Continuous measurement of cardiac output during aortic cross-clamping by the oesophageal Doppler monitor ODM 1

British Journal of Anaesthesia 1995; 74: 655-660

Continuous measurement of cardiac output during aortic crossclamping by the oesophageal Doppler monitor ODM 1 K.-F. KLOTZ, S. KLINGSIEK, M. SINGER, H. WENK, S. ELEFTHERIADIS, H. KUPPE

AND P . SCHMUCKER

We have compared the Doppler against the thermodilution technique for measurement of cardiac output in six patients during aortic surgery. The correlation coefficient between the two methods was between 0.76 and 0.84 during the different periods of the operation. Using the integral nomogram instead of direct calibration, the Doppler system underestimated cardiac output in atherosclerotic patients. However, the Doppler method did register accurately significant changes in cardiac output. (Br. J. Anaesth. 1995; 74: 655-660) Key words Measurement techniques, cardiac output. techniques, Doppler echocardiography.

Measurement

Patients requiring surgery on the aorta are frequently affected by degenerative vascular disease and are elderly [1]. As they are recognized to be at high risk of circulatory decompensation [2, 3], a high level of haemodynamic monitoring is recommended [2], including frequent evaluation of cardiac output [4, 5]. In addition to thermodilution (TD) there are other methods which measure cardiac output, such as transtracheal [6] or suprasternal [7] Doppler of aortic blood flow, or thoracic bioimpedance [8, 9]. Oesophageal Doppler monitoring (DOP) of cardiac output is a safe, non-invasive method [10] which allows continuous measurement and provides additional information on circulatory state [11]. Easy handling and successful use are described, even in elderly patients [12]. It has also been shown to provide a good indication of alteration in cardiac output caused by changes in the level of PEEP [13] or intravascular volume shifts during transurethral prostatectomy [14]. However, it is not clear if the particular oesophageal Doppler device used in the

above studies [10-14] is appropriate in situations considered difficult by authors using earlier devices of different design [15, 16]. One such situation is during the clamping phase of aortic surgery. Patients and methods After approval by the Ethics Committee of the Medical University of Lubeck, we studied six patients (table 1) who gave written, informed consent. Patients were excluded if they had known cardiac valve defects or any history of oesophageal disorder. Despite the small number of patients, a wide range of age, height, weight and body surface area was covered. All patients (two female, four male), ASA II and III, were undergoing reconstructive surgery because of aneurysmal disease of the abdominal aorta and common iliac arteries. Past medical history included coronary artery disease, previous myocardial infarction and peripheral vascular disease. No patient with diabetes mellitus was included in the study. All were scheduled for replacement of the diseased vessel segments by a Yshaped prosthesis via an open transabdominal approach using a median laparotomy. The aorta was exposed up to the left renal vein and down to the iliac vessels as far as necessary. In all cases it was possible to clamp distal to the renal artery. Heparin 5000 u. was given i.v. for anticoagulation. A proximal endto-end anastomosis was performed using either a Dacron or PTFE-graft, usually of diameter 16 mm. In case of arterial occlusive disease a proximal end to side anastomosis was preferred. The method of anaesthesia and perioperative monitoring is shown in table 2. Anaesthesia was given by an independent anaesthetist who was not involved in the study. Routine premedication with flunitrazepam 1.0 mg was given 1 h before surgery. On arrival in the operating theatre, peripheral i.v. and radial arterial cannulae were inserted. Fluid

Table 1 Patient characteristics. BSA = Body surface area

Mean SD

Median Max. Min.

Age (years)

Height (cm)

Weight (kg)

BSA (m2)

66.2 — 65.5 75 55

171.3 8.6 173 180 156

71.3 15.1 75.5 88 50

1.8 0.2 1.915 2.03 1.55

KARL-FRIEDRICH KLOTZ, MD, STEPHAN KUNGSIEIC, SAVAS ELEFTHERIADIS, MD, HERMANN KUPPE, MD, PETER SCHMUCKER, MD

(Department

of

Anaesthesiology);

HEINER

WENK,

MD

(Department of Surgery); Medical University of Lubeck, Ratzeburger Allee 160, 23538 Lubeck, Germany. MERVYN SINGER, MD, Bloomsbury Institute of Intensive Care Medicine, University College London, Medical School, Rayne Institute Building, University Street, London WC1E 6JJ. Accepted for publication: December 20, 1994.

