Local anaesthetics are often administered during general anaesthesia and the myocardial toxicity of bupivacaine has been well defined [1,2]. Halothane and isoflurane are known to worsen the cardiac toxicity of bupivacaine by a negative inotropic effect [3-5], while several studies reported a decrease of systemic toxicity of intravenous bupivacaine by propofol, isoflurane and sevoflurane [6,7]. The increase in toxic thresholds of bupivacaine with isoflurane and sevoflurane are explained by a direct action on the myocardium and by neurogenically mediated cardiac effects [6-8]. The present study used perfused isolated rat hearts to exclude the systemic effects of sevoflurane and bupivacaine and to reveal their direct effects on myocardial contractility and rhythm.
Following the approval of our institutional committee, 30 adult male Wistar albino rats weighing 150-200 g were used. The rats were killed by cervical dislocation without the use of anaesthetic agents. The hearts were excized and immersed immediately in cold heparinized (10 000 IU L−1) modified Tyrode buffer solution (NaCl 128 mmol, KCl 4.7 mmol, CaCl2 2.36 mmol, NaHCO3 20 mmol, NaH2PO4 0.36 mmol, MgCl2 1 mmol, glucose 10 mmol), then mounted on a stainless steel cannula in a Langendorff perfusion apparatus. Perfusion was obtained with a Minipuls 3® peristaltic pump (Gilson Medical Electronics, Middleton, WI, USA) with a constant flow of 8 mL min−1; the solution was equilibrated with 95% oxygen and 5% carbon dioxide. The gas mixture was continuously bubbled through the bathing solution. To minimize evaporation, the jacketed reservoir was sealed with a thin paraffin sheet as described by Hanouz and colleagues . A heart chamber was used to prevent drying and keep the temperature constant. Less than 1 min elapsed between excision and aortic cannulation; if not, the hearts were discarded. The perfusate and bath temperature was maintained at 37°C by means of a heat exchanger.
Isovolumetric left ventricular pressure was continuously recorded with a pressure transducer connected to a thin saline-filled latex balloon size 3 (Hugo Sachs Elektronik KG, Berlin, Germany) inserted into the left ventricle through the mitral valve from a cut in the left atrium. The balloon was filled to achieve a pressure of 5 mmHg at the start of the experiment. The mean perfusion pressure was in the range 70-80 mmHg. Biopac Systems (Santa Barbara, CA, USA) MP100® data acquisition system and its peripherals provided an analogue to digital conversion and recording at 100 samples s−1 on a Pentium 200 MMX computer with a Microsoft® Windows95® operating system. Pressure recordings were obtained by a TSD104A® pressure transducer using a DA100B® general-purpose transducer amplifier. Two electrodes were placed on the aorta and apex of the heart to monitor the electrocardiogram (ECG). Real time bipolar ECG recordings - with left ventricular pressure on a different channel - were obtained using an EL400® unipolar needle electrode via an ECG100B® electrocardiogram amplifier. All measurements were obtained after saving data to files. Derivatives of pressure curves were then produced by using Acqknowledge Software® version 3.2.6 to measure dp/dt curves.
The modified multiple-column Langendorff perfusion apparatus was used as described previously ; stopcocks connected the multiple columns. In the sevoflurane group (Group S, n = 10) after stabilization for 20 min; baseline measurement, and 20 min later a second set of measurements was recorded (Stage 1); sevoflurane 1.4% (1 MAC) (Sevorane®; Abbott Laboratories, Queensborough, Kent, UK), delivered from a Sevotec® (Ohmeda, Chesham, UK), was then added to the vaporizing gas mixture (Sevo 1.4). After 20 min, the sevoflurane concentration was raised to 2.8% (2 MAC). In the bupivacaine groups, following the baseline recordings, bupivacaine-HCl (Marcain® 5%; AstraZeneca, Istanbul, Turkey) was introduced from its column. The bupivacaine doses were 5 μmol (1.44 μg mL−1) in Group B5 (n = 10), and 10 μmol (2.88 μg mL−1) in Group B10 (n = 10). Following a 20 min stabilization period (Stage 1), sevoflurane 1.4% was added to the gas mixture (Sevo 1.4) and raised to 2.8% (Sevo 2.8) as in sevoflurane group. All measurements were taken at the last minute of each 20 min experimental period.
