Peter Slinger, MD
Department of Anaesthesia
Toronto Hospital (General) and University of Toronto
Objectives:
A 60-year-old female was the driver of a car in a head-on collision. She was wearing a seat-belt, the car did not have an air-bag.
Her injuries are:
The patient is alert and had no loss of consciousness. There are no other abnormality on complete physical, laboratory or radiologic examination. The patient is scheduled for emergency left thoracotomy and repair of the descending thoracic aorta. Intra-operatively, the patient experienced arterial oxygen desaturation during one-lung ventilation that was exacerbated when the descending thoracic aorta was cross-clamped. This desaturation was resistant to manipulations of CPAP to the non-ventilated lung and PEEP to the ventilated lung.
A survey of published articles in the past 30 years, in comparable surgical populations, gives the impression that the incidence of hypoxaemia during one-lung anaesthesia with an inspired oxygen concentration (FiO2) of 1.0 has declined from levels of 20-25% in the 1970's to less than 10% today.
It is not certain what may have led to this trend, however two parallel advances in thoracic anaesthesia in the same period could affect oxygenation.
Kerr et al2 noted that the incidence of hypoxaemia was different between different surgical populations. Patients having one-lung anaesthesia for thoracotomies for non-pulmonary operations (vascular or oesophageal) had more hypoxaemia than those having pulmonary resection surgery. Of particular note is the comparatively high incidence of hypoxaemia (27%) which has been reported with thoracoscopic surgery for emphysema5. Also of note are reports of more frequent hypoxaemia in bilateral thoracoscopic procedures during surgery on the second lung.
During the course of one-lung anaesthesia mean PaO2 values using an FiO2 of 1.0 will fall from approximately 400 mmHg to a nadir of 200 mmHg at 20-30 minutes3. Since the individual values become much more widely dispersed than during two-lung ventilation, a few patients can be expected to become hypoxaemic.
The major cause of this hypoxaemia is the pulmonary ateriovenous shunt of the de-oxygenated blood through the non-ventilated lung. Factors which influence this shunt are hypoxic pulmonary vasoconstriction, gravity, the pressure differential between the thoraces and physical lung collapse. In individual clinical situations it is not possible, at present, to determine the relative contribution of each of these four factors.
Hypoxic pulmonary vasoconstriction in the non-ventilated lung6 is felt by most investigators to be the most important variable determining PaO2 during one-lung anaesthesia. The lung, unlike any other organ, vasoconstricts in response to hypoxia. HPV can decrease the regional pulmonary blood flow by up to 50% and is triggered primarily be a decrease in alveolar oxygen tension. Mixed venous oxygen tension is a lesser trigger to HPV and the HPV reflex may fail during severe falls in PaO2. HPV functions best during normal homeostasis and is inhibited by a wide variety of physical disturbances (acid/base imbalance, temperature changes, surgical trauma, etc.). There are interspecies and intraspecies variations in the HPV response that may account for such findings as differential tolerance to high altitudes. HPV is inhibited by essentially all volatile anaesthetics and there is a difference between anesthetics in the extent to which they inhibit HPV. Isoflurance seems to be less inhibitory than enflurance or halothane7. It has been estimated that 1 MAC isoflurance will inhibit 20% of the HPV response. However, this does not seem to have a clinically significant effect on arterial oxygenation during one-lung anaesthesia. During surgery in the lateral position, gravity will usually ameliorate the decrease in oxygenation due to one-lung anaesthesia. Blood flow to a lung will decrease 10% turning from the supine to the non-dependent lateral position for open thoracotomy. This benefit will be lost during supine thoracoscopic procedures or lung resections via median sternotomy. The physical collapse of the operative lung also decreases its blood flow since pulmonary vascular resistance is lowest at functional residual capacity and increases as long volume decreases.
These tendencies to redistribute perfusion to the non-operative lung are opposed by the pressure differential between the thoraces during thoracotomy. Positive pressure ventilation applied only to one-lung causes unequal pressure transmission to the pulmonary vasculative and increases the ipsilateral resistance to blood flow. Manipulating the ventilating pressures and tidal volumes during one-lung anaesthesia can improve the oxygenation for certain patients8. However, it is not yet possible to predict the optimal ventilatory settings for an individual patient. The use of a 10 ml/kg tidal volume while limiting plateau airway pressures to 25 cm H2O at end inspiration seem to be useful initial parameters for one-lung ventilation in the majority of patients.
