Hepatic vascular exclusion (HVE) involves portal triad clamping (PTC) (Fig. 1) and occlusion of the inferior vena cava below and above the liver. It completely isolates the liver and the retrohepatic vena cava from the rest of the circulation. Its aim is to reduce the risk of massive hemorrhage and air embolism caused by a tear of the vena cava or a hepatic vein during the removal of large or posterior liver tumors. The physiological adaptation to PTC has been extensively studied (1–3). In contrast, although HVE (4) has been performed for approximately 20 yr, either during liver transplantation or major liver resections, still little is known (5,6) about the complex mechanisms that contribute to explaining the extraordinary physiological adaptation that results in its relatively good hemodynamic tolerance. Liver resection is regarded as major abdominal surgery that requires a deep level of anesthesia with a subsequent attenuation of the baroreceptor reflex (7). Surprisingly, mean arterial blood pressure (MAP) is maintained during HVE (6). This suggests that adaptation of arterial blood pressure in response to the decrease in venous return may involve the intervention of one or more of the three hormonal systems: the sympathetic system, the renin-angiotensin system (RAS), and the arginine vasopressin (AVP) system (8,9). Our study concerns the hormonal contribution to hemodynamic stability during HVE.
Over the years, our team has developed extensive experience with HVE (10,11), and it is now widely and extensively used in our unit to perform liver resections. We therefore decided to prospectively investigate hormonal and hemodynamic changes in response to HVE in a carefully selected population of patients undergoing major liver resection for secondary liver tumors developed in noncirrhotic livers. In this study, we report on the results of the first comprehensive study of the hormonal and hemodynamic responses to HVE. We further propose an explanation for the circulatory adaptation to HVE, with particular emphasis on the hierarchy of hormonal mechanisms that are involved in maintaining adequate arterial blood pressure in anesthetized patients submitted to a sudden and major decrease in venous return provoked by HVE.
From January 1999 to June 2000, 23 consecutive patients undergoing liver resection for metastatic tumor were studied according to a protocol approved by the local committee on human research. Written, informed consent was obtained before surgery. Patients with cirrhosis, chronic hepatitis, or other causes of portal hypertension were excluded from the study, as were patients with impaired renal function, defined as serum creatinine concentration >120 μmol/L, and patients with a history of alteration of cardiac function (systolic left ventricular [LV] dysfunction), absence of sinus rhythm, a clinical history of congestive heart failure (New York Heart Association Class II, III, or IV), and/or LV fractional diameter shortening <0.30 on preoperative echocardiography. Patients with hormonally active tumors were excluded. Diastolic LV dysfunction was defined as a ratio of Doppler transmitral waves E/A <1.
Anesthetic management and intraoperative care were standardized throughout the study. The same team (two surgeons and two anesthesiologists) was in charge of the patients during the study. Patients were premedicated with hydroxyzine 2 mg/kg orally, 2 h before surgery. General anesthesia was induced with IV thiopental 4–6 mg/kg and sufentanil 0.5 μg/kg. Muscular relaxation was obtained with atracurium 0.6 mg/kg IV at induction then maintained with a continuous infusion of 0.3–0.5 mg/kg IV. Anesthesia was maintained with sufentanil 0.5 μg · kg−1 · h−1, midazolam 0.03–0.08 mg/kg until unclamping, and 0.1%–0.7% end-tidal isoflurane with 50% nitrous oxide in oxygen. Ventilation was controlled throughout anesthesia by using a Julian ventilator (Dräger Medical, Antony, France) and circle absorber system. Tidal volume was set at 10 mL/kg and was frequency-adjusted to maintain expired end-tidal carbon dioxide at 4.5 ± 0.5 kPa. Throughout the period of study, warming therapy was applied to minimize intraoperative hypothermia. Heating blankets (Bair Hugger; Augustine Medical, Inc., Eden Prairie, MN) were placed on the legs, upper thoracic body, and head to maintain esophageal temperature >35.5°C. Electrocardiogram, heart rate, and arterial pressure (AP) (systolic, diastolic, and MAP) by radial artery cannulation were monitored continuously (Solar 8000; Marquette SA, France) in all patients.
Pulmonary artery pressure, pulmonary artery occlusion pressure (PAOP), continuous cardiac output (CO), and mixed venous oxygen saturation (Svo2) were monitored. Systemic vascular resistance index (SVRI) was calculated. A transesophageal echocardiography probe with a 5-MHz transducer (HP Image Point; Hewlett-Packard, Andover, MA) was inserted after induction and positioned to obtain LV cross-sectional images at the midpapillary muscle level. That probe allowed the end-systolic and end-diastolic areas to be measured and the fractional area change (FAC) to be calculated as previously described (12). For investigation purposes, hemodynamic variables were noted 2 min before clamping (T0); 5, 15, and 30 min after clamping (T1, T2, and T3, respectively); and 15 min after unclamping (T4). No patient received vasopressors during surgery. Homologous packed red blood cells were transfused to maintain hematocrit >29%.
