Single level posterolateral lumbar fusion in a New Zealand White rabbit (Oryctolagus cuniculus) model: Surgical anatomy, operative technique, autograft fusion rates, and perioperative care

Abstract Introduction The posterolateral lumbar fusion (PLF) New Zealand White (NZW) (Oryctolagus cuniculus) rabbit model is a long‐standing surgical technique for the preclinical evaluation of materials for spinal fusion. A detailed understanding of lumbar spine anatomy and perioperative care requirements of rabbits is imperative for correct execution of the model both scientifically and ethically. This study describes the preoperative procedures and surgical techniques used in single level PLF in a NZW rabbit model as it pertains to the animal husbandry, lumbar spine anatomy, anesthesia, surgical approach, and perioperative care of rabbits in a research setting. Materials and Methods We describe the surgical technique and perioperative patient care for single level PLF in a NZW rabbit model. Medical records from a single research facility were retrospectively reviewed for adult NZW rabbits that underwent single level PLF (L4‐L5) between January 2016 and December 2019. The number of lumbar vertebrae per rabbit, fusion rates at 12 weeks using iliac crest autograft and complications are reported. Skeletal maturity was confirmed by preoperative fluoroscopic and radiographic documented closure of hindlimb physes. Results The PLF rabbit surgical model and perioperative patient care is described. PLF was performed in 868 adult female entire NZW rabbits. The majority of rabbits had seven lumbar vertebrae (620/868; 71.4%), followed by six (221/868; 25.5%), and eight (27/868; 3.1%). Fusion rates at 12 weeks for PLF using iliac crest autograft as assessed by manual palpation and radiographic assessment was 76.9% and 70.0%, respectively. Postoperative complications included occasional partial autograft site wound dehiscence due to self‐trauma. Conclusions For PLF rabbit models, a detailed understanding of the surgical technique, rabbit lumbar anatomy including number of lumbar vertebrae, and dietary and husbandry requirements of rabbits, is essential for execution of the model and animal welfare.

respectively. Postoperative complications included occasional partial autograft site wound dehiscence due to self-trauma.
Conclusions: For PLF rabbit models, a detailed understanding of the surgical technique, rabbit lumbar anatomy including number of lumbar vertebrae, and dietary and husbandry requirements of rabbits, is essential for execution of the model and animal welfare.  1 The gold standard for successful spinal fusion is traditional autogenous iliac bone graft, which has a reported fusion rate ranging from 50% to 90%. [2][3][4][5][6] The NZW rabbit PLF model is not stabilized with any posterior fixation and is considered a noninstrumented model that has been used for evaluation of the in vivo performance of bone graft substitutes and biological materials, [2][3][4][5]7 as well as the effect of medications such as opioids, 8 nonsteroidal anti-inflammatories, 9 and antibiotics 10 on spinal fusion. A complete list of the studies using this model is beyond the scope of this manuscript.
Regardless of the purpose of the study, the investigator(s) and surgeon(s) must have a comprehensive understanding of the surgical technique, the anesthetic protocol, and pre and postoperative care requirements of rabbits, specifically diet and husbandry. Failure to address these aspects of the study may lead to complications and associated morbidity and/or mortality which results in increased costs and time requirements of the experiment. 11 Common complications of PLF include incorrect level of fusion, neurological impairment, hemorrhage, surgical site infection and anesthesia-related issues. 11,12 Palumbo et al have described detailed surgical anatomy for PLF in the rabbit. 11 Valdes et al reported a 26% complication rate for PLF in a cohort of 48 rabbits. 13 The authors adapted this knowledge to describe a refined experimental protocol without complication. The authors attributed the reduction in complication rate to a more

| Experimental protocol
Our experimental protocol was implemented in accordance with the ASTM Standard Guide for in vivo evaluation of rabbit lumbar intertransverse process spinal fusion models 14 following institutional ethics approval (Animal Ethics Committee, UNSW, Sydney).

| Physical examination
Prior to study enrolment, all rabbits were physically examined by a veterinarian familiar with rabbit husbandry, anatomy, physiology, and surgery. Routine parameters were assessed including mucous membrane color, capillary refill time, heart rate, respiratory rate, presence of gastrointestinal sounds, and rectal temperature. Close attention was paid to mentation, body condition, coat condition and ambulation. All rabbits experienced a minimum 1-week acclimatization period prior to surgical intervention to allow for adjustment to their environment, including pen mates and attending staff.

