Thrombosis of microvascular anastomoses in traumatized vessels: fibrin versus platelets

Thrombosis is the end result of two closely interrelated processes: the coagulation cascade and the platelet aggregation process. To determine their relative contribution, we used pharmacologic agents that selectively block each process. The specific effect of each pharmacologic agent on either fibrin deposition or platelet activity was confirmed morphologically by scanning electron microscopy and was substantiated with ADP-induced platelet aggregation and blood clotting time determinations. Forty-two rats had both femoral arteries subjected to a standardized crush-avulsion injury. A total of 84 femoral microvascular anastomoses were subsequently performed. None of the 24 control anastomoses treated with saline remained patent, whereas 6 of 24 of the anastomoses treated with dazmagrel (a selective thromboxane synthetase and platelet aggregation inhibitor), 2.5 mg/kg IV, remained patent and 18 of 24 of those treated with a single dose of heparin, 200 U/kg IV, remained patent. All 12 anastomoses treated with both drugs remained patent but developed a 33 percent hematoma rate. We conclude that in this microvascular model, fibrin mesh deposition is a more significant factor than platelet aggregation in the pathogenesis of occlusional thrombosis within traumatized arteries. Its temporary inhibition with a single dose of heparin yielded a 75 percent improvement in patency rate.     

Therapeutic value of intravenous heparin in microvascular surgery: an experimental vascular thrombosis study

In an attempt to decrease a 10 to 15 percent vascular thrombosis rate leading to graft occlusion, low-dose human-grade heparin was studied to determine if carefully monitored intravenous therapy would increase 7-day patency in a known potent thrombosis model. In New Zealand white rabbits, the type of infusate administered intravenously, either saline (30 animals) or heparin (35 animals), was selected at random after completing a 2-mm arterial inversion graft in the femoral artery. A 72-hour infusion was used in all animals; the control group received sterile saline and the experimental group received a heparin infusion at 45 microliters per hour after a 500-unit bolus. All grafts in both groups were patent at the time of groin closure. Patency in the heparin-perfused group was 67 percent (24 of 35) as compared to 19 percent (6 of 30) in the control group (p less than 0.05) 1 week postoperatively. Scanning electron microscopy showed significantly less dense fibrin deposition and a decrease in the number of aggregated platelets in the heparin-perfused grafts. Partial tissue thromboplastin time values in the experimental group ranged between 55 and 75 seconds (control 20 to 25 seconds). We have shown that heparin, an inexpensive and readily available agent, maintains 1-week microarterial patency and results in few complications in a reliable, reproducible, and versatile thrombosis model. The clinical ramifications of using an antiplatelet agent that diminishes fibrin deposition in microsurgery are apparent.     

Another method to prevent venous thrombosis in microsurgery: an in situ venous catheter

Free-flap failure is in the order of 4 to 10 percent. Heparin is more effective at preventing venous thrombosis than arterial thrombosis. This study was undertaken to investigate the efficacy of delivering heparin at a high dose locally but low dose systemically (heparin infusion via a catheter placed proximal to the venous anastomosis) to prevent venous thrombosis in microsurgery. A model of venous thrombosis was first established by a venous inversion graft in the rat femoral vein (this was performed in seven animals and resulted in 100 percent thrombosis). Saline and heparin were delivered proximal to the inverted vein graft to assess the effect of each in preventing venous thrombosis. Flow/patency distal to the inverted vein graft was assessed by observation under the microscope, the milk test, and rate of flow (flowmeter). Saline infused via a catheter proximal to the venous inversion graft resulted in 100 percent thrombosis in 10 animals. Heparin (100 U/ml at 2 to 3 ml/hour) infused through a catheter for 2 hours proximal to the anastomosis resulted in flow in all 10 animals during the infusion. Blood was also taken before beginning the procedure (control) and after the heparin infusion distal to the anastomosis (local partial thromboplastin time) as well as in the contralateral femoral vein (systemic). The control for all animals that received heparin was <3 minutes. The systemic partial thromboplastin time after heparin infusion was <3 minutes in seven animals, 3.3 minutes in two animals, and >7 minutes in one animal. The local partial thromboplastin time distal to the inverted vein graft was >10 minutes in nine animals and 3.7 minutes in one animal. The study also had a clinical component, in which a catheter was placed in a vein of the free flap, and heparin was infused over 5 days. This technique has been used in 83 consecutive free flaps. In three recent free flaps performed on the limbs, the local partial thromboplastin time (close to the anastomosis) was raised but the systemic time was normal. This technique offers a method in preventing venous thrombosis in microsurgery. It is simple to implement and is not associated with the systemic complications of heparin.     

