The Effects of Hormone Resuscitation on Cardiac Function and Hemodynamics in a Porcine Brain-Dead Organ Donor Model

Authors


* Corresponding author: Alfred J. Hing, a.hing@victorchang.unsw.edu.au

Abstract

We compared the effects of hormone resuscitation (HR) with a norepinephrine-based protocol on cardiac function, hemodynamics and need for vasopressor support after brain death in a porcine model. Following brain death induction, animals were treated with norepinephrine and fluids for 3 h. In the following 3 h, they continued on norepinephrine and fluids (control) or received additional HR (triiodothyronine, methylprednisolone, vasopressin, insulin). Data were collected pre-brain death, 3 and 6 h post-brain death. At 6 h, median norepinephrine use was higher in controls (0.563 vs. 0 μg/kg/min; p < 0.005), with 6/8 HR animals weaned off norepinephrine compared with 0/9 controls. Mean arterial pressure was higher in HR animals at 6 h (74 ± 17 vs. 54 ± 14 mmHg; p < 0.05). Cardiac contractility was also significantly higher in HR animals at 6 h (stroke work index 1.777 vs. 1.494). After collection of 6 h data, all animals were placed on the same low dose of norepinephrine. At 6.25 h, HR animals had higher stroke work (3540 ± 1083 vs. 1536 ± 702 mL.mmHg; p < 0.005), stroke volume (37.2 ± 8.2 vs. 21.5 ± 9.8 mL; p < 0.01) and cardiac output (5.8 ± 1.4 vs. 3.2 ± 1.2 L/min; p < 0.005). HR in a porcine model of brain death reduces norepinephrine requirements, and improves hemodynamics and cardiac function. These results support the use of HR in the management of the brain-dead donor.

Introduction

Despite the advances made in the medical management of congestive heart failure, cardiac transplantation remains the only potentially definitive treatment for patients with end-stage cardiac disease (1). The evergrowing demand for transplantation, however, is not matched by the supply of suitable donor hearts (2). This supply shortage is due to the relatively small pool of organ donors, compounded by the fact that a significant number of hearts are not recovered and transplanted, often due to organ dysfunction and an unwillingness to use suboptimal donors (3–6). Additionally, the transplant waiting list mortality rate has been reported to be as high as 17–22% per year (1,3). Clearly, there is a need to optimize organ usage from the currently available donor pool to reduce transplant waiting lists and waiting list mortality.

The success of transplantation is critically dependant upon the quality of the donor organ, which is significantly influenced by the process of donor brain death (6–11). Reported consequences of brain death include hemodynamic instability and donor organ dysfunction (7,9,10,12,13), rejection of organs for transplantation, and an overall increase in posttransplant complications (6,7,13,14). A characteristic feature of brain death is the ‘sympathetic/autonomic storm’ causing extreme hypertension and tachycardia, followed by loss of sympathetic tone and massive vasodilatation leading to hemodynamic instability (13). The substantial increase in endogenous catecholamines that accompanies the ‘storm’ (15) causes organ ischemia (7,10,13) and increases cardiomyocyte intracellular calcium, which in turn initiates a cascade of events leading to disturbed metabolism, cellular injury and death (7,10,11,16). There is also associated myocardial structural damage (17), impaired ATP production (13,18) and free radical-mediated damage (13).

Another consequence of brain death is disruption of the hypothalamic-pituitary axis, resulting in decreased circulating levels of triiodothyronine (8,16,18,19), arginine vasopressin (8,12,20), cortisol/adrenocorticotrophic hormone (8,13) and insulin (8,13). These changes have been associated with impaired aerobic metabolism, increased anaerobic metabolism, depletion of high energy phosphates and increased lactate production (8,13,16,21). Endothelial, platelet and leukocyte activation is also seen, along with up-regulation of MHC class II antigens, and pro-inflammatory mediators such as cytokines (e.g. TNF-α IFN-γ, IL-6), chemokines (e.g. MIP-1α MCP-1) and adhesion molecules (13,22–26). This immune system activation enhances donor organ immunogenicity, stimulating the recipient's immune system and potentially precipitating graft rejection, as well as affecting medium- to long-term graft function (13,22,23,27).

