Thoracic Anesthesia


The Principles of Thoracic Anesthesia


Physiology of One-Lung Ventilation

Raquel R. Bartz
Richard E. Moon

Optimal operating conditions for many cardiothoracic procedures require collapse of one lung, producing a challenge for the anesthesiologist who must maintain arterial PO2, PCO2, and hemodynamics within tolerable levels while ventilating the single remaining lung. One-lung ventilation (OLV) for thoracic surgery is usually performed while the patient is in the lateral decubitus position, with the nondependent lung collapsed. For other types of procedures, for example whole lung lavage, the patient may be in the supine position and the nonventilated lung remains inflated with saline resulting in different physiologic consequences. Therefore, a thorough understanding of pulmonary physiology will help facilitate the delivery of anesthesia for procedures requiring OLV.


Flow of gas into the lungs is dependent upon the pressure difference between the upper airway (or endotracheal tube) and the alveoli. This pressure change can be induced either by inspiratory muscle contraction (spontaneous breathing) or positive pressure ventilation. Time-related variation in lung and chest wall mechanical properties induce variable tidal volumes and airway pressures during both spontaneous breathing and in some modes of positive pressure ventilation. To expand the lungs and overcome elastic recoil and resistive loads, a sufficient transpulmonary pressure gradient must either be generated with the use of positive pressure breathing or from negative pressure by the contraction of the diaphragm and muscles of inspiration as occurs during spontaneous breathing.

Elastic Effects

The lungs can be considered as many small, interconnected balloons inside the chest cavity, the elastic properties of which produce a recoil force. During lung expansion, recoil force of the lung is generated due to smooth muscle, elastin, and collagen as well as from other surface forces. The latter are due to surface tension generated by liquid within the alveoli. This is attenuated by surfactant produced by type II pneumocytes present at the alveolar-air interface. Changes in lung volume are accompanied by parallel changes in the volume of the chest wall, which has its own elastic properties. There is also a resistive component, mostly due to the properties of the airways, discussed below.

The lungs expand due to an increase in transpulmonary pressure (Ptp), equal to the difference between alveolar pressure (PA) and pleural pressure (Ppl) (Ptp = PA – Ppl). The pressure generated within the lungs to produce lung inflation must also be great enough to overcome forces independent of the lung including the chest wall, diaphragm, and abdomen (defined collectively as “chest wall”).

The change in a given unit volume (eg, lung volume) per unit change in pressure is termed compliance. Compliance (C) is calculated by the following equation:


Where V is intrapulmonary gas and P is transthoracic pressure. Lung compliance (CL) in a normal awake human is typically in the range 150 to 250 mL/cm H2O. Because elastic recoil forces increase at maximal lung volumes, compliance decreases as the lung inflates. Lung compliance is also decreased in diseases that result in scarring, fibrosis, and pulmonary edema. Although lung compliance normally increases with aging, diseases that disrupt the lung architecture of the alveolar septa, such as emphysema, will accelerate this process. Changes of lung compliance also occur with general anesthesia. Measurements before and after induction of anesthesia have typically revealed a 15% to 50% decrease in CL.1

Because transpulmonary pressure is difficult to measure, the most clinically relevant ΔP is transrespiratory pressure, PTR, defined as the pressure required to expand both lungs and the chest wall, which in a mechanically ventilated patient is equal to airway pressure. PTR can be used to calculate total respiratory system compliance (CRS), which is typically 90 to 120 mL/cm H2O. Compliances for individual hemithoraces in normal anesthetized volunteers in the lateral decubitus position have been reported as 39 mL/cm H2O (nondependent hemithorax) and 29 mL/cm H2O (dependent hemithorax).2

Several factors affect lung compliance including lung volume, patient position, pulmonary blood volume, age, restriction of chest wall expansion, smooth muscle tone, lung diseases, and general anesthesia. After induction of general anesthesia there is a 20% to 35% reduction in CRS. In common clinical usage, respiratory system compliance is defined as flow-dependent inflation (dynamic compliance) and non-flow-dependent inflation or static compliance. Dynamic and static compliance can be easily measured during positive pressure ventilation. Dynamic compliance represents the function of airway resistance on pressure changes whereas static compliance is measured when gas flow is static at the end of inspiration and thus mostly represents the compliance of the alveolar units. Changes in peak airway pressure measured by the mechanical ventilator at the end of inspiration reflect changes of dynamic compliance whereas changes of the plateau pressure (alveolar pressure) reflect changes in static compliance (Figure 3–1). Measurement of lung expansion and deflation, however, show that deflation pressures exceed inflation pressures. This difference in the volumes between inflation and deflation within the lung represents the affects of airway resistance and hysteresis, which is mostly due to the air–water surface forces at the beginning of inflation. The change in compliance during inflation and deflation can also be graphically plotted as pressure/volume loops, which many mechanical ventilators on anesthesia machines now display. The conventional clinical definition of dynamic compliance described above for mechanically ventilated patients differs from the classic definition, in which inspiratory and expiratory pressures are obtained at zero flow.


