Activity and transport of antithrombin during acute limb ischemia

ORIGINAL ARTICLE

Activity and transport of antithrombin during acute limb ischemia Paul E. Burke, M D , F R C S , F R C S ( I ) , C o l u m F. Harvey, F R C S ( E d ) , Charlene L. Phythian, BS, Carol A. Gervin, P h D , and Lazar J. Greenfield, M D ,

Dublin, Ireland Republic, Belfast, Northern Ireland, Richmond, Va., and Ann Arbor, Mich. Antithrombin (AT III), a major circulating anticoagulant, may be influenced by ischemiainduced changes in microvasoxlar integrity and contribute to localized hypercoagulability. In a nonheparinized intact canine hindlimb model we determined AT III activity by chromogenic substrate assay (S-2238); coagulation changes with fibrinogen, activated partial thromboplastin time (aPTr), and prothrombin time (PT); and transvas~aflar exchange by lymph-to-plasma total protein concentration ratio. Femoral venous plasma and lymph samples were assayed during I hour o f steady state (C), 6 or 8 hours o f aortoiliac occlusion (I), and I or 3 hours o f reperfusion (R). Four groups were studied: GI, sham operated (n = 5); GII, moderate ischemia (n = 7), arterial pressure 30% to 45% C, GIII, 6 hours o f severe ischemia (n = 7), arterial pressure 5% to 20% C; and GIV, 8 hours of severe ischemia (n = 5), arterial pressure 5% to 20% C. All parameters varied near baseline in the control group and the group with moderate ischemia. Fibrinogen decreased after 3 hours o f ischemia in GIII from 218 --- 38 to 175 ± 46 mg/dl (mean ± SEM) and in GIV from 254 - 39 to 201 - 44 mg/dl (p < 0.005) as a I ' T r and PT increased. All parameters returned to baseline on R in GIII only. Plasma AT H I decreased in GIV from 89% -+ 4.6% to 53.6% - 16.2% (p < 0.005) after 3 hours and remained low during late I and R. The decrease was less severe in GIII. Lymph AT III decreased after 3 hours from 33.5% --- 6.6% to 6.4% ± 4.2% (p < 0.01) in GIII and from 40% ± 3.9% to 15.5% ± 6.6% (p < 0.01) in GIV, remained low during late I in both groups, and returned to baseline on R only in GIII. The changes in AT H I occurred as lymph-to-plasma total protein ratio increased progressively throughout I (from 0.42 ± 0.02 to 0.64 ± 0.07 [p < 0.001] at 6 hours in GIII and from 0.46 ± 0.03 to 0.85 ± 0.20 [p < 0.01] at 8 hours in GIV). A sustained reduction in plasma and lymph AT III levels is observed during 8 hours ofischemia in contrast to the transient reduction during 6 hours o f ischemia. These changes appear to be associated with coagulation changes within the limb but independent of the microvascular permeability changes. (J VASe SURG 1989;9:740-6.)

Reduced levels o f antithrombin III (AT HI), a major circulating plasma protein anticoagulant, have been associated with thrombotic complications in vascular surgery. 13 McDaniel and Bergan 4 showed that AT III and other coagulation factors were reduced after distal bypass procedures. Ischemia o f From St. Vincent's Hospital, Dublin, Ireland Republic (Dr. Burke), Mater Hospital, Belfast, Northern Ireland (Dr. Harvey), the Depamnent of Surgery, Medical College of Virginia, Richmond,Va. (Ms. Phythian and Dr. Gervin),and the Department of Surgery, University of Michigan, Ann Arbor, Mich. (Dr. Greenfield). Supported in part by National Institutes of Health Grant HL32619. Reprint requests: Carol A. Gervin,PhD, Department of Surgery, Medical Collegeof Virginia, P.O. Box 6, MCV Station, Richmond, VA 23298-0006. 740

the primate limb, induced experimentally by tourniquet occlusion, was associated with a signh'icant depletion o f coagulation factors in the limb, including A T III. s This was assumed to be consistent with disseminated intravascular coagulation induced by ischemia. Whether the changes described in the systemic circulation are greater in locally ischemic tissue or are the prime determinant o f intravascular thrombosis leading to unsuccessful revascularization remains unknown. The post-capillary venule has been identified as a site at which thrombosis occurs on reperfusion o f ischemic tissue. 6-s The pathogenic mechanisms at a microvascular level leading to these changes are not understood. The effects o f ischemia on capillary structure and endothelial function are not often con-

