Primary anastomotic bonding in polytetrafluoroethylene grafts?

Primary anastomotic bonding in polytctrafluorocthylcnc grafts? William J. Quifiones-Baldrich, M.D., Stanley Ziomek, M.D., Theodore Henderson, B.A., and Wesley S. Moore, M.D.,

Los Angeles, Calif. Previous studies have demonstrated that standard knitted and woven fabric grafts are forever dependent on the suture material for anastomotic tensile strength. Clinical experience with polytetrafluoroethylene (PTFE) and double velour knitted grafts have shown that there is extensive fibrous capsular bonding between the graft and the surrounding tissues. This would lead to increased anastomotic tensile strength. To test this theory, 34 mongrel dogs underwent replacement of their infrarenal aortas with grafts made of PTFE (10 dogs), of double velour knitted Dacron (DVD, 11 dogs), of single velour knitted Dacron (SVD, 5 dogs), and of woven Dacron (WD, 8 dogs). One anastomosis was constructed with 5-0 Prolene and the opposite anastomosis was constructed with 5-0 Dexon (average absorption time, 21 days). In five grafts each of PTFE and DVD, as well as in all eight WD grafts, the midgraft was divided and resutured with 5-0 Dexon. All grafts were harvested together with adjacent proximal and distal aorta between 3 and 10 months from the time of implantation. The tensile strength of each anastomosis was measured with a tensiometer. The mean graft-to-artery (absorbable suture) anastomotic tensile strength, in pounds, for PTFE (14.3) and DVD (I2.6) was significantly higher than that for SVD (6.9) or WD (7.2) (p < 0.003). Graft-to-graft anastomotic tensile strength for PTFE (mean I7.3) was significantly better than that for DVD (mean 9.0; p < 0.03) or WD (mean 7.9; p < 0.001). Analysis ofanastomotic tensile strength as a function of time revealed continued increase in PTFE in contrast to a slow declirm with time in DVD. All 10 dogs in the PTFE and WD groups with graft-to-graft anastomosis (Dexon) were alive after 10 months, with intact anastomoses. Three of five dogs with DVD grafts died between 19 and 35 days following operation because of anastomotic disruption. On the basis of tensile strength and survival data, we conclude that there is evidence of improved primary anastomotic bonding in PTFE grafts and DVD grafts when contrasted to single velour knitted and woven Dacron grafts. Graft-to-graft anastomotic bonding appears superior in PTFE grafts when compared with DVD and VIFD. (J VAse SURG 1987;5:311-8.)

Late disruption of prosthctic graft-to-artery anastomosis remains an unsolved problem, with an estimated incidence from 3% to 24%. ~-4 Clearly, this complication is multifactorial and corrcction of factors identified in thc earlier era of vascular surgery, such as the use of silk sutures, s has resulted in improved long-term results. Somc of the etiologic factors in anastomosis breakdown and pseudoaneurysm formation, such as hypertension, proximity to a From the Section of Vascular Surgery, University of California at Los Angeles School of Medicine, Los Angeles, Calif. Presented at the Thirty-fourth Scientific Meeting of the North American Chapter, International Society for Cardiovascular Surgery, New Orleans, La., June i0-11, I986. Reprint requests: William J. Quifiones-Baldrich, M.D., Section of Vascular Surgery, Department of SurgeDr/General CHS 72-248, School of Medicine, University of California at Los Angeles, 10833 LeConte Ave., Los Angeles, CA 90024.

joint, 6 and infcction,s,7 are beyond the control of the surgeon. Other factors, such as endarterectomy of the recipient vesscl 7 and postoperativc anticoagalation, 8 may bc rcquired to obtain the best possible initial result. There rcmains a common factor in the process of pseudoaneurysm formation, which is disruption of the integral bond between the artery and the graft. Expcrimental and clinical studies have suggested that a graft-to-artery anastomosis is forever depcndent on the suture material for its integrity. 9,1° Using expcrimental animals, Moore et al. H confirmed the greater tensilc strength of the anastomosis afforded by the suture; however, some degrcc of strength was present without the suture. This suggested a fibrous tissue bond bctween the graft and thc rccipient artetT. In a study of graft healing, Malonc et al.~2 concluded 311