| Skeletal maturity
Skeletal maturity was confirmed prior to enrolling an animal for PLF by fluoroscopic and radiographic documentation of distal femoral, proximal tibial (epiphyseal and apophyseal), and proximal fibular physeal closure. A portable fluoroscopy unit (Shanghai Bojin Electric Instrument & Device Co, Shanghai, China) was used for physeal closure screening at the time of physical examination. Immediately prior to surgery, mediolateral radiographs were taken to further document physeal closure. Radiographs were taken using a POSKOM model PXP-60HF portable machine. Digital cassettes (AGFA CR MD4.0 Cassette, AGFA, Germany) were processed by an AGFA CR 75.0 Digitizer and workstation.

| Anesthesia and perioperative care
All rabbits were weighed prior to surgery to ensure correct calculation of all medications to be administered. The rabbits were premedicated with Buprenorphine (0.03 mg/kg) and Midazolam (0.5 mg/kg) via intramuscular injection using a 23 g needle. The rabbits were preoxygenated for 10 minutes prior to masked isoflurane induction and maintained between 2% and 3%, titrated to effect. A balanced crystalloid solution (Hartmann's solution) was administered subcutaneously (20 mL/kg) prior to commencement of surgery. Continuous multiparameter cardiopulmonary monitoring (Datalys V7, Lutech, Ronkonkoma, New York) was used throughout anesthesia and recorded every 10 minutes during the procedure. The isoflurane setting was weaned nearing the completion of the procedure. The rabbits received external heat support in the form of a heat pad. Following completion of surgery, the rabbits were wrapped in a clean towel for added warmth and received masked oxygen supplementation until righting before being returned to their housing pen. All rabbits received perioperative prophylactic antibiotics (Enrofloxacin  The area was aseptically prepared with chlorhexidine-alcohol soaked gauze swabs followed by povidine-iodine spray. The surgical field was then draped in a sterile fashion.

| Bone graft harvest (iliac crest)
A linear skin incision was made along the spine of the iliac crest (left or right) in a caudomedial to craniolateral direction. Subcutaneous and fascial tissue was incised using a combination of sharp and blunt dissection. The gluteal muscles were incised along their fascial origin and elevated from the cranial and caudal aspects of the iliac crest, taking care to stay as close to the ilium as possible to minimize tissue trauma and associated hemorrhage. The authors find it useful to place a gauze swab on either side of the iliac crest using a Freer periosteal elevator 16 to assist in hemostasis and removal of muscle and connective tissue prior to harvesting of the corticocancellous graft. Graft harvest should be confined to the cranial two thirds of the iliac crest as manipulation of the caudal third, close to the sacroiliac junction, may result in damage to the lumbosacral trunk and cranial gluteal neurovasculature.
Bone graft was harvested using bone rongeurs ( Figure 2). Care was taken to ensure that the corticocancellous bone harvested was free of soft tissue as well as excessive blood which can be a source of variation when quantifying the amount. Sharp dissection was used to remove redundant soft tissue and the graft was morselized into small pieces (<5 mm) 3 ( Figure 3). For cases where an alternative graft material was being evaluated, 2 cc (which equates to approximately 1.7 g) of autograft was harvested unilaterally. This quantity is used to ensure a reproducible measurement of autograft as particle shape and packing can influence volume measurements. Autograft was then com-

