Secondary Logo

Journal Logo

Original Article

Prospective, randomized, controlled evaluation of the preventive effects of positive end-expiratory pressure on patient oxygenation during one-lung ventilation

Mascotto, G.; Bizzarri, M.; Messina, M.; Cerchierini, E.; Torri, G.; Carozzo, A.; Casati, A.

Author Information
European Journal of Anaesthesiology: September 2003 - Volume 20 - Issue 9 - p 704-710


The currently used strategy to prevent hypoxaemia during one-lung ventilation includes the increase of the inspired oxygen concentration, the application of continuous positive airway pressure (CPAP) to the non-ventilated lung and the selective application of a positive end-expiratory pressure (PEEP) to the dependent, ventilated lung [1,2]. The application of CPAP to the non-dependent lung is usually effective but may interfere with surgery [3]; while the use of high values of PEEP to the dependent lung paradoxically can impair patient oxygenation by shifting blood to the non-dependent non-ventilated lung [4].

Although data on the usefulness of PEEP during one-lung ventilation are still inconclusive, Slinger and colleagues [5] recently demonstrated that titrating the value of PEEP according to the lung-chest wall compliance of each patient could improve oxygenation during one-lung ventilation. We therefore conducted a prospective, randomized, controlled study to test the hypothesis that applying a value of PEEP optimized on the lung-chest wall compliance can improve intraoperative patient oxygenation during one-lung ventilation.


With the approval of our institutional Ethics Committee and patients' written informed consent, 50 patients undergoing open chest one-lung ventilation in the lateral decubitus position were enrolled into the study. The same team of surgeons always performed all surgical procedures, which included pneumonectomy, lobectomy and atypical lung resections. Patients with severe cardiovascular or respiratory diseases or ASA physical status >III, as well as patients with contraindications to the application of PEEP or to the use of double-lumen endobronchial tube were excluded. Before surgery, all patients underwent routine pulmonary function testing including arterial blood-gas analysis. Patients with a predicted postoperative FEV1 < 40% also underwent perfusion pulmonary scintiphorography to assure a value of predicted postoperative FEV1 > 35%.

Standard premedication was given 30 min before arrival in the operating room (diazepam 0.1 mg kg−1 orally and atropine 0.01 mg kg−1 intramuscularly). Then, a thoracic epidural catheter was placed at the T4-T5 or T5-T6 interspace. After an initial test dose with lidocaine 2% 40 mg, an epidural block was induced with lidocaine 1% 5 mL and fentanyl 1 μg kg−1. Epidural block was then maintained with further 5 mL boluses of lidocaine 1% every 60 min.

General anaesthesia was induced with thiopental 5 mg kg−1 and fentanyl 1 μg kg−1 intravenously (i.v.). After adequate muscle relaxation was produced with atracurium 0.6 mg kg−1 i.v., a double-lumen endobronchial tube was placed. According to our routine practice, we used a left-sided endobronchial tube both for right and left thoracic procedures. Only two patients (one in each group) undergoing left pneumonectomy received a right-sided endobronchial tube. The adequate position of the endobronchial tube was confirmed with fibreoptic bronchoscopy before and after turning the patient to the lateral decubitus position. In case of unstable or very high airway pressure, the correct position of the endobronchial tube was checked again with fibreoptic bronchoscopy.

Patients' lungs were mechanically ventilated using a nitrous oxide in oxygen mixture (FiO2 = 0.5) with a Cato® volume cycled ventilator (Dräger, Lübeck, Germany), using a tidal volume (VT) of 9 mL kg−1, an inspiratory:expiratory time ratio of 1:1, and an inspiratory pause of 10%. The respiratory rate was adjusted to maintain an arterial CO2 partial pressure ranging between 4.6 and 5.6 kPa. The same parameters were maintained also during one-lung ventilation.

General anaesthesia was maintained with isoflurane (end-tidal concentration ranging between 0.5 and 1%), titrated to maintain systolic arterial pressure and heart rate within ±20% of baseline values. Clinical hypotension (decrease of systolic arterial pressure >30% from baseline values) was treated with volume expansion (i.v. infusion of gelatin 500 mL). Muscle relaxation was maintained with a continuous i.v. infusion of atracurium (0.01 mg kg−1 min−1).