Figure 3–1. Measurement of respiratory system (thoracic compliance). The pressure-volume relationship (peak and plateau pressures, positive end-expiratory pressure (PEEP), and tidal volume, VT) during delivery of a positive pressure tidal volume can be used to calculate clinically defined static and dynamic respiratory system compliances (CDyn, CStat).

Resistive Effects

The airways of the respiratory system not only serve as the conducting vessels for gas transport but also contribute to the overall homeostasis of the lungs. In the noninstrumented patient, air is humidified in the upper airway and ciliated epithelial cells clear particles as they move down respiratory passages. Airway diameters, and thus resistance (R), change in response to sympathetic, cholinergic and nonadrenergic noncholinergic (NANC) systems, which act on bronchial smooth muscle and mucus glands. NANC mediators include substance P, neurokinins A and B (which cause bronchoconstriction) and vasoactive intestinal peptide (VIP, which causes bronchodilation). Bronchodilation also occurs as a result of local generation of nitric oxide (NO), which induces smooth muscle relaxation through soluble guanylate cyclase-dependent mechanisms. Like compliance, airway resistance changes with lung volume and differs during inspiration and expiration. During inspiration, lung expansion tends to increase airway diameter and thus decrease resistance; during expiration the decreasing lung volumes can result in compressed airways and increased airway resistance. Clinically, this may be an issue for patients with pathological narrowing of the airways, such as asthma, which may result in intrinsic positive end-expiratory pressure at the end of expiration or auto PEEP. Because inhaled anesthetics inhibit smooth muscle contraction of the airways, they produce a significant decrease in airway resistance and in the past have been used successfully to treat severe asthma exacerbations.


Lung volumes can be divided into different components (Figure 3–2). Normal, resting breathing occurs with a tidal volume (VT) of approximately 5 to 7 mL/kg of ideal body weight. The volume of gas that can be exhaled beyond a resting expiration is termed the expiratory reserve volume (ERV), which is approximately 1000 mL in males and 600 mL in females. A small amount of gas is always present within the lungs at the end of maximal expiration and is termed the residual volume (RV). The summation of RV and ERV make up the functional residual capacity (FRC), which represents the point where alveolar pressure equals ambient pressure and the expansive chest wall forces are balanced by the elastic recoil of the lungs. Several factors influence FRC. Stiffening of the chest wall from aging and a decrease in the elastic forces of the lung leads to a gradual increase in RV and FRC over time. FRC is linearly related to height and decreases significantly due to obesity. In women, FRC is about 10% lower than in men. FRC is also decreased by the supine position and is affected by general anesthesia. FRC decreases during general anesthesia by 10% to 20% primarily due to a change in the shape of the ribcage.3 In awake volunteers in the lateral decubitus position, the dependent lung has a lower FRC than the nondependent lung, which during anesthesia can lead to atelectasis and regions of poor ventilation. In studies in anesthetized volunteers after being turned from supine to lateral decubitus position, the normal reduction in FRC due to anesthesia was partially reversed, due to an increase in the nondependent lung FRC.2 In the supine position each lung contributes approximately the same proportion of ventilation. However, in the lateral decubitus position during spontaneous breathing ventilation of the dependent lung is greater than the nondependent lung,2,4 while during positive pressure ventilation the reverse is true.2


Figure 3–2. Subdivision of lung volumes. Several factors can different subdivisions of the lung volumes especially Functional Residual Capacity (FRC). FRC is linearly related to height, which increases FRC. Obesity and the supine position cause a reduction in FRC. (Adapted with permission from Shier D, Butler J, Lewis R. Hole’s Human Anatomy and Physiology. New York: McGraw-Hill; 2004).


Ventilation exchanges air from the environment to the lungs by producing a transpulmonary pressure gradient that facilitates transport of oxygen from the upper airway to the alveoli and movement of carbon dioxide in the opposite direction. Several factors influence both the alveolar oxygen and carbon dioxide pressure. The main determinants of alveolar PO2 (PAO2) are inspired PO2 and alveolar ventilation image. Arterial PO2 (PaO2) also depends on ventilation/perfusion matching, right-to-left (intrapulmonary and intracardiac) and alveolar-capillary diffusion, although the latter is rarely a limiting factor in clinical medicine except during exercise in patients with interstitial disease.

After leaving the right heart, deoxygenated blood enters the pulmonary capillaries and flows into the pulmonary veins. Hypoxemia results from desaturated blood (venous admixture) emanating from abnormal gas exchange units with low ventilation/perfusion ratio <1) or shunt image units. Pulmonary arteriovenous right-to-left shunts have been described within the normal lung and are detectable when pulmonary artery flow and pressure are high, such as during exercise. Two other physiological sources of deoxygenated blood include the bronchial and thebesian veins, which drain directly into the pulmonary veins and left atrium, respectively (“left-to-left” or post-pulmonary shunt). Normally these account for <1% of the cardiac output (CO) and therefore do not significantly affect arterial oxygenation. However, in patients with severe bronchiectasis, where the bronchial circulation can increase several fold, left-to-left shunt may contribute significantly to arterial hypoxemia.

