Anesthetic regimens used for thoracic surgery include total-IV anesthesia (TIVA), general anesthesia (GA) with volatile anesthetics, and thoracic epidural anesthesia (TEA) combined with GA. They have different effects on hypoxic pulmonary vasoconstriction (HPV), pulmonary and systemic hemodynamics, and the incidence of hypoxemia during one-lung ventilation (1–3).
TIVA has been compared to GA with inhaled anesthetics in multiple studies with respect to oxygenation and shunt fraction during one-lung ventilation. Although it is generally accepted that volatile anesthetics inhibit HPV and may promote hypoxemia in a dose-dependent manner during one-lung ventilation (1,2), IV anesthetics including propofol inhibit HPV to a minor degree (2,4).
With TEA, Garutti et al. (3) observed higher shunt fractions and lower Pao2 values compared with TIVA. In experimental studies, TEA has not influenced HPV and has minimally influenced pulmonary and systemic hemodynamics (5–8). In addition, a metaanalysis of Ballantyne et al. (9) confirmed that clinical measures of pulmonary outcome (incidence of infections, atelectasis, or other complications) are significantly improved by epidural local anesthetic treatment.
However, the preferred regimens, TEA combined with GA and TIVA, have not been compared with respect to their intraoperative clinical relevance (HPV, hemodynamic variables, and hypoxemia). The purpose of the study was to investigate arterial oxygenation and shunt fraction during one-lung ventilation with respect to hemodynamic variables throughout surgery in patients undergoing thoracotomies with pulmonary resections.
Fifty patients, ASA physical status II–III, scheduled for elective pulmonary surgery, gave their written informed consent to participate in this prospective, controlled, randomized study approved by the local Ethics Committees. All patients required thoracotomies for pulmonary resections in the lateral decubitus position and one-lung ventilation. Preoperative history and examination, as well as laboratory results, elec-trocardiogram, chest radiograph, preoperative pul-monary function tests, and perfusion scans were obtained.
Before induction of anesthesia hemodynamic monitoring was established with a radial artery catheter contralateral to the operated side for invasive blood pressure monitoring, arterial blood gas sampling and hemoglobin determinations. A pulmonary artery catheter was inserted in the right jugular vein to the pulmonary capillary wedge position for pulmonary artery pressure (PAP) monitoring, mixed venous blood gas sampling and thermodilution cardiac output measurements. The correct position (pulmonary artery of the dependent lung) of the pulmonary artery catheter was confirmed by preoperative chest radiographs. Heart rate (HR), arterial blood pressure (systolic and diastolic), and PAP (systolic and diastolic) were continuously monitored and recorded (Solar 8000; Marquette Hellige, Freiburg, Germany). Arterial oxygen saturation (Spo2) was continuously monitored by pulse oximetry. Inspired oxygen fraction (Fio2) and end-tidal isoflurane concentration as well as end-tidal CO2 were measured (Solar 8000, Marquette Hellige). Additional monitoring in all patients included esophageal temperature, electrocardiogram, and tidal volume measurements.