Intraoperative monitoring included continuous electrocardiogram (leads II and V5) and heart rate, invasive arterial and central venous pressures, arterial blood-gas analysis (PaO2, PaCO2; IL BGM 1312®, Instrumentation Laboratory, Lexington, MA, USA), pulse oximetry, inspired oxygen fraction (FiO2), end-tidal CO2 (ETCO2) and end-tidal isoflurane concentrations, peak airway pressure, and minute ventilation.

Using a computer-generated sequence of numbers, patients were then randomly allocated to two groups. In the control group, no PEEP was applied (Group ZEEP, n = 22); in the treatment group, PEEP was preventively applied after the baseline observation time with two-lung ventilation and zero PEEP (Group PEEP, n = 28). The value of PEEP to be applied was determined in the absence of surgical manipulations using a proper trial with increasing PEEP steps. During each trial, the ventilatory parameters were set to maintain a constant inspiratory flow (0.2 Ls−1) with a tidal volume of 9 mL kg−1. Then, starting from zero PEEP, the PEEP was progressively increased with a first step of 3 cm H2O, followed by further steps of 2 cm H2O until no improvement in lung-chest compliance was observed with further increasing the PEEP value [6]. For each step we calculated the lung-chest wall compliance after a 2 min equilibration period before the next increase in the value of PEEP. We therefore defined as best PEEP the lowest value of PEEP resulting in the highest value of lung-chest wall compliance measured during the trial. The lung-chest wall compliance was calculated by dividing the expired tidal volume by the static plateau pressure (airway pressure with zero flow at the end of a 0.4-0.6 s end-inspiratory pause) [7]. The trial to determine the best PEEP value in patients of Group PEEP was repeated at the following time points: (1) after induction of general anaesthesia with the patient supine and two-lung ventilation; (2) when starting closed chest one-lung ventilation with the patient in the lateral decubitus position; and (3) when starting open chest one-lung ventilation. Each trial required approximately 10 min, while the values of PEEP were modified in Group PEEP after each trial according to the best PEEP value determined with the trial.

The baseline value of lung-chest wall compliance was recorded after induction of general anaesthesia during two-lung ventilation with the patient supine and zero PEEP. Lung-chest wall compliance was further measured after 10, 30 and 60 min of open chest one-lung ventilation, and then at the end of surgery with two-lung ventilation and the patient in the lateral decubitus position.

Arterial blood for blood-gas analysis was drawn after induction of general anaesthesia during two-lung ventilation and zero PEEP with the patient in the supine position (baseline), after 10 min of closed chest one-lung ventilation, after 10 and 30 min of open chest one-lung ventilation, at the end of surgery with two-lung ventilation and the patient in the lateral decubitus, and finally in the postanaesthesia care unit (PACU) with the patient receiving oxygen therapy by facemask. The ratio between arterial partial pressure of oxygen and the inspired oxygen concentration was then calculated (PaO2:FiO2).

In instances of intraoperative desaturation (SPO2 < 92%), the position of the endobronchial tube was first verified with fibreoptic bronchoscopy, and then the FiO2 was increased to 100%. If this treatment was ineffective in restoring SPO2 ≥ 92%, surgery was temporary interrupted and the non-dependent lung was re-inflated manually with 100% oxygen. The need for 100% oxygen ventilation and manual re-inflation of the non-operated lung was also recorded.

The atracurium infusion was stopped when it was estimated that there was 20 min before the last skin suture. After the last skin suture, the volatile anaesthetic agents were discontinued and residual neuromuscular block antagonized (neostigmine 2 mg and atropine 1 mg i.v.), while the lungs were ventilated manually with 100% oxygen using a fresh gas flow of 6 L min−1 until spontaneous ventilation resumed. Extubation was performed when the patient was judged to be awake (making purposeful movements), breathing regularly and with adequate oxygenation (SPO2 > 92% breathing supplementary oxygen). In the PACU, an anaesthesia nurse, blinded to the study's aim and design, evaluated the patients every 10 min until readiness to discharge. Criteria for PACU discharge included a modified Aldrete score ≥9 [8], stable vital signs, adequate airway, and being alert and responsive with pain and nausea controlled.

