Methods

References for this review were identified by searches of PubMed database. Search terms were “rabbit” and “subarachnoid hemorrhage.” Only papers published in English between 1969 and 2013 were included.

Historical Background and Evolution

In 1969, Offerhaus and van Gool reported the first rabbit subarachnoid hemorrhage (SAH) model using a blood shunt method to investigate cardiac dysfunction after SAH [1]. In 1977, Svendgaard et al. reported an increased reactivity of the isolated basilar artery in the rabbit SAH model using a single blood injection method into the cisterna magna [2]. As far as we know, this is the first report of a rabbit SAH model using the cisterna magna blood injection method. In 1982, Edvinsson et al. first reported rabbit cerebral vasospasms after SAH using a single blood injection method into the chiasmatic cistern [3]. Since the early 1980s, the rabbit SAH model using a single blood injection method into cisterna magna has been used for the research of SAH or cerebral vasospasms (CVS) [4–7]. Since the late 1980s, the double blood injection method, which induces a more severe and prolonged vasospasm than the single blood injection method, and which is well established in rat and dog models [8], has been used for investigating CVS in rabbits [9, 10]. However, the double-injection method is not popular in rabbits, possibly because the enhancement of CVS is ineffective and carries a high mortality rate [11, 12]. Recently, Zhou et al. reported a lower mortality rate and significantly more pronounced CVS in the double blood injection method in comparison to the single blood injection method in the rabbit SAH model [13].

Rabbits Employed

The age of rabbit used for experimental SAH ranges from 80 days to 2 years, although few reports actually describe the age with any precision [14–16]. Nakajima et al. investigated the effect of aging on CVS in the rabbit single-injection model by comparing the time course of CVS and the vasodilating effect of papaverine among three groups: young (2–3 months), adult (6–9 months), and old (20–40 months) [17]. With advancing age, the degree of CVS was augmented, while both the resolution of CVS after maximal constriction and the vasodilating effect of papaverine were impaired [17]. The weight of the rabbit used for the experimental SAH model ranges from 1.5 to 5 kg (depending on the age of rabbits used) [18–20], although most of the rabbits weigh about 2–4 kg (Table 1). As for the strain, the most popular rabbits for the SAH model are the New Zealand white rabbit and the Japanese white rabbit. Fauve de Bourgogne rabbits [21–23] and Burgundy rabbits [24, 25] have rarely been used for the rabbit SAH model. Male rabbits are often preferred to female rabbits [10, 26–28, 15] because of their general pathogenic susceptibility compared with the biologically resilient female [29]. However, the rationale for choosing the gender is not clearly described in the literature.

Table 1 Physiological parameters of rabbit
Full size table

Anesthesia, Analgesia, Perioperative Care, and Monitoring

Ketamine is the most preferred drug for anesthesia (20 mg/kg [30, 31, 15], 25 mg/kg [32, 19, 33, 34, 13, 35], 30 mg/kg [20, 36–42], 35 mg/kg [43–47], 40 mg/kg [48–57], 50 mg/kg [58, 59, 14, 60, 61, 10, 62–67, 18, 68–71], 55 mg/kg [72, 73], 70 mg/kg [74], intramuscularly (i.m.); 3 mg/kg [75, 76], intravenously (i.v.)). Ketamine is used alone or in combination with xylazine (2.5 mg/kg [77], 5 mg/kg [43, 44, 30, 45, 74, 46, 31, 39, 41], 6 mg/kg [49, 52, 36–38, 40, 55, 42], 8 mg/kg [50, 51, 62, 53, 63, 56, 57], 9 mg/kg [72, 73], 10 mg/kg [58, 59, 61, 10, 66, 67, 18, 47, 69, 70], 15 mg/kg [19], i.m.), droperidol (1.0 mg/kg [32–34, 78, 13, 35], i.m.) or pentobarbital (20 mg/kg [65, 54, 68], 30 mg/kg [15], i.v.). Intravenous injection of pentobarbital is often used for anesthesia (20 mg/kg [4, 21, 12], 25 mg/kg [79–81], 30 mg/kg [82–84], 45 mg/kg [85, 5]). Acepromazine [21, 22, 31, 23], diazepam [75, 76], ethomidate [22], or alcuronium [24] are occasionally used in combination with other anesthesia. In a few cases, urethane has been used intraperitoneally [86] or intravenously [16, 87]. In intubated and ventilated cases, inhalation of isoflurane or halothane is often used for maintenance anesthesia [9, 88–91, 24]. During anesthesia, rabbits are often endotracheally intubated and ventilated, although spontaneous respiration without intubation has been achieved in many studies. In some intubated cases, rabbits were mechanically ventilated if necessary until spontaneous respiration resumed [50, 57, 42]. Tracheostomy was performed in a few cases [92, 93, 86]. Oxygen was supplied in some cases [79, 19, 24, 55]. In experiments requiring perfusion fixation, rabbits were often ventilated during anesthesia. The depth of anesthesia is usually evaluated by the presence of body movement during surgery, such as that following pain stimulation by a periodical toe-pinch [37]. Postoperative pain relief is managed by subcutaneous administration (0.1–0.2 mg/kg) or intramuscular injection (0.04 mg/kg) of buprenorphine [20, 77, 37, 38].

