The efficacy of dextran 40 in preventing early postoperative thrombosis following difficult lower extremity bypass

The efficacy of dextran 40 in preventing early postoperative thrombosis following difficult lower extremity bypass R o b e r t B. R u t h e r f o r d , M.D., Darrell N. Jones, Ph.D., Sven-Erik Bergentz, M.D., David Bergqvist, M.D., Allastair M. K a r m o d y , M.D., H e r b e r t Dardik, M.D., Wesley S. M o o r e , M.D., J e r r y Goldstone, M.D., William R. Flinn, M.D., A n t h o n y J. C o m e r o t a , M.D., William J. Fry, M.D., and Dhiraj M. Shah, M.D.,

Denver, Colo., Malmb) Sweden, Albany, N.Y., Englewood, NJ., Los Angeles, Calif., Tucson, Ariz., Chicago, Ill., Philadelphia, Pa., Dallas, Tex. In a randomized, mtdticenter trial the efficacy" of intravenous dextran 40 (D-40; Rheomacrodex) in preventing early postoperative thrombosis was tested in the following difficult lower extremity bypasses: (1) femoropopfiteal with poor runoff using autologous vein (AV), (2) femoropopliteal using grafts other than AV, (3) single or sequential bypasses to infrapopliteal arteries, and (4) the above-mentioned bypasses with adjunctive procedures that destroy adjacent endothelial surfaces (e.g., thrombectomy and endarterectomy). Five units of D-40 was administered to the experimental group at 75 to 100 ml/hr, two during and immediately after operation and one each 3 days postoperatively. Antiplatelet drugs were withheld until i week after surgery. The overall 1-week occlusion rate was 6.9% (5 of 73) with D-40 and 20.5% (17 of 83) for controls, which is statistically significant. Particularly significant was the difference in group 3: 0% (0 of 28) for D-40 and 27.8% (10 of 36) for controls. In the same time period there were no occlusions in group I and only one occlusion each with AV grafts in both D-40 and control groups, 2.6% (1 of 38). When grafts other than AV were used, the ocdusion rate was significantly lower at I week for D-40, 11.4% (4 of 35), than controls, 35.6% (16 of 45). By i month the protective effect of D-40 was partially lost-- 15.3% (11 of 72) for D-40 and 20.7% (17 of 82) for controls (no statistical significance). Use of antiplatelet drugs during this period had no discernible effect. (J VASC SURG 1984; 1: 765-773.)

Certain peripheral arterial bypasses carry an unacceptably high risk o f early postoperative thrombosis but, once early patency is achieved, fail at a much lower rate. Characteristically these procedures _.ivolve anastomosing into small-caliber distal arteries, implanting grafts whose luminal surfaces are initially quite thrombogenic or are associated with adjunctive procedures (e.g., endarterectomy and thrombectomy) that disturb or destroy adjacent endothelial surfaces. Such difficult bypasses carry a From the University of Colorado Health Services Center, Denver, the University of Lund-Malm6 General Hospital, Maim6, Sweden, Albany Medical College, Albany, N.Y., Englewood Hospital, Englewood, N.J., the University of California in Los Angeles Health ServicesCenter, the Universityof Arizona Health Services Center, Tucson, Northwestern University Medical School, Chicago, Temple University Health Services Center, Philadelphia, and the University of Texas Health Services Center, Dallas. Presented at the Thirty-eighth Annual Meeting of the Societyfor Vascular Surgery, Atlanta, Ga., June 7-8, 1984. Reprint requests: Robert B. Rutherford, M.D., U.C.H.S.C., Box C312, 4200 E. 9th Ave., Denver, CO 80262.

postoperative thrombosis rate close to 20%. A number o f measures have been tried in an attempt to minimize this risk in addition to the routine use o f completion angiography to rule out technical errors. These measures are generally intended either to increase flow (e.g., arteriovenous fistulas and sympathetic blockade) or decrease coagulability (e.g., heparin and antiplatelet drugs). Infusions o f dextran 40 (D-40; Rhcomacrodex) produce both these effects, and in a limited clinical trial their administration has been found to protect early patency in lower extremity bypasses under selected circumstances. 1 This report presents the initial results o f a randomized, multicenter trial o f D-40 in difficult loveer extremity arterial bypasses. MATERIAL AND METHODS Four categories o f bypass to the popliteal or infrapopliteal arteries considered to be at high risk for early postoperative thrombosis were selected for this trial. These are outlined in Table I. F r o m January 765

766

Journal of VASCULAR SURGERY

Rutherford et al.