Heart rate (beats min−1), PR duration (AV conduction time, ms), QRS duration (intraventricular conduction time, ms) and ventricular dysrhythmias were recorded. Ventricular dysrhythmias were scored as: 0 = rhythmic contraction; 1 = at least one ventricular premature beats min−1; 2 = bigeminy; 3 = three consecutive ventricular premature beats min−1; 4 = ventricular tachycardia; and 5 = ventricular fibrillation (defined by The Lambeth Conventions ). Left ventricular systolic pressure (LVSP, mmHg), contraction (+dp/dtmax mmHg s−1), and relaxation (−dp/dtmax mmHg s−1), time to reach peak systolic pressure (the time between the initiation of contraction until the left ventricular pressure reaches maximum; TPSP, ms), change in left ventricular diastolic pressure from baseline and pressure-rate product (RPP = ventricular rate × peak systolic pressure) were recorded and calculated. The coronary effluent was collected during the course of experiment for biochemical determination of creatine kinase (CK, IU L−1), creatine kinase muscle band (CKmb, IU L−1) and lactate dehydrogenase (LDH, IU L−1); Hitachi 717 Autoanalyzer® (Roche, Tokyo, Japan) and Olympus Au 800® (Olympus, Tokyo, Japan) analysers were used respectively.
All data are the mean ± SEM. Statistical analysis for parametric data was performed by repeated measures of ANOVA (for analysis within groups between periods) and post hoc comparisons were made by Dunnett's significant difference test. One-way ANOVA (for comparing groups) and post hoc the Tukey-Kramer significant difference test was used. Paired and unpaired t-tests were used when required (heart rate, PR and QRS duration in Group B10). Statistical analysis for non-parametric data was performed by Friedman's non-parametric repeated measures test and post hoc comparisons were made by Dunn's multiple comparisons. Differences among means were considered significant when P < 0.05.
Heart rate and rhythm
Heart rate was stable and regular in Group S at all times and there was no change in the duration of the PR interval (P > 0.05) (Table 1). In Group B5, heart rate decreased 59% from the basal rate of 314 beats min−1 after the addition of bupivacaine. A further decrease occurred with the addition of sevoflurane reaching a maximum of 67% (P = 0.0001) (Table 1). The addition of bupivacaine 5 μmol caused ventricular dysrhythmias (score 1-4 - one of each). In Group B10, the changes in heart rate were not significant. The addition of bupivacaine 10 μmol caused ventricular tachycardia and fibrillation (score 4 and 5) in nine (90%) of the hearts. This dysrhythmia persisted with the addition of sevoflurane in concentrations of 1.4 and 2.8%. In the only remaining heart, ventricular tachycardia commenced when sevoflurane 1.4% was added. The differences between groups were statistically significant at Stages 1, Sevo 1.4 and Sevo 2.8 (Friedman; P < 0.05).
Both the AV and intraventricular conduction times were stable in the sevoflurane group (Table 1). While the addition of bupivacaine 5 μmol prolonged the PR interval, the addition of sevoflurane shortened this interval in Group B5 (P < 0.05). In Group B10, the addition of bupivacaine 10 μmol shortened the PR interval, but because ventricular tachycardia and fibrillation developed, it was impossible to quantify (Table 1). The differences between groups were significant at Stages 1, Sevo 1.4 and Sevo 2.8 (P < 0.05). The QRS duration increased significantly both in the B5 and B10 groups. In Group B5, a further prolongation was observed with Sevo 2.8 (Table 1).
In Group S, only at Sevo 2.8 did the contractility variables (+dp/dtmax and LVSP) differ from the baseline measurements. In the bupivacaine groups, the addition of bupivacaine 5 or 10 μmol both lowered +dp/dtmax significantly (P < 0.05); further decreases were observed with the addition of sevoflurane 1.4 and 2.8% (Fig. 1). There were significant differences between the groups (P < 0.01) in terms of contraction and relaxation except for baseline measurements. In Group B5, the change in LVSP was not significant while LVSP fell dramatically in Group B10 with the addition of bupivacaine to the perfusate. There were significant differences between Groups B5 and B10 at Stages 1 and Sevo 2.8 (Fig. 2). While TPSP stayed constant in Group S, TPSP increased in both bupivacaine groups and later decreased with the addition of sevoflurane (P < 0.001) and the differences between groups were significant at all stages other than baseline (P < 0.05) (Fig. 2).
The change in left ventricular diastolic pressure was highest in Group B10 (11.6 ± 17.2 mmHg) and lowest in Group S (−0.06 ± 1.3 mmHg). The difference between groups was significant (P < 0.05).
Overall, cardiac function, evaluated by the rate-pressure product, revealed significant deterioration by bupivacaine in Groups B5 and B10 (67-90% decrease). There was no change in Group S at any stage. Rate-pressure product was significantly different between groups (P < 0.05) (Fig. 3).
Cellular injury was determined by CK, CKMB and LDH measurements. In all three groups, baseline CK, CPKMB and LDH concentrations were 1.9 IU L−1, 1.2-2.6 IU L−1 and 3.4-3.8 IU L−1, respectively. The changes at different periods were not significant in any of the groups (P > 0.05).