Volume 9 Number 5 May 1989

1.0

Antithrombin activity in the ischemic limb 741

steady tate

I

0.8 ~

ischemia

I

reperfusion

steady

1.0 , statel

o Group I ® Group II • Group III

"-

ischemia

I reperfusion

Iv

0.6 0.4 0.2 ~ *p
vshrl

0

0

0

TIME (hours) Fig. 1. Lymph-to-plasma total protein concentration ratio during steady state, 6 hours of ischemia, and 3 hours of reperfusion comparing moderate ischemia (group II) and severe ischemia (group III) to control dogs (group I). sidered as factors that contribute significantly to the thrombotic changes in the ischemic limb. Pearce9 suggested that endothelial damage secondary to vascular disease leads to platelet aggregation and a stimulation of the extrinsic clotting system through release ofthromboplastin. Zdeblick et al.1° showed that failure of limb reimplantation was associated with a higher concentration of fibrin aggregates in the venous effluent and postulated an imbalance in the coagulation-fibrinolytic system. AT III is the principal plasma circulating anticoagulant. 11 There are a number of mechanisms by which ischemia could alter AT III activity through an effect on capillary structure and endothelial cell function. A generalized increase in intravascular coagulation would lead to a consumption of AT III. It has been shown that endothelial bound heparin proteoglycans accelerate the inactivation ofthrombin by AT III. 12Any endothelial damage could remove this stimulus, thus preventing the inhibition of subcoagulant amounts ofthrombin and allowing an earlier and more rapid activation of the coagulation cascade. Finally, AT III is a protein with a molecular weight of 58,000 daltons. An increase in capillary permeability could lead to increased transvascular movement of this protein, as This could lead to intravascular deficiency of antithrombin, which would predispose to thrombosis. Lymph flow protein analysis during experimental limb ischemia provides us with a method of detecting changes in capillary integrity and increased leakage of proteins from plasma into the interstitium.~4 These

1

2

3

4

5

6

7

8

9

10

TIME (hours) Fig. 2. Lymph-to-plasma total protein concentration ratio comparing the 8-hour severe ischemia group (group IV) with the control group (group I) over 10 hours.

changes may also serve as an indicator of general endothelial function and injtny. We have used this technique in a model of an intac~ ischemic limb that closely resembles the clinical setting of acute arterial occlusion. By studying protein and AT III in plasma and lymph, as well as the general parameters of coagulation, we have been able to analyze the effects of ischemia on AT III activity and transport and its role in the induction of coagulation changes during limb ischemia. MATERIAL A N D M E T H O D S Mongrel dogs of either sex (20 to 25 kg) were anesthetized with sodium pentobarbital ( 30 mg/kg), supplemented as required. The animals were intubated with a cuffed endotracheal tube and allowed to breathe room air. Dogs were ventilated (Edco Scientific, Inc., Chapel Hill, N.C.) if their arterial oxygen tension fell below 80 torr. Body temperature was maintained at 39 ° C with a thermal heat blanket. Fluid loss was replaced by I to 2 L of saline solution (0.9%) during the 12 to 14 hours of the study. All animal experiments were approved by the animal care committee of the Medical College of Virginia and were in accordance with the "Principles of Laboratory Animal Care" and the !'Guide for the Use of Laboratory Animals" (NIH publication No. 80-23, revised 1985). Surgical preparation. The right hindlimb was used as the experimental limb in all studies. The infrarenal aorta and iliac and femoral arteries were exposed through a long retroperitoneal incision, taking

742

Journal of VASCULAR SURGERY

Burke et al.