312 Quihones-Baldrich et al.

Fig. 1. A, Specimen mounted in tensiometcr. B, P e a k tensile strength and anastomosis separation. C, Partial disruption of capsule. D, Complete capsule disruption.

that significant differences exist in healing characteristics between Dacron and polytetrafluoroethylene (PTFE) prostheses. The velour type of Dacron prosthesis appeared to have improved healing compared with the nonvelour Dacron prosthesis. The purpose of this experiment was to determine the strength of the anastomotic fibrous bond in various prosthetic grafts that are currently in use. In addition, we evaluated the bond that occurs in a graft-to-graft anastomosis. We selected PTFE, double velour knitted Dacron (DVD), single velour knitted Dacron (SVD), and woven Dacron (WD) grafts. An attempt was made to determine the healing characteristics at the anastomosis as a function of time. MATERIAL A N D METHODS In thirty-four adult mongrel dogs (weight, 15 to 25 kg) of either sex the infrarenal aorta was replaced from the level of the renal arteries to the aortic trifurcation with one of four different prosthetic grafts. Group I (n = 10) had replacement with 6 mm PTFE grafts (Gore-Tex, W. L. Gore & Assoc., Huntington Beach, Calif.), group II (n = 11) with double velour knitted Dacron grafts (knitted Cooley, Meadox Medicals, Inc., Oakland, N.J.), group III (n = 5) with single velour knitted Dacron grafts (Exs Vascular Prosthesis, USCI, Billerica, Mass.), and group

]'ournal of VASCULAR SURGERY

IV (n = 8) with woven Dacron grafts (DeBakey arterial prosthesis, USCI). An end-to-end anastomosis (proximal or distal) was performed with 5.0 polypropylene (nonabsorbable) suture, and the opposite anastomosis was performed in an identical fashion with 5-0 braided polyglycolic acid (absorbable) suture (Dexon, Davis & Geck, Los Angeles, Calif.). In addition, five grafts in group I, five grafts in group II, and all grafts (n = 8) in group IV had a third graft-to-graft anastomosis in the midportion of the prosthesis done with the use of a 5-0 polyglycolic acid (absorbable) suture. This midanastomosis was performed before implantation. The graft was transected and reanastomosed. All anastomoses were done with a single continuous suture that was started posteriorly. All experiments were done with the animals under general endotracheal anesthesia (halothane), and aseptic conditions wcrc maintained. 3+Systemic heparinization was achieved with intravenous heparin (1 mg/kg) before aortic clamping. At the conclusion of graft implantation, the retroperitonemn was closed with a continuous O-Dexon suture. The animals were allowed to recover and maintained in the animal care facility until the time of harvest. If any abnormality suggestive of a problem with the graft was observed during the maturity period, the animal was killed, the graft was harvested, and findings were recorded. After an interval of between 3 and 10 months, the animals were reanesthetized and their grafts were harvested along with at least 2 cm of native aorta at either end. The entire graft and the surrounding fibrous capsule were included in the specimen. Particular care was taken to prevent disruption of the 1~ brous capsule or traction at any of the anastomoses. Any unusual findings, such as contained hematoma or pseudoaneuusm, were noted. The animal was killed at the end of the harvesting procedure (intravenous Euthanol 5 ml/10 kg). Explantcd grafts were divided transversely into two or three pieces; each piece contained an anastomosis in the middle of the specimen. Tensile strength of each anastomosis was measured with an Instron model CC tcnsiometer (Universal Testing Instrument, Instron Corp., Los Alamitos, Calif.) properly calibrated to 50 pounds (22.7 kg) full scale. Pulling speed was calibrated at 0.2 inch (0.508 cm) ~This research was conducted under the guidelines of the "Principles of Laboratory Animal Care" and the "Guide for Care and Use of Laboratory Animals" (NIH Publication No. 80-23, revised 1978).