Table I. Types of bypass tested in D-40 trial

Table III. Case distribution by participating center

Category

Description

I

Femoropopliteal bypass using autologous vein with poor rtmofff Femoropopliteal bypass using grafts other than autologous vein (PTFE, umbilical vein, Dacron, composite

II III IV

Center

grafts) Single or sequentialbypassto infrapoplitealarteries Bypass to Popliteal or infrapopliteral arteries associated with adjunctiveprocedurest (endarterectomy, thrombectomy,angioplasty,transluminaldilatation)

*Less than two patent infrapopliteal arteries. t Sympathectomy, arteriovenous fistulas excluded.

Table II. Disqualifications in D-40 trial Reasons

D-40

Procedure not performed D-40 stopped D-40 administered Antiplatelet drugs given Not properly randomized Improper D-40 administration Improper secondary operation Total

3 11 -8 -3 2 27

Control 7 -4 11 4 -2 28

Total 10 11 4 19 4 3 4 55

1983 to May 1984, 211 patients from nine participating centers who were expected to undergo one of these four types o f bypass were considered for entry into this study. After obtaining informed consent, randomization to control or D-40 treatment was accomplished by opening one of a series of sequentially numbered envelopes color-coded according to category of operation. Each category was separately numbered and separately randomized. The control group was treated in routine fashion in terms of blood replacement and intravenous fluid administration. The experimental group received five bottles (500 ml/bottle) of D-40 according to the following protocol: bottle 1 was infused at 100 m l / h r after induction of anesthesia; bottle 2 was given at 75 m l / h r immediately following bottle 1; and bottles 3, 4, and 5 were administered on each of the next three postoperative days, also at 75 ml/hr. In both groups, heparin was administered intravenously (100 U/kg) prior to cross-clamping and placement o f the graft. Its effect was reversed by protamine sulfate only if excessive bleeding was noted at the time o f wound closure or additional doses of heparin had been administered. No further heparin was given and antiplatelet drugs were withheld from 2 days before surgery to the seventh postoperative day and then administered according to each center's routine.

Albany Medical College, Albany, N.Y. Englewood Hospital, Englewood, N.J. Maim6 General Hospital, Malm6, Sweden University of California, Los Angeles University of Colorado, Denver University of Arizona, Tucson Northwestern University, Chicago Temple University Hospital, Philadelphia University of Texas, Dallas Total

Cases 58 21 15 13 12 11 10 10 6 156

Specific patient exclusion criteria included: (1) known or suspected bleeding disorders (including preoperative use of antiplatelet or anticoagulant drugs), (2) recent or recurrent history of congestive heart failure, (3) significantly decreased renal f u ~ tion, (4) history of hypersensitivity reaction to D-40, and (5) previous entry into this study. Causes for premature termination from the study (death, graft occlusion, graft removal for sepsis, amputation, hemorrhage, pulmonary edema, drug reaction, etc.) were identified, as were protocol violations, failure to randomize, or other disqualifying irregularities. The D-40 and control groups were compared for patency at 1 day, 1 week, and 1 month, with further breakdown according to operation category and graft type. Patency was accepted without anglographic proof only if a significant increase in the ankle/brachial index (>0.15) had been achieved and maintained. Patency maintained by thrombectomy or any further operative procedure was disallowed regardless of successful outcome. Intergroup comparisons were also made with regard to risk factors, clinical status (indications), anteriographic rune52, preoperative ankle/brachial index, and location of distal anastomosis. Finally, the treatment and control groups were compared with regard to treatment-specific complications, particularly for evidence of excessive bleeding (intraoperative hypotension, estimated blood loss, number of transfusions, and hemorrhagic wound complications), signs of volume overload (congestive heart failure, pulmonary edema, restriction of intravenous fluids, and use of diuretics), healing complications or wound infections, as well as major nonspecific complications such as myocardial infarction, renal failure, or pulmonary embolism. The time and cause of death, graft failure, and amputation or reoperation for any cause were recorded in each case.

Volume 1 Number6 November1984

Dextran 40 following di~cult lower extremity bypass 767

Table IV. Occlusion rate at 1 week: Comparison by operative category (156 cases)

Table V. Occlusion rate at 1 week: Comparison by graft type (156 cases)

Occlusion at 1 week D-40 Category

Description

No.

Occlusion at 1 week

Control %

No.