After blood injection, the intracranial pressure (ICP) and the blood pressure transiently increase before gradually recovering. Using the ICP-controlled extra-intracranial blood shunt model, Marbacher et al. reported that within 1–2 min after SAH induction, the ICP rose to a peak (about 8-fold higher than the baseline value), and within 5–10 min, the ICP returned to a steady state that was significantly higher than the baseline [94]. In the single blood injection model, a 2-fold increase of ICP was seen after SAH, even following 1 mL of autologous blood injection [25]. Marbacher et al. also reported that the mean arterial blood pressure (MABP) increased steadily from 70 to 90 mmHg until reaching the ICP peak and then slowly decreased toward the baseline under subcutaneous injection of ketamine (30 mg/kg), xylazine (6 mg/kg), and continuous intravenous anesthesia [94]. Further, in the single blood injection model, 4 mL of autologous blood produced a 1.3-fold increase of the MABP (from 89 to 115 mmHg) within the first minute following SAH, which returned to the baseline 15 min after SAH under anesthesia (intramuscular injection of ketamine (50 mg/kg) and xylazine (5 mg/kg)) [67]. In the same model, 1 mL of autologous blood has been shown to induce a 1.3-fold increase of MABP (from 65 to 85 mmHg) after SAH while under intravenous anesthesia using urethane (8 mg/kg) [87]. These results suggest that the amount of clot influences ICP, but not MABP.

The mortality of the single blood injection model is comparatively low (0–8 %) [61, 43, 36, 37, 46, 84, 13], with very few exceptions [53, 90] (Table 2). In the double blood injection model, the mortality is comparatively high (5.56 % [13], 16.7 % [78], 20 % [12], 25.8 % [9], and 30 % [95]) (Table 2). In the triple blood injection model, Black et al. reported a mortality rate of 46 % [85]. Rabbits often die owing to acute respiratory distress after the subarachnoidal injection of blood [12], angiographic procedure [12], or brain stem injury during cisternal injection [85]. Above all, respiratory arrest after blood injection is the most important cause of the death after SAH. Zhou et al. reported that respiratory arrest occurred immediately or within 5 min of blood injection with intervals between 8 s to 1 min in 13 % of rabbits in the single blood injection model and 22 % of rabbits in the double blood injection model [13]. Spallone et al. reported that apnea was observed in 50 % of rabbits in the double-injection model with intervals between 8 and 30 s [12]. In the double blood injection model, death tends to occur immediately after the second injection of blood [78, 13]. Therefore, the volume of blood in the second injection has often been reduced to about 60 % of the first injection [10, 52, 30, 64, 31, 39, 16].

Table 2 Pathophysiological parameters of rabbit to be expected
Full size table

Regardless of the presence or absence of respiratory support, PaCO2 and PaO2 are generally maintained around 35–45 and 90–150 mmHg, respectively [49, 60, 10, 36, 74, 67, 96, 90, 24, 56, 25, 87] (Table 1). In spontaneous respiration, body temperature and pH are maintained between 38 and 40 °C and between 7.3 and 7.5 [49, 60, 36, 74, 67, 96, 90, 38, 24, 56, 25], respectively (Table 1). There is great variability in reported heart rates: 140–300 beats per minute (bpm) [74, 67, 37, 24, 25].

The neurobehavior of the rabbit following SAH has been evaluated using grading systems by Endo et al. [10, 53, 92, 70], Strong et al. [90, 76, 16, 97], and Zhou et al. [13] (Table 2). A few reports have performed neurobehavioral scoring of rabbits using a system commonly employed for scoring dogs [61, 98, 99, 35]. Using the grading system by Strong et al., Song et al. reported that the neurobehavioral score gradually deteriorated after the second blood injection in the double blood injection model, which peaked at day 5 postinjection [16]. After the first blood injection, rabbits presented with an apathetic mood, reduced locomotor activity and sleepiness, decreased feeding and drinking, and occasional neck stiffness [95]. These signs were significantly aggravated after the second blood injection [95]. Spallone et al. and Tang et al. reported that hemiparesis was observed in 5 and 40 % of SAH rabbits with the double blood injection model, respectively [12, 70].

