- Email: [email protected]
Pre-arterialisation of the arterialised venous flap: an experimental study in the rat
British Journal of Plastic Surgery (2001), 54, 621~630
9 2001 The British Association of Plastic Surgeons doi: 10.1054/bjps.2001.3675
BRITISH
JOURNAL
OF
[ ~ ]
PLASTIC
SURGERY
Pre-arterialisation of the arterialised venous flap: an experimental study in the rat B. Wungcharoen,W. Pradidarcheep*,Y. Santidhananon and V. Chongchet Plastic and Reconstructive Surgery Unit, Department of Surgery, Bangkok Metropolitan Administration Medical College and Vajira Hospital; and *Department of Anatomy, Faculty of Medicine, Srinakharinwirot University, Bangkok, Thailand SUMMARY. Arteriovenous fistulae cause haemodynamic and morphological changes to the local venous channels. We have used the concept of preformed arteriovenous fistulae to study the viability improvement of arterialised venous flaps. Five groups of flaps were created using the abdominal skin of the Wistar rat (n = 10 in each group) with a silastic sheet implanted underneath. Group 1 (control) contained a flap without a vascular supply, group 2 (venous perfusion flap) contained a single pedicled skeletonised vein and a draining vein, and group 3 (arterialised venous flap) contained an arteriovenous shunt proximal to the single pedicled skeletonised vein and a draining vein; in group 4 (7 day pre-arterialised flap) the arteriovenous shunt was performed 7 days before the flap was raised in the same procedure as group 3, and in group 5 (14 day pre-arterialised flap) the arteriovenous shunt was performed 14 days before the flap was raised. The surviving surface areas of the flaps in each group, assessed 7 days after raising, were 0%, 22.21%, 54.32%, 62.21% and 97.47%, respectively. There was a statistically significant difference in survival between venous perfusion flaps and arterialised venous flaps (P < 0.05). Only the 14 day pre-arterialised flaps had a statistically significantly larger area of survival than arterialised venous flaps (P < 0.05). Microangioarchitecture of the pre-arterialised group, studied by the microvascular corrosion-cast technique combined with scanning electron microscopy and transmission electron microscopy, revealed dilatation of veins, numerous small neo-vessels and a decrease in or total absence of functioning valves. We conclude that 14-day pre-arterialisation in the rat model improved the survival of arterialised venous flaps by increasing collateral pathways for arterialised blood flow through the flap. 9 2001 The British Association of Plastic Surgeons
Keywords: pre-arterialisation, arterialised venous flap, arteriovenous shunt, rat, flap survival, microangioarchitecture.
A venous flap is a non-conventional vascular perfusion skin flap, through which the main transit of blood flow is the venous system. It can be perfused by either venous or arterialised blood. Arterialised venous flaps have been shown to be more reliable than venous perfusion flaps both in experiments and in clinical studies. 1-3 Small arterialised venous flaps can be successfully employed clinically.4 A number of ways to improve the survival of large arterialised venous flaps have been studied, because such flaps tend to give less-favourable results. 5-8 These studies have investigated adequate draining veins from the venous flap, 9,1~the size of the venous network in the centre of the flap 1~ and neovascularisation from the recipient bed. 12 All of these factors are likely to limit the application of these flaps or lead to unfavourable outcomes if the degree to which they are present is suboptimal. Nakayama et al first described, using rat models, successful large arterialised venous flaps in which the primary arteriovenous anastomosis of the feeding vein was performed 2 weeks prior to flap harvest. 13 Although this was not a controlled experiment, comparing results
with conventional arterialised venous flaps, it demonstrated the possible improvement in venous-flap survival. The results of the authors' previous study of pre-arterialisation in clinical cases also support this concept. 14 We designed this study to investigate the effects of pre-arterialisation on the survival of arterialised venous flaps in an animal model. Pre-arterialisation was achieved by preforming an arteriovenous fistula of the vein within the flap for different periods of time.
Materials and methods Animal model and design of the flaps In this study, we used 50 male Wistar rats weighing between 300 g and 350 g. General anaesthesia was induced by injection of intraperitoneal sodium pentobarbital (30 mg/kg). The ventral abdominal wall was shaved using an electric shaver, and prepared with 70% chlorhexidine in alcohol. The outline of the flap was a rectangle on the fight side of the abdominal wall measuring 4 cm in length and 3 cm in width. The medial border of the flap was the axis between the xiphoid process and the symphysis pubis. The lower border of the flap was the axis formed by a perpendicular line from the midline axis to the
Presented at the 25th Annual Congress of the Royal College of Surgeons of Thailand, 14-17 July 2000.
62l
622
British Journal of Plastic Surgery
lateral concavity between the abdomen and the leg (Fig. 1). The femoral vessels were located underneath the medial portion of the lower border of the flap. All surgical procedures were conducted using aseptic techniques.
Experimental groups Five groups of rats (n = 10 in each group) were used in the study (Fig. 2, Table 1). The rats were randomly assigned to groups.
Group 1 (control group). The skin flap was raised without a vascular supply. A silastic sheet was placed under the flap. The skin flap was immediately sutured back into the donor site. Group 2 (venous perfusion venous flap). The skin flap was elevated fed only by the skeletonised right inferior epigastric vein. The inferior epigastric artery was ligated. Outflow was provided by the long thoracic vein at the upper lateral border of the flap. A silastic sheet was inserted under the flap and the flap was sutured back into the donor site.
Long thoracic vein.";
Group 3 (arterialised perfusion venous flap). A sideto-side arteriovenous anastomosis was performed between the fight femoral artery and the fight femoral vein just proximal to the origin of the inferior epigastric vein. The skin flap was elevated fed by the skeletonised right inferior epigastric vein. The inferior epigastric artery was ligated. Outflow was provided by the long thoracic vein. A silastic sheet was inserted under the flap and the flap was sutured back into the donor site. Group 4 (7 day pre-arterialised venous flap). Prearterialisation was performed through a 1 cm skin incision at the medial lower border of the planned flap. A sideto-side anastomosis was performed between the right femoral artery and the right femoral vein proximal to the origin of the inferior epigastric vein. The incision was closed with simple interrupted sutures using 5/0 nylon. Seven days after the anastomosis, the flap was raised based on the skeletonised right inferior epigastric vein, with the long thoracic vein acting as the draining vein. A silastic sheet was inserted and secured underneath the flap and the flap was sutured back into the donor site. Group 5 (14 day pre-arterialised venous flap). Prearterialisation was performed in the same way as in group 4. The flap was raised 14 days after the arteriovenous shunt was performed. The flap was supplied by the completely skeletonised right inferior epigastric vein and drained by the long thoracic vein. A silastic sheet was inserted under the flap and the flap was sutured back into the donor site.
A-Vshunt
,,' immediate
raising
v~n Group 1
Group 2
Group 3
Inferior WS.
~ ~ A - V
~L 7 day interval
~
~
shunt
lap harvest
~
Group 4
Figure 1--Animal model for the experiment. Schematic illustration of flap location.
Table 1
i ~ ~ A-Vshunt 14 day Interval
Flap harvest
Group 5
Figure 2---Schematic illustration of the five flap groups: group 1, control group; group 2, venous perfusion venous flap; group 3, arterialised perfusion venous flap; group 4, 7 day pre-arterialised venous flap; group 5, 14 day pre-arterialised venous flap.
Experimental protocol
Group
Day 0
Day 7
1. control (n = 10) 2. venous perfusion (n = 10) 3. arterialised perfusion (n -- 10) 4.7 day pre-arterialised (n = 10) 5.14 day pre-arterialised (n = 10)
flap elevation flap elevation arteriovenous shunt, flap elevation arteriovenous shunt arteriovenous shunt
flap assessment flap assessment flap assessment flap elevation
Day 14
Day 21
flap assessment flap elevation
flap assessment
Pre-arterialisation of the arterialised venous flap
Surgical technique The panniculus carnosus was included with the skin flap in all rats. The silastic sheets inserted in all groups measured 4.5 x 3.5cm and were secured in place using quilted 5/0 nylon sutures at all borders. All skin flaps were sutured back into the donor sites using simple interrupted 5/0 nylon sutures. Side-to-side anastomosis was performed by making 1 mm slit incisions with a number 11 blade on adjacent sides of the artery and vein, and completed with eight interrupted 10/0 nylon sutures. The retrograde flow of arterial blood through the fistula was tested by temporarily occluding the femoral artery proximal to the shunt. When the femoral artery was occluded by microforceps, the inferior epigastric vein looked cloudy blue. When the artery was released, the inferior epigastric vein turned bright red and markedly dilated due to the arterial blood flow. All nerves and other vessels attached to the flap except for the inferior epigastric vein and the long thoracic vein were cauterised and transected. The bleeding sites were controlled with bipolar cauterisation.
Exclusion criteria Rats were excluded from the study if the vascular pedicle thrombosed, the patency of the arteriovenous shunt could not be demonstrated during flap harvest or the animal died during the experiment.
Flap assessment The assessment included survival assessment, microangioarchitecture and histological studies.
Survival assessment.
