A number of studies have investigated the use of positive end-expiratory pressure (PEEP) in the dependent lung during one-lung ventilation (OLV) for thoracic surgery. The results, however, are conflicting. Some studies reported a beneficial effect of PEEP on oxygenation [1-7], some did not [3,7,8] and some even reported detrimental effects [1,3,9]. To explain different responses to the same manoeuvre (i.e. application of PEEP), we hypothesized that differences should be searched for in patients who may differ in their underlying condition. Indeed, in a preliminary study of 18 consecutive patients we found that younger patients, with higher preoperative values of forced expiratory volume in 1 s (FEV1) showed greater changes of the pressure-volume (PV) curve shape when lateral decubitus was instituted, suggestive of greater collapse of the dependent lung. As the intraoperative changes of the PV curve were significantly correlated with preoperative FEV1, we decided to randomize 48 consecutive patients to receive PEEP (PEEP 10 cmH2O) or zero end-expiratory pressure (ZEEP) and stratify them according to their preoperative value of FEV1 (cut-off value 72%, median according to our preliminary study). The aim of the study was to test the hypothesis that PEEP applied to the dependent lung during OLV is more likely to be beneficial in patients with rather healthy lungs as judged preoperatively by FEV1, as they undergo more lung collapse in the lateral decubitus position.
The study population included 48 patients (33 males/15 females; age 63 ± 8 yr; body mass index 24.8 ± 3.3 kg m−2) scheduled for thoracic surgery (including 33 lobe resections and 7 pneumonectomies). Exclusion criteria were age less than 18 yr, severe bronchial obstruction or large bullae, lung transplantation, lung volume reduction surgery or procedures planned to be shorter than 30 min. The Institutional Ethics Committee approved the study, and informed consent was obtained from subjects.
Patients were first stratified according to their preoperative FEV1. Thereafter, if eligible, they were randomized (closed envelope) to ZEEP or PEEP groups. Four sub-groups were thus defined according to the combination of FEV1 (high or low) and PEEP (0 or 10 cmH2O). As the primary outcome was PaO2 change during PEEP application, we planned to power the study to detect a difference of about 85% of variability with a power of 0.80 using t-test for unpaired data at a significance level of 0.05 (two tailed). This led to the enrolment of 48 patients.
Anaesthesia was induced with thiopental (3-7 mg kg−1) and the patients were paralyzed with succinylcholine (1 mg kg−1) or vecuronium (0.1 mg kg−1). The patients were intubated with a left-sided double lumen catheter with the tube position confirmed by auscultation and fibreoptic bronchoscopy both in supine and lateral decubitus. General anaesthesia was maintained with isoflurane (1.5 maximum allowed concentration) plus intravenous fentanyl (50 μg boluses). Patients had central venous, radial artery and bladder catheters sited. They were ventilated with tidal volumes of 10 mL kg−1, with constant inspiratory flow and inspiratory time accounting for 33% of total time (Siemens Servo 900C; Siemens-Elema, Solna, Sweden). Respiratory rate was set to maintain arterial PCO2 within the patient's normal range. No PEEP was initially added. Ventilator parameters were kept constant between supine and lateral positions. The inspired fraction of oxygen (FiO2) in the supine position was set at 60% or higher, as needed to maintain the oxygen saturation by pulse oximetry (SPO2) greater than 92%.
Gas exchange and haemodynamics
Arterial blood was drawn for gas analysis (IL 1312 BGM; Instrumentation Laboratory Company, Lexington, MA, USA). Invasive arterial pressure and heart rate were recorded (Hewlett Packard 71034; Boeblinger, Germany). End-tidal CO2 and inspired oxygen concentration were also recorded (Ohmeda 5250 RGM; Louisville, CO, USA).
