Giorgio De Santis,
and Bernard McC. O'Brien
THE DELETERIOUS EFFECT
OF ARTERIOVENOUS FLOW
REVERSAL DURING
EXPERIMENTAL FREE
MUSCLE TRANSFER
ABSTRACT
Arteriovenous flow reversal (AVR) has been used experimentally to salvage ischemic limbs and to create novel skin flaps with some success. The clinical applicability of AVR in muscle by way of two arteriovenous anastomoses in the rabbit was investigated. Twenty-four rabbits were divided into two groups. In Group 1 (control), the rectus femoris muscle was harvested and transplanted in the opposite thigh, anastomosing the donor femoral artery to the recipient femoral artery, and the donor rectus femoris vein to the recipient femoral vein. In Group 2 (flow reversal), the same procedure was done except the donor artery was anastomosed to the recipient vein and vice versa. Six and 24 hr postoperatively, specimens were compared macroscopically and by weight and histology. Reversed flow muscles were significantly heavier than control muscles at 6 hrand at 24 hr. Histologically, 6 hr of AVR caused edema, intramuscular hemorrhage, neutrophil infiltration, and thrombosis of most vessels. By 24 hr muscle cell degeneration was well advanced. All control muscles were viable, with only mild edema and slight peripheral necrosis. Possible reasons for the failure of AVR in muscle are discussed. On the basis of these results, AVR in free muscle transfer is not advocated.
The technique of arteriovenous flow reversal (AVR) has been the subject of research for over 80 years. Its first application was in experimental1 and clinical2
attempts to salvage ischemic limbs. The most success-ful recent experimental results have involved staged AVR by end-to-side arteriovenous anastomoses in the femoral or popliteal vessels. Amir-Jahed3 found that for
one-stage AVR by end-to-end anastomoses to suc-cessfully nourish the distal leg, vascular tone must be diminished by lumbar sympathectomy. Other ischemic limb studies4-8 using H-type or end-to-side
anasto-moses as a method of staging AVR were generally
successful, but doubts still exist concerning the clini-cal applicability of AVR in ischemic limb salvage9
In skin flaps, AVR has been studied by several groups. AVR in loose-skinned animals (rats,1011
rab-bits12) achieved approximately 95 percent survival
with or without staging. Most recently, however, Ger-mann et a!.13 using a porcine skin flap model, found
that of 12 AVR flaps elevated, none survived beyond 48 hr, casting doubt on the clinical applicability of this technique.
Limb ischemia AVR models can take advantage of extensive subcutaneous arteriovenous
communica-367
Microsurgery Research Centre, St. Vincent's Hospital, Melbourne, Australia Reprint requests-. Mr B.McC. O'Brien, Microsurgery Research Centre, St. Vincent's Hospital, 41 Victoria Parade, Fitzroy, Melbourne 3065, Australia Accepted for publication May 29, 1989 © 1989 by
JOURNAL OF RECONSTRUCTIVE MICROSURGERY/VOLUME 5, NUMBER 4 OCTOBER 1989
tions that allow blood to pass into the normal arterial supply of muscle.14 But AVR in skeletal muscle alone
has not been studied. During a clinical free muscle transfer, spatial factors and pedicle lengths may make it technically easier to do two arteriovenous moses rather than one arterial and one venous anasto-mosis. In this study using the rabbit rectus femoris muscle, a well-defined experimental model,15 the
fea-sibility of AVR in skeletal muscle was investigated.
MATERIALS AND METHODS
Twenty-two Australian outbred rabbits weighing 2.0 to 3.0 kg were used. Anesthesia was induced by sodium pentobarbital (30 mg/kg) and maintained by oxygen, nitrous oxide, and halothane.
Surgical Procedure. After a vertical skin incision,
the aponeurosis between the vasti muscles was in-cised, after which the rectus femoris muscle was clearly visible and was bluntly separated from the sur-rounding tissue. Its distal tendinous insertion was se-vered interiorly and dissection continued superiorly. Perforators to the vasti muscles were cauterized. The femoral vessels were then defined. The vascular pedicle to the rectus femoris muscle arises from the femoral vessels between the profunda and epigastric vessels. Once the neurovascular pedicle had been defined, the motor nerve was divided and the tendinous attach-ments of the muscle on its superior and deep margins were sharply divided. This totally freed the muscle on its vascular pedicle. The proximal femoral artery and distal rectus femoris vein were used for subsequent anastomoses (Fig. 1). Femoral vessels in the opposite leg were exposed.
Group 1 (Control). The muscle was harvested,
weighed, and transferred to the opposite thigh (Fig. 2A). The donor artery was anastomosed end-to-end to the recipient femoral artery and donor vein to the recipient femoral vein in a standard fashion. The mus-cle was evaluated at 6 hr (n = 4) and 24 hr (n = 4) postoperatively.