The abdomen was opened via a bilateral subcostal incision with a midline extension. Dissection of the inferior vena cava below and above the liver was performed as already described elsewhere in detail (11). At the end of dissection, fluid volume was adjusted to obtain a PAOP of approximately 9–12 mm Hg, and response to a 5-min trial of HVE was noted. This test is routinely performed to obtain a preliminary estimation of subsequent tolerance to prolonged HVE (4). Poor tolerance to HVE may be anticipated if, 5 min after the test clamping, MAP decreases to <55 mm Hg, Svo2 decreases to <60%, or CO decreases by >60% of pre-clamp values.
In all patients, blood samples were collected 2 min before the HVE trial test; 5, 15, and 30 min after clamping; and 15 min after unclamping, synchronized with hemodynamic and transesophageal echocardiography points (T0, T1, T2, T3, and T4). Samples were drawn in chilled glass tubes containing EDTA (7.5 mg/5 mL), centrifuged at 2000 g for 15 min, and immediately deep-frozen and stored at −20°C until atrial natriuretic peptide (ANP), AVP, plasma renin activity (PRA), and catecholamine (epinephrine [E], norepinephrine [NE], and dopamine) concentration determinations were performed as described in detail elsewhere (13–15). Mean plasma concentrations in healthy subjects of either sex were previously found to be 12 ± 1 pg/mL, 1.4 ± 1.0 pg/mL, and 1.6 ± 0.5 ng/mL for ANP, AVP, and PRA, respectively.
All results are expressed as mean ± sd. Statistical analysis used repeated-measures analysis of variance and Dunnet’s least significant difference test. All values were two tailed, and P < 0.05 was considered significant.
Twenty-three consecutive patients were included in the study. One 78-yr-old ASA physical status III patient did not tolerate the HVE test. This patient was chronically receiving antihypertensive therapy. Pre- and intraoperative patient characteristics are reported in Table 1.
MAP was always maintained >55 mm Hg during HVE without any vasoconstrictor but with colloid infusion (Table 1). Systolic arterial blood pressure, diastolic arterial blood pressure, and MAP returned after unclamping to values significantly less than baseline values (Table 2). The blood concentration of hemoglobin significantly decreased from 11.2 ± 1.6 g/dL at T0 to 10.7 ± 1.5 g/dL at T4 (P < 00.1). Five patients were transfused during HVE. Cardiac index (CI) decreased after HVE, remained stable during this period, and then increased after unclamping to a value more than baseline.
LV end-diastolic area (LVEDA), LV end-systolic area, and FAC significantly decreased 5 min after clamping, remained low during HVE, and then returned to their initial values 15 min after unclamping (Fig. 2). Central venous pressure (CVP) and PAOP followed the same pattern (Table 2 and Fig. 3). Mean pulmonary artery pressure significantly decreased from 19 ± 4 mm Hg before HVE to 11 ± 5 mm Hg during HVE (P < 0.0001) and returned to a level (22 ± 5 mm Hg) not significantly different from the initial level after unclamping. SVRI increased significantly from T0 to T1 (×2.4 increment), remained increased during HVE, and then reached a value significantly less than the baseline value (T0) (Fig. 4). Svo2 significantly decreased from 82% ± 6% before HVE to 75% ± 10% (P < 0.0001), 74% ± 9%, and 75% ± 6% at 5, 15, and 30 min, respectively, after clamping and then returned to a level (86% ± 6%) not significantly different from the initial value before HVE.
Dopamine concentration and PRA did not change significantly during or after HVE (Table 3). In contrast, both the ANP and AVP concentration increased significantly from T0 to T1 in all patients. Blood concentrations of both E and NE also increased significantly during HVE. NE remained significantly increased at T3. Although the increase in NE and E concentration was not significant at T1, 18 patients had increase (<20%) in the concentration of these hormones. The blood concentration of both hormones rapidly returned to baseline levels after unclamping, and the ANP plasma concentration was even less than baseline after unclamping.
LVEDA and ISVR were significantly correlated with NE (r = 0.52, P < 0.0001 and r = 0.32, P < 0.01, respectively) and AVP (r = 0.37, P < 0.0001 and r = 0.44, P < 0.0001, respectively) (Fig. 5). The maximum change in AVP (i.e., the increment in AVP levels between T0 and T1) was significantly correlated (r = 0.45, P < 0.001) with the change in LVEDA (LVEDA decrement between T0 and T1).
