Sheep Model of Hemodialysis Arteriovenous Fistula Using Superficial Veins Marius C. Florescu,* Kirk W. Foster,† Andrew R. Sacks,* John Lof,‡ Elizabeth A. Stolze,‡ Gretchen M. Fry,‡ Derek P. Bumgardner,‡ Tara Tysinger,§ Melanie J. Kuchta,§ Henry J. Runge,¶ William B. Hadley,¶ and Michael C. Morris** *Section of Nephrology, University of Nebraska Medical Center, Omaha, Nebraska, †Pathology Department, University of Nebraska Medical Center, Omaha, Nebraska, ‡Large Animal Laboratory, University of Nebraska Medical Center, Omaha, Nebraska, §Vascular Laboratory, University of Nebraska Medical Center, Omaha, Nebraska, ¶UNEMED, University of Nebraska Medical Center, Omaha, Nebraska, and **Section of Kidney and Pancreas Transplant Surgery, University of Nebraska Medical Center, Omaha, Nebraska

ABSTRACT Current models of animal arteriovenous fistula (AVF) are swine models of femoral vein terminolaterally anastomosed to femoral artery, creating a deep AVF. This feature sets it aside from human AVFs using superficial veins. Our AVF model uses sheep superficial veins to create an AVF almost identical to human model. AVFs were created in six sheep using basilic veins sutured terminolaterally to brachial artery. Presurgery vein and artery diameters were measured. We measured AVFs and feeding arteries blood flows and diameters at 1, 3, and 5 weeks postsurgery. At 5 weeks we performed angiograms, euthanized animals, and harvested AVFs. Four

animals completed the study. Three AVFs developed and were patent at 5 weeks; one thrombosed. Animal weight and presurgery vessels diameters predicted AVFs blood flows and diameters. Despite using vessels with diameters smaller than the ones recommended for human AVF, the Fistulas developed. Two animals died before the study conclusion for causes unrelated to surgery. This AVF model is anatomically almost identical to the human AVF and has a good maturation rate. It is a viable model for studying AVF maturation, devices intended to improve AVF maturation, AVF related procedures and can even support hemodialysis needles.

Arteriovenous fistulae (AVF) are presently considered the best vascular access for hemodialysis (1) but have a poor maturation rate. Failure rates have been reported as high as 60%, often requiring patients to undergo multiple endovascular or open salvage procedures. Despite this, a high proportion still does not mature and are ultimately abandoned (2,3). To better understand and ultimately improve AVF, it is necessary to create a similar model in large animals. Existing models have employed a variety of animals and vessels (4). Swine models are among the most common due to their relatively simple care needs and low possibility of the animal disturbing the wound (4–7). Other animal models that have been used less commonly include dogs, sheep, and nonhuman primates (4). The type of fis-

tula has been equally varied. These have included femoral artery to femoral vein, carotid artery to jugular vein, axillary artery to brachial vein, and aorta to inferior vena cava. Moreover, both side-to-side and end-to-side techniques have been used on native vessels; several studies have used synthetic material grafts. Studies using the aforementioned models often cite the similar geometry and flow parameters to human AVF as advantageous features. However, they differ from the ones used in humans because they are often deeper, larger in the context of the animal’s circulatory system, and they are using deep veins surrounded by muscles instead of superficial veins as the human AVF. In addition, while many investigations have demonstrated patency greater than 1 month, a review of the available literature has shown that fistula maturation and growth have not been a major endpoint. Rather, the primary goals of investigations thus far have been pathologic analysis of the native vessels after creation and mechanisms of stenosis (5–7). This has contributed to the popularity of the swine model, since inflammation and stenosis develop predictably and quickly. As a result, less attention in large animal

Address correspondence to: Marius C. Florescu MD, Associate Professor of Medicine, 983040 Nebraska Medical Center, Omaha, NE 68198-3040, Tel.: +402-559-9227, Fax: +402-559-9504, or e-mail: [email protected]. Seminars in Dialysis—Vol 28, No 6 (November–December) 2015 pp. 687–691 DOI: 10.1111/sdi.12407 © 2015 Wiley Periodicals, Inc. 687

688

Florescu et al.