Secondary factors affecting PaO2 include CO and second gas effect. The alveolar/arterial PO2 difference in a young healthy adult breathing air is usually < 10 mm Hg but increases with age and lung disease. This difference results from physiological shunt or venous admixture. In diseased lungs this difference may increase significantly and is the main cause of arterial hypoxemia.

In the steady state, PCO2 in arterial blood is determined by the amount of CO2 produced from metabolism and the amount eliminated by ventilation. Total ventilation (minute ventilation, [image], which is the product of respiratory rate [f] and tidal volume [VT]), is made up of alveolar ventilation and dead space ventilation (image). Dead space is composed of the volume of the upper airway and conducting airways (“anatomic” or “series” dead space) and unperfused or hypoperfused distal gas exchange units (physiological or parallel dead space). Alveolar PCO2 is determined by the following equation:


Alveolar ventilation can be written as:


Where VD/VT is dead space/tidal volume ratio. Therefore, since image is a function of both image and VD/VT, any variation in image image, or VD/VT will also affect PCO2. O2 consumption image, and hence image, decreases slightly during general anesthesia, to a large extent because of the frequently associated reduction in body temperature (Figure 3–3).


Figure 3–3. Alveolar ventilation is determined by the amount of anatomic and physiologic dead space as well as the minute ventilation, which is a function of respiratory rate and tidal volume.

During OLV, maintenance of PaCO2 at preinduction levels will require image at or slightly higher than during two-lung ventilation. This is to account for the effect of right-to-left shunt and less efficient image matching. Indeed, the majority of individuals requiring OLV have lung disease, which may limit the degree to which isocapnia can be maintained. In particular, obstructive lung disease is associated with increased VD/VT. Airways obstruction may limit the adequacy of expiration and lead to dynamic hyperinflation also known as auto PEEP, particularly during OLV5, which in turn can reduce pulmonary blood flow and lead to systemic hypotension.


The pulmonary arteries divide several times to form arterioles that have larger diameters and thinner vessel walls than their counterparts in the systemic circulation. This interconnected network of vessels leads to small single-cell layered capillaries where alveolar-capillary exchange of O2 and CO2 occurs via diffusion. While the majority of pulmonary gas exchange occurs in the capillaries, a significant portion also occurs in the larger arterioles of the pulmonary vasculature. During air breathing, pulmonary blood may acquire as much as 15% of its O2 load before reaching the capillaries; during 100% oxygen breathing the pulmonary arterial blood may be fully oxygenated before traversing the capillaries.6

Because the pulmonary circulation is a large and highly compliant interconnected network of vasculature under relatively low pressure and normal physiologic conditions, the right ventricle requires less contractile force to generate blood flow into the capillary network of the lungs than the force generated by left ventricle for the systemic circulation. For this reason, the right ventricle has less musculature than its counterpart on the left side. Adequate ejection of blood from the right ventricle and the distensibility or resistance of the pulmonary vasculature are the main determinants of pulmonary arterial blood flow.

Positional changes in pulmonary blood flow distribution would be expected when the patient is turned to the lateral decubitus position. While animal studies suggest in fact that very little flow redistribution occurs; human data support increased blood flow in the dependent lung.4,7,8 This facilitates image matching and maintenance of arterial oxygenation during OLV in the lateral decubitus position.


The relatively low-pressure pulmonary arterial system contains approximately 450 mL of blood at any given time or about 9% of total circulatory blood volume. Normal pulmonary arterial pressure is typically 25/5 mm Hg, mean pressure 10 to 12 mm Hg. Pulmonary blood flow or perfusion exhibits significant heterogeneity among lung regions. This was described in the 1960s and thought to be primarily related to gravity.9 West first proposed the idea of gravity-dependent forces being responsible for regional heterogeneity of pulmonary perfusion, with blood in the erect position flowing preferentially to the most dependent areas of the lung and the least amount to the lung apices (Figure 3–4).


Figure 3–4. The Zones of West. In the uppermost zones (zone 1) image, thus there is no blood flow. In the middle zone (zone 2) image, thus flow depends on the pressure difference between arterial and alveolar pressures. In the basal lung units (zone 3) image, thus blood flow is continuous, and ventilation and perfusion are well matched. (Redrawn permission from West JB, Dollery CT, Naimark A.10)

West proposed the existence of three lung zones. In the uppermost zones (zone 1) at apices of the lung, pulmonary blood flow is lacking given the pressure exerted by the alveoli (PA) is greater than pressure in the pulmonary arteries (Pa) or veins (PV); image. This zone is therefore largely dead space. In the middle zone (zone 2) image, flow depends on the pressure difference between arterial and alveolar pressures. In the basal lung units (zone 3) image, blood flow is continuous, and ventilation and perfusion are well-matched. The practical implications of West zones are that if lung volume in zone 3 is atelectatic then better oxygenation may ensue if it can be recruited. Some authors have added a zone 4 region to reflect those areas in the lung where interstitial pressure (Pis) is higher than PV, and blood flow is disproportionately reduced.