Experimental reclamping of free-flap pedicles: the effect of prolonged stasis on the anastomoses and clamp sites

The effect of static blood in direct contact with areas of microvascular anastomoses and previous clamp application for prolonged periods of time has been investigated. The free groin flap was used as a model in 27 white rabbits. The flap pedicle vessels were reclamped proximal to the anastomoses and areas of previous clamp application for periods of time varying between 30 minutes and 4 hours after 15 minutes of blood flow over these areas. A 100 percent patency rate was achieved despite the long periods of reclamping. Histologically, minor intimal damage was visible in the immediate period following anastomoses and clamping of the vessels. After 2 weeks, despite a thickened myofibroblastic intimal lesion, an intact endothelial layer was observed. No evidence of thrombosis could be demonstrated in either period. We postulate that vessels carefully treated and with technically well-performed anastomoses can be regarded as "normal" vessels after 15 minutes of blood flow over the anastomoses and clamp sites. We suggest that when required, microvascular clamps may then be reapplied without risk for prolonged periods of time despite static blood being in contact with these areas.     

Microvascular surgical experimental thrombosis model: rationale and design

The rationale for the design of surgical models of microvascular thrombosis is discussed, and a new model, the arterial inversion graft (AIG), is described and evaluated in the New Zealand white rabbit. Femoral artery segments of predetermined length are excised, gently turned inside-out, and resutured into their native position. Blood flow is restored, and at varying time intervals, vessel patency is assessed through the direct "milking test." In this study, three groups of 20 arterial inversion grafts of 2, 5, and 10 mm in length are created and evaluated for patency at 1 hour and again at 7 days. The incidence of femoral artery occlusion in this model appears to be an increasing function of arterial inversion graft length both at 1 hour--30 percent (2 mm), 80 percent (5 mm), and 100 percent (10 mm)--and at 7 days--65 percent (2 mm), 90 percent (5 mm), and 100 percent (10 mm). This proportionality suggests the arterial inversion graft may be adjusted in length to provide an incidence of vessel occlusion best suited to the needs of any particular experiment.     

Anastomosis of vessels of unequal diameter using an interpositional vein graft

In this study, 100 rabbits were used to assess the efficacy of five different methods of microvascular anastomosis where a vessel diameter discrepancy of 5:1 existed. The inferior vena cava of the rabbit was used as a graft in the femoral artery. In 50 percent of the rabbits the graft was reversed to assess the effects on flow. When explored between 7 and 10 days after anastomosis, an overall patency rate of 96 percent was recorded. Three grafts were not patent in the reversed group and one was not patent in the nonreversed group. There was no significant statistical difference in patency rates between any of the groups, as calculated by the Fisher's exact probability test. The tapered end-to-end and side-to-end anastomoses were found to be the most rapid and simplest methods to perform.     

End-to-end versus end-to-side microvascular anastomosis patency in experimental venous repairs

In this experimental study, venous end-to-end and end-to-side microvascular anastomoses in similar and diameter-discrepant vessels were compared. In 50 rats, end-to-end microvascular repair of the divided epigastric vein and end-to-side repair of the epigastric vein into the femoral vein showed 5-day patency rates of 75 and 88 percent, respectively. These data are not statistically different. In 20 rats, microvascular repair of end epigastric to end femoral veins (size discrepant) and end epigastric to side femoral veins showed 5-day patency rates of 50 and 85 percent, respectively. These data are statistically different (p less than 0.05). We conclude from these experimental data that end-to-side venous repairs may be useful in lowering the anastomosis thrombosis rate seen when size-discrepant veins are repaired.     