Postoperative analgesia consisted of a continuous epidural infusion of ropivacaine 0.2% with fentanyl 2 μg mL−1 (infusion rate ranging between 4 and 6 mL h−1) and ketoprofen 100 mg i.v. every 8 h. Rescue analgesia with tramadol 100 mg i.v. was also available if required.

In the PACU, a chest radiograph was also performed, and a radiologist, who was blinded to the use or not of PEEP during surgery, graded the severity of the injury at the non-operated lung by comparing the PACU radiograph to the preoperative one using a three-point scale, where 1 = no signs of atelectasis on the non-operated lung, with no differences from preoperative radiograph; 2 = partial atelectasis on the non-operated lung not present at the preoperative radiograph; and 3 = lobar atelectasis of the non-operated lung.

Statistical analysis

To calculate the required sample size, we took into account the mean and standard deviation of arterial partial pressure of oxygen reported during one-lung ventilation in previous investigations [5-9]. Twenty-two patients per group were required to detect a 30% difference in the intraoperative arterial partial pressure of oxygen between the two groups, accepting a two-tailed α error of 5% and a β error of 20% [10]. Statistical analysis was performed using the program Systat® 7.0 (SPSS, Inc., Chicago, IL, USA). The U-test was used to compare continuous variables, while categorical data were analysed using the contingency table analysis with Fisher's exact test. The Kaplan-Meier's log-rank analysis was also used to compare the time required to achieve PACU discharge. Unless otherwise indicated, results are the mean (±SD) or as number (percentage). P ≤ 0.05 was considered as significant.


All surgical procedures were successfully completed. No differences in anthropometric variables, preoperative pulmonary function tests, duration of surgery, intraoperative blood loss and total amount of fluid volume infused during surgery were reported between the two groups (Table 1).

Table 1
Table 1:
Demographic variables, preoperative pulmonary function test with predicted postoperative FEV1, and blood-gas variables in patients undergoing one-lung ventilation with the preventive application of PEEP (Group PEEP, n = 28) or not (Group ZEEP, n = 22)

According to the specific trial, the PEEP applied to the non-operated lung in Group PEEP was 4.3 ± 2 cmH2O after general anaesthesia induction with two-lung ventilation and the patient supine, 6.2 ± 1.8 cmH2O during closed chest one-lung ventilation, 6.0 ± 2 cmH2O during open chest one-lung (P = 0.007 and 0.005, respectively, as compared with the value measured during two-lung ventilation with the patient in supine position).

The preventive application of PEEP did not result in clinically relevant changes in heart rate or arterial pressure compared with patients in Group ZEEP. Figure 1 shows the lung-chest wall compliance measured in the two groups throughout the study. Lung-chest wall compliance decreased in both groups during one-lung ventilation compared with the baseline value. However, patients of Group PEEP had higher lung-chest wall compliance during one-lung ventilation than patients of Group ZEEP. At the end of surgery with two-lung ventilation, the values of lung-chest wall compliance were similar in the two groups but still lower than the first measurement performed before surgery with the patient supine and two-lung ventilation.

Figure 1
Figure 1:
Evolution of lung-chest wall compliance after induction of general anaesthesia with the patient supine (baseline), after 10, 30 and 60 min of one-lung ventilation with the patient in the lateral decubitus position, and then at the end of surgery in the lateral position in patients undergoing elective resection of lung neoplasm with the preventive application of PEEP (Group PEEP,n = 28) or without PEEP (Group ZEEP, n = 22). □: PEEP; ▪: ZEEP. Baseline: after general anaesthesia induction during two-lung ventilation and zero PEEP with the patient in the supine position; end of surgery: during two-lung ventilation and the patient in the lateral decubitus before completing closure of the chest wall. *P < 0.05 compared with Group ZEEP. †P < 0.05 compared with baseline during two-lung ventilation with the patient supine.

