Reduction in deposition of indium 111—labeled platelets after autologous endothelial cell seeding of Dacron aortic bifurcation grafts in humans: A preliminary report

Reduction in deposition of indium 111-labeled platelets after autologous endothelial cell seeding of Dacron aortic bifurcation grafts in humans: A preliminary report Per ()rtenwall, M.D., Hans Wadenvik, M.D, Jack Kutti, M.D., Ph.D., and Bo Risberg, M.D., Ph.D., Giiteborg, Sweden Autologous endothelial seeding (AES) of vascular prostheses in dogs increases thrombusfree surface and improves prosthetic prostacyclin production, patency, and the ability to withstand hematogenous challenge with bacteria. No such information is available in human subjects. In the present study one limb of an aortic Dacron bifurcation prosthesis was seeded with autologous endothelial cells (ECs) harvested from the distal portion of the saphenous vein by enzymatic treatment. The deposition of indium 111-labeled platelets on the vascular prostheses was studied 1 and 4 months after operation. In seven of nine patients seeding resulted in decreased accumulation of radiolabeled platelets compared with sham-seeded control limbs (p < 0.04), when studied 1 month after surgery. A decrease in platelet accumulation occurred over the whole prosthesis between 1 and 4 months, and no significant difference was noted at 4 months between seeded and nonseeded graft limbs. Although the seeding density was very low (440 ECs/cm2), the observed difference in platelet accumulation for AES-treated graft limbs in the early postoperative course merits further investigation of this technique in human beings. (J VASCSURG 1987;6:17-25.) Present Dacron or microporous polytetrafluoroethylene (PTFE) vascular prostheses (VPs) do not spontaneously acquire a vascular endothelial cell (VEC) lining after implantation in humansJ '2 External gamma imaging studies of indium 111-labeled platelet deposition in patients suggests that lack of VEC linings makes VPs highly reactive with platelets and therefore liable to failure from thrombotic occlusion for long periods of time after implantation. s-s The method of autologous endothelial seeding (AES), which was developed to attack these problems, harvests VECs from peripheral veins and attaches these to the VP lining by embedding them in a fibrin clot formed on the VP at the time of surgery. From the Departments of Surgery (Drs. Ortenwall and Risberg) and Internal Medicine, (Drs. Wadenvikand Kurti), University of G6teborg. Supported by grants from the Medical Faculty, University of G-6teborg, The Swedish Societyof Medicine, The GOteborg MedicalSociety,and The SwedishNationalAssociationagainst Heart and Chest Diseases. Reprint requests: Per Ortenwall, M.D., Department of Surgery, University of G6teborg, (~stra Sjukhuset, CK, S-416 85 G6teborg, Sweden.

In dogs the AES procedure results in formation of complete luminal VEC linings of the VP within 4 to 6 weeks after VP implantation. 6"7 In canine models, AES sharply curtails the interaction of indium i l l - l a b e l e d platelets with VPs 8 and improves both patency rates and resistance to flow reductioninduced thrombosis. 9-H An unanswered question about the clinical use of AES is whether EC coverage of large VPs could be achieved with the relatively small amounts of autogenous vein available as VEC sources in human subjects. Although VEC-like linings have been observed by scanning electron microscopy in one AES-treated human VP so far,12 single morphologic results cannot assess the effect of AES on platelet-VP interaction in patients. To assess the effect of AES on platelet-VP interactions in humans, we studied the accumulation of indium 11 i-labeled platelets on seeded and nonseeded limbs of Dacron aortic bifurcation grafts 1 and 4 months after VP implantation. We report that significant reduction in labeled platelet accumulation occurs in the AES-treated segments of these VPs in the early postoperative phase. 17

Journal of VASCULAR SURGERY

18 Ortenwall et al.

Table I. Description of patient population studied GraJ~ Diameter type (ram)

Preop ABI seeded/ control limbs

PostopABI seeded/ control limbs

Patient

Age (yr)

Sex

Previous vascularsurgery

Indication for surgery

1 2

57 62

M M

-TEA (ICA)

C A

B B

16 18

0.50/0.50 --

0.80/0.75 --

3 4

61 61

M M

-T E A (ICA)

A L

B V

18 16

0.80/0.80 0.20/0.20

0.90/0.85 0.55/0.65

5 6 7 8

66 71 70 52

M M F M

---TEA (CIA)

