**2. PEEP optimization**

### **2.1 History**

356 Front Lines of Thoracic Surgery

Fig. 3. Three pressure-volume curves obtained in a lung healthy patient undergoing orthopedic surgery. Paw (blue line): airway pressure; Ptrach (red circle): tracheal pressure; Palv

When patients are generally anesthetized, the alveoli in the dependent lung regions may collapse while non-dependent regions remain open. With the help of sufficient external pressure delivered by a ventilator during inspiration, some of the collapsed lung regions may be opened up (i.e. they are recruited) but already open regions may be overinflated. Neither atelectasis nor hyperinflation of lung regions is beneficial in most clinical cases, however, one or the other or a mixture of both processes is inevitable to a certain degree. During expiration, while gas is exhaled, the alveolar pressure drops, which may lead to alveolar collapse (i.e. derecruitment) of the dependent lung regions. Different types of ventilator-induced lung injury (VILI) are therefore observed during mechanical ventilation (Dreyfuss and Saumon, 1998, Uhlig and Frerichs, 2008), such as shear stress trauma caused by cyclic recruitment/derecruitment, barotrauma and pulmonary edema caused by high ventilation pressure. As one of many perioperative complications, acute respiratory distress syndrome (ARDS) was found developed in 3.1% of patients after thoracic surgery and carrying a high mortality rate of over 30% (Grichnik and D'Amico, 2004, Phua *et at.*, 2009). Various lung protective ventilation strategies have therefore been proposed, including high positive end-expiratory pressure (PEEP) combined with low Vt (Brower *et al.*, 2004, Brochard *et al.*, 1998, The Acute Respiratory Distress Syndrome Network, 2000), permissive hypercapnia (Hickling *et al.*, 1990), and recruitment maneuvers (Lachmann, 1992), to reduce the adverse consequences of mechanical ventilation. In this chapter, we focus on

PEEP was introduced to maintain the once recruited atelectatic areas open and thereby reduce the risk of hypoxemia, cyclic recruitment/derecruitment and biotrauma (Gattinoni *et al.*, 2001, Slutsky and Tremblay, 1998). Figure 4 illustrates the effect of PEEP on keeping the lung open. In healthy subjects, although the pulmonary alveoli increase their size at the end of inspiration and decrease at the end of expiration, the shape of alveoli doesn't change

(black dashed line): alveolar pressure.

optimization of PEEP.

**1.2 Lung protective ventilation strategy** 

In 1960's, Ashbaugh *et al.* proposed the use of PEEP to improve oxygenation in a clinical syndrome characterized by atelectasis and hypoxemia (Ashbaugh *et al.*, 1967). The use of PEEP has become widespread ever since that study. Suter and his colleagues later published the concept of "optimal" PEEP (Suter *et al.*, 1975). Because at that time, cardiac output and blood gas measurements were not always available, they suggested that maximizing tidal compliance could be used to identify a PEEP level, at which oxygen delivery was optimized. In the past three decades, a multitude of physicians and scientists dedicated themselves to identify the best PEEP levels for patients under surgeries (Beiderlinden *et al.*, 2003, Berendes *et al.*, 1996, Bensenor *et al.*, 2007), as well as patients with variable diseases, such as ALI or ARDS (Badet *et al.*, 2009, Huh *et al.*, 2009), morbid obesity (Bohm *et al.*, 2009, Erlandsson *et al.*, 2006), chronic obstructive pulmonary disease (COPD) (Glerant *et al.*, 2005, Mancebo *et al.*, 2000), brain-injury (Shapiro and Marshall, 1978, Huynh *et al.*, 2002), including infants (Greenough *et al.*, 1992, Dimitriou *et al.*, 1999). Although different terminologies and endpoints for optimizing PEEP were used (Villar, 2005), most of the approaches tried to obtain the best oxygenation while minimizing VILI as outcome. A lower mortality rate and a better quality of life would be the most desirable goals of therapies. While PEEP has experimentally been shown to reduce VILI, there is no consent in the literature if a suitable PEEP is able to reduce mortality (Miller *et al.*, 1992, DiRusso *et al.*, 1995, Brower *et al.*, 2004), due to the fact that the effect of PEEP is hard to be assessed independently.

It remains under debate how to titrate an adequate PEEP level in individual patients, despite the widely used application of PEEP in clinical practice (Rouby *et al.*, 2002). Increase of PEEP may prevent alveolar derecruitment in dependent areas but may lead to hyperinflation in the non-dependent areas, which may trigger pulmonary inflammation (Terragni *et al.*, 2007). Besides, high PEEP levels may reduce cardiac output (Baigorri *et al.*, 1994) and impair the hemodynamic stability (Herff *et al.*, 2008). Therefore, as also stated by Rouby and Brochard in an editorial (Rouby and Brochard, 2007), one goal of setting PEEP is to find a suitable level, high enough to keep the lung open while minimizing adverse side effects.

