Article Text

Download PDFPDF

Original research
Focused post mortem dissection technique for harvest of rete mirabile in domestic swine (Sus scrofa)
  1. Javed Khader Eliyas1,
  2. Marek Niekrasz2,
  3. Craig Wardrip2,
  4. Seon-Kyu Lee3
  1. 1Section of Neurosurgery, University of Chicago, Chicago, Illinois, USA
  2. 2Animal Resources Center, University of Chicago, Chicago, Illinois, USA
  3. 3Section of Neurosurgery, Department of Surgery and Radiology, University of Chicago, Chicago, Illinois, USA
  1. Correspondence to Dr Seon-Kyu Lee, Neurointerventional Radiology, Department of Radiology, 5841 S Maryland Avenue, MC 2026, Chicago, Illinois 60637, USA; sklee{at}


Background Rete mirabile (RM) of the domestic pig is a popular animal model of arteriovenous malformations. The RM (Latin for ‘wonderful net)’ comprises the arterioarterial portal connecting ascending pharyngeal arteries and the internal carotid arteries, which exists in the skull base of even-toed ungulates. Although angiographic access of the RM is relatively easy, its post mortem procurement is complicated and its detailed technique has not been well described.

Objective To present our focused post mortem dissection technique for undamaged and complete harvest of the RM.

Materials and methods Fourteen domestic (40–70 lb (18–32 kg)) swine were used in this study. Angiographies were performed under general anesthesia in all animals. A 5F Berenstein catheter was used for angiography and a 014 microcatheter was used to obtain superselective angiography. A stepwise surgical dissection technique has been developed to efficiently harvest RM. Angiographic and surgical anatomy were also compared.

Results The RM was supplied by bilateral ascending pharyngeal arteries. Bilateral anterior cerebral arteries, middle cerebral arteries, and the basilar system were identified rostral to the RM. Our surgical dissection technique was developed during a project to streamline harvesting of the RM and a stepwise description is as follows: (1) decapitate the swine by removing the head through the plane of the occiput and C1 vertebral body; (2) remove the tongue and oropharynx via a ventral approach; (3) dissect through the posterior pharyngeal wall identifying bilateral tympanic bullae and the basisphenoid bone; and (4) remove the basisphenoid bone about one and half inches above the rostral end of the tympanic bullae to fully expose the rete.

Conclusions The RM can be procured efficiently and effectively with our technique, without requiring any sophisticated surgical devices.

  • Arteriovenous Malformation
  • Dissection
  • Technique
View Full Text

Statistics from


The rete mirabile (RM) of the domestic pig has been used as an animal model of arteriovenous shunting in many studies.1 ,2 Being a high-volume arterioarterial portal existing in the epidural skull base of even-toed ungulates, it is a suitable, naturally occurring animal model for arteriovenous malformations (AVMs).3 Some studies have further increased flow through the RM by creating a sump effect with the unilateral common carotid artery (CCA) and internal jugular vein anastomosis.4 Preclinical experimental investigations for various therapeutic procedures such as developing a new embolic material, or evaluation of the therapeutic effects of a new treatment method such as focused radiation, require both histopathological effects and catheter angiographic findings to prove its efficacy. Therefore, undamaged and complete harvest of swine RM is one of most critical components of these experimental investigations. Swine RM can be relatively easily identified on catheter angiography; however, owing to its deep intracranial location, undamaged and complete harvest of swine RM can be challenging and time consuming. To our knowledge, there has been no report describing a detailed post mortem dissection technique for the harvest of swine RM. We present a focused, relatively simple, and time-efficient post mortem dissection technique for undamaged and complete harvest of swine RM that can be used for research.

Materials and methods

Fourteen domestic swine, aged 3–6 months and weighing between 40 and 70 lb (18–30 kg), were included in this study. All animals were subject to an Institutional Animal Care and Use Committee approved transarterial catheter embolization of a new liquid embolic agent using a swine RM AVM model. All procedures were performed under general anesthesia with endotracheal intubation.