D-40 %

I

Femoropopliteal,AV, 0/10 0 0/9 0 poor runoff II Femoropopliteal, 2/17 11.8 3/20 15.0 graft other than AV IU Femorotibial/peroneal0/28 0 10/36 27.8* (single/sequential) IV Any above bypass 3/I8 1 6 . 7 4/18 22.2 with adjunctive procedure Total 5/73 6.9 17/83 20.5*

Graft type

2,,70.

%

AV PFFE Umbilical vein (UV) Other than AV? All grafts

1/38 0/19 4/i5 4/35 5/73

2.6 0 26.7 11.4 6.9

%

1/38 2.6 4/20 20 12/22 54.5 1 6 / 4 5 35.6* 1 7 / 8 3 20.5*

Table VI. Occlusion rate vs. postoperative interval Postoperative interval

Disqualifications. The reasons for disqualification are listed in Table II. In 10 patients no procedure or a procedure other than one represented by the four categories described previously was performed. In nine, D-40 was stopped for perceived cause and in two, for failure to obtain valid informed consent. In four, D-40 was given to patients randornized to the control group. In 19 (11 controls and eight D-40 patients), antiplatelet drugs were inadvertently administered within the period prohibited by protocol. Four patients were not properly randomized, in three the protocol for D-40 administration was not followed, and four were disqualified for improper secondary operations (e.g., arteriovenous fistulas). Graft patency data were available for 39 of the 55 disqualified patients. Occlusion ares at 1 week in these cases do not favor either treatment group (controls = 2 o f 17 occlusions, D-40 patients = 2 o f 22). Except for analysis o f the ,complications o f D-40 infusion, the following observations concern only the remaining 156 qualified patients. The distribution of qualified patients entered by each of the nine participating medical centers is indicated in Table III. Patency. Table IV compares the occlusion rate at 1 week for D-40 patients and controls with a breakdown according to operation category. Table V also compares the 1-week occlusion rates for the two groups broken down according to graft type. Table VI compares the occlusion rates for the two groups at 1 day, 1 week, and 1 month. There is no difference in the occlusion rate between controls and those receiving D-40 in category I (discontinued

No.

*Statisticallysignificant. *PTFE + UV + Dacron.

*Statisticallysigni6cant.

OBSERVATIONS

Control

1 day 1 week 1 month?

D-40 No. 0/73 5/73

11/72

Control %

No.

%

0

8/73 17/83 17/82

9.6" 20.5* 20.7

6.8 15.3

*Statis6cally significant. ~One death in each group. after 19 cases because o f the lack o f thrombosis) or in procedures in which autologous vein was employed. There are mean differences in the occlusion rate at 1 week, favoring D-40, for the other operation categories and for other graft types. These differences are statistically significant (chi square, p < 0.05) for category UI operations and grafts other than autologous vein. Overall patency is statistically significantly better for the D-40 group at 1 day and I week but not at I month. The pattern o f occlusion for D-40 vs. control patients is shown in Fig. 1. The median day o f occlusion for the control group is day 3, whereas the median day o f occlusion for the D-40 group is day 8 (p < 0.001 by MannWhitney U test). Autologous vein grafts used in situ or reversed clearly had a significantly lower early postoperative thrombosis rate than other grafts. Patency differences between polytetrafluoroethylene (PTFE) and umbilical vein grafts were, however, also associated with marked differences in location of distal anastomosis, as noted below. Comparability analyses. Tables VII through X compare the makeup of the D-40 and control groups by distribution o f operative category, distribution o f graft type, and site o f distal anastomosis and by clinical classification, runoff, and preoperative ankle/brachial index. They show that randomization by operative category also resulted in no sig-

768

Journal of VASCULAR SURGERY

Rutherford et al.

10--

o

"" o

5

median= day 3

•Q

~

5

median

= day

8

10 1

5

10

15

20

2:5

30

Fig. 1. Postoperative occlusions for D-40 and control groups graphically compared in regard to postoperative day of occurrence.

Table VII. Distribution by category D-40 (n =73)

Control (n =83)

Category

No.

%

No.