Volume 5 Number 2 February 1987

Primary anastomotic bonding in PTFE grafls? 313

!Fi

. . . . . . . .

Z72'5"L~ ]?iti~ i :~ > ! + { ................

+}4:, ............ ~: 7k~ 22 2Z:Z

~4--l---g-g-

b-,~

', a

~

a

e~ ; } ~ :

;

4-..+- ÷~~--

,2 ->

7.i- Z~

-2-++277, -gTT~ +-+-.~-

4. ~+--+-+ 4

~-~

.~-÷.--~*-4 ~ - ~ - t - - + -

W

R

Fig. 2. Tensiometer recording of tensile strength for PTFE gra•to-artery absorbable-suture (/eft) anastomosis and nonabsorbable-suture (right) anastomosis (0 to 1 = 5 pounds). Table I. Anastomotic tensile strength--PTFE

Graft-to-graft

Dog No.

Time (too)

Nonabsorbable

Absorbable

(absorbable)

1 2 3 4 5 6 7 8 9 10

9.5 9 8 7.5 3.5 6 6 6 5.5 4

27 22.5 20.0 14.5 18.0 16.0 21.0 18.0 20,5 23.5

20.0 15.0 13.5 14.0 12.0 10.0 18.0 NA 17.0 10.0

21~5 16.0 --18.0 ---20.0 11.0

N A = not available. NOTE: Tensile strength is in pounds per 1.88 cm.

per minute. Tensile strength readings were recorded continuously until disruption of the anastomosis occurred, at which time peak tensile strength was determined. This was expressed in pounds per circumference of the graft (6 mm diameter × v = 1.88 cm). The specimen was mounted with the anastomosis intact, appropriate-sized bolts on both ends, and held in place with steel bands (Fig. 1). To avoid slippage of the tissue, sandpaper was placed between the tissue and the bands. Peak tensile strength was recorded for each anastomosis (Fig. 2) and analyzed according to graft, suture material, location, and time of implantation. An analysis of variance with pairwise comparison was carried out. A difference was considered significant

at the 95% confidence level (p < 0.05). Tensile strength as a function of time was analyzed with the use of a general linear model calculated by least square fitting. RESULTS All animals survived thc operative procedure. Three dogs in group II (double velour knitted Dacron) died suddenly at 17, 19, and 37 days following implantation. Necropsy revealed a disrupted proximal anastomosis (one dog) and midanastomoses (two dogs). These anastomoses were performed with absorbable (Dexon) sutures. The remaining 31 grafts were available Ibr harvest 3 to 10 months following implantation. One graft in group I (PTFE) was

Journal of VASCULAR SURGERY

314 Qui~ones-Baldrich et aI.

Table II. Anastomotic tensile strength of double velour knitted Dacron Dog No.

Time (mo)

Nonabsorbable

Absorbable

11 12 13 14 15 16 17 18

3 7 7 6 9 9 9 19 days

22.5 18.0 18.5 28.0 19.0 20.0 20.0 32.0

18.0 8.0 10.0 14.0 6.0 17.5 13.0 14.5

19 20

3.5 17 days

17.5 NA

21

35 days

NA

Graft-to-graft (~bsorbable) ------8.0 Disruption and death 10.0 NA

13.0 Disruption and death NA

Disruption and death

NA = not available. NOTE: Tensile strength is in pounds per 1.88 cm.