Reactive thrombocytosis alone does not affect the patency of microvascular anastomosis in the splenectomy rat

Vascular thrombosis is a harbinger of failure in microsurgery. However, there is still controversy regarding the correlation of the complications of thrombocytosis and thrombosis. Some evidence indicates that patients with elevated platelet counts tend to have a higher flap failure rate, and surgeons usually hesitate to operate on patients with thrombocytosis. Nevertheless, the authors have experienced successful free tissue transfer in seven patients with thrombocytosis resulting from traumatic splenectomy or multiple trauma. On the basis of clinical observation, the authors investigated whether reactive thrombocytosis contributes to the patency of a microvascular anastomosis. In a rodent splenectomy-induced thrombocytosis model (n = 40), stable reactive thrombocytosis occurred after postoperative days 5 to 10, with the peak on postoperative day 7. Femoral artery division and reanastomosis was performed in rats with or without splenectomy-induced thrombocytosis, and vascular patency was assessed. Platelet counts and platelet activation were studied in correlation to microvascular patency. Platelet activation as demonstrated by CD62P expression on platelets was not significantly different between rats with and without thrombocytosis (6.41 +/- 0.95 percent versus 4.51 +/- 0.55 percent, respectively; p = 0.089). As immature platelets were not increased (2.86 +/- 0.33 percent versus 1.99 +/- 0.32 percent, p = 0.074), it seems that the splenectomy-induced thrombocytosis is the result of redistribution of platelets instead of an increase in bone marrow production. There were no significant differences in the patency rates or perfusion units of femoral artery after arterial anastomosis between rats with and without thrombocytosis (90 percent and 95 percent, respectively; p = 0.561). In conclusion, this study demonstrates that microvascular anastomosis can be performed safely in patients with reactive thrombocytosis without platelet activation.     

Experimental microvascular polytetrafluoroethylene grafts: 6-month patency

The efficacy of 1-mm internal-diameter polytetrafluoroethylene as a microvascular prosthesis is unclear. In this study, 8-mm-long segments of this material were implanted into the femoral arteries of 30 rats. Animals were examined every 2 weeks up to 6 months by Doppler ultrasound. Cumulative patency by the life-table method was 86.7 percent at 6 months. There were 26 patent grafts, 2 occlusions, and 2 deaths. Intimal hyperplasia adjacent to the anastomoses was seen in the native arteries. The pseudointima lining the grafts was cellular near the anastomoses but usually acellular in midgraft regions. It is concluded that if early failure does not occur, then good long-term patency is possible with 1-mm polytetrafluoroethylene in this setting and that patency is not dependent on development of a cellular pseudointima. Longer graft segments should be evaluated in future studies.     

The effect of established infection on microvascular surgery

The success of microvascular anastomoses in the presence of staphylococcal infection was studied using rat femoral arteries. There was a spontaneous thrombosis rate of 19 percent in normal vessels that traversed the area of infection. Vessels with an anastomosis outside the area of infection had a similar thrombosis rate, but if the anastomotic site was within the infected area itself, the thrombosis rate increased to 75 percent. Inflammatory changes with subsequent fibrosis in the media and adventitia appeared responsible for the thrombosis. The intima was unaffected by the presence of infection. This study suggests that when a microvascular anastomosis is necessary in the presence of infection, the anastomosis should be placed outside the area of infection with a pedicle to traverse the infected area.     

Effect of total venous occlusion on capillary flow and necrosis in skeletal muscle

Limb replantation and microvascular transfer of flaps are sometimes complicated by postoperative venous thrombosis. Total venous occlusion can lead to complete shutdown of microvascular perfusion, resulting in failure of the transfer or replantation. Once venous return stops, it must be restored within a critical period of time for tissue survival. The purpose of this experiment was to delineate this critical period of time at which no reflow and irreversible muscle necrosis occurs by the use of a rat gracilis flap microcirculation model. The gracilis muscle of 40 male Wistar rats (135.3 +/- 37.2 g) was elevated on its vascular pedicle and mounted on a raised platform for videomicroscopic analysis. Animals were randomly assigned to one of four groups: (1) sham (no total venous occlusion), (2) 10 minutes of total venous occlusion, (3) 30 minutes of total venous occlusion, and (4) 60 minutes of total venous occlusion. Total venous occlusion was established by placing a microvascular clamp across the femoral vein at the junction of the gracilis pedicle. The number of flowing capillaries in five consecutive high-power fields (832x) were counted at baseline and at 5, 15, 30, 60, 120, 180 minutes, and 24 hours after reperfusion. At 24 hours after reperfusion, the gracilis muscles were harvested and stained with nitroblue tetrazolium. Percentage of muscle necrosis was measured by using computer planimetry. The data were reported as mean +/- standard error of mean and were compared between groups by analysis of variance and appropriate post hoc comparisons. Total venous occlusion for 10, 30, and 60 minutes showed a significant decrease in the number of flowing capillaries through 24-hour postreversal. There was a significant drop (p < 0.01) in the number of flowing capillaries from 30 minutes of total venous occlusion to 60 minutes of total venous occlusion at all times. Muscle necrosis was significantly increased in all three groups of total venous occlusion compared with the sham group (36.1 +/- 1.7 percent, 45.5 +/- 3.4 percent, 74.1 +/- 4.7 percent versus 14.3 +/- 1.7 percent, and p < 0.01). These results indicate that irreversible tissue damage occurs in a very short time interval (60 minutes) in this model, making the early detection of venous occlusion critical to the successful correction of this complication.     