Approximately a third of the 35-40% shunt during one-lung anaesthesia is due to ventilation-perfusion mismatch in the ventilated dependent lung. Several factors under the control of the anaesthetist can influence this dependent lung shunt.
First, an excess of intravenous crystalloids can rapidly cause desaturation of the pulmonary venous blood draining the dependent lung9. Even though the downstream vascular pressures for both lungs are identical during unilateral hypoxia, because of HPV and pre-capillary vasoconstriction, the hypoxic lung has a lower capillary pressure and less tendency to accumulated interstitial fluid. It is important to avoid intravenous fluid excess in these patients.
Second, the use of nitrous oxide will lead to increased dependent lung atelectasis since it causes greater instability of poorly ventilated lung regions than oxygen10,11.
Third, the use of very small tidal volumes in the relatively non-compliant dependent lung with its reduced functional residual capacity will increase ipsilateral pulmonary vascular resistance and promote alveolar collapse8.
Because the risk of intraoperative hypoxaemia is higher during one-lung anaesthesia for thoracic surgery than in any other types of common elective surgical procedure, it has long been recognized that a continuous method of monitoring oxygenation is needed. Pulse oximetry for one-lung ventilation was described in the 1980's and has been one of the landmarks in patient safety for thoracic anaesthesia12. However, there are limitations to the technology. Pulse oximetry is prone to errors due to various artifacts and occasional malfunctions when it is most needed13. Also, pulse oximetry does not give any early warning of the rapidly falling PaO2 that occurs during one-lung anaesthesia before any change in saturation. It is the patient whose PaO2 declines most rapidly who is most likely to become hypoxaemic. A more reliable and sensitive method of detecting oxygenation changes prior to actual desaturation is still required for safe thoracic anaesthesia.
Side-stream spirometry permits on-line monitoring of pulmonary mechanics. This technology can provide an early warning of loss of lung isolation or accidental lobar obstruction. It may be possible to use this information to select the optimal ventilatory parameters for an individual patient during OLV.
Several factors have been identified which allow prediction of the risk of hypoxaemia developing during one-lung anaesthesia.
The PaO2 during two-lung ventilation either preoperatively when breathing air or, more specifically, during ventilation with a high FiO2 in the lateral position intra-operatively correlates with the PaO2 during one-lung ventilation3. The patients who are most capable of matching regional lung perfusion to ventilation during two-lung anaesthesia retain this ability when only one lung is ventilated.
The side of lung collapse affects the PaO2 during one-lung anaesthesia. The larger right lung receives approximately 10% more pulmonary blood flow than the left. When the right lung is collapsed for surgery there is a larger shunt and a higher incidence of hypoxaemia. The nadir of mean PaO2 values during one-lung anaesthesia for right thoracotomies is approximately 165 mmHg versus 235 mmHg for left-sided surgery15. When doing bilateral procedures, such as thoracoscopy for sympathectomy or bullae it is logical to operate on the right side first. Since the oxygenation tends to be less during the second period of one-lung anaesthesia, probably due to mechanical trauma to the first lung, it is the optimal management to ventilate the lung with best gas-exchange potential during surgery to the second side.
Patients with good preoperative spirometric pulmonary function tests tend to have lower PaO2 values during one-lung anaesthesia than patients with poor spirometry. This has been demonstrated in several studies, for both forced vital capacity and forced expiratory volume in one second3,8. As yet, it is not possible to explain this almost paradoxical phenomenon with certainty. However, it is now known that patients with hyperinflation develop a higher level of intrinsic positive expiratory pressure (auto-PEEP) during one-lung ventilation16, this may lead to a more favourable FRC in the relatively non-compliant dependent lung.
Patients who preoperatively have diminished perfusion or ventilation to the operative lung, due most often to anatomic factors such as the mass effects of a tumour or a giant bulla, will tend to have better oxygenation during one-lung anaesthesia. Coversely, patients with a proportionately high perfusion or ventilation of the operative lung are at increased risk of hypoxaemia17. This information is available for some thoracic surgical patients from a preoperative radionuclide ventilation/perfusion lung scan.