All rats were observed for 7 days after raising the flap. They were then anaesthetised before the assessments. The surviving parts of the flaps were closely traced onto translucent plastic sheets. Surviving parts were defined as areas without any necrosis. The surface area traced was calculated by scanning the sheet, along with a standard surface-area sheet, using a graphic computerised scanner. The pixels in the scanned picture were then counted using histogram analysis in Adobe Photoshop v. 5.0 software. The number of pixels was converted into surface area by comparison with the standard surface sheet. All mapping was performed blindly.
Microangioarchitecture study. The microvasculature was studied by transmission electron microscopy and, threedimensionally, by a vascular corrosion cast in conjunction with scanning electron microscopy. Six rats from each group were studied immediately after the survival assessments. Transmission electron microscopy was performed after sacrifice of the animals, but vascular corrosion casting was performed while the animals were still alive. For transmission electron microscopy, surviving areas of flaps were randomly selected and fixed by immersion in 2.5% glutaraldehyde in cacodylate buffer, pH 7.4, and left overnight in the same fixative at 4 ~ The samples were post-fixed in 1% osmium tetroxide, dehydrated in a graded series of ethanols, infiltrated and embedded in
623 Araldite 502 resin. Semi-thin sections (1 p~m) were stained with 1% toluidine blue and examined under a light microscope; 70-100nm sections were stained with 1% uranyl acetate, then lead citrate, and examined under a transmission electron microscope at an accelerating voltage of 80 kV. For vascular corrosion casting, the animal preparation, the ventricular injection of Batson's number 17 plastic mixture (casting medium) and the preparation for obtaining the skin-flap vascular casts were as according to the method previously described by Promwikorn et al. 15 After corroding the soft tissue with 40% KOH, the remaining vascular cast was stuck onto a brass stub with silver paint, and coated with gold-palladium prior to being examined and photographed under a scanning electron microscope at accelerating voltages of 10kV and 15 kV. Plastic mixtures, which were infused through the left ventricle while the animal was still alive, travelled along the femoral artery and passed through the arteriovenous fistula. From the fistula, the infusate was distributed through the inferior epigastric vein to the whole of the surviving flap. The casts represent the venous microarchitecture of the venous flaps. Scanning electron microscopy of these vascular casts revealed the detailed architecture of the inner walls.
Histological study. The remaining rats from each group were used for histological analysis. Mid-portions of the surviving parts of the flaps were harvested for histological study and stained with haematoxylin and eosin. These sections were examined under a light microscope. Sections were obtained from four rats in group 1, three rats in each of groups 2, 3 and 5 and two rats in group 4.
Statistical analysis A Median Test was used to analyse the differences in survival between groups 2, 3, 4 and 5 using SPSS v. 9.0 software. Mann-Whitney U-tests were used for post-hoc comparisons between groups 2 and 3, groups 3 and 4, groups 4 and 5, and groups 3 and 5. Statistical significance was defined as P < 0.05.
Results We were able to assess survival in 45 out of the 50 rats. Five rats were excluded from the experiment: one rat in group 2 died, the anastomosis failed in one rat in each of groups 3, 4 and 5, and the pedicle thrombosed in one rat in group 4.
Results of the survival assessment All flaps in the control group necrosed. The mean surviving area in the venous-perfusion group was 22.21%. The mean surviving area in the arterialised perfusion group was 54.32%. The mean surviving area of the arterialised venous flaps that were pre-arterialised for 7 days was 62.21%. The arterialised venous flaps that were pre-arterialised for 14 days showed the highest surviving area, with a mean of 97.47% (Table 2, Figs 3 and 4). Statistical
624
British Journal of Plastic Surgery Table 2
Survival areas in each group
Group
Mean surface-area survival (cm 2)
1 ( n = 10) 2 (n = 9) 3 (n = 9) 4 (n = 8) 5 (n =9)
0.00 2.67 6.52 7.46 11.70
s.e.m.
-
0.25 0.69 0.79 0.10
13
v
0.00
22.21 54.32 62.21 97.47
Groups compared
'Z
Asymptotic significance
--3.576 --0.866 -3.595 --3.486
<0.05* 0.386 <0.05* <0.05*
10 8
m
-
2.09-3.24 4.92-8.11 5.59-9.34 11.46-11.93
12
2 and 3 3 and 4 3 and 5 4 and 5
n," f/} < ILl
Mean surface area survival (%)
Table 3 Comparisons of mean survival areas between groups using the Mann-Whitney U-test
14'
X
95% confidence interval of the mean
6 4
2
I1" 2~ if)
0
*Significant difference.
-2 N=
1
2
3
4
5
EXPERIMENTAL GROUP
Figure 3--Box plot showing the surviving surface areas of the flaps in each group. The box length is the interquartile range (Q3-Q1). The dark line within each box is the median (Q2). The asterisk indicates the extreme ( > Q 3 + 3 ( Q 3 - Q O ) and o indicates outliers (between 1.5(Q3-Q1) and 3(Q3- Q1)).
Figure 5--Stereomicrograph of vascular corrosion cast before coating with gold-palladium. This is an example of a group 4 vascular corrosion cast.
Figure 4--An example of a 14 day pre-arterialised venous flap (group 5). Note that survival of the flap is nearly complete. Dilated tortuous superficial veins are visible throughout the flap.
analysis by M e d i a n Tests revealed significant differences b e t w e e n groups 2, 3, 4 and 5 (median = 6.58, ) 2 = 19.487, asymptotic significance less than 0.05). M a n n - W h i t n e y U-tests showed significant differences in survival b e t w e e n the venous perfusion group and the arterialised perfusion group, b e t w e e n the arterialised perfusion group and the 14 day pre-arterialised group, and b e t w e e n the 7 day pre-arterialised group and the 14 day pre-arterialised group (asymptotic significance less than 0.05) (Table 3). There was no significant difference between the arterialised group and the 7 day pre-arterialised group (asymptotic significance of 0.386).
Pre-arterialisation of the arterialised venous flap
Result of the microangioarchitecture study Vascular corrosion casting could be conducted only in groups 3, 4 and 5 (Fig. 5) because the vessels in the control group and the venous perfusion group were all thrombosed. On gross examination of the casts, the venous pattern of conventional arterialised venous flaps (group 3) was a homogeneously distributed network with straight main veins. The 7 day pre-arterialised group (group 4) showed dilated tortuous main veins with a heterogeneous distribution of venous networks. In the 14 day pre-arterialised group (group 5) the dilated tortuous interconnecting venous network was prominent throughout the whole flap. Under scanning electron microscopy, group 3 (Fig. 6) revealed a normal inner microarchitecture of the main small collecting veins (100-300lxm), 16 consisting of indentations along the smooth inner vascular wall. These indentations were endothelial nuclear imprints. The tributaries of these collecting veins were muscular venules (50-100txm) and collecting venules (30-50 lxm). These venules were uniformly distributed and gradually decreased in size.
625 The postcapillary venules (8-30 txm) and venous capillaries ( < 8 txm) were sparse. Venous valves were usually found where venules branched off from the collecting vein. In group 4 the collecting veins became dilated and tortuous. The inner architecture of the collecting veins disappeared and was replaced with rough wrinkled surfaces (Fig. 7). Small vessels, including venous capillarysized and postcapillary venule-sized vessels, were found immediately branching off from the collecting vein. Venous valves were still seen, but only scarcely. The microarchitecture of group 5 was distinct (Fig. 8). The collecting veins, including the muscular venules and the collecting venules, were markedly distended and tortuous. The inner walls of these dilated vessels were rough and velvety with wrinkled surfaces. Numerous small vessels (venous capillary-sized vessels and postcapillary venule-sized vessels) sprouted from the collecting veins and the venules. They ran along or angled away from the collecting veins and venules in every direction. These small vessels formed vascular plexuses, interconnecting the dilated collecting veins and venules. There were no
Figure 6--Scanning electron micrograph of a vascular cast from group 3. (A) A venous network with straight smooth collecting veins ( 100-300 Fun) and venules gradually decreasing in size. There are very few venous capillaries ( < 8 ~um) and postcapillary venules (8 30 ~m) connecting the larger vessels. (B) Indentations on the smooth surfaces of the collecting veins represent the imprints of nuclei of endothelial cells. Venous valves, which were distributed throughout the network, can be seen where 9enules branch off from the collecting vein.
Figure 7--Scanning electron micrograph of a vascular cast from group 4. (A) Occasional venous valves are seen at the branching sites of venules from the collecting vein. (B) A few small vessels (venous capillary-sized vessels and postcapillary venule-sized vessels) abruptly branch off from the collecting vein. The normal nuclear indentations on the inner walls of the collecting veins are lost and replaced by a rough wrinkled surface. Compared with group 3, the collecting veins are tortuous and very dilated.
626
British Journal of Plastic Surgery
Figure 8--Scanning electron micrograph of a vascular cast from group 5. (A) Numerous small vessels (venous capillary-sized vessels and postcapillary-sized vessels) sprout from the collecting veins and venules in a multidirectional pattern. These small vessels form the vascular plexus interconnecting the collecting veins and venules. (B) The collecting vein has become markedly dilated and tortuous. The nuclear indentations seen in normal veins have been lost and replaced by a rough wrinkled surface. Abundant venous capillary-sized vessels branch off from these dilated veins.
Figure 9--Light micrograph showing a semi-thin section of a vein from a group 3 specimen. The endothelial cells are flat, elongated and scarce, with heterochromatic nuclei: va, venous valve (toluidine blue, • 70).