%

I II I]7i IV

10 17 28 18

13.7 23.3 38.4 24.6

9 20 36 18

10.8 24.1 43.4 21.7

Table VIII. Distribution by graft type Graft type

D-40

Control

AV, reversed AV, in situ PTFE Umbilical vein Dacron/composite

15 23 19 15 1 73

11 27 20 22 3 83

nificant differences between the D-40 and control groups in distribution o f graft type, severity of disease as represented by clinical classification, runoff, and ankle/brachial index or by site of distal anastomosis. Graft-type groups are comparable in all but two factors: (1) average clinical classification is significantly higher (worse) for autologous vein than for PTFE and (2) there are significant differences in the level of distal anastomosis between the umbilical vein and the PTFE groups. A total o f 61.5% of

PTFE grafts were placed to the above-knee popliteal segment, whereas 83.8% of umbilical grafts were carried to either the below-knee popliteal or infrapopliteal arteries. Complications. O f the 201 patients in whom one of the four categories o f lower extremity bypasses was performed (i.e., excluding the 10 with no or othcr procedures; Table II), 101 received D-40 and 100 did not. The reportcd complications and deaths in these two groups are listed in Table XI. Five patients died during the 30-day period of observation; two controls and three who received D-40. Two deaths occurred in each group from myocardial infarction. An additional death in the D-40 group was attributed to respiratory failure on day 10. This patient exhibited respiratory distrc~ immediately after surgery so that the second bottle of D-40 was not given and the remaining three were administered at only 25 ml/hr, which wag cause for disqualification. One additional D-40 patient had a nonfatal myocardial infarction. Five D-40 patients and three control patients had evidence o f probabl~ fluid overload (intravenous fluids restricted, diuretics administered, or frank pulmonary edema observed). Ten patients in each group had actual or probable bleeding complications (wound hematoma or excessive wound bleeding intraoperatively, gastrointestinal bleeding, hypotension, or excessiw transfusions). The average amount of blood transfused was higher for controls (not significant), b~ the only two episodes o f gastrointestinal bleeding

Volume 1 Number 6 November 1984

Dextran 40 following difficult lower extremity bypass 769

Table IX. Site of distal anastomosis

Table X. Comparison of disease severity

Site

D-40

Control

Total

Popliteal (above knee) Popliteal (below knee) Peroneal-tibial mink Posterior tibial Anterior tibial Peroneal Dorsalis pedis Total

16 26 3 9 8

18 23 3 10 14

i0

iI

1 73

4 83

34 49 6 19 22 21 5 156

occurred in the D-40 group. Hypotension was less common in the D-40 group, but healing complications and/or wound infections were the same. Among the other complications singly represented were a transient ischemic attack and paraplegia appearing on day 3 in the D-40 group and multiple~-rgan failure in a control patient. Although the listed complications do not appear to be significantly different in number or kind, it should be noted that D-40 infusion was stopped in nine instances for perceived cause (myocardial infarction, three; fluid overload, three; and hypotension, excessive wound bleeding, and paraplegia, one each). DISCUSSION

Although for practical purposes early postoperative thrombosis of bypass grafts is considered to be caused by technical error until proved otherwise, it is well recognized that spontaneous thrombosis may also occur when factors classically considered to be part of Virchow's triad are operative. Disturbed intimal surfaces (after endarterectomy, angioplasty, thrombectomy, or transluminal dilatation) or thrombogenic graft surfaces lacking an active endothelium combined with the low flows [ommonly seen when small-caliber bypasses are carried down to distal arteries would seem challenge enough. Added to this has been the recent demonstration that patients undergoing distal bypass have increased coagulability, with increased platelet reactivity to collagen, increased factor VIII antigen activity, and decreased levels of antithrombin III.~ It is no wonder many distal bypasses carry a much higher early thrombosis rate than can be explained on the basis o f technical error; the 1-week occlusion rate in controls in this study of 20.5% is typical. Attempts to increase flow by creating an arteriovenous fistula at or beyond the distal anastomosis not only introduces significant technical complexities but has produced mixed results, a,4 Some support can be found for concomitant sympathectomy

Poor runoff* Average clinical classificafiont Average preoperative anlde/brachial index

D-40

Control

47% 3.02 0.34

47% 3.49 0.39

(No significant differences) * More than half of distal arteries significantly obstructed. 1 = clandication; 2 = rest pain; 3 = threatened tissue loss; 4 = tissue loss; 5 = tissue loss with infection.

