Worsening of cerebral hyperemia by the administration of terlipressin in acute liver failure with severe encephalopathy

Authors


Abstract

There is increasing evidence that terlipressin is useful in patients with cirrhosis and hepatorenal syndrome, but there are no data of its use in patients with acute liver failure (ALF) in whom hepatorenal syndrome is common. Although terlipressin produces systemic vasoconstriction, it produces cerebral vasodilatation and may increase cerebral blood flow (CBF). Increased CBF contributes to intracranial hypertension in patients with ALF. The aim of this study was to evaluate the safety of terlipressin in patients with ALF with respect to cerebral hemodynamics. Six successive patients with ALF were ventilated electively for grade IV hepatic encephalopathy. Patients were monitored invasively and CBF was measured (Kety-Schmidt technique). Measurements were made before and at 1, 3, and 5 hours after intravenous (single bolus) administration of terlipressin (0.005 mg/kg), median, 0.25 mg (range, 0.2–0.3 mg). There was no significant change in heart rate, mean arterial pressure, or cardiac output. CBF and jugular venous oxygen saturation both increased significantly at 1 hour (P = 0.016). Intracranial pressure increased significantly at 1 hour (P = 0.031), returning back to baseline values at 2 hours. In conclusion, administration of terlipressin, at a dose that did not alter systemic hemodynamics, resulted in worsening of cerebral hyperemia and intracranial hypertension in patients with ALF and severe hepatic encephalopathy. These data suggest the need to exercise extreme caution in the use of terlipressin in these patients in view of its potentially deleterious consequences on cerebral hemodynamics. (HEPATOLOGY 2004;39:471–475.)

Acute liver failure (ALF) is characterized by rapid deterioration in the level of consciousness, and a mortality rate of approximately 90% in patients who fulfill criteria for poor prognosis.1 In the later stages of encephalopathy, elevation in intracranial pressure (ICP) is common as a result of the development of cerebral edema. The exact pathogenic mechanism of the increased ICP is unknown, but current literature supports a multiple hit hypothesis in which ammonia is thought to be central, with a later increase in cerebral blood flow (CBF) accentuating the rise in ICP.2–4 In patients with ALF, autoregulation is impaired, but can be restored if cerebral arteriolar tone is increased by hyperventilation, suggesting that cerebral vasodilatation is responsible for the impairment in autoregulation.5

Hepatorenal syndrome develops in approximately 55% of all patients referred to specialized units with ALF, and frequently, these patients require renal support.6–8 Terlipressin is an inactive vasopressin analog that is metabolized slowly in vivo to its active form, lysine vasopressin. It has been used to treat norepinephrine-resistant hypotension when associated with septic shock,9 and there is an increasing body of evidence that it may be useful in patients with cirrhosis and hepatorenal syndrome.10–14

Vasopressin is a potent systemic vasoconstrictor thought to exert its effects through peripheral V1 receptors. In animal studies, the administration of vasopressin, however, also increases CBF by dilating cerebral arteries.15–17 Vasopressin also may regulate regional CBF by balancing the effects of increased flow mediated by nitric oxide released from the endothelium, with decreased flow in vessels contracted by direct stimulation of smooth muscle.18 It has been postulated that the increase in CBF may be mediated through cerebral V2 receptors.19 Given the distribution of V2 receptors in the brain, it is possible that terlipressin may accentuate cerebral hyperemia and may worsen intracranial hypertension in ALF. Our aims in performing this study were to evaluate whether terlipressin has any deleterious effect on cerebral hemodynamics before conducting a suitable randomized controlled trial of the use of terlipressin in patients with hepatorenal syndrome in ALF. In this study, we evaluated the effect of administration of terlipressin at a subtherapeutic dose to patients with ALF and grade IV hepatic encephalopathy and measured serial changes in systemic hemodynamics, ICP, and CBF.

Abbreviations

ALF, acute liver failure; ICP, intracranial pressure; CBF, cerebral blood flow; JVOS, jugular venous oxygen saturation; MAP, mean arterial pressure.