Generally speaking, the current available methods of PEEP titration can be mainly divided into three categories: They are based on 1) arterial blood gases such as partial pressure of oxygen in arterial blood (PaO2) and oxygen saturation (SpO2); 2) lung mechanics such as dynamic compliance and static pressure-volume (P/V) curve; 3) imaging techniques such as computed tomography (CT) and electrical impedance tomography (EIT). In the following, representative methods within these three categories are introduced and their assets and drawbacks discussed.

Optimizing Perioperative Ventilation Support with

thus, not suitable for continuous bedside monitoring.

**2.3 Optimizing PEEP with lung mechanics** 

Adequate Settings of Positive End-Expiratory Pressure 359

Fig. 5. Change of PaO2 during a decremental PEEP trial in a porcine model of acute lung injury. Vertical dashed line indicates the optimal PEEP level with respect to PaO2.

Luecke *et al.* argued that improving only PaO2 was not good enough, the elevation of PaCO2 should not be ignored (Luecke *et al.*, 2005). Girgis *et al.* have shown in twenty ALI/ARDS patients that the PaO2/FiO2 ratio was improved by tuning FiO2 after a recruitment maneuver and monitoring the SpO2 changes during decremental PEEP titration (Girgis *et al.*, 2006). SpO2 values were measured by pulse oximetry, which is a noninvasive method, however, less precise than direct measurement of arterial oxygen saturation. Rouby concluded in a review that the highest PaO2 and SaO2 at the lowest FiO2 indicated the right PEEP level (Rouby *et al.*, 2002). Caramez *et al.* have compared ten different parameters for setting PEEP following a recruitment maneuver, including blood gas analysis and lung mechanics (Caramez *et al.*, 2009). The results of PEEP selection using Crs, PaO2 with or without PaCO2 were statistically indistinguishable (Caramez *et al.*, 2009). Statistically significant differences may have not been revealed due to the small number of studied subjects (n = 14) and high variation. Although these studies indicate that PaO2 is a possible criterion for setting PEEP, precise blood gas analysis is invasive and discontinuous and,

The P/V curve has been introduced to individualize Vt and PEEP settings in patients with ARDS by Matamis *et al.* in 1984 (Matamis *et al.*, 1984). In this concept, a lower inflection point (LIP) and an upper inflection point (UIP) are identified on the inflation limb of the P/V curve (Fig. 6B). The LIP was considered to be the pressure level at which massive alveolar recruitment occurs (Jonson and Svantesson, 1999); UIP was considered to be the pressure level indicating alveolar overdistension (Roupie *et al.*, 1995). In consequence, a ventilation strategy was developed to keep the lung open (by setting PEEP above LIP) and to minimize overdistension (by restricting Vt such that peak pressure is smaller than UIP)

Fig. 4. The effect of positive end-expiratory pressure (PEEP) on keeping the alveoli open. A: Under controlled mechanical ventilation without PEEP, alveoli of a healthy subject stay open throughout the whole breathing cycle, while the alveoli of a patient with lung disease collapse at the end of expiration. B: When PEEP is applied, the recruited alveoli will no longer collapse at the end of expiration in a patient with lung disease.

#### **2.2 Optimizing PEEP with blood gas analysis**

One of the main goals of PEEP selection is to optimize oxygenation. Therefore, it is reasonable to guide the PEEP settings by analyzing blood gases (Girgis *et al.*, 2006, Borges *et al.*, 2006, Luecke *et al.*, 2005). It was suggested that "best" PaO2 (maximum value) indicates the "optimal" PEEP in many studies (Borges *et al.*, 2006, Suarez-Sipmann *et al.*, 2007). Toth *et al.* suggested setting PEEP at the level where PaO2 starts to drop rapidly during a decremental PEEP trial (Toth *et al.*, 2007). A typical course of PaO2 values obtained during a decremental PEEP trial in an experimental model of ALI is shown in Fig. 5. PEEP decreased from 30 cmH2O to 5 cmH2O in steps of 5 cmH2O. Decrease of PaO2 implies worse aeration and oxygenation. Optimal PEEP is defined at the pressure level before PaO2 decreases significantly.