Respiratory system PV curve of the dependent hemithorax were measured as follows: with full muscle paralysis the patients were disconnected from the ventilator and the lungs allowed to deflate passively to atmospheric pressure. The dependent lung was then connected to a super-syringe (Compli 80 System; Kontron Instruments, Italy) and inflated stepwise (100 mL per step of pure oxygen with a flow rate of 1.5 L min−1) starting from atmospheric pressure and continuing to a total volume of 1 L. The lung was then allowed to deflate passively. PV curves were analysed as described elsewhere . The total system compliance was divided into starting compliance (Cstart - the ratio between the first 100 mL inflation and the corresponding pressure), and inflation compliance (Cinf - slope of the PV curve after the first 100 mL). Non-linearity of the PV curve was estimated from the Cinf/Cstart ratio close to one indicates a linear PV curve. Static compliance (Cstatic) was calculated as the ratio of expiratory tidal volume delivered to the corresponding pressure on the PV curve. Effective compliance (Ceff) was calculated from ventilator parameters as tidal volume divided by the difference between plateau and total expiratory airway pressure. Total expiratory pressure was calculated as the PEEP set according to randomization, or as intrinsic PEEP if exceeding the set PEEP.
A schematic protocol is shown in Figure 1. Blood for gas analysis and PV curve were taken in the supine position and in lateral decubitus when the chest was opened. In the supine position and the lateral position before randomization to ZEEP or PEEP, lung history was normalized (three consecutive breaths at 45 cmH2O). If PEEP was applied, the lung was deflated to PEEP from end inspiration of the last volumetric manoeuvre.
All results are presented as mean ± SD. Differences between the supine position and lateral decubitus were analysed by means of analysis of variance (ANOVA) for repeated measurements. The Student-Newman-Keuls test was applied for all pair-wise comparisons. Primary outcome (changes of PaO2 after PEEP or ZEEP) was analysed by ANOVA for repeated measurements. We first considered the effect of ZEEP or PEEP. Subsequently, we considered the effect of preoperative FEV1 on response to PEEP. This was done by stratifying patients according to FEV1 values (high or low group) and running ANOVA for repeated measurements. Fishers exact test was used to assess the rate of responders and non-responders to PEEP between patients with high or low FEV1. Linear regression was used to correlate variables. P < 0.05 was accepted as significant. Data were analysed using Sigma Stat 2.03 (Jandel Corporation; SPSS Inc., Chicago, IL, USA).
A total of 48 patients were studied. Two of them were excluded from analysis because of technical problems during data collection. Patients in the ZEEP (n = 23) and PEEP (n = 23) groups were similar (age 64 ± 9 vs. 62 ± 8 yr, body mass index 24.3 ± 4 vs. 25.3 ± 3 kg m−2, FEV1 78.9 ± 23% vs. 78.9 ± 17% of predicted value, respectively). The ratio of males to females, type of operation, endobronchial tube size, side of decubitus and ventilator settings were also similar in the two groups. Patients were recruited consecutively, therefore the number of patients was different in the two stratification groups (high vs. low FEV1). There were a total of 15 patient with low (57.3 ± 12%, range 28-72) FEV1 and 31 with high (89.6 ± 13%, range 74-135) FEV1, however, their proportion within each randomization group was similar (ZEEP: 15 high, 8 low; PEEP: 16 high, 7 low). Patient characteristics, type of interventions and ventilator settings did not differ between patients with high or low FEV1 within each randomization group (ZEEP or PEEP).
Arterial pressure, heart rate and central venous pressure did not differ between the two randomization groups and in the high/low FEV1 strata and no substantial changes were noted throughout the protocol.