Decreased thrombogenicity of vascular prostheses following gas denucleation by hydrostatic pressure

The high rate of thrombosis of 1.0-mm polytetrafluoroethylene (PTFE) grafts has limited their use in microvascular surgery. One possible reason for this is the blood-gas interface due to entrapped air in the interstices. The present study examines the effect on patency rates of elimination of this blood-gas interface by high pressurization. Comparing pressurized and nonpressurized grafts in the same animals showed a patency rate of 100 percent at 7 days for treated grafts, while the control (nonpressurized) grafts had all clotted by 1 hour. The implications for microvascular surgery as well as vascular surgery in general are discussed.     

The use of prostacyclin analogue-containing suture for the prevention of postoperative venous thrombosis in the rat

The ability of prostacyclin analogue incorporated into a controlled-release suture to prevent postoperative venous thrombosis was investigated. Thirteen rats underwent bilateral transection and anastomosis of the common femoral vein. In each animal, polycaprolactone suture containing 0.25 micrograms/cm of the prostacyclin analogue Iloprost (Schering Ag, Berlin, West Germany) was used to perform the anastomosis on one vessel. Similar suture without prostacyclin analogue was used on the contralateral vessel, which served as a control. Functional patency and luminal surface morphology were assessed 24 hours postoperatively. All anastomoses performed using suture containing prostacyclin analogue were patent. Among controls, five anastomoses were patent and eight were occluded. This difference was highly significant (p less than 0.005). All anastomoses performed with prostacyclin analogue-containing suture exhibited a uniform absence of thrombosis. In contrast, eight control veins exhibited a dense, well-organized fibrinous clot that filled the entire lumen, effectively sealing off the vessel. These results suggest that the prostacyclin analogue released from the suture was highly effective in inhibiting thrombus formation without adversely affecting the vessel's ability to achieve hemostasis.     

Low-molecular-weight heparin (dalteparin) effectively prevents thrombosis in a rat model of deep arterial injury

Unfractionated heparin is often used to prevent thrombosis in microvascular surgery, but a major drawback of heparin therapy is increased bleeding. Low-molecular-weight heparins prevent venous thrombosis as effectively as heparin and have better bioavailability and a longer plasma half-life, which explains the increased use of low-molecular-weight heparins as substitutes for heparin in clinical practice. However, the ability of low-molecular-weight heparins to prevent arterial thrombosis has been debated. In this study, the authors compared the antithrombotic and antihemostatic effects of heparin and the low-molecular-weight heparin dalteparin in a rat model of arterial thrombosis. A segment of the left common carotid artery was isolated between vascular clamps and opened longitudinally. An endarterectomy was performed and the arteriotomy was closed with a running suture. The antithrombotic effect (vascular patency 31 minutes after reperfusion) and the surgical bleeding were measured. Groups of 10 rats were treated in a blind, random fashion with intravenous injection of one of the following substances 1 minute before clamp release. Three groups received a bolus of heparin (20, 60, or 180 IU anti-factor Xa/kg), three groups received dalteparin (60, 180, or 540 IU anti-factor Xa/kg), and one group was treated with vehicle (saline). Heparin 180 IU/kg produced a distinct antithrombotic effect compared with the control group (p = 0.03), but it also significantly increased the surgical bleeding to 2.0 g compared with 1.5 g in the control group (medians, p = 0.01). Dalteparin 180 and 540 IU/kg also produced a powerful antithrombotic effect (p = 0.01 and p = 0.03, respectively). In contrast to heparin, 180 IU/kg dalteparin did not increase the surgical bleeding (median, 1.5 g; p = 0.37 versus controls). Dalteparin 540 IU/kg increased the median surgical bleeding to 2.6 g (p = 0.06 versus controls). The nonsignificant difference may be explained by the great interindividual variation of surgical bleeding in the high-dose dalteparin group. Dalteparin prevented arterial thrombosis as effectively as unfractionated heparin. In contrast to heparin, dalteparin did not increase the surgical bleeding, which indicates that dalteparin instead of heparin can be used to prevent thrombosis in microvascular surgery.     