Fig. 4. The effect of positive end-expiratory pressure (PEEP) on keeping the alveoli open. A: Under controlled mechanical ventilation without PEEP, alveoli of a healthy subject stay open throughout the whole breathing cycle, while the alveoli of a patient with lung disease collapse at the end of expiration. B: When PEEP is applied, the recruited alveoli will no

One of the main goals of PEEP selection is to optimize oxygenation. Therefore, it is reasonable to guide the PEEP settings by analyzing blood gases (Girgis *et al.*, 2006, Borges *et al.*, 2006, Luecke *et al.*, 2005). It was suggested that "best" PaO2 (maximum value) indicates the "optimal" PEEP in many studies (Borges *et al.*, 2006, Suarez-Sipmann *et al.*, 2007). Toth *et al.* suggested setting PEEP at the level where PaO2 starts to drop rapidly during a decremental PEEP trial (Toth *et al.*, 2007). A typical course of PaO2 values obtained during a decremental PEEP trial in an experimental model of ALI is shown in Fig. 5. PEEP decreased from 30 cmH2O to 5 cmH2O in steps of 5 cmH2O. Decrease of PaO2 implies worse aeration and oxygenation. Optimal PEEP is defined at the pressure level before PaO2 decreases

longer collapse at the end of expiration in a patient with lung disease.

**2.2 Optimizing PEEP with blood gas analysis** 

significantly.

Fig. 5. Change of PaO2 during a decremental PEEP trial in a porcine model of acute lung injury. Vertical dashed line indicates the optimal PEEP level with respect to PaO2.

Luecke *et al.* argued that improving only PaO2 was not good enough, the elevation of PaCO2 should not be ignored (Luecke *et al.*, 2005). Girgis *et al.* have shown in twenty ALI/ARDS patients that the PaO2/FiO2 ratio was improved by tuning FiO2 after a recruitment maneuver and monitoring the SpO2 changes during decremental PEEP titration (Girgis *et al.*, 2006). SpO2 values were measured by pulse oximetry, which is a noninvasive method, however, less precise than direct measurement of arterial oxygen saturation. Rouby concluded in a review that the highest PaO2 and SaO2 at the lowest FiO2 indicated the right PEEP level (Rouby *et al.*, 2002). Caramez *et al.* have compared ten different parameters for setting PEEP following a recruitment maneuver, including blood gas analysis and lung mechanics (Caramez *et al.*, 2009). The results of PEEP selection using Crs, PaO2 with or without PaCO2 were statistically indistinguishable (Caramez *et al.*, 2009). Statistically significant differences may have not been revealed due to the small number of studied subjects (n = 14) and high variation. Although these studies indicate that PaO2 is a possible criterion for setting PEEP, precise blood gas analysis is invasive and discontinuous and, thus, not suitable for continuous bedside monitoring.

#### **2.3 Optimizing PEEP with lung mechanics**

The P/V curve has been introduced to individualize Vt and PEEP settings in patients with ARDS by Matamis *et al.* in 1984 (Matamis *et al.*, 1984). In this concept, a lower inflection point (LIP) and an upper inflection point (UIP) are identified on the inflation limb of the P/V curve (Fig. 6B). The LIP was considered to be the pressure level at which massive alveolar recruitment occurs (Jonson and Svantesson, 1999); UIP was considered to be the pressure level indicating alveolar overdistension (Roupie *et al.*, 1995). In consequence, a ventilation strategy was developed to keep the lung open (by setting PEEP above LIP) and to minimize overdistension (by restricting Vt such that peak pressure is smaller than UIP)

Optimizing Perioperative Ventilation Support with

level, at which Crs is maximum.

**2.4 Optimizing PEEP with imaging techniques** 

handling.

Adequate Settings of Positive End-Expiratory Pressure 361

Fig. 7. Mean dynamic respiratory system compliances (Crs) of a sedated patient under mechanical ventilation at different PEEP levels. Dashed line indicates the "optimal" PEEP

compliance-volume curves are classified into three categories: 1) a decrease in slope indicates overdistension; 2) an increase in slope indicates recruitment; 3) a quasi-horizontal compliance-volume curve indicates a suitable PEEP setting (Mols *et al.*, 1999). However, the method has not yet been evaluated for clinical relevance. Ranieri *et al.* used the pressuretime curve as an index to predict pulmonary stress (Ranieri *et al.*, 2000). This method requires phases with constant air flow which limits its applicability. Nevertheless, these methods have brought the importance of pulmonary mechanical stress into focus. Talmor *et al.* estimated the transpulmonary pressure with help of esophageal balloon catheters and set PEEP to such a level that transpulmonary pressure stayed between 0 and 10 cmH2O during end-expiratory occlusion, and less than 25 cmH2O during end-inspiratory occlusion (Talmor *et al.*, 2008). They observed improvement of PaO2/FiO2 ratio and Crs compared to the group guided according to the ARDS network standard-of-care recommendations. This finding is interesting, but the placement of esophageal balloon catheters needs additional effort in clinical care. Therefore, this method will only become accepted if advantages over other methods using Crs and blood gas analysis are outweighing the extra burden of complex

