Ross DA, Chow JY, Shin J, Restifo R, Joe JK, Sasaki CT, Ariyan S. Arterial Coupling for Microvascular Free Tissue Transfer in Head and Neck Reconstruction. Arch Otolaryngol Head Neck Surg. 2005;131(10):891-895. doi:10.1001/archotol.131.10.891
To demonstrate the efficacy of arterial coupling.
We report our experience in head and neck reconstruction with the Unilink Microvascular Anastomotic System (Synovis MCA, Birmingham, Ala). Data were collected in a consecutive series of 49 patients undergoing composite resection of head and neck tumors followed by free tissue transfer.
All patient care took place at Yale-New Haven Hospital, New Haven, Conn, a university-based tertiary care facility.
Forty-nine consecutive patients aged 43 to 85 years underwent a total of 50 microvascular free tissue transfers using the Unilink coupling device. There were 18 women and 31 men, and the following 3 types of flaps were performed: radial forearm (n = 36), fibula (n = 12), and rectus abdominus (n = 2).
The Unilink coupling device was used in this case series. Each arterial and venous anastomosis was performed with the coupling device. Free tissue transfers were monitored clinically and outcomes were recorded.
Main Outcome Measures
Flap survival and thrombosis of the arterial anastomoses were determined, as was median length of stay.
There were no flap failures in the series. Of the 50 coupled arterial anastomoses, the predominant coupler size used was 2.5 mm in diameter. Reconstructions included 36 radial forearm, 12 fibular osteocutaneous, and 2 rectus abdominus myocutaneous free flaps. One intraoperative arterial thrombosis occurred, requiring hand-sewn anastomosis, and another pulled away from the intact coupler in a steroid-dependent patient. There were no complications related to technical performance of the coupling device. The median length of stay was 14 days.
While hand-sewn anastomoses in free tissue transfer remain the preferred technique for many microsurgeons, use of the coupler is a viable alternative to sutured anastomoses.
Microvascular free tissue transfers (FTTs) have become a mainstay of reconstruction following oncologic resections in the head and neck. Ever since 1960, when Jacobson and Suarez1 first reported their observations on the use of 7-0 silk to complete carotid artery anastomoses in dogs, the gold standard for performing microvascular anastomoses has been the penetrating suture with attached needle. Unfortunately, immediate reconstruction following extirpative cancer surgery is both labor and time intensive. Success of FTTs relies on the quality of the microvascular anastomosis; failure can lead to life-threatening flap necrosis, wound breakdown, and fistula formation.
Numerous advancements have been made in the past 4 decades. Patency rates increased after the introduction of the operating microscope, facilitating the use of microvascular FTTs. Our group has previously demonstrated success in microsurgery when using loupe magnification rather than microscopy because the former has significantly decreased both operative time and cost without marked change in patency rates.2
Other groups have developed alternatives to sutured anastomoses to improve patency rates. The objectives of such alternatives were to minimize vessel wall damage, decrease ischemia and surgical time, and ease the challenge of microsurgery. Nonpenetrating anastomosis techniques would allow for minimal endothelial disruption, which might favorably influence the patency rate. Released in 1962, the Nakayama device consisted of 2 metallic rings and 12 interlocking pins with corresponding holes.3 However, the device did not gain widespread acceptance until 1986, when Östrup and Berggren4 introduced their modified version as the Unilink Microvascular Anastomotic System (Synovis MCA, Birmingham, Ala). To use this device, the ends of 2 vessels are each passed through rings made of high-density polyethylene. The 6 stainless steel pins on each ring impale the vessel walls and evert them at a 90° angle. The rings may then be mechanically approximated when the pins of each ring insert into the holes of the other to create a secure vascular anastomosis.
This system is available with inner diameters ranging from 1.0 to 3.0 mm at 0.5-mm intervals and can be applied successfully to vessels in the same range of sizes. Approved by the Food and Drug Administration, the Unilink coupler is the most commonly used nonsuture anastomotic system. The device can be used without fear of foreign body reaction or long-term sequelae.5,6 Blair et al7,8 have previously shown that the postoperative histologic features of healing following use of the microvascular coupling device are identical to those of sutured vessels. In addition, mechanically coupled anastomoses show a 50% increase of burst strength 16 weeks after surgery compared with sutured vessels.9 Couplers in laboratory animal anastamoses have been more successful than hand-sewn anastamoses, independent of previous irradiation. Notably, in both laboratory and clinical studies, the coupled anastomosis is more expedient than hand suturing vessels.10- 15
The Unilink coupling device has been accepted for use in venous, but not arterial, anastomoses. Some authors have observed vessel wall thinning following arterial coupling without noting any clinically significant sequelae.5,6 Shindo et al16 have reported a small series of arterial anastomoses, in which 2 of 16 coupled arteries thrombosed and none ruptured or developed pseudoaneurysms. DeLacure et al17 described 7 attempted arterial anastomoses for head and neck reconstructions with FTTs but abandoned 2 secondary to perceived vessel wall thickness, impliability, and narrowed lumen diameter. Concerns regarding technical performance of the anastomoses, including vessel wall thickness, small diameters, and high-pressure flow across the anastomosis, have severely limited the acceptance of arterial microvascular coupling.