Figure 2 shows the changes in PaO2:FiO2 ratio (a) and arterial to end-tidal CO2 gradient (b) throughout the study. The PaO2:FiO2 ratio decreased in both groups during one-lung ventilation. At the end of surgery and in the PACU, the PaO2:FiO2 ratio recovered but remained lower than baseline values. During one-lung ventilation with the chest still closed, the PaO2:FiO2 ratio was higher in patients of Group ZEEP than in patients with selective PEEP. However, after the chest had been opened, no further differences in the PaO2:FiO2 ratio were observed between the two groups. The arterial to end-tidal CO2 gradient significantly increased in Group PEEP during one-lung ventilation compared with the baseline measurement during two-lung ventilation and zero PEEP; and the means measured in Group PEEP during one-lung ventilation were higher than those recorded in patients of group ZEEP (Fig. 2b). Ventilation with 100% oxygen because of an SPO2 decrease <92% was required in 10 patients of Group ZEEP (45%) and in 14 patients of Group PEEP (50%) (P = 0.78). Manual re-inflation of the operated lung was required in two patients of Group ZEEP (9%) and in three patients of Group PEEP (10%) (P = 0.78).

Figure 2
Figure 2:
Changes in the ratio between arterial partial pressure of oxygen and inspired oxygen concentration (PaO2/FiO2) (a), and arterial to end-tidal CO2 gradient (Pa-ET)CO2 gradient) (b) measured at different time points during the study in patients undergoing lung resection with (Group PEEP, n = 28) or without (Group ZEEP, n = 22) the selective application of PEEP. □: PEEP; ▪: ZEEP. The PaO2: FiO2 ratio was also calculated in the PACU with the patients spontaneously breathing, while the (Pa-ET)CO2 gradient was not calculated at this observation time. OLV-CC: one-lung ventilation with closed chest; OLV-OC: one-lung ventilation with open chest. *P < 0.05 compared with Group ZEEP; †P < 0.05 compared with baseline during two-lung ventilation with the patient supine.

PACU discharge criteria were achieved after 48 min (25th-75th percentiles: 32-58 min) in Group PEEP and 45 min (30-57 min) in Group ZEEP (P = 0.60). The radiological evaluation of the non-operated lung performed in the PACU was similar to the preoperative examination in 23 patients of Group PEEP (82%) and in 18 patients of Group ZEEP (82%) (P = 0.98). No evidences of atelectasis were reported in either group, while nine patients - five in Group PEEP (18%), four in Group ZEEP (18%) - showed partial atelectasis on the non-operated lung not present on the preoperative radiograph.


Application of CPAP to the non-ventilated operated lung is known to be effective in preventing hypoxaemia during one-lung ventilation, but it frequently interferes with surgery [1,3,11]. However, selective PEEP applied to the dependent, ventilated lung will usually not disturb the surgeon [1,2,9]. In the literature, contradictory results have been reported when a predetermined level of PEEP has been applied during one-lung ventilation in order to improve patient oxygenation [2,12,13].

To minimize cardiovascular effects of PEEP and optimize the advantages in terms of oxygenation, Carroll and colleagues [14] suggested the application of only a minimum amount of PEEP to the dependent lung during one-lung ventilation. Others have suggested that the PEEP level should be equal to the measured intrinsic PEEP [15]. More recently, in a study evaluating the relationship between static compliance curve, PEEP and oxygenation, Slinger and colleagues [5] suggested that PEEP should be applied to the dependent lung according to its mechanical properties.

The results of this prospective, randomized, controlled study demonstrated that the preventive use of an optimized and patient-specific PEEP level during one-lung ventilation resulted in a 10-15% increase in lung-chest wall compliance throughout surgery. However, this effect was not associated with a clinically relevant improvement in patient oxygenation.

The method we used to optimize the value of selective PEEP was not based on the measurement of the whole pressure-volume curve, but we selected the minimum value of PEEP resulting in the best static compliance during a stepwise trial. A similar stepwise method with incremental levels of PEEP has been described by Putensen and colleagues [6], who derived a quasi-static compliance value in patients with acute lung injury by increasing PEEP with progressive 3 cmH2O steps. In the present study, the value of lung-chest wall compliance was derived by simply dividing the expiratory tidal volume for the plateau pressure at the end of an inspiratory pause with zero flow of 0.4-0.6 s. This method may be less accurate than the other methods, such as the occlusion method, to provide a complete evaluation of pulmonary mechanics, but can be reliably used for clinical purposes [7] and has been described in previous clinical studies [16,17].

