EFFECT OF ANASTOMOSIS AND GEOMETRY OF VESSEL

CURVATURE ON BLOOD FLOW VELOCITY AND PATENCY

IN MICROVESSELS
WING YUNG CHEUNG,1 FENG ZHANG, M.D., Ph.D.,1,2
URS BOSCH, M.D.,2 HARRY J. BUNCKE, M.D.,2 and
WILLIAM C. LINEAWEAVER, M.D.1*

The effect of the geometry of the vessel and the number of

anastomoses on the blood flow was studied. Four different
shapes of the vessel were constructed by using a 6-cm-long
double vein graft model with three anastomoses: (1) an alpha
loop, (2) an omega loop, (3) a sigmoid curve, and (4) straight.
Blood flow was measured by an ultrasound Doppler flowmeter. The result showed no alternation in blood flow across
different geometry and through three patent microanastomoses. However, six out of seven vein grafts were thrombosed
at 24 hr postoperative due to vascular kinks. This model

demonstrates potential sites of kinking at the dissection end

of the femoral artery, the microanastomoses, the side
branches of the vein graft, and the adventitial adhesions.
This model is recommended to microvascular trainees for the
study of kinking and the management of redundant pedicles
and vein grafting.

The most commonly identifiable technical errors in micro-

experience is the single most important factor,1 but it is

certainly more desirable if there is a practical model of
kinking to smooth out the learning curve. The resultant
redundant vessel has been stressed by some authors to be
placed in a gentle curve with fixation.4 However, experimental study of laxity is lacking. Nevertheless, this curvature has not yet been defined. Furthermore, the effect of the
geometry of the vessel curvature to the blood flow has not
been analyzed. Therefore, it is the objective of this study to
quantify the difference of blood flow of the vessel curvature
in different geometry and number of microanatomoses using a double vein grafting model.

vascular surgery are closure under tension or twists and

kinks of the pedicle, especially with utilization of vein grafting.1,2 Salvage rate is disappointingly low when both artery
and vein are thrombosed. Pedicle laxity may be one of the
critical factors increasing the risk of kinking. Although criticism has been that laxity and microanastomoses causing
turbulence may diminish blood flow, some authors advocated an operative plan to have too much vessel length
rather than just enough. Failure to select a flap with a vascular pedicle that can comfortably reach the anticipated recipient vessels could lead to flap failure.1
Occasionally, a lax pedicle is unavoidable particularly
in head and neck reconstruction. It is commonly seen in the
free fibular osteocutaneous graft and the free jejunum graft.
Sometimes a redundant loop of interposition vein graft is
deliberately reconstructed, for example, in the intra-arterial
chemotherapy in head and neck oncology and in kidney
dialysis.3 To make the final pedicle just right without tension or laxity requires good clinical judgment. Operative

Division of Plastic and Reconstructive Surgery, Stanford University Medical

Center, Stanford, CA.
Department of Microsurgical Transplantation and Replantation, Davies Medical Center, San Francisco, CA.

Contract grant sponsor: Microsurgery Foundation of the Davies Medical Center, Kwong Wah Hospital, Hong Kong; Contract grant sponsor: Division of
Plastic and Reconstructive Surgery, Stanford University, CA.
*Correspondence to: William C. Lineaweaver, M.D., Division of Plastic and
Reconstructive Surgery, Stanford University Medical Center, NC104, Stanford,
CA 94305.
1997 Wiley-Liss, Inc.

1997 Wiley-Liss, Inc.

MICROSURGERY

17:491494

1996

MATERIALS AND METHODS

Twenty-two Sprague-Dawley rats weighing between

400500 g were anesthetized with intraperitoneal Phenobarbital, 50 mg/kg. All of the procedures were performed under
the National Research Councils guidelines for the care and
use of laboratory animals.
Following general anesthesia, the lower abdomen and
groin area were shaved with an electric clipper. Transverse
inguinal incisions were made to expose bilateral femoral
and epigastric vessels. The epigastric in continuity with the
femoral vein were harvested as a graft. The femoral artery
was freed carefully to minimize vasospasm. Branches to the
gracilis muscle and the distal femoral vein were ligated. The
vein graft consisted of two segments of femoral (average 1.0
mm in diameter) and epigastric (average 0.8 mm in diam-

492

Cheung et al.