Figure 10--Light micrograph showing a semi-thin section of a vein from a group 5 specimen. The endothelial cells (e) are numerous, with euchromatic nuclei: *, internal elastic lamina ( x 90).
demonstrable valves in any specimen in this group. There was no demonstrable retrograde flow of the plastic mixture into arterioles or arteries in any of the groups. In the semi-thin section, the endothelial cells of group 3 (Fig. 9) demonstrated a flat elongated heterochromatic nucleus. There were fewer endothelial cells in the veins of this group than in the veins of groups 4 and 5. By contrast, the veins of groups 4 and 5 had many more endothelial cells (Fig. 10). Ultrastructural detail from transmission electron microscopy of these endothelial cells revealed active young cells with euchromatic nuclei (Fig. 11). The internal elastic lamina in the veins in these groups was discontinuous. There were also numerous young endothelial cells, which connected to form new vessels in the subcutaneous tissue in groups 4 and 5. Transmission electron microscopy of these areas demonstrated a neovascutarisation process. The active young endothelial cells with euchromatic nuclei and rich rough endoplasmic reticulum within the cytoplasm (Fig. 12) had encircled the newly formed lumina. This was seen in both the pre-arterialised groups, but was much more pronounced in group 5.
Histologicalfindings All the flaps in group 1 had necrosed from the skin through to the subcutaneous tissue. In group 2, the surviving parts of the flaps were composed of scanty small vessels and atrophic subcutaneous tissue. In group 3, histology showed oedematous subcutaneous tissue with dilated superficial veins (Fig. 13). In the pre-arterialised groups, the subcutaneous tissue was less oedematous but the superficial veins were massively dilated (Fig. 14). Numerous small vessels were scattered through all layers of the flaps.
Discussion As described by Nakayama et al, an abdominal skin flap in the rat can be supplied by arterial blood through the venous system. 13 In their study the arterialised venous flap was pre-arterialised, and the authors named it 'the delayed fS flap'. It showed a promising survival rate. Our study was designed as a controlled trial to test the hypothesis that pre-arterialisation of the venous flap
Pre-arterialisation of the arterialised venous flap
627
Figure 13--Light micrograph showing a section of the survivingpart of a group 3 flap. Oedematous subcutaneoustissue is prominent.Note the few dilated superficialveins (V) (haematoxylinand eosin, • 9).
Figure 11 Transmissionelectron micrograph of an endothelial cell from a vein in group 5. The euchromaticnucleus(N) is prominentin the endothelial cell, which projects into the lumen (L). The internal elastic lamina(*) forms a brokenzigzag line (scale bar: 2 ixm).
Figure 14~Light micrograph showing a section of the survivingpart of a group 5 flap. Huge dilated veins (V) and numerous small vessels are seen withinthe subcutaneoustissue. Compared with group 3, the subcutaneous tissue appears less oedematous (haematoxylinand eosin, • 9).
Figure 12~Transmission electron micrograph of the wall (tangential cut) of a newly formed blood vessel within the subcutaneoustissue of a group 5 specimen.Youngendothelialcells with euchromaticnuclei (N) are rich in rough endoplasmicreticulum(r) (scale bar: 2 p~m).
yields a better survival than can be obtained with the conventional arterialised venous flap. We used a side-toside anastomosis between the femoral artery and the femoral vein to perfuse the venous flap. Our pilot study using end-to-end anastomosis resulted in the postoperative
death of nearly all the animals. This is consistent with Nakayama et al's experiment, in which 13 out of 15 animals died after the end-to-end anastomosis of the femoral artery and the epigastric vein in the conventional arterialised venous flap. 13 We concluded that the side-to-side anastomosis of the femoral artery and the femoral vein, without the ligation of any main vessels, would not cause significant physiologic change, and most of the animals survived until the end of the experiment. The arteriovenous shunt, which was situated proximal to the epigastric vessels, could shunt arterial blood into the venous system of the flap, and its action could be demonstrated immediately after the anastomosis by observing dilated epigastric veins containing bright red blood. The shunt effect was confirmed later by vascular castings, which demonstrated the comparative microarchitecture of the venous systems in each flap. Indeed, side-to-side anastomosis was one of the classic procedures for vascular access in haemodialysed patients in the past, and has been proven to secure maximal fistula flow rate. 17 The insertion of a silastic
628 sheet underneath the flap aimed to exclude neovascularisation from the bed and the external effect of surrounding vessels, in both survival assessment and the vascularcasting study. Complete skeletonisation of the nutrient veins, including the transection of all nerves, reflected the actual survival of the arterialised venous flap, and thus demonstrated the effects of pre-arterialisation. Small vascular networks within the perivenous areolar tissue and nerve might unintentionally improve flap survival. 18'19 This is why the venous island flaps in early clinical reports had high survival rates, 2~ while the total venous perfusion free flap had comparatively poor survival. 3'24 The lower surviving area of the venous perfusion flap (group 2) (22.21%), in which the pedicle was completely skeletonised, confirmed this. The arterialised venous flap, which had a mean survival area of 54.32%, was statistically significantly superior to the venous perfusion flap (Mann-Whitney U-test between groups 2 and 3: Z = --3.576, asymptotic significance less than 0.05). This result runs contrary to some excellent outcomes in other experiments with arterialised venous flaps. 9,25 However, this may be due to the preclusion of plasmatic imbibition and early neovascularisation from the bed by the silastic sheet. Studies have confirmed significant roles of both factors. 26'27 The survival of small arterialised venous flaps was satisfactory in an early clinical report,4 but larger flaps have demonstrated less predictable survival. 5,s Neovascularisation may account for the lower survival rate of larger venous flaps: 12 small flaps are less dependent on neovascularisation from the bed because the marginal blood supply and/or plasmatic imbibition are adequate for survival. Apart from neovascularisation, a balanced blood flow, including arterialised inflow, through the centrally located venous network together with adequate draining veins are essential for survival, w These factors have been manipulated to improve survival in various recent experiments, which have included pre-fabrication of the venous flap, 28 expansion of the venous flap 29 and surgical delay of the venous flap, 30'31 including our pre-arterialisation experiment. Our model of pre-arterialisation aimed to improve survival by optimising the trade-off between sufficient oxygenated inflow and adequate outflow. The mean survival of the 14 day pre-arterialised flap (97.47%) was statistically significantly better than that of the conventional arterialised venous flap (Mann-Whitney U-test between groups 3 and 5: Z = -3.595, asymptotic significance less than 0.05). However, the mean survival of the 7 day prearterialised flap (62.21%) was not statistically significantly different from that of the conventional arterialised venous flap (Mann-Whitney U-test between groups 3 and 4: Z = - 0 . 8 6 6 , asymptotic significance of 0.386). This implies that pre-arterialisation influenced survival over the longer duration. Pre-arterialisation of the venous flap was achieved by using an arteriovenous fistula. So, the same haemodynamics and pathophysiology could be expected: shunts cause venous channels to dilate, and promote collateral venous channels. 16 This, in turn, leads to a decreased resistance of the venous network in the flap. Microangioarchitecture, demonstrated by vascular casting and scanning electron microscopy, also supported this hypothesis. Pre-arterialisation causes dilatation of the
British Journal of Plastic Surgery venous system, and stretches and expands the endothelium, which showed a rough wrinkled surface in both the 7 day and the 14 day pre-arterialised groups. The very distinct feature of the 14 day pre-arterialised group was the frequency of small vessels branching from the main veins. Although, in the vascular corrosion casting and scanning electron microscopy study, these small vessels could not be clearly identified either as neovascularisation or as existing collateral vessels opening under the high retrograde pressure, the semi-thin section and transmission electron microscopy study demonstrated numerous active young endothelial cells in the 14 day pre-arterialised group. Some of these cells were seen to be forming new lumen. This evidence supports angiogenesis within the flap. In the conventional arterialised venous flap, limited vascular channels may lead to inadequate oxygenation of the flap even if it is perfused by arterialised blood. Selection of a donor site that contains abundant venous networks would increase the survival rate of the arterialised venous flap 1~ because oxygenation would be improved, although relative hypoxia would still induce angiogenesis, 32 which is necessary for long-term perfusion. The small vessels that were demonstrated in the pre-arterialised groups increased the pathway for microcirculation. The absence of venous valves in the 14 day pre-arterialised group might result from the effect of retrograde pressure or from venous-channel dilatation. The absence of valves certainly resulted in the multidirectional flow of arterialised blood through the venous network and the newly formed vessels. The resistance in the flap then decreased and, consequently, the stagnation of blood also decreased. Oxygenation of the flap tissue was then easily accomplished. According to the experiment of Wolff et al, 33 the capillary system in all types of venous flap is reached by some of the inflowing oxygenated haemoglobin over the entire flap surface. The arterialised venous flap had poorer intracapillary oxygenated haemoglobin than the arterial flap, and the venous flow-through flap had a lower oxygen supply than the arterialised venous flap. Although very few capillary-sized vessels were demonstrated in conventional arterialised venous flaps (group 3) by vascular casting, the increased survival in the pre-arterialised venous flap groups supports this 'intracapillary haemoglobin oxygenation of venous flap' concept. Capillary-sized vessels were much more frequent in the pre-arterialised groups, especially in the 14 day pre-arterialised group, and the mean survival area increased. Although this microangioarchitecture study was descriptive, it could, in part, be used to explore the morphological changes of the venous structure within the venous flap. The hypothesis that pre-arterialised venous flaps have better survival is based on this evidence. The exact morphological and haemodynamic changes in different kinds of venous flaps need to be explored further. From the results of our experimental model in the rat, we draw the following conclusions. Without the influences of plasmatic imbibition and neovascularisation from the recipient bed: first, arterialised venous flaps had a statistically significantly better survival than venous perfusion venous flaps; second, pre-arterialisation improved the survival of arterialised venous flaps; third,
Pre-arterialisation of the arterialised venous flap only the 14 day pre-arterialised flaps showed statistically significantly better survival than conventional arterialised venous flaps; fourth, pre-arterialisation caused dilatation of the superficial veins, and abundant small vascular plexuses were found along the main collecting veins of the flap - functioning venous valves were absent or decreased after pre-arterialisation; and fifth, the microangioarchitectural changes resulting from pre-arterialisation might explain the increase in survival, in terms of a decrease in resistance to arterial blood flow, an increase in oxygenated blood exchange through numerous small vessels and plexuses, and a decrease in the volume of stagnant deoxygenated blood within the flap. These results are based on the rat flap model, which includes the panniculus carnosus and which is different from human skin. It is therefore not possible to extrapolate these findings directly to human skin, but it would seem appropriate to evaluate the application of pre-arterialisation in the clinical situation.