CT has a very good spatial resolution and is able to show the distribution of the tissue density in the chest, thereby providing primarily morphological data. Hence, CT is the gold standard for assessment of tidal volume distribution in injured lungs and many validation studies were done by comparing various methods with the CT results (Gattinoni *et al.*, 2006, Carvalho *et al.*, 2008, Suarez-Sipmann *et al.*, 2007, Meier *et al.*, 2008). Figure 8 shows two chest CT images of a patient under mechanical ventilation at two different PEEP levels. Higher aeration and reversal of lung collapse in the dependent lung regions were detected

at a PEEP level of 15 cmH2O (Fig. 8, right) compared to that of 5 cmH2O (Fig. 8, left).

(Dambrosio *et al.*, 1997, Hermle *et al.*, 2002). Takeuchi *et al.* proposed that setting PEEP at LIP + 2 cmH2O might be more appropriate (Takeuchi *et al.*, 2002). However, studies indicate that LIP is only the beginning of recruitment and the UIP is not a reliable marker of overdistension (Crotti *et al.*, 2001, Hickling, 2002, Downie *et al.*, 2004). New findings suggest that it may be better to set PEEP according to UIP on the deflation limb of the P/V curve (Albaiceta *et al.*, 2004). In order to obtain quasi-static P/V curves, the normal ventilation process has to be interrupted by performing a specific respiratory maneuver such as lowflow inflation (Servillo *et al.*, 1997), super-syringe inflation (Matamis *et al.*, 1984) or SCASS, i.e. static compliance by automated single steps (Sydow *et al.*, 1991), (Fig. 6A). As pointed out by LaFollette *et al.,* the key to bedside application is acquiring a dynamic curve, which is easier and more applicable, instead of a static one (LaFollette *et al.*, 2007).

Respiratory system compliance Crs or elastance Ers (Crs=1/Ers) can be measured quasistatically by airway occlusion (D'Angelo *et al.*, 1991) or dynamically by applying linear leastsquares regression on the first order equation of motion (Eq. 1) (Iotti *et al.*, 1995). Considering the limitation of static Crs (Stenqvist *et al.*, 2008) and the significant difference between static and dynamic Crs (Stahl *et al.*, 2006), it is the dynamic lung mechanics that should be monitored. Many studies have shown that "optimal" PEEP may be obtained by identifying maximal dynamic Crs, or minimal Ers (Fig. 7) (Suter *et al.*, 1975, Carvalho *et al.*, 2008, Stahl *et al.*, 2006). Hickling demonstrated with a mathematical model of an ARDS lung that maximizing Crs during a decremental PEEP trial may be more suitable to indicate the "optimal" PEEP (Hickling, 2001). Several studies support this result (Suarez-Sipmann *et al.*, 2007, Hanson *et al.*, 2009).

Fig. 6. A: Airway pressure (Paw) of an ARDS patient during the SCASS maneuver to determine static compliance by automated single steps (Sydow *et al.*, 1991) and B: the corresponding P/V curves with lower inflection point (LIP) and upper inflection point (UIP) marked on both inflation and deflation limbs.

Methods other than P/V curve and maximum dynamic compliance are rarely used in clinical practice. Mols *et al.* suggested that the intra-tidal compliance-volume curve, calculated by the SLICE method (Guttmann *et al.*, 1994), was able to indicate the ongoing recruitment and overdistension of alveoli in the lung (Mols *et al.*, 1999). The shapes of the

(Dambrosio *et al.*, 1997, Hermle *et al.*, 2002). Takeuchi *et al.* proposed that setting PEEP at LIP + 2 cmH2O might be more appropriate (Takeuchi *et al.*, 2002). However, studies indicate that LIP is only the beginning of recruitment and the UIP is not a reliable marker of overdistension (Crotti *et al.*, 2001, Hickling, 2002, Downie *et al.*, 2004). New findings suggest that it may be better to set PEEP according to UIP on the deflation limb of the P/V curve (Albaiceta *et al.*, 2004). In order to obtain quasi-static P/V curves, the normal ventilation process has to be interrupted by performing a specific respiratory maneuver such as lowflow inflation (Servillo *et al.*, 1997), super-syringe inflation (Matamis *et al.*, 1984) or SCASS, i.e. static compliance by automated single steps (Sydow *et al.*, 1991), (Fig. 6A). As pointed out by LaFollette *et al.,* the key to bedside application is acquiring a dynamic curve, which is