There have been reports that a structured donor management algorithm incorporating hormone replacement may ameliorate many of the negative effects of brain death, and improve cardiac function and donor hemodynamics in both animal models and humans (3,6,16,21,28). Based on these reports, a deceased donor management protocol for hemodynamically unstable donors was developed in the United States, incorporating invasive hemodynamic monitoring and hormone resuscitation (HR) (3,29). However, to our knowledge, there have been no prospective, randomized controlled studies of the complete HR protocol advocated in the United Network for Organ Sharing (UNOS) Critical Pathway. Therefore, we conducted a randomized controlled study to examine the impact of the complete HR protocol on hemodynamics, cardiac function and the need for inotrope support in a porcine model of brain death.

Materials and Methods

A porcine model of the brain-dead organ donor which closely mimics the deceased human donor was used, as previously described (30). This model was approved by our animal ethics committee and was utilized in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (31).

Animals and anesthesia

Twelve Landrace and 10 Westran pigs (31.85–60.00 kg) were used. Pigs were premedicated with an intramuscular injection of ketamine 10 mg/kg, midazolam 1 mg/kg and atropine 0.05 mg/kg. Once sedated, general anesthesia was induced using intravenous (IV) thiopentone 2–10 mg/kg. Animals were then intubated and ventilated with 100% oxygen. Anesthesia was maintained with isoflurane 1–4% inhaled gas and fentanyl 100–150 μg IV, as required. Intravenous saline (0.9%) was infused at 10 mL/kg in the first hour, followed by 5 mL/kg/h titrated to a central venous pressure (CVP) of 0–5 mmHg. Continuous monitoring of animals included: electrocardiogram, invasive arterial blood pressure (via a femoral arterial catheter), CVP (via an internal mammary central venous catheter), pulse oximetry, end tidal expired CO2 and core temperature. Arterial blood gas and glucose analyses were performed at hourly intervals, and when clinically indicated, using an i-STAT Portable Clinical Analyzer (i-STAT Corporation, USA).

Surgery for brain death induction and data acquisition

A burrhole was made in the right frontoparietal skull and a Foley catheter placed in the subdural space. Lignocaine 1 mg/kg IV was given for arrhythmia prophylaxis, a median sternotomy was performed and the heart exposed. Arrhythmias were treated with additional lignocaine and internal direct DC defibrillation (10–30 J) if required.

A flow probe (Transonic Systems Inc., USA) was placed around the left anterior descending coronary artery (LAD) and a micromanometer-tipped catheter (Millar Instruments Inc., USA) was placed into the left ventricle (LV). Six 2-mm ultrasonic dimension transducers (Sonometrics Corp., Canada) were sewn onto the epicardium to measure the base-apex major axis, anterior–posterior minor axis, and the left ventricular free wall-right ventricular free wall minor axis diameters of the heart. Data from these devices were digitised at 200 Hz.

Brain death induction

After acquisition of baseline data, brain death was induced by inflating the Foley catheter balloon with water in 3 mL increments every 30 s to a total of 24 mL. This increased the intracranial pressure, causing an autonomic storm similar to that seen clinically. Brain death was confirmed by the typical hemodynamic changes of brain death, the absence of response to painful stimuli, and the absence of pupillary and corneal reflexes. Anesthesia was ceased 15 min after balloon inflation to also confirm brain death.

Donor management

If despite IV fluid resuscitation, the mean arterial blood pressure (MAP) fell below 60 mmHg following brain death, an IV norepinephrine infusion (20 μg/mL) was commenced and titrated to achieve MAP 60–70 mmHg. Three hours after brain death induction, animals were randomly assigned to commence on a HR protocol (Table 1) (in addition to norepinephrine) or remain on norepinephrine only (control) for a further 3 h. In HR animals, vasopressin was increased incrementally to a maximum of 4 U/h to maintain MAP 60–70 mmHg, and then if required, norepinephrine increased. Attempts were made to wean animals off norepinephrine in both groups whilst maintaining MAP. If successful, attempts were then made to wean animals off vasopressin, again whilst maintaining MAP 60–70 mmHg.

