Universitätsklinik Charité, Abteilung für Anaesthesiologie und operative Intensivmedizin, Schumannstrasse 20-21, D-10117 Berlin, Germany*Corresponding author
Accepted for publication: August 29, 2000
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Br J Anaesth 2001; 86: 3843
Keywords: lung, hypoxic pulmonary vasoconstriction; anaesthetics i.v., propofol; anaesthetics volatile, sevoflurane; lung, pulmonary shunt fraction; surgery, thoracic
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Clinical investigations in patients undergoing one-lung ventilation (OLV) have been less conclusive. Two studies failed to detect significant differences in arterial oxygenation between propofol and isoflurane3 or between isoflurane and sevoflurane anaesthesia.4 In contrast, Kellow and colleagues5 found significantly greater shunt fractions during isoflurane anaesthesia than during propofol anaesthesia.
Sevoflurane has useful effects during thoracic surgery. It is a potent bronchodilatator and its low bloodgas partition coefficient allows rapid adjustment of the depth of anaesthesia. Rapid emergence from anaesthesia allows rapid return of spontaneous respiration and avoids the risks of postoperative mechanical ventilation. We compared the effects of sevoflurane and propofol on pulmonary shunt fraction in patients requiring OLV for thoracic surgery.
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
All patients were premedicated with oral midazolam 0.1 mg kg1 in the ward. In the operating theatre, a radial artery catheter was placed under local anaesthesia. Etomidate 0.2 mg kg1 and fentanyl 0.10.2 mg were used to induce anaesthesia. Neuromuscular block was achieved with cisatracurium 0.15 mg kg1, followed by endobronchial intubation with a left-sided double-lumen tube (Broncho-Cath, Mallinckrodt Medical, Athlone, Ireland) in all patients. The correct position of the tube was initially confirmed by auscultation and the absence of a leak from the lumen connecting to the non-ventilated lung, and by direct observation of the atelectatic, non-ventilated lung after thoracotomy. After induction of anaesthesia, a central vein was cannulated and a flow-directed thermodilution catheter (Arrow International, Reading, PA, USA) was placed in the pulmonary artery.
Propofol was infused continuously at an initial rate of 9 mg kg1 h1 which was reduced to 6 mg kg1 h1 after 10 min. Sevoflurane was given to maintain an end-expiratory concentration of 1.8 vol%. Arterial pressure was maintained within 20% of baseline values by administration of crystalloids and fentanyl. Increments of cisatracurium were given to maintain suppression of the second twitch using a train-of-four stimulation.
We used a mixture of oxygen and air to avoid increased venous admixture from absorption atelectasis. The fractional inspired concentration of oxygen was initially set at 0.5 and adjusted to maintain arterial haemoglobin saturation above 91%, measured by pulse oximetry. During two-lung ventilation (TLV) and OLV, tidal volumes of 10 ml kg1 were used with the ventilatory rate adjusted to maintain end-tidal PCO2 at 3540 mm Hg. After blood gas analysis, the rate was adjusted to obtain an arterial PCO2 of 3545 mm Hg. During OLV, the lumen of the non-ventilated lung remained open to atmosphere; tidal volumes were decreased if peak airway pressure exceeded 35 cm of water. Positive end-expiratory pressure was not applied. The ratio of inspiratory to expiratory time was 1:2.
At the time of measurement, inspiratory and expiratory gas concentrations had been stable for 15 min and were not allowed to change by >10%. All measurements were performed before surgical manipulation of the lung. Cardiac output was measured and mixed-venous and arterial blood gas analysis done (i) after 30 min of TLV with the patient in the supine position; (ii) after 30 min of stable OLV in the supine position (OLV-1); and (iii) after opening of the pleura in the lateral decubitus position and before surgical manipulation of the lung (OLV-2). In patients undergoing thoracoscopy, trocars were left open to atmosphere at the time of measurement (OLV-2). Mixed venous and arterial blood samples were collected in duplicate. Mixed venous blood was drawn over
30 s to avoid inadvertent arterialization. The samples were analysed immediately using a blood gas analyser (ABL 505; Radiometer, Copenhagen, Denmark), which was calibrated daily according to the manufacturers instructions. The mean of two measurements was recorded. Thermodilution cardiac output was measured by forcible injection of 10 ml of normal saline at room temperature. Measurements were repeated until the difference between three successive readings was <10%. The mean of the readings was recorded as cardiac index. At the same time, haemodynamic variables were recorded, including heart rate, mean arterial and pulmonary artery pressure, pulmonary artery occlusion and central venous pressure (Marquette Electronics, Milwaukee, WI, USA). Inaccuracies in cardiac output measurements could have occurred, because the position of the pulmonary artery catheters was not verified by imaging techniques.