Table XI. Comparison of reported complications and death Reported complications Myocardial infarction Fluldoverload, congestive

failure Bleeding Intraoperative wound Postoperative wound Gastrointestinal Hypotension Healing complications, infection Transient ischemicattack Paraplegia Generalized rash Multiple-organ failure Total

D-40 (n = 101)

Control (n = 100)

3 (2)* 5 (1)*

2 (2)* 3

2 4 2 2 2

4 2 -4 2

1 1 1 --

---1

22

18

*No. of deaths in parentheses.

in experimental 5,6 and clinical 7 reports, but this adds a separate operative procedure. On the other hand, turning to commonly used antithrombotic drugs presents the problem of finding an agent that is both effective and safe. Experimental studies have shown improved short-term patency with antiplatelet drugs, 8,9 but after numerous clinical trials, only two have been able to show improved short-term patency, one in above-knce femoropopliteal bypass 1° and the other in coronary artery bypasses, n None have shown patency protection for extremity bypasses during the critical immediate postoperative period. At the other extreme, whereas heparin is considered to be the most effective agent in prevcnting postoperative thrombosis, maintaining full heparinization for several days beyond the actual period of opcration carries a scrious risk of hemorrhagic complications, one most vascular surgeons are willing to accept only after graft thrombectomy. Warfarin is theoretically attractive because it is known to improve antithrombin III

770

Rutherford et al.

levels, which may be reduced in as much as 20% of patients undergoing distal bypass 12 and which have been implicated in patients with repeated graft thrombosis. 13 However, this slow-acting oral medication is awkward to use in the perioperative period, and it has not been shown to be effective in protecting patency of femoropopliteal bypass grafts? 4 The initial results o f this multicenter trial indicate that infusions of D-40 offer significant protection against early postoperative thrombosis following difficult lower extremity bypasses without significantly increasing the risk of bleeding or other treatment-related complications, thus offering an attractive compromise between heparin and antiplatelet drugs. D-40 is a high molecular weight polysaccharide produced from sucrose by the action of Leuconostoc mesenteroides. Fractionation and hydrolysis produce a product with a mean molecular weight of 40,000. D-40 decreases platelet adhesiveness, 15 reduces factor VIII activity, 16 and increases clot lysability. 1~ In addition, it decreases viscosity, reduces surface thrombogenicity by either a coating effect or change in surface electronegativity,a8 and as a volume expander, significantly increases peripheral flow. x9-21 There has been considerable debate over whether the demonstrated benefits of D-40 in preventing thrombosis in experimental trials and other clinical settings is because of its increasing flow or decreasing coagulability. Regardless of the relative contributions o f these two effects, it has been demonstrated to be effective experimentally in improving patencies with small-caliber arterial anastomoses, ~2 graft interpositions, '3 and arterial thrombosis/embolectomy preparations. 24"25 Despite its demonstrated effectiveness experimentally in the arterial system, D-40's major clinical application as an antithrombotic agent has been in the prevention o f deep venous thrombosis, in which its efficacy is now well established, z6 Its clinical effectiveness in preventing thrombosis after arterial reconstruction has not been thoroughly tested. Foster et alY noted its value in protecting arterial reconstructions in a 1966 report, and one limited randomized trial in Switzerland was carried out by WaibeP in 1976. In the latter report, which included 64 D-40 patients and 67 controls, a protective effect for D-40 could be demonstrated in patients with distal bypasses and with poor runoff. This multicenter trial deliberately focused on those types of distal bypass that might have a high enough spontaneous rate of thrombosis to allow demonstration of a statistically significant effect for D-40 with a reasonable number of cases. Femoro-

Journal of VASCULAR SURGERY

popliteal bypasses using autologous vein in patients with good runoff were excluded from the outset, and after the absence of postoperative thrombosis in the first 19 cases, such bypasses in those with poor runoff (category I) were discontinued. Even bypasses to the peroneal and tibial arteries using autologous vein rarely thrombosed in this study, 3.5% (2 of 57). Yet, despite the inclusion of 76 bypasses in which autologous vein was used (66% with the in situ technique), a statistically significant overall protection against graft thrombosis was demonstrated for D-40 at 1 day and 1 week. In addition to this threefold reduction in the overall occlusion rate 3 days beyond its last infusion, D-40 reduced thrombosis during the same time period in peroneal-tibial bypasses and in bypasses using PTFE grafts from 27.8% and 20%, respectively, to zero and roughly halved the thrombosis rate of umbilical vein g r a ~ The I-week occlusion rate for controls for all bypasses in which a graft other than saphenous vein was used was 26.7% compared with 11.4% in those receiving D-40. The generally higher early thrombosis rate for umbilical vein than for PTFE grafts probably can be explained by the fact that the majority (64.5%) of the latter were used in above-knee femoropopliteal bypass compared with only 16.2% of umbilical vein grafts placed above the knee. This different utilization of the two grafts may reflect the demonstration in two major multicenter trials that, on the one hand, above-knee femoropopliteal bypass grafts with PTFE give patency rates equivalent to autologous vein~8 and, on the other hand, that umbilical vein gave better patency than PTFE grafts when carried to the below-knee popliteal artery. 29 Although it may be possible to show a significant protective effect for D-40 extending up t~'~ month with a larger number o f cases, comparison of patency rates at 1 week and 1 month suggests that as much as half of the protective effect of D-40 is lost during that interval period. Whether or not antiplatelet drugs begun preoperatively and/or low-dose warfarin begun on the first or second postoperative day could consolidate this benefit is speculative but deserves study. Our data revealed only that no protection was afforded those receiving antiplatelet drugs beginning on the seventh postoperative day. If one accepts the evidence presented here that infusions of D-40 provide significant patency protection in the immediate postoperative period, a risk-benefit question still remains--does one pay an unreasonable price for this benefit in terms of corn-