Acknowledgements The microvascular corrosion castings were conducted at the research laboratory of the Department of Anatomy, Faculty of Science, Mahidol University, Bangkok, Thailand. The authors thank Dr Reon Somana and all the research fellows there for their valuable cooperation. The electron microscopy examinations were conducted at the electron microscopy research laboratory of the Department of Anatomy, Faculty of Medicine, Srinakharinwirot University, Bangkok, Thailand. The pathology preparation and results were the contribution of Dr Somnuek Jesdapatarakul and Dr Nantana Kaewpila from the Department of Pathology, Bangkok Metropolitan Administration Medical College and Vajira Hospital, Bangkok, Thailand.
References 1. Suzuki Y, Isshiki N, Ishikawa K, Koyama H. Viability and quantitative dermofluorometry of experimental arterialised and non-arterialised venous flaps. Br J Plast Surg 1993; 46: 273-8. 2. Wolff K-D, Telzrow T, Rudolph K-H, Franke J, Wartenberg E. Isotope perfusion and infrared thermography of arterialised, venous flow-through and pedicled venous flaps. Br J Plast Surg 1995; 48: 61-70. 3. Chen H-C, Tang Y-B, Noordhoff MS. Four types of venous flaps for wound coverage: a clinical appraisal. J Trauma 1991; 31: 1286-93. 4. Inoue G, Maeda N, Suzuki K. Resurfacing of skin defects of the hand using the arterialised venous flap, Br J Plast Surg 1990; 43: 135-9. 5. Yilmaz M, Menderes A, Karata ~, Karaca C, Barut~u A. Free arterialised venous forearm flaps for limb reconstruction. Br J Plast Surg 1996; 49: 396-400. 6. Sakai S. Arterialised venous gtroin flap: case report. Br J Plast Surg 1996; 49: 90-2. 7. Klein C, Kowics A, Stuckensen T. Free arterialised venous forearm flaps for intraoral reconstruction. Br J Plast Surg 1997; 50: 166-71. 8. Inoue G, Suzuki K. Arterialized venous flap for treating multiple skin defects of the hand. Plast Reconstr Surg 1993; 91: 299-302. 9. lnada Y, Fukui A, Tamai S, Mizumoto S. The arterialised venous flap: experimental studies and a clinical case. Br J Plast Surg 1993; 46: 61-7. 10. Woo S-H, Kim S-E, Lee T-H, Jeong J-H, Seul J-H. Effects of blood flow and venous network on the survival of the arterialized venous flap. Plast Reconstr Surg 1998; 101: 1280-9.
629 11. Lee WPA. Arterialized venous flap for treating multiple skin defects of the hand. Plast Reconstr Surg 1993; 91: 303-6. 12. Pittet B, Chang P, Cederna E Cohen MB, Blair WF, Cram AE. The role of neovascularization in the survival of an arterialized venous flap. Plast Reconstr Surg 1996; 97: 621-9. 13. Nakayama Y, Soeda S, Kasai Y. Flaps nourished by arterial inflow through the venous system: an experimental investigation. Plast Reconstr Surg 1981; 67: 328-34. 14. Wungcharoen B, Santidhananon Y, Chongchet V. Pre-arterialisation of an arterialised venous flap: clinical cases. Br J Plast Surg 2001; 54: 112-16. 15. Promwikorn W, Thongpila S, Pradidarcheep W, Mingsakul T, Chunhabundit P, Somana R. Angioarchitecture of the coeliac sympathetic ganglion complex in the common tree shrew (Tupaia glis). J Anat 1998; 193: 409-16. 16. Rhodin JAG. Ultrastructure of mammalian venous capillaries, venules, and small collecting veins. J Ultrastruct Res 1968; 25: 452-500. 17. Sumner DS. Hemodynamics and pathophysiology of arteriovenous fistulae. In Rutherford RB, ed. Vascular Surgery. Philadelphia: W.B. Saunders, 1995: 1166-91. 18. Murata K, Tamai S, Inada Y, Fukui A, Miyamoto S. Transfer of a pedicled venous flap containing perivenous areolar tissue and nerve: an experimental study. Br J Plast Surg 1999; 52: 223-9. 19. Noreldin AA, Fukuta K, Jackson IT. Role of perivenous areolar tissue in the viability of venous flaps: an experimental study on the inferior epigastric venous flap of the rat. Br J Plast Surg 1992; 45: 18-22. 20. Chavoin JP, Rouge D, Vachaud M, Boccalon H, Costagliola M. Island flaps with an exclusively venous pedicle. A report of eleven cases and a preliminary haemodynamic study. Br J Plast Surg 1987; 40: 149-54. 21. Thatte MR, Kumta SM, Purohit SK, Deshpande SN, Thatte RL. Cephalic venous flap: a series of 8 cases and a preliminary report on the use of 99mTc labelled RBCs to study the saphenous venous flap in dogs. Br J Plast Surg 1989; 42: 193-8. 22. Thatte RL, Thatte MR. The saphenous venous flap. Br J Plast Surg 1989; 42: 399-404. 23. Inada Y, Fukui A, Tamai S, Kakihana T, Maeda M. The sliding venous flap for covering skin defects with poor blood supply on the lateral aspects of fingers. Br J Plast Surg 1991; 44:368-71. 24. Amarante J, Costa H, Reis J, Soares R. Venous skin flaps: an experimental study and report of two clinical distal island flaps. Br J Plast Surg 1988; 41: 132-7. 25. Nichter LS, Haines PC. Arterialized venous perfusion of composite tissue. Am J Surg 1985; 150: 191-6. 26. Fukui A, Tamai S, Maeda M, Inada Y, Mii Y, Mine T. The pedicled venous flap. An experimental study. Br J Plast Surg 1993; 46: 116-21. 27. Matsushita K, Firrell JC, Ogden L, Tsai T-M. Blood flow and tissue survival in the rabbit venous flap. Plast Reconstr Surg 1993; 91: 127-35. 28. Takato T, Komuro Y, Yonehara H, Zuker RM. Prefabricated venous flaps: an experimental study in rabbits. Br J Plast Surg 1993; 46: 122-6. 29. Mutaf M, Tasaki Y, Fujii T. Expansion of venous flaps: an experimental study in rats. Br J Plast Surg 1998; 51: 393-401. 30. Byun JS, Constantinescu MA, Lee WPA, May JW Jr. Effects of delay procedures on vasculature and survival of arterialized venous flaps: an experimental study in rabbits. Plast Reconstr Surg 1995; 96: 1650-9. 31. Cho B-C, Lee M-S, Lee J-H, Byun J-S, Baik B-S. The effects of surgical and chemical delay procedures on the survival of arterialized venous flaps in rabbits. Plast Reconstr Surg 1998; 102: 1134-43. 32. Stavri GT, Zachary IC, Baskerville PA, Martin JF, Erusalimsky JD. Basic fibroblast growth factor upregulates the expression of vascular endothelial growth factor in vascular smooth muscle cells: synergistic interaction with hypoxia. Circulation 1995; 92: 11-14. 33. Wolff K-D, Uekermann B, Matthes G, Wartenberg E. Intracapillary haemoglobin oxygenation and interstitial pO 2 in venous flaps: an experimental study in rats. Microsurgery 1998; 18: 324-30.
630 The Authors
Boonthin Wungcharoen MD, Consultant Plastic Surgeon Yingyos Santidhananon MD, Chief of Plastic Surgery Vasant Chongchet MD, FRCS, Professor of Plastic Surgery Plastic and Reconstructive Surgery Unit, Department of Surgery, Bangkok Metropolitan Administration Medical College and Vajira Hospital, 681 Samsen Road, Dusit, Bangkok 10300, Thailand.
British Journal of Plastic Surgery
Wisuit Pradidarcheep PhD, Biological Research Scientist and Lecturer Department of Anatomy, Faculty of Medicine, Srinakharinwirot University, Sukhumvit 23, Bangkok 10110, Thailand. Correspondence to Dr Boonthin Wungcharoen. Paper received 17 January 2001. Accepted 27 June 2001, after revision.