All patients received 0.5–1.0 mg flunitrazepam orally 1 h before their arrival in the operating room. Anesthesia was induced in both groups with IV doses of thiopental (3–5 mg/kg), fentanyl (5–10 μg/kg), and pancuronium (0.1 mg/kg). For maintenance of anesthesia, the patients were randomized to either the TIVA group or the TEA (TEA combined with GA) group. In the TIVA group anesthesia was maintained with IV propofol at continuous infusion rates of 6–10 mg · kg-1 · h-1 and IV fentanyl (5–10 μg/kg boli intermittently until 1 h before end of surgery). In the TEA group, an epidural catheter was placed at the T6-7 or T7-8 interspace using the paramedian approach before induction of anesthesia. The epidural space was identified by the loss of resistance technique using a 10 mL-glass syringe filled with 0.9% saline. After placement of the catheter a test dose of 15 mg bupivacaine 0.5% isobar was given to exclude intrathecal position. Then the initial dose of bupivacaine 0.5% depending on the age and size of the patient (range 15–25 mg) was injected via the epidural catheter. In total, 30–40 mg bupivacaine 0.5% (test dose plus initial dose) was given before surgical incision. The level of anesthesia was determined by loss of pinprick sensation. For intraoperative use, a dosage interval of 80 min was chosen. The individual dose for every patient was titrated depending on the initial required dose, age and size (range 15–25 mg bupivacaine 0.5%). Epinephrine as addition to local anesthetics was not given with respect to possible influences on shunt fraction during one-lung ventilation. Anesthesia was maintained with an end-tidal concentration of 0.3–0.5 vol % isoflurane. Relaxation was provided with a single dose of IV pancuronium 0.05–0.15 mg/kg in both groups. Blood temperature was kept constant within 0.5°C and >35.5°C. Vasoactive drugs were not given and would have been considered an exclusion criterion. Volume treatment was controlled in both groups with crystalloids and colloids to keep the patient in stable fluid balance. Central venous pressure and pulmonary wedge pressure were not allowed to differ more than 10% from baseline values. In addition, after change of position the values were compared to previous values and taken as baseline data. If hemoglobin levels decreased below 8 g/dL, erythrocyte transfusions were administered to maintain a hemoglobin level of 10 g/dL.
After induction of anesthesia, a left-sided double-lumen tube (Broncho-cath; Mallinckrodt Inc., Argyle, NY) was inserted. The correct position of the tube was determined by auscultation and confirmed by fiberoptic bronchoscopy before and after the patient was in the lateral decubitus position. The patients’ lungs were ventilated with intermittent positive pressure (Aestiva 3000; Ohmeda GmbH, Erlangen, Germany). Ventilation was controlled with Fio2 1.0 and a tidal volume of 10 mL/kg at a rate to maintain Paco2 (arterial) within 35 to 40 mm Hg. Continuous positive airway pressure or positive end expiratory pressure ventilation were not applied before finishing step 4 of the experimental sequence.
Our experimental protocol consisted of seven steps. During each step, hemodynamic measurements were taken and arterial and mixed venous blood gases obtained.
- 1. Before induction of anesthesia while the patients were supine and breathing room air.
- 2. 20 min after induction, supine position, chest closed.
- 3. During two-lung ventilation 20 min after surgical opening of the chest, while patients were in the lateral decubitus position.
- 4. During one-lung ventilation 20 min after collapse of the nondependent lung, in the lateral decubitus position. This experimental sequence was completed before ligation of any major pulmonary vessel. There was no surgical manipulation during measurements step 2-4.
- 5. Rechanging to two-lung ventilation, lateral decubitus position.
- 6. During two-lung ventilation, supine position.
- 7. After extubation while the patients were supine and breathing 2 L/min oxygen per mask.
Each set of hemodynamic measurements, obtained at the end of expiration, consisted of HR, mean arterial pressure (MAP), central venous pressure, mean pulmonary artery pressure, pulmonary capillary wedge pressure, and cardiac output (CO). CO was measured by thermodilution technique and expressed as the mean of three consecutive measurements in one patient at each step (Δ ≤ 0.3 L/min). Arterial and mixed venous blood gases were immediately analyzed by the OSM3-Hemoxymeter (Radiometer Inc., Copenhagen, Denmark) during each set of measurements. After surgery, time to extubation (discontinuation of propofol infusion/inhaled anesthesia after step 6 measurement until extubation) was documented. Postoperative analgesic treatment was controlled and consisted of IV piritramide (Dipidolor; Janssen, Neuss, Germany) given at the request of the patient (patient-controlled analgesia, Vygon GmbH, Aachen, Germany, continuous 1–2 mg/h, boli 1.5–3 mg/lockout time 10 min) in the TIVA group and of epidural administration of 6–10 mL/h bupivacaine 0.25% via the thoracic epidural catheter in the TEA group. Pain intensity was evaluated every hour on a 100 mm pain visual analog scale ranging from 0 (no pain) to 100 mm (worst pain imaginable) the day of surgery immediately after arrival at the intensive care unit (ICU) and on the following three postoperative days. ICU stay, Acute Physiology and Chronic Health Evaluation II score as well as respiratory infections, in particular pneumonia were documented. Pneumonia was diagnosed according to Centers for Disease Control criteria (10). A pulmonary infiltrate was mandatory for the diagnosis.