Effect of lateral decubitus
Randomization groups. In the supine position, patients in the ZEEP and PEEP groups were ventilated similarly (tidal volume: 10.2 ± 0.8 mL kg−1, respiratory rate: 10.1 ± 0.7 breaths min−1) blood gases were similar in the two groups in the supine position (FiO2. 61.4 ± 5.5% vs. 63.1 ± 8.6%; PaO2: 30.2 ± 10.3 vs. 29.5 ± 9.2 kPa; PaCO2: 5.35 ± 0.75 vs. 5.29 ± 0.75 kPa, ZEEP vs. PEEP, respectively). There were no significant differences in Ceff (ZEEP: 56.9 ± 13 mL cmH2O, PEEP: 0.3 ± 13 mL cmH2O−1). When lateral decubitus was instituted, Ceff decreased and oxygenation deteriorated to a similar extent in both groups (Table 1). Intrinsic PEEP increased similarly in the two randomization groups (from 0.13 ± 0.5 cmH2O in supine position to 1.57 ± 2.2 cmH2O in lateral decubitus, P < 0.05).
FEV1stratification. When positioned in lateral decubitus all patients were characterized by impaired respiratory mechanics, as assessed by PV curve analysis (Table 2). Interestingly, the decrease in compliance was mainly due to a decrease in the Cstart, whereas the Cinf was unaffected. This resulted in an increase of Cinf/Cstart ratio, suggesting a change of shape of the PV curve. However, when lateral decubitus was instituted, patients with high FEV1 showed significantly greater changes of respiratory mechanics (Table 2 and Fig. 2). In the lateral decubitus position, patients with high FEV1 required a higher FiO2 to maintain an oxygenation similar to that of patients with low FEV1. In fact, FiO2 was increased from 61.4 ± 5.5% in the supine position to 69.3 ± 15% in the lateral position in the high FEV1 group (P < 0.05). In the low FEV1 group, the change was from 63.1 ± 8.5% to 69.5 ± 14% (not significant). Interestingly, intrinsic PEEP increased from 0.13 ± 0.5 cmH2O in the supine position to 1.13 ± 1.9 cmH2O in lateral decubitus in patients with high FEV1. It increased significantly more (P < 0.05) in patients with low FEV1 (from 0.14 ± 0.5 to 2.5 ± 2.6 cmH2O).
Effect of PEEP
Randomization groups. The application of 10 cmH2O of PEEP to the dependent lung significantly improved oxygenation compared to the control. Moreover, the application of PEEP was associated with a significant improvement of respiratory system compliance (Table 1).
FEV1stratification. The application of PEEP increased PaO2 significantly in patients with high FEV1 as compared to patients with low FEV1, despite similar FiO2 before and after treatment. Similarly, Ceff improved significantly more in patients with high FEV1 (Table 1, high vs. low FEV1). Moreover, the rate of responders to PEEP was greater in patients with high FEV1 (13/16) as compared to patients with low FEV1 (3/7, P = 0.05).
The main finding of this study was that the application of 10 cmH2O of PEEP to the dependent lung during OLV in an unselected population induces a modest improvement in oxygenation. However, this effect was mainly due to the effects of PEEP on patients with 'better lungs', as assessed by preoperative FEV1. The oxygenation improvement was also associated with changes in mechanical characteristics of the respiratory system, once again mainly observed in patients with high FEV1.
We decided to stratify the patients according to their preoperative FEV1 based on our unpublished preliminary data obtained in 18 consecutive unselected patients. In that pilot study we found that patients experienced different patterns of changes of PV curve when moved from supine to lateral position. Some of them had great decreases in compliance, whereas others did not. We also found a strong association between the decrease in compliance and the preoperative FEV1; the higher the FEV1, the greater the decrease in compliance. These patients were also younger and better oxygenated preoperatively. The available data suggest the hypothesis that the better the lung, the greater is the potential for lung collapse during anaesthesia in the open-chest lateral decubitus position. Therefore, we decided to use the preoperative FEV1 to stratify patients and, as a cut off, we choose arbitrarily the median FEV1 value (72%) of the 18 patients in the pilot study.