9 2001 The British Association of Plastic Surgeons doi: 10.1054/bjps.2001.3675
BRITISH
JOURNAL
OF
[ ~ ]
PLASTIC
SURGERY
Pre-arterialisation of the arterialised venous flap: an experimental study in the rat B. Wungcharoen,W. Pradidarcheep*,Y. Santidhananon and V. Chongchet Plastic and Reconstructive Surgery Unit, Department of Surgery, Bangkok Metropolitan Administration Medical College and Vajira Hospital; and *Department of Anatomy, Faculty of Medicine, Srinakharinwirot University, Bangkok, Thailand SUMMARY. Arteriovenous fistulae cause haemodynamic and morphological changes to the local venous channels. We have used the concept of preformed arteriovenous fistulae to study the viability improvement of arterialised venous flaps. Five groups of flaps were created using the abdominal skin of the Wistar rat (n = 10 in each group) with a silastic sheet implanted underneath. Group 1 (control) contained a flap without a vascular supply, group 2 (venous perfusion flap) contained a single pedicled skeletonised vein and a draining vein, and group 3 (arterialised venous flap) contained an arteriovenous shunt proximal to the single pedicled skeletonised vein and a draining vein; in group 4 (7 day pre-arterialised flap) the arteriovenous shunt was performed 7 days before the flap was raised in the same procedure as group 3, and in group 5 (14 day pre-arterialised flap) the arteriovenous shunt was performed 14 days before the flap was raised. The surviving surface areas of the flaps in each group, assessed 7 days after raising, were 0%, 22.21%, 54.32%, 62.21% and 97.47%, respectively. There was a statistically significant difference in survival between venous perfusion flaps and arterialised venous flaps (P < 0.05). Only the 14 day pre-arterialised flaps had a statistically significantly larger area of survival than arterialised venous flaps (P < 0.05). Microangioarchitecture of the pre-arterialised group, studied by the microvascular corrosion-cast technique combined with scanning electron microscopy and transmission electron microscopy, revealed dilatation of veins, numerous small neo-vessels and a decrease in or total absence of functioning valves. We conclude that 14-day pre-arterialisation in the rat model improved the survival of arterialised venous flaps by increasing collateral pathways for arterialised blood flow through the flap. 9 2001 The British Association of Plastic Surgeons
Keywords: pre-arterialisation, arterialised venous flap, arteriovenous shunt, rat, flap survival, microangioarchitecture.
A venous flap is a non-conventional vascular perfusion skin flap, through which the main transit of blood flow is the venous system. It can be perfused by either venous or arterialised blood. Arterialised venous flaps have been shown to be more reliable than venous perfusion flaps both in experiments and in clinical studies. 1-3 Small arterialised venous flaps can be successfully employed clinically.4 A number of ways to improve the survival of large arterialised venous flaps have been studied, because such flaps tend to give less-favourable results. 5-8 These studies have investigated adequate draining veins from the venous flap, 9,1~the size of the venous network in the centre of the flap 1~ and neovascularisation from the recipient bed. 12 All of these factors are likely to limit the application of these flaps or lead to unfavourable outcomes if the degree to which they are present is suboptimal. Nakayama et al first described, using rat models, successful large arterialised venous flaps in which the primary arteriovenous anastomosis of the feeding vein was performed 2 weeks prior to flap harvest. 13 Although this was not a controlled experiment, comparing results
with conventional arterialised venous flaps, it demonstrated the possible improvement in venous-flap survival. The results of the authors' previous study of pre-arterialisation in clinical cases also support this concept. 14 We designed this study to investigate the effects of pre-arterialisation on the survival of arterialised venous flaps in an animal model. Pre-arterialisation was achieved by preforming an arteriovenous fistula of the vein within the flap for different periods of time.
Materials and methods Animal model and design of the flaps In this study, we used 50 male Wistar rats weighing between 300 g and 350 g. General anaesthesia was induced by injection of intraperitoneal sodium pentobarbital (30 mg/kg). The ventral abdominal wall was shaved using an electric shaver, and prepared with 70% chlorhexidine in alcohol. The outline of the flap was a rectangle on the fight side of the abdominal wall measuring 4 cm in length and 3 cm in width. The medial border of the flap was the axis between the xiphoid process and the symphysis pubis. The lower border of the flap was the axis formed by a perpendicular line from the midline axis to the
Presented at the 25th Annual Congress of the Royal College of Surgeons of Thailand, 14-17 July 2000.
62l
622
British Journal of Plastic Surgery
lateral concavity between the abdomen and the leg (Fig. 1). The femoral vessels were located underneath the medial portion of the lower border of the flap. All surgical procedures were conducted using aseptic techniques.
Experimental groups Five groups of rats (n = 10 in each group) were used in the study (Fig. 2, Table 1). The rats were randomly assigned to groups.
Group 1 (control group). The skin flap was raised without a vascular supply. A silastic sheet was placed under the flap. The skin flap was immediately sutured back into the donor site. Group 2 (venous perfusion venous flap). The skin flap was elevated fed only by the skeletonised right inferior epigastric vein. The inferior epigastric artery was ligated. Outflow was provided by the long thoracic vein at the upper lateral border of the flap. A silastic sheet was inserted under the flap and the flap was sutured back into the donor site.
Long thoracic vein.";
Group 3 (arterialised perfusion venous flap). A sideto-side arteriovenous anastomosis was performed between the fight femoral artery and the fight femoral vein just proximal to the origin of the inferior epigastric vein. The skin flap was elevated fed by the skeletonised right inferior epigastric vein. The inferior epigastric artery was ligated. Outflow was provided by the long thoracic vein. A silastic sheet was inserted under the flap and the flap was sutured back into the donor site. Group 4 (7 day pre-arterialised venous flap). Prearterialisation was performed through a 1 cm skin incision at the medial lower border of the planned flap. A sideto-side anastomosis was performed between the right femoral artery and the right femoral vein proximal to the origin of the inferior epigastric vein. The incision was closed with simple interrupted sutures using 5/0 nylon. Seven days after the anastomosis, the flap was raised based on the skeletonised right inferior epigastric vein, with the long thoracic vein acting as the draining vein. A silastic sheet was inserted and secured underneath the flap and the flap was sutured back into the donor site. Group 5 (14 day pre-arterialised venous flap). Prearterialisation was performed in the same way as in group 4. The flap was raised 14 days after the arteriovenous shunt was performed. The flap was supplied by the completely skeletonised right inferior epigastric vein and drained by the long thoracic vein. A silastic sheet was inserted under the flap and the flap was sutured back into the donor site.
A-Vshunt
,,' immediate
raising
v~n Group 1
Group 2
Group 3
Inferior WS.
~ ~ A - V
~L 7 day interval
~
~
shunt
lap harvest
~
Group 4
Figure 1--Animal model for the experiment. Schematic illustration of flap location.
Table 1
i ~ ~ A-Vshunt 14 day Interval
Flap harvest
Group 5
Figure 2---Schematic illustration of the five flap groups: group 1, control group; group 2, venous perfusion venous flap; group 3, arterialised perfusion venous flap; group 4, 7 day pre-arterialised venous flap; group 5, 14 day pre-arterialised venous flap.
Experimental protocol
Group
Day 0
Day 7
1. control (n = 10) 2. venous perfusion (n = 10) 3. arterialised perfusion (n -- 10) 4.7 day pre-arterialised (n = 10) 5.14 day pre-arterialised (n = 10)
flap elevation flap elevation arteriovenous shunt, flap elevation arteriovenous shunt arteriovenous shunt
flap assessment flap assessment flap assessment flap elevation
Day 14
Day 21
flap assessment flap elevation
flap assessment
Pre-arterialisation of the arterialised venous flap
Surgical technique The panniculus carnosus was included with the skin flap in all rats. The silastic sheets inserted in all groups measured 4.5 x 3.5cm and were secured in place using quilted 5/0 nylon sutures at all borders. All skin flaps were sutured back into the donor sites using simple interrupted 5/0 nylon sutures. Side-to-side anastomosis was performed by making 1 mm slit incisions with a number 11 blade on adjacent sides of the artery and vein, and completed with eight interrupted 10/0 nylon sutures. The retrograde flow of arterial blood through the fistula was tested by temporarily occluding the femoral artery proximal to the shunt. When the femoral artery was occluded by microforceps, the inferior epigastric vein looked cloudy blue. When the artery was released, the inferior epigastric vein turned bright red and markedly dilated due to the arterial blood flow. All nerves and other vessels attached to the flap except for the inferior epigastric vein and the long thoracic vein were cauterised and transected. The bleeding sites were controlled with bipolar cauterisation.
Exclusion criteria Rats were excluded from the study if the vascular pedicle thrombosed, the patency of the arteriovenous shunt could not be demonstrated during flap harvest or the animal died during the experiment.
Flap assessment The assessment included survival assessment, microangioarchitecture and histological studies.
Survival assessment.
All rats were observed for 7 days after raising the flap. They were then anaesthetised before the assessments. The surviving parts of the flaps were closely traced onto translucent plastic sheets. Surviving parts were defined as areas without any necrosis. The surface area traced was calculated by scanning the sheet, along with a standard surface-area sheet, using a graphic computerised scanner. The pixels in the scanned picture were then counted using histogram analysis in Adobe Photoshop v. 5.0 software. The number of pixels was converted into surface area by comparison with the standard surface sheet. All mapping was performed blindly.