Downloaded from http://bja.oxfordjournals.org/ at Lahore University of Management Sciences on May 21, 2015

Summary

British Journal of Anaesthesia

656 Table 2 Anaesthesia and monitoring used for management of reconstructive aortic surgery Induction of anaesthesia

Fentanyl 0.2 mg Midazolam 3.0 mg Etomidate 20 mg Vecuronium 8.0 mg Maintenance of anaesthesia

Inspired gas mixture N 2 O:O 2 = 70:30 Inhalation agent 0.5-1.0% isoflurane Supplementary doses of fentanyl 0.1 mg as necessary Supplementary doses of vecuronium 2 mg every 30 min Monitoring

management was guided by cardiac filling pressures and consisted of infusions of electrolyte solutions and plasma expanders (hydroxyethyl starch and serum albumin solutions). For monitoring purposes a central venous and a 7French gauge flow-directed pulmonary artery catheters (93A-131-7F, Baxter Edwards Laboratories, Irvine, CA, USA) were inserted via the right internal jugular vein. Ventilator settings were adjusted according to data obtained from repeated arterial blood-gas samples. METHODS OF HAEMODYNAMIC MONITORING

Continuous monitoring of the various blood pressures was performed with the mid-heart level taken as the reference point. The rate—pressure product was calculated by multiplication of systolic arterial pressure by heart rate. All data were obtained using the Sirecust 1281 monitor (Siemens, Erlangen, Germany). THERMODILUTION MEASUREMENTS

Cardiac output was measured intermittently by injection of 10 ml of ice-cold saline (0-5 °C) using the Sirecust 1281 haemodynamic monitoring computer. All measurements were distributed randomly over the whole respiratory cycle. Cardiac output was taken as the mean of three measurements showing close agreement. However, for comparative purposes, all single injectate values were taken into account. OESOPHAGEAL DOPPLER MONITOR FOR CONTINUOUS CARDIAC OUTPUT MEASUREMENT

The DOP measurements were performed by an anaesthetist not involved in the anaesthetic management of the patient. After trachea! intubation a Doppler probe of 7 mm in diameter was inserted via the mouth into the oesophagus. The probe was connected to the Doptek ODM 1 Computer (Deltex Ltd, Chichester, West Sussex, UK). This device uses 4-MHz continuous wave Doppler ultrasound

EXPERIMENTAL DESIGN

All data from both Sirecust and ODM 1 monitors were recorded onto videotape for offline evaluation. For each determination of cardiac output by TD, a corresponding value for DOP was made by averaging the individual beats over the approximate 10 s needed for TD measurements. The use of single injectate TD measurements for comparison was necessary because of the rapid changes in cardiac output during aortic surgery. The period of surgical clamping and unclamping was analysed closely. From 1 min before to 5 min after these procedures all haemodynamic data were noted on a second-bysecond basis. STATISTICAL ANALYSIS

The paired values obtained by TD and DOP were divided into three groups, namely the period before clamping of the aorta (pre-clamp), the period with the aorta completely occluded by the clamp (clamp), and then following removal of the aortic clamp (post-clamp). Haemodynamic data from each patient were averaged to obtain a single mean value for each period. Data from all patients were used to compute the mean (SD). Cardiac output data were compared by regression analysis and the correlation coefficient (r) computed. Bias and precision of DOP compared with TD were computed by the Bland-Altman

Downloaded from http://bja.oxfordjournals.org/ at Lahore University of Management Sciences on May 21, 2015

Elcctrocardiography Pulse oxymetry Arterial pressure Central venous pressure Pulmonary artery and wedge pressure Analysing expired gas Intermittent evaluation of arterial blood-gas tensions

and presents both an acoustic signal and a visual spectral analysis of the Doppler frequency shifts on a real-time colour display. This dual output allows, after a short learning period, easy manipulation of the Doppler probe into an appropriate position in the oesophagus to obtain clear Doppler signals of blood flow in the descending thoracic aorta. The monitor performs online estimations of total cardiac output, either by primary calibration of the monitor using another evaluation method, for example, TD, or by using a fixed stored nomogram from manually imported biometric data, including age, sex, body weight and height. Measurement of absolute left ventricular output is correct only when using an appropriate nomogram, and assuming a constant aortic cross-sectional area during systole, and a constant distribution of cardiac output between the vessels originating from the aortic arch and the remainder which constitutes flow in the descending aorta. These assumptions may be incorrect in patients with aortic degenerative disease as the diameter of the descending aorta may be aneurysmatically increased or degeneratively decreased. The relative distribution of blood flow between the supra-aortic vessels and the descending aorta may be changed, especially during aortic crossclamping. These potential problems thus formed the basis of this study. The technique and handling of the probe is described in detail elsewhere [12]. We used the nomogram method for measurement of total cardiac output to allow determination of the validity of this particular device. Cardiac output was measured on a beat-to-beat basis.