A C L C

B B B B

18 16 14 16

0.80/0.80 0.40/0.90 0.20/0 0.50/0.55

0.90/0.90 0.80/0.90 0.40/0.40 0.95/0.90

9

74

M

--

A

B

18

0.60/1.00

0.80/1.00

Postop course Uneventful Bowel obstruction 3 wk after surgery Uneventful Bowel obstruction 3 rno after surgery Uneventful Uneventful Uneventful Serorna in groin; healed in 2 wk Temporary postop renal insufficiency

ICA = internal carotid artery; CIA = common iliac artery; TEA = thromboendarterectomy; C = disabling claudication; A = aneurysm; L = limb salvage; B = Bionit (USCI); V = Vascutek VP 1200k (Vascutek).

MATERIAL A N D METHODS Patients and control subjects. Nine patients (eight men and one woman) scheduled for abdominal aortic reconstruction in whom we planned to use a bifurcated VP have been studied. Their mean age was 63.7 ___ 7.1 years (range 52 to 74 years). All patients had preoperative evaluation of cardiac function, assessed by cardiac sonography (M-mode, two-dimensional sector scan, and pulsed Doppler), isotope cardiography (MUGA), and myocardial scintigraphy to assess left ventricular function. Electrocardiograms were obtained at rest and during exercise. The ankle-brachial pressure index (ABI) was calculated for each limb pre- and postoperatively. Five of the patients were operated on because of occlusive disease (two for limb salvage and three for disabling claudication) and four because of aneurysms. Three patients had had previous vascular surgery (one carotid and two iliac reconstructions). Five patients had evidence of previous myocardial infarctions and two had a history of angina pectoris. The left ventricular ejection fraction (LVEF) was impaired in two patients. Eight patients had a history of smoking and one had diabetes. Platelet survival studies were carried out in 13 healthy male volunteers (aged 20 to 45 years; mean 29 - 8 years) with autologous indium 111-labeled platelets. In five of them the interactions between platelets and "normal" vascular wall were also studied. All studies were reviewed and approved by the Ethical Committee of the University of Gtteborg. Informed consent was obtained from all patients and volunteers.

Surgical procedure. Patients were anesthetized with a mixture of pentothal and fentanyl (Leptanal, Janssen Pharmaceutica, Beerse, Belgium) and were mechanically ventilated with a mixture of oxygen and nitrous oxide. Postoperatively all patients also had an epidural block with an indwelling catheter for analgesia for 2 to 3 days. Intraoperative monitoring was performed by continous measurement of arterial blood pressure, electrocardiography, and central hemodynamics by the use ofa Swan-Ganz catheter. The antibiotic cefuroxime (Zinacef, Glaxo Operations UK Ltd., Greenford, U.K.) was given, 1.5 gm intravenously three times during the day of surgery, starting 3 hours before the procedure. Routine anticoagulation was dextran 40 (Rheomacrodex, Pharmacia Inc.) 500 ml perioperatively, 500 ml during the first 8 postoperative hours, followed by 500 ml daily for 4 days. All patients were operated on by the same two vascular surgeons, starting simultaneously by exposing the aorta and removing a part of the greater saphenous vein, just above the ankle. The cellharvesting procedure (see later) was performed in the operating room and took between 45 and 60 minutes. During this time the proximal anastomosis was performed with inlay technique, with a running 4-0 polypropylene suture. Noncrushing vascular clamps were then placed across the graft limbs, just distal to the bifurcation, and the proximal aortic clamp was released after the central portion of the graft was preclotted. After a randomized choice was made of the graft limb to be seeded and of the order in which to construct the distal anastomoses, the cell suspension was divided into two 2 ml aliquots. These aliquots were sequentially mixed with equal volumes