Microangioarchitecture study. The microvasculature was studied by transmission electron microscopy and, threedimensionally, by a vascular corrosion cast in conjunction with scanning electron microscopy. Six rats from each group were studied immediately after the survival assessments. Transmission electron microscopy was performed after sacrifice of the animals, but vascular corrosion casting was performed while the animals were still alive. For transmission electron microscopy, surviving areas of flaps were randomly selected and fixed by immersion in 2.5% glutaraldehyde in cacodylate buffer, pH 7.4, and left overnight in the same fixative at 4 ~ The samples were post-fixed in 1% osmium tetroxide, dehydrated in a graded series of ethanols, infiltrated and embedded in
623 Araldite 502 resin. Semi-thin sections (1 p~m) were stained with 1% toluidine blue and examined under a light microscope; 70-100nm sections were stained with 1% uranyl acetate, then lead citrate, and examined under a transmission electron microscope at an accelerating voltage of 80 kV. For vascular corrosion casting, the animal preparation, the ventricular injection of Batson's number 17 plastic mixture (casting medium) and the preparation for obtaining the skin-flap vascular casts were as according to the method previously described by Promwikorn et al. 15 After corroding the soft tissue with 40% KOH, the remaining vascular cast was stuck onto a brass stub with silver paint, and coated with gold-palladium prior to being examined and photographed under a scanning electron microscope at accelerating voltages of 10kV and 15 kV. Plastic mixtures, which were infused through the left ventricle while the animal was still alive, travelled along the femoral artery and passed through the arteriovenous fistula. From the fistula, the infusate was distributed through the inferior epigastric vein to the whole of the surviving flap. The casts represent the venous microarchitecture of the venous flaps. Scanning electron microscopy of these vascular casts revealed the detailed architecture of the inner walls.
Histological study. The remaining rats from each group were used for histological analysis. Mid-portions of the surviving parts of the flaps were harvested for histological study and stained with haematoxylin and eosin. These sections were examined under a light microscope. Sections were obtained from four rats in group 1, three rats in each of groups 2, 3 and 5 and two rats in group 4.
Statistical analysis A Median Test was used to analyse the differences in survival between groups 2, 3, 4 and 5 using SPSS v. 9.0 software. Mann-Whitney U-tests were used for post-hoc comparisons between groups 2 and 3, groups 3 and 4, groups 4 and 5, and groups 3 and 5. Statistical significance was defined as P < 0.05.
Results We were able to assess survival in 45 out of the 50 rats. Five rats were excluded from the experiment: one rat in group 2 died, the anastomosis failed in one rat in each of groups 3, 4 and 5, and the pedicle thrombosed in one rat in group 4.
Results of the survival assessment All flaps in the control group necrosed. The mean surviving area in the venous-perfusion group was 22.21%. The mean surviving area in the arterialised perfusion group was 54.32%. The mean surviving area of the arterialised venous flaps that were pre-arterialised for 7 days was 62.21%. The arterialised venous flaps that were pre-arterialised for 14 days showed the highest surviving area, with a mean of 97.47% (Table 2, Figs 3 and 4). Statistical
624
British Journal of Plastic Surgery Table 2
Survival areas in each group
Group
Mean surface-area survival (cm 2)
1 ( n = 10) 2 (n = 9) 3 (n = 9) 4 (n = 8) 5 (n =9)
0.00 2.67 6.52 7.46 11.70
s.e.m.
-
0.25 0.69 0.79 0.10
13
v
0.00
22.21 54.32 62.21 97.47
Groups compared
'Z
Asymptotic significance
--3.576 --0.866 -3.595 --3.486
<0.05* 0.386 <0.05* <0.05*
10 8
m
-
2.09-3.24 4.92-8.11 5.59-9.34 11.46-11.93
12
2 and 3 3 and 4 3 and 5 4 and 5
n," f/} < ILl
Mean surface area survival (%)
Table 3 Comparisons of mean survival areas between groups using the Mann-Whitney U-test
14'
X
95% confidence interval of the mean
6 4
2
I1" 2~ if)
0
*Significant difference.
-2 N=
1
2
3
4
5
EXPERIMENTAL GROUP
Figure 3--Box plot showing the surviving surface areas of the flaps in each group. The box length is the interquartile range (Q3-Q1). The dark line within each box is the median (Q2). The asterisk indicates the extreme ( > Q 3 + 3 ( Q 3 - Q O ) and o indicates outliers (between 1.5(Q3-Q1) and 3(Q3- Q1)).
Figure 5--Stereomicrograph of vascular corrosion cast before coating with gold-palladium. This is an example of a group 4 vascular corrosion cast.
Figure 4--An example of a 14 day pre-arterialised venous flap (group 5). Note that survival of the flap is nearly complete. Dilated tortuous superficial veins are visible throughout the flap.
analysis by M e d i a n Tests revealed significant differences b e t w e e n groups 2, 3, 4 and 5 (median = 6.58, ) 2 = 19.487, asymptotic significance less than 0.05). M a n n - W h i t n e y U-tests showed significant differences in survival b e t w e e n the venous perfusion group and the arterialised perfusion group, b e t w e e n the arterialised perfusion group and the 14 day pre-arterialised group, and b e t w e e n the 7 day pre-arterialised group and the 14 day pre-arterialised group (asymptotic significance less than 0.05) (Table 3). There was no significant difference between the arterialised group and the 7 day pre-arterialised group (asymptotic significance of 0.386).
Pre-arterialisation of the arterialised venous flap
Result of the microangioarchitecture study Vascular corrosion casting could be conducted only in groups 3, 4 and 5 (Fig. 5) because the vessels in the control group and the venous perfusion group were all thrombosed. On gross examination of the casts, the venous pattern of conventional arterialised venous flaps (group 3) was a homogeneously distributed network with straight main veins. The 7 day pre-arterialised group (group 4) showed dilated tortuous main veins with a heterogeneous distribution of venous networks. In the 14 day pre-arterialised group (group 5) the dilated tortuous interconnecting venous network was prominent throughout the whole flap. Under scanning electron microscopy, group 3 (Fig. 6) revealed a normal inner microarchitecture of the main small collecting veins (100-300lxm), 16 consisting of indentations along the smooth inner vascular wall. These indentations were endothelial nuclear imprints. The tributaries of these collecting veins were muscular venules (50-100txm) and collecting venules (30-50 lxm). These venules were uniformly distributed and gradually decreased in size.
625 The postcapillary venules (8-30 txm) and venous capillaries ( < 8 txm) were sparse. Venous valves were usually found where venules branched off from the collecting vein. In group 4 the collecting veins became dilated and tortuous. The inner architecture of the collecting veins disappeared and was replaced with rough wrinkled surfaces (Fig. 7). Small vessels, including venous capillarysized and postcapillary venule-sized vessels, were found immediately branching off from the collecting vein. Venous valves were still seen, but only scarcely. The microarchitecture of group 5 was distinct (Fig. 8). The collecting veins, including the muscular venules and the collecting venules, were markedly distended and tortuous. The inner walls of these dilated vessels were rough and velvety with wrinkled surfaces. Numerous small vessels (venous capillary-sized vessels and postcapillary venule-sized vessels) sprouted from the collecting veins and the venules. They ran along or angled away from the collecting veins and venules in every direction. These small vessels formed vascular plexuses, interconnecting the dilated collecting veins and venules. There were no
Figure 6--Scanning electron micrograph of a vascular cast from group 3. (A) A venous network with straight smooth collecting veins ( 100-300 Fun) and venules gradually decreasing in size. There are very few venous capillaries ( < 8 ~um) and postcapillary venules (8 30 ~m) connecting the larger vessels. (B) Indentations on the smooth surfaces of the collecting veins represent the imprints of nuclei of endothelial cells. Venous valves, which were distributed throughout the network, can be seen where 9enules branch off from the collecting vein.
Figure 7--Scanning electron micrograph of a vascular cast from group 4. (A) Occasional venous valves are seen at the branching sites of venules from the collecting vein. (B) A few small vessels (venous capillary-sized vessels and postcapillary venule-sized vessels) abruptly branch off from the collecting vein. The normal nuclear indentations on the inner walls of the collecting veins are lost and replaced by a rough wrinkled surface. Compared with group 3, the collecting veins are tortuous and very dilated.
626
British Journal of Plastic Surgery
Figure 8--Scanning electron micrograph of a vascular cast from group 5. (A) Numerous small vessels (venous capillary-sized vessels and postcapillary-sized vessels) sprout from the collecting veins and venules in a multidirectional pattern. These small vessels form the vascular plexus interconnecting the collecting veins and venules. (B) The collecting vein has become markedly dilated and tortuous. The nuclear indentations seen in normal veins have been lost and replaced by a rough wrinkled surface. Abundant venous capillary-sized vessels branch off from these dilated veins.
Figure 9--Light micrograph showing a semi-thin section of a vein from a group 3 specimen. The endothelial cells are flat, elongated and scarce, with heterochromatic nuclei: va, venous valve (toluidine blue, • 70).