Figure 1. The microvessels placed in different geometry of vessel curvature.

eter) veins between the transected femoral artery forming a

6-cm-long loop. The first segment was anastomosed with
the epigastric vein to the femoral artery. The second segment was reversed with an end-to-end femoral venovenous
anastomosis. The epigastric vein of the second segment was
anastomosed to the distal femoral artery. This arrangement
minimized size discrepancies between the vein graft and the
artery. There were no intervening valves detected in this
segment of vein. Bipolar cautery was not used throughout
the procedure. Lidocaine was used when vasospasm was
observed. Intraluminal heparin was applied for cleaning the
vessel ends. All of the anastomoses were performed with the
standard end-to-end technique using 10-0 nylon suture under 1625 microscopic magnification.
The experiment was divided into five groups with the
microvessels placed in different geometries of vessel curvature (Fig. 1). Four geometries were constructed using the
double epigastric vein graft model with three microanastomoses: the alpha loop (group 1, n 4 6, Fig. 2), the
omega loop (group 2, n 4 6, Fig. 3), the sigmoid
curve (group 3, n 4 6), and the straight (group 4, n 4
4). The contralateral femoral artery with a single anastomosis was used as a control group (group 5, n 4 22).
Patency was checked at 20 min after vascular anastomosis with the standard strip test. The blood flow across
these four geometries of vessels and the control group was

Figure 2. Double vein graft model, the alpha loop.

recorded at 20, 60, and 180 min using the ultrasonic Doppler flowmeter (Transonic Flowprobe, Probe 1R293D). The
blood flow gradient along each segment of the vessel with
double vein grafts (groups 1, 2, 3, and 4) at different time
intervals was also recorded. In addition, the wounds of
seven rats (two each from group 1, 2, and 3, and one from
group 4) were closed with running suturing. These rats were
kept 24 hr for patency and flow examination. Statistical
analysis was performed using the t-test.

Anastomosis and Geometry on Blood Flow and Patency

493

Table 1. Patency Rates of Vein Graftings in Different Geometries

20 Minutes After Surgery*
Group

Patent graft

Patency rate (%)

Control
Alpha
Omega
Sigmoid
Straight

22
8
9
8
4

22
6
6
6
4

100
75
67
75
100

*P > 0.05.

Figure 3. Double vein graft model, the omega loop.

RESULTS

The patency rate of this double vein graft 20 min after

surgery is shown in Table 1.
Blood flow varied in the individual rat. The reading of
the blood flow at each measurement in each group was
slightly increased after anastomosis. The mean blood flow
at proximal and distal to vein grafts of each group from 20
to 180 min is shown in Table 2. Compared with the proximal blood flow, the distal flow of each group was not significantly decreased (P > 0.05). The blood flow change rates
(distal flow/proximal flow) were 1.99, 6.85, 7.93, 2.3, and
4.14% in control, alpha, omega, sigmoid, and straight
groups.
In all double vein grafting groups (groups 1, 2, 3, and 4),
the mean blood flow gradient along each vascular segment
(proximal artery, proximal vein graft, distal vein graft, and
distal artery) at different time intervals is shown at Figure 4.
There was no significant difference of blood flow between
sections of the vessel at 20 and 180 min after anastomosis.
Only one (from group 4) out of seven vein graftings
remained patent. The thrombosed vein grafts were observed
to be kinked in the compressed space of the inguinal area
due to abduction of the hind limb of the animal. Sites of
kinking were identified at the side branches of the vein
grafts, venovenous microanastomoses, proximal and distal
end of the femoral artery, and the fibrinous adhesions. The
blood flow gradient along each vasculer section from the
patent vessel is also shown in Figure 4. Dilation of the site
at femoral vein-to-femoral vein anastomosis was observed.
In this one case, the blood flow was significantly increased
compared with the flow at 180 min after anastomosis.
DISCUSSION