| Bone marrow aspiration (proximal tibia)
Bone marrow aspirate (BMA), when required for the study, was harvested from the proximal tibia for hydration of select bone graft substitutes prior to implantation. A 25 mm, 18 g trochar needle (Sternobell, Biopsybell, Mirandola, Italy) was manually inserted into the cancellous bone of the proximo-medial tibia ( Figure 4). The insertion site was approximately at the midpoint of the tibial tuberosity cranially and the caudal cortex of the tibia caudally. A release in pressure can be felt once the trochar penetrates the near cortex and reaches the bone marrow cavity. The stylet was removed, a sterile Luer-lock 3 mL syringe attached to the hub of the trochar and negative pressure applied to harvest 2 mL of bone marrow. The trochar was removed and firm manual pressure placed over the site for hemostasis. BMA hydration of the materials for a study was performed immediately following harvest to avoid the use of any antithrombotic agents. A muscular window was then created by carefully placing Weitlaner self-retaining retractors between the two muscle bellies, elevating the longissimus muscle dorsomedially and the iliocostalis muscle laterally to access the middle third of the transverse processes ( Figure 6). It can be useful to pack the surgical site with a gauze swab for hemostasis and removal of additional soft tissue overlying the transverse processes. Using a Freer periosteal elevator, the transverse processes and intertransverse ligament were exposed by manually elevating remaining muscle fibers, taking care to avoid the segmental artery which courses laterally from the cranial aspect of the transverse process. Decortication was performed using a high-speed pneumatic surgical drill (Midas Rex, Medtronic, Memphis, Tennessee) with an M-8 Matchstick Burr. The central portion of the transverse processes were carefully decorticated for a distance of 10 mm from the vertebral body and pars to a level where bleeding bone beds were visually F I G U R E 3 Two cc (~1.7 g) of corticocancellous bone graft harvested from the iliac crest F I G U R E 4 Bone marrow aspiration from the proximal tibia using a 25 mm, 18 g trochar needle (Sternobell, Biopsybell, Mirandola, Italy)) F I G U R E 2 Corticocancellous bone graft harvest from the iliac wing using bone rongeurs present to the surgeon 3 (Figures 7 and 8). Great care was taken to decorticate the bone, creating a bleeding bone bed in preparation for placement of graft material over the transverse processes. Given the oblique cranioventral angulation of the transverse processes, ambidexterity is advantageous for decortication. The same procedure was performed on the contralateral side. To facilitate accurate bilateral approach, the location of the L4 and L5 transverse processes were marked with sterile needles through the longissimus muscle to ensure the correct level was treated on both sides.

| Material placement
Material (bone graft +/− synthetic material of 2 cc per side) for implantation was placed into 3 cc syringes with the tip removed to provide a simple means to accurately place the materials in the prepared posterolateral beds. The material was implanted bilaterally on the decorticated transverse processes and overlying the intertransverse ligament between the L4 and L5 transverse processes ( Figure 9). Care was taken to ensure the material was placed midline and directly on the decorticated transverse processes.

| Wound closure
The muscle planes were closed using 3-0 monofilament absorbable suture in a simple continuous pattern. Care was taken to take deep bites of the muscle to ensure closing of dead space to prevent graft migration and seroma formation. The superficial fascia of the longissimus and multifidus muscles was included in this layer to assist in closing the dead space. The subcutaneous layer was closed in a simple continuous buried pattern with 4-0 monofilament absorbable material. The skin was closed with 4-0 monofilament absorbable material in either a buried simple continuous or Ford-interlocking pattern. A transparent adherent film spray (Op-site, Smith & Nephew) was applied to all surgical wounds to encourage formation of a fibrin seal and discourage self-trauma.    All rabbits were fed a high-quality diet, of which the majority was high-fiber timothy or oaten hay, supplemented with leafy greens such as silver beet, spinach, bok choy, pak choy, and broccoli. High-quality commercial pellets, herbs (dill, parsley, coriander, mint, etc) and fruits (strawberries, apple, etc) were provided in moderation. Water was provided ad libitum via self-automated bottles. Calm, slow-paced music was played during daylight hours.

| Radiographic analysis
All animals were confirmed to be skeletally mature prior to surgery as verified by closure of the distal femoral, proximal tibial and fibular physes. Representative radiographs are presented in Figure 11. Accurate implant placement at the L4-L5 level was confirmed in all cases based on postoperative radiographs ( Figure 12). Reporting of surgical times for each synthetic material is beyond the scope of this manuscript. The traditional rabbit PLF spinal fusion model utilizes the Wiltse approach that is, the intermuscular plane between the multifidus and longissimus muscles, which facilitates direct access to the pars, transverse processes, and facet joints of the lumbar spine. 21 The surgical approach we describe is via the intermuscular plane between the longissimus and iliocostalis muscles; lateral and ventral relative to the Wiltse approach, as described by Zunariah et al. 19 We have found that Confirming skeletal maturity prior to enrolling an animal in a study is important for study integrity. 14 The literature remains inconsistent regarding skeletal maturity in rabbits and methods to document closure of hindlimb physes. [22][23][24][25] Variations in breeder stock and genetics make age and body weight imprecise parameters for determining skeletal maturity in NZW rabbits. Skeletal maturity must be confirmed by preoperative fluoroscopic or radiographic assessment to avoid enrolment of immature animals which could artificially skew the results of fusion given the potential increased healing capacity of younger animals. Figure 11 demonstrates the radiographic appearance of distal femoral, proximal tibial (apophyseal and epiphyseal) and proximal fibular physes of NZW rabbits at 3, 6, and 12 months of age.