Some ventilation/perfusion scans are often not available, one simple clinical method of obtaining comparable data is to examine the differential contribution of each lung to gas exchange. This can be done intraoperatively for all patients when a double-lumen tube is used. Side-stream spirometry permits the comparison of the proportional contribution of each lung to respiration during two-lung ventilation18. As with ventilation perfusion scans, if the lung to be collapsed has decreased gas-exchange function, the risk of hypoxaemia is less.
On the basis of data routinely available for most thoracic surgical patients, it is possible to estimate the risk of hypoxaemia3. For example, a patient having a right-sided thoracotomy who had relatively poor PaO2 values during two-lung ventilation but good preoperative spirometry is highly likely to desaturate during one-lung anaesthesia. Knowing which patients are most likely to desaturate allows the anaesthetist to institute prophylactic therapy at the initiation of one-lung anaesthesia.
When hypoxaemia occurs during one-lung anaesthesia, the major cause is shunt in the non-ventilated lung. However, other potential causes such as malposition of an endobronchial tube or inadequate oxygen delivery should be ruled out. Since treatment and prophylaxis of hypoxaemia in this setting have the same underlying principles, they will be discussed together.
The use of the highest possible FiO2 during one-lung anaesthesia improves oxygenation. Historically, studies using an FiO2 less than 0.9 showed a much higher incidence of hypoxaemia. However, it is not always advisable to use high FiO2 levels in these patients. The recent administration of a variety of drugs, which includes Bleomycin19, Mitomycin20, and Amiodarone21 has been associated with pulmonary oxygen toxicity when an FiO2 > 0.3 was used intraoperatively for thoracic surgery. High inspired oxygen concentrations are first-line therapy for treatment and prevention of hypoxaemia. Since FiO2 levels > 0.9 cannot always be used safely, it is important for the anaesthetist to be aware of other potential methods to avoid hypoxaemia during one-lung anaesthesia.
Continuous positive airway pressure (CPAP) to the non-ventilated lung is the other first-line of defence and treatment22. When compared to a variety of other mechanical techniques to improve oxygenation such as PEEP to the ventilated lung or simple oxygen insufflation of the non-ventilated lungs, CPAP is consistently superior. A wide variety of devices have been described to provide CPAP to the non-dependent lung. These devices are sold commercially or can be assembled simply from equipment available in most anaesthetic departments. They should incorporate: an air/oxygen source separate from the anaesthetic machine supply to the patient, a simple anaesthetic circuit (such as a Mapleson D) with a variable exhaust pressure and a manometer. It is useful if the non-ventilated lung can be re-inflated when required for surgical purposes then returned to CPAP without dismantling the circuit23.
A CPCP of FiO2 1.0 is usually used but this can be decreased using air/oxygen mixtures if required. Since the usual transpulmonary distending pressure of the lung is 5 cmH2O, a CPAP of 5 cm or higher results in a fully inflated lung which impedes surgery. Useful increases in oxygenation can be achieved with 1-2 cmH2O CPAP24.
The most important caveat regarding the use of CPAP is that it must be applied to the fully inflated lung25. Even short periods of lung collapse impair the efficiency of CPAP since the opening pressure of atelectatic lung units exceed 20 cmH2O. Because of the problems which re-inflation may cause at an in-opportune surgical moment, it is useful to predict which patients are most at risk of hypoxaemia and apply CPAP prophylactically at the onset of one-lung anaesthesia.
CPAP will not be beneficial if adequate alveolar pressure cannot be developed in the non-ventilated lung. In cases of bronchopleural fistula, bronchial obstruction, bronchial sleeve resection, etc., CPAP cannot be relied upon to improve oxygenation. During open thoracotomy the surgical access is not significantly impeded by the application of 2 cm CPAP to the lung. However, during thoracoscopy even low level CPAP may obstruct the surgical field.
The anaesthetic technique has an impact on oxygenation during one-lung ventilation. Among the volatile anaesthetics, Isoflurane has been shown to provide better PaO2 levels than Halothane or Enflurane26. At present no intravenous technique such as propofol-alfentanil27 or other infusions have been shown to be superior to 1 MAC Isoflurane.