Several authors have postulated a central site of action of bupivacaine causing ventricular dysrhythmias, tachycardia and hypertension mediated by an increase in autonomic system outflow - itself caused by blockade of inhibitory GABA neurons [3,12,13]. In this study, the isolated heart was used to avoid possible indirect cardiac effects of bupivacaine and sevoflurane mediated by autonomic nervous system and hormonal influences, as well as pre- and after-load factors.
Sevoflurane, as a single agent, did not cause any changes in heart rate or rhythmicity. Heart rhythm was severely disturbed by bupivacaine, as expected. The decrease in heart rate in our study was more severe than that of Graf and colleagues (17% decrease with bupivacaine 10 μmol) . Bupivacaine blocks the sodium channels; the block is slow and reversible. The electrophysiological disturbance leads to a decreased conduction velocity throughout the conducting system [8,15]. Depression of conduction is dangerous for it can result in re-entrant dysrhythmias likely to degenerate into often fatal ventricular fibrillation. Such accidents may sometimes occur at lower concentrations than the 0.4-1.2 μg mL−1 blood concentrations in human beings; concentrations achieved after diffusion into the systemic circulation from the injection site [16,17]. A re-entrant dysrhythmia phenomenon - including ventricular fibrillation - was observed in Group B10 in our study.
Graf and colleagues reported a moderate effect of sevoflurane on dromotropy . Fukuda and colleagues and Ohmura and colleagues reported antidysrhythmic effects of sevoflurane on bupivacaine-induced dysrhythmias, but attributed this effect to the central nervous system effects of both bupivacaine and sevoflurane [6,7]. In our study, bupivacaine induced dysrhythmia by a direct action but sevoflurane had no untoward effect on ventricular dysrhythmias.
Cellular injury of the perfused heart may cause electrophysiological and mechanical changes. However, the enzyme study results exclude this possibility. Enzyme release was low within the experimental period and values were far too low to show any sign of cardiac injury. Furthermore, there was a statistically insignificant tendency for enzyme leakage to decrease with time.
Contractility was affected by sevoflurane 2 MAC, while at 1 MAC there was almost no change. The addition of bupivacaine (either 5 or 10 μmol) to the perfusate depressed the contractility variables more severely in Group B10; bupivacaine suppresses cardiac contraction mainly due to blockade of sodium channels. In human beings it has been reported that bupivacaine at high blood concentrations (3.0-4.0 μg mL−1) has a direct effect on the strength of cardiac contraction . In our study, 90% of rats in Group B10 exhibited major dysrhythmias and all after the addition of sevoflurane.
Hanouz and colleagues reported that clinically relevant concentrations of sevoflurane do not significantly modify myocardial relaxation . In our study, relaxation (−dp/dtmax) was affected by sevoflurane at 2 MAC. The smaller changes in contractility with the addition of sevoflurane are in contrast with Park and colleagues' study reporting a 65-85% decrease and that of Harkin and colleagues' study pointing to a 58% decrease [18,19].
Both A-V and intraventricular conduction, and the time to peak systolic pressure were severely affected by the lower dose of bupivacaine (Group B5). With the higher dose, re-entrant mechanisms causing ventricular fibrillation make it impossible to use these variables. The following points should be considered in the assessment of the clinical relevance of this study. The first point is the applicability of the bupivacaine doses we used to the doses used in clinical practice. Assuming a bupivacaine blood:plasma concentration ratio of about 0.75, and 75% binding of bupivacaine to plasma proteins at normal pH, the ratio of the concentration of bupivacaine in whole blood to that in protein-free in artificial perfusate is approximately 3 : 1 . Using these approximations, the range of bupivacaine 5-10 μmol (1.44-2.88 μg mL−1) in the perfusate solution at normal pH corresponds to 4.5-9 μg mL−1 bupivacaine in whole blood in human beings. This concentration range is clinically achievable, as bupivacaine 200 mg given intravascularly to a 70 kg individual would produce peak blood concentrations of bupivacaine 1.5-10.5 μg mL−1. It has been shown that the lowest blood concentrations of bupivacaine that cause cardiotoxicity in human beings may be as low as 1.5-3 μg mL−1[17,21]. The second point is that this study was performed in vitro and there is a lack of systemic effects of both drugs. The third point that should be borne in mind is that the rat myocardium has higher contractility than human myocardium: this difference makes it more difficult to extrapolate to the outcome in human beings .