The fundamental properties in fluid mechanics are summarized in Bernoullis priciples.5 Despite the fact that
the condition of frictionless fluid is not found in living

models, major variables associated with fluid energy loss

are identified to be turbulence and resistance. It is generally
accepted that fluid energy is dissipated and laxity diminishes blood flow; and also likely to imply that the number
of anastomoses creating turbulence would decrease blood
flow.6,7 However, the effect of the curvature geometry
and its resultant resistance to blood flow was shown not to
be significant in this model with no significant drop of
blood flow passing through three patent microanastomoses. The blood flow was more dependent on the final
curvature without a kink or twist rather than the geometry
per se.
The potential sites of kinking of the vessel were readily
demonstrated in this double vein graft model. The dissection end of the femoral artery was a weak point for kinking.
Inadequate mobilization of the vessel would lead to a kink.
The microanastomoses and the ligated side branches of the
vein graft were also sensitive to twist and kink. Especially,
when the axis was twisted and the microanastomoses were
performed under torsion, a kink would be manifested following the restoration of blood flow. The vessel was also
likely to kink when there was incomplete clearance of the
adventitia. All these potential sites of kinking were readily
reproduced in changing the geometry of the vessel loop.
Twists in vessel or graft will spiral downstream to the first
fixed point where blockage occurs.
In view of the mobility of the redundant vessel, it was
important to note that fixation of the curvature was extremely important and usually critical in clinical settings.
Only one of the seven vein grafts withstood compression
in the limited space of the groin in this study. Methods
of fixation that may overcome external forces kinking onto
the vessels definitely deserve further investigation. This
posed another important clinical issue in using pedicle
monitoring devices and its risk of disturbing the flow of the
vessel. Our model in the rat groin was not suitable for
long-term studies.
According to Poiseuilles Law, the velocity of flow of a
liquid is inversely proportional to the cross-sectional area of
the tube and the viscosity of the fluid.5 This vein graft
model showed a gradual increase of blood flow after the
procedure. The disproportional increase of flow may be
related to the opening up of collateral venous outflow, recruitment of arteriovenous shunts, or the resolution of va-

494

Cheung et al.
Table 2. Mean Blood Flow at Proximal and Distal to Vein Grafts of Each Group From 20 to 180 Minutes

Group
Control
Alpha
Omega
Sigmoid
Straight

Proximal flow SD
(ml/min)

Distal flow SD
(ml/min)

Flow change
(ml/min)

Percentage

P value

0.803 0.468
0.467 0.119
0.593 0.268
0.520 0.183
0.483 0.191

0.787 0.468
0.435 0.119
0.546 0.219
0.508 0.180
0.463 0.203

0.016
0.032
0.047
0.012
0.02

1.99
6.85
7.93
2.3
4.14

>0.05
>0.05
>0.05
>0.05
>0.05

Figure 4. Mean blood flow gradient along each segment at different time interval (ml/min).

sospasm. It has been shown that the vein only regained 60%
of the original diameter at 24 hr.8 The wider segment of the
vein graft and slower blood flow in this segment were observed in our study. Further study is required to clarify these
observations since the study number was small and the rheology of blood vessel is complex.9
In conclusion, this study showed that blood flow velocity in microvessels was not affected by the vessel curvature
geometry, and that three patent microanastomoses did not
affect blood flow in situ on short observation. This double
vein graft model was a reproducible model for the study of
kinking and a practical laboratory exercise to learn the management of a redundant pedicle.

REFERENCES
1. Khouri RK: Avoiding free flap failure. Clin Plast Surg 19:773781,
1992.

2. Oliva A, Yim K, Lineaweaver WC, Buncke GM, Alpert BS, Siko P,

Buncke HJ: Vein grafting in microsurgical transplantation. Plast Surg
Forum 13:2829, 1990.
3. Ho CM, Lam LK, Wei WI, Choi TK, Lam KH: Regional chemotherapy for recurrent squamous head and neck cancers through a saphenous vein interposition graft. Arch Otolaryngol Head Neck Surg
119:608611, 1993.
4. Urken ML, Vickery C, Weinberg H, Buchbinder D, Biller HF: Geometry of the vascular pedicle in free tissue transfers to the head and neck.
Arch Otolaryngol Head Neck Surg 115:954960, 1989.
5. Sumner DS: Essential hemodynamic principles, in Vascular Surgery.
Philadelphia Saunders, 1989.
6. Acland RD: Microvascular surgical experimental thrombosis model:
Rationale and design (discussion). Plast Reconstr Surg 83:873874,
1989.
7. Zhang F, Lineaweaver WC, Buntic R, Walker R: Mechanical evaluation of anastomotic tension and patency in arteries. J Reconstr Microsurg 12:121126, 1996.
8. Cooley BC, Gould JS: The effect of venous flow alterations upon
patency of rat femoral vein anastomoses. Microsurgery 13:138142,
1992.
9. Liu SQ, Fung YC: Rheology of blood vessels in diabetes. Biorheology
29:443457, 1992.