Methods that appeared to confirm the West’s original gravitational hypothesis were quite crude at the time,11 and newer evidence has led to other models. Glenny and his colleagues, using microsphere studies of the lungs during the 1990s, were able to determine that gravity dependent mechanisms, while present, play a lesser role, the major factor being branching and path length of the vessels.12 Due to these factors, pulmonary blood flow is highest in hilar regions compared with the lung periphery. Previously, recruitment of apical pulmonary vessels was thought to occur during increased cardiac output (CO) with exercise, however recent studies suggest that increasing CO above resting levels results in a very modest redistribution of blood flow.13 A gravitational effect will be most evident in a well-expanded lung, with pulmonary vessels dilated and low pulmonary artery pressure, and least evident in a low volume, vasoconstricted lung with high pulmonary artery pressure. For example, no gravitational effect has been demonstrated within the dependent lung in the lateral decubitus position.


Pulmonary arteriolar vasodilation and vasoconstriction play an important role in distribution of pulmonary blood flow and maintain matching of pulmonary ventilation image with pulmonary perfusion image. Pulmonary arterioles constrict in response to low oxygen tension (hypoxic pulmonary vasoconstriction or HPV). This is unique to the lung vessels; blood vessels in the systemic circulation dilate in response to hypoxemia. HPV occurs in response to alveolar hypoxia at an PAO2 less than 50 mm Hg14 and are also somewhat dependent on mixed venous PO2. The precise intracellular mechanisms leading to pulmonary arteriole smooth muscle contraction in response to oxygen sensing have yet to be defined. However, multiple components for intracellular smooth muscle contraction are thought to play a role including L-type calcium channels, Kv channels, nonspecific cation channels, Rhokinase, reactive oxygen species, and release or uptake of NO by hemoglobin.15-17 Hypoxia has also been shown to stimulate the release of endothelin, a powerful vasoconstrictor.18-20 In response to activation of the endothelin (ET-A and ET-B) receptors by endothelin, the pulmonary arterial smooth muscle myocytes constrict leading to increased pulmonary vascular resistance.18,21 Proposed mechanisms are shown in Figure 3–5. There is little evidence that HPV plays a major role in matching under normal conditions. However, in the presence of lung pathology, HPV facilitates the diversion of pulmonary blood flow away from hypoxic regions to regions with better oxygenation.22 Due to the same mechanism, whole lung hypoxia can cause an increase in pulmonary artery pressure, which if severe enough can lead to decreased right ventricular function.


Figure 3–5. Pathways involved in hypoxic pulmonary vasoconstriction. Acute hypoxia results in an increase of intracellular calcium in pulmonary arterial smooth muscle cells and thus contraction. This increase in calcium is achieved by inflow of extracellular calcium through plasmalemmal calcium channels and release of intracellularly stored calcium. Hypoxic effects could be mediated or modulated by a decrease (left side) or increase (right side) of reactive oxygen species (ROS). NADPH: reduced nicotinamide adenine dinucleotide phosphate; NSCC: nonspecific cation channels; TRP: transient receptor potential; NADH: reduced nicotinamide adenine dinucleotide; NAD: nicotinamide adenine dinulceotide; NADP: nicotinamide adenine dinucleotide phosphate; CCE: capacitative calcium entry; ATP: adenosine triphosphate; IP3: inositol triphosphate; cADPR: cyclic ADP-ribose; SR: sarcoplasmatic reticulum. (Adapted with permission of the European Respiratory Society Sommer, N., A. Dietrich, et al. (2008). Regulation of hypoxic pulmonary vasoconstriction: basic mechanisms. Eur Respir J. 32(6):1639-1651. doi:10.1183/09031936.00013908.)

Pulmonary arterial smooth muscle contraction and dilation also occur in response to systemic factors other than low or normal arterial oxygen tensions. For instance, the pulmonary vessels dilate in response to a high pH or alkalosis and contract in response to a low pH or acidosis. Several drugs, such as inhaled anesthetics, sodium nitroprusside, nitroglycerin, and calcium channel blockers, can also cause vasodilation and inhibition of HPV. In the presence of lung pathology or collapse, these agents can reduce arterial oxygenation. Some halogenated anesthetic agents can attenuate HPV by as much as 30%. However, most commonly used anesthetics such as isoflurane, sevoflurane, and desflurane appear to have modest inhibition of HPV at clinically relevant levels close to 1 MAC.23


If nitrogen has been washed out of the lung and the airway is connected to a source of 100% oxygen, apneic oxygenation can occur. Adequate oxygenation occurs despite apnea due to the relatively slow rise in PCO2 (typically 8 to 12 mm Hg during the first minute, 2 to 3 mm Hg/min thereafter24 and ongoing convective delivery of oxygen from airway to alveoli due to the ratio of CO2 production:O2consumption (R = respiratory exchange ratio) less than 1 (typically 0.8). For a normal individual approximately 250 mL/min of oxygen leave the alveoli and enter the blood. In parallel, the body produces 200 mL of carbon dioxide each minute. Some of this enters the alveoli, with the rest remaining in dissolved form in the blood. Provided the airway remains open and connected to an oxygen source, the discrepancy between oxygen uptake and carbon dioxide release will force oxygen to flow continuously from airway to alveolus despite the lack of active ventilation. When using apneic oxygenation, it has been shown that arterial hemoglobin-O2 saturation of 100% can be maintained for nearly an hour.25 During OLV this principle can be used to augment arterial PO2 if the nonventilated lung remains connected to a source of oxygen.