Respiratory system compliance Crs or elastance Ers (Crs=1/Ers) can be measured quasistatically by airway occlusion (D'Angelo *et al.*, 1991) or dynamically by applying linear leastsquares regression on the first order equation of motion (Eq. 1) (Iotti *et al.*, 1995). Considering the limitation of static Crs (Stenqvist *et al.*, 2008) and the significant difference between static and dynamic Crs (Stahl *et al.*, 2006), it is the dynamic lung mechanics that should be monitored. Many studies have shown that "optimal" PEEP may be obtained by identifying maximal dynamic Crs, or minimal Ers (Fig. 7) (Suter *et al.*, 1975, Carvalho *et al.*, 2008, Stahl *et al.*, 2006). Hickling demonstrated with a mathematical model of an ARDS lung that maximizing Crs during a decremental PEEP trial may be more suitable to indicate the "optimal" PEEP (Hickling, 2001). Several studies support this result (Suarez-Sipmann *et al.*,

Fig. 6. A: Airway pressure (Paw) of an ARDS patient during the SCASS maneuver to determine static compliance by automated single steps (Sydow *et al.*, 1991) and B: the corresponding P/V curves with lower inflection point (LIP) and upper inflection point (UIP)

Methods other than P/V curve and maximum dynamic compliance are rarely used in clinical practice. Mols *et al.* suggested that the intra-tidal compliance-volume curve, calculated by the SLICE method (Guttmann *et al.*, 1994), was able to indicate the ongoing recruitment and overdistension of alveoli in the lung (Mols *et al.*, 1999). The shapes of the

marked on both inflation and deflation limbs.

easier and more applicable, instead of a static one (LaFollette *et al.*, 2007).

2007, Hanson *et al.*, 2009).

Fig. 7. Mean dynamic respiratory system compliances (Crs) of a sedated patient under mechanical ventilation at different PEEP levels. Dashed line indicates the "optimal" PEEP level, at which Crs is maximum.

compliance-volume curves are classified into three categories: 1) a decrease in slope indicates overdistension; 2) an increase in slope indicates recruitment; 3) a quasi-horizontal compliance-volume curve indicates a suitable PEEP setting (Mols *et al.*, 1999). However, the method has not yet been evaluated for clinical relevance. Ranieri *et al.* used the pressuretime curve as an index to predict pulmonary stress (Ranieri *et al.*, 2000). This method requires phases with constant air flow which limits its applicability. Nevertheless, these methods have brought the importance of pulmonary mechanical stress into focus. Talmor *et al.* estimated the transpulmonary pressure with help of esophageal balloon catheters and set PEEP to such a level that transpulmonary pressure stayed between 0 and 10 cmH2O during end-expiratory occlusion, and less than 25 cmH2O during end-inspiratory occlusion (Talmor *et al.*, 2008). They observed improvement of PaO2/FiO2 ratio and Crs compared to the group guided according to the ARDS network standard-of-care recommendations. This finding is interesting, but the placement of esophageal balloon catheters needs additional effort in clinical care. Therefore, this method will only become accepted if advantages over other methods using Crs and blood gas analysis are outweighing the extra burden of complex handling.

#### **2.4 Optimizing PEEP with imaging techniques**

CT has a very good spatial resolution and is able to show the distribution of the tissue density in the chest, thereby providing primarily morphological data. Hence, CT is the gold standard for assessment of tidal volume distribution in injured lungs and many validation studies were done by comparing various methods with the CT results (Gattinoni *et al.*, 2006, Carvalho *et al.*, 2008, Suarez-Sipmann *et al.*, 2007, Meier *et al.*, 2008). Figure 8 shows two chest CT images of a patient under mechanical ventilation at two different PEEP levels. Higher aeration and reversal of lung collapse in the dependent lung regions were detected at a PEEP level of 15 cmH2O (Fig. 8, right) compared to that of 5 cmH2O (Fig. 8, left).