research has been devoted to studying the nature of AVF maturation or means of improvement. This can only be accomplished using native vessels (i.e., cannot be studied with a graft), and to better reflect the conditions in human patients should be done with vessels that are both superficial and peripheral. We present a sheep model with the goal of creating a fistula more closely resembling those used in humans than previous large animal models. To this end, we employed an end-basilic vein to side-brachial artery configuration in a black faced sheep. This model uses superficial veins located under the skin and if needed can even support needles placement for HD. Methods Because of the similar anatomy of the forelimb of large sheep to the human arm, and the presence of organized superficial veins we decided to create the model in sheep. All AVFs were created using the same vessels: brachial arteries and basilic veins. We called the large superficial vein we used, basilic vein because of its medial location similar with the human basilic vein. The study was approved by the University of Nebraska Medical Center Institutional Animal Care and Use Committee (IACUC). Using Doppler ultrasound, we measured the blood flow through the AVF and the feeding artery pre-and postanastomosis. We also measured vessel diameters at surgery, 1, 3, and 5 weeks postsurgery. At 5 weeks we performed AVF angiogram. Then the animals were euthanized, and the AVF harvested for pathology evaluation. We used anticoagulation with aspirin and clopidogrel (Plavix). Initially, we used clopidogrel 150 mg/ day, a dose similar to humans. Later in the study, an article (8) was published reporting that the sheep cannot effectively absorb clopidogrel, and an effective dose would be 225 mg po BID. We changed to that dose 3 weeks after AVF surgery in animal number 3. At that time, we added aspirin 325 mg/day. We used anticoagulation because each animal served as its own control in a study evaluating a new device to improve AVF maturation and needed anticoagulation. Each animal had two AVF created: on one forelimb the AVF was created using the experimental device and on the other forelimb the AVF was created using the usual technique we will describe here. Both surgeries were performed the same day, by the same team, under general anesthesia. The primary operator is a transplant surgeon with a vast expertise in hemodialysis vascular access surgery. Animal Care Sheep were housed communally, fed Teklad Ruminant Diet 7060 (Harlan Laboratories, Madi-

son, WI) at approximately 2% of body weight daily, and given one handful of alfalfa daily as an enrichment treat. Prior to surgery, sheep were fasted for 24 hours and water was withheld for 12 hours. On the day of surgery sheep were given an IM injection of ketamine (10 mg/kg—KetaVed, Vedco, St. Joseph, MO) and xylazine (0.1 mg/kg—AnaSed, Lloyd Laboratories, Shenandoah, IA) followed by IM atropine (0.15 mg/kg). The sheep were weighed, then placed on isoflurane inhalation anesthesia via cone mask, to accomplish intubation, and placement of orogastric tube. Once sufficiently anesthetized, a cuffed endotracheal tube was placed in the trachea with the aid of a laryngoscope and the cuff inflated to avoid aspiration of rumen contents. The orogastric tube was then placed. Animals were kept in sternal recumbency until after intubation and orogastric tube placement. Animals were ventilated at 10– 15 ml/kg, at 8–12 breaths per minute (BPM) once intubated. Preparation continued with shaving of the hind leg for mid saphenous vein vascular access placement, both forelegs for surgeries, and areas for placement of EKG leads and cautery grounding pad. The intravenous catheter was secured with suture. Once IV placement was completed, the animal was placed on the surgery table. Surgery was done under isoflurane anesthesia (induction at 4.0%, maintained at 1.75–2.0%). Sheep were monitored continuously with Sp02, heart rate, and EKG. The surgical site was prepped with the application of betadine scrub and prep followed by chlorhexidine. Lactated Ringers with 5% dextrose was administered at 2–4 ml/kg/hour, IV. Prior to the opening incision sheep received cefazolin (20 mg/kg, IV) and carprofen (2 mg/kg, SQ). Postsurgery sheep received penicillin G procaine and penicillin G benzathine (10,000 U/kg, IM). Animals were carefully extubated and continuously monitored, maintaining sternal recumbency, until they could stand independently. Carprofen (2 mg/ kg, SQ) was given once daily for the first 2 days postsurgery for pain management. Penicillin G procaine and penicillin G benzathine (10,000 U/kg, IM) administration was repeated on the 3rd day postsurgery. Sheep were observed at least once daily for the 3 days following surgery, then three times per week until euthanized. Each animal received clopidogrel (225 mg, BID, PO) and aspirin (325 mg/day PO), until euthanized. AVF Creation—Surgical Technique 1 The study animal was sedated and intubated and placed on the surgical table in lateral decubitus. Preparation was done as described previously. The location of the artery and vein were marked on the skin (Fig. 1). Ultrasound evaluation of feeding artery and veins is encouraged for optimal results.

689

ANIMAL MODEL OF HEMODIALYSIS ARTERIOVENOUS FISTULA

Fig. 1. Medial aspect of the sheep forelimb with the location of the artery and vein clearly marked.