When transitioning from two- to one-lung ventilation, ventilation of the lung on the operative side will be discontinued, either with the aid of an endobronchial blocker or double-lumen endotracheal tube that allows independent lung ventilation. The isolated lung will then gradually collapse of its own accord or due to external compression by the surgeon, facilitating exhalation of gas via the bronchial tree. Lung collapse can be accelerated by absorption of soluble gases into the bloodstream. Predictably, this occurs faster if the lung has been ventilated with image or 100% image compared with N2-O2mixtures (Figure 3–6).


Figure 3–6. Nondependent lung deflation is best (high “lung deflation score”) when a ventilating gas of high blood solubility is used, either 100% O2 or O2-N2O, compared with air. (Adapted from with permission Ko R, McRae K, Darling G, et al.26)

Collapse of the nonventilated lung results in a reduction in its blood flow, due to the combination of mechanical effects and HPV. The degree of right-to-left shunt through the nonventilated lung to a large extent governs the PaO2. A high blood flow through the nonventilated lung reduces PaO2 and vice-versa (see below). Factors that reduce blood flow through the nonventilated lung include surgical retraction,27 and in the case of whole lung lavage, the hydrostatic pressure of the lavage fluid (typically 50 cm H2O at the end of each fill cycle).


Hemodynamic effects can be seen when the nonventilated lung collapses due to the reduction in available pulmonary vasculature. However, the reduction in pulmonary blood flow through the nonventilated lung during OLV, in most individuals, is accompanied by only a modest increase in pulmonary vascular resistance and mean pulmonary artery pressure (typically a 2-4 mm Hg increase); however, the increase may be augmented in the presence of hypoxemia or hypercapnia.


Because right-to-left shunt plays a dominant role in blood oxygenation during OLV, the major factors that affect PaO2 during OLV are fractional blood flow through the nonventilated lung and mixed venous oxygen content image.

Mixed Venous Oxygen Content

image is determined by oxygen consumption (image) and tissue oxygen delivery (the product of CO and arterial oxygen content). Since image and hemoglobin concentration are relatively constant over short-time periods, changes in image reflect O2 delivery. In Figure 3–7, arterial Hb-O2 saturation is plotted as a function of shunt fraction image and, for convenience, image. Due to the absence of an increase in CO, anemia will cause a decrease in image. In the presence of VA/Q mismatch or right-to-left shunt, any change in O2 delivery, caused by changes in CO or Hb, will influence image and hence PaO2. An increase in CO or Hb, by delivering more oxygen to tissues, will cause an increase in image, and hence a rise in PaO2 and vice-versa.


Figure 3–7. At any given shunt fraction, arterial oxygenation depends on mixed venous oxygen content (displayed here as mixed venous Hb-O2 saturation). In the presence of right-to-left shunt (or image mismatch) low image is associated with arterial hypoxemia. In the presence of shunt or image mismatch a reduction in arterial oxygenation can also be caused by a drop in cardiac output (see text). Anemia would have the same effect as low image.


During OLV, lateral decubitus position improves oxygenation compared with the supine position due to gravitational diversion of blood from the nondependent to the dependent lung, thus decreasing shunt fraction. The effect of position on arterial PO2 at different inspired oxygen fractions is shown in Figure 3–8. VA/Q relationships and the effect of selective PEEP are shown in Figure 3–9. Shunt may also be affected during OLV by head-up or head-down tilt. Head-down tilt during OLV has been associated with worsening of arterial oxygenation, and head-up tilt, an improvement.28 The increase in PaO2 during head-up tilt has been attributed to a decrease in shunt.


Figure 3–8. Arterial PO2 during OLV as a function of position and inspired oxygen fraction (0.4, 0.6, and 1.0). (Reproduced with permission from Bardoczky GI, Szegedi LL, d’Hollander AA, Moures JM, de Francquen P, Yernault JC.30)



Figure 3–9. CT scans and VA/Q distributions during anesthesia (open circles: ventilation, closed circles: blood flow, l/min) in a male (age 52, non-smoker). Upper panel: controlled ventilation (CV) with ZEEP supine; Middle panel: CV with zero end-inspiratory pressure (ZEEP), lateral; Lower panel: differential ventilation (DV) and selective PEEP, lateral. There are atelectatic areas in both lungs while supine, with a large shunt. Atelectatic areas in non-dependent lung are diminished in the lateral position, but shunt has not decreased and the number of regions with low VA/Q ratios has increased. Atelectasis in the dependent lung and shunt flow were both reduced by DV + selective PEEP. (Reproduced from Klingstedt C, Hedenstierna G, Baehrendtz S, Lundqvist H, Strandberg A, Tokics L, Brismar B. Ventilation-perfusion relationships and atelectasis formation in the supine and lateral positions during conventional mechanical and differential ventilation. Acta Anaesthesial Scand. 1990;34:421-429. With permission of John Wiley & Sons, Inc.)