Table I. Femoral venous plasma fibrinogen, aPTT, and prothrombin time Group I

II

Ill

IV

Time (hr) 1 4 7 8 9 10 1 4 7 8 9 10 1 4 7 8 9 10 1 4 7 8 9 10

F ibrinogen (mg / dl) 196 177 196 212 294 238 170 154 168 176 171 181 218 175 191 218 217 247 254 201 183 192 190 195

-+ 34 + 33 + 38 -+ 34 + 116 -+ 47.8* +- 37.5 +- 29 -+ 31 + 43 + 36 + 34.5 -+ 39 -+ 46 + 48 +- 34 + 36 -+ 43 + 39 +- 4 5 J -+ 42~z _+ 47 + 475 -+ 5 1 t

aPTT (see) 13.9 14.9 16.0 15.6 16.0 16.1 13.0 14.1 14.8 16.2 15.5 14.1 15.8 41.4 20.3 16.8 18.1 20.2 13.9 32.0 32.0

_+ 0.5 + 0.7 + 0.3 + 1.0 + 0.7* _+ 0.6* _+ 1.1 _+ 1.1 + 1.2 + 1.0 +_ 1.7 _+ 1.1" _+ 1.2 +_ 15.2" -+ 5.0 + 0.8 _+ 1.8 -+ 4.1 + 0.6 _+ 17.0 __ 17.0 -32.6 + 16.9 36.8 _+ 21.1

PT (sec) 7.8 8.1 7.5 7.6 7.4 8.3 8.3 8.0 9.0 7.6 7.0 8.4 36.1 14.7 8.1 8.0 8.2 7.5 25.7 25.4 23.8 24.0 23.8

-2-_0.1 + 0.4 +- 0.4 -+ 0.4 -+ 0.4 + 0.3 + 0.3 _+ 0.3 -+ 1.2 + 0.4 + 1.0 -+ 0.3 _+ 16.5 -+ 6.9 _+ 0.5 + 0.5 _+ 0.5 ~ 0.8 + 18.6 -- 18.7 _+ 19.1 + 19.0 _+ 19.1

PT, Prothro mb in time. Values are mean -+ SEM. *p < 0.01; t p < 0,005; ~p < 0.001.

care to preserve all adjacent lymphatics. All infrarenal branches of the aorta, right external iliac artery, and right femoral artery were ligated. The distal iliac vessels were cannulated via a large medial side branch (the deep femoral artery and vein) and the Venous cannula was directed down the femoral vein to midthigh level. Cannulas were also placed in the common carotid artery and internal jugular vein. All cannulas were used for pressure monitoring and blood sampiing and were kept patent by a nonheparinized saline flush. A detailed description of the lymph model and hemodynamic monitoring techniques has been given previously.13 Lymph was collected from a femoral lymphatic of the deep lymphatic system. Lymph flow was maintained by passive flexion of the hindpaw by a pulley system (50 cycles/rain). Lymph flow was recorded at 15-minute intervals unless the volume was less than 0.25 m g - - t h e n 30- and 60-minute intervals were used. The lymph was collected in preweighed graduated vials containing 20 txl of sodium citrate (3.8%) and was used to determine total protein concentration (in grams per deciliter) and AT III activity (in percent). Femoral arterial blood was sampled every 15 minutes (30 minutes during ischemia) for total protein concentration and hematocrit. Femoral venous samples were collected hourly to determine

plasma fibrinogen levels (in milligrams per deciliter), activated partial thromboplastin time (in seconds), prothrombin time (in seconds), and AT III activity (in percent). Experimental protocol. All animals were studied for a 10-hour period: 1 hour of steady state, 6 or 8 hours of ischemia, and 1 or 3 hours of reperfusion. The animals were divided into four groups: group I (n = 5), controls: sham-operated animals in which lymph was collected over 10 hours without alterations in femoral pressure; group II (n = 7): moderate ischemia (6 hours), complete occlusion of the infrarenal aorta and both external iliac arteries, which reduced femoral artery pressure to 30% to 35% of steady-state levels; group III (n = 7): severe ischemia (6 hours), femoral artery pressure was reduced to 5% to 20% of steady-state levels by occlusion of the right external iliac artery (additional aortic and left external iliac artery occlusion was required in four of seven animals); and group IV (n = 5): severe ischemia (8 hours), pressure reduction was similar to that of group III, with severe ischemia followcd by 1 hour ofreperfusion. Additional clamping was required in two of the five animals. Blood and lymph analysis. Blood samples were centrifuged at 3000 rpm for 10 minutes at - 4 ° C and the plasma was stored at - 7 0 ° C. Lymph was

Volume 9 Number 5 May 1989

Antithrombin activity in the ischemic limb 743

Table II. Coagulation parameters of central venous plasma Group I

II

III

IV

Time (hr) 1 4 7 8 9 10 1 4 7 8 9 10 1 4 7 8 9 10 1 4 7 8 9 10

F ibrinogen (mg/ dl) 177.0 139.2 161.4 212.6 294.0 238.8 155.0 148.0 149.3 172.0 155.2 158.7 154.3 170.1 125.8 101.3 128.7 276.7 239.4 195.4 145.9 175.0 190.4 157.8