Table III. Anastomotic tensile strength of single velour knitted Dacron Dog No, 22 23 24 25 26

Table IV. Anastomotic tensile strength of woven Dacron

Time Graft-to-graft (mo) Nonabsorbable Absorbable (absorbable) 5 4.5 4.5 4.5 4.5

8.0 10.5 19.0 23.0 19.0

6.5 9.0 6.5 7.0 6.0 ~

m

NOTE: Tensile strength is in pounds per 1.88 cm. ~Anastomotic pseudoaneurysm.

thrombosed at the time of harvcst. Thc anastomotic tensile strength of this particular graft was similar to that of patent grafts in this group and therefore is included in the data. The cause of thrombosis is unclear, and the thrombosis may have occurred at the time of harvest. A pseudoaneurysm of the distal (Dexon) anastomosis was seen in one graft in group III (single velour knitted Dacron). No pseudoaneurysms or anastomotic disruption was seen in either group I (PTFE) or group IV (woven Dacron). Tensile strength measurements are presented, according to suture material and graft type, in Tables I through IV. Mean tensile strength for each prosthesis according to the suture used is presented in Table V. A statistically significant difference between the tensile strength of PTFE and double velour knitted Dacron mid graft-to-graft anastomosis was seen, with superior strength evident in the PTFE graft. In this regard, PTFE graft-to-graft anastomosis was significantly stronger than any of the other tested prostheses (p < 0.03). The survival data correlate, to some degree, with this finding. Whereas

Dog. No. 27 28 29 30 31 32 33 34

Graft-to-graft Time (too) Nonabsorbable Absorbable (absorbable) 10 10 8.5 8 8 8 8 7

12.5 14.5 15.5 12.0 18.0 15.0 10.7 20.0

6.5 8.5 8.5 8.7 6.0 11.0 8.5 5.0

10,5 9.5 3.0 12.0 ~+ 11.2 8.5 1,7 7.0

NOTE: Tensile strengths in pounds per 1.88 cm. *Midanastomosis may have been included in tensiometer clamp.

there were three deaths from anastomotic disruption in the double velour knitted Dacron group, there were no such events in the PTFE group. Although one would expect a similar high incidence of disruption in the other two graft types, only one pseudoaneurysm in the single velour knitted group and no anastomotic disruption in the woven Dacron group were observed. There was no significant difference in the tensile strength of graft-to-artery anastomosis (absorbable suture) between PTFE and double velour knitted Dacron prostheses (p = 0.249). Both grafts, however, had superior graft-to-artery anastomotic tensile strength when compared with single velour Dacron (p < 0.003) or woven Dacron grafts (p < 0.0002). When anastomosis done with nonabsorbable suture material is considered, no statistically significant difference between PTFE and DVD was observed (p = 0.83). DVD had superior tensile strength

Volume 5 Number 2 February 1987

PrimaG anastomotic bonding in PTFE grafis? 315

E~ Z

20

O

Z

•

A

0 O_

16

II

A

A

I

C9 Z LU n~ 1-09 LL] _d O3 Z hJ

12 tt

•

A

A

8 PTFE o ~

4

A

DVD A - - -

t

I

2

4

T - -

1-"

6

8

t

i

!0

MONTHS Fig. 3. Tensile strength as a function of time in PTFE and double velour knitted Dacron (DVD) grafts. Note increase in tensile strength with time for PTFE.

when compared to SVD (p < 0.03) and WD (p < 0.02). Comparison of PTFE with DVD and SVD revealed a trend toward superior tensile strength for PTFE, although the difference was not statistically significant (p < 0.12 and p < 0.06, respectively). Analysis of tensile strength as a function of location of the anastomosis (proximal or distal) revealed no difference in either group (p > 0.3). Analysis of healing characteristics as a function of time reveals a gradual increase in anastomotic tensile strength between 3 and 9 months for PTFE grafts in contrast to a slow and gradual decrease in tensile strength throughout the study period for double velour knitted Dacron grafts (Fig. 3). The difference in the slope does not achieve statistical significance, given the number of observations in each group. This trend occurs regardless of suture type, which suggests that the graft material is a factor in this observation. This analysis was not performed for single velour knitted or woven grafts. All anastomoses that were done with absorbable sutures were disrupted at the graft-to-artery or graftto-graft junction during tensiometer measurements. Anastomotic disruption in those done with nonabsorbable sutures was caused by arterial tearing at suture holes, with intact graft and suture line, which suggests adequate strength in the graft and suture components. The anastomotic strength in this situation thus reflects the aortic tissue strength. Separate