Inhibition of microsurgical thrombosis by the platelet glycoprotein IIb/IIIa antagonist SR121566A

Despite major improvements in tools and significant refinements of techniques, microsurgical anastomosis still carries a significant risk of failure due to microvascular thrombosis. The key to improving the success of microvascular surgery may lie in the pharmacologic control of thrombus formation. Central to pathologic arterial thrombosis are platelets. Glycoprotein IIb/IIIa is a highly abundant platelet surface receptor that plays a major role in platelet aggregation by binding platelets to each other through the coagulation factor fibrinogen. To explore the ability of antithrombotic agents to prevent microvascular thrombosis, a rabbit ear artery model was used in which a standardized arterial injury results in predictable thrombus formation. This model was used to examine whether SR121566A, a specific and potent glycoprotein IIb/IIIa inhibitor, can successfully prevent microsurgical thrombosis.Using a coded, double-blind experimental design, 20 rabbits (40 arteries) were assigned to four treatment groups: (1) saline injection (n = 10), (2) acetylsalicylic acid 10 mg/kg (n = 10), (3) heparin 0.5 mg/kg bolus with subsequent intermittent boluses of 0.25 mg/kg every 30 minutes (n = 10), and (4) SR121566A 2 mg/kg bolus (n = 10). After vessel damage and clamp release, arteries were assessed for patency at 5, 30, and 120 minutes by the Acland refill test. Coagulation assays, in vivo bleeding times, and ex vivo platelet aggregation studies were also conducted. Scanning electron microscopy was used to examine mural thrombus composition.A significant, fourfold increase in vessel patency following administration of SR121566A over saline control (80 percent versus 20 percent patency, respectively, at 35 minutes after reperfusion, p < 0.01) was noted. This was correlated with marked inhibition of ex vivo platelet aggregation. This antiplatelet treatment did not prolong coagulation assays (mean international normalized ratio: saline, 0.66 +/- 0.04; SR121566A, 0.64 +/- 0.03; mean thromboplastin time: saline, 19.63 +/- 0.67; SR121566A, 17.87 +/- 3.27) and bleeding times (mean bleeding time: saline, 42 +/- 4; SR121566A, 48 +/- 6). Scanning electron microscopy demonstrated extensive platelet and fibrin deposition in control vessel thrombi. In contrast, thrombi from SR121566A-treated vessels demonstrated predominance of fibrin with few platelets when examined under scanning electron microscopy.Administration of SR121566A was associated with a significant increase in vessel patency, without deleterious effects on coagulation assays or bleeding times. The increase in vessel patency was correlated with inhibition of platelet aggregation and decreased platelet deposition, as demonstrated by scanning electron microscopy. Glycoprotein IIb/IIIa antagonists represent a new class of anti-platelet agents that may be suited for inhibiting microsurgical thrombosis. This study supports further investigation into the use of these agents in microsurgery.     

Endovascular operations on leg arteries

The paper presents the results color duplex scanning (CDS) diagnosis during 42 endovascular operations (40 balloon angioplasties and 2 stentings) on leg arteries with stenoocclusive changes. The operations were successful in 95.5% of cases, as evidenced by angiography and in 83.3% as shown by CDS; there were 7 (16.7%) cases of hemodynamic events, such as angioplastic site thrombosis, intimal dissection, and residual stenosis. Late follow-ups revealed 6 (16%) cases of hemodynamic events, such as angioplastic area thrombosis and reocclusion. CDS permitted evaluation of the patency of leg arterial endovascular intervention areas, which was 44.96+/-19.77% at a follow-up of up to 60 months (mean follow-up period 18.76+/-2.58). Two stents were patent within 24 months.     

Prevention of microvascular thrombosis with low-dose tissue plasminogen activator.