_+ 55.5 -+ 44.1 ~ 47.4 _ 34.2 -+ 116.0 _+ 47.8* + 26.2 -+ 28.7 _+ 24.5 -+ 43.2 + 40.5 + 42.9 _+ 25.4 -+ 25.3 + 32.5 +_ 26.8 + 65.7 + 54.8~ + 34.2 _+ 43.6 + 43.0* -+ 56.3 + 47.6 _+ 67.8

aPTT (sec) 12.6 13.3 14.5 14.8 15.1 14.6 13.5 13.1 12.6 13.4 13.2 13.3 13.1 13.5 13.0 13.6 15.5 14.6 13.8 13.4 15.8 16.3 20.0 19.6

+ 0.6 + 1.0 _+ 0.6* _+ 0 . 5 t + 1.9 _+ 1.1" _+ 1.5 + 1.1 + 1.6 _+ 1.9 + 1.0 + 0.9 + 0.5 _+ 0.8 _+ 0.6 + 2.1 + 0.4 _+ 0.5 + 0.9 + 0.8 + 1.9 _+ 1.5 + 3.0t _+ 3 . 3 t

PT (sec), 7.8 8.1 7.5 7.6 7.4 9.3 8.9 8.5 8.8 8.3 7.6 8.6 9.7 8.9 9.2 8.5 7.5 7.4 8.8 8.0 8.9

+ 0.1 + 0.4 +_ 0.4 -+ 0.4 -+ 0.4 + 0.6 +- 0.4 + 0.5 + 0.6 -+ 0.7 + 1.3 -+ 0.5 + 1.4 + 0.5 _+ 0.7 -+ 0.5 + 0.2 + 0.3 + 1.4 + 0.8 -+ 1.5

I T , P r o t h r o m b i n time. Values are mean _+ SEM.

*p < 0.01; tp < 0.005; :~p < 0.001.

centrifuged before storage only if red cells appeared to be present. Total protein concentration was determined by Bradford's method.15 Lymph protein concentrations were mathematically corrected according to the volume ratio of l)nnph to citrate present, and the corrected values were consistent with dilution curves for hindlimb lymph in citrate determined in our laboratory. Jugular and femoral venous samples were collected (4.5 ml) in vials containing 0.5 ml of sodium citrate. The venous plasma samples were analyzed for fibrinogen as described by Clauss, ~6 prothrombin times as described by Quick, ~7 activated partial thromboplastin time (alXlT) as described by Proctor and Rappaport, 18 and AT III activity as described by the two-stage method of Odegard et al.19 Lymph AT III activity was also determined by this method. Statistics. Data are expressed as means and standard errors. Comparisons were made between steady state (hour 1) and subsequent hours in each group. The analysis of variance repeated-measures model with Bonferroni's correction was used to compare all data. 2° RESULTS

The lymph-to-plasma total protein concentration ratio varied near baseline in the control (group I)

and moderate ischemia (group II) groups. In groups III and IV (Figs. 1 and 2) there was a significant increase in the lymph-to-plasma total protein concentration ratio, progressing from 0.42 ___ 0.02 to 0.64 - 0.07 (p < 0.001) after 5 hours of ischemia in group III and from 0.46 _+ 0.03 to 0.91 _+ 0.22 (p < 0.01) after 7 hours in group IV, with no further increase occurring during reperfusion. General coagulation parameters of femoral venous plasma and central venous plasma are summarized in Tables I and II, respectively. All parameters in femoral venous plasma (Table I) varied near baseline in the control and moderate ischemia groups. In the 6-hour severe ischemia group (group III) fibrinogen decreased transiently afi:er 3 hours as alXl-T increased transiently but significantly. In the 8-hour severe ischemia group (group IV) fibrinogen decreased significantly after 3 hours and remained low throughout the ischemia period. Prothrombin time and aPTT showed a corresponding increase, which also was constant but never reached statistical significance. Coagulation parameters in central venous plasma (Table II) remained within normal limits throughout ischemia and reperfusion in the four groups except for fibrinogen in group IV, which decreased in parallel with the femoral results from 239.4 ___ 36.2 to

744

Journal of VASCULAR SURGERY

Burke et al.