Table V. Least square mean tensile strength of anastomosis

Prolene Dexon Graft-to-graft anastomosis

PTFE

DVD

SVD

WD

20.1 14.33* 17.3 ~

21.7 12.6 * 9.0

16.0 6.9 --

15.4 7.2 7.9

NOTE: Tensile strength is in pounds per 1.88 cm. *p < 0.01.

measurements o f aortic tissue "alone revealed an average tensile strength of 24.5 pounds. Thus the aorta

at the anastomosis is weaker and probably reflects the "perforated line" effect of the anastomosis. DISCUSSION

Anastomotic disruption is a late complication of arterial grafting. Its multifactorial origin suggests that we will continue to observe this complication despite improvements in technique, graft material, suture material, prevention of infection, and overall patient care. Regardless of the specific etiologic factors, a strong bond between the recipient artery and the prosthetic graft is desirable. Clearly, the suture material is important in maintaining the integrity of an arterial-graft anastomosis. Our experiments suggest, however, that in addition there occurs a fibrous tissue anastomotic bond that, in dogs, is capable of maintaining the integrity of the repair. The strength

316 Qui~ones-Baldrich et al.

of this bond appears to be dependent on the nature of the prosthesis and its healing characteristics. Our results suggest that PTFE and double velour knitted Dacron produce a stronger anastomotic bond between artery and graft than single velour knitted Dacron or woven Dacron. Tensile-strength measurements suggest superior graft-to-graft bonding in PTFE grafts. This correlates with survival data in which two of five dogs in the double velour knitted Dacron group died as a result of graft-to-graft anastomotic disruption. The observation of three anastomotic disruptions in the double velour knitted Dacron group as compared with none in the PTFE group must be interpreted with great caution. No anastomotic disruptions occurred, despite a longer overall observation period, in the woven-graft group. The latter had the lowest anastomotic tensile strength of any of the tested prostheses. If one accepts the premise that tensile strength correlates with healing, a longer observation period might have revealed a similar incidence of disruption in the various groups. On the other hand, all disruptions occurred early (19 to 35 days); thus it is not possible to reach any valid conclusion on the basis of these data alone. Species differences in healing characteristics must also be considered in the interpretation of our results. Edwards et al.9 reported their observations on anastomotic tensile strength in human beings. No evidence of fibrous tissue bond was found on qualitative testing of human aorta-prosthetic graft (Teflon and knitted Dacron) anastomosis. The method used (hand pulling) did not allow for quantitation. In an analysis of 1084 arterial prostheses implanted in human subjects and in experimental animals, Sauvage et al. 13 concluded that man has the slowest rate of transinterstice ingrowth. In this respect, dogs approximate human healing more closely than do pigs, calves, and baboons. A review of the literature reveals a large number of reports documenting the incidence and etiologic factors of pseudoaneurysms.7"8a4-18Of 401 reported cases of graft-to-artery pseudoaneurysms, 92% were associated with knitted or woven Dacron prostheses. This, however, is likely due to the fact that most surgeons prefer to use Dacron prostheses for aortofemoral reconstructions. The latter accounts for the majority of reported pseudoaneurysms. On the other hand, pseudoaneurysm formation with PTFE prostheses appears very rare. Haimov et a1.I9reported two pseudoaneurysms in 362 cases of femoropopliteal reconstruction. Both may have been associated with infection. This suggests an incidence of 0.5 %. In the UCLA series of 154 femoropopliteal bypasses with

