In this randomised controlled trial we examined the effects of four different levels of positive end expiratory pressure (PEEP at 0, 5, 8 or 10 cmH2O), added to the dependent lung, on respiratory profile and oxygenation during one lung ventilation. Forty-six patients were recruited to receive one of the randomised PEEP levels during one lung ventilation. We did not find significant differences in lung compliance, intra-operative or postoperative oxygenation amongst the four different groups. However, the physiological deadspace to tidal volume ventilation ratio was significantly lower in the 8 cmH2O PEEP group compared with the other levels of PEEP (p < 0.0001). We concluded that the use of PEEP (≤10 cmH2O) during one lung ventilation does not clinically improve lung compliance, intra-operative or postoperative oxygenation despite a statistically significant reduction in the physiological deadspace to tidal volume ratio.
Anaesthesia for lung resection routinely involves one lung ventilation (OLV) to provide optimum surgical operating conditions and to isolate and protect the lungs during surgery. However, one lung ventilation is known to create an obligatory right to left shunt  and increase ventilation perfusion (V/Q) mismatch, both of which are attributable to hypoxia and/or hypoxaemia. The mechanisms involved in V/Q mismatch are associated with an increase either in physiological deadspace or in areas of hypo-alveolar ventilation. It is not known whether the application of positive end expiratory pressure (PEEP) to the dependent lung alters lung compliance to improve alveolar ventilation.
Earlier observations [2–5] from the intensive care setting established that the use of adequate levels of PEEP [6–9] throughout the respiratory cycle could greatly improve oxygenation in patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) by increasing functional residual capacity and recruiting alveoli to improve lung compliance. In the present study we investigated the effects of PEEP added to the dependent lung during one lung ventilation on respiratory compliance, physiological deadspace (Vd), intra-operative and postoperative oxygenation. To evaluate adequate levels of PEEP during one lung ventilation, we investigated the effects of different levels of PEEP on these respiratory parameters.
Following Local Research Ethics Committee approval, written informed consent was obtained from each patient prior to recruitment into the study. Criteria for inclusion in the study were as follows: patients undergoing elective thoracotomy for lung resection requiring one lung ventilation and direct arterial monitoring, age ≥ 18 years, ASA II or III. Patients with ASA IV or greater or patients undergoing emergency surgery were not studied.
Respiratory parameters were measured using mainstream infrared capnography and a fixed orifice pneumotachograph connected to a respiratory mechanics monitor (CO2SMO Plus Respiratory Profile Monitor, Novametrix Medical Systems, Wallingford, CT) and computer software capable of both on-line and off-line analysis.
Patients were randomly allocated to receive one of four PEEP levels – 0, 5, 8, 10 cmH2O – using computer-generated numbers. Allocations to different levels of PEEP were concealed in opaque, sequentially numbered sealed envelopes until the patient had consented to participate in the study. All patients underwent lung spirometry prior to surgery, the spirometry was performed in the sitting position. Patient's height and weight were recorded. Premedication was administered if thought necessary by the anaesthetist. In the anaesthetic room an arterial line was inserted using local anaesthesia and an arterial blood gas sample was taken with the patient breathing room air. After connection of electrocardiography, non-invasive blood pressure monitor and pulse oximetry, anaesthesia was induced in the supine position with an adequate anaesthetic dose of propofol, fentanyl and non-depolarising muscle relaxant drugs. An appropriately sized Robertshaw double lumen tracheal tube was inserted and the correct position was confirmed using a fibreoptic scope. Anaesthesia was maintained using oxygen, air and isoflurane. Baseline ventilator settings whilst on two lung ventilation were pressure controlled ventilation (PCV) with preset inspiratory pressure to achieve a tidal volume of 7–9 ml.kg−1, 5 cmH2O PEEP, inspiratory and expiratory ratio 1 : 2, respiratory rate to maintain the end tidal carbon dioxide (EtCO2) at 4.5–5.5 kPa with Fio2 0.5. The patient was turned into the lateral decubitus position and OLV was commenced with respiratory profile monitoring. Ventilator settings on OLV were PCV with preset inspiratory pressure to achieve tidal volume 7–9 ml.kg−1, randomised PEEP level, inspiratory and expiratory ratio 1 : 2, respiratory rate to maintain EtCO2 at 4.5–5.5 kPa and Fio2 1.0. Arterial blood gases were taken every 20 min whilst on OLV and the arterial carbon dioxide (Paco2) from the blood gas sample was then entered into the CO2SMO monitor to derive the alveolar and physiological deadspace. Other details recorded included the patient's demographic data, pre-operative spirometry data, the duration of OLV, duration of surgery, the indication for surgery and the extent of surgery.