Table 1.  Hormone resuscitation protocol
HormoneDose
Methylprednisolone15 mg/kg IV bolus
Triiodothyronine4 μg IV bolus, followed by 4 μg/h  IV infusion
Vasopressin0.5–4.0 units/h titrated to mean  arterial pressure 60–70 mmHg
InsulinTitrated to blood sugar level 6–10  mmol/L at minimum 1 unit/h IV

Data acquisition and analysis

Hemodynamic data were recorded immediately before and during transient inferior vena cava occlusion. Data were recorded prior to brain death (baseline), and then at hourly intervals for 6 h after brain death induction. After data were recorded at 6 h, norepinephrine was infused at 3.3 μg/min for 15 min in both control and HR animals and another set of data taken (6.25 h post-brain death induction) to exclude norepinephrine dose as a confounder.

Data acquisition and analyses were undertaken using SonoSOFT Software versions 3.1.6 and 3.3.2 (Sonometrics Corp.). The prolate ellipsoid model was used to calculate epicardial left ventricular volume (LVVepi) from the cardiac dimension data (LVVepi=π a.b2/6, where a = major axis length and b = minor axis length) and pressure–volume (PV) loops were derived (Figure 1) (30). Stroke work (SW) was calculated as the PV loop area for each heart beat. The relationship between SW and end-diastolic volume, termed the preload recruitable stroke work (PRSW) relationship (Figure 2), was determined using linear regression analysis of the data acquired during vena caval occlusion (32).

Figure 1.

Left ventricular (LV) pressure-volume loops. Representative loops obtained during transient vena caval occlusion and taken from the same hearts at baseline and at 6.25 h post-brain death in Control and Hormone Resuscitation groups. Note the decline in PV loop area (stroke work) in the control animal compared with the hormone resuscitation animal between baseline and 6.25 h.

Figure 2.

Preload recruitable stroke work (PRSW) relationship. Representative PRSW relationships from the Control and Hormone Resuscitation groups are compared.

Primary outcomes for the study were norepinephrine requirements, LV contractile function (PRSW relationship) and hemodynamic indices. LAD flow, arterial blood gases, blood sugar and troponin I were also assessed. Troponin I was assayed using the AxSYM microparticle enzyme immunoassay platform (Abbott Laboratories, USA).

Statistical analysis

Statistical analyses were performed using SPSS for Windows 12.0.1 (SPSS Inc., USA). Continuous variables were reported as mean ± SD and categorical variables as actual prevalence in the group. Differences between groups with normally distributed data were compared using Student's t-test for independent samples. The proportion of animals in each group weaned off norepinephrine was compared using Fisher's exact test. Non-parametrically distributed data were reported as median (range) and compared using the Mann-Whitney U test. Differences were considered statistically significant at p < 0.05.

The significance of differences in the PRSW relationship was determined using a multiple linear regression (MLR) implementation of two-way analysis of covariance with repeated measures (30). To reduce potential confounders, LVVepi and SW were normalized within individual animals to their baseline steady state values. The normalized volume axis intercept (nVw) and stroke work index (SWI) were derived from the MLR model of the PRSW relationship. Overall change in LV contractility as reflected in the PRSW relationship is determined by changes in slope and nVw. Ischemia causes a decrease in slope and an increase in nVw. SWI is the regression estimate of the group's mean normalized SW at baseline steady state end-diastolic volume. It represents the net effect of the interaction between changes in slope and nVw on normalized SW at the normal operating volume of the heart (the most physiologically relevant end-diastolic volume).

Results

Twenty-two animals were used in this study: nine in the control group (five Landrace and four Westran) and eight in the HR group (five Landrace and three Westran). Five animals were excluded for technical reasons: one had a large atrial septal defect, one had cardiomyopathy, two had anesthetic-related deaths, and one sustained a significant LAD injury. Both Westran and Landrace pigs were used due to supply issues.