Figure 10--Light micrograph showing a semi-thin section of a vein from a group 5 specimen. The endothelial cells (e) are numerous, with euchromatic nuclei: *, internal elastic lamina ( x 90).
demonstrable valves in any specimen in this group. There was no demonstrable retrograde flow of the plastic mixture into arterioles or arteries in any of the groups. In the semi-thin section, the endothelial cells of group 3 (Fig. 9) demonstrated a flat elongated heterochromatic nucleus. There were fewer endothelial cells in the veins of this group than in the veins of groups 4 and 5. By contrast, the veins of groups 4 and 5 had many more endothelial cells (Fig. 10). Ultrastructural detail from transmission electron microscopy of these endothelial cells revealed active young cells with euchromatic nuclei (Fig. 11). The internal elastic lamina in the veins in these groups was discontinuous. There were also numerous young endothelial cells, which connected to form new vessels in the subcutaneous tissue in groups 4 and 5. Transmission electron microscopy of these areas demonstrated a neovascutarisation process. The active young endothelial cells with euchromatic nuclei and rich rough endoplasmic reticulum within the cytoplasm (Fig. 12) had encircled the newly formed lumina. This was seen in both the pre-arterialised groups, but was much more pronounced in group 5.
Histologicalfindings All the flaps in group 1 had necrosed from the skin through to the subcutaneous tissue. In group 2, the surviving parts of the flaps were composed of scanty small vessels and atrophic subcutaneous tissue. In group 3, histology showed oedematous subcutaneous tissue with dilated superficial veins (Fig. 13). In the pre-arterialised groups, the subcutaneous tissue was less oedematous but the superficial veins were massively dilated (Fig. 14). Numerous small vessels were scattered through all layers of the flaps.
Discussion As described by Nakayama et al, an abdominal skin flap in the rat can be supplied by arterial blood through the venous system. 13 In their study the arterialised venous flap was pre-arterialised, and the authors named it 'the delayed fS flap'. It showed a promising survival rate. Our study was designed as a controlled trial to test the hypothesis that pre-arterialisation of the venous flap
Pre-arterialisation of the arterialised venous flap
627
Figure 13--Light micrograph showing a section of the survivingpart of a group 3 flap. Oedematous subcutaneoustissue is prominent.Note the few dilated superficialveins (V) (haematoxylinand eosin, • 9).
Figure 11 Transmissionelectron micrograph of an endothelial cell from a vein in group 5. The euchromaticnucleus(N) is prominentin the endothelial cell, which projects into the lumen (L). The internal elastic lamina(*) forms a brokenzigzag line (scale bar: 2 ixm).
Figure 14~Light micrograph showing a section of the survivingpart of a group 5 flap. Huge dilated veins (V) and numerous small vessels are seen withinthe subcutaneoustissue. Compared with group 3, the subcutaneous tissue appears less oedematous (haematoxylinand eosin, • 9).
Figure 12~Transmission electron micrograph of the wall (tangential cut) of a newly formed blood vessel within the subcutaneoustissue of a group 5 specimen.Youngendothelialcells with euchromaticnuclei (N) are rich in rough endoplasmicreticulum(r) (scale bar: 2 p~m).
yields a better survival than can be obtained with the conventional arterialised venous flap. We used a side-toside anastomosis between the femoral artery and the femoral vein to perfuse the venous flap. Our pilot study using end-to-end anastomosis resulted in the postoperative
death of nearly all the animals. This is consistent with Nakayama et al's experiment, in which 13 out of 15 animals died after the end-to-end anastomosis of the femoral artery and the epigastric vein in the conventional arterialised venous flap. 13 We concluded that the side-to-side anastomosis of the femoral artery and the femoral vein, without the ligation of any main vessels, would not cause significant physiologic change, and most of the animals survived until the end of the experiment. The arteriovenous shunt, which was situated proximal to the epigastric vessels, could shunt arterial blood into the venous system of the flap, and its action could be demonstrated immediately after the anastomosis by observing dilated epigastric veins containing bright red blood. The shunt effect was confirmed later by vascular castings, which demonstrated the comparative microarchitecture of the venous systems in each flap. Indeed, side-to-side anastomosis was one of the classic procedures for vascular access in haemodialysed patients in the past, and has been proven to secure maximal fistula flow rate. 17 The insertion of a silastic
628 sheet underneath the flap aimed to exclude neovascularisation from the bed and the external effect of surrounding vessels, in both survival assessment and the vascularcasting study. Complete skeletonisation of the nutrient veins, including the transection of all nerves, reflected the actual survival of the arterialised venous flap, and thus demonstrated the effects of pre-arterialisation. Small vascular networks within the perivenous areolar tissue and nerve might unintentionally improve flap survival. 18'19 This is why the venous island flaps in early clinical reports had high survival rates, 2~ while the total venous perfusion free flap had comparatively poor survival. 3'24 The lower surviving area of the venous perfusion flap (group 2) (22.21%), in which the pedicle was completely skeletonised, confirmed this. The arterialised venous flap, which had a mean survival area of 54.32%, was statistically significantly superior to the venous perfusion flap (Mann-Whitney U-test between groups 2 and 3: Z = --3.576, asymptotic significance less than 0.05). This result runs contrary to some excellent outcomes in other experiments with arterialised venous flaps. 9,25 However, this may be due to the preclusion of plasmatic imbibition and early neovascularisation from the bed by the silastic sheet. Studies have confirmed significant roles of both factors. 26'27 The survival of small arterialised venous flaps was satisfactory in an early clinical report,4 but larger flaps have demonstrated less predictable survival. 5,s Neovascularisation may account for the lower survival rate of larger venous flaps: 12 small flaps are less dependent on neovascularisation from the bed because the marginal blood supply and/or plasmatic imbibition are adequate for survival. Apart from neovascularisation, a balanced blood flow, including arterialised inflow, through the centrally located venous network together with adequate draining veins are essential for survival, w These factors have been manipulated to improve survival in various recent experiments, which have included pre-fabrication of the venous flap, 28 expansion of the venous flap 29 and surgical delay of the venous flap, 30'31 including our pre-arterialisation experiment. Our model of pre-arterialisation aimed to improve survival by optimising the trade-off between sufficient oxygenated inflow and adequate outflow. The mean survival of the 14 day pre-arterialised flap (97.47%) was statistically significantly better than that of the conventional arterialised venous flap (Mann-Whitney U-test between groups 3 and 5: Z = -3.595, asymptotic significance less than 0.05). However, the mean survival of the 7 day prearterialised flap (62.21%) was not statistically significantly different from that of the conventional arterialised venous flap (Mann-Whitney U-test between groups 3 and 4: Z = - 0 . 8 6 6 , asymptotic significance of 0.386). This implies that pre-arterialisation influenced survival over the longer duration. Pre-arterialisation of the venous flap was achieved by using an arteriovenous fistula. So, the same haemodynamics and pathophysiology could be expected: shunts cause venous channels to dilate, and promote collateral venous channels. 16 This, in turn, leads to a decreased resistance of the venous network in the flap. Microangioarchitecture, demonstrated by vascular casting and scanning electron microscopy, also supported this hypothesis. Pre-arterialisation causes dilatation of the
British Journal of Plastic Surgery venous system, and stretches and expands the endothelium, which showed a rough wrinkled surface in both the 7 day and the 14 day pre-arterialised groups. The very distinct feature of the 14 day pre-arterialised group was the frequency of small vessels branching from the main veins. Although, in the vascular corrosion casting and scanning electron microscopy study, these small vessels could not be clearly identified either as neovascularisation or as existing collateral vessels opening under the high retrograde pressure, the semi-thin section and transmission electron microscopy study demonstrated numerous active young endothelial cells in the 14 day pre-arterialised group. Some of these cells were seen to be forming new lumen. This evidence supports angiogenesis within the flap. In the conventional arterialised venous flap, limited vascular channels may lead to inadequate oxygenation of the flap even if it is perfused by arterialised blood. Selection of a donor site that contains abundant venous networks would increase the survival rate of the arterialised venous flap 1~ because oxygenation would be improved, although relative hypoxia would still induce angiogenesis, 32 which is necessary for long-term perfusion. The small vessels that were demonstrated in the pre-arterialised groups increased the pathway for microcirculation. The absence of venous valves in the 14 day pre-arterialised group might result from the effect of retrograde pressure or from venous-channel dilatation. The absence of valves certainly resulted in the multidirectional flow of arterialised blood through the venous network and the newly formed vessels. The resistance in the flap then decreased and, consequently, the stagnation of blood also decreased. Oxygenation of the flap tissue was then easily accomplished. According to the experiment of Wolff et al, 33 the capillary system in all types of venous flap is reached by some of the inflowing oxygenated haemoglobin over the entire flap surface. The arterialised venous flap had poorer intracapillary oxygenated haemoglobin than the arterial flap, and the venous flow-through flap had a lower oxygen supply than the arterialised venous flap. Although very few capillary-sized vessels were demonstrated in conventional arterialised venous flaps (group 3) by vascular casting, the increased survival in the pre-arterialised venous flap groups supports this 'intracapillary haemoglobin oxygenation of venous flap' concept. Capillary-sized vessels were much more frequent in the pre-arterialised groups, especially in the 14 day pre-arterialised group, and the mean survival area increased. Although this microangioarchitecture study was descriptive, it could, in part, be used to explore the morphological changes of the venous structure within the venous flap. The hypothesis that pre-arterialised venous flaps have better survival is based on this evidence. The exact morphological and haemodynamic changes in different kinds of venous flaps need to be explored further. From the results of our experimental model in the rat, we draw the following conclusions. Without the influences of plasmatic imbibition and neovascularisation from the recipient bed: first, arterialised venous flaps had a statistically significantly better survival than venous perfusion venous flaps; second, pre-arterialisation improved the survival of arterialised venous flaps; third,
Pre-arterialisation of the arterialised venous flap only the 14 day pre-arterialised flaps showed statistically significantly better survival than conventional arterialised venous flaps; fourth, pre-arterialisation caused dilatation of the superficial veins, and abundant small vascular plexuses were found along the main collecting veins of the flap - functioning venous valves were absent or decreased after pre-arterialisation; and fifth, the microangioarchitectural changes resulting from pre-arterialisation might explain the increase in survival, in terms of a decrease in resistance to arterial blood flow, an increase in oxygenated blood exchange through numerous small vessels and plexuses, and a decrease in the volume of stagnant deoxygenated blood within the flap. These results are based on the rat flap model, which includes the panniculus carnosus and which is different from human skin. It is therefore not possible to extrapolate these findings directly to human skin, but it would seem appropriate to evaluate the application of pre-arterialisation in the clinical situation.