Optimizing Perioperative Ventilation Support with

Adequate Settings of Positive End-Expiratory Pressure 363

homogeneous distribution of regional Crs and ventilation was observed in healthy, injured and surfactant-treated lungs (Dargaville *et al.*, 2010). Zhao and colleagues applied the global inhomogeneity (GI) index (Zhao *et al.*, 2009) to facilitate the PEEP titration in mechanically ventilated patients undergoing orthopedic surgery (Zhao *et al.*, 2010) (Fig. 9). Lowhagen et al proposed the assessment of intratidal ventilation distribution using EIT to identify optimal PEEP level in patients with ALI/ARDS (Lowhagen *et al.*, 2010). These results are promising but they still need to be confirmed in further larger studies on lung injured patients. Other EIT-based methods assessing regional lung filling characteristics (Grant *et al.*, 2009, Hinz *et al.*, 2007) have also shown potential to guide PEEP setting. As stated by Dueck in a review article, EIT is helpful in achieving the balance between alveolar recruitment and hyperinflation for patients with severe lung injury (Dueck, 2006). Although the use of EIT is limited to scientific research and clinical experiments, EIT has the potential to gain acceptance from more physicians and become a useful tool in clinical routine in the future.

Fig. 9. PEEP titration guided by ventilation homogeneity based on electrical impedance tomography (EIT) (Zhao *et al.*, 2010). Top: EIT images at different PEEP levels (from left to right: PEEP = 0, 15, 28 cmH2O). The color bars at the right side of each image indicate the magnitude of change in relative impedance during ventilation. Lung regions with the highest ventilation are coded in red. Bottom: global inhomogeneity (GI) index (Zhao *et al.*, 2009) versus PEEP. The minimum value of GI index implies the PEEP level at which

Individualized PEEP titration is important, especially in patients with severe lung injury (Kallet and Branson, 2007). Methods discussed in this chapter focused on different aspects: Crs and P/V curves represent the global mechanical properties of the respiratory system;

ventilation distribution is most homogeneous.

**2.5 Combination of PEEP setting indices** 

Fig. 8. CT images of a patient under mechanical ventilation at two different PEEP levels. Left: PEEP=5 cmH2O; Right: PEEP=15 cmH2O. Pneumonic infiltrates in the left lung and intrapleural fluid accumulation are discernible. Lung aeration is increased, e.g. in the anterior region of the right lung (white circle) and dependent lung regions are recruited (black circles) at the higher PEEP level.

Unfortunately, application of CT imaging for bedside monitoring is limited due to complex handling (e.g. large equipment) and radiation exposure of patients. Hertzog *et al.* have reported a case study using mobile CT scanners to optimize PEEP in a 6-month-old premature infant (Hertzog *et al.*, 2001). However, even with the development of low-dose CT, radiation exposure makes it practically impossible to use CT to guide PEEP titration at the bedside.

In contrast to CT, the relatively new imaging technique, electrical impedance tomography (EIT) is noninvasive and radiation-free. EIT utilizes the phenomenon that changes in regional air content modify electrical impedance of lung tissue (Nopp *et al.*, 1993). Small alternating electrical currents are applied at the chest wall surface during examination and the resultant potential differences are measured. The distribution of electrical impedance within the chest can be determined from these data. Although EIT has a relatively low resolution, it has the potential to monitor regional lung aeration and to visualize regional ventilation distribution dynamically at the bedside (Zhao *et al.*, 2010). Thus, EIT may provide additional information to individualize protective ventilation strategies by titrating PEEP.

Several applications of EIT for selecting PEEP were recently proposed. Erlandsson and colleagues used EIT to set PEEP in morbidly obese patients by maintaining a stable endexpiratory lung volume, and suggested that the corresponding PEEP level was optimal (Erlandsson *et al.*, 2006). Although the PaO2/FiO2 ratio and Crs increased in these patients, this "optimal" PEEP need not be the best oxygenation point. Besides, the identification of stable, horizontal end-expiratory EIT values may be difficult. Luepschen and colleagues modified the centre of gravity index from Frerichs *et al.* (Frerichs *et al.*, 1998) to evaluate functional lung opening and overdistension of the lung tissue in an animal model of lavageinduced acute lung injury (Luepschen *et al.*, 2007). Dargaville *et al.* have applied EIT during an incremental and decremental PEEP trial to identify the PEEP level at which the most

 Fig. 8. CT images of a patient under mechanical ventilation at two different PEEP levels. Left: PEEP=5 cmH2O; Right: PEEP=15 cmH2O. Pneumonic infiltrates in the left lung and intrapleural fluid accumulation are discernible. Lung aeration is increased, e.g. in the anterior region of the right lung (white circle) and dependent lung regions are recruited

Unfortunately, application of CT imaging for bedside monitoring is limited due to complex handling (e.g. large equipment) and radiation exposure of patients. Hertzog *et al.* have reported a case study using mobile CT scanners to optimize PEEP in a 6-month-old premature infant (Hertzog *et al.*, 2001). However, even with the development of low-dose CT, radiation exposure makes it practically impossible to use CT to guide PEEP titration at