Cardiac output measurement with ODM 1

technique. Bias indicates the mean deviation of DOP from TD whereas precision shows the distribution of the DOP values around the mean DOP value. All consecutive TD measurements, where an increase or decrease of more than 2 litre min"1 occurred, were included in a further evaluation which included correlation and regression analysis. This was performed to verify consistency in the direction of change in cardiac output measured by the two techniques.

Discussion Castor and colleagues [17] simultaneously compared three methods of evaluating cardiac output, namely TD, thoracic electrical bioimpedance and aortic Doppler ultrasound by the suprasternal approach. Whereas bioimpedance and TD showed a good correlation, the Doppler device differed significantly. This was caused mostly by problems with the handling of the suprasternal Doppler probe. Wong and colleagues [18] measured cardiac output with an oesophageal Doppler flow probe in pigs. Their results were unsatisfactory although this may be related to both the device used and the increased aortic wall elasticity in the young animals monitored. Siegel, Fitzgerald and Engstrom [6] compared transtracheal Doppler with TD and found a correlation coefficient of 0.63. This was inferior to our results with the Doptek oesophageal Doppler system which we feel solves many of the methodological problems associated with the oesophageal approach. The good auditory and visible display of the ultrasound signal helps considerably in determining correct probe placement, a feature lacking in all other machines.

It is noteworthy that we did not calibrate the DOP measurements at any time during the study and that we compared the DOP values with single injectate measurements of cardiac output by the TD technique. The inherent bias and variation of the TD method, which may be as high as 50% [19], is included in all the correlation data presented. This variation may adversely affect the comparison with DOP. During the pre-clamp period, cardiac outputs measured by DOP correlated well with that obtained by TD. Our correlation coefficient of 0.84 is comparable with the values indicated by other authors [15] for more complicated methods. The pre-clamp bias of 0.96 litre min"1 may be related to the integral stored nomogram for estimation of aortic diameter which might not be applicable in patients with severe atherosclerotic disease of the aorta. The DOP measurements were, on average, approximately 20 % lower; this may result from underestimation of the greater aortic diameter in this particular group of patients because of pathological widening. However, as we studied only six such patients, more data are required before recommending any change in the integral nomogram. During aortic cross-clamping, the correlation coefficient of 0.79 was similar to the pre-clamp period and sufficiently accurate, although the bias did increase to 1.51 litre min"1. Theoretically, this may be caused by aortic dilatation induced by the clamp and also redistribution of blood flow. The diameter of the aorta may possibly be influenced by mean arterial pressure although this is more likely to affect the elastic aorta of the young compared with the elderly patient whose aorta is far more rigid [20, 21]. We observed a slight decrease in mean arterial pressure from 85.4 to 79.0 mm Hg, thus pressurerelated changes in aortic diameter would be small and tend to allow dilatation rather than constriction if they occurred at all. The ODM 1 measurements are based on a fixed blood flow distribution of 30 % to the supra-aortic vascular structures and 70 % to the descending aorta. This distribution is assumed by the device to be constant. However, during cross-clamping there may be redistribution to the supra-aortic regions. This leads to underestimation of the complete cardiac output by the DOP method, as seen in our study. After declamping of the aorta, the correlation between DOP and TD in our study decreased to 0.74 whereas Perrino, Fleming and LaMantia [15] found a correlation coefficient of 0.85. However, the precision of their measurements was decreased compared with ours. Gelman and colleagues [22] investigated blood flow distribution and metabolic changes produced by declamping of the aorta. There seemed to be a significant release of vasoactive mediators from the intermittently malperfused tissues to the systemic circulation with a complex vascular and therapeutic response. This may possibly lead to redistribution of blood flow to the upper and lower body, although this hypothesis required confirmation. The advocates of the DOP method claim that this technique does not indicate "true" cardiac output,

Downloaded from http://bja.oxfordjournals.org/ at Lahore University of Management Sciences on May 21, 2015