SAH Induction

The most commonly used site for blood injection is the cisterna magna. The lateral prepontine cistern has rarely been used as a blood injection site (using a silastic tube) [100, 71]. Typically in the prone position, cerebrospinal fluid (CSF) is aspirated percutaneously or through a surgically opened wound from the cisterna magna using a needle, and autologous arterial blood is injected into the cistern magna. For head fixation, a stereotactic head frame has been used in some cases [85, 63, 100, 5, 12, 42]. The occipital protuberance is commonly used as a palpable landmark to identify the injection site [90]. Autologous blood from the central ear artery (auricular artery) was used in most cases, while that from the femoral artery was used in some cases [101, 27, 21, 44, 102, 66, 103, 15, 84, 104]. Rarely, venous autologous blood was used for this model [91, 87]. Autologous blood is rarely heparinized before injection [4, 23]. Usually, the blood injection is performed manually without using any pumping device. The injection pressure depends on the time period of the blood injection. The most popular blood volumes have been 1.0 mL/kg in a weight-adapted volume injection and 1.0, 1.5, 2.0, 2.5, and 3.0 mL in a fixed volume injection, while a wide range of weight-adapted (0.3 mL/kg [95], 0.5 mL/kg [82, 91], 0.9 mL/kg [19, 92], 1.25 mL/kg [4], 1.5 mL/kg [90, 76]) and fixed (0.5 mL [101, 86], 4.0 mL [49, 36, 67], and 5.0 mL [48, 50, 62, 53, 36, 40, 55, 56, 69]) volumes of blood have been used in the cisterna magna blood injection model. As discussed above, the blood volume in the second injection was often reduced compared with the first injection volume (2.5 mL + 1.5 mL [10], 5 mL + 3 mL [30, 31, 39], 1.5 mL/kg + 1.0 mL/kg [16], 0.8 mL/kg + 0.5 mL/kg [64]). While the most popular injection devices are 23- and 25-gauge butterfly needles, others include 16-gauge [49, 60], 21-gauge [9, 85, 66], 22-gauge [101, 24], 24-gauge [68], and 27-gauge needles [79, 19, 44, 12, 83]. Lumber puncture needles [34], silicone tubes [26, 100, 71], polyethylene catheters [25], trocar needles [16] are among the other devices used for injection. Blood injection is performed over a period of time, which is commonly within several minutes, to avoid respiratory arrest (for 10 s [27, 70], 20 s [60, 34, 12], 30–45 s [77, 23], 30–90 s [22], 1 min [22], 2 min [74, 47], 2–3 min [21, 87], 3–5 min [90, 76], and 4 h [71]; over 10 s [51, 40, 56, 57], 20 s [4, 91], 20–30 s [62], 30 s [36], 1 min [101, 61, 82, 15, 78], 2 min [59, 64, 24, 42], and 30–45 min [77]). After blood injection, the rabbit is kept in a prone position with the head tilted down (30° [79, 9, 59, 61, 33, 77, 82, 36, 54, 67, 90, 34, 95, 76, 15, 78, 83, 104, 13, 42], 45° [102, 88, 26, 25, 41], 65° [20, 37]) (Fig. 1) for some time (3 min [44], 5 min [21, 22, 102, 12], 10 min [85, 19, 20, 95, 37, 91, 92, 23], 15 min [79, 48, 58, 27, 4, 43, 52, 45, 82, 72, 73, 55, 56, 83, 41, 42], 20 min [50, 51, 60, 66, 70], 30 min [59, 32, 77, 65, 88, 74, 54, 89, 18, 15, 68, 16, 13, 35], 40 min [84], 45 min [46], 60 min [105, 104]). To spread the blood across the cranium, it is important to place the head in a downward position. In the double blood injection model, the second blood injection is usually performed following a 24- or 48-h interval [82, 12].

Fig. 1
figure 1

Prone position with the head in downward. Tilting the head 30° down on an adjustable bed with the feet of the rabbit strapped to the bed frame using silicone tubes

Full size image

Technical Considerations

For accurate puncture, the neck should be flexed as much as possible to maximize the craniocervical junction. Furthermore, touching the inion and C2 spinous process is important for the accurate puncturing of cisterna magna (Fig. 2a). For improved visualization, the posterior neck of the rabbit should be sprayed with alcohol (Fig. 2b), or the hair of the posterior neck should be trimmed. To avoid puncture failure, the following factors may be important: (1) having an appropriate posture for injection, (2) not evacuating too much CSF, and (3) directing the needle slightly rostrally and puncturing as the tip of a needle runs through under the foramen magnum. During manual blood injection, the size of the syringe is important, because the resistance of the ICP can be felt through the appropriate-sized syringe, such as 2.5 cc. After the injection of blood, the respiratory condition of the rabbits should be observed for around 10 min, because respiratory arrest mostly occurs in the first few minutes after blood injection. Accidental movement of the needle tip induced by body movement during puncture causes the erroneous injection of blood into a space other than the subarachnoid space. To avoid body movement, appropriate control of the depth of anesthesia is important. Immediate withdrawal of the needle after injection causes leakage of injected blood with CSF from a dural pinhole.

