Congenital heart defects occur in 0.8% of all newborns.1 The correction of these defects often requires the implementation of a heart–lung machine (HLM). However, the use of a HLM involves inherent risks. Partly due to the disparity between the small blood volume of newborns and the comparatively huge foreign body surface area of a HLM, severe complications can ensue, which can culminate in multiorgan failure and death. For this reason, recent endeavors have been made to miniaturize the HLM, with the aim of reducing the filling volume and foreign body surface area, thereby lessening the complications. A further objective is complete avoidance of foreign blood administration.
Following the design and construction of our miniaturized HLM (MiniHLM) and preliminary in vitro testing, testing by means of an animal model was necessary. Although a suitable animal model does not exist for every disease and application, adequate testing through the use of an animal model has become indispensible to modern medical research.2
We designed a rabbit animal model to test our newly developed MiniHLM.3–5 For this purpose, we had many stipulations pertaining to anatomy, including size, weight, and also the blood volume, which should reflect the newborn patients for whom the MiniHLM was constructed. These specifications are multifactorial.6 Furthermore, as several of the complications of HLMs are induced by a rise in inflammatory parameters,5,7 the tested animals should demonstrate an expression of inflammatory parameters that is similar to humans. Further blood testing with reference to hemolysis and fibrinolysis4 should also be feasible. To prevent falsification of the results through foreign blood administration, it should be possible to complete the trials completely without foreign blood.
In relation to the aforementioned stipulations, the rabbit fulfills the requirements for serving as the small animal model. It is similar with regard to its anatomy, including blood volume, to newborn children. Although the rabbit possesses a persistent left-sided superior vena cava, this plays only a subordinate role for cannulation and connection to the HLM. The blood volume of infants is approximately 75 ml/kg body weight, and the blood volume of rabbits is 58 ml/kg. Thus, testing of the HLM using a rabbit model would mean testing under more strict and difficult conditions. In contrast to the rat model,8–10 the ascending aorta and the right atrium can be cannulated after sternotomy exactly as done in the case of a newborn, and it is possible to sustain adequate flows. Also, comparability of hemolysis,4,11 as well as inflammatory5,12,13 parameters between rabbits and humans has been reported.
Material and Methods
In the first series, 13 New Zealand white rabbits (n = 13; 4.04 ± 0.52 kg) were initially used for testing of the MiniHLM 01 (first version of MiniHLM). The differences between the MiniHLM 01 and the MiniHLM 02 (second version) were described in detail by Arens et al.14 (Table 1).
As two of the 13 rabbits had preexisting heart disease preoperatively (one rabbit had high-grade aortic valve stenosis and the other had sustained an old myocardial infarction), we made the decision to transition to a different, more robust, breed of rabbit.
In the second series, we operated on a total of 19 female Chinchilla Bastard rabbits. Due to various problems, four animals could not be included in the analysis (one animal on account of protamine administration, one animal as a result of thoracic surgical bleeding, and two animals due to bleeding at the groin site), and a total of 15 rabbits were included in the final analysis.
Seven Chinchilla Bastard rabbits were perfused with the MiniHLM (dynamic priming volume 127 ml; Figure 1). Eight animals serving as a control were perfused using Dideco Kids D100 oxygenator (Sorin Group, Milan, Italy) and a Stöckert roller pump (modified dynamic priming volume 149 ml). The modifications made on the Dideco Kids D100 systems for the purpose of comparison are not suitable for normal operations, as the perfusionist was required to sit halfway beneath the operating table. The mean total body weight of the animals was 3.65 ± 0.44 kg (Dideco Kids group: 3.63 ± 0.43 kg; MiniHLM group: 3.69 ± 0.48 kg [n.s.]).
All animals were treated in accordance with the German guidelines for the care of laboratory animals, which correspond with the National Institutes of Health guidelines. The in vivo testing was authorized by the state governmental Animal Care and Use committee of North Rhine-Westphalia, Germany.