Volume 1 Number 6 November1984

plications caused by D-40? In this study there was no statistically significant difference between the control and the experimental groups in the rate of death or serious complications (e.g., myocardial infarction) or even, for that matter, in the incidence of Needing complications or the frequency of fluid overload. Furthermore, there were no instances o f anaphylactic reaction. Nevertheless, the possibility of the latter must be accepted, even though severe reactions occur only in 1 out of 10,000 administrations and are almost eliminated by the use of the hapten dextran 1000 (Promit). Furthermore, fluid overload in elderly patients with known cardiac disease, even if they have not recently been in congestive failure, must be carefully monitored. This complication resulted in the discontinuation of D-40 in 3% of patients in this study, and in others the infus;on rate had to be slowed and/or diuretics administered. However, when the stated reasons for patient exclusion are observed and the protocol for its administration carefully followed, the risk of D-40 administration would seem reasonably small and far outweighed by the patency protection it offers to patients undergoing difficult distal bypass. Particularly if its protective effect can be consolidated with the concomitant use of antiplatelet drugs or warfarin, it would appear that a very significant patency protection benefit is offered by the use of D-40 infusions in certain patients undergoing difficult lower extremity arterial bypass, particularly those in whom grafts other than autologous vein must be used or the distal anastomosis is carried down to an infrapopliteal artery or a diffusely diseased artery with poor runoff. REFERENCES

2.

3.

4.

5.

6.

Waibel P. Antikoagulation in der ge fasschirurgie. Vasa 1976; 5:107. McDaniel MD, Pearce WH, Yao JST, Rossi EC, Fahey VA, Green D, Flinn WR, Bergan JJ. Sequential changes in coagulation and platelet function following femorofibial bypass. J VASe SURG1984; 1:261-8. Snyder Jr SO, Gregory RT, Wheeler JR, Bayle RG. Failure of arteriovenous fistulas at distal tibial bypass anastomotic sites. Presented at the Sixteenth World Congress, The International Society for Cardiovascular Surgery, Rio de Janeiro, Sept 19, 1983. Dardik H, Sussman B, Ibrahim IM, et al. Distal arteriovenous fistula as an adjunct to maintaining arterial and graft patency for limb salvage. Surgery 1983; 94:478. Castcn DF, Sadley AH, Foreman D. An experimental study of sympathectomy on patency of small vessel anastomoses. Surg Gynecol Obstet 1962; 114:462. Williams GD, Crumpler JB, Campbell GS. Effects of sympathectomy on the severely traumatized artery. Arch Surg 1970; 101:704.