Systemic vascular resistance (SVR), pulmonary vascular resistance, and arteriovenous oxygen difference (C(a-v)O2) were calculated using standard formulas. The following formulas were used to calculate venous admixture (QsQt):
- QsQt [%] = [(Cc‘O2 – CaO2)/(Cc‘O2 – CvO2)]
- Oxygen content of pulmonary capillary blood/Cc‘O2 = (Hb · 1.39) SaO2 + PAo2 · 0.0031
- Alveolar oxygen pressure/PAo2 [mm Hg] = [Fio2 · (PB–PH2O)–Paco2]
- PB = 760 mm Hg; PH2O = 47 mm Hg
- Oxygen content arterio/venous/C(a/v)O2 [mL/dL] = [Hb · 1.39 · S(a/v)O2 + P(a/v)O2 · 0.0031]
All data were expressed as median and range. Statistical analysis was performed using the Mann-Whitney U-test for determining intergroup differences. For intragroup analysis, the Friedman test for global significances was used. If there was a significant difference detected by the Friedman test globally, Wilcoxon’s signed rank sum test for matched pairs was used to analyze the difference locally. Changes in hemodynamic data were analyzed using the Bonferroni correction for multiple comparisons. Spearman correlation coefficients were calculated. A P < 0.05 and the Bonferroni corrected P, respectively, were considered significant.
Fifty patients were studied. Basic patient characteristics did not differ between groups (Table 1). Before induction (Step 1) of GA significant differences concerning hemodynamic and oxygenation variables were not observed between groups (Table 2). After induction of anesthesia during two-lung ventilation in the supine and lateral decubitus position (Steps 2–3) CO decreased significantly in the TIVA group, whereas it remained unchanged in the TEA group (Fig. 1). MAP decreased significantly in both groups after induction of anesthesia, but no differences between groups were found (Table 2). Changing from two-lung ventilation to one-lung ventilation (Step 4), Pao2 decreased significantly (P < 0.05) in both groups, but Pao2 values remained significantly (P < 0.05) higher in the TEA group compared with the TIVA group (Fig. 1). However, no case of hypoxemia (defined as Pao2 < 75 mm Hg) was observed in either group. Furthermore shunt fraction increased significantly in both groups to the same extent (Fig. 1). Changing from two-lung to one-lung ventilation, CO increased significantly in the TIVA group (Fig. 1). Returning to two-lung ventilation (Steps 5–6), MAP and SVR remained significantly higher in the TIVA group (Table 2).
After extubation (Step 7) hemodynamic as well as oxygenation variables did not differ significantly between groups except for MAP and SVR (Table 2). Time to extubation was significantly shorter in the TEA group (13 min, range 10–30 min) than in the TIVA group (45 min, range 25–60 min) in our protocol design. Postoperative pain relief was superior in the TEA group compared with the TIVA group (Fig. 2). Complications associated with the TEA technique such as bleeding, infections, or postspinal headache were not observed. Acute Physiology and Chronic Health Evaluation II scores on admission to ICU did not differ between groups. ICU stay was significantly shorter in the TEA group (median, 2 days; range 1–8 days) than in the TIVA group (median 3 days; range 1–16 days) (P < 0.04). Twenty-eight percent of the patients in the TIVA group developed pneumonia during their ICU stay compared with 12 percent of the patients in the TEA group (P < 0.16; Power 29%).
The most important findings of our study were that TEA combined with GA showed improved arterial oxygenation during one-lung ventilation compared with TIVA, that CO was maintained stable in the TEA group throughout surgery compared with the TIVA group, and that time to extubation was significantly shorter in the TEA group in our protocol design.