Vasodilators such as calcium channel blockers, nitrates, and nitroprusside tend to inhibit HPV, and thus may disproportionately increase flow to the nonventilated lung. Inhaled anesthetics can also inhibit HPV.29 However, this is not normally seen clinically, perhaps because the concentrations of inhaled agents commonly used (≈1 MAC) are insufficient to offset the degree of HPV actually encountered. Alternatively, HPV may be relatively unimportant in determining PaO2 during anesthesia. An additional possibility is that while the effects of inhaled anesthetics on pulmonary blood flow may tend to impair image matching, this could be offset by their bronchodilatory effects, which may engender a parallel improvement in intrapulmonary gas distribution. In any event, although inhaled anesthetics may attenuate the HPV response, their effect is usually reduced over time.

Intravenous anesthetics do not directly affect HPV, although high doses could reduce oxygenation due to a reduction in CO, causing a decrease in mixed venous oxygen content. Thoracic epidural anesthesia has no significant effect except at high doses (eg, 6-8 mL 0.75% ropivacaine at T7-8); this tends to cause a reduction in arterial oxygenation,31 which could similarly be due to decreased CO.

Improved oxygenation can be achieved through the use of an agent that increases HPV. One such agent is almitrine, a pulmonary vasoconstrictor, which has been shown during OLV to enhance arterial oxygenation when used in conjunction with an α1-adrenergic receptor agonist. These agents have been suggested as useful adjuncts in conjunction with inhaled nitric oxide to treat severe hypoxemia during OLV. Nitric oxide (NO), because it is inhaled and works locally at gas exchange sites where it is distributed, is a selective vasodilator of well-ventilated gas exchange units and thus improves imagematching. The combination of inhaled NO and intravenous almitrine may be especially effective at raising PaO2 during OLV.

Gas Exchange Effects of pH Manipulation

Acidosis produces pulmonary vasoconstriction; alkalosis causes, pulmonary vasodilation. Hypocapnia has been shown in animals to inhibit HPV32-34 and in dogs can be virtually abolished at a PaCO2 of 20 mm Hg (pH ≈ 7.6). However, in humans more modest degrees of alkalosis (pH ≈ 7.5) induced by administration of tris buffer or sodium bicarbonate appear to have little or no effect.35 Although acute respiratory acidosis induces pulmonary vasoconstriction, it does not appear to augment HPV.33

Mechanical Effects of Ventilation

Mechanical ventilation will increase pressure on the dependent hemithorax and thus tend to redistribute blood flow toward the nondependent lung, an effect that is augmented by hyperventilation and high levels of PEEP. If there is insufficient time to exhale fully, auto-PEEP can occur, which has the same effect (Figure 3–10). Auto-PEEP is more likely to occur when airways resistance is high. Low levels of PEEP or auto-PEEP in the dependent lung may actually be helpful in reducing atelectasis, however auto-PEEP can increase dead space fraction, and hence PaCO2, especially in patients with underlying emphysema. Auto-PEEP can also lead to decreased venous return and hypotension. It may also further augment redistribution of blood flow to the nonventilated lung, leading to a decrease in PaO2. Auto-PEEP can be transiently eliminated by removal of the patient from the ventilator circuit and allowing adequate time for complete exhalation. It can be minimized by decreasing the amount of time spent in inspiration, by increasing the inspiratory:expiratory time (I:E ratio) or decreasing the minute ventilation and allowing the PaCO2 to rise (permissive hypercapnia).



Figure 3–10. Detection of auto-PEEP. Flow and airway pressure (Paw) waveforms from a patient with chronic obstructive pulmonary disease, receiving mechanical ventilation.

Positive pressure ventilation with high tidal volumes and unrecognized auto-PEEP can cause unintended pulmonary parenchymal damage. Studies on mechanically ventilated ICU patients who have underlying lung injury in the form of ARDS have consistently shown that minimizing lung volumes not only decreases inflammatory cytokines within the circulation but also decreases mortality associated with ARDS.36-38 This may be a factor even within the relatively short time-frame of a surgical procedure, particularly if the transition from two-lung ventilation to OLV is not accompanied by a reduction in tidal volume. During OLV, maintenance of PaCO2 within an acceptable range requires approximately the same ventilation as when both lungs are ventilated. However, after the transition from two-lung ventilation to OLV if tidal volume is not decreased, stretch injury is very likely to occur. This is particularly true for most surgical patients since the majority of thoracic procedures are performed on patients with significant underlying pulmonary parenchymal disease.