The mean animal weight in the control group was 48.8 ± 5.5 kg, compared with 45.7 ± 6.9 kg in the hormone group (p = 0.33). Mean LV volume was 116.1 ± 11.4 mL in the control group and 111.8 ± 17.3 mL in the hormone group (p = 0.56). Mean LV volume to body weight ratio was 2.4 ± 0.2 in the control group and 2.5 ± 0.2 in the hormone group (p = 0.58).

One HR and two control animals had ventricular fibrillation arrests prior to brain death induction. These arrests occurred during cardiac instrumentation for data acquisition, and all were successfully cardioverted to sinus rhythm. Additional doses of lignocaine were administered to animals that had arrested, and to one additional control and two HR animals for ventricular ectopic beats post-brain death induction.

Inotrope requirements

Norepinephrine doses required to maintain MAP were non-parametrically distributed. At 1 h post-brain death induction, all animals required norepinephrine to support blood pressure, with a median dose of 0.066 (0.033–0.146) μg/kg/min in control and 0.111 (0.053–0.209) μg/kg/min in HR animals. By 3 h post-brain death, 8/9 control and 8/8 HR animals were still requiring norepinephrine, with a median dose in the control group of 0.065 (0.000–1.083) μg/kg/min and 0.219 (0.060–1.560) μg/kg/min in the HR group (p = 0.093). At 6 h post-brain death, median norepinephrine dose was 0.563 (0.017–9.333) μg/kg/min in the control group and 0 (0–0.314) μg/kg/min in the HR group (p < 0.005), with the control animal off norpinephrine at 3 h re-commencing norepinephrine (Figure 3).

Figure 3.

Norepinephrine usage in individual animals: (A) control (n = 9) and (B) hormone resuscitation (n = 8).*Norepinephrine dose at 6 h was 9.333 μg/kg/min; #Norepinephrine dose at 6 h was 6.279 μg/kg/min. At 1 h post-brain death, all animals required norepinephrine to support blood pressure. By 3 h post-brain death, 8/8 hormone resuscitation and 8/9 control animals required norepinephrine. By 6 h, 2/8 hormone resuscitation and 9/9 control animals required norepinephrine. See text for further details on norepinephrine requirements.

Norepinephrine was weaned off within 6 h of brain death in 0/9 control and 6/8 HR animals (p < 0.005). All animals weaned off norepinephrine were also weaned off vasopressin. The two HR animals still requiring norepinephrine were on a lower dose at 6 h post-brain death, compared with 3 h post-brain death (0.314 vs. 0.576 μg/kg/min and 0.071 vs. 0.284 μg/kg/min, respectively). These two animals were also receiving vasopressin infusions (0.0005 and 0.0012 units/kg/min).

Ventricular contractility

Representative PV loops from the control and HR groups in Figure 1 demonstrate the decline in stroke work (PV loop area) in a control animal compared with a HR animal between baseline and 6.25 h. Representative PRSW relationships in Figure 2 demonstrate a decrease in slope in a control animal and an increase in slope in a HR animal between baseline and 6.25 h. Results of the MLR analysis of the PRSW relationship are shown in Table 2. There was no difference in LV contractility pre-brain death between control (SWI = 0.974) and HR (SWI = 1.012) groups (p = 0.08). By 6 h post-brain death, SWI had increased 53% to 1.494 in the control group and 76% to 1.777 in the HR group. At a fixed norepinephrine dose at 6.25 h, the SWI was maintained in the HR group at 1.770, but declined in the control group to 0.918. MLR analysis confirmed that LV contractility was significantly higher in the HR group at both 6 and 6.25 h post-brain death induction (p < 0.0001).

Table 2.  Preload recruitable stroke work (PRSW) relationship; control (CON; n = 9) and hormone resuscitation (HR; n = 8) groups
GroupSlope (nMw)y-Axis interceptNormalized volume-axis intercept (nVw)Stroke work index (SWI)
  1. nMw and y-axis intercept are the group mean ± standard error of the normalized PRSW. nVw and SWI calculated from nMw and y-axis intercept.

  2. *Norepinephrine infusion fixed at 3.3 μg/min.