Acknowledgements The microvascular corrosion castings were conducted at the research laboratory of the Department of Anatomy, Faculty of Science, Mahidol University, Bangkok, Thailand. The authors thank Dr Reon Somana and all the research fellows there for their valuable cooperation. The electron microscopy examinations were conducted at the electron microscopy research laboratory of the Department of Anatomy, Faculty of Medicine, Srinakharinwirot University, Bangkok, Thailand. The pathology preparation and results were the contribution of Dr Somnuek Jesdapatarakul and Dr Nantana Kaewpila from the Department of Pathology, Bangkok Metropolitan Administration Medical College and Vajira Hospital, Bangkok, Thailand.
References 1. Suzuki Y, Isshiki N, Ishikawa K, Koyama H. Viability and quantitative dermofluorometry of experimental arterialised and non-arterialised venous flaps. Br J Plast Surg 1993; 46: 273-8. 2. Wolff K-D, Telzrow T, Rudolph K-H, Franke J, Wartenberg E. Isotope perfusion and infrared thermography of arterialised, venous flow-through and pedicled venous flaps. Br J Plast Surg 1995; 48: 61-70. 3. Chen H-C, Tang Y-B, Noordhoff MS. Four types of venous flaps for wound coverage: a clinical appraisal. J Trauma 1991; 31: 1286-93. 4. Inoue G, Maeda N, Suzuki K. Resurfacing of skin defects of the hand using the arterialised venous flap, Br J Plast Surg 1990; 43: 135-9. 5. Yilmaz M, Menderes A, Karata ~, Karaca C, Barut~u A. Free arterialised venous forearm flaps for limb reconstruction. Br J Plast Surg 1996; 49: 396-400. 6. Sakai S. Arterialised venous gtroin flap: case report. Br J Plast Surg 1996; 49: 90-2. 7. Klein C, Kowics A, Stuckensen T. Free arterialised venous forearm flaps for intraoral reconstruction. Br J Plast Surg 1997; 50: 166-71. 8. Inoue G, Suzuki K. Arterialized venous flap for treating multiple skin defects of the hand. Plast Reconstr Surg 1993; 91: 299-302. 9. lnada Y, Fukui A, Tamai S, Mizumoto S. The arterialised venous flap: experimental studies and a clinical case. Br J Plast Surg 1993; 46: 61-7. 10. Woo S-H, Kim S-E, Lee T-H, Jeong J-H, Seul J-H. Effects of blood flow and venous network on the survival of the arterialized venous flap. Plast Reconstr Surg 1998; 101: 1280-9.
629 11. Lee WPA. Arterialized venous flap for treating multiple skin defects of the hand. Plast Reconstr Surg 1993; 91: 303-6. 12. Pittet B, Chang P, Cederna E Cohen MB, Blair WF, Cram AE. The role of neovascularization in the survival of an arterialized venous flap. Plast Reconstr Surg 1996; 97: 621-9. 13. Nakayama Y, Soeda S, Kasai Y. Flaps nourished by arterial inflow through the venous system: an experimental investigation. Plast Reconstr Surg 1981; 67: 328-34. 14. Wungcharoen B, Santidhananon Y, Chongchet V. Pre-arterialisation of an arterialised venous flap: clinical cases. Br J Plast Surg 2001; 54: 112-16. 15. Promwikorn W, Thongpila S, Pradidarcheep W, Mingsakul T, Chunhabundit P, Somana R. Angioarchitecture of the coeliac sympathetic ganglion complex in the common tree shrew (Tupaia glis). J Anat 1998; 193: 409-16. 16. Rhodin JAG. Ultrastructure of mammalian venous capillaries, venules, and small collecting veins. J Ultrastruct Res 1968; 25: 452-500. 17. Sumner DS. Hemodynamics and pathophysiology of arteriovenous fistulae. In Rutherford RB, ed. Vascular Surgery. Philadelphia: W.B. Saunders, 1995: 1166-91. 18. Murata K, Tamai S, Inada Y, Fukui A, Miyamoto S. Transfer of a pedicled venous flap containing perivenous areolar tissue and nerve: an experimental study. Br J Plast Surg 1999; 52: 223-9. 19. Noreldin AA, Fukuta K, Jackson IT. Role of perivenous areolar tissue in the viability of venous flaps: an experimental study on the inferior epigastric venous flap of the rat. Br J Plast Surg 1992; 45: 18-22. 20. Chavoin JP, Rouge D, Vachaud M, Boccalon H, Costagliola M. Island flaps with an exclusively venous pedicle. A report of eleven cases and a preliminary haemodynamic study. Br J Plast Surg 1987; 40: 149-54. 21. Thatte MR, Kumta SM, Purohit SK, Deshpande SN, Thatte RL. Cephalic venous flap: a series of 8 cases and a preliminary report on the use of 99mTc labelled RBCs to study the saphenous venous flap in dogs. Br J Plast Surg 1989; 42: 193-8. 22. Thatte RL, Thatte MR. The saphenous venous flap. Br J Plast Surg 1989; 42: 399-404. 23. Inada Y, Fukui A, Tamai S, Kakihana T, Maeda M. The sliding venous flap for covering skin defects with poor blood supply on the lateral aspects of fingers. Br J Plast Surg 1991; 44:368-71. 24. Amarante J, Costa H, Reis J, Soares R. Venous skin flaps: an experimental study and report of two clinical distal island flaps. Br J Plast Surg 1988; 41: 132-7. 25. Nichter LS, Haines PC. Arterialized venous perfusion of composite tissue. Am J Surg 1985; 150: 191-6. 26. Fukui A, Tamai S, Maeda M, Inada Y, Mii Y, Mine T. The pedicled venous flap. An experimental study. Br J Plast Surg 1993; 46: 116-21. 27. Matsushita K, Firrell JC, Ogden L, Tsai T-M. Blood flow and tissue survival in the rabbit venous flap. Plast Reconstr Surg 1993; 91: 127-35. 28. Takato T, Komuro Y, Yonehara H, Zuker RM. Prefabricated venous flaps: an experimental study in rabbits. Br J Plast Surg 1993; 46: 122-6. 29. Mutaf M, Tasaki Y, Fujii T. Expansion of venous flaps: an experimental study in rats. Br J Plast Surg 1998; 51: 393-401. 30. Byun JS, Constantinescu MA, Lee WPA, May JW Jr. Effects of delay procedures on vasculature and survival of arterialized venous flaps: an experimental study in rabbits. Plast Reconstr Surg 1995; 96: 1650-9. 31. Cho B-C, Lee M-S, Lee J-H, Byun J-S, Baik B-S. The effects of surgical and chemical delay procedures on the survival of arterialized venous flaps in rabbits. Plast Reconstr Surg 1998; 102: 1134-43. 32. Stavri GT, Zachary IC, Baskerville PA, Martin JF, Erusalimsky JD. Basic fibroblast growth factor upregulates the expression of vascular endothelial growth factor in vascular smooth muscle cells: synergistic interaction with hypoxia. Circulation 1995; 92: 11-14. 33. Wolff K-D, Uekermann B, Matthes G, Wartenberg E. Intracapillary haemoglobin oxygenation and interstitial pO 2 in venous flaps: an experimental study in rats. Microsurgery 1998; 18: 324-30.
630 The Authors
Boonthin Wungcharoen MD, Consultant Plastic Surgeon Yingyos Santidhananon MD, Chief of Plastic Surgery Vasant Chongchet MD, FRCS, Professor of Plastic Surgery Plastic and Reconstructive Surgery Unit, Department of Surgery, Bangkok Metropolitan Administration Medical College and Vajira Hospital, 681 Samsen Road, Dusit, Bangkok 10300, Thailand.
British Journal of Plastic Surgery
Wisuit Pradidarcheep PhD, Biological Research Scientist and Lecturer Department of Anatomy, Faculty of Medicine, Srinakharinwirot University, Sukhumvit 23, Bangkok 10110, Thailand. Correspondence to Dr Boonthin Wungcharoen. Paper received 17 January 2001. Accepted 27 June 2001, after revision.