Fig. 2
figure 2

Improvement of visibility of rabbit neck and identifying the landmark for puncturing. a Touching the inion and spinous process of C2 for identifying the puncturing point. b For improved visualization, the posterior neck of the rabbit was sprayed with alcohol

Full size image

The most common method of sacrificing rabbits is by perfusion fixation. In some cases, rabbits have been sacrificed by bolus injection of sodium thiopenthal (40 mg/kg [37], 20 mg/kg [19]), intravenous injection of potassium chloride [36, 46], intraperitoneal injection of sodium pentobarbital (200 mg/kg) [18], intracardiac injection of 15 mL of alcohol [44], exsanguination under intravenous injection of sodium pentobarbital (120 mg/kg [54], 60 mg/kg [89], 25 mg/kg [81]) and decapitation [21]. Perfusion fixation is commenced with 200–500 mL of flushing solution (physiological saline solution [43, 45, 36, 66, 74, 31, 15, 47, 83] which is occasionally heparinized [15, 83], Hank’s balanced salt solution [50, 58, 59, 67, 72, 39, 56, 41, 57], or physiological phosphate buffer solution [78, 13]) under an appropriate perfusion pressure (100 cm H2O [58, 59, 74, 39], 120 cm H2O [60, 32, 43, 33, 45, 78, 47, 55, 57, 13, 35], 75 mmHg [50, 56, 83, 41], 100 mmHg [85]) followed by 200–1000 mL of fixative solution (3 % [49, 60, 47] or 4 % [82, 66, 18, 15, 68, 83] paraformaldehyde, 2.5 % glutaraldehyde [85, 5], 10 % formaldehyde [33, 74, 28, 34, 78, 13, 35, 42], or a mixture of paraformaldehyde and glutaraldehyde (2 % + 2.5 % [59, 36, 67, 72, 73, 31], 2 % + 2 % [58], 2 % + 1.5 % [39], 1 % + 1.5 % [50, 41]).

Discussion

Table 3 summarizes the advantages and disadvantages/limitations of the rabbit cisterna magna blood injection SAH model. Compared with large animals such as pigs, dogs, and monkeys, the size of rabbit can be easily handled in addition to its inexpensiveness and docile nature. The size of rabbit is large enough to perform angiography unlike small-size animals such as rat or mouse. Therefore, rabbit SAH model allows us to evaluate the chronological change of the arterial diameter by repetitive angiography in the same animal. Furthermore, the size of rabbit basilar artery is easy to handle, and the amount of the tissue of basilar artery is sufficient for performing gene or protein expression analysis. For these reasons, rabbit SAH model seems to be appropriate for the study of delayed CVS. In addition, rabbit has safe, less invasive, and highly reliable vascular access in the ear, central auricular artery, and marginal auricular vein. This is one of the most important advantages in surgical procedure of SAH induction. On the other hand, using rabbit for SAH model seems to have limitation in the research using molecular biological approach because of the less availability of useful antibodies or genetically modified animals in rabbit.

Table 3 Advantage and disadvantage of using rabbit for SAH model
Full size table

Currently, there is no evidence confirming the appropriateness of rabbit cisterna magna blood injection SAH model in the study of early brain injury. In volume-controlled blood injection method, ICP elevation seems to be insufficient. Rapid increase in ICP followed by decreased cerebral perfusion pressure (CPP) is one of the most important factor causing early brain injury. Therefore, pressure-controlled blood injection method, such as endovascular perforation, appears to be more appropriate than volume-controlled blood injection method for the study of early brain injury. Recently, Marbacher et al. demonstrated that the rabbit blood shunt model can be used for the study of early brain injury after SAH. In their rabbit ICP-controlled blood shunt model, the bleeding provoked rapid ICP increase, causing CPP decrease to almost zero, and consistent early damage to the hippocampus, basal cortex, and cerebral vasculature [94]. Therefore, the rabbit ICP-controlled blood shunt model might be better than the cisterna blood injection model to study early brain injury. However, there is insufficient accumulation of research knowledge regarding early brain injury using rabbit single or double blood injection model. Further study using rabbit blood injection method is needed to evaluate the suitability of this model in the study of early brain injury.

Conclusion

In this review, detailed features of animals used, detailed methods and settings of SAH induction and various physiological and pathological parameters of rabbit cisterna magna single and double blood injection model were demonstrated, and technical issues of making models as well as the advantages, disadvantages, and limitations of using this model were discussed. This review might deepen the understanding of rabbit cisterna magna blood injection model and be helpful for choosing the appropriate animal model in accordance with each investigation for various pathological conditions after SAH.