Premedication was performed with the use of intramuscular injections of medetomidine hydrochloride (0.2 ml/kg, Domitor, Pfizer, Berlin, Germany), midazolam (1.0 mg/kg, Dormicum, Roche, Austria), and fentanyl citrate (0.025 mg/kg, Fentanyl- Jansen, Jansen Cilag, Neuss, Germany). Placement of a urinary catheter was not possible for anatomical reasons, thus the urine was collected in a diaper and weighed.
After premedication, an electrocardiogram was connected to monitor the heart rate. While spontaneously breathing, the rabbits were shaved in the sternal region, the back (for placement of the electrocautery grounding pad), and the groin. After induction of anesthesia (fentanyl citrate [0.016–0.032 mg/kg/h], medetomine hydrochloride [0.04–0.08 mg/kg/h], and ketamine hydrochloride [87 mg/kg/h] Ketamin 10%, Ceva, Düsseldorf, Germany), the animals were mechanically ventilated. Catheters were placed via open surgical access in the left femoral artery (Leaderflex 22 gauge, Vygon, Aachen, Germany), for blood pressure monitoring and blood sampling, and percutaneously into the ear vein (Vasofix 20 gauge, Braun, Melsungen, Germany), for drug administration. ECG, arterial blood pressure, and rectal temperature were monitored continuously with a surveillance monitor (Datex-Ohmeda OM 3; General Electrics, Solingen, Germany).
The heart was exposed via a standard median sternotomy. Following systemic anticoagulation with intravenous heparin (3 mg/kg), cardiopulmonary bypass (CPB) was established through cannulation of the ascending aorta (8 F Fem-Flex pediatric cannula, Edwards Lifesciences Irvine, CA) and the right atrium (DLP 12 F angulated metal tip cannula, Medtronic, Meerbusch, Germany; Figure 2). Venous blood was drained kinetically. During CPB, the lungs were not ventilated. Cardiopulmonary bypass was conducted in the setting of mild hypothermia (32˚ C). The persistent left-sided superior vena cava was clamped off.
After initiation of the HLM, the aorta was clamped. Cardiac arrest was achieved with cardioplegia. In the first two animals, we induced cardiac arrest with Bretschneider cardioplegia solution (170 ml; Custodiol; Koehler Chemie, Bensheim, Germany). However, because of massive problems with decreased hematocrit, we consequently changed the method. Blood cardioplegia was used for the subsequent rabbits. For this purpose, a solution of 20 ml potassium chloride (14.9%, Braun Melsungen, Melsungen, Germany), 5 ml magnesium chloride (50%, Inresa, Freiburg, Germany), and 25 ml sodium chloride (0.9%, Braun Melsungen, Melsungen, Germany) was prepared. Initially, 20 ml venous rabbit blood was infused with 2 ml of this solution by means of a pediatric aortic root cannula (DLP, Medtronic, Minneapolis, MN) inserted into the clamped off ascending aorta, and 20 ml rabbit blood was infused with 1 ml of solution through the pediatric root cannula every 10 minutes as a maintenance dose. After 1 hour aortic clamp time, the HLM was gradually reduced, the vessels decannulated, and the sternum reclosed.
Arterial pressure (mm Hg) and perfusion parameters consisting of blood flow (l/min) and arterial line pressure (mm Hg) were monitored continuously with a surveillance monitor (Datex-Ohmeda OM 3; General Electrics, Solingen, Germany).
Blood samples were collected before skin incision (pre-CPB), 15 minutes after opening the aorta (during CPB), and 30 minutes after CPB (post-CPB) to check for hemoglobin (g%), plasma-free hemoglobin (mg%), fibrinogen (g/L), and blood gas analysis (pCO2 in mm Hg, pO2 in mm Hg, pH, ABEc in mmol/L, and lactate in mmol/L, ABL 510, Radiometer, Willich, Germany).
Blood for examination of inflammatory response (tumour necrosis factor -α, interleukin [IL]-1β, IL-6, IL-8, and IL-10) was taken before skin incision, 5 minutes before opening the aorta, 15 minutes after opening the aorta, and 4 hours after the initiation of CPB. The parameters of inflammation were expressed by means of the comparative CT method (ΔΔCT method).