Our practice not to decrease the tidal volumes at the commencement of one-lung ventilation is widely used and accepted in the literature [1]. Nonetheless, some authors may consider excessive it or even suboptimal [1]. Tidal volumes of 10 mL kg−1 have been reported as possibly correlated with excessive inflation pressure and barotrauma to the lungs [1,18]; while the application of PEEP further increases the airway pressure, potentially overdistending the alveoli. This increases pulmonary vascular resistance in the dependent lung, shifting blood flow toward the non-ventilated, operated lung, thus increasing shunt [1]. For this reason, we used the minimum tidal volume reported to be associated with an adequate maintenance functional residual capacity during one-lung ventilation [1].

The lack of positive effects of PEEP on patient oxygenation may be related to the use of a relatively large tidal volume and an inspiratory to expiratory ratio of 1 : 1, which probably maintained an adequate distension of the ventilated lung also in patients of Group ZEEP. This could also account for the relatively low values of 'optimized' PEEP measured in patients in Group PEEP, with values of best PEEP during one-lung ventilation very close to the value of intrinsic PEEP reported by Inomata and colleagues [15] in similar clinical setting. These findings are also supported by Tokics and colleagues [19], who demonstrated that PEEP reduces the size of lung atelectasis - diagnosed with computed tomography - during inhalational anaesthesia, but not necessarily the pulmonary shunt. However, the anaesthesia technique used in this study could also be a factor. In fact, volatile anaesthetics produce a dose-dependent inhibition of hypoxic pulmonary vasoconstriction [20], and the use of an integrated epidural-general anaesthesia technique resulted in the use of lower concentration of isoflurane. On the other hand, it has been demonstrated that even though it does not have a direct influence on the primary hypoxic-induced increase in tone of the pulmonary arterioles, thoracic epidural anaesthesia enhances blood flow diversion from the non-ventilated lung to other well-oxygenated areas of the lung, thus improving arterial oxygenation [21]. Furthermore, Ishibe and colleagues [21] also demonstrated that cardiovascular effects induced by thoracic epidural anaesthesia enhance the hypoxic pulmonary vasoconstriction reflex by decreasing the venous partial pressure of oxygen.

Interestingly, the preventive use of PEEP during closed chest one-lung ventilation resulted in a clinically relevant worsening of patient oxygenation as compared with patients of Group ZEEP. This effect was reasonably related to the diversion of blood toward the non-dependent, non-ventilated lung increasing pulmonary shunt, and suggests that selective PEEP should not be applied to the dependent lung until the chest has been opened.

The application of PEEP to the dependent lung also increased the gradient between the arterial and end-tidal CO2 partial pressures, which is closely related to the physiological dead space [22]. In a dependent lung ventilated with a relatively large tidal volume and an inspiratory to expiratory ratio of 1 : 1, PEEP reasonably reduced the blood flow in over-distended lung regions, increasing the number of high VA/Q lung units compared with patients with zero PEEP [23].

In conclusion, the results of this prospective, randomized, controlled study demonstrated that although the preventive application of a selective PEEP optimized in each patient on pulmonary mechanics may improve lung-chest wall compliance during one-lung ventilation, it does not result in a clinically relevant improvement of patient oxygenation, provided a large tidal volume is given with an inspiratory : expiratory time ratio of 1 : 1. If selective PEEP is to be used to the dependent lung during one-lung ventilation, it should not be applied until the surgeon opens the chest, otherwise it may worsen the oxygenation.