Table IH. AT III activity of systemic plasma, limb plasma, and hindlimb lymph A T IH (%) Group I

II

III

IV

Time (hr) 1 4 7 8 9 10 1 4 7 8 9 10 1 4 7 8 9 10 1 4 7 8 9 10

Systemic plasma 64.6 63.8 50.3 54.1 40.8 70.6 56.8 54.0 51.4 42.7 72.9 63.5 54.1 42.8 45.4 46.5 40.8 32.5 24.0 20.6

_+ 13.3 ~+ 6.5 + 6.3 _ 7.8 -+ 9.5 ± 29.6 _+ 28.5 + 31.1 + 23.0 -_+ 23.5 _+ 16.6 + 15.8 _+ 15.3 + 24.2 -+ 16.5 + 25.4 _+ 20.7 + 18.9 _+ 18.8 -_+ 12.8

Femoral plasma 96.9 91.8 93.3 85.9 89.6 79.6 77.4 77.7 73.3 64.4 70.4 65.0 92.1 80.1 85.9 73.1 86.9 73.1 89.0 53.6 40.0

+ 3.1 + 5.6 + 4.6 _+ 3.4 +_ 5.6 +_ 10.0 + 5.2 _ 7.8 _+ 6.0 + 7.4 + 6.2 + 8.0* _+ 5.6 + 8.8 - 5.1 + 8.0t _+ 5.5 _+ 8.1~: _+ 4.6 _+ 16.2~ + 11.05 -52.9 + 2 . 2 t 52.8 _+ 6 . 1 t

Hindlimb lymph 30.3 28.1 24.3 26.0 17.1 32.0 21.7 21.8 26.9 23.8 30.4 26.3 33.5 6.4 19.3 33.7 24.4 18.0 40.0 15.5 17.5

-+ 5.2 -+ 7.8 -+ 9.2 + 6.7 -+ 9.9 +- 4.7 _+ 6.5 -+ 3.0 -+ 5.3 + 4.3 + 2.5 + 0.9 -+ 6.6 +- 4.2* +- 8.0 + 9.6 + 7.8 -+ 6.0 -+ 3.9 -+ 6.5* -+ 7.2*

18.0 + 6.0* 13.0 + 8.3

Values are mean -+ SEM. *p < 0.01; t p < 0.005; :~p < 0.001.

145.4 + 43.0 (p < 0.01) at hour 7. Fibrinogen levels, like those in femoral plasma, did not return fiflly to baseline. A progressive rise in aPTT levels from 13.8 + 0.9 to 20.0 + 3.0 seconds also was noted in this group after 8 hours of ischemia, which paralleled the femoral venous Plasma. AT III activity. Systemic venous plasma, femoral venous plasma, and lymph AT III activity are summarized in Table III. The values in systemic plasma were consistently lower than the correspondmg femoral plasma in all groups. Control and moderate ischemia caused minimal changes in femoral venous plasma and lymph AT III activity during the 10 hours of study (Fig. 3). Plasma and lymph AT III levels decreased wtihin 3 hours of ischemia in group IV and remained significantly low for the remainder of ischemia and reperfusion (Fig. 4). In group III (Fig. 3) lymph AT III levels also decreased during ischemia but returned to baseline on reperfusion. Femoral plasma levels in group III remained within normal range. DISCUSSION The multiple-ligation model of limb ischemia is preferable to the tourniquet model when studying intravascular coagulation changes. Wu and

Mansfield21 showed that fibrinolytic activity was increased significantly during tourniquet occlusion, and Miller et al. s have acknowledged that the trauma of tourniquet occlusion was probably responsible for tissue thromboplastin release and intravascular coagulation in their tourniquet model. Cafferrata et al.6 used a similar model to that described in this report for the study of coagnlation changes in plasma. Their findings agree with our results confirming the value of the model for studies of intravascular coagulation. The simultaneous study of coagulation parameters in lymph, as well as plasma, during limb ischemia has not been reported previously, although Leach and Browse22 have looked at other parameters of hemostasis in hindlimb lymph. The reduction in lymph AT III activity in association with plasma levels during 8 hours of severe ischemia appears to confirm that intravascular coagulation is occurring in the peripheral venous plasma, as well as at the microcirculatory level. Several groups 6.s have shown experimentally that microvascular thrombosis occurs in the early reperfusion period and may prevent reflow. Our results suggest that the changes leading to occlusive thrombosis may begin during ischemia at a microcirculatory level. The relevance of lymph and plasma changes to intravascular thrombosis was