Overall lateral decubitus induced a decrease in compliance of the dependent hemithorax (PV curve analysis). However, this decrease was significantly greater, as in the pilot study, for the patients with higher preoperative FEV1. We did not measure pleural pressure (the oesophageal pressure in lateral decubitus would be difficult to interpret anyway). Consequently, we could not estimate the relative role of the chest wall and of the lung in decreasing respiratory system compliance. The fact that lung volume decreases during anaesthesia and paralysis in supine position is well known , with possible mechanisms being chest wall modification , upward shift of the diaphragm  and compression atelectasis . Atelectasis has also been observed in the lateral position [14,15]. Constitutional factors contribute to lung collapse during anaesthesia [16,17]. Interestingly, the appearance of compression atelectasis is limited in patients with chronic obstructive pulmonary disease, compared to normal subjects .
To explain our results we may speculate that a more flexible chest wall in healthier subjects deforms more in the lateral position, possibly leading to locally decreased transpulmonary pressure with formation of atelectasis . It is also possible that healthier lungs are more prone to collapse, leading to a decrease in respiratory system compliance as a consequence rather than the cause of atelectasis. The two phenomena may coexist. Whatever the cause, one fact holds true: healthier subjects deteriorate more in the lateral position than patients with lower baseline compliance and preoperative FEV1.
As previously discussed, conflicting results have been reported with the use of PEEP to the dependent lung during OLV. In fact, the application of PEEP in a consecutive unselected population will give highly variable results depending on the proportion of PEEP responders and non-responders. Our results show that it is possible to identify the patients more likely to respond to PEEP with a single, non-invasive preoperative test (FEV1). We do not have any direct proof of the mechanism involved in the improvement of oxygenation. However, some indirect data may help us to interpret our results. PEEP may improve oxygenation through two basic mechanisms or their combination: haemodynamic interference, resulting in redistribution/decrease of blood flow from shunted to normally aerated regions; keeping previously collapsed and perfused lung regions open. The association between the significant changes in compliance and oxygenation improvement suggest that the second mechanism is predominant. Moreover, we could not find any significant change in haemodynamics in our patients. The bulk of the data suggests that healthier subjects (with a more flexible chest wall and softer lung) undergo greater amounts of compression atelectasis in the lateral position and that PEEP may restore both oxygenation and compliance, likely associated with recruitment. This is in line with the data obtained by Slinger and colleagues .
In subjects with a lesser amount of compression atelectasis (chronic obstructive pulmonary disease and older patients) PEEP is likely to be useless. Its primary function, to keep the lung open, is not a problem in these patients. It is even possible that, in this condition, PEEP is detrimental by overdistending the lung and diverting blood flow to the dependent lung regions.
Limitations and clinical implications
The limited number of patients with low FEV1 investigated in this study prevents conclusions on this population. The greater response to PEEP in patients with a high FEV1 may seem unimportant for two reasons: first, the surgical procedure is often very swift, so that OLV is not a real problem and second, the ventilator settings are not of major concern in patients with high FEV1. However, during difficult procedures OLV may be necessary for a long time. In that setting, even minor contributions to correct hypoxaemia, such as those observed in this study, may become relevant. Given the possibly prevalent atelectasis in patients with a high FEV1, PEEP to the dependent lung is worth trying, as we have shown. It is easy and inexpensive, particularly if compared with the pharmacological adjuncts that have proved effective in this setting [20-22].
In conclusion, we have shown that PEEP to the dependent lung during OLV is more likely to improve oxygenation in patients with high as compared to low FEV1. This observation may help the clinician tailor ventilator settings during OLV in the lateral decubitus position.