In light of the laboratory successes in arterial anastomoses, we have begun to perform all anastomotic FTTs with the Unilink coupling device in the belief that faster, more reliable techniques will further the evolution of head and neck reconstruction. This study therefore describes a single institution’s experience with fully coupled microvascular anastomoses. In this retrospective review of 50 microvascular FTTs, we found that the use of microvascular coupling for FTT reconstruction following head and neck oncological surgery is equivalent to, or better than, sutured anastomoses.
A retrospective review was undertaken of all cases of microsurgical free flap reconstruction performed by the Yale Head and Neck Reconstruction team from November 2001 through December 2003. Inpatient and outpatient medical records were reviewed and evaluated; all relevant information was extracted and filed into the Excel spreadsheet program (Microsoft, Redmond, Wash). Data strata included patient age, sex, tumor location, history of radiation therapy, type of FTT, time of operation, length of stay (LOS), recipient vessels, and coupling device diameter.
The surgical teams included an ablative team (led by C.T.S. or J.K.J.), a reconstructive team (D.A.R., J.S., S.A., and R.R.) and, when necessary, a radiation oncology team for intraoperative brachytherapy implantation. The ablative team typically completed surgical extirpation prior to the start of reconstruction, allowing for the exact extent of the defect to be known before harvesting and fashioning of the donor flap. Operative data, complications, and outcomes were reviewed and compared with prior historical controls and results in the published literature.
Forty-nine consecutive patients underwent a total of 50 microvascular FTTs using the Unilink coupling device. There were 18 women and 31 men, with an average age of 64 years (range, 43-85 years). Three types of flaps were performed: radial forearm (n = 36), fibula (n = 12), and rectus abdominus (n = 2). The reconstructions were for a variety of defects, including floor of mouth, tongue, tonsil, skin resurfacing, orbit, cervical esophagus, and mandible only (Table 1). Prior to surgery, signed informed consent was obtained from each patient.
A total of 50 arterial anastomoses were completed. Each reconstruction was performed by the same lead surgeon (D.A.R.), who was assisted by 1 of 2 assistant surgeons (J.S or S.A.). Loupes with original magnification ×3.5 were used to perform the microanastomoses. All but 1 FTT had double venous and a single arterial anastomoses. The 2.5-mm coupler was used in 41 cases, while the 3.0-mm coupler was used in 8, and the 2.0-mm coupler in just 1. The recipient vessels were the superior thyroid (n = 31) and the facial (n = 19) arteries. Indications for reconstruction are listed in Table 2.
Forty-five operative times were available for the 50 coupled anastomoses. Because this was a retrospective review, only total operative times including both surgical extirpation and ensuing graft harvest and transfer were obtainable. In the last 3 cases, the arterial anastomotic coupling times were also noted.
Although use of the coupler with venous anastomoses has previously been well described, mounting the artery has not. All arteries were dilated with the tips of a curved microneedle holder because its spring-action handle provided increased dilating force compared with the standard vessel dilator. Even after significant dilation, the thicker walls of arteries sometimes remain more resistant to piercing by the coupler pins. While those arteries with the thickest vessel walls may have appeared to have intimal tearing when mounted on the coupler ring, the vessel ends were generally easily approximated with satisfactory flow. In contrast to the hand-sewn technique, minimal preparation was required prior to anastomosis. Only nominal removal of vessel adventitia was necessary prior to mounting on the coupler rings. Each FTT was performed with a single arterial and 2 venous anastomoses.
The total operative times were evaluated. Two groups were then analyzed: the first 23 cases performed compared with the last 22. The 2-tailed, unpaired t test was used to measure statistical differences.
All patients were treated with 81 mg/d of aspirin for 2 weeks postoperatively. No intravenous dextran 40 or heparin sodium was used, except during revascularization in those patients with venous thromboses. Patients were kept well hydrated with crystalloid fluids to maintain adequate blood volume and pressure.