For angiography, femoral access was obtained through open surgical cut-down to insert a 5F short sheath (25 cm). The initial angiogram was obtained with a 5F Berenstein catheter, and embolization was performed with the coaxial use of a 014 microcatheter. We did not routinely heparinize the animals and flushing of the catheters was performed manually and periodically. Pre- and postembolization catheter angiograms of bilateral carotid arteries and ascending pharyngeal arteries were obtained on the day of embolization. Based on a scheduled follow-up period (1, 2, and 4 weeks after embolization), follow-up catheter angiograms of these arteries were also obtained. Subsequently, the animals were euthanized.

The euthanized animals were methodically dissected to obtain the rete, which forms the basis of our post mortem dissection technique. For the first three animals, a bone saw (Stryker, Kalamazu, Michigan, USA) was used for decapitation at the level of the upper cervical spine (C2–C3 level), and the ascending dissection was performed after the CCA and subsequently, ascending pharyngeal artery, as described previously.5 With the fourth animal, we developed our focused, time-efficient post mortem dissection technique and applied it to all the remaining animals. The approximate time from skin incision until harvest of the RM was measured. Sequential steps were photographed during the dissection and compared with catheter angiograms obtained ante mortem.


The RM was successfully obtained from all 14 animals. The average time needed for harvesting for the first three animals using the previously described ascending dissection technique was about 2 h, and for the animals in which we used our newly described dissection technique was about 25 to 30 min. The expected location of the rete was found accurately in all cases and was found within the epidural space and posterior to the sella, supplied primarily by bilateral ascending pharyngeal arteries. For RM procurement by our technique, no microscope or pneumatic drill was required and the procedure did not entail use of an electric saw to decapitate the animal or mandibular advancement to reach the skull-base region. Also, this being a transclival approach, our technique did not involve transcranial removal of the brain to harvest the RM.

The RM measured an average of 2.5×1.5 cm with a definite midline connection between the sides, ventral (inferior) to the sella. The efferent pathway of the RM was an internal carotid artery that subsequently divided into cranial and caudal divisions on each side, as seen on the angiograms. The retial vessels became intradural by entering the cavernous sinus while the origins of a single artery are often in the distal cavernous sinus. Anastomotic branches from the external carotid and internal maxillary artery (ramus anastomoticus and arteria anastomotica) were too small to be identified during our dissection (figure 1).

Figure 1

Focused dissection of the RM, shown through sequential steps. (A) The post-auricular incision (shown in black) is started behind both ears and brought down to the anterior neck, hugging the lower margin of the mandible. By meeting its counterpart from the other side, a complete, circular incision is made. (B) and (C) The decapitated head is disconnected from the neck at the atlanto-occipital joint, presenting the full face of the occipital condyle (thin black arrows). The severed spinal cord can be seen posteriorly (thin white arrow). The styloid processes are marked on the magnified picture (C) with asterisks. The RPST in the unmagnified view (B) has been cleared over the styloid process in the magnified view and demonstrates clivus. (D) The remaining both occipital condyles (black arrows) and the clivus are removed with rongeurs. (E) Further cranial removal of clivus (small white arrows) is limited by bilateral AB. (F) Once they are removed, the distal part of the clivus is now taken out to provide a glimpse of the inferior part of the RM. (G) Harvesting of the rete requires circumferential dissection, especially the ventral aspect of the sella turcica (white arrow heads). (H) Another specimen in which we have injected Onyx 18 through the right ascending pharyngeal artery and RM shows the relationship among bilateral ascending pharyngeal arteries, RM, and AB (partially resected). AB, auditory bullae; AscPA, ascending pharyngeal artery; RM, rete mirabile; RPST, retropharyngeal soft tissue.

Sequential harvesting steps of RM in the domestic swine are summarized in the following five steps:

  1. Decapitation of the pig/removal of the head (figure 2A,B).

    A curvilinear post-auricular incision is made, which is carried around the anterior neck to complete a circle. Dissection is continued sharply with a scalpel until the craniovertebral junction is reached, initially from the posterior neck. Then, ligamentous connections of the horizontally oriented atlanto-occipital joints are severed, to disconnect the cranium from the rest of the body. At this time the head remains attached only by prevertebral, retromandibular tissue and anterior neck soft tissue. Subsequent division of soft tissue and strap muscles completes the decapitation.