In this study we describe hemodynamic and hormonal changes that occur during HVE for major liver resections performed in noncirrhotic patients. HVE is responsible for a major decrease in venous return reflected by a decrease in LVEDA, PAOP, and CO. This decrease is stable during HVE, and MAP is maintained by a marked increase in SVR associated with an increase in vasopressin, NE, and E secretion. At unclamping, hemodynamic and hormonal variables returned to their initial values. Despite the marked decrease in venous return after caval clamping that is responsible for a rapid and stable decrease in LVEDA, PAOP, and CI, HVE in patients undergoing hepatic resection with nonpathologic liver is mostly well tolerated, as previously reported by Delva et al. (5,6). MAP was usually maintained at >55 mm Hg during the HVE period, with Svo2 maintained at >70%. This study gives precise insight into the mechanisms that are involved to explain the stability of AP. Maintained systemic AP is afforded for an increase in vascular resistance. A sudden decrease in preload, as demonstrated by a reduction in LVEDA and PAOP and an increase in arterial vascular tone reflected by SVRI, leads to a significant decrease in LV FAC. This reduction in LV FAC was stable throughout the HVE period, without development of mixed venous desaturation. Interestingly, after unclamping, we observed a peripheral vasodilation reflected by a significant increase in CI and a decrease in SVRI from T0 to T4. That observation could be due to fluid loading during HVE, with consecutive hemodilution, or to a vasodilation of the splanchnic circulation after hepatic and splanchnic ischemia secondary to HVE (ischemia reperfusion). Our findings also showed that AVP, NE, and E are important hormonal factors responsible for the maintenance of MAP during HVE, whereas PRA appears to play no role in this adaptation.
Our results clearly showed an increase in NE and E, but that could be due to hepatic exclusion from the bloodstream during HVE, because there is some evidence for substantial mesenteric and hepatic NE removal (13,16,17). By contrast, there is no physiologic evidence for substantial hepatic AVP removal, and Lentschener et al. (3) even found a significant decrease in AVP concentration after PTC when the hepatic pedicle was infiltrated 15 minutes before PTC. The dramatic increase in AVP blood levels observed in our study cannot be explained by a reduction in the hepatic catabolic rate. The link of causality between AVP blood concentrations and maintenance of a satisfactory MAP has been noted by Lentschener et al. (3) for PTC. The trigger of this hormonal secretion might be found in relation to the splanchnic bed, with particular emphasis on modifications of the portal pressure, as suggested in previous studies (14,15,18). Such a mechanism could theoretically also be involved in HVE. The abrupt increase in AVP concentration that was noted in our study, to a point even larger than that measured in PTC patients by Lentschener et al., is probably explained by a stronger splanchnic trigger in HVE patients. Because retrohepatic IVC is completely mobilized and clamped just below the right cardiac cavities, the potential for venous return to the heart through collateral veins (e.g., diaphragmatic veins) is less than in PTC patients, and splanchnic pressure is probably higher. In contrast to PTC patients, however, in whom the increase in ISVR is more than what would be anticipated as a consequence of reduction in venous return (2), we observed in our study that the increase in ISVR was a compensatory phenomenon perfectly suited to the decrease in CI due to HVE. Because of ethical considerations, we did not measure portal pressure during surgery, and we are therefore unable to confirm our hypothesis that portal pressure is higher in HVE than in PTC patients. As another explanation to the observed hemodynamic and hormonal differences between HVE and PTC patients, one may hypothesize that HVE, in contrast to PTC, is characterized by a major and acute reduction in CVP that may trigger the secretion of AVP by the posterior pituitary gland in response to the Henry-Gauer reflex (19–21). Indeed, we noted that the sympathetic system and AVP were sufficient to maintain systemic AP during HVE. Previous studies demonstrated that blockade of one or two of the three hormonal systems (sympathetic, angiotensin, and vasopressin) could be compensated for by the activation of the others (8,9). During HVE, RAS is not actually blocked, but is, rather, not triggered. With respect to HVE, the demonstration of a favored AVP hormonal response appears particularly well adapted, because this hormone greatly reduces mesenteric blood flow (8). The demonstration of a significant increase in ANP blood levels in HVE patients may be a confirmation of the central role of the atrium but is difficult to explain. In other situations of hypovolemia, some authors also found increases in ANP blood concentration. The sudden decrease in CVP triggers hyperactivity of the atrium to better fill the ventricle, and this stretch possibly triggers ANP secretion (22,23).
Our results suggest that the extraordinary adaptation to HVE is not sustained by the RAS, but rather by vasopressin and NE intervention. These findings indicate that pharmacological studies in patients who do not tolerate HVE during liver transplantation need to test agonists of the AVP system in association with catecholamines.