Results Haemodynamic data are shown in table 3. There were no statistically significant differences in any value between the three periods, although rate-pressure product, DOP, stroke volume, mean arterial pressure and mean pulmonary artery pressure decreased during cross-clamping while systemic vascular resistance increased slightly. All variables returned to pre-clamp values after unclamping of the aorta. It was possible to measure cardiac output in all patients at all attempts and it proved very easy to obtain adequate Doppler signals. In figure 1 the data of 55 measurements of cardiac output obtained during the pre-clamping period by TD are plotted against the simultaneous DOP values. The correlation coefficient was 0.84 and the regression line followed the equation DOP = 0.91 x(TD)-0.38. The complete results of the statistical evaluation of all three periods are shown in table 4. For comparison of 15 large (more than 2 litre min"1) differences between consecutive measurements of TD with corresponding DOP values, measurements from all surgical periods were included in the evaluation. The values correlated with a correlation coefficient of r = 0.89. The regression equation was delta DOP = 0.69 A TD+0.53.

657

658

British Journal of Anaesthesia Table 3 Haemodynamic data during the three surgical periods (mean (SD)). HR = Heart rate, MAP = mean arterial pressure, MPAP = mean pulmonary arterial pressure, SV = stroke volume, DOP = Doppler (ODM 1) cardiac output, SVR = systemic vascular resistance, RPP = rate-pressure product Clamp

Pre-clamp 1

HR (beat min" ) MAP (mm Hg) MPAP (mm Hg) SV(ml) DOP (litre mirr1) SVR (dyn cm' 5 s"1) RPP

85.2 (23.6) 82.7 (22.6) 85.4(15.4) 79.0(11.5) 21.6(4.9) 17.9 (5.3) 55.5(20.1) 44.9(19.4) 4.73 (2.22) 3.64(1.82) 1394(412) 1739(571) 10693(3811) 9384 (2583)

150

10

_ 125" a> x 100 E E

T

c 8 E | 6

82.9(18.5) 88.7(13.8) 24.0 (5.8) 60.6(15.2) 5.01 (1.86) 1290 (269) 10582(2623)

75

a. <

0-

O 4 Q

-40

4 6 8 TD (litre mirr1)

10

12

Figure 1 Comparison of oesophagcal Doppler (DOP) measurements with simultaneous thermodilution (TD) evaluations during the pre-clamping period (n = 55). (DOP = 0.91xTD-0.38, r = 0.84.)

but rather a reasonable estimate. However, it does provide a good qualitative guide to changes [18]. We found that combining the results of the correlation tests, a very high r value of 0.89 was obtained for cardiac output changes. All marked changes in measurements of cardiac output by TD were accompanied by appropriate changes in DOP measurements. This was not seen in the study performed by Wong and colleagues [18] who found that 5—7% of changes occurred in the opposite direction. However, in our study we selected only paired measurements where there were large changes in cardiac output measured by TD. This excludes "false small differences" caused by the variability of the TD technique; a 14% change in cardiac output measured by TD is needed to be confident that this is a true change [23]. MECHANISMS OF COMPENSATION OF THE UNCLAMPING MANOEUVRE

The advantage of the DOP method for monitoring patients during reconstructive surgery on the aorta is that we may be able to rapidly identify deterioration

40 Time (s)

80

Figure 2 Changes in mean arterial pressure (MAP) during declamping (D) in patient A (male 63 yr of age, 178 cm height and 76 kg weight) and in patient B (male, 74 yr of age, 166 cm height and 73 kg weight).

and initiate therapy. Bush and colleagues [24] described a method of preventing hypotension after declamping of the aorta by optimal volume loading. This loading may not be guided directly by the filling pressure before the declamping procedure because there is often no change in left ventricular filling pressure, even with marked changes in left ventricular end-diastolic volume loading [22, 25]. Gregoretti and colleagues [26] showed that healthy animals responded to the release of aortic blood flow with a marked change in cardiac output. This is in contrast with our findings in patients and may be caused by the differences in elasticity and compliance of the arterial system. To illustrate the difference in the cardiovascular behaviour and the possibilities of rapid diagnosis, the original haemodynamic recordings are shown from two of our six patients. Figure 2 shows the change in mean arterial pressure during declamping and this was similar in both patients. However, as shown by the DOP measurements given in figure 3, there was marked difference in the cardiac output response; the cardiac output of patient A increased immediately after release of the cross-clamp, whereas in patient B cardiac output, after a small and shortlived elevation, stayed at the same level. The reason for this difference could be because of the different