Calculation of shunt fraction
The shunt fraction was computed using a standard formula based on the three-compartment model proposed by Riley and colleagues:6
Qs/Qt=(Cc'O2CaO2)/(Cc'O2CvO2)
where Qs=shunt flow, Qt=cardiac output and Cc'O2, CaO2 and CvO2 represent the oxygen content of pulmonary end-capillary, arterial and mixed venous blood, respectively. Arterial and mixed venous oxygen content were calculated according to the formula
CO2=PO2x0.0031+(Hbx1.34SO2/100)
where PO2 and SO2 represent the partial pressure (mm Hg) and oxyhaemoglobin saturation in the arterial (CaO2) or mixed venous (CvO2) blood. Arterial and mixed venous PO2 and SO2 can be directly measured. In the case of Cc'O2 the relevant partial pressure and saturation have to be derived from the alveolar oxygen tension (PAO2):
PAO2=FIO2x(PBPH2O)(PaCO2/R)
where FIO2 is the fractional inspired oxygen concentration, PB is barometric pressure, PH2O is the saturated vapour pressure of water (47 mm Hg) at body temperature, PaCO2 is arterial carbon dioxide partial pressure and R is the respiratory quotient (assumed to be 0.8). Therefore, Cc'O2=PAO2x0.0031+(1.34Hb). The actual barometric pressure was recorded before each measurement. PACO2 tension was assumed to equal PaCO2 and an assumption was made that the haemoglobin would be 100% saturated.
The sample size was estimated using the data of a previous investigation.16 A difference of 8.5% in the mean increase of shunt fraction between the groups and a standard deviation of 9.3% were used for the calculation. Forty patients would be required to give 80% probability (power) of demonstrating this difference at the 5% significance level.
Statistical analysis
One-way analysis of variance (ANOVA) was used to test the difference between the means of normally distributed data. Simple linear regression was used to analyse the relationship between cardiac index (predictor variable) and shunt fraction (dependent variable). We applied best subsets linear regression analysis to assess the association between five predictor variables and the dependent variable (shunt fraction). This technique allows a systematic search of the different combinations of predictor veriables, selecting those subsets that best contribute to the variation of the dependent variable. We tested five variables: cardiac index, mean arterial pressure (MAP), mean pulmonary artery pressure (MPAP), PO2) and PaCO2. We decided not to test derived variables such as pulmonary and systemic arterial resistance, since their calculation uses haemodynamic variables included in the analysis (cardiac output, MAP and MPAP). We finally did a post hoc power analysis to assess the sensitivity of the study to detect a true difference between the groups.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Inhibition of the HPV response by inhalational anaesthetics is well established in animals. Depression of HPV has a typical sigmoid doseresponse curve with an effective dose (ED50) slightly less than twice the minimal alveolar concentration (MAC) and an ED90 of 3 MAC. There is no major difference between the volatile anaesthetics.1 7 During inhalational anaesthesia at 1 MAC (1.8 vol%) of sevoflurane, the HPV response would be reduced by approximately 25%.2
Cardiac output, mixed venous PO2 and lung perfusion
There are several reasons for the failure to reproduce these findings in the clinical context. The conflicting results obtained from human studies comparing intravenous and volatile anaesthetic agents in patients undergoing OLV have been widely attributed to haemodynamic changes, particularly the reduction of cardiac output, which may be more pronounced especially with the older volatile agents. The efficacy of HPV is inversely related to the cardiac output,8 but interactions between several physiological variables can obscure the effects of cardiac output on HPV and shunt fraction in the clinical setting. A reduction in cardiac output will decrease PO2 and increase pulmonary vasoconstriction. Volatile agents inhibit HPV by direct action, but at the same time augmentation of the HPV response occurs from decreasing cardiac output. Consequently, pulmonary shunt fraction and, by inference, HPV will appear to be unaffected.9
The HPV response is a function of both mixed venous and alveolar oxygen tension (PAO2).10 During OLV, the PAO2 of the non-ventilated, atelectatic lung can be assumed to equilibrate with the PO2. In animal experiments, low and normal P
O2 values between 25 and 46 mm Hg (3.26.1 kPa) cause maximal diversion of the blood flow (4050%) away from the collapsed lung.11 Hambraeus-Jonzon and colleagues,12 however, showed that during unilateral hypoxic ventilation (FIO2=0.05) and hyperoxic ventilation of the other lung (FIO2 =1.0) in humans, the moderate decrease in P
O2 made only a minor contribution to blood flow diversion compared with the reduction in alveolar PO2.