In the first series with 13 New Zealand White rabbits, CPB was performed with all animals for 1 hour. However, postoperative cardiac failure was encountered for varying reasons in eight animals.
In the next series involving 19 Chinchilla Bastard rabbits, 15 of the 19 rabbits were successfully weaned from CPB. There were no HLM hardware-related technical problems. Visual inspection ruled out thrombus formation in the bypass circuit. Foreign blood was not administered in all cases.
Priming volume of the Aachen MiniHLM 02 (127 ± 3 ml; p = 0.001) was lower than the priming volume of Dideco Kids (149 ± 12 ml; p = 0.001). The absolute and body weight-adjusted maximum blood flow during aortic cross-clamping was not statistically different between the two perfusion systems (absolute maximum blood flow: Aachen MiniHLM 184.33 ml/min ± 45.42 ml/min vs. Dideco Kids 222.0 ± 67.32 ml/min; p = 0.491).
During full-flow conditions, blood gas analyses revealed pO2 values ranging between 90 and 400 mm Hg, pCO2 between 17 and 47 mm Hg, and pH between 7.16 and 7.69 for the MiniHLM. For the Dideco Kids series, pO2 values ranged between 77 and 380 mm Hg, pCO2 ranged between 15 and 54 mm Hg, and pH ranged between 7.03 and 7.7. The necessity of preserving blood was allowed only for sporadic blood samples, leading to this wide range of blood gas parameters. However, these aberrations were easily adjusted through drug administration and/or changes in the gas flow of the HLM.
Hemoglobin values dropped after the initiation of CPB in both groups, but the drop was significantly larger in the Dideco Kids group than that in the MiniHLM group (Figure 3). The use of the MiniHLM resulted in a statistically significant lower decrease in blood fibrinogen than Dideco Kids (Figure 4).
Of further interest was the progression of the anti-inflammatory cytokine IL-10, as well as the depicted ratio of IL-10 to IL-6.5 This ratio was considerably larger in the animals perfused with the MiniHLM than those in the Dideco Kids group (Figure 5). A few rabbits demonstrated very strong reactions to the operations; a wide range can be observed with respect to the 95% confidence interval.
The testing of a newly developed MiniHLM initially in vitro and then in animal trials before its implementation in humans is absolutely necessary to confirm correct functioning of the device, to enable adjustment of laboratory setup, and to ensure the safety of its use in a living organism. The goal is to reduce the rate of complications encountered after its implementation, particularly if it is to be used in newborns. A previous study comparing the MiniHLM with a standardized HLM has already established the MiniHLM in terms of priming volume, fibrinogen consumption, and hemolysis.14
Of further interest is the progression of the inflammatory and anti-inflammatory cytokines. The stipulations for a well-established animal model are complex. It should resemble its target human counterpart regarding anatomical characteristics, including size, weight, and other important properties related to the study’s purpose (in our case, blood volume), as well as the measurability of certain laboratory values. In addition, the financial cost is also a significant factor to consider.
Furthermore, we required of an animal model that it be comparable with regard to operative management. After performing sternotomy, cannulation of the ascending aorta and the right atrium should be performed. However, previous rabbit models describe cannulation by means of the umbilical vein and the carotid artery without sternotomy15 or aortic cannulation performed ex vivo.13 In rat models,10,16 the carotids and the right atrium were likewise cannulated via the jugular vein.
The rabbit is also a good candidate for meeting the demand of comparability of inflammation parameters because of its adequate inflammatory reaction.12 The animals is this study reacted very individually in response to the HLM, which resulted in a wide range of data. Due to the significant variability in the expression of biomarkers from one animal to the other, the number of animals studied is not sufficient to draw final conclusions. With regard to fibrinolysis, there are great similarities between rabbits and humans.11
The testing of the MiniHLM with the use of an animal model was surgically very challenging. During the first series, we had to make some methodical adjustments. For this reason, the results of this series are not comparable.