Materials and Methods

Studies were undertaken at The Scottish Liver Transplantation Unit, Edinburgh between June and December 1999. Approval was obtained by the Lothian Research Ethics Committee, and written informed assent was obtained from the next of kin of each patient in accordance with the Declaration of Helsinki (1989) of the World Medical Association.

Study Design.

Six unselected patients with ALF were enrolled into the study. Patients admitted to the Intensive Care Unit with grade IV hepatic encephalopathy were included into the study. Patients were excluded if they required inotropic support or any specific treatment for elevated ICP before entry into the study. Cardiovascular hemodynamics, ICP, jugular venous oxygen saturation (JVOS), CBF, and blood gases were measured before and then for 5 hours after administration of 0.005 mg/kg of terlipressin (median, 0.25 mg; range, 0.2–0.3 mg) as a single intravenous bolus (terlipressin acetate injection; Ferring Pharmaceuticals Ltd., UK). This low dose was chosen to prevent significant changes in the systemic hemodynamics so as to allow assessment of changes in CBF regardless of the changes in mean arterial pressure (MAP).

Patients.

We studied six patients with ALF (median age, 27 years; range, 22–46 years; four females; causes included acetaminophen [n = 4], acute fatty liver of pregnancy [n = 1], non-A, non-B viral [n = 1]) and four fulfilled the Kings College Criteria for poor prognosis.1 The median concentration of serum bilirubin was 156 μmol/L (range, 102–345 μmol/L), prothrombin time was 67 seconds (range, 49–123 seconds), plasma lactate was 4.5 mmol/L (2.9–6.9 mmol/L), arterial ammonia was 212 μmol/L (range, 167–311 μmol/L), and the arterial pH was 7.32 (range, 7.20–7.43). Median serum creatinine was 312 μmol/L (range, 189–365 μmol/L), and four patients required renal support with continuous veno-venous hemofiltration for hepatorenal syndrome. Hemofiltration was started before and was continued throughout the duration of the study.

Monitoring and Measurements.

All the patients were ventilated mechanically after sedation with propofol. ICP was monitored continuously in five of the patients using a subdural fiberoptic system (Camino; Camino Laboratories, San Diego, CA). The ICP value obtained was recorded at hourly intervals. Cardiovascular hemodynamics were monitored continuously with a Swan-Ganz catheter (Edwards Lifesciences, Irvine, CA). Heart rate, MAP, and cardiac output were recorded at hourly intervals.

Measurement of Cerebral Blood Flow and Jugular Venous Oxygen Saturation.

To measure CBF, an arterial catheter was inserted into the femoral artery and a jugular bulb catheter was inserted into the left internal jugular vein (4F Opticath, U440, Abbott Laboratories, Chicago, IL). Correct positioning of the jugular bulb catheter was confirmed with a lateral head-and-neck radiograph.20 Cerebral venous monitoring via a jugular bulb catheter allows assessment of global oxygen delivery adequacy and does not exacerbate intracranial hypertension.21 JVOS was monitored continuously via the reverse jugular catheter and was recorded hourly. CBF was determined only if the patient was hemodynamically stable, defined as a variation of less than 10% in the mean arterial pressure. Ventilation was adjusted to achieve an arterial carbon dioxide tension of 4 to 4.5 kiloPascals (kPa) and was not altered again during the study to prevent PaCO2 becoming a confounding factor of CBF. Modification of the Kety-Schmidt technique22 was used to measure CBF, which is dependent on the rate of uptake of nitrous oxide by the brain as detailed previously.23 Measurements were made before and then after 1 and 5 hours after the administration of terlipressin.

Statistics.

Data are expressed as median and range. Differences in measured variables between individual time points after administration of terlipressin were calculated using the Wilcoxon signed rank test. Significance was accepted at P < 0.05.

Results

Patients.

Three of the four patients who fulfilled criteria for poor prognosis underwent successful OLT a median of 46 hours (range, 37–67 hours) after admission to the intensive care unit. The fourth patient, who had psychosocial contraindications to orthotopic liver transplantation, died 4 days after admission to the intensive care unit of multiorgan failure. The other two patients recovered without need for OLT and could be discharged from the intensive care unit 73 and 84 hours, respectively, after first presentation with grade IV encephalopathy.