Variations in cardiac output not only influence PO2 but also affect pulmonary perfusion. Domino and colleagues13 demonstrated that marked increases in cardiac output will increase pulmonary perfusion and worsen ventilationperfusion mismatch; a 50% increase in pulmonary blood flow increased V/Q heterogeneity by 25%. On the other hand, low cardiac output can cause inhomogeneous distribution of blood flow in the ventilated lung with similar consequences for the ventilationperfusion ratio. The effects of changes in cardiac output and subsequent alterations in P
O2 and lung perfusionare difficult to separate during OLV, but P
O2 values are usually within the range at which maximal stimulation of the HPV response occurs. Only dramatic changes in cardiac output and pulmonary perfusion will attenuate this response. In the present study, cardiac index and P
O2 values were similar for both treatment groups and are therefore unlikely to cause differences in shunt fractions between the groups.
V/Q scatter and bronchodilation
The method we used to derive shunt fraction does not allow distinction between venous admixture resulting from perfusion of non-ventilated lung regions (true shunt) with ventilationperfusion ratios of zero (V/Q=0) and lung regions that are perfused, but poorly ventilated and therefore have low, but not zero, V/Q ratios. Ventilationperfusion mismatch is often present in patients undergoing lung resection surgery. Sevoflurane may have caused more uniform distribution of ventilation to the dependent lung during OLV, thereby decreasing shunt fraction. This could have opposed or even overcome the increase in venous admixture from inhibition of the HPV response.
The increase in shunt fraction by direct inhibition of HPV by inhalational anaesthetics administered in concentrations of 1 MAC can be expected to be rather small.1 Despite maximal stimulation of the HPV response, there will be an obligatory shunt fraction of 25% during OLV.12 14 Attenuation of the maximal HPV response by 25% would increase shunt fraction from 25% to 30% in the presence of 1 MAC sevoflurane, provided other confounding variables remain unchanged. Thus, direct inhibition of HPV could cause a 5% increase in shunt fraction.
Shunt fraction in clinical investigations
Propofol in doses of 612 mg kg1 h1 had no significant effect on PaO2 or shunt fraction.15 Reports comparing propofol with volatile agents overestimated the effects of direct inhibition of HPV by volatile agents, since shunt fractions increased more than expected from animal data. Shunt fractions during OLV were greater in patients who had received enflurane16 (1 MAC) and isoflurane5 (1 MAC). In both studies, shunt fractions were derived from oxygenation indices in a similiar manner to the present study.
Carlsson and colleagues17 18 applied multiple inert gas elimination techniques to investigate the influence of isoflurane and enflurane on HPV in human lungs. This complex method allows accurate measurement of the true shunt fraction. The authors found a 23% increase in shunt fraction after administration of 1.5 vol% isoflurane, which corresponded to an attenuation of the HPV response by 20%. They found that effects on hypoxia-induced pulmonary vasoconstriction were almost immeasurable when isoflurane was administered in clinically used concentrations. Similarly, up to 2 MAC of enflurane caused no significant change in shunt or arterial oxygenation.
Abe and colleagues19 investigated the effects of propofol, sevoflurane and isoflurane in patients undergoing oesophageal surgery and concluded that the administration of propofol was associated with significantly improved oxygenation and lower shunt fractions. However, the experimental sequence was such that the volatile agents always preceded propofol. The reduction in shunt fraction may be because propofol was always given at a later stage of surgery and after a longer period of OLV.
The timing of the measurements and their relation to the stage of the surgical procedure is important since surgical trauma can temporarily inhibit HPV.20 We made measurements before surgical manipulation of the non-ventilated lung. Beside the direct effects of surgical trauma to lung tissues, the efficacy of HPV response may vary with the duration of OLV. Rees and colleagues found a maximal increase in shunt fraction after 30 min of OLV during enflurane anaesthesia.21 In patients who required prolonged OLV for oesophageal surgery, oxygenation improved significantly over time. Minimal PaO2 values occurred after 30 min and reached a maximum after 90 min of OLV.22 23
For our patients, simple linear regression showed a positive trend, but no significant correlation, between cardiac index and shunt fraction during OLV, consistent with the postulated inverse relationship between cardiac output and HPV response.8 Other investigators,24 however, reported improved arterial oxygenation with increasing cardiac output. Best subsets regression analysis showed that only cardiac index and PaCO2 contributed significantly to the variations in shunt fraction. Mixed venous oxygen tension, which is also thought to affect the HPV response, did not explain changes in shunt fraction. This supports the view that the attenuation of the HPV response occurs at much lower PO2 values than those usually achieved during OLV.