  2. Retropharyngeal and preclival soft tissue dissection (figure 2B,C).

    Attention is then directed to the anterior and cranial aspect of the craniovertebral junction. The retropharyngeal soft tissue is swept off the skull base on either side, again using sharp dissection. Skeletonization of the skull base and further removal of bilateral styloid processes is performed to obtain sufficient space. Subsequently, dissected retropharyngeal tissue is shifted laterally to the styloid process on either side. The dissection plane is then advanced anteriorly, lateral to the styloid processes, to release lingual attachments on either side. This allows swinging of the tongue out of the mouth with better visualization of the central skull base.

  3. Occipital condylectomy and clivectomy (figure 2D and E).

    The next step involves bilateral occipital condylectomy, with care taken to preserve medially present dura. The clivus is freed from the dura with a Penfield-like instrument and rongeured until auditory bullae obstruct further advancement. Occasionally, the lower level of the rete is visible at the level of the auditory bullae.

  4. Auditory bullectomy (mastoidectomy) (figure 2F,G).

    Further exposure of the rete requires removing the auditory bullae on either side, which is easily achieved by biting with a rongeur-like instrument. This provides more room to circumvent the RM, especially the distal part that enters the cavernous sinus. Often we remove only the medial part of the bullae that is adjacent to the rete, without the need to perform a complete bullectomy.

  5. Dissection of the rete free from dura and sella (figure 2H).

    The rete now is clearly seen in the epidural space and careful separation from the dura is carried out. In the midline a small bony element is found dorsal to the rete and corresponds to the sella tursica. Once a good separation plane is obtained all around the rete, it is removed with a short segment of the ascending pharyngeal artery on one side, to help label the sides correctly. The rete is completely free once the distal tributary (internal cerebral artery) arising from it is sectioned on either side.

Figure 2

Angiographic anatomy of the rete mirabile (RM) and intracranial circulation of the domestic pig. (A) The left ascending pharyngeal injection shows the entire rete with intracranial circulation. Note the faintly visualized right internal maxillary artery. (B) The unsubtracted image after embolization shows the cast of the new liquid embolic material fillings, the left ascending pharyngeal artery, and the left-side RM. The relationship between the RM and auditory bullae is well shown. (C) Right ascending pharyngeal injection shows bilateral anastomotic channels between the RM and the external ophthalmic branch of the internal maxillary artery. (D) After embolization of the left side of the RM, the left common carotid artery injection shows retrograde filling of the rete through the ramus anastomaticus arising from the left internal maxillary artery. 1, right ascending pharyngeal artery; 2, left ascending pharyngeal artery; 3, RM; 4, right and left internal carotid artery; 5, arachnoid bullae; 6, embolized rete; 7, basilar artery; 8, internal maxillary artery; 9, arteria anastomotica of Daniel, Dawes, and Prichard; 10, ramus anastomaticus; 11, left common carotid artery; 12, external ophthalmic branch of the internal maxillary artery; 13, occipital artery.


The RM has been widely used by investigators who are interested in understanding and developing AVM-related in vivo experiments.6 It serves as an animal model for evaluation of newer embolization materials and techniques. Some authors also use it to study the effects of radiation on the AVM nidus.2 ,7 Although many studies have used the RM for their experimental model, to the best of our knowledge, the detailed dissection technique of its procurement has not been described. The few dissection techniques that have been described require surgical expertise, including complete knowledge of swine skull-base and neck surgical anatomy, and thus appear to be complicated. For example, Akin et al5 harvested the RM by performing an ascending dissection technique, starting with a CCA cut-down and then following the carotid artery to the ascending pharyngeal artery and, eventually, the RM. Their method required surgical expertise and an operating microscope, in addition to other instruments such as pneumatic drills. Other investigators harvested the RM after transcranial removal of the brain and then dissection of the sellar and parasellar dura intracranially.2 ,8 This method requires a wide craniotomy of the cranial vault and fixation of the brain in a good medium, so that it can be lifted out without damaging RM structures. Another way of obtaining the rete is by injecting acrylic material to create a vascular corrosion cast.3 A vascular corrosion cast might be good for gross and microvascular anatomical studies but histopathological examinations cannot be performed. Therefore, this technique cannot be used for new embolic agent developments or for studying the effects of radiation on retial vessels, since those experiments require histopathological examinations.