In a blind, randomized study, two groups, each of seven rabbits, were treated with either a very low dose of human melanoma cell line-derived tissue-type plasminogen activator (t-PA) or isotonic saline. t-PA (0.067 mg/kg of body weight) was administered intraaortically, 20 percent being given as a 30-second "bolus" infusion just prior to the reperfusion of intimectomized central ear arteries and the rest as a continuous infusion during the next 2 hours. Arteriotomic bleeding times, accumulations of 32P-labeled platelets, patency, and sizes of thrombus deposits 2 hours after reperfusion were recorded. To confirm the presence of tissue plasminogen activator in plasma, fibrin-plate lysis assays of arterial plasma were performed immediately before and 1/2 hour and 2 hours after starting drug infusion. Arteriotomic bleeding times were similar in both groups. Transient "oozing" from wound edges occurred in 40 percent of rabbits treated with tissue plasminogen activator. Patency was significantly increased and thrombus deposits were smaller in the tissue plasminogen activator group. Plasma from animals treated with tissue plasminogen activator caused massive lysis of fibrin plates, whereas plasma from control animals caused little or no lysis. Platelet accumulations were very similar in both groups, indicating that occlusive thrombi mainly consisted of other elements than platelets (e.g., fibrin and red cells). Scanning electron microscopy showed normally adhering and aggregating platelets in both groups. This study shows that mild fibrinolytic stimulation with tissue plasminogen activator significantly improves patency in severely traumatized small-caliber arteries and indicates that such treatment may be one approach to prevent thrombosis at microvascular anastomotic sites.     

Reconstruction of rat femoral veins with microvascular prostheses

Synthetic conduits have not been suitable for microvascular reconstruction owing primarily to their high thrombogenicity. Vein replacements are the most vulnerable to thrombosis because of their low shear rates and low pressure. Experimental replacement of microvenous segments with prosthetic segments has shown little success. Recent technological advances in biomaterials and control of thrombogenesis provide the potential for success in the development of venous prostheses. The purpose of this study was to assess the use of nonbiodegradable composite polyurethane microvascular prostheses for reconstruction of rat femoral veins. Rat femoral venous defects of 10 mm were reconstructed with autogenous vein (n = 12), unprocessed plain polyurethane (n = 5), and nonbiodegradable composite polyurethane (n = 31). Patency was evaluated by direct observation and proximal venous milking tests. The patency rate of composite grafts was not significantly different from that of isotopic vein (p = 0.5, Fisher's exact test), and both had higher patency than unprocessed polyurethane (p less than 0.01). Composite grafts were examined sequentially using light and scanning electron microscopy. Grafts were fully endothelialized between the first and third months. The neointimal, neomedial, and neoadventitial layers could be seen more distinctly over time. New opportunities in reconstructive microsurgery may be opened by microvascular prostheses that are complaint and thromboresistant.     

An absorbable catheter system for use in microvascular and peripheral vascular surgery: an experimental study in the canine.

An absorbable catheter for use in regional anticoagulation in microvascular and peripheral vascular surgery was studied in 20 sites in 10 adult beagle dogs to answer three questions: (1) Could the polyglycolic acid and trimethylene carbonate catheter withstand intraarterial pressures of infusion and completely absorb over a predictable time interval? (2) Could the catheter be filled with heparin and maintain patency for reuse after a 24-hour interval? (3) Could the catheter be placed in a side branch of a major artery and, after catheter dissolution, maintain long-term patency of the primary feeding artery? The catheters were completely absorbed from 24 to 34 weeks following implantation. The catheters were able to withstand intraarterial pressures, and no evidence of significant thrombosis of the primary feeding artery was seen in any animal studied. No complications of catheter leak, hematoma formation within the catheter placement sites, or sepsis were noted in any of the 20 catheter sites studied.     

The thigh flap: an osteomyocutaneous free-flap model in the rat

A new experimental model for free-flap transfer has been developed in the rat. This "thigh flap" is an osteomyocutaneous free flap of bone (femur), muscle (thigh), and skin (groin) based on the femoral vessels. The flap is harvested from the left groin and thigh of an inbred female rat and is transferred to a subcutaneous pocket in the left groin of a male rat of the same inbred strain. The femoral vessels supplying the flap are anastomosed end-to-end with the femoral vessels of the recipient. Thirty flaps have been transferred, with 5 technical failures. Three of the remaining 25 flaps developed necrosis within 24 hours. The other 22 flaps remained viable until the rat was sacrificed at 7 days. The survival rate of the thigh flap was thus 88 percent. The model is suitable for use in metabolic, vascular, and immunologic studies of composite free flaps.