Cardiovascular Hemodynamics.

As shown in Table 1, all patients showed evidence of a hyperdynamic circulation with increased cardiac output and heart rate, and reduced MAP and systemic vascular resistance. After administration of terlipressin, there was minimal increase in the MAP (Fig. 1A) and systemic vascular resistance and an insignificant reduction in cardiac output and heart rate, but none of these changes were either clinically or statistically significant. An increase in hydrogen ion concentration was found to be significant at 5 hours after terlipressin (P = 0.031), although this corresponds to a change in arterial pH from 7.41 to 7.39, which was not thought to be clinically relevant because both values remained within normal limits. There were no other significant changes found (Table 1).

Table 1. Acid-Base Status and Systemic Hemodynamics After Administration of 0.005 mg/kg of Terlipressin Intravenously
 Before Administration30 Minutes After Terlipressin60 Minutes After Terlipressin5 Hours After Terlipressin
  • NOTE. Data expressed as median (range).

  • *

    P < 0.05 compared with that before terlipressin administration. Significance was tested using Wilcoxon signed rank test.

Hydrogen ion concentration (nmol/L)39.1 (37-42) 40.0 (38-43)43.1 (37-45)*
pCO2 (kPa)4.5 (3.9-4.7) 4.4 (4.0-4.5)4.5 (4.2-4.6)
pO2 (kPa)13.1 (11.2-16.1) 12.9 (11.1-17.0)11.7 (10.9-16.3)
Heart rate (per min)94 (73-103)91 (73-108)90 (73-103)93 (77-99)
Cerebral perfusion pressure (mmHg)63 (56-66) 61 (53-66)64 (52-68)
Central venous pressure (mmHg)12 (8-14)11 (9-14)11 (9-15)10 (9-13)
Cardiac output (L/min)10.7 (8.8-13.1)10.3 (9.1-12.9)10.2 (9.1-12.8)10.6 (8.8-13.0)
Systemic vascular resistance (dyn.sec/cm5)464.8 (389-627)502 (396-691)485.6 (418-641.7)502 (402-628)
Figure 1.

Changes in (A) mean arterial pressure (MAP); (B) intracranial pressure (ICP); (C) cerebral blood flow (CBF); and (D) jugular venous oxygen saturation (JVOS) before and after administration of 0.005 mg/kg intravenous terlipressin. Individual patients are represented: patient 1 (⧫); patient 2(□); patient 3 (▴); patient 4 (▪); patient 5 (•); and patient 6 (○). P values were calculated using the Wilcoxon signed rank test. Normal values (used by authors' institution): MAP, 93–100 mmHg; CBF, 45–50 mL/100 g per minute; ICP, 0–15 mmHg; JVOS, 55%–75%.

Cerebrovascular Hemodynamics.

There was a significant increase in CBF from a median of 69 mL/100 g per minute (range, 48–83 mL/100 g per minute) to 81 mL/100 g per minute (range, 62–97 mL/100 g per minute; P = 0.016) 1 hour after administration of terlipressin (Fig. 1C). The CBF returned to baseline values at 5 hours. This was associated with an increase in ICP in all patients from a median of 15 mmHg (range, 13–18 mmHg) to 20 mmHg (range, 16–23 mmHg; P = 0.031) after 1 hour (Fig. 1B), returning to baseline values at 2 hours. In keeping with the increased CBF, the JVOS increased from a median of 75% (range, 67%–89%) to 87% (range, 75%–94%; P = 0.016) 1 hour afterward and remained significantly elevated for 3 hours, returning to baseline values at 5 hours 80% (range, 66%–90%; Fig. 1D).

Discussion

This study demonstrates that administration of low-dose terlipressin produces no significant changes in systemic hemodynamics, but is associated with increases in CBF and a resultant increase in ICP, suggesting the need for extreme caution in the use of this drug in patients with ALF with grade IV hepatic encephalopathy.