The FTTs were monitored every 2 hours for the first 24 hours, then every 3 hours until the third postoperative day, followed by examinations every 6 hours, then every 12 hours, and finally every 24 hours in successive days. Color, turgor, capillary refill, Doppler ultrasonographic findings, pulse palpation, and dermal bleeding to needlestick were used to clinically assess vascular patency. Following discharge, patients were followed up for 1 to 17 months.
No flap failures were encountered in our series. None of the patients manifested any short- or long-term foreign body reaction. Previously irradiated patients did not present with an increased rate of anastomotic complications, though their average LOS was increased.
Of the 50 arterial anastomoses, 1 intraoperative thrombosis of a 2.0-mm ring size coupler occurred, requiring resection of the coupler and repeated coupling following further dilation of the vessels to 2.5 mm in diameter. Postoperatively, the patient did well without any further difficulties. A single postoperative hematoma occurred in this same patient where the 2.0-mm coupler did not provide adequate flow, and the anastomosis was hand sewn. The only other arterial complication occurred in a steroid-dependent, immunosuppressed renal transplant patient who, on postoperative day 12, had a rupture of the anastomosis requiring emergent neck exploration and ligation of the bleeding vessel. The coupler was found intact, but the radial (recipient) artery had torn away from the coupler pins, allowing the superior thyroid (donor) artery to bleed into the neck potential space. Fortunately, though the vessel was tied off, adequate neovascularization had taken place, and the FTT was able to survive without its initial donor arterial supply.
We performed a total of 100 venous couplings and encountered only 2 venous thromboses, each occurring on the second postoperative day. In the first patient, lack of ipsilateral venous access required a venous anastomosis to the contralateral neck. When venous congestion was later noted, surgical exploration revealed pedicle compression as the vein traveled over the thyroid cartilage to the contralateral internal jugular vein. The other observed venous complication involved a twist in the venous pedicle, which propagated to the coupling site, causing compression, stasis, and subsequent thrombosis. Both patients were emergently reexplored and underwent successful thrombectomy, recoupling, and 5 days of intravenous heparin anticoagulation without further complications.
Median LOS was 14 days (mean, 16.1 days). Those patients receiving previous radiation had a median LOS of 22 days (mean, 21.2 days). All patients were discharged only after tracheal decannulation and either demonstration of adequate oral nutrition or per enterogastric tube placement.
Available total operative times (45 of 50 cases) were also examined (Table 3). Duration of the first 23 cases were compared with the last 22. The average and median total operative times decreased between the former and the latter, from 661 minutes and 645 minutes to 597 minutes and 606 minutes, respectively. Unfortunately the t test did not show statistical difference between the 2 groups (P = .10). Although times for anastomoses were not recorded early in our experience, the last 3 cases were timed. Including vessel preparatory time, the last 3 arterial anastomoses took 7, 6, and 7 minutes, respectively.
Free tissue transfers for head and neck reconstruction have grown in popularity over the past 20 years. This enthusiasm has been attributed to advances in techniques and instrumentation as well as more reliable donor sites.18 Yet the success of such reconstructions is dependent on multiple variables, including the quality and reliability of the vessel anastomoses.
In fact, these anastomoses comprise the critical portion of free flap reconstruction. Use of the traditional hand-sewn method presents several potential perils. Inadequate eversion of the vessel may lead to luminal exposure of adventitia, which is highly thrombogenic. Uneven placement of sutures may lead to either anastomotic leak or stasis and increased risks for hematoma or thrombosis, respectively. Unrecognized back-walling of a stitch may cause lumen stenosis and thrombosis. Each of these risks are effectively addressed with the use of the coupler.
When mounted on the coupler rings, the vessel walls are each everted 90°, permitting complete visualization of the 2 lumens. This detailed view of the intima allows for accurate and even placement on the 6 pins of each coupler ring, thereby eliminating the risk of compromised blood flow secondary to uneven suturing or inadvertent suture placement through the back wall. In our series, no leaks occurred at the coupling site. The male-to-female interlocking design of the couplers, each with 6 pins that insert into the other, serves to hold the anastomosis tightly in place. In addition, the rigid plastic rings form a stent at the site of anastomosis, preventing vessel spasm and thrombosis at a common and important site for both complications.