2 Skin incision and dissection of the brachial artery and basilic vein (Fig. 2). 3 The basilic vein was mobilized for a length of approximately 5 cm with side branches ligated as needed. The distal vein was ligated and a long venotomy created on the posterior aspect of the vein approximating the anastomosis. The vein was flushed with 1000 units/l of heparinized saline solution. 4 The brachial artery was similarly mobilized with ligation of side branches. An anterior arterotomy of 7 mm length was made and the artery was flushed proximally and distally with 1000 units/l heparinized saline solution. Vascular anastomosis of the fistula is accomplished using a 7.0 prolene suture on a CV-1 needle (Fig. 3). The anastomosis is started in a side-toside technique at the cephalad end. The back wall is done first. As the back wall is completed, the vein is trimmed to convert the anastomosis to an end-to-side configuration (Fig. 4). Vessels are clamped using Fine Pediatric Vascular

Fig. 3. The basilic vein is sutured to the brachial artery in a termino-lateral fashion using a continuous suture.

Fig. 4. Final AVF.

Clamps. At the completion of the anastomosis, distal clamps are released first. Proximal clamps are then released. The suture is tied after the clamps are released. Hemostasis is completed with the use of a gelatin compressed sponge soaked in thrombin. 5 Skin closure using continuous sutures (Fig. 5).

Results

Fig. 2. Brachial artery and basilic vein exposed after skin incision and dissection.

In our four animals who completed the study 3 AVFs developed and were patent at 5 weeks and only one failed. Table 1 describes the AVFs blood flow and inner diameters at 1, 3, and 5 weeks as well as brachial arteries and basilic veins inner diameters at the time of surgery.

690

Florescu et al.

Fig. 6. Doppler ultrasound image of AVF. Fig. 5. Skin closure.

The surgery took an average of 1 hour to complete. One animal (#1) died of aspiration pneumonia 1 week post surgery and animal #4 died the next day after surgery from a preexisting liver abscess. Both animals had autopsies that established the causes of deaths. Neither of the animal’s deaths could be considered direct consequences of AVF surgery. The only complication related to the surgery occurred in animal # 2 who developed postsurgical wound dehiscence at the right front leg that resolved with supportive treatment and did not became infected. The AVF matured successfully. The animals recovered well after the surgery and were not bothered by the surgical procedure. They did not disturb the surgical wound by excessive scratching. There were no surgical wound infections. There was no evidence of distal ischemia in the front limbs and no evidence of congestive cardiac failure in any of the animals. Figure 6 shows the Doppler ultrasound of AVF and Fig. 7 shows the same AVF angiogram. Discussion Our AVF model showed a good maturation rate (75%) and blood flow 5 weeks post surgery.

Fig. 7. AVF angiogram.

Because the number of animals is small, we did not consider necessary to present mean values for AVF blood flow or diameter. The AVF is anatomically similar to a human brachio-basilic AVF. However, the feeding artery diameters ranging between 1.8 mm and 2.5 mm is smaller than the human brachial artery and closer in size to the human radial artery. The vein used for AVF creation has a smaller diameter (2–3 mm) than the human counterpart. Similar to the human AVF, the feeding artery diameter and the vein diameter seem to predict AVF maturation and 5 weeks blood flow. The diameters of the arteries and veins used for AVF in animals number 5 and 6

TABLE 1. Presurgery artery and vein diameters and postsurgery AVF blood flow and diameter Animal #

Weight (kg)

Artery diameter at surgery

Vein diameter at surgery

2

72

2.5 mm

3

68

5 6

AVF week 1

AVF week 3

AVF week 5

3 mm

Not measured

Not measured

2.3 mm

2.4 mm

Q = 115 ml/ minute Ø = 2.9 mm

Q = 46 ml/ minute Ø = 2.4 mm

Q = 550 ml/minute Ø = 5 mm Q = 127 ml/minute Ø = 3.4 mm

55

1.8 mm

2 mm

thrombosed

thrombosed

thrombosed

59

1.9 mm

2 mm

Q = 54 ml/ minute Ø = 2 mm

Q = 68 ml/ minute Ø = 2 mm

Q = 76 ml/minute Ø = 2 mm

Plavix – clopidogrel; Ø – diameter; Q – blood flow.

Anticoagulation Plavix 150 mg qd Plavix 150 mg qd After week 3 Plavix 225 mg BID Aspirin 325 mg qd Plavix 225 mg BID Aspirin 325 mg qd Plavix 225 mg BID Aspirin 325 mg qd