A major variable in the choice of anaesthetic technique is maintenance of cardiac output. Previously, it was thought than an augmentation of cardiac output during one-lung anaesthesia would cause passive dilatation of the pulmonary vasculature and oppose HPV in the non-ventilated lung, thereby lowering PaO2. Although this may happen to a degree, the net effect of increasing cardiac output in this situation is to increase PaO2 via an increase in mixed venous oxygen content since these patients have such a large shunt (35-40%)26. The current popularity of combining neuraxial blockade with low dose volatile/narcotic anaesthesia for thoracic surgery may have an indirect benefit on PaO2 during one-lung anaesthesia by maintaining cardiac output.
High frequency jet ventilation (HFJV) to the operative lung provides superior oxygenation. However, HFJV tends to increase the diameter of central airways and can impede surgery during pulmonary resections. HFJV is useful for non-pulmonary intrathoracic surgery such as thoracic aortic aneurysm or oesophageal resections28.
Positive and end-expiratory pressure to the ventilated lung decreases PaO2 in the majority of patients during one-lung anaesthesia probably by exacerbating the pressure differential between the thoraces and re-distribution of pulmonary blood flow to the non-ventilated lung. However, several studies suggest that a small minority of patients, often those with the poorest PaO2 values, benefit from dependent-lung PEEP29. It is conceivable that those patients with normal lung volumes who do not develop auto-PEEP during one-lung anaesthesia may maintain a more favourable FRC in the ventilated lung with the addition of extrinsic PEEP. When oxygenation remains unsatisfactory after application of CPAP to the non-ventilated lung, a trial of adding 5 cm PEEP to the ventilated lung is warranted.
Intentional obstruction of the blood flow to the non-ventilated lung is possible in several situations to treat hypoxaemia. During pneumonectomy or lung transplantation the ipsilateral pulmonary artery can be clamped by the surgeon. Pulmonary artery balloon-tipped floatation catheters can be placed under fluoroscopic control and inflated to decrease regional pulmonary blood flow. This can be useful in patient with pulmonary arteriovenous malformations30.
Re-inflation of the non-dependent lung is a rapid and reliable method of managing hypoxaemia during one-lung anaesthesia. This can be done prophylactically every 5 minutes and may allow a procedure to continue that would otherwise have to be aborted. Due to the individual surgical circumstances, it is not always possible to re-inflate the non-ventilated lung. Depending on the case, it is sometimes viable to individualize a method of oxygen delivery to the non-ventilated lung such as with a separate catheter passed through the operative field into the distal lung during a sleeve or carinal resection.
Various pharmacological methods of modulating the unilateral pulmonary vascular tone such as prostaglandin E7 are now available and have appeared to be useful in animal studies31. The HPV response can be augmented by administration of cyclooxygenase inhibitors such as non-steroidal anti-inflammatory drugs. This will increase the resistance of HPV to the vasodilatory effects of volatile anaesthetics.
Nitric oxide (NO) is an important regulator of pulmonary blood flow. When administered to the ventilated lung regions of pigs, NO did not cause a further decrease in the blood flow to a hypoxic lobe. However, when NO was delivered to the ventilated lung regions in combination with the systemic administration of a NO synthase inhibitor (which augments HPV) the blood flow to a hypoxic lobe decreased from 27% to 3% of baseline values32. The safety and increasing availability of NO may make such pharmacological manipulation of pulmonary blood flow clinically possible. Early human studies have shown mixed results.
On-line monitoring systems using physical (Doppler) and chemical (gas exchange) data to evaluate the differential pulmonary blood flow between the ventilated and non-ventilated lungs are feasible and are currently undergoing investigation. Such systems could permit determination of ventilatory parameters to optimize tidal volume and PEEP and allow modification of anaesthetic technique in individual patients. Since the majority of patients have auto-PEEP during OLV, transoesophageal echocardiography may identify the 10% of patients who develop right-left shunt through a patent foramen ovale from PEEP during anaesthesia35.
Questions
Answers: 1(e); 2(b); 3(c); 4(b)