Use of the mechanical coupling device for arterial anastomoses is a controversial topic. Although some authors believe that the technical challenges and learning curve associated with its use outweigh potential benefits, we have found the coupler to be effective and efficient as a method for microvascular anastomoses. Shindo et al16 reported 2 thromboses in 16 arterial anastomoses, while Berggren et al19 reported 1 in 5 and Ahn et al20 described 5 of 29. However, we experienced only a single intraoperative thrombotic event during the course of 49 coupled arterial anastomoses. Other authors have also noted difficulty manipulating the thick vessel walls when everting them onto the pins and have suggested that this may have led to decreased laminar blood flow and possibly even complete obstruction of the vessel lumen following completion of the coupled anastomosis.20 We share this concern regarding increased thickness and decreased pliability of arterial vessel walls, particularly when couplers measuring less than 2.5 mm were used. Our experience has indicated that couplers below this diameter were more likely to restrict blood flow, as was the case in our single intraoperative thrombotic event, in which a 2.0-mm coupler was used. Since that time, we have dilated donor and recipient arteries to accommodate a ring size no smaller than 2.5 mm. We believe that this minimum diameter permits reasonably rapid flow across the anastomosis, preventing the formation of large thrombi and quickly clearing smaller clots. We did not encounter any technical difficulties or clinically relevant complications while dilating the donor (superior thyroid or facial) or recipient (radial forearm, fibular, or rectus abdominus FTTs) arteries. Traumatic tearing or cracking of the intima has also been previously reported. Although this also occurred in our experience, we chose to continue with the coupling and observed that in each case except 1, the anastomosis remained patent and without leakage. The notable exception occurred in the steroid-dependent patient, who developed a postoperative hematoma requiring emergency neck exploration and ligation of the involved vessel. This patient had been a recipient of a cadaveric renal transplant and therefore had been receiving long-term, high-dose prednisone and immunosuppressants. Given her predisposition for impaired wound healing, she may have had the same result even following a hand-sewn anastomosis. We postulate that anastomotic thromboses secondary to intimal tears were unlikely to occur because the trauma was likely restricted to the areas of eversion, which were approximated to one another and therefore never exposed to blood flow following coupling.
Others have previously demonstrated the utility and safety of coupled venous anastomoses. For our FTTs, we used 2 coupled end-to-side venous anastomoses to the internal jugular veins or its large tributaries. Therefore, we cannot comment on the actual rate of thromboses because a thrombotic event in a single vein may not have been clinically significant owing to adequate compensation of the other venous anastomosis. However, neither of the 2 venous complications we experienced was attributable to the coupler device. The first complication was caused by compression of the flap pedicle as it traveled over the thyroid cartilage to gain vascular access to the contralateral neck. Once repositioned, the improved pedicle geometry allowed the flap to be salvaged after thrombectomy and recoupling. The second complication occurred as a result of an unrecognized twist in the vascular pedicle. Rotational correction and recoupling reestablished venous outflow, and the FTT survived without further complication.
Our experience suggests that the technique is also valid for arterial anastomoses. Our series clearly demonstrates the reliability of arterial coupling, and we recommend that it be considered as a valid alternative to hand-sewn anastomoses. At our institution, skilled microsurgeons have had no difficulty in quickly mastering the technique. In fact, most find it less challenging than hand sewing vessels. In addition to the low rate of complications, our median LOS was 2 weeks, which falls at the low end of the average LOS of 14 to 28 days following FTT. Although Ryan and Hochman21 reported a median LOS of just 9 days following FTT reconstruction after head and neck surgery, they chose to discharge most of their patients prior to resumption of eating by mouth and tracheal decannulation.
Over the course of 2 years, there was a decrease in total operative time, though the validity of this data is limited (P = .10) because it reflects both extirpation and FTT rather than FTT alone. Given the wide variation in operative times for head and neck surgery (eg, the degree of difficulty depending on site and extent of disease, the need for bilateral rather than unilateral neck dissection, the adequacy of initial margins, and the varying difficulty of donor site harvests), the total operative times presented herein provide only a general impression of the learning curve associated with the use of couplers for arterial anastomoses. We believe that the operative time required for coupling will be less than that for hand-sewn anastomoses, though we currently do not have enough data to prove this hypothesis. To fully examine the question of operative time, we are continuing to collect data in a prospective trial to adequately compare coupler vs hand-sewn anastomosis operative times.
In conclusion, we have shown couplers to be a safe alternative to hand-sewn arterial anastomoses. Our complication rate is far lower than many previously reported in the literature,16,19,20 with just 2 complications in 50 coupled arterial anastomoses while maintaining an average LOS similar to that previously reported.
Correspondence: Douglas A. Ross, MD, Section of Otolaryngology, Yale University School of Medicine, 333 Cedar St, Campus Box 208041, New Haven, CT 06520-8041 (firstname.lastname@example.org).
Submitted for Publication: December 29, 2004; final revision received March 28, 2005; accepted May 2, 2005.
Financial Disclosure: None.
Funding/Support: This study was supported in part by the McFadden, Harmon, and Mirikitani Endowments, all located in New Haven, Conn.