservative therapy after cardiac operations. Crit Care Med 1984; 12: 969

101. Smetana GW, Lawrence VA, Cornell JE: Preoperative pulmonary risk stratification for noncardiothoracic surgery: systematic review for the American College of Physicians. Ann Intern Med 2006; 144: 581

102. Chumillas S, Pace JL, Delgado F et al: Prevention of postoperative pulmonary complications through respiratory rehabilitation: A controlled clinical trial. Arch Phys Med Rehab 1998; 79: 5

103. Brooks-Brunn JA: Validation of a predictive model for postoperative pulmonary complications. Heart Lung 1998; 27: 151

104. Gracey DR, Divertie MB, Didier EP: Preoperative pulmonary preparation of patients with chronic obstructive pulmonary disease: A prospective study. Chest 1979; 76: 123

105. Petty TL, Brink GA, Miller NW, Corsello PR: Objective functional improvement in chronic airway obstruction. Chest 1970; 57: 216

Editors: Barash, Paul G.; Cullen, Bruce F.; Stoelting, Robert K.; Cahalan, Michael K.; Stock, M. Christine