Pre-brain death (CON)2.727 ± 0.088−1.753 ± 0.0760.6430.974
Pre-brain death (HR)2.849 ± 0.065−1.837 ± 0.0560.6451.012
3 h post-brain death (CON)4.464 ± 0.177−2.704 ± 0.1360.6061.760
3 h post-brain death (HR)4.360 ± 0.102−2.715 ± 0.0750.6231.645
6 h post-brain death (CON)4.035 ± 0.258−2.541 ± 0.2070.6301.494
6 h post-brain death (HR)4.478 ± 0.119−2.701 ± 0.0860.6031.777
6.25 h post-brain death* (CON)2.430 ± 0.108−1.512 ± 0.0810.6220.918
6.25 h post-brain death* (HR)4.833 ± 0.146−3.063 ± 0.1110.6341.770

Hemodynamic changes

Hemodynamic data are shown in Table 3. At baseline and 3 h post-brain death induction (and before HR was commenced), there were no differences in MAP between groups. Once HR was commenced, MAP was higher at 6 and 6.25 h compared with 3 h post-brain death in the hormone group, whereas MAP was lower in the control group at 6 h and lower still at 6.25 h, compared with 3 h post-brain death. Consequently, MAP was higher in the HR group compared with the control group at 6 h (74 ± 17 vs. 54 ± 15 mmHg; p < 0.05) and 6.25 h post-brain death (72 ± 21 vs. 38 ± 11 mmHg; p < 0.005).

Table 3.  Hemodynamics: control (CON; n = 9) and hormone resuscitation (HR; n = 8) groups
GroupMean arterial pressure (mmHg)Heart rate (bpm)Cardiac output (L/min)Stroke work (mL.mmHg)LAD flow (mL/min)
  1. *Norepinephrine infusion fixed at 3.3 μg/min. p < 0.05. p < 0.005. LAD = left anterior descending coronary artery.

Pre-brain death (CON)64 ± 1899 ± 264.7 ± 1.13146 ± 108723 ± 10
Pre-brain death (HR)59 ± 1391 ± 124.1 ± 0.52916 ± 47328 ± 5
3 h post-brain death (CON)63 ± 9161 ± 255.6 ± 0.93356 ± 72435 ± 14
3 h post-brain death (HR)61 ± 7166 ± 274.6 ± 1.42634 ± 90344 ± 17
6 h post-brain death (CON)54 ± 15186 ± 415.1 ± 1.92980 ± 152847 ± 23
6 h post-brain death (HR)74 ± 17157 ± 195.3 ± 2.23224 ± 115845 ± 25
6.25 h post-brain death* (CON)38 ± 11155 ± 253.2 ± 1.21536 ±70231 ± 14
6.25 h post-brain death* (HR)72 ± 21156 ± 155.8 ± 1.43540 ± 108351 ± 15

There were no differences in heart rate between groups at baseline (p = 0.429) and after brain death induction (Table 3; p = 0.698, 0.092 and 0.880 at 3, 6 and 6.25 h post-brain death, respectively). Cardiac output (CO) was higher in the HR group at 6.25 h post-brain death (5.8 ± 1.4 vs. 3.2 ± 1.2 L/min; p < 0.005), but not at any other time-points (Table 3; p = 0.189, 0.119 and 0.858 at baseline, 3 and 6 h post-brain death). Similar trends were seen in SW (Table 3) and stroke volume (SV): SW was higher in the HR group at 6.25 h post-brain death (3540 ± 1083 vs. 1536 ± 702 mL.mmHg; p < 0.005), as was SV (37.2 ± 8.2 vs. 21.5 ± 9.8 mL; p < 0.01). There were no significant differences in SW or SV between groups at any other time-points. LAD flow was not significantly different between groups at any time-point (Table 3).

Troponin I, blood sugar levels and pulmonary function

Troponin I levels increased progressively in both groups over time, but no significant differences between groups were identified at any time-point (Table 4).

Table 4.  Troponin I (μg/L; median and range): control (n = 7) and hormone resuscitation (n = 7) groups
TimeControlHormone resuscitation
  1. Results from three animals excluded due to change in laboratory assay — results from the new assay not comparable to the original AxSYM assay.