Dextran 40 following difficult lower extremity bypass 771

7. Cutler BS, Thompson JE, Kleinsasser LB, Hcnpel GK. Autologous saphenous vein femoropopliteai bypass. Analysis of 298 cases. Arch Surg 1967; 94:592. 8. Oblath RW, Buckler FO, et al. Prevention ofplatelet aggregation and adherence to prosthetic grafts by aspirin and dipyridamole. Surgery 1978; 84:37. 9. Allen BT, Long JA, Welch MJ, et al. Effect of aspirin therapy and its withdrawal on control and endothelial cell-seeded grafts. Surg Forum 1983; 34:470. 10. Green RM, Roedersheimer LB, DeWeese JA. Effects of aspirin and dipyridamole on expanded polytetraftuoroethytene patency. SurgeD, 1982; 92:1016. 11. Chesebro JH, Clements IP, Fnrter V, et al. A platelet inhibitor drug trial in coronary artery bypass operations. N Engl J Med 1982; 307:73. 12. Pearce WH. Discussion of McDaniel et al. J VAsc SUF.G 1984; 1:268. 13. Towne JB, Bernhard VM, Hussey C, Garancis JC. Antithrombin deficiency--A cause of unexplained thrombosis in vascular surgery. Surgery 1981; 89:735. 14. Lewis JL, Salzman EW. Antithrombotic therapy. In: Rutherford RB, ed. Vascular surgery, ed 2. WB Saunders Co, 1984. 15. Cronberg S, Robertson B, Nilsson IM, Nilehn JE. Suppressive effect of dextran on platelet adhesiveness. Thromb Diath Haemorrh 1966; 16:384. 16. Aberg M, Hedner U, Bergentz BE. Effect of dextran on factor VIII (antihemophilic factor) and platelet fimction. Ann Surg 1979; 189:243. 17. Abcrg M, Bergentz SE, Hedncr U. The effect ofdextran on the lysability ofex vivo thrombi. Ann Surg 1975; 181:342. 18. Ross S, Ebert R. Microelectrophoresis of blood platelets and the effect of dcxtran. J Clin Invest 1955; 38:135. 19. Bergan JJ, Trippcl OH, Kaupp A, Kukral JC, Nowlin WF. Low molecular weight dextran in treatment of severe ischemia. Arch Surg 1965; 91:338. 20. Bienenstock J, Harding ELT. Low-molecular-weight dextran (Rheomacrodex) in ischemic ulceration of the skin. Lancet 1964; 1:524. 21. Brown P. The effect of low molecular-weight dextran on the microcirculation of skin. Br J Hast Surg 1967; 20:272. 22. Sasamota Y. Experimental studies on continuous local infusion of anticoagulants for reconstruction of smaii arteries. Kumamoto Med J 1970; 44:839. 23. Eiken O. Thrombotic occlusion of experimental grafts as a function of the regional blood flow. Acta Chit Scand 1961; 121:210. 24. Laha R, Israeli J, Dujovny M, Barrionuevo P, Hellstrom H. Low molecular weight dextran in experimental embolectomy. Stroke 1980; 11:59. 25. Hunt PS, Reeve TS. Experimental arterial thrombus prevention and renal fxmetion in sheep after infusion with low molecular weight dextran and dextran 70,000, Ann Surg 1967; 166:51. 26. Bergentz SE. Dextran prophylaxis of venous thromboembolism in venous problems. In: Bergan JJ, Yao JST, eds. Chicago: Year Book Medical Publishers Inc, 1978; 529-40, 27. Foster J, Killen D, Jolly P, Kirtley J. Low molecular weight dextran in vascular surgery. Prevention of early thrombosis following arterial reconstruction in 85 cases. Ann Surg 1966; 163:764. 28. Bergan JJ, Veith FJ, Bernhard VM, Yao JS, Flinn WR, Gupta SK, Scher LA, Samson RH, Towne lB. Randomiza-

772

Rutherford et al.

tion of autogenous vein and polytetrafluoroethylene grafts in femorodistal reconstruction. Surgery 1982; 92:921. 29. Eickhoff IH, Buchardt-Hansen HJ, Brom6 A, Ericsson BF, Kondt KF, Mouritzen C, Myhre HO, Norgren L, Rostad H,

DISCUSSION Dr. W. Sterling Edwards (Albuquerque, N.M.). All of us would agree that this is an impressive study on the use o f dextran to keep prosthetic grafts open in difficult cases. There are two comments I would like to make and one o f them is rather obvious. Why don't the authors do a study in which they continue the dextran a little longer if occlusions occur on the fifth or seventh day? Perhaps this would make an even more significant addition to the patency rate of these difficult grafts. The second comment relates to a study that we were performing about 20 years ago stimulated by the work o f Milton Bryant who developed an animal model for studying drugs such as dextran. This animal model was the simple performance o f an endarterectomy on the femoral artery o f dogs. After a 2 cm endarterectomy o f the femoral artery, in 100% of the dogs the arteries occlude in a few hours, a very good model. When the dogs were given perioperative dextran, about 80% o f the arteries stayed open indefinitely. Heparin and streptokinase were not as effective if they were used perioperatively. If, in addition, we irrigated the inside of the endarterectomized artery at the time o f operation with dextran, we increased the patency rate to perhaps 90%, which simply leads me to wonder if the patency rate would increase if you soaked the synthetic graft in dextran before implanting it. Another thought comes to mind. Would perioperative intravenous dextran plus local irrigation with dextran help prevent the occasional carotid endarterectomy thrombosis, a problem we see occasionally? Dr. Frank J. Veith (New York, N.Y.). We at Montefiore participated in Dr. Rutherford's original study, but our data were not included because all 32 o f our patients received perioperative aspirin and dipyridamole (Persantine). The data from our 32 patients alone are inconclusive with regard to patency, although our dextran patients had a higher incidence o f systemic complications. Therefore, before we all switch to the dextran bandwagon, I would like to be sure that the effect of dextran cannot be reproduced by aspirin and dipyridamole alone. Certainly these agents may be safer than dextran. This means that Dr. Rutherford will have to continue his study and compare bypasses in patients receiving perioperative aspirin/dipytidamole and dextran with those in patients getting aspirin and dipyridamole alone. I would also hope that studies with longer follow-up are done to see if the dextran effect persists. I would like to ask Dr. Rutherford if he is planning these additional stud-