Volume 9 Number 5 May 1989

120

Antithrombin activity in the ischemic limb 745

ischemia

m

100 80

120

plasma

o Group I 0 Group II • Group Ill

I *p <0.01 **p < 0.005 + p<0.001

q'--.."~'L-+-_*"V

8o-

"1"-':1

60~

40-

.o

2Oo r0

1

", 2

ischemia

I

I

loo

1-

60

I

3

4 5 6 TIME (hours)

7

8

~ ~ . . ~ . .~. • ~ * *** ,, I. lymph

l

o Groupl ***~<'0u.'&'5 o'Gr°up'v+P<°°°'

9

10

0

1

2

3

I *

~ I *

4 5 6 7 TIME (hours)

.'~"~

8

"

1

9

10

Fig. 3. AT III activityof femoral venous plasma and lymph in control (group I), moderate ischemia (group II), and 6-hour severe ischemia groups (group III).

Fig. 4. AT III activityin femoral venous plasma and lymph in control (group I) and 8-hour severe ischemia groups (group IV).

given added significance when iliac artery blood flow could not be resumed and a "no reflow" phenomenon developed in two of the five animals in the 8-hour severe ischemia group. In the absence of any change in plasma activity the transient reduction in lymph AT III in the 6-hour severe ischemia group cannot be explained fully on the basis of generalized intravascular coagulation. The changes in hemostasis leading to increased coagtdability of blood in ischemic tissue may begin in the capillaries and not appear in femoral plasma because of hemodiludon unless very large changes occur. Therefore the reduction in lymph AT III may reflect changes in microvascular hemostasis that are consistent with intravascular thrombosis. The changes in ~ and prothrombin time along with the transient hypofibrinogenemia suggest that intravascttlar coagulation and perhaps an increased fibrinolysis in response to the thrombosis were occurring at the time of the reduction in lymph AT III levels. This suggests that lymph AT III reflected intravascular coagulation at the microcirculatory level. We can only assume that the site of microvascular thrombosis was distal. This would support our conclusion that lymph AT III is a more sensitive indicator of coagulation changes than plasma AT III. Our initial hypothesis was that the reduction in plasma AT ][II activity was secondary to increased transvascular movement of AT III from plasma to lymph during ischemia. The generalized increase in transvascular protein movement documented by the increased lymph-to-plasma total protein concentration ratio la should occur with the small-molecule AT III. However, the reduction in AT III activity in

lymph during ischemia despite the increase in total protein concentration suggests that transvascular protein movement was not the mechanism by which plasma AT III activity was decreased. The general changes in other coagulation parameters suggest that intravascular coagulation was the primary determinant of AT III activity. The antithrombin/thrornbin interaction is said to be accelerated by the presence of endogenous heparin proteoglycans that lie along the endothelium. t2 The increased lymph.-to-plasma total protein concentration ratio implies capillary leakage and endothelial damage. Impaired function of the endothelial lining for a period exceeding 8 hours of ischemia may contribute to a significant and irreversible reduction in AT III levels. Rapid inactivation of subcoagulant amounts of thrombin by AT III may alter the balance between the activators and inhibitors of the coagulation cascade. An overriding thrombosis and a generalized consumptive coagulopathy could occur. Further study of the effect of normal and ischemia altered cndothelium on AT III activity is warranted to determine the importance of endothelial-botmd heparin and AT III in the prevention of thrombosis. The significant reduction in plasma and lymph AT III activity during 8 hours of ischemia was associated with sustained, although insignificant, changes in the other coagulation parameters. This would be consistent with reports of inappropriate thrombosis after vascular procedures in patients demonstrating signs of acquired .AT III deficiency) The maintenance of AT III activity may therefore be important in the prevention of irreversible coagulation

746

Burke et al.