Pre-surgery0.0 (0.0–6.0)0.0 (0.0–3.5)
Post-instrumentation1.3 (0.6–13.5)2.4 (0.5–22.5)
1 h post-brain death5.1 (1.8–20.2)6.4 (2.5–31.2)
3 h post-brain death12.9 (5.4–31.9)14.1 (8.8–38.3)
6 h post-brain death22.7 (16.3–500)30.5 (13.9–106.1)

There was no significant difference in blood sugar levels between groups at baseline. Three hormone animals required insulin for hyperglycemia: one requiring 10 units over the final 3 h of management and the other two requiring 5 and 10 units. After correction with insulin, there were no significant differences in blood sugar levels at 6 h post-brain death induction (7.45 ± 2.44 mmol/L in hormone animals vs. 10.64 ± 4.53 mmol/L in controls; p = 0.096), despite three control animals having levels >10 mmol/L at 6 h.

There were no differences between groups with respect to PaO2, PaCO2 and alveolar-arterial (Aa) gradient at baseline or at 3 h post-brain death (i.e. before commencing HR). As FiO2= 1.0 in all animals, PaO2/FiO2= PaO2, and hence PaO2 was comparable between animals. By 6 h post-brain death, there were no significant differences between groups in PaO2 (346 ± 131 mmHg in HR vs. 336 ± 160 mmHg in control; p = 0.896), PaCO2 (62 ± 13 vs. 56 ± 20 mmHg; p = 0.481), or Aa gradient (289 ± 133 vs. 306 ± 168 mmHg; p = 0.822).

Discussion

We used a porcine model to compare the combined HR protocol endorsed in the UNOS Critical Pathway (3,29) with a norepinephrine-based protocol for deceased donor management. Norepinephrine was used because it is the most commonly used inotrope in donor management in Australia and New Zealand. In 2005, 94% of donors in Australia and New Zealand received vasopressor support with norepinephrine being used in 81% of cases (4). Similarly high rates of catecholamine administration have been reported in other registries such as the Eurotransplant registry, where 91% of brain-dead donors were maintained with catecholamines and one-third donors receiving more than one catecholamine infusion (33).

One of the most striking results from our study was the impact of HR on the need for norepinephrine. Six of eight hormone-treated animals were weaned off norepinephrine, with the remaining two requiring lower doses of norepinephrine at 6 h compared with 3 h post-brain death when hormonal therapy was commenced. In contrast, none of the control animals were able to be weaned off norepinephrine. Rather, their dosage requirements increased significantly over the final 3 h of donor management, indicating the development of tachyphylaxis. Indeed, several control animals were unable to maintain the target MAP specified in the study protocol despite rapid escalation in the norepinephrine dose. The hemodynamic status of these animals closely mimicked that of human donors that become hemodynamically unstable after brain death. Such patients are often managed with further catecholamines and additional intravenous fluids, both of which can further impair donor heart function (13,14,33). Administration of norepinephrine, in particular, has been associated with myocardial damage and initial nonfunction after cardiac transplantation (33–35). As a result, hearts from donors receiving high doses of catecholamine, particularly in the presence of hemodynamic instability, are frequently rejected for use in transplantation (3,4,28).