In contrast to CT, the relatively new imaging technique, electrical impedance tomography (EIT) is noninvasive and radiation-free. EIT utilizes the phenomenon that changes in regional air content modify electrical impedance of lung tissue (Nopp *et al.*, 1993). Small alternating electrical currents are applied at the chest wall surface during examination and the resultant potential differences are measured. The distribution of electrical impedance within the chest can be determined from these data. Although EIT has a relatively low resolution, it has the potential to monitor regional lung aeration and to visualize regional ventilation distribution dynamically at the bedside (Zhao *et al.*, 2010). Thus, EIT may provide additional information to individualize protective ventilation strategies by titrating

Several applications of EIT for selecting PEEP were recently proposed. Erlandsson and colleagues used EIT to set PEEP in morbidly obese patients by maintaining a stable endexpiratory lung volume, and suggested that the corresponding PEEP level was optimal (Erlandsson *et al.*, 2006). Although the PaO2/FiO2 ratio and Crs increased in these patients, this "optimal" PEEP need not be the best oxygenation point. Besides, the identification of stable, horizontal end-expiratory EIT values may be difficult. Luepschen and colleagues modified the centre of gravity index from Frerichs *et al.* (Frerichs *et al.*, 1998) to evaluate functional lung opening and overdistension of the lung tissue in an animal model of lavageinduced acute lung injury (Luepschen *et al.*, 2007). Dargaville *et al.* have applied EIT during an incremental and decremental PEEP trial to identify the PEEP level at which the most

(black circles) at the higher PEEP level.

the bedside.

PEEP.

homogeneous distribution of regional Crs and ventilation was observed in healthy, injured and surfactant-treated lungs (Dargaville *et al.*, 2010). Zhao and colleagues applied the global inhomogeneity (GI) index (Zhao *et al.*, 2009) to facilitate the PEEP titration in mechanically ventilated patients undergoing orthopedic surgery (Zhao *et al.*, 2010) (Fig. 9). Lowhagen et al proposed the assessment of intratidal ventilation distribution using EIT to identify optimal PEEP level in patients with ALI/ARDS (Lowhagen *et al.*, 2010). These results are promising but they still need to be confirmed in further larger studies on lung injured patients. Other EIT-based methods assessing regional lung filling characteristics (Grant *et al.*, 2009, Hinz *et al.*, 2007) have also shown potential to guide PEEP setting. As stated by Dueck in a review article, EIT is helpful in achieving the balance between alveolar recruitment and hyperinflation for patients with severe lung injury (Dueck, 2006). Although the use of EIT is limited to scientific research and clinical experiments, EIT has the potential to gain acceptance from more physicians and become a useful tool in clinical routine in the future.

Fig. 9. PEEP titration guided by ventilation homogeneity based on electrical impedance tomography (EIT) (Zhao *et al.*, 2010). Top: EIT images at different PEEP levels (from left to right: PEEP = 0, 15, 28 cmH2O). The color bars at the right side of each image indicate the magnitude of change in relative impedance during ventilation. Lung regions with the highest ventilation are coded in red. Bottom: global inhomogeneity (GI) index (Zhao *et al.*, 2009) versus PEEP. The minimum value of GI index implies the PEEP level at which ventilation distribution is most homogeneous.

#### **2.5 Combination of PEEP setting indices**

Individualized PEEP titration is important, especially in patients with severe lung injury (Kallet and Branson, 2007). Methods discussed in this chapter focused on different aspects: Crs and P/V curves represent the global mechanical properties of the respiratory system;

Optimizing Perioperative Ventilation Support with

*Anesth,* 10**,** pp. 598-602.

*Engl J Med,* 351**,** pp. 327-36.

*Care Med,* 35**,** pp. 740-7.

*Med,* 34**,** pp. 2291-9.

*Appl Physiol,* 70**,** pp. 2602-10.

(PEEP). *Intensive Care Med,* 29**,** pp. 944-8.