Pattern of Ventilation

Customary ventilatory strategy during general anesthesia incorporates constant tidal volume and rate (constant inter-breath interval). This however is very different from the normal ventilatory pattern, in which tidal volume and inter-breath interval vary from breath to breath. On cursory inspection this pattern appears random, although it is not. In fact, breathing usually exhibits fractal properties, where the inter-breath interval and tidal volume of a given breath are influenced by characteristics of breaths in the distant past.39 Successful incorporation of a physiological breathing pattern into mechanical ventilation has been demonstrated in a porcine model of acute lung injury, using a pattern derived from a spontaneously breathing anesthetized pig.40 The advantage of such “biologically variable ventilation (BVV)” has subsequently been shown for low tidal volume ventilation in ARDS.41 In humans undergoing abdominal aortic aneurysm repair, BVV produced higher PaO2, lower PaCO2, lower dead space ventilation, increased compliance and lower mean peak inspiratory pressure.42 Improved PaO2 using BVV during OLV has been demonstrated in pigs.43 Human studies during OLV have not yet been published, however evidence thus far available suggests that of biologically variable ventilation may provide an additional method of improving arterial oxygenation.


Pulmonary gas exchange during OLV is governed by several factors, including mechanical factors that govern the distribution of pulmonary blood flow, such as ventilation of the single lung and collapse of the nondependent lung, hypoxic pulmonary vasoconstriction, CO, and mixed venous oxygen content. Vasodilators can attenuate HPV and thus may lower PaO2, although this may be offset by increased CO. Manipulation of these factors during OLV can usually achieve acceptable PaCO2 and PaO2. Stretch injury of the ventilated lung is a serious concern when inappropriately high tidal volumes are used.


1. Don H. The mechanical properties of the respiratory system during anesthesia. Int Anesthesiol Clin. 1977;15(2):113-136.

2. Rehder K, Hatch DJ, Sessler AD, Fowler WS. The function of each lung of anesthetized and paralyzed man during mechanical ventilation. Anesthesiology. 1972;37(1):16-26.

3. Warner DO, Warner MA. Human chest wall function while awake and during halothane anesthesia. II. Carbon dioxide rebreathing. Anesthesiology. 1995;82(1):20-31.

4. Kaneko K, Milic-Emili J, Dolovich MB, Dawson A, Bates DV. Regional distribution of ventilation and perfusion as a function of body position. J Appl Physiol. 1966;21(3):767-777.

5. Ducros L, Moutafis M, Castelain MH, Liu N, Fischler M. Pulmonary air trapping during two-lung and one-lung ventilation. J Cardiothorac Vasc Anesth. 1999;13(1):35-39.

6. Conhaim RL, Staub NC. Reflection spectrophotometric measurement of O2 uptake in pulmonary arterioles of cats. J Appl Physiol. 1980;48(5):848-856.

7. Mure M, Domino KB, Robertson T, Hlastala MP, Glenny RW. Pulmonary blood flow does not redistribute in dogs with reposition from supine to left lateral position. Anesthesiology. 89(2):483-492.

8. Chang H, Lai-Fook SJ, Domino KB, et al. Spatial distribution of ventilation and perfusion in anesthetized dogs in lateral postures. J Appl Physiol. 2002;92(2):745-762.

9. West JB. Effects of ventilation-perfusion inequality on over-all gas exchange studied in computer models of the lung. J Physiol. 1969;202(2):116P+.

10. West JB, Dollery CT, Naimark A. Distribution of blood flow in isolated lung: relation to vascular and alveolar pressures. J Appl Physiol. 1964;19:713-724.

11. West JB. Studies of pulmonary and cardiac function using short-lived isotopes oxygen-15, nitrogen-13 and carbon-11. Prog At Med. 1968;2:39-64.

12. Glenny RW. Blood flow distribution in the lung. Chest. 1998;114(1 Suppl):8S-16S.

13. Robertson HT, Hlastala MP. Microsphere maps of regional blood flow and regional ventilation. J Appl Physiol. 2007;102(3):1265-1272.

14. Sommer N, Dietrich A, Schermuly RT, et al. Regulation of hypoxic pulmonary vasoconstriction: basic mechanisms. Eur Respir J. 2008;32(6):1639-1651.

15. Archer SL, Weir EK, Reeve HL, Michelakis E. Molecular identification of O2 sensors and O2-sensitive potassium channels in the pulmonary circulation. Adv Exp Med Biol. 2000;475:219-240.

16. Sham JS, Crenshaw BR Jr, Deng LH, Shimoda LA, Sylvester JT. Effects of hypoxia in porcine pulmonary arterial myocytes: roles of K(V) channel and endothelin-1. Am J Physiol Lung Cell Mol Physiol. 2000;279(2):L262-272.

17. McMahon TJ, Moon RE, Luschinger BP, et al. Nitric oxide in the human respiratory cycle. Nat Med. 2002;8(7):711-717.

18. O’Brien RF, Robbins RJ, McMurtry IF. Endothelial cells in culture produce a vasoconstrictor substance. J Cell Physiol. 1987;132(2):263-270.