Increased ICP and brain herniation is a major cause of mortality in patients with ALF if this is not controlled by repeated mannitol treatments and ultrafiltration.1 At the time this study was planned (1998), there was increasing concern that increased CBF may underlie the pathogenesis of increased ICP in patients with ALF. In particular, data from Aggarwal et al.24 showed that those patients with high ICP have elevated CBF. Since then, there has been an increasing body of literature suggesting a crucial role for cerebral hyperemia as being critical in the development of intracranial hypertension in ALF. Data supporting the above are derived from observations in experimental animals and also from studies in patients with ALF. Studies in portacaval shunted rats administered an ammonia load have demonstrated a rise in CBF that paralleled the increase in ICP and correlated directly with brain water content.25, 26 Cerebral hyperemia is of crucial importance in the development of increased ICP in ALF. Normal cerebral vascular resistance is essential for the maintenance of cerebral autoregulation. Reactive vasodilatation or vasoconstriction ensures constant cerebral perfusion.27 When methods are introduced to control the rise in CBF, such as methionine-sulfoximine (an inhibitor of glutamine synthase)25 and mild hypothermia,28 ammonia-induced cerebral edema is prevented. It also has been shown that the cerebral vasoconstrictor indomethacin blunts the rise in CBF in portacaval shunted rats receiving an ammonia load and that the consequent reduction in CBF leads to a disproportionate reduction of ammonia uptake by the brain, which may reduce brain edema.29

Direct evidence for the role of cerebral hyperemia being important in the pathogenesis of increased ICP has been derived from studies in patients with ALF that have shown that an increase in CBF, induced by a rise in MAP, was associated with an increase in ICP.30 The rise in ICP did not occur if the cerebral hyperemia was prevented from occurring using hypothermia.30, 31 Although data in humans are variable, they support the notion that increased CBF is crucial in the pathogenesis of increased ICP.5, 30–33 Indeed, in the present study the rise in CBF induced by terlipressin was associated with an increase in ICP.

The loss of CBF autoregulation is likely to be mediated through cerebral vasodilatation.34 However, the increase in CBF after administration of terlipressin is unlikely to reflect the effects of altered CBF autoregulation because administration of terlipressin was not associated with any significant changes in the mean arterial pressure and therefore cerebral perfusion pressure. In addition, the effects of a low-dose bolus of terlipressin on CBF and consequent ICP were of short duration and completely reversible, with all values returning to baseline values at 5 hours, which is entirely consistent with the pharmacokinetics of terlipressin, which has a half-life of 50 minutes.35

The exact mechanism by which terlipressin leads to an increase in CBF cannot be elucidated from this study, but we postulate that it is likely to be mediated through cerebral V2 receptors. If the effect had been preferentially mediated through V1 receptors, then we would have expected a rise in MAP as a result of its peripheral vasoconstrictor effects, which was not the case in this study. The results of a recent study by Chung et al.,36 using an experimental animal model of cerebral edema (ammonia infusion after portacaval anastomosis), demonstrating that vasopressin administration results in an increase in CBF and worsening of brain edema support this suggestion. They observed an increase in CBF when vasopressin was administered both in the presence of V1 and V2 receptor antagonists. With V1 receptor antagonism the increase in CBF after vasopressin occurred without a simultaneous increase in the MAP indicating a V2 receptor-dependent mechanism, as was observed in our study. Importantly, they also observed that the animals that were treated with vasopressin had higher ammonia concentrations. Although the exact mechanism of this hyperammonemia is not clear, it may represent reduced perfusion of critical ammonia-removing organs.

In conclusion, the results of this study suggest that administration of even a single subtherapeutic dose of terlipressin to patients with ALF and grade IV hepatic encephalopathy may have deleterious consequences through worsening of cerebral hyperemia and intracranial hypertension. The effect of terlipressin on CBF in patients with ALF and mild hepatic encephalopathy, however, cannot be ascertained from the results of this study. These data suggest that extreme caution should be exercised and that close monitoring is required if this drug is used in patients with ALF and severe hepatic encephalopathy.

Acknowledgements

The authors thank Joseph Eliahoo (joseph.eliahoo@uclh.nhs.uk) for his help and advice with the statistical analysis of the data and the generous support of The Sir Siegmund Warburg Voluntary Settlement.