Journal of VASCULAR SURGERY

Trippestad A, A randomized trial of PTFE versus human umbilical vein for femoropopliteal bypass surgery: Preliminary results. Br 1 Surg 1983; 70:85.

ies and to commend him and his group on a very interesting presentation. Dr. Philip N. Sawyer (Brooklyn, N.Y.). As part of our 35-year study o f the vascular tree, in 1968 we started to look at the effect o f a whole series o f anticoagulant agents on the physical chemistry of the vascular tree. While looking at dextran, we wondered why everybody had given up using it as an electrochemically active agent. At the same time Dr. Chou at Columbia described why dextran worked. It produced a pseudo yon Willebrant's disease. So we decided to use dextran from about 1968 to 1973, at which time Dr. Kuntz asked me why we did not do a randomized study. Finally, we did. We put an entire series o f patients on dextran who had had vascular reconstructions, randomizing them by statistical techniques vs. no anticoagulant therapy postoperatively. Then on the fifth postoperative day we switched them to heparin, 20,000 units subcutaneously once a day, to get the electrochemical effects of heparin. The effects of this procedure in a statistically significant study were reported in 1978. Not only did we have an effective decrease in postoperative thrombosis in an entire series o f different types of grafts in man, but also most postoperative complications were eliminated (postoperative myocardial infarctions, strokes, and thrombophlebitis). You will note that we used very small doses o f dextran, between 15 and 20 ml/hr, to get electrochemical nonanticoagulating effects. Scanning electron micrographs show dextran absolutely prevents adhesion of platelets to injured tissue. platelets adhered to the animals perfused with dextran. We believe that there is incontrovertible evidence that dextran does what Dr. Rutherford says. I think it can be used in much smaller dosages than he is now using. If you switch to a subcutaneous heparin regimen following dextran, keep the patient on that therapeutic regimen postoperatively to prevent late thrombosis, that is, 4 days to 3 or' 4 months, the incidence of thrombosis again decreases. Simultaneously you can start to taper heparin and switch to aspirin and dipyridamole as heparin tolerance decreases about the fourth to fifth week. We discharge our patients on a dosage of 20,000 units of heparin a day subcutaneously and monitor them at 2, 4, and 8 weeks. It alb seems to work well. We have now used this procedure for 4500 to 5000 patients. The incidence o f severe complications with dextran at the low-dosage regimen of 15 to 30 m l / h r seems to be tolerable, the incidence o f pulmonary edema being very small.

Volume 1 Number 6 November1984

Dr. R u t h e r f o r d (closing). Dr. Edwards, we used a Swedish protocol because the study was conceived after discussions with Dr. Sven-Erik Bergentz and that particular protocol had been quite effective in their hands. In fact, they do arterial surgery without systemic heparin. They just use heparin solutions for irrigation. We have considered administering dextran an extra day or two to see whether we could prevent those thromboses that occur just after discontinuation o f dextran but could not change our protocol in mid-study. I might add a comment in regard to your study in dogs. One o f the effects o f dextran is an ability to increase clot lysability, and this may have contributed to the improved patency in your study. Dr. Veith, you have raised a valid question about the relative safety and efficacy o f these two different ap-

Dextran 40following difficult lower extremity bypass 773

proaches. Based on our previous discussions and discussions with others, we plan in our next study to compare three groups. One group will receive aspirin and dipyridamole, another will receive dextran, and a third group will receive both dextran and aspirin and dipyridamole. In that study we also plan a longer follow-up, probably 6 months to a year. Dr. Sawyer, I thank you for your supportive remarks and sharing an experience in which dextran was found to be effective. You bring up the question whether we could not use smaller doses. As I said before, we followed the protocol used in Malm6, Sweden, which deliberately used a more rapid infusion rate to obtain a flow effect in addition to its effects on factor VIII and platelets. When given this way, it stays in the circulation for at least 6 to 8 hours after the end o f the infusion.