Journal of VASCULAR SURGERY

PTFE, no pseudoaneurysms have been observed (unpublished data). Although the site of reconstruction may be an important factor in this apparent difference, our experimental data suggest an additional mechanism by which this difference may be explained. Previous experiments have determined that maximum anastomotic strength for Dacron prostheses is achieved at 3 to 4 weeks, with a slight decline thereafter.2° Our results confirm these findings for Dacron prostheses; however, PTFE anastomosis showed a continued slow increase in tensile strength for the duration of the experiment (9 months). The number of experiments is small, and therefore the reliability of this analysis is limited. This study in no way suggests that a graft-(', artery anastomosis with absorbable suture is a clinically viable technique. It has; however, allowed us to quantitate the contribution of the healing process to the strength of the anastomosis. It is interesting to compare the anastomotic tensile strengths of the various grafts when nonabsorbable suture material is used. When nonabsorbable suture material is used, both PTFE and double velour knitted Dacron have higher tensile anastomotic strength than single velour knitted Dacron and woven Dacron. The same phenomenon is observed when absorbable suture material is used, which suggests that the increased strength seen in the former may result from improved anastomotie healing. Other investigators have reported no difference in anastomotic tensile strength between woven and knitted Dacron in up to 12 months of0bservation. 21 Our results are similar in that no difference was see~ between WD and knitted SVD. Double velour kni'cted Dacron, however, had superior tensile strength compared with SVD and WD. Thus improved healing is suggested in the double velour knitted prosthesis. The observation of improved graft-to-graft bonding in PTFE grafts requires further investigation. It is interesting to note a higher tensile strength in the PTFE graft-to-graft anastomosis than in the graftto-artery union. This may be the result of better tissue incorporation at both sides of the anastomosis in the graft-to-graft setting, whereas perhaps the artery, not being a foreign body, may be less adherent to the capsule. We conclude, first, that under the experimental conditions described, primary anastomotic bonding between prosthetic and recipient arteries occurs in the dog model and its strength is dependent on the graft type. In this regard, the types of PTFE and

Volume 5 Number 2 February 1987

double velour knitted D a c r o n tested p r o d u c e d an anastomosis with the highest tensile strength a m o n g the prostheses investigated. Second, under the experimental conditions described, primary anastomotic graft b o n d i n g is better with P T F E grafts tested than with the types o f double velour knitted Dacron, single velour knitted Dacron, and w o v e n D a c r o n used in o u r investigations. Third, clinical studies are warranted to d o c u m e n t that the incidence o f anastomotic aneurysm formation with P T F E grafts is lower, as suggested by observations with the types o f grafts described. REFERENCES

1. Curler EC, Dunphy JE. The use of silk in infected wounds. N Engl J Mcd 1941;224:101. 2. Donovan DJ~ Buckman CA. Aortoenteric fistula. Arch Surg 1967;95:810. 3. Olsen WR, DeWeese NS, Fry WJ. False aneurysm of abdominal aorta. Arch Surg 1966;92:123. 4. Stoney RJ, Albo RJ, Wylie EJ. False aneu~sms occurring after arterial grafting operations. Am J Surg 1965;110:153. 5. Starr DS, Weathcrford SC, Lawrie GM, Morris Jr GC. Suture material as a factor in the occurrence of anastomotic false aneurysms: An analysis of twenty-six cases. Arch Surg 1979;114:412-5. 6. Richardson JV, McDowell Jr HA. Anastomotic aneurysms following arterial grafting: A ten year experience. Ann Surg 1976;184:179-82. 7 Knox WG. Peripheral vascular anastomotic aneurysms: A fifteen year experience. Ann Surg 1976;183:120~3. 8. Christensen RD, Bematz PE. Anastomotic aneurysms of the femoral artery. Mayo Clin Proc 1972;47:313-7. 9. Edwards WS, Doughton Jr D, Quatrlcbaum R. Anastomosis