In summary, sevoflurane administered in clinical concentrations of 1 MAC resulted in similar changes in shunt fraction as did propofol. Cardiac index, mixed PO2 and PaCO2 did not differ between the groups. Much of the overall shunt fraction during OLV may result from sources other than the attenuation of the HPV response. Haemodynamic stability and appropriate ventilatory manoeuvres are probably far more important for achieving optimal arterial oxygenation during OLV than is the choice of the anaesthetic agent.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 Ishibe Y, Gui X, Uno H, Shiokawa J, Umeda T, Suekane K. Effects of sevoflurane on hypoxic pulmonary vasoconstriction in the perfused rabbit lung. Anesthesiology 1993; 79: 134853[ISI][Medline]
3 Reid CW, Slinger PD, Lenis S. A comparison of the effects of propofolalfentanil versus isoflurane anesthesia on arterial oxygenation during one-lung ventilation. J Cardiothorac Anesth 1996; 10: 86063[ISI]
4 Wang JYY, Russell GN, Page RD, Jackson M, Pennefather SH. Comparison of isoflurane and sevoflurane on arterial oxygenation during one lung ventilation. Br J Anaesth 1998; 81: 85053
5 Kellow NH, Scott AD, White SA, Fenech RO. Comparison of the effects of propofol and isoflurane anaesthesia on right ventricular function and shunt fraction during thoracic surgery. Br J Anaesth 1995; 75: 57882
6 Riley RL, Lilienthal JL, Proemmel DD, Franke RE. On the determination of the physiologically effective pressures of oxygen and carbon dioxide in alveolar air. Am J Physiol 1946; 147: 191[ISI]
7 Domino KB, Borowec L, Alexander CM et al. Influence of isoflurane on hypoxic pulmonary vasoconstriction in dogs. Anesthesiology 1986; 64: 42329[ISI][Medline]
8 Eisenkraft JB. Effects of anaesthetics on the pulmonary circulation. Br J Anaesth 1990; 65: 6378[ISI][Medline]
9 Nunn JF. Nunns Applied Respiratory Physiology, 4th Edn. Oxford: ButterworthHeinemann, 1993; 414
10 Marshall C, Marshall BE. Site and sensitivity for hypoxic pulmonary vasoconstriction. J Appl Physiol 1983; 55: 71116
11 Domino KB, Wetstein L, Glasser SA et al. Influence of mixed-venous oxygen tension (PO2) on blood flow to atelectatic lung. Anesthesiology 1983; 59: 42834[ISI][Medline]
12 Hambraeus-Jonzon K, Bindslev L, Mellgard AJ, Hedenstierna G. Hypoxic pulmonary vasoconstriction in human lungs: a stimulusresponse study. Anesthesiology 1997; 86: 30815[ISI][Medline]
13 Domino KB, Eisenstein BL, Tran T, Hlastala MP. Increased pulmonary perfusion worsens ventilationperfusion mismatch. Anesthesiology 1993; 79: 81721
14 Marshall BE, Marshall C, Benumof J, Saidman LR. Hypoxic pulmonary vasoconstriction: effects of lung segment size and alveolar oxygen tension. J Appl Physiol 1981; 38: 84650
15 Kerr LV, Aken HV, Vandermeersch E, Vermaut G, Lerut T. Propofol does not inhibit hypoxic pulmonary vasoconstriction in humans. J Clin Anesth 1989; 1: 2848[Medline]
16 Spies C, Zaune U, Pauli MHF, Boeden G, Martin E. Comparison of enflurane and propofol during thoracic surgery. Anaesthesist 1991; 40: 1418[ISI][Medline]
17 Carlsson AJ, Hedenstierna G, Bindslev L. Hypoxic-induced pulmonary vasoconstriction in human lungs exposed to enflurane anaesthesia. Acta Anaesthesiol Scand 1987; 31: 5762[Medline]
18 Carlsson AJ, Bindslev L, Hedenstierna G. Hypoxia-induced pulmonary vasoconstriction in the human lung: the effect of isoflurane anesthesia. Anesthesiology 1987; 66: 31216[ISI][Medline]
19 Abe K, Shimizu T, Takashina M, Yoshiya I. The effects of propofol, isoflurane and sevoflurane on oxygenation and shunt fraction during one-lung ventilation. Anesth Analg 1998; 87: 11649[Abstract]
20 Chen L, Williams JJ, Alexander CM, Roy RJ, Marshall C. The effect of pleural pressure on the hypoxic pulmonary vasoconstrictor response in closed chest dogs. Anesth Analg 1988; 67: 6739
21 Rees ID, Gaines GY. One-lung anesthesiaa comparison of pulmonary gas exchange during anesthesia with ketamine and enflurane. Anesth Analg 1984; 63: 5215[Abstract]
22 Ishikawa S, Makita K, Nakazawa K, Amaha K. Continuous intraarterial blood gas monitoring during oesophagectomy. Can J Anaesth 1998; 45: 2736[Abstract]
23 Ishikawa S. Oxygenation may improve with time during one lung ventilation. Anesth Analg 1999; 89: 258
24 Slinger P, Scott WAC. Arterial oxygenation during one-lung anesthesia: a comparison of isoflurane and enflurane. Anesthesiology 1996; 82: 94046[ISI]