In the second series with the Chinchilla Bastard rabbits, it was shown that the rabbit is comparable to humans with regard to the operative methodology. Testing of the parameters hemolysis4 and inflammation,5 and of the performance of the MiniHLM3,14 were successfully completed.
The newly developed MiniHLM prototype was successfully tested in comparison with an industrially manufactured HLM in terms of technical function, expression of inflammation, and hemolysis in a rabbit animal model. Therefore, its biological safety was effectively demonstrated in vivo.
It was shown that the Chinchilla Bastard rabbit is very suitable for examining different HLMs with regard to surgical methods, hemolysis, and the expression of inflammatory parameters.
1. Roger VL, Go AS, Lloyd-Jones DM, et al.American Heart Association Statistics Committee, and Stroke Statistics Subcommittee. Heart disease and stroke statistics–2011 update: A report from the American Heart Association. Circulation. 2011;123(4):e18–e209
2. Mortell A, Montedonico S, Puri P. Animal models in pediatric surgery. Pediatr Surg Int. 2006;22:111–128
3. Arens J, Schnoering H, Pfennig M, et al. The Aachen miniHLM–a miniaturized heart-lung machine for neonates with an integrated rotary blood pump. Artif Organs. 2010;34(9):707–713
4. Schnoering H, Arens J, Sachweh JS, et al. The Aachen miniaturized heart-lung machine–first results in a small animal model. Artif Organs. 2009;33(11):935–940
5. Schnoering H, Arens J, Terrada E, et al. A newly developed miniaturized heart-lung machine–expression of inflammation in a small animal model. Artif Organs. 2010;34(11):911–917
6. Carney E, Litwak K, Weiss WAnimal Models Working Group:. . Animal models for pediatric circulatory support device pre-clinical testing: National heart, lung, and blood institute pediatric assist device contractor’s meeting animal models working group. ASAIO J. 2009;55(1):6–9
7. Seghaye MC, Engelhardt W, Grabitz RG, et al. Multiple system organ failure after open heart surgery in infants and children. Thorac Cardiovasc Surg. 1993;41:49–53
8. Dong G-H, Xu B, Wang C-T, et al. A rat model of cardiopulmonary bypass with excellent survival. J Surg Res. 2005;123(2):171–175
9. Jungwirth B, de Lange F. Animal models of cardiopulmonary bypass: Development, applications, and impact. Semin Cardiothorac Vasc Anesth. 2010;14(2):136–140
10. Moehrlen U, Stammberger U, Moehrlen C, Schmid RA. Partial cardiopulmonary bypass in rats using a hollow fibre oxygenator. Interact Cardiovasc Thorac Surg. 2003;2(4):603–606
11. Siller-Matula JM, Plasenzotti R, Spiel A, Quehenberger P, Jilma B:. Interspecies differences in coagulation profile. Thromb Haemost. 2008;100(3):397–404
12. Mabika M, Laburn H. The role of tumour necrosis factor-alpha (TNF-alpha) in fever and the acute phase reaction in rabbits. Pflugers Arch. 1999;438:218–223
13. Yamawaki H, Lehoux S, Berk BC. Chronic physiological shear stress inhibits tumor necrosis factor-induced proinflammatory responses in rabbit aorta perfused ex vivo. Circulation. 2003;108(13):1619–1625
14. Arens J, Schnöring H, Reisch F, Vázquez-Jiménez JF, Schmitz-Rode T, Steinseifer U. Development of a miniaturized heart-lung machine for neonates with congenital heart defect. ASAIO J. 2008;54(5):509–513
15. Skrzypchak AM, Miskulin J, Bartlett RH, Annich GM. Developing a small animal model of extracorporeal circulation. ASAIO J. 2006;52(3):296–302
16. de Lange F, Yoshitani K, Podgoreanu MV, Grocott HP, Mackensen GB. A novel survival model of cardioplegic arrest and cardiopulmonary bypass in rats: A methodology paper. J Cardiothorac Surg. 2008;3:51
small animal model; extracorporeal circulation; rabbit