Adequate Settings of Positive End-Expiratory Pressure 365

Baigorri, F., De Monte, A., Blanch, L., Fernandez, R., Valles, J., Mestre, J., Saura, P. & Artigas,

Beiderlinden, M., Groeben, H. & Peters, J. (2003). Safety of percutaneous dilational

Bensenor, F. E., Vieira, J. E. & Auler Junior, J. O. (2007). Thoracic sympathetic block reduces

Berendes, E., Lippert, G., Loick, H. M. & Brussel, T. (1996). Effects of positive end-expiratory

Bohm, S. H., Maisch, S., Von Sandersleben, A., Thamm, O., Passoni, I., Martinez Arca, J. &

Brochard, L., Roudot-Thoraval, F., Roupie, E., Delclaux, C., Chastre, J., Fernandez-Mondejar,

Brower, R. G., Lanken, P. N., Macintyre, N., Matthay, M. A., Morris, A., Ancukiewicz, M.,

Caramez, M. P., Kacmarek, R. M., Helmy, M., Miyoshi, E., Malhotra, A., Amato, M. B. &

Carvalho, A. R., Spieth, P. M., Pelosi, P., Vidal Melo, M. F., Koch, T., Jandre, F. C., Giannella-

Crotti, S., Mascheroni, D., Caironi, P., Pelosi, P., Ronzoni, G., Mondino, M., Marini, J. J. &

D'angelo, E., Robatto, F. M., Calderini, E., Tavola, M., Bono, D., Torri, G. & Milic-Emili, J.

Dambrosio, M., Roupie, E., Mollet, J. J., Anglade, M. C., Vasile, N., Lemaire, F. & Brochard,

alveolar recruitment and hyperinflation. *Anesthesiology,* 87**,** pp. 495-503.

failure: a clinical study. *Am J Respir Crit Care Med,* 164**,** pp. 131-40.

obstructive pulmonary disease. *Crit Care Med,* 22**,** pp. 1782-91.

respiratory system compliance. *Sao Paulo Med J,* 125**,** pp. 9-14.

distress syndrome. *Am J Respir Crit Care Med,* 174**,** pp. 268-78.

reduction in ARDS. *Am J Respir Crit Care Med,* 158**,** pp. 1831-8.

A. (1994). Hemodynamic responses to external counterbalancing of auto-positive end-expiratory pressure in mechanically ventilated patients with chronic

tracheostomy in patients ventilated with high positive end-expiratory pressure

pressure ventilation on splanchnic oxygenation in humans. *J Cardiothorac Vasc* 

Tusman, G. (2009). The effects of lung recruitment on the Phase III slope of volumetric capnography in morbidly obese patients. *Anesth Analg,* 109**,** pp. 151-9. Borges, J. B., Okamoto, V. N., Matos, G. F., Caramez, M. P., Arantes, P. R., Barros, F., Souza,

C. E., Victorino, J. A., Kacmarek, R. M., Barbas, C. S., Carvalho, C. R. & Amato, M. B. (2006). Reversibility of lung collapse and hypoxemia in early acute respiratory

E., Clementi, E., Mancebo, J., Factor, P., Matamis, D., Ranieri, M., Blanch, L., Rodi, G., Mentec, H., Dreyfuss, D., Ferrer, M., Brun-Buisson, C., Tobin, M. & Lemaire, F. (1998). Tidal volume reduction for prevention of ventilator-induced lung injury in acute respiratory distress syndrome. The Multicenter Trail Group on Tidal Volume

Schoenfeld, D. & Thompson, B. T. (2004). Higher versus lower positive endexpiratory pressures in patients with the acute respiratory distress syndrome. *N* 

Harris, R. S. (2009). A comparison of methods to identify open-lung PEEP. *Intensive* 

Neto, A. & De Abreu, M. G. (2008). Ability of dynamic airway pressure curve profile and elastance for positive end-expiratory pressure titration. *Intensive Care* 

Gattinoni, L. (2001). Recruitment and derecruitment during acute respiratory

(1991). Pulmonary and chest wall mechanics in anesthetized paralyzed humans. *J* 

L. (1997). Effects of positive end-expiratory pressure and different tidal volumes on

blood gas analysis provides a direct view on the oxygenation status; CT and EIT evaluate the local ventilation distribution. Obviously, it is rational to combine these different variables to guide PEEP titration. We suggest selecting PEEP according to a weighted combination of Crs, GI index (EIT analysis) and SpO2 (or PaO2) to include all available information on the patient's lung status. The disease state of the patient and strategic treatment goals may lead to different weighted combinations. A practical way to define these weighting factors is still warrant and should be achieved in the future with further studies.

Besides, ventilator settings, such as tidal volume (Suter *et al.*, 1978) and inspired oxygen concentration (FiO2) (Rouby *et al.*, 2002) may strongly influence the "optimal" level of PEEP. The National Institutes of Health's ARDS Network has developed a recommendation in form of a PEEP/FiO2 titration table to adjust these variables (Brower *et al.*, 2004). As mentioned before, lung protective ventilation strategies are more than just PEEP optimization. The patients will also benefit from adequate tidal volumes and body positioning which may additionally limit hyperinflation and reduce the amount of nonaerated lung tissue.