Table 4 Results of the stanstical evaluation of the comparative measurements during the three different surgical periods Period

n

Correlation coefficient

Pre-clamp Clamp Post-clamp

55 75 65

0.84 0.79 0.76

120

Regression equation

Bias (litre min"1) (mean DOP —mean T D )

Precision (litre min"1) (95% confidential intervals)

DOP = 0.91 x (TD)-0.38 DOP = 0.74 x (TD)-0.14 DOP = 0.43 x(TD)+1.75

-0.96 - 1 51 -1.47

1.33 to -3.24 -0.02 to -2.99 0.86 to -3.79

Downloaded from http://bja.oxfordjournals.org/ at Lahore University of Management Sciences on May 21, 2015

12

Post-clamp

Cardiac output measurement with ODM 1

659

D

Figure 3 Changes in cardiac output (CO), measured by ODM 1, in patients A and B (for details see legend to fig. 2).

B

'" 3000 If) 1

c 2000 ^

A

1000

CO

n -40

D 1

1

j 1

1

1

1

1

40 Time (s)

80

120

Figure 4 Changes in systemic vascular resistance (SVR) evaluated from the traces of figures 2 and 3 for patients A and B (for details see legend to fig. 2).

behaviour in systemic vascular resistance shown in figure 4. Increases in stroke volume are effective mainly via the Frank-Starling mechanism [27] but when this mechanism is fully effective changes in systemic vascular resistance may be the only mechanism available for maintaining organ perfusion pressure in an appropriate range [28, 29]. Acknowledgements We thank Deltex Ltd, Chichester, for the loan of the ODM 1 monitor.

References 1. Silverstein PR, Caldera DL, Cullen DJ, Davison JK, Darling RC, Emerson CW. Avoiding the hemodynamic consequences of aortic cross-clamping and unclamping. Anesthesiology 1979; 50:462^66. 2. Attia RR, Murphy JD, Snider M, Lappas DG, Darling RC, Lowenstein E. Myocardial ischemia due to infrarenal aortic cross-clamping during aortic surgery in patients with severe coronary artery disease. Circulation 1976; S3: 961-965. 3. Gelman S, Rabbani S, Bradley EL. Inferior and superior vena caval blood flows during cross-clamping of the thoracic aorta in pigs. Journal of Thoracic and Cardiovascular Surgery 1988; 96: 387-397. 4. Prys-Roberts C. Cardiovascular monitoring in patients with vascular disease. British Journal of Anaesthesia 1981; S3: 767-776. 5. Gelman S, Patel K, Bishop SP, Fowler KL, Smith LR. Renal and splanchnic circulation during infrarenal aortic crossclamping. Archives of Surgery 1984; 119: 1394-1399. 6. Siegel LC, Fitzgerald DC, Engstrom RH. Simultaneous intraoperative measurement of cardiac output by thermodilution and transtracheal Doppler. Anesthesiology 1991; 74: 664-669.

Downloaded from http://bja.oxfordjournals.org/ at Lahore University of Management Sciences on May 21, 2015