Improved arterial oxygenation in the TEA group was achieved despite equal shunt fractions during one-lung ventilation. Although CO was significantly lower in the TIVA than in the TEA group during two-lung ventilation, it significantly increased after one-lung ventilation. This increase even exceeded the CO values in the TEA group during one-lung ventilation. CO correlated slightly but significantly (one-lung ventilation;r = 0.34;P = 0.02) with the shunt fraction in the TIVA group. An increase in CO is usually associated with an increased shunt fraction while Pao2 is unchanged or decreased (11). Decreases in CO are associated with decreased PAP, which can potentiate HPV and reduce shunt (11). In the presence of regional atelectasis, Pao2 is significantly affected by CO (11). Previous clinical studies have shown controversial results with regard to oxygenation, shunt fraction, and hemodynamic variables during one-lung ventilation (2,4,12,13). Van Keer et al. (4) studied 10 patients requiring thoracotomy. Anesthesia was maintained with continuous IV propofol infusion (10 mg/kg/h). During two-lung ventilation and one-lung ventilation no change in CO, shunt fraction, and Pao2 was observed. This might be a result of methodological differences because one-lung ventilation measurements were started before opening the chest. Kellow et al. (12) investigated patients undergoing thoracotomy and observed a significant increase of cardiac index and shunt fraction changing from two-lung ventilation to one-lung ventilation. However, the interpretation of the shunt fraction is limited because patients were ventilated with 50% nitrous oxide in oxygen and Pao2 was not determined. Steegers et al. (13) studied 14 patients requiring lobectomy and continuous IV propofol infusion (6–9 mg/kg/h). Shunt fraction and Pao2 did not differ during one-lung ventilation compared to two-lung ventilation. Their study did not include any baseline data such as CO. Changes in these hemodynamic variables would cause secondary changes in the pulmonary circulation (12). Spies et al. (2) compared TIVA with propofol (10 mg/kg/h) versus 1 MAC enflurane in patients undergoing thoracotomy. CO and shunt fraction increased significantly changing from two-lung to one-lung ventilation, whereas Pao2 per definition decreased. This was in accordance with our results. Changing to one-lung ventilation caused significantly higher increases of CO and lower Pao2 values in the TIVA group compared with the TEA group.
TEA has not inhibited HPV in experimental studies (5,6). Ishibe et al. (5) demonstrated an enhanced HPV response and improved arterial oxygenation during one-lung ventilation with TEA in dogs, which resulted from decreased PvO2 and low CO because of sympathetic nerve activity blockade. Brimioulle et al. (6) observed increased HPV during epidural blockade but no effect from previous α- or β-blockade, suggesting that all its effects on pulmonary circulation are related to sympathetic blockade. In contrast, Garutti et al. (3) observed higher shunt fractions (39.5%) and lower Pao2 values (120 mm Hg) during one-lung ventilation in a TEA group compared with a TIVA group in patients undergoing thoracotomy. They concluded that TEA could not be recommended for thoracic surgery requiring one-lung ventilation (3). However, their study has major limitations. CO and mixed-venous oxygen tension, which are important factors for assessing the impact of HPV, were not measured (14). Venous blood gas analysis to determine shunt fraction was obtained using a central venous catheter (3). TEA was combined with propofol. Kasaba et al. (15) reported that the hypotensive effects of propofol are additive to those of epidural anesthesia. Therefore, in our study design, we decided not to use TEA in combination with propofol to avoid vasoactive support because of expected hypotension. Garutti et al. (3) used IV ephedrine only in the TEA group when systolic arterial pressure decreased to < 100 mm Hg. Ephedrine is a partial α and β agonist (16). Because of the fact that β-adrenergic subtype transcripts in lung and left ventricular porcine tissues (β1: 67/72; β2: 33/28; β3: 2/25) were found (17), it cannot be eliminated that increases in CO via β-receptor activity may be responsible for the increased shunt fraction and impaired oxygenation in the study of Garutti et al. (3). In accordance with our results, Hachenberg et al. (18), using multiple inert gas elimination to analyze ventilation-perfusion inequality showed that TEA did not influence development of shunt before and after induction of GA.