ANIMAL MODEL OF HEMODIALYSIS ARTERIOVENOUS FISTULA

were inferior to the ones recommended for human AVF: diameter of at least 2 mm for the artery and 2.5 mm for the vein at the site of anastomosis (9,10). We chose to perform AVF surgeries in these rather small animals to evaluate whether or not the AVF could mature using such small vessels and to compare the performance of the experimental device in such a challenging environment. In our study, the only fistula that thrombosed was the one placed in the smallest animal with the smallest vessels, animal number 5. In our model, the sheep’s artery diameter is relatively small and the AVF should be created as high on the front leg as possible. This way the brachial artery diameter is slightly higher and maximizes the chances of AVF maturation. The animal’s weight seems to be proportional to the size of the vessels and hence predicts AVF maturation. We noted that large animals with larger arteries and veins (animals #2 and 3) successfully developed AVF and had higher blood flow compared with smaller animals (#5 and 6). We suggest using larger animals to increase the chance of AVF maturation. The surgery took about 1 hour to complete and besides one AVF with skin dehiscence, there were no local complications. The surgery is very similar to the one performed in humans and uses an identical technique. It is important to have a team and facility able to handle large animals. The team should be trained in the special techniques of anesthesia in sheep and continuous gastric aspiration during the surgery to avoid aspiration pneumonia. We lost one animal because of aspiration pneumonia, but after changing our anesthesia technique this problem did not happen again. The second animal fatality was caused by a preexisting liver abscess that could not have been predicted. As presented before, we used anticoagulation during the study. A human study (3) showed that the use of clopidogrel for 6 weeks post surgery did not improve AVF maturation rate and we don’t expect that our results to be affected by the use of anticoagulation. This AVF model is very similar to the AVF used in humans. Compared with the swine, the sheep has long legs which cause the superficial veins to organize into large veins with valves similar to the human arm to facilitate the blood return. The presence of those large superficial veins, as well as a brachial

691

artery that runs both superficial and close to one such large vein allowed us to create an AVF that anatomically is almost identical to human AVF. The superficial veins we used are easily accessible compared with AVF models using deep veins, meaning our model can be used for research involving HD needles placement, dialysis vascular access procedures, and devices for AVF maturation in addition to studying the AVF hemodynamics and pathology. Conclusion The AVF we present is anatomically almost identical with the human AVF, has a good maturation rate, and reasonably low complication rate. Odds of successful maturation improve with larger animals and vessels. We consider that this AVF is a viable model for further research of AVF maturation and devices intended to improve AVF maturation.

References 1. Besarab A, Work J, Brouwer D, Bunchman TE, Dinwiddie LC, Goldstein SL, et al.: KDOQI clinical practice guidelines and clinical practice recommendations for 2006 updates: hemodialysis adequacy, peritoneal dialysis adequacy and vascular access. Am J Kidney Dis 48: S1–S322, 2006 2. Huijbregts HJ, Bots ML, Wittens CH, Schrama YC, Moll FL: Blankestijn PJ; CIMINO study group. Hemodialysis arteriovenous fistula patency revisited: results of a prospective, multicenter initiative. Clin J Am Soc Nephrol 3(3):714–719, 2008 3. Dember LM, Beck GJ, Allon M, et al. ; Dialysis Access Consortium Study Group: Effect of clopidogrel on early failure of arteriovenous fistulas for hemodialysis: a randomized controlled trial. JAMA 299 (18):2164–2171, 2008 4. Terry CM, Dember LM: Novel therapies for hemodialysis vascular access dysfunction: myth or reality? Clin J Am Soc Nephrol 8 (12):2202–2212, 2013 5. Loveland-Jones CE, Jayarajan S, Fang J, Monroy A, Zhang HM, Holt-Bright L, et al.: A new model of arteriovenous fistula to study hemodialysis access complications. J Vasc Access 15(5):351–357, 2014 6. Roy-Chaudhury P, Khan R, Campos B, Wang Y, Kurian M, Lee T, et al.: Pathogenetic role for early focal macrophage infiltration in a pig model of arteriovenous fistula (AVF) stenosis. J Vasc Access 15 (1):25–28, 2014 7. Krishnamoorthy MK, Banerjee RK, Wang Y, Choe AK, Rigger D, Roy-Chaudhury P: Anatomic configuration affects the flow rate and diameter of porcine arteriovenous fistulae. Kidney Int 81(8):745–750, 2012 8. Weigand A, Boos AM, Ringwald J, Mieth M, Kneser U, Arkudas A, et al.: New aspects on efficient anticoagulation and antiplatelet strategies in sheep. BMC Vet Res 9:192, 2013 9. Feldman HI, Joffe M, Rosas SE, Burns JE, Knauss J, Brayman K: Predictors of successful arteriovenous fistula maturation. Am J Kidney Dis 42(5):1000–1012, 2003 10. Silva MB Jr, Hobson RW II, Pappas PJ, Jamil Z, Araki CT, Goldberg MC, et al.: A strategy for increasing use of autogenous hemodialysis access procedures: Impact of preoperative noninvasive evaluation. J Vasc Surg 27(2):302–308, 1998