Our post mortem dissection technique does not require extensive swine skull-base dissection or wide craniectomy and brain removal. Thus, there is no need to use large cumbersome tools such as air drills, an operative microscope, or microdissection tools. Our focused dissection through the retropharyngeal space will lead the dissection plane directly to the RM. In all the 11 animals in which we used the focused dissection technique, the RM was successfully harvested, which demonstrates that our technique does not need extensive training.

Even-toed ungulates of the Artiodactyla order have unique cerebral vascular anatomy compared with other mammals. The brain in these species (pigs, camels, deer) is supplied from the RM, which is formed from branches of the external carotid artery (ascending pharyngeal artery in swine).9 The RM is an extensive arborization of small (250–700 µm) vessels that eventually coalesce, after piercing the cavernous sinus dura, to form the internal carotid artery.10 ,11 The internal carotid artery then divides into cranial and caudal branches and often, the basilar artery is constructed from bilateral caudal divisions meeting at the midline. The RM functions synonymously to the circle of Willis in humans and is suggested to benefit the hemodynamic stability and thermoregulation of the animal.3 ,9 The mean arterial pressure beyond the RM has been measured as only one-fifth of the pressure of the supplying parent vessel.12 There have also been some unsubstantiated reports that the RM assists the venous return to the heart via its pulsation in the cavernous sinus.13

Despite uncertainty about its real function—that is, whether it serves to regulate cerebral blood flow and temperature or aids venous return by transmitting pulsation to the cavernous side, the structural similarity of RM to the AVM nidus is clearly evident. For example, the caliber of retial vessels is similar to that of nidal vessels in AVMs.10 Blood flow through the RM moves to a distal artery, while flow in the AVM nidus is directed into arterialized veins. Thus, the RM remains an attractive animal model for AVM-related experimental investigations.14 Our focused, relatively simple and time efficient dissection technique might facilitate future investigations using the RM as an animal AVM model.

Potential limitations of our study include lack of objective evaluation of the time efficiency, since we could not objectively compare the required surgical dissection time of our technique with that of previously described techniques. However, based on our initial experience the first three animals took approximately 2 h per animal and the focused dissection technique took about 25 to 30  min of harvest time for each animal, indicating substantially improved efficiency. Another limitation of this technique is that it might not be applicable for use by investigators who wish to evaluate other adjacent anatomical structures, such as the CCA and its branches, proximal ascending pharyngeal artery, epidural space, subarachnoid space, dural coverage, and venous drainage, since most of above-described anatomical structures would probably be severed or ignored during our dissection.

In summary, our focused RM dissection technique takes only a quarter of the time required by other techniques, yet can safely procure the RM with minimal risk of damage. Moreover, in spite of its effectiveness, our technique did not require sophisticated equipment or extensive knowledge or surgical technique for skull-base surgery and dissection. With our focused dissection technique, the RM was accurately located in all animals without any difficulty and removed with a preserved microangiographic structure and histopathological characteristics.


The RM of the domestic pig is an excellent animal model for AVM experiments, especially for evaluation of newer treatments or for understanding the effect of pre-existing modalities. Since most of these studies require comprehensive histopathological examination of the RM, its undamaged and complete harvest is one of most important parts of the experiment. Using our focused, relatively easy, and time-efficient post mortem dissection technique, successful harvesting of the RM can be obtained, with little surgical training and without using sophisticated tools such as an operating microscope, pneumatic drills, air saws, or brain fixation. Our technique can easily be used investigators who are working on the swine rete.


View Abstract


  • Contributors JKE and S-KL designed and performed the post mortem dissections, collected and analyzed data, drafted and revised the manuscript. MN and CW performed the post mortem dissections, critically reviewed and edited the manuscript. All authors read and approved the final manuscript.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.