During thoracic surgery it is important to maintain the competence of HPV. This is preserved if cardiac function is maintained as close as possible to preoperative values (12). Even if MAP and SVR values were lower in the TEA group, hemodynamic stability with regard to CO was better maintained in the TEA group than in the TIVA group. These results have already been reported in preliminary clinical studies (7,8). TEA produces minor reductions in CO, HR, and blood pressure (7). Tanaka et al. (8) measured CO using the suprasternal Doppler method and the thermodilution method with a Swan-Ganz-catheter in 13 patients undergoing thoracotomy with small-dose TEA. Only minor decreases of MAP after endotracheal intubation were observed. Cardiac index and pulmonary wedge pressure were essentially unchanged during the study period. With propofol, Spies et al. (2) observed a significant decrease of MAP and CO after induction of anesthesia in patients undergoing thoracotomy. In addition, Larsen et al. (19) observed significant decreases in cardiac index attributable to a negative inotropic effect of propofol.
Extubation after surgery was performed significantly earlier in the TEA group than in the TIVA group in our protocol design. No previous study has compared TEA combined with GA versus TIVA in patients with lung resections with regard to time to extubation. Boldt et al. (20) reported extubation times of 31 ± 10 min in patients after thoracotomy with propofol and fentanyl. This is shorter than what we observed in our patients. However, in our study patients received continuous propofol infusion/inhaled isoflurane in the same dose range until Step 6 measurements were finished to have comparable hemodynamic and oxygen-transport related variables. Because of the fact that larger doses of propofol are required for thoracic surgery (2), the prolonged infusion might have accounted for the prolonged time to extubation in the TIVA group. In addition, Hughes et al. (21) reported an increased context-sensitive half-time of propofol with infusion duration. Therefore, it cannot be eliminated that accumulation might have biased the results with regard to extubation time. In addition, early extubation is clinically relevant because of a decreased incidence of pulmonary infection rates and shortened postoperative stay on the ICU resulting in reduction of costs, in particular after cardiac and thoracoabdominal surgery (22,23)
The TEA procedure may also be relevant in postoperative pain control. In our study, postoperative pain relief was superior and ICU stay shorter in the TEA group with epidural bupivacaine compared with the TIVA group with IV piritramide. This has been shown already in previous clinical studies (9,24). A cumulative metaanalysis confirmed that postoperative epidural pain control can significantly decrease the incidence of pulmonary morbidity (9).
In conclusion, both anesthetic regimens are safe intraoperatively. TEA combined with GA did not impair arterial oxygenation. However, Pao2 decreased significantly in the TIVA group compared with the TEA group. This might be attributable to the increase in CO in the TIVA group after one-lung ventilation, whereas CO remained stable in the TEA group. Therefore, patients with cardiopulmonary disease might profit from TEA combined with GA with respect to oxygenation and stable hemodynamics during one-lung ventilation.
We thank Dr. U. Mansmann, MD, representive Head of the Department of Statistical Medicine, Free University of Berlin, for his assistance in statistical analysis. We also thank Dr. Michael Martin, MD, of our Department of Anesthesiology and Intensive Care, Charité, for helping to correct the paper as a native speaker.
1. Pagel PS, John JL, Damask MC, et al. Desflurane and isoflurane produce similar alterations in systemic and pulmonary hemodynamics and arterial oxygenation in patients undergoing one-lung ventilation during thoracotomy. Anesth Analg 1998; 87: 800–7.
2. Spies C, Zaune U, Pauli G, Martin E. Comparison of enflurane and propofol during thoracic surgery. Anaesthesist 1991; 40: 14–8.
3. Garutti I, Quintana B, Olmedilla L, et al. Arterial oxygenation during one-lung ventilation: combined versus general anesthesia. Anesth Analg 1999; 88: 494–9.
4. Van Keer L, van Aken H, Vandermeersch E, Vermaut G. Propofol does not inhibit HPV in humans. J Clin Anesth 1989; 1: 284–8.