Group 2 {Flow Reversal). The muscle was harvested,
weighed, and transferred to the opposite thigh. The donor vein was anastomosed to the recipient femoral artery. Once venous blood was noted to flow from the donor artery, it was anastomosed to the recipient vein (Fig. 2B). The muscle was evaluated at 6 hr (n - 7) and 24 hr (n = 9) postoperatively.
Evaluations. Muscles were weighed at initial
ele-vation from their donor site and at the time of explora-tion, when patency of the vessels and macroscopic appearance of the muscle were also assessed. Speci-mens were fixed in buffered formalin and processed for histology using standard techniques. Histologic
sec-Figure 1. Schematic diagram of the isolated rectus femoris experimental model. The donor artery is the femoral and the donor vein is the rectus femoris.
tions of the specimens were blindly reviewed, espe-cially noting tissue necrosis and edema.
Student's t-tests were used for statistical analysis.
RESULTS
All control muscles were viable with patent anas-tomoses. At 6 hr, all reversed flow muscles appeared viable although edematous, and all anastomoses were patent. Congestion worsened over time. Thirty-three percent of the reversed flow anastomoses were throm-bosed at 24 hr.
Reversed flow muscles were significantly heavier than corresponding control muscles at 6 hr (9.0 g ± 0.2 vs 7.1 g ± 0.2, p < 0.01) and 24 hr (9.8 g ± 0.5 vs 7.3 g ± 0.6, p < 0.01) (Table 1). There was no significant ence between the muscle weights after the two differ-ent times of AVR.
Histologically, all control muscles were normal apart from mild edema and some slight peripheral necrosis, probably due to surgical handling (Fig. 3). Reversed flow muscles at 6 hr displayed thrombosis of most vessels, along with neutrophil infiltration, intra-muscular hemorrhage, edema and evidence of muscle cell degeneration (Fig. 4). By 24 hr, all vessels appeared thrombosed, massive hemorrhage had occurred, and necrotic muscle fibers were present throughout the muscle (Fig. 5).
368
REVERSE FLOW
CONTROL
A ^ — * r ^ * - ^ B Figure 2. Schematic diagrams of test groups. A, Control: artery anastomosed to artery and vein to vein. B, Reversed flow: recipient femoral artery anastomosed to rectus femoris vein, and donor femoral artery anastomosed to femoral vein.
Table 1. Comparison of Muscle Weights
Experimental Croups 1. Control 2. Flow reversal 0 hr 6.17 ± 0.24 (n = 8) 6.06 ± 0 . 1 4 (n = 16) NS Biopsy time 6 hr 7.10 ± 0.24 (n = 4) 9.00 ± 0 . 1 8 (n = 7) p < 0.01 2 4 hr 7.32 ± 0.55 (n = 4) 9.79 ± 0.48 (n = 9) p < 0.01
DISCUSSION
Rectus femoris muscle weights (g) (mean ± SEM) in control and flow reversal groups measured before the experimental period and postoperatively at 6 or 24 hr. Levels of significance are included for comparisons between control and flow reversal groups at 0, 6 and 24 hr (NS = not significant).
: ; ; : ; .
-Figure 3. Histology of control muscle 24 hr postopera-tively. The muscle is predominantly normal and healthy, except for some slight edema. (H & E, 60x)
AVR has been successful in skin flaps of loose-skinned animals10^12 and in experimental salvage of
ischemic limbs,4-8 but it has not been attempted in a
free muscle flap. The results from this study show that one-staged, end-to-end AVR in the rectus femoris mus-cle caused tissue death, similar to the effects of AVR in pig skin flaps.13 Six hours postoperatively,
intramuscu-lar edema, hemorrhage, and neutrophil infiltration had occurred, and eventually muscle fiber necrosis was widespread. There are several possible explanations for these results.
Factors such as a lack of intramuscular arterio-venous communications,14 constricted precapillary
sphincters,3 and venous valves16 may have acted as
barriers to reversed blood flow. The resultant lack of venous drainage would have caused ischemic muscle damage.
Spence et a!.,14 while demonstrating significant
subcutaneous arteriovenous anastomoses in the ca-nine hind limb, failed to demonstrate any in the skele-tal muscles. Therefore, if blood was entering the mus-cle by way of the vein, the only available route would be through the capillaries and the precapillary sphincter into the arterial system. The precapillary sphincter has been shown to constrict in response to high venous pressure,17 and so in this experimental situation, its
resistance would be great.
. . . " ••",
JOURNAL OF RECONSTRUCTIVE MICROSURGERY/VOLUME 5, NUMBER 4 OCTOBER 1989
lems that could be investigated. If in muscle AVR is possible only in a true retrograde direction to normal blood flow, then the precapillary sphincter would need to be overcome. Abolishing the smooth muscle vascu-lar tone within the rectus femoris by local sympathec-tomy3 would allow easier passage of blood through the
precapillary sphincter.