1. Tarhan S, Lundborg RO. Effects of increased expiratory pressure on blood gas tensions and pulmonary shunting during thoracotomy with use of the Carlens catheter. Can Anaesth Soc J
2. Brown DR, Kafer ER, Roberson VO, Wilcox BR, Murray GF. Improved oxygenation during thoracotomy with selective PEEP to the dependent lung. Anesth Analg
3. Katz JA, Laverne RG, Fairley HB, Thomas AN. Pulmonary oxygen exchange during endobronchial anesthesia: effect of tidal volume and PEEP. Anesthesiology
4. Hedenstierna G, Baehrendtz S, Klingstedt C, et al.
Ventilation and perfusion of each lung during differential ventilation with selective PEEP. Anesthesiology
5. Baehrendtz S, Klingstedt C. Differential ventilation and selective PEEP during anaesthesia in the lateral decubitus posture. Acta Anaesthesiol Scand
6. Cohen E, Eisenkraft JB, Thys DM, Kirschner PA, Kaplan JA. Oxygenation and hemodynamic changes during one-lung ventilation: effects of CPAP 10, PEEP 10, and CPAP 10/PEEP 10. J Cardiothorac Anesth
7. Aalto-Setala M, Heinonen J, Salorinne Y. Cardiorespiratory function during thoracic anaesthesia: a comparison of two-lung ventilation and one-lung ventilation with and without PEEP 5. Acta Anaesthesiol Scand
8. Khanam T, Branthwaite MA. Arterial oxygenation during one-lung anaesthesia. 1. A study in man. Anaesthesia
9. Capan LM, Turndorf H, Patel C, Ramanathan S, Acinapura A, Chalon J. Optimization of arterial oxygenation during one-lung anesthesia. Anesth Analg
10. Gattinoni L, Pesenti A, Avalli L, Rossi F, Bombino M. Pressure-volume curve of total respiratory system in acute respiratory failure. Computed tomographic scan study. Am Rev Respir Dis
11. Strandberg A, Tokics L, Brismar B, Lundquist H, Hedenstierna G. Atelectasis during anaesthesia and in the postoperative period. Acta Anaesthesiol Scand
12. Westbrook PR, Stubbs SE, Sessler AD, Rehder K, Hyatt RE. Effects of anesthesia and muscle paralysis on respiratory mechanics
in normal man. J Appl Physiol
13. Froese AB, Bryan AC. Effects of anesthesia and paralysis on diaphragmatic mechanics in man. Anesthesiology
14. Larsson A, Jonmarker C, Jogi P, Werner O. Ventilatory consequences of the lateral position and thoracotomy in children. Can J Anaesth
15. Klingstedt C, Hedenstierna G, Lundquist H, Strandberg A, Tokics L, Brismar B. The influence of body position and differential ventilation on lung dimensions and atelectasis formation in anaesthetized man. Acta Anaesthesiol Scand
16. Strandberg A, Tokics L, Brismar B, Lundquist H, Hedenstierna G. Constitutional factors promoting development of atelectasis during anaesthesia. Acta Anaesthesiol Scand
17. Gunnarson L, Tokics L, Gustavsson H, Hedenstierna G. Influence of age on atelectasis formation and gas exchange impairment during general anaesthesia. Br J Anaesth
18. Gunnarson L, Tokics L, Lundquist H, et al.
Chronic obstructive pulmonary disease and anaesthesia: formation of atelectasis and gas exchange impairment. Eur Respir J
19. Slinger PD, Kruger M, McRae K, Winton T. Relation of the static compliance curve and positive end-expiratory pressure to oxygenation during one-lung ventilation. Anesthesiology
20. Moutafis M, Liu N, Dalibon N, et al.
The effects of inhaled nitric oxide and its combination with intravenous almitrine on PaO2
during one-lung ventilation in patients undergoing thoracoscopic procedures. Anesth Analg
21. Chen TL, Ueng TH, Huang CH, Chen CL, Huang FY, Lin CJ. Improvement of arterial oxygenation by selective infusion of prostaglandin E1
to ventilated lung during one-lung ventilation. Acta Anaesthesiol Scand
22. Rocca GD, Passariello M, Coccia C, et al.
Inhaled nitric oxide administration during one-lung ventilation in patients undergoing thoracic surgery
. J Cardiothor Vasc Anesth