5. Ishibe Y, Shiokawa Y, Umeda T, et al. The effect of thoracic epidural anesthesia on hypoxic pulmonary vasoconstriction in dogs: an analysis of the pressure-flow curve. Anesth Analg 1996; 82: 1049–55.
6. Brimioulle S, Vachiéry JL, Brichant JF, et al. Sympathetic modulation of hypoxic pulmonary vasoconstriction in intact dogs. Cardiovascular Research 1997; 34: 384–92.
7. Blomberg S, Emanuelsson H, Ricksten S. Thoracic epidural anesthesia and central hemodynamics in patients with unstable angina pectoris. Anesth Analg 1989; 69: 558–62.
8. Tanaka K, Harada T, Dan K. Low-dose thoracic epidural anesthesia induces discrete thoracic anesthesia without reduction in cardiac output. Reg Anesth 1991; 16: 318–21.
9. Ballantyne JC, Carr DB, de Ferranti S, et al. The comparative effects of postoperative analgesic therapies on pulmonary outcome: cumulative meta-analyses of randomized, controlled trials. Anesth Analg 1998; 86: 598–612.
10. Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, Am J Infection Control 1988; 16: 128–40.
11. Cheney FW, Colley S. The effect of cardiac output on arterial blood oxygenation. Anesthesiology 1980; 52: 496–503.
12. Kellow NH, Scott AD, White SA, Feneck RO. Comparison of the effects of propofol and isoflurane anesthesia on right ventricular function and shunt fraction during thoracic surgery; Br J Anaesth 1995; 75: 578–82.
13. Steegers PA, Backs PJ. Propofol and. Alfentanil during one-lung ventilation. J Cardiothorac Anesth 1990; 4: 194–9.
14. Benumof JL, Wahrenbrock EA. Local effects of anesthetics on regional hypoxic pulmonary vasoconstriction. Anesthesiology 1975; 43: 525–32.
15. Kasaba T, Kondou O, Yoshimura Y, et al. Hemodynamic effects of induction of general anesthesia with propofol during epidural anesthesia. Can J Anaesth 1998; 45: 1061–5.
16. Vansal SS, Feller DR. Direct effect of ephedrine isomers on human beta-adrenergic receptor subtypes. Biochem Pharmacol 1999; 58: 807–10.
17. Mc Neel RL, Mersmann HJ. Distribution and quantification of βeta1-, βeta2-, and β3-adrenergic receptor subtype transcripts in porcine tissues. J Anim Sci 1999; 77: 611–21.
18. Hachenberg T, Holst D, Ebel C, et al. Effect of thoracic epidural anesthesia on ventilation-perfusion distribution and intrathoracic blood volume before and after induction of anesthesia. Acta Anesthesiol Scand 1997; 41: 1142–8.
19. Larsen R, Rathgeber J, Bagdahn A, et al. Effects of propofol on cardiovascular dynamics and coronary blood flow in geriatric patients. A comparison with etomidate. Anesthesia 1988; 43: 25–31.
20. Boldt J, Müller M, Uphus D, et al. Cardiorespiratory changes in patients undergoing pulmonary resection using different anesthetic management techniques. J Cardiothorac Vasc Anesth 1996; 10: 854–9.
21. Hughes M, Glass P, Jacobs J. Context-sensitive half time in multi compartment pharmacokinetic models for intravenous anesthetic drugs. Anesthesiology 1992; 76: 334–41.
22. Brodner G, Pogatzki E, Van Aken H, et al. A multimodal approach to control postoperative pathophysiology and rehabilitation in patients undergoing abdominothoracic esophagectomy. Anesth Analg 1998; 86: 228–34.
23. Zehr KJ, Dawson PB, Yang SC, Heitmiller RF. Standardized clinical pathways for major thoracic cases reduces hospital costs. Ann Thorac Surg 1998; 66: 914–9.
24. Azad SC, Groh J, Beyer D, et al. Continuous peridural analgesia versus patient-controlled intravenous analgesia for pain therapy after thoracotomy. Anaesthesist 2000; 49: 9–17.