tuuu

7. Donovan KD, Dobb GJ, Newman MA, Hockings BES, Ireland M. Comparison of pulsed Doppler and thermodilution methods for measuring cardiac output in critically ill patients. Critical Care Medicine 1987; 15: 853-857. 8. Bernstein DP. Continuous noninvasive real-time monitoring of stroke volume and cardiac output by thoracic electrical bioimpedance. Critical Care Medicine 1986; 14: 898-901. 9. Young JD, McQuillan P. Comparison of thoracic electrical bioimpedance and thermodilution for the measurement of cardiac index in patients with severe sepsis. British Journal of Anaesthesia 1993; 70: 58-62. 10. Daniel WG, Erbel R, Kasper W, Visser CA, Engberding R, Sutherland GR, Grube E, Hanrath P, Maisch B, Dennig K, Schartl M, Kremer P, Angermann C, Iliceto S, Curtius JM, Mugge A. Safety of transesophageal echocardiography. A multicenter survey of 10,419 examinations. Circulation 1991; 83: 817-821. 11. Singer M, Bennett ED. Noninvasive optimization of left ventricular filling using esophageal Doppler. Critical Care Medicine 1991; 19: 1132-1137. 12. Singer M, Clarke J, Bennett ED. Continuous hemodynamic monitoring by esophageal Doppler. Critical Care Medicine 1989; 17: 447-152. 13. Patel M, Singer M. The optimal time for measuring the cardiorespiratory effects of positive end-expiratory pressure. Chest 1993; 104: 139-142. 14. Evans JWH, Singer M, Chappie CR, Macartney N, Wai JM, Milroy EJG. Haemodynamic evidence for cardiac stress during transurethral prostatectomy. British Medical Journal 1992; 304: 666-671. 15. Perrino AC, Fleming J, LaManria KR. Transesophageal Doppler cardiac output monitoring: performance during aortic reconstructive surgery. Anesthesia and Analgesia 1991; 73: 705-710. 16. Bernstein DP. Noninvasive cardiac output, Doppler flowmetry, and gold-plated assumptions. Critical Care Medicine 1987, 15: 886-888. 17. Castor G, Klocke RK, Stoll M, Helms J, Niedermark I. Simultaneous measurement of cardiac output by thermodilution, thoracic electrical bioimpedance and Doppler ultrasound. British Journal of Anaesthesia 1994; 72: 133-138. 18. Wong DH, Watson T, Gordon I, Wesley R, Tremper KK, Zaccari J, Stemmer P. Comparison of changes in transit time ultrasound, esophageal Doppler, and thermodilution cardiac output after changes in preload, afterload, and contractility in pigs. Anesthesia and Analgesia 1991; 72: 584-588. 19. Stevens JH, Raffin TA, Mihm FG, Rosenthal MH, Stetz CW. Thermodilution cardiac output measurement: effects of the respiratory cycle on its reproducibility. Journal of the American Medical Association 1985; 253: 2240-2242. 20. List W, Gravenstein N, Banner T, Caton D, Davis RF. Interaction in sheep between mean arterial pressure and cross-sectional area of the descending aorta: implications for esophageal Doppler monitoring. Anesthesiology 1987; 67: A178. 21. Hillel Z, Zane E, Thys D. Systemic arterial blood pressure influences the cross-sectional area of the descending thoracic aorta. Anesthesiology 1988; 69: All. 22. Gelman S, Khazaeli MB, Orr R, Henderson T. Blood volume redistribution during cross-clamping of the descending aorta. Anesthesia and Analgesia 1994; 78: 219-224. 23. Stetz CW, Miller RG, Kelly GE, Raffin TA. Reliability of the thermodilution method in the determination of cardiac output in clinical practice. American Review of Respiratory Diseases 1982; 126: 1001-1004. 24. Bush HL, LoGerfo FW, Weisel RD, Mannick JA, Hechtman HB. Assessment of myocardial performance and optimal volume loading during elective abdominal aortic aneurysm resection. Archives of Surgery 1977; 112: 1301-1306. 25. Roizen MF, Beaupre PN, Alpert RA, Kremer P, Cahalan MK, ShiUer N, Sohn YJ, Cronnelly R, Lurz FW, Ehrenfeld WK, Stoney RJ. Monitoring with two-dimensional transesophageal echocardiography—Comparison of myocardial function in patients undergoing supraceliac, suprarenal—infraceliac, or infrarenal aortic occlusion. Journal of Vascular Surgery 1984; 1: 300-305. 26. Gregoretti S, Gelman S, Henderson T, Bradley EL. Hemodynamics and oxygen uptake below and above aortic

660 occlusion during crossclamping of the thoracic aorta and sodium nitroprusside Infusion. Journal of Thoracic and Cardiovascular Surgery 1990; 100: 830-836. 27. Stokland O, Miller MM, Ilebekk A, Kiil F. Mechanism of hemodynamic responses to occlusion of the descending thoracic aorta American Journal of Physiology 1980; 238: H423-H429. 28. Gelman S, Reves JG, Fowler K, Samuelson PN, Lell WA,

British Journal of Anaesthesia Smith LR. Regional blood flow during cross-clamping of the thoracic aorta and infusion of sodium nitroprusside. Journal of Thoracic and Cardiovascular Surgery 1983; 85: 287-291. 29. Gold MS, DeCrosta D, Rizzuto C, Ben-Harari RR, Ramanathan S. The effect of lumbar epidural and general anesthesia on plasma catecholamines and hemodynamics during abdominal aortic aneurysm repair. Anesthesia and Analgesia 1994; 78: 225-230.

Downloaded from http://bja.oxfordjournals.org/ at Lahore University of Management Sciences on May 21, 2015