Several investigators have demonstrated experimentally that cardiac contractility as assessed by the PRSW relationship deteriorates in untreated animals after brain death (36–38). The reasons for this decline are likely to be multifactorial. The fall in cardiac contractility is biphasic with an initial transient decrease which may be caused by myocardial ischemia (38) or rapid desensitization of the β-adrenergic signalling pathway (39) during the ‘autonomic storm’. This is followed by a sustained reduction in contractility which may be due to ongoing catecholamine-induced myocardial ischemia (10, 37), impaired β-adrenergic signaling (39), impaired myocardial metabolism associated with the altered neurohormonal environment after brain death (16,37,40) or upregulation of pro-inflammatory mediators (7,37). Analysis of the PRSW relationship in our study demonstrated that hormone-treated animals had increased cardiac contractility at 6 h post-brain death, both in comparison with baseline (pre-brain death) and with norepinephrine-treated animals. The difference in contractility between the two groups was further accentuated at 6.25 h post-brain death when both groups were placed on same dose of norepinephrine (Table 2, Figure 2). In a similarly designed study using cross-bred pigs, Lyons et al. reported that bolus administration of methylprednisolone prior to or soon after brain death prevented any deterioration in cardiac contractility and preserved the PRSW at baseline levels for up to 6 h postinduction of brain death (37). Considering this finding in relation to our study, it suggests that the methylprednisolone given as part of the HR protocol in our study may have contributed to but cannot explain the increased contractility observed in hormone-treated animals. Furthermore, as 6 of 8 animals in the hormone-treated group were weaned off vasopressin by 6 h post-brain death, we speculate that infusion of triiodothyronine was responsible for the increase in contractility observed in these animals. Triiodothyronine has been shown to have direct positive inotropic effects in isolated hearts in vitro (41,42), and in normal and cardiomyopathic hearts in vivo (43). The positive inotropic action of triiodothyronine appears to be mediated via up-regulation of sarco-endoplasmic reticulum Ca2+ ATPase (SERCA) (44–46) and is independent of the β-adrenergic signaling pathway (42).

Norepinephrine treatment after brain death also increased cardiac contractility in our study; however, the positive inotropic response at 6 h post-brain death was less than that observed at 3 h despite up-titration of the norepinephrine dose, indicating tachyphylaxis to norepinephrine. This finding is consistent with other reports of rapid desensitization of the myocardial β-adrenergic signalling pathway after brain death (39,47), probably triggered by the high circulating levels of endogenous norepinephrine post-brain death. Administration of norepinephrine as in our study is likely to further desensitize myocardial β-adrenergic signalling. Another potential explanation for the loss of contractile responsiveness to norepinephrine with time post-brain death is the development of catecholamine-induced myocardial injury. This phenomenon has been demonstrated in a variety of experimental and clinical settings (33–35) and appears to be mediated via β-adrenergic stimulation, as it can be abrogated by pharmacological beta blockade (48).

Chiari et al. reported a small increase in troponin I levels after brain death in pigs, consistent with myocardial ischemic injury (49). Troponin I levels recorded 3 h after brain death in their study were comparable to those recorded in our study, allowing for differences in assay techniques and in the degree of surgical preparation of the heart between the two studies. Serial measurements of plasma troponin I levels in our study revealed no significant differences between treatment groups at any time-point post-brain death. This observation should be interpreted cautiously for several reasons. A small rise in troponin I levels was evident in both groups of animals prior to induction of brain death, indicating a contribution of surgical preparation of the heart to the rise in troponin I level after brain death. Other variables that could affect troponin I levels such as cardioversion occurred in some animals but not others. Finally, the range in troponin I values was wide and non-parametrically distributed indicating that the study was underpowered to detect differences in troponin I release between the two treatment protocols. Bearing in mind these caveats, it is noteworthy that the enhanced contractility observed in animals that received HR was not associated with increased myocardial injury as judged by troponin I release. The possibility that both treatment protocols increased myocardial troponin I release cannot be excluded. However this is unlikely, as we have found in ongoing experiments that the rate of increase in troponin I after brain death was similar for both treatment protocols when compared with saline treated control animals up to 6 h post-brain death (50).

Our study showed that overall hemodynamic status was better in hormone-treated animals. Cardiac output was well maintained by both treatment protocols; however, there were marked differences in arterial blood pressure responses. By 6 h post-brain death, control animals had a significantly lower MAP despite up-titration of the norepinephrine dose. As the cardiac output was similar in both groups at 6 h (Table 3), the difference in MAP can only be explained by differing actions of the treatment protocols on systemic vascular resistance. This indicates the development of rapid desensitization to norepinephrine at the vascular level, which occurred in parallel with the myocardial desensitization described above. The cause for this vascular desensitization is unclear but could be explained by high circulating levels of endogenous catecholamines after brain death (37,49) acting directly on the α1-adrenergic receptor (51), or by the loss of other vasoconstrictor mechanisms after brain death (14,20,52). Regardless of the mechanism, our findings indicate that norepinephrine alone is a poor choice to maintain blood pressure in the brain dead donor.