Sample size calculation predicted a minimal sample size of 10 patients per group was sufficient to achieve a power of 90% with an α error of 0.05 to detect a mean different in lung compliance of 5 ml.cmH2O−1 with standard deviation of 2.5 ml.cmH2O−1 for each PEEP level. GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, CA, statistical package) was used. The data were assessed for normality of distribution. Two-way anova was used to compare the mean values in the study groups. A p-value < 0.05 was considered statistically significant.
Forty-six patients were recruited into the study; however, four patients did not complete the study for the following reasons: monitor failure, inoperable cancer and problems with isolation of the lung. Forty-two patients were therefore included in the intention to treat analysis of the study.
The characteristics of the patients are summarised in Table 1; all of the variables were comparable. There were no significant changes in arterial blood pressure or heart rate with the application of PEEP. Analysis of total respiratory compliance (Fig. 1) at 5, 20, 40 and 60 min revealed no differences between the different levels of PEEP nor was there any change in respiratory compliance over time. The physiological deadspace to tidal volume ratio (Vd/Vt) was significantly lower (Fig. 2) in the 8 cmH2O PEEP group (p < 0.0001) than at other PEEP levels and this effect was sustained from 20 min to 40 and 60 min. There was no significant difference between arterial oxygen to fractional inspired oxygen ratio (Pao2/Fio2) with each of the different PEEP levels over time (Fig. 3).
Table 1. Characteristics of patients receiving PEEP. Values quote are mean (SD).
No. of patients
pO2 air; kPa
pCO2 air; kPa
Duration OLV; h:min
Duration of surgery; h:min
We have found that the application of PEEP 8 cmH2O to the dependent lung during OLV significantly reduced the physiological deadspace to tidal volume ratio (p < 0.0001). However, this effect was not reflected in any significant improvement in lung compliance or oxygenation. There are several explanations for this; either PEEP does not work on OLV because of an increase in the transpulmonary shunt  or the application of PEEP was not optimised to the best lung compliance .
High levels of PEEP can exert cardiovascular effects by increasing intrathoracic pressure, causing intra-alveolar vessel compression and therefore total pulmonary vascular resistance increases. Cardiac output decreases, transpulmonary shunt increases and Pao2 will decrease due to diversion of blood flow away from the ventilated dependent lung to the non-ventilated lung. This could explain why Vd is reduced at 8 cmH2O, whereas at higher levels of PEEP (10 cmH2O) there is an increase in Vd, indicating that overdistention of alveoli has occurred with a resultant an increase in wasted ventilation. It is known that reductions in cardiac output will result in an increase in Vd/Vt . However, in our study we did not measure cardiac output changes during OLV, which might have further increased our understanding of the changes in Vd.
It is known that significant reduction in lung volume and functional residual capacity occurs in the dependent lung during OLV due to the effects of anaesthesia, the weight of the mediastinum and abdominal contents. Applying PEEP would help to splint the alveoli throughout the respiratory cycle and thus improve oxygenation. However, several studies on the effects of PEEP during OLV produced conflicting evidence, some showing improvement [12–16] and others no benefit or worsening of Pao2[17–21].
Interestingly, there appears to be an identifiable patient subgroup that would benefit from PEEP – those with low arterial oxygen tensions during OLV.
At present there are no other published studies looking at the effects of PEEP on physiological deadspace during OLV. It is interesting to note that in our study PEEP of 8 cmH2O provided the lowest Vd/Vt, suggesting that the areas of hypo-alveolar ventilation are minimised.
Suter and co-workers  demonstrated that varying PEEP levels would alter the position of tidal ventilation on the pressure volume curve, resulting in an increase in compliance up to a point, where overdistention occurs and compliance decreases. Macotto and colleagues  titrated the PEEP according to the best chest–lung wall compliance for each patient to the dependent lung during OLV. They found that selective application of PEEP increases lung–chest wall compliance but did not improve patient oxygenation. By contrast, we did not find an improvement in lung–chest compliance, possibly because the application of PEEP was randomised at four different levels rather than optimised for each patient individually.
Our study has certain limitations. The anaesthetic technique was not standardised for pain relief, i.e. epidural, intrathecal morphine or intravenous morphine, and this could affect lung chest wall compliance. However, limiting patient choice of analgesia may not always be a feasible option when anaesthesia is delivered to such a heterogeneous population. Cross-over design may have minimised intersubject variation; however, this may have contributed to the difficulty in achieving an adequate number of respiratory measurements on OLV as some of the subjects were on OLV for only 1 h.
In conclusion, the application of PEEP remains appealing with its availability on most modern anaesthetic machines and the ease which it can be added to any mode of ventilation. However, PEEP seems to provide little benefit with oxygenation and lung compliance during OLV, despite improvement in physiological deadspace to tidal volume ventilation.