Successful salvage of ischemic limbs has gener-ally involved a staged procedure, either by end-to-side or H-type arteriovenous anastomoses.4"8 This allows
venous valves to be overcome and collateral circula-tions to develop, while simultaneously allowing blood flow in the normal direction. A combination of these two techniques may improve the chances of survival of the reversed flow muscle. It is also conceivable that muscle survival might be improved in a model with a second vein available for venous drainage or with the draining vessel spatially removed from the inflow ves-sel.1011
Further investigation is required to determine the microvascular nature of the failure of AVR in skeletal muscle. But, on the basis of this study, one-stage arte-riovenous flow reversal in muscle is to be discouraged. Such steps as staging the procedure and securing adequate drainage through another vein could be pur-sued for scientific reasons. Nonetheless, at this time it appears that the clinical applicability of the procedure is doubtful.
Figure 4. Reversed flow muscle, six hours postopera-tively. Hemorrhage, thrombosis and degeneration of muscle fibers have occurred. (H & E, 60x)
Figure 5. Reversed flow muscle, 24 hr postoperatively. The degree of hemorrhage and thrombosis has increased. Necrotic muscle fibers are present throughout. (H & E, 60 x )
REFERENCES
370
In canine hind limb AVR,478 venous valves are
rendered incompetent by arterial pressure, at least in large veins. This may happen to intramuscular valves also. Partially competent venous valves can act as sites for microthrombus formation,13 which in this study
may have been another obstacle to blood flow. No valves were observed in the rectus femoris vein be-tween its origin at the muscle and the anastomosis, but more study is required to define their presence within the muscle.
The blood flow obstruction caused an increase in intramuscular pressure from edema and hemorrhage. The resultant ischemic compartment syndrome18 was
responsible for the rapid degeneration of the muscle fibers. Skin is much more resistant to ischemia than muscle19 and more time is available for alternative
circulation patterns to develop before permanent tis-sue damage occurs. This is another factor in the greater success of AVR in skin and whole limb experiments, compared to one-stage AVR in skeletal muscle.
There are some possible solutions to these
prob-10
Carrel A, Guthrie CC: The reversal of the circulation in a limb. Ann Surg 43:203, 1906
San Martin y Satrustegui A: Chirurgie de l'appareil circulatoire: L'anastomose arterio-veneuse. Sem Med 22:395, 1902 Amir-lahed AK: Revascularization of lower extremities by
rever-sal of blood flow with and without lumbar sympathectomy: An experimental study. Surgery 59:243, 1966
Gerard DF, Gausewitz SH, Dilley RB, Bernstein EF: Acute physi-ologic effects of arteriovenous anastomosis and fistula in revascularizing the ischemic canine hind limb. Surgery 89: 485, 1981
Graham AM, Sniderman A, Jothy S, el a!.: Staged reversal of venous flow for revascularization of the severely ischemic limb. I Surg Res 35:11, 1983
lohansen K, Bernstein EF: Revascularization of the ischemic canine hind limb by arteriovenous reversal. Ann Surg 190: 243, 1979
Root HD, Cruz AB, |r: Effects of an arteriovenous fistula on the devascularized limb. IAMA 191:645, 1965
Matolo NM, Cohen SE, Wolfman EF Ir Use of an arteriovenous fistula for treatment of the severely ischemic extremity: Experimental evaluation. Ann Surg 184:622, 1976 Warren R: Can the venous system be made to act in situ for the
arterial system"5 Arch Surg 112:1238, 1977
Nakayama Y, Soeda S, Kasai Y: Flaps nourished by arterial inflow through the venous system: An experimental inves-tigation Plast Reconstr Surg 67:328, 1981
Voukidis T: An axial pattern flap based on the arterialised venous network: An experimental study in rats. Br I Plast Surg 35:524, 1982
12. Ji SY, Chia SL, Cheng HH: Free transplantation of venous net- 16. Watterson PA, Taylor Gl, Crock |G: The venous territories of work pattern skin flap: An experimental study in rabbits. muscles: Anatomical study and clinical implications. Br I Microsurgery 5:151, 1984 Plast Surg 41:569, 1988
13. GermannGK, Eriksson E, Russell RC, Mody N: Effect of arterio- 17. Cohen SE, Matolo NM, Wolfman EF: Arteriovenous fistula for venous flow reversal on blood flow and metabolism in a revascularization of the ischemic extremity. Vase Surg 10: skin flap. Plast Reconstr Surg 79:375, 1987 238, 1976
14. Spence RJ, Rhodes BA, Wagner HN ]r. Regulation of arterio- 18. Heppenstall RB, Scott R, Sapega A, etal.-. A comparative study of venous anastomotic and capillary blood flow in the dog leg. the tolerance of skeletal muscle to ischemia: Tourniquet Am 1 Physiol 222:326, 1972 application compared with acute compartment syndrome 15. Frey M, Gruber H, Havel M, el al.: Experimental free-muscle I Bone It Surg 68A:820, 1986
transplantation with microneurovascular anastomoses. 19. Strock PE, Majno G: Vascular responses to experimental tour-Plast Reconstr Surg 71:689, 1983 niquet ischemia. Surg Gynecol Obstet 129:309, 1969