The authors wish to thank Professor B. Riou and Doctor E. Delva for their critical reviews of the manuscript and substantive comments.
1. Pringle JH. Notes on the arrest of hepatic hemorrhage due to trauma. Ann Surg 1908; 48: 541–9.
2. Delva E, Camus Y, Paugam C, et al. Hemodynamic effects of portal triad clamping in humans. Anesth Analg 1987; 66: 684–8.
3. Lentschener C, Franco D, Bouaziz H, et al. Haemodynamic changes associated with portal triad clamping are suppressed by prior hepatic pedicle infiltration with lidocaine in humans. Br J Anaesth 1999; 82: 691–7.
4. Heaney JP, Stanton WK, Halbert DS, et al. An improved technic for vascular isolation of the liver. Ann Surg 1966; 163: 237–41.
5. Delva E, Camus Y, Nordlinger B, et al. Vascular occlusion for liver resections. Ann Surg 1989; 209: 211–8.
6. Delva E, Barberousse JP, Nordlinger B, et al. Hemodynamic and biochemical monitoring during major liver resection with use of hepatic vascular exclusion. Surgery 1984; 95: 309–18.
7. Marty J, Reves JG. Cardiovascular control mechanisms during anesthesia. Anesth Analg 1989; 69: 273–5.
8. Brand P, Metting P, Britton S. Support of arterial pressure by major pressure systems in conscious dogs. Am J Physiol 1988; 255: H843–91.
9. Carp H, Vadhera R, Jayaram A, Garvey D. Endogenous vasopressin and renin-angiotensin systems support blood pressure after epidural block in humans. Anesthesiology 1994; 80: 1000–7.
10. Hannoun L, Vaillant JC, Borie D, Delva E. Techniques de l’exclusion vasculaire du foie et des hépatectomies extrèmes (chirurgie “ex situ ex vivo” et “in situ ex vivo” du foie). Encyclopédie Médico-Chirurgicale: Techniques chirur- gicales—Appareils digestifs. Paris: Editions Scientifiques et Medicales, Elsevier, 1994:40–766, 16 P.
11. Hannoun L, Borie DC, Vaillant JC, et al. Liver resection with normothermic ischemia exceeding one hour. Br J Surg 1993; 80: 1161–5.
12. Eyraud D, Mouren S, Teugels K, et al. Treating anesthesia-induced hypotension by angiotensin II in patients chronically treated with angiotensin-converting enzyme inhibitors. Anesth Analg 1998; 86: 259–63.
13. Aneman A, Eischofer G, Olbe L, et al. Sympathetic discharge to mesenteric organ norepinephrine spillover. J Clin Invest 1996; 97: 1640–6.
14. Andrews CJH, Andrews WHH, Orbach J. A sympathetic reflex elicited by distension of mesenteric venous bed. J Physiol (Lond) 1972;226:119–31.
15. Greenway CV, Stark RD. Hepatic vascular bed. Physiol Rev 1971; 51: 23–65.
16. Irita K, Okamoto H, Sakaguchi Y, Takahashi S. A possible increase in plasma norepinephrine by removal of the liver. Acta Anaesthesiol Scand 1998; 42: 1164–7.
17. Eisenhofer G, Rundquist B, Äneman A, et al. Regional release and removal of catecholamines and extraneuronal metabolism to metanephrine. J Clin Endocrinol Metab 1995; 80: 3009–17.
18. Sawchenko PE, Friedman MI. Sensory functions of the liver: a review. Am J Physiol 1979; 236: R5–20.
19. Zoller RP, Mark AL, Abboud FM, Heistad DD. The role of low pressure baroreceptors in reflex vasoconstrictor responses in man. J Clin Invest 1972; 51: 2967–72.
20. Van de Buuse M. Role of the mesolimbic dopamine system in cardiovascular homeostasis: stimulation of the ventral tegmental area modulates the effect of vasopressin on blood pressure in conscious rat. Clin Exp Pharmacol Physiol 1998; 25: 661–8.
21. Berkenstadt H, Rosenthal T, Peleg E, et al. Elevated plasma atrial natriuretic peptide levels after occlusion of the thoracic aorta. Chest 1999; 15: 130–4.
22. Putensen C, Mutz N, Pomaroli A, et al. Atrial natriuretic factor release during hypovolemia and after volume replacement. Crit Care Med 1992; 20: 984–9.
23. Frajewicki V, Kahana L, Yechieli H, et al. Effects of severe hemorrhage on plasma ANP and glomerular ANP receptors. Am J Physiol 1997; 273 (5 Pt 2):R1623–30.