| Complications
Rabbits can present an anesthetic challenge to researchers and clinicians, 26,27 with a mortality risk 14 times higher than dogs. 28 Being a prey species, rabbits are easily stressed which can have significant adverse cardiopulmonary effects. Adequate acclimatization of their surrounds, animal care staff and researchers interacting with the animals should be followed to reduce stress on the animal prior to surgery. Buprenorphine and midazolam premedication combination provides good sedation prior to induction which facilitates stress-free handling of the rabbit. Appropriate premedication significantly reduces the amount of isoflurane required for general anesthesia. We premedicate at least 45 minutes prior to commencing surgery given the prolonged onset of action of buprenorphine in rabbits; ensuring the rabbits have effective analgesia at the time of surgical intervention. As a benzodiazepine, midazolam provides sedation and muscle relaxation prior to induction. This drug combination reduces the isoflurane requirements for general anesthesia to help reduce the cardiovascular side effects of isoflurane. Additionally, the combined sedative effects of both drugs facilitate stress-free handling of the rabbit for isoflurane induction. Preoxygenation prior to induction creates an oxygen reservoir that can delay the onset of desaturation which may occur due to isoflurane induction induced apnoea. 29 Masked isoflurane maintenance has not been associated with adverse events in our experience; however, endotracheal intubation may be of use in establishing and maintaining an airway.
The vasodilative effects of inhalation anesthetics such as isoflurane in combination with their large surface area, make rabbits extremely susceptible to hypothermia. 29 Hypothermia may result in bradycardia, increased intraoperative hemorrhage, prolonged anesthetic recovery, and postoperative surgical site infection. 29 We reduce hypothermia as much as possible in our anesthetic protocol by using external heat support (heat mat, towel wrapping, etc) throughout the perioperative period and titrating the isoflurane setting to effect.
Additional heat support methods that we could employ include hot water bottles, forced warm-air blowers, and enclosed humidicribs/ incubators.
Rabbits are monogastric herbivores that utilize hindgut fermentation for digestion. Rabbits require a high-feed intake diet, rich in insoluble fiber to promote gastrointestinal motility and maintain their aradicular hypsodont dentition (high crowned, continuously growing teeth. 30  Our rabbit housing facilities were developed in accordance with the guidelines of the ARRP. 18 Housing pens are floored with good quality hay, which has been shown to reduce the incidence of pododermatitis and trichobezoars in rabbits. 31 We seldom see respiratory disease in our rabbit cohort, which may be due to the superior ventilation of pens compared to cages, in conjunction with regular daily cleaning and appropriate screening of symptomatic rabbits. 32 Laboratory rabbits prefer to be housed with littermates, spending approximately 79% of their time in each other's company. 31 We maintain 3 to 4 rabbits per pen, each with their own shelter to help minimize conflict. When noted, offending rabbits may be housed singularly with clear vision of other rabbits and/or rotated between pens. In our experience, appropriate stocking densities, environmental enrichment and strictly using female rabbits, reduces conflict between rabbits and encourages the rabbits to display their innate behaviors.
Pasteurella multocida, a virulent and readily transmitted gramnegative coccobacillus, may manifest as multiple clinical syndromes in rabbits including dacrocystitis, rhinitis, pneumonia, otitis media, pyometra, orchitis, abscesses, and septicaemia. 33,34 Most adult rabbits are believed to be infected with P. multocida, 35 although infection may be asymptomatic. 36 P. multocida has variable and inconsistent antimicrobial sensitivities. Enrofloxacin is a relatively safe, concentrationdependent fluoroquinolone with activity against P. multocida in rabbits 37 ; however, multiple studies have questioned its effectiveness. 35,38 Anecdotally, despite prophylactic Enrofloxacin administration, we have observed occasional Pasteurella-associated abscessation in rabbit PLF studies. For these reasons, we recently updated our perioperative antibiotic choice to Procaine penicillin, which has proven efficacy against P. multocida. 39 Regardless of the antibiotic choice, strict environmental control, good husbandry conditions and adherence to principles of aseptic surgical technique, are in our opinion, the best control measures of P. multocida-associated disease in experimental rabbit cohorts.

| CONCLUSION
We describe an updated protocol on the long-standing PLF model of spinal fusion from the preoperative to postoperative setting, the variation in the number of lumbar vertebrae in rabbits and 12 week autograft fusion rates. Our techniques accommodate for the unique anesthetic, dietary and husbandry requirements of rabbits, supported by meticulous surgical technique to minimize surgical complications to achieve optimal experimental outcomes. We hope this information is useful for other researchers using this model with the aim of animal welfare preservation and the principles of reduction, refinement and replacement.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of this article.