19. Hieda HS, Gomez-Sanchez CE. Hypoxia increases endothelin release in bovine endothelial cells in culture, but epinephrine, norepinephrine, serotonin, histamine and angiotensin II do not. Life Sci. 1990;47(3):247-251.

20. Rakugi H, Tabuchi Y, Nakamaru M, et al. Evidence for endothelin-1 release from resistance vessels of rats in response to hypoxia. Biochem Biophys Res Commun. 1990;169(3):973-977.

21. MacLean MR, McCulloch KM, Baird M. Endothelin ETA- and ETB-receptor-mediated vasoconstriction in rat pulmonary arteries and arterioles. J Cardiovasc Pharmacol. 1994;23(5):838-845.

22. Domino KB, Hlastala MP, Eisenstein BL, Cheney FW. Effect of regional alveolar hypoxia on gas exchange in dogs. J Appl Physiol. 1989;67(2):730-735.

23. Rogers SN, Benumof JL. Halothane and isoflurane do not decrease PaO2 during one-lung ventilation in intravenously anesthetized patients. Anesth Analg. 1985;64(10):946-954.

24. Eger EI, Severinghaus JW. The rate of rise of PaCO2 in the apneic anesthetized patient. Anesthesiology. 1961;22:419-425.

25. Frumin MJ, Epstein RM, Cohen G. Apneic oxygenation in man. Anesthesiology. 1959;20:789-798.

26. Ko R, McRae K, Darling G, et al. The use of air in the inspired gas mixture during two-lung ventilation delays lung collapse during one-lung ventilation. Anesth Analg. 2009;108(4):1092-1096.

27. Ishikawa S, Nakazawa K, Makita K. Progressive changes in arterial oxygenation during one-lung anaesthesia are related to the response to compression of the non-dependent lung. Br J Anaesth. 2003;90(1):21-26.

28. Park SY, Kim DH, Kim JS, Lee SS, Hong YW. Effects of head-up tilt on intrapulmonary shunt fraction and oxygenation during 1-lung ventilation in the lateral decubitus position. J Thorac Cardiovasc Surg. 2010;139(6):1436-1440.

29. Bjertnaes LJ, Hauge A, Nakken KF, Bredesen JE. Hypoxic pulmonary vasoconstriction: inhibition due to anesthesia. Acta Physiol Scand. 1976;96(2):283-285.

30. Bardoczky GI, Szegedi LL, d’Hollander AA, Moures JM, de Francquen P, Yernault JC. Two-lung and one-lung ventilation in patients with chronic obstructive pulmonary disease: the effects of position and FiO2Anesth Analg. 2000;90(1):35-41.

31. Xu Y, Tan Z, et al. Effect of thoracic epidural anesthesia with different concentrations of ropivacaine on arterial oxygenation during one-lung ventilation. Anesthesiology. 2010;112(5):1146-1154.

32. Lloyd TC, Jr. Influence of blood pH on hypoxic pulmonary vasoconstriction. J Appl Physiol. 1996;21(2):358-364.

33. Benumof JL, Wahrenbrock EA. Blunted hypoxic pulmonary vasoconstriction by increased lung vascular pressures. J Appl Physiol. 1975;38(5):846-850.

34. Benumof JL, Mathers JM, Wahrenbrock EA. Cyclic hypoxic pulmonary vasoconstriction induced by concomitant carbon dioxide changes. J Appl Physiol. 1976;41(4):466-469.

35. Bergofsky EH, Lehr DE, Fishman AP. The effect of changes in hydrogen ion concentration on the pulmonary circulation. J Clin Invest. 1962;41:1492-1502.

36. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med. 2000;342(18):1301-1308.

37. Kallet RH, Jasmer RM, Pittet JF, et al. Clinical implementation of the ARDS network protocol is associated with reduced hospital mortality compared with historical controls. Crit Care Med. 2005;33(5):925-929.

38. MacIntyre N. Ventilatory Management of ALI/ARDS. Semin Respir Crit Care Med. 2006;27(4):396-403.

39. Fadel PJ, Barman SM, Phillips SW, Gebber GL. Fractal fluctuations in human respiration. J Appl Physiol. 2004;97(6):2056-2064.

40. Lefevre GR, Kowalski SE, Girling LG, Thiessen DB, Mutch WA. Improved arterial oxygenation after oleic acid lung injury in the pig using a computer-controlled mechanical ventilator. Am J Respir Crit Care Med. 1996;154(5):1567-1572.

41. Boker A, Graham MR, Walley KR, et al. Improved arterial oxygenation with biologically variable or fractal ventilation using low tidal volumes in a porcine model of acute respiratory distress syndrome.Am J Respir Crit Care Med. 2002;165(4):456-462.

42. Boker A, Haberman CJ, Girling L, et al. Variable ventilation improves perioperative lung function in patients undergoing abdominal aortic aneurysmectomy. Anesthesiology. 2004;100(3):608-616.

43. McMullen MC, Girling LG, Graham MR, Mutch WA. Biologically variable ventilation improves oxygenation and respiratory mechanics during one-lung ventilation. Anesthesiology. 2006;105(1):91-97.