Lyons et al. reported that both cardiac output and arterial blood pressure fell over 6 h after brain death induction in pigs (37). Interestingly, methylprednisolone did not prevent the fall in cardiac output, even though it maintained cardiac contractility (37). This finding indicates that the fall in cardiac output resulted from altered loading conditions on the heart, either reduced preload or increased afterload. Their findings suggest that methylprednisolone at least partly preserved peripheral vascular resistance, although this was not measured directly. In comparison, the combined hormonal protocol used in our study maintained both blood pressure and cardiac output at or above the levels measured prior to brain death.

High-dose steroids have been shown to block up-regulation of inflammatory mediators (53), prevent ischemic damage (54), and improve organ function and graft survival after transplantation (23,27,55). High dose steroids was associated with improved oxygenation and lung recovery in a large retrospective analysis of potential lung donors (27). We did not detect any difference between the effect of combined hormonal therapy and norepinephrine treatment on lung function after brain death. Donor lung function as assessed by serial measurements of PaO2, PaCO2 and Aa gradient remained stable over the 6 h after induction of brain death. In the study of Follette et al. (1998), donors were managed for 23.5 h after receiving high dose steroid therapy (27). Possibly, a longer period of observation post-brain death may have revealed treatment differences in our study. The use of high dose steroids in the combined hormonal therapy did not adversely affect blood sugar control. Indeed, blood sugar levels tended to be lower in animals that received hormonal therapy (7.5 ± 2.4 vs. 10.6 ± 4.5 mmol/L; p = 0.096).

There are a number of limitations in our study to be considered. The duration of our study was limited to 6 h after brain death, although clinically the majority of multiorgan donors are managed for longer periods after brain death (4). However, in light of the trends that were apparent by 6 h post-brain death, we believe that the differences between treatments would have been even greater had the animals been followed for longer. Also, we did not assess the impact of individual hormones on the donor, nor did we assess the effects of the treatment regimes on intra-abdominal organs. Posttransplantation outcomes with these treatments were also not assessed. There have been concerns that vasopressin may impair function and preservation of abdominal organs such as the liver, pancreas and kidney because of its nonselective vasoconstrictive effects (14). Observational data suggest that these concerns may be unfounded, however, as the retrieval of all intra-abdominal organs is increased and the risks of primary graft dysfunction and graft loss appear to be reduced when the donor has received combined hormonal therapy after brain death (5,55–57). Nonetheless, the lack of a prospective randomized study limits any conclusions that can be made regarding the impact of combined hormonal therapy on the quality of intra-abdominal donor organs.

To our knowledge, this study is one of the few, if only, prospective, randomized controlled studies testing combined HR, and examining its effects on donor cardiac function, hemodynamics and vasopressor requirements. Previous experimental studies have tested single hormonal replacement or a combination of hormonal therapies (5, 21) but not the complete hormone replacement protocol advocated in the UNOS Critical Pathway. Clinical studies of the hormone replacement cocktail have been observational, retrospective and used nonmatched controls (28, 55). In addition, the contribution of other changes in donor management such as invasive hemodynamic monitoring make it difficult to judge the relative contribution of combined hormonal resuscitation to the improvements in donor organ outcomes reported in these studies. The results of our study support the use of combined HR to enhance donor heart function; however, there is a clear need for the impact of combined HR on intra-abdominal donor organs to be assessed. This is being addressed in an ongoing randomized study in our laboratory.

In conclusion, this study has demonstrated that combined HR as advocated in the UNOS Critical Pathway increases donor cardiac contractility, stabilizes hemodynamics and markedly reduces the need for catecholamine support. The study also highlights the limitations and potential adverse consequences of norepinephrine on cardiac function and hemodynamic status after brain death. Translation of these findings to clinical practice would be expected to improve the quality of donor hearts including ‘sub-optimal’ hearts that would otherwise be rejected for transplantation.

Acknowledgments

This study was supported by a National Health and Medical Research Council of Australia (NHMRC) grant. Dr. Hing received scholarship support from the Cardiac Society of Australia and New Zealand, NHMRC and the Royal Australasian College of Surgeons.

Ancillary