DISCUSSION Dr. Bruce M. Smith (Nashville, Tenn.). While you cannot properly deny that the clinical findings of pseudoaneurysms were different in your groups, the tensile strength, as I understand it, includes the measure of the cross-sectional area of the structures that you are disrupting on the tensiometer. If you corrected your data for crosssectional area, would the PTFE and Dacron groups become more similar in their tensile strength? Dr. Quifiones-Baldrich. All the grafts were 6 mm in diameter and, in reality, within the 9-month study period there was no dilatation of the grafts. We thus assumed that the circumference of the anastomosis was similar for all tested prostheses. This is really the best way to measure tensile strength as opposed to using cut strips by which one introduces a significant amount of variation between the tested strips. Dr. H o w a r d D. Greisler (Maywood, Ill.). There is nothing wrong, as I see it, with an absorbable suture line. However, this is true only if the materials being sutured

Primary anastomotic bonding in PTFE grafts? 317

between synthetic graft and artery: A study of tensile strength. Arch Surg 1963;86:477-9. 10. Moore WS, Hall AD. Late suture failure in the pathogenesis of anastomotic fa~,se aneurysms. Ann Surg 1970;172: 1064-8. 11. Moore WS, Hall AD, Allen RE. Tensile strength of a~erial prosthetic anastomosis. J Surg Res 1972;13:209-14. 12. Malone JM, Misiorowski RL, Moore WS, Chvapil M. Influence of prosthetic graft fabrication on graft healing, It. J Surg Res 1981;30:443-8. 13. Sauvage LR, Berger KE, Wood SJ~ Yates SG~ Smith JC, Mansfield PB. Interspedes healing of porous arterial prostheses observations, 1960 to 1974, Arch Surg 1974;109: 698-705. 14. Stoney RJ, Albo RI, Wylie EJ. False aneu~,sms occurring after arterial grafting operations. Am J Surg I965;110: 153-61. 15. Szilagyi DE, Smith RE, Elliott JP, et al. Anastomotic aneurysms after vascular reconstruction: Problems of incidence, etiology, and treatment. Surgery 1975;78:800-16. 16. Youkey JR, Clagett GP, Rich MM, Brigham RA, Orecchia PM, Salander JM. Femoral anastomotic false aneurysms: An eleven year experience analyzed with a case control study. Ann Surg 1984;199:703-9. 17. Dennis JW, Littooy FN, Greisler HP, B~er WH. Anastomotic pseudoaneurysms: A continuing late complication of vascular reconstructive procedures. Arch Surg 1986;121: 314-7. 18. Chavez CM. False aneurysms of the femoral artery: A challarge in management. Ann Surg I976;183:694-700. 19. Haimov H, Iiron S, Iacobson JH. The expanded polytetrafluoroethylene graft: Three years experience with 362 grafts. Arch Surg 1979;114:673-7. 20. Hohf RP. Tensile strength of arterial-prosthesis anastomosis during healing. Ann Surg 1962;156:805-i0. 21. Kottmeier CA, Wheat MW. Strength of anastomosis in aortic prosthetic graft. Am Surg 1965;31:128-34.

permit the establishment of adequate regenerative and fibroblasfic reactions to bridge the suture line with enough strength to withstand various hemodynamic stresses. In my own view, Dacron clearly permits a less extensive tissue reaction than does PTFE. I do not know why that is. It may be related to differences in complement activation by these biomaterials. There may be differences in the dectronegativity of the surfaces or differences in the inflammatory reactions to these materials. My question for the authors is, did they evaluate the differences in the histologic reactions to these materials? And, if they did, did they do so before or after the tensile strength measurements which, of course, might alter these histologic appearances? Dr. Qtfifiones-Baldrich. We also have the impression that PTFE has better incorporation and better adhesion to the recipient artery, and it has been our clinical impression that the incidence of pseudoaneu~sms with PTFE prostheses is extremely low. In regard to your question, I believe that, in order to

318

Qui~ones-Baldrich et al.