1. Brodsky JB. Approaches to hypoxemia during single-lung ventilation. Curr Opin Anaesthesiol 2001; 14: 71-76.
2. 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. Acta Anaesthesiol Scand 1975; 19: 287-295.
3. Cohen E. One lung ventilation: prospective from an interested observer. Minerva Anestesiol 1999; 65: 275-283.
4. Benumof JL, Alfery DD. Anesthesia for thoracic surgery. In: Miller RD, ed. Anesthesia, 5th edn. New York, USA: Churchill Livingstone, 2000: 1665-1752.
5. 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 2001; 95: 1096-1102.
6. Putensen C, Baum M, Hormann C. Selecting ventilator settings according to variables derived from the quasi-static pressure/volume relationship in patients with acute lung injury. Anesth Analg 1993; 77: 436-447.
7. Casati A, Valentini G, Ferrari S, Senatore R, Zangrillo A, Torri G. Cardiorespiratory changes during gynaecological laparoscopy by abdominal wall elevation: a comparison with carbon dioxide pneumoperitoneum. Br J Anaesth 1997; 78: 51-54.
8. Aldrete JA, Kroulik D. A postanesthetic recovery score. Anesth Analg 1970; 49: 924-934.
9. Cohen E, Eisenkraft JB. Positive end-expiratory pressure during one-lung ventilation improves oxygenation in patients with low arterial oxygen tension. J Cardiothoracic Vasc Anesth 1996; 10: 578-582.
10. Browner WS, Black D, Newman B, et al. Estimating sample size and power. In: Designing Clinical Research - An Epidemiologic Approach. Baltimore, USA: Williams & Wilkins, 1988: 139-150.
11. Slinger P, Triolet W, Wilson J. Improving arterial oxygenation during one-lung ventilation. Anesthesiology 1988; 68: 291-295.
12. Eisenkraft JB, Cohen E, Neustein SM. Anesthesia for thoracic surgery. In: Barash PG, Cullen BF, Stoelting RK, eds. Clinical Anesthesia. Philadelphia, USA: Lippincott-Raven 1997: 779-784.
13. Katz JA, Laverne RG, Fairley HB, Thomas AN. Pulmonary oxygen exchange during endobronchial anesthesia: effects of tidal volume and PEEP. Anesthesiology 1982; 56: 164-171.
14. Carroll GC, Tuman KJ, Braverman B, et al. Minimal positive end-expiratory pressure (PEEP) may be 'best PEEP'. Chest 1988; 93: 1020-1025.
15. Inomata S, Nishikawa T, Saito S, Kihara S. 'Best' PEEP during one-lung ventilation. Br J Anaesth 1997; 78: 754-756.
16. Larsson A, Malmkvist G, Werner O. Variations in lung volume and compliance during pulmonary surgery. Br J Anaesth 1987; 59: 585-591.
17. Klingstedt C, Baehrendtz S, Bindslev L, Hedenstierna G. Lung and chest wall mechanics during differential ventilation with selective PEEP. Acta Anaesthesiol Scand 1985; 29: 716-721.
18. Rimensberger PC, Pristine G, Mullen BM, Cox PN, Slutsky AS. Lung recruitment during small tidal volume ventilation allows minimal positive end-expiratory pressure without augmenting lung injury. Crit Care Med 1999; 27: 1940-1945.
19. Tokics L, Hedenstierna G, Strandberg A, Brismar B, Lundquist H. Lung collapse and gas exchange during general anesthesia: effects of spontaneous breathing, muscle paralysis and positive end-expiratory pressure. Anesthesiology 1987; 66: 157-167.
20. Eisenkraft JB. Hypoxic pulmonary vasoconstriction. Curr Opin Anaesthesiol 1999; 12: 43-48.
21. Ishibe Y, Shiokawa Y, Umeda T, Uno H, Nakamura M, Izumi T. The effect of thoracic epidural anesthesia on hypoxic pulmonary vasoconstriction in dogs: an analysis of the pressure-flow curve. Anesth Analg 1996; 82: 1049-1055.
22. Casati A, Salvo I, Torri G, Calderini E. Arterial to end tidal carbon dioxide gradient and physiological dead space monitoring during general anesthesia: effects of patients' position. Minerva Anestesiol 1997; 63: 177-182.
23. Vigil AR, Clevenger FW. The effects of positive end-expiratory pressure of intrapulmonary shunt and ventilatory deadspace in nonhypoxic trauma patients. J Trauma 1996; 40: 618-622.


© 2003 European Academy of Anaesthesiology