The changes in the rate-pressure product-a global variable, almost summarizing the whole study-clearly showed that bupivacaine causes both rhythm and contractility disturbances in rat hearts. These effects of bupivacaine are exaggerated at high dose. Sevoflurane does not have any untoward effect on bupivacaine-induced cardiotoxicity in clinically relevant doses in isolated rat hearts. These results leads us to assume that the findings of in vivo studies reported by Fukuda and colleagues  and Ohmura and colleagues  are solely related to the systemic effects of sevoflurane in a bupivacaine-toxicity setting.
The work was supported by the Research Fund of the University of Istanbul, Project Nos 0395/301297 and 0469/280998, and Abbott Laboratories, Istanbul.
1. Albright GA. Cardiac arrest following regional anesthesia with etidocaine or bupivacaine. Anesthesiology
2. Singh P, Lee JS. Cardiovascular and central nervous system toxicity of local anesthetics. Seminars in Anesthesia, Perioperative Medicine and Pain
3. Heavner JE, Badgwell JM, Dreyden SF, Flinders C. Bupivacaine toxicity in lightly anesthetized pigs with respiratory imbalances plus or minus halothane. Reg Anesth
4. Umuroğlu T, Eti Z, Göğüş FY, Ay B, Yayci A. The effects of sevoflurane and halothane anesthesia on bupivacaine cardiotoxicity in rabbits. J Turk Soc Algology
5. Badgwell JM, Heavner JE, Kytta J. Bupivacaine toxicity in young pigs is age dependent and is affected by volatile anesthetics. Anesthesiology
6. Fukuda H, Hirabayashi YY, Shimizu R, Saitoh K, Mitsuhata H. Sevoflurane is equivalent to isoflurane for attenuating bupivacaine induced arrhythmias and seizures in rats. Anesth Analg
7. Ohmuro S, Ohta T, Yamamoto K, Kobayashi T. A comparison of the effects of propofol and sevoflurane on the systemic toxicity of intravenous bupivacaine in rats. Anesth Analg
8. Clarkson CW, Hondeghem LM. Mechanism for bupivacaine depression of cardiac conduction. Fast block of sodium channels during the action potential with slow recovery from block during diastole. Anesthesiology
9. Hanouz JL, Vivien B, Guegniaud PY, Lecarpentier Y, Coriat P, Riou B. Comparison of the effects of isoflurane and halothane on rat myocardium. Br J Anaesth
10. Bozkurt AK, Süzer Ö, Kaynar M. Benefits of supplementing St. Thomas' Hospital cardioplegic solution with tetraethyl-ammonium on functional and metabolic recovery of isolated rat hearts. Cardiovasc Surg
11. Walker MJ, Curtis MJ, Hearse DJ, et al.
The Lambeth Conventions: guidelines for the study of arrhythmias in ischaemia infarction, and reperfusion. Cardiovasc Res
12. Thomas RD, Behbehani MM, Coyle DE, Denson DD. Cardiovascular toxicity of local anesthetics an alternative hypothesis. Anesth Analg
13. Nishikawa K, Terai T, Morimotu O, Yukioka H, Asada A. Bupivacaine does not suppress cardiac sympathetic nerve activity during halothane anesthesia in the cat. Acta Anaesthesiol Scand
14. Graf BM, Martin E, Nosnkaj ZJ, Stowe DF. Stereospecific effect of bupivacaine isomers on atrioventricular conduction in the isolated perfused guinea pig heart. Anesthesiology
15. Eledjam JJ, de la Coussaya JE, Bassoul B, Brugeda J. Mechanisms of cardiac toxicity of bupivacaine. Ann Fr Reanim
16. Freysz M, Timour Q, Bertrix L, Loufoua-Moundanga J, Omar S, Faucon G. Enhancement by ischemia of the risk of cardiac disorders, especially fibrillation, in regional anesthesia with bupivacaine. Acta Anaesthesiol Scand
17. Timour Q, Gaillard P, Bui-Xuan B, Vial T, Evreux JC, Freysz M. Cardiac accidents of locoregional anesthesia: experimental study of risk factors with bupivacaine. Bull Acad Natl Med
18. Park WP, Pancrazio J, Suh CK, Lynch C. Myocardial depressant effects of sevoflurane. Anesthesiology
19. Harkin CP, Pagel PS, Kersten JR, Hetitrick D, Waltier D. Direct negative inotropic and lusitropic effects of sevoflurane. Anesthesiology
20. Bosnjak ZJ, Stowe DF, Kampine JP. Comparison of lidocaine and bupivacaine depression of sinoatrial nodal activity during hypoxia and acidosis in adult and neonatal guinea pigs. Anesth Analg
21. Knudsen K, Beckman S, Blomberg S, Sjovall J, Edvardsson N. Central nervous and cardiovascular effects of i.v. infusions of ropivacaine, bupivacaine and placebo in volunteers. Br J Anaesth