Appendix
Materials and methods
Animal model
Animals were pretreated with acetyl salicylic acid (500 mg/day) and Amiodarone (100 mg twice a day) starting four days prior to intervention. Prior to surgery sedation was achieved with injection mixture of ketamine (20 mg/kg IM), azaperone (1 mg/kg IM) and atropine (0.01 mg/kg IM). Anesthesia was induced with propofol (3-5 mg/kg iv), and the animals were intubated after local anesthesia of the larynx. Animals received furthermore buprenorphine (Temgesic®, Essex Chemie, Lucerne, Switzerland, 10 mcg/kg) and carprofen (Rimadyl®, Pfizer, Switzerland, 4 mg/kg) intravenously. Anesthesia was then maintained via inhalation with a maximum volume of 1.5 % isoflurane in oxygen and intravenous propofol (0.01-0.04 mg/kg/min). Intraoperative monitoring was performed including electrocardiogram, heart rate, and directly measured blood pressure, respiratory rate, inspired and expired gases (CO2, O2 and isoflurane) and core body temperature.
In this model, the right femoral artery was percutaneously punctured under ultrasound-guidance and a 6F introducer sheath introduced . An intravenous bolus of heparin (50 IU/kg) was given to achieve systemic anticoagulation. The pharyngeal artery was superselectively catheterized with a 5F Envoy (Codman Neurovascular, Raynham, MA, USA) guiding catheter. A micro-catheter, either a Marathon (Covidien/ ev3 Neurovascular, Irvine, CA, USA) or a balloon-tip micro-catheter Scepter C (Microvention Inc., Tustin, CA, USA) was positioned into the proximal portion of the rete, possibly in wedge position. Super-selective angiographies were performed to depict the vascular anatomy of the rete and the origin of the internal carotid artery. After filling up the dead volume of the micro-catheter with DMSO, Easyx was slowly injected under continuous fluoroscopic control. Embolization was continued until a satisfactory filling of the rete was achieved without embolization of the ICA, or until no more penetration of the material into the rete could be achieved, and reflux into the pharyngeal artery occurred. After removal of the micro-catheter, control angiographies were performed to demonstrate effectiveness.
In the complex AVM-model the side-to-side arterio-venous anastomosis between the right common carotid artery and the right external jugular vein was created without prior endovascular occlusion of the distal external carotid artery. This was done in order to reduce animal sufferance and study costs. Briefly, under sterile conditions, the right side of the neck was longitudinally incised parallel to the sternocleidomastoid muscle. Both the external jugular vein and the common carotid artery (CCA) were surgically exposed and cleaned of adventitia. Four vascular clamps were applied on the artery and the vein and an elliptical arteriotomy and venotomy was performed at the same level. A side-to-side anastomosis was sutured with a running 7.0 prolene suture. The CCA was then permanently ligated proximal to the fistula, allowing only retrograde flow from the rete to the vein. Fistula flow was verified by marked arterial pulsation and dilatation of the jugular vein. Wound closure was achieved by subcutaneous tissue and skin suture in layers.At least three to four weeks after anastomosis surgery the endovascular procedure was performed similar to the previous model: an 5F envoy guiding catheter was positioned in the left pharyngeal artery to demonstrate components of the AVM model and blood diversion through both retia toward the contralateral jugular vein. In this model embolization was performed only from the left side.
The left rete was embolized in all animals, whereas the right rete was occluded directly only in the animals without a complex AVM model, since in these cases there was no more access available to the right pharyngeal artery due to the AV-shunt creation. Concerning performance in target vessel occlusion, the left rete was considered therefore as basis for comparison. Four animals with complex AVM models served as a control group and were treated with Onyx.
Histological analysis
Tissue harvest and processing
Animals were sacrificed by intravenous injection of pentobarbital (150mg/kg). The kidney was extracted after euthanasia and fixed immediately in 4% formaldehyde. Additionally an untreated kidney (animal 13 and 23) was taken as a control organ. To avoid the mechanical artefacts, each rete was carefully exposed and dissected from the cavernous sinus first after fixation of the cranial base in 4% formaldehyde. The fixed retia mirabile from 20 animals (No. 2-17 and 23-26) were cut in a way that both right and left retia were divided into two halves through a middle incision. One half from each left and right rete was embedded in a cassette of paraffin, sectioned at about 4 μm and stained with hematoxylin-eosin (HE) and elastica-van Gieson (vG-E).
After fixation the kidneys of the first 16 animals (No. 2-17) were rinsed in water and further dehydrated in an ascending series of ethanol (50%, 70%, 80%, 90% 96%, 100%) before being placed in xylene under vacuum. The infiltration of the samples was processed under vacuum at 4°C in methyl methacrylate. Embedding of specimens in methyl metacrylate (Methacrylacid-methylester - Fluka Chemie GmbH, Buchs, Switzerland; Dibutylphthalat - Merck-Schuchardt OHG, Hohenbrunn, Germany; Perkadox 16 - Dr. Grogg Chemie AG, Stetten, Switzerland) was performed in customized forms. After polymerization, the blocks were mounted on plastic frames and cut with a precision saw (Leica SP 1600®, Leica Instruments GmbH, Nussloch, Germany). Sections were mounted on acropal slides (Perspex GS Acrylglas Opal 1013, Wachendorf AG, Basel, Switzerland) and polished to 100 - 150μm sections (Exakt® Mikroschleifsystem 400 CS, Exakt Apparatebau GmbH, Norderstett, Germany). Ground sections were surface stained with Giemsa.
The kidneys from the last 4 animals (No. 23-26) were split into symmetrical halves and cut transversally in the area of upper pole, middle part and lower pole respectively, so that every slice contained all anatomical layers of the kidney (cortex, medulla and pelvis). The slices were processed for paraffin embedding, sectioned at about 4 μm and stained with hematoxylin-eosin (HE) and elastica-van Gieson (vG-E).
Gross and histological analysis
Macroscopic assessment of the treated organs was made focusing on various post-embolization changes, involving texture/consistency, extravasation, thrombosis, inflammatory reactions and infarction pattern (hypervascular model, renal artery). Further parameters were in the long survival time animals also assessed, namely organ atrophy and remodelling processes as fibrosis and scarring. Additional emphasis was made on physical appearance of the embolic material in the tissue and its location in embolized vessels.
Histological analysis provided further information about the vessel occlusion (incomplete/complete) and the alignment and adhesion of the material to vascular wall (low/moderate/high-regarding, if any intravascular cells and structures were to be found between the material and the vessel wall). This was performed excluding all possible artefacts due to histological processing (material clefts, holes, shrinkage).
Essential evaluation of the vessels and their surrounding involved: changes in the arterial wall, as absence or damage to its layers, including intramural or perivascular haemorrhage; extravasation of the embolic material; type of inflammatory reaction with location (vascular, peri-vascular and peri-material) and its severity (mild, moderate, severe); and cellular infiltration.
The chronic changes implicated remodelling reactions: material encapsulation, fibrosis and vascular neoformation. All histological slides were evaluated using a light microscope (macroscope Leica 420, microscope Leica DMR system). Photographs were taken from the collected organs and representative histological slides (digital camera, Leica DFC 420 and DFC 320).