carry out tensile strength measurements that are reliable, one should not disturb the tissues in the area of the anastomosis. Therefore it is mutually exclusive to perform tensile strength measurement and to obtain tissue for histologic examination. Clearly, one could examine these tissues after the disruption, but I am not sure that this would allow a good indication of exactly what may be occurring histologically. Therefore we did not perform histologic studies in this experiment. Dr. Roger M. Hayashi (Los Gatos, Calif.). Actually, I have two questions: (1) What is the role of infection? Were cultures taken from any of these anastomoses either at the time of explantation or following disruption? Did infection influence anastomotic strength? (2) Are you considering the possibility of studying for fibroblast anastomotic ingrowths in both the Dcxon and the PTFE models that you used here? Dr. Quifiones-Baldrich. We did not perform any cultures in any of the animal experiments. It is clearly possible that in some of the disrupted grafts that we observed, spontaneous disruptions occurred at 21 days and later, and these may have been associated with infection. There was no gross evidence of this at the time of necropsy. It is interesting that the disruptions all occurred within the same group and that these animals were all implanted at random. The grafts were not implanted in each group at a particular interval. It would thus be unlikely that all three of the disruptions would have infection, but clearly that is a possibility. We are not planning to do fibroblastic studies on these anastomoses. However, we are very interested in pursuing our initial findings, which suggest that the tensile strength of these anastomoses increases with time in the PTFE graft as opposed to the Dacron grafts. Dr. Peter F. Noyes (Medford, Ore.). Have you looked at the bonding between normal arteries as compared with arteries in grafts? My mentor, Dr. Wylie, always believed that the bonding between two normal arteries was very strong as compared with grafts. Furthermore, did you have a chance to observe this phenomenon? Dr. Quifiones-Baldrich. We have not looked at the bonding between normal arteries. However, one thing that is apparent from these data is that the bonding that occurs in the PTFE graft between the graft-to-graft anastomosis as opposed to the graft-to-artery anastomosis is stronger, and that was a surprising finding. One would expect the graft-to-graft bonding to be weaker than a graft-to-artery

Journal of VASCULAR SURGERY

bonding. However, in the PTFE graft, that does not appear to be the case; I suspect it may be related to the presence of a foreign body, with tissue incorporation at both sides of the anastomosis, as opposed to situations in which an artery and a graft are bonded. Dr. Jonathan E. Hasson (Boston, Mass.). I believe I understood you to say that the failure mode was disruption by suture pull-through of the artery, either proximal or distal to the anastomosis, which implies that in a strain analysis you are pulling through at the weakest link of the repair. Your data would then suggest that in some way this bonding is affecting the mechanical strength of.the artery that was removed from the anastomosis, because if this is where it is tearing out, this has to be the weakest link. I therefore wonder if you could speculate on some mechanism as to how that would happen. It is a little puzzling to me how a graft on one side of the suture line will imp,,t some mechanical strength to the artery on the other side. In addition, I wonder whether you have tested any of these anastomoses by carefully removing the Prolene continuous suture to actually ascertain the anastomotic bonding as an isolated factor. Dr. Quifiones-Baldrich. Regarding the first question, I do not think that the strength of the anastomosis occurs at the interface between the artery and the graft. I believe that the strength of the anastomosis is imparted by the fibrous tissue capsule that covers both the artery at the site of the anastomosis and the graft. The fact that the capsule is very adherent to the graft and also to the artery is, in my opinion, what accounts for the tensile strength. I therefore do not think that primary bonding is at the exact interface of the artery and the graft. In regard to your second question, we did not remove the Prolene suture so as not to disturb the anastomosis and could therefore test it intact. This was the reason we did one of the anastomoses with an absorbable suture. When one tries to remove the suture, some of the fibrous caps~, at the anastomotic site will inevitably be disrupted. Dr. Hayashi. What is the tensile strength of Dexon over time in this implanted site model? Dr. Quifiones-Baldrich. We base our estimate on what the manufacturer suggests, which is that the halfqife of the tensile strength of Dexon is about 21 days. After approximately 90 days, the suture has completely disappeared from the anastomosis. That is why we chose 3 months as our earliest time of testing.