changes, and methods that help to maintain AT III activity, such as warfarin therapy, may have an important role in maintaining patency of vessels in ischemic tissue. 2s Further studies should help to determine whether AT III activity is related to the maintenance of prolonged patency rates as was demonstrated recently in distal bypass procedures. 23 In summary, when the intact hindlimb model of limb ischemia is used it appears that intravascular coagulation is the primary determinant of AT III activity. There is tittle evidence to support increased transvascular movement of AT III across the damaged endothelial capillaries despite its small molecular size. The significant reduction in lymph AT III levels during ischemia suggests that coagulation may be occurring at a microvascular level, even though it is not apparent in peripheral plasma. When the latter does occur it may be associated with irreversible changes in intravascular coagulation. REFERENCES

1. Towne JB, Bernhard VM, Hussey C, Garancis JC. Antithrombin deficiency: a cause of unexplained thrombosis in vascular surgery. Surgery 1981;89:735-42. 2. Karl R, Garlick I, Zarins C, Cheng E, Chadiak J. Surgical implications of antithrombin III deficiency. Surgery 1981; 89:429-33. 3. Shapiro ME, Rodvien R, Bauer KA, Salzman EW. Acute aortic thrombosis in antithrombin III deficiency. JAMA 1981;245:1759-61. 4. McDaniel WR, Bergan JJ. Sequential changes in coagulation and platelet fimcfion following femoro-tibial bypass. J VASC SURG 1984;1:261-8. 5. Miller SH, Eyster ME, Saleem A, Gottlieb L, Buck D, Graham WP. Intravascular coagulation and fibrinolysis within primate extremities during tourniquet ischemia. Ann Surg 1979;190:227-30. 6. Cafferrata I-IT, Robinson AJ, Blaisdeli FW. Coagulation changes in regional ischemia. Surg Forum 1968;19:31-2. 7. Dunant JH, Edwards WS. Small vessel occlusion in the extremity after various periods of arterial occlusion: an experimental study. Surgery 1973;73:240-5.

Journal of VASCULAR SURGERY

8. Eriksson E, Anderson WA, Replogle RL. Effects of prolonged ischemia on muscle microcirculation in the cat. Surg Forum 1974;25:254-5. 9. Pearce WH. Coagulation abnormalities in patients undergoing lower extremity revascularization. In: Kempczinski RE, ed. The ischemic leg. Chicago: Year Book Medical Publishers, 1985:233-45. 10. Zdeblick TA, Shaffer JW, Field GA. An ischemia-induced model of revascularization failure of replanted limb. J Hand Surg 1985;10;125-31. 11. Taylor FV. Survey and new description of the hemostasis system. Surv Synth Pathol Res 1983;1:251-73. 12. Marcum JA, McKenney JB, Rosenberg RD. Acceleration of thrombin-anti-thrombin complex formation in rat hindquarters via heparinlike molecules bound to the endotheliurn. J Clin Invest 1984;74:341-50. 13. Burke PE, Harvey CF, Murphy LM, Gervin CA, Greenfield LJ. Transvascular protein movement in the intact ischemic hindlimb. J Surg Res 1987;43:351-9. 14. Taylor AE. Capillary fluid filtration: starling forces and lymph flow. Circ Res 1981;49:557-75. 15. Bradford MM. A rapid and sensitive method for quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72:248-54. 16. Clauss A. Gerrinnungsphysiologische Schnellrnethode zur Bestimmung des Fibrinogens. Acta Haematol (Basel) 1957; 17:237-46. 17. Quick AJ. Hemorrhagic diseases and thrombosis. 2nd ed. Philadelphia: Lea & Fcbiger, 1966;391-5. 18. Proctor RR, Rappaport SI. The partial thromboplastin test with kaolin, a simple screening test for first stage plasma dotting factor deficiencies. Am J Clin Pathol 1961;36: 212-9. 19. Odegard OR, Lie M, AbildgaardU. Heparin cofactor activity measured with an amidolytic method, Thromb Res 1975;6: 287-94. 20. WaUenstein S, Zucker CL, Fleiss JL. Some statistical methods useful in circulation research. Circ Shock 1981;8:503-17. 21. Wu AVO, Mansfield A. The fibrinolytic activity of the vein following venous stasis. Br J Surg 1979;66:637-9. 22. Leach RE, Browse NL. Effect of venous hypertcusion on canine hindlimb lymph. Br J Surg 1985;72:275-8. 23. Kretschmer G, Wenzl E, Wagner O, et al. Influence of anticoagulant treatment in preventing graft occlusion following saphenous vein bypass for femoropopliteal occlusive disease. Br J Surg 1986;73:689-92.