Volume 6 Number 1 July 1987

of autologous whole blood and were injected 5 minutes apart into the graft limb to be seeded, which was then clamped distally. The cells were allowed to settle in the graft limb during completion of the distal anastomosis. The distal anastomoses were preferably sutured end to end to the common iliac artery, but if this was not feasible, and end-to-side anastomosis to the external iliac or common femoral artery was performed. Just before completion of the anastomosis, both clamps were released and blood allowed to flush through the graft limb. Fogarty catheters were used only if flow was considered impaired. The nonseeded (control) graft limb was treated the same way, except that blood was mixed with culture medium, omitting the ECs. Eight Bionit grafts (USCI Surgical Products, C. R. Bard Inc., Billerica, Mass.) and one Vascutek VP 1200K graft (Vascutek Ltd., Ayr, Scotland) were implanted. A femoropopliteal reversed vein bypass was performed in one patient at the time of the aortic reconstruction (patient 4). For further description of patients and surgical procedures, see Table I. An estimation ofrunoffconditions, judged by postoperative ABI values, showed no difference between seeded and control graft limbs. Cell harvesting and culture procedure. Approximately 10 cm of distal calf saphenous vein was excised. The vein was caunulated at both ends with Tibbs' olive head arterial cannulas, flushed with chilled buffered saline solution (Hanks' BSS-CMF, Flow Labs, Richmansworth, England). It was then filled to slight distension with prewarmed collagenase solution (0.1%, type I, batch 45F-0676, Sigma Chemical Co., St. Louis, Mo.) in phosphate-buffered saline solution (PBS), clamped at both ends and incubated in sterile PBS at 37 ° C for 10 minutes. The vein contents were flushed into a sterile centrifuge tube with 10 ml of complete medium (medium 199 with Earle's salts, 25 mmol/L HEPES buffer, L-glutamine (Gibco, Paisley, Scotland), 100 ~g/ml porcine heparin (H-3125, Sigma Chemical Co.), 25 IU/ml penicillin with 25 ~g/ml streptomycin (Flow Labs), 150 ~g/ml endothelial cell growth factor (CR-ECGS, Collaborative Research, Waltham, Mass.) and 20% fetal calf serum (Gibco), at 4 ° C. All fluids were sterilized by passage through 0.22 ~m filters (Millex, MiUipore, Molsheim, France). After cell harvest the vein segment was cut open, and the luminal area of the vein used was measured. The effluent was flushed through the vein two more times, then centrifuged at 180 × g for 10 minutes. The resulting cell pellet was resuspended in 5 ml of complete medium. Four milliliters of this

Human endothelial seeding

19

Fig. 1. Computer-generated picture of nuclear scan of a patient with a 16 × 8 mm knitted Dacron prosthesis. All anastomoses were performed end to end. The seeded graft limb (12 cm in length) is on the right side ofthc picture, while the control limb (17 cm in length) is to the left. The number of counts is presented in a color-coded scale, where highest activity is white, followed by green, yellow, red, and blue. Note lower activity on seeded graft limb. Areas with high activity in uppermost part of picture represent liver and spleen, and in lower part the penile area can be visualized. suspension was used for seeding of the VP. The remaining 1 ml was stored at room temperature and taken to the cell culture laboratory after surgery was completed. Two hundred microliters was taken for cell counting and viability staining in a hemocytometer, with the trypan blue exclusion dye test. The other portion of the cell suspension was inoculated onto plastic tissue culture wells (Falcon, BectonDickinson Labware, Oxnard, Calif.), precoated with human fibronectin (10 i~g/cm 2 of dish area) (Collaborative Research Inc.). The culture medium was changed every third day, and the cultures held at 37 ° C in a humidified 5% carbon dioxide environment. m4 The cultures were studied with an inverted phase-contrast microscope, and the identity of the cells was checked with a peroxidase-antiperoxidasestaining test for factor VIII-related antigen, t5 Radioisotopic procedures. These procedures were performed 1 and 4 months after surgery. Platelet separation and labeling procedure. All subjects to have their platelets labeled were prohibited to ingest any aspirin or other nonsteroidal antiinflammatory drugs 10 days before the study. Seventeen milliliters of venous blood was obtained with

Jotunal of VASCULAR SURGERY

20 Ortenwall et al.

Table II. Platelet labeling and survival data Labeling efficiency

Platelet recovery

Platelet mean life span

29.4 ± 7.4

81 ± 7.5

53 ± 8.8

8.43 ± 0.81

6 3 . 7 -+ 7.1

85 ± 4.8

60 _+ 9.7

7.41 ± 1.16

64.5 _+ 5.5

87 ± 5.0

64 _+ 10.4

8.34 ± 0.99

Age (yr) Control (N = Patients (N = Patients (N =

subjects 13) 1 m o postop 9) 4 m o postop 6)

(%)

(%)

(days)

NOTE: Data expressed as mean ± standard deviation.

a 19-gauge butterfly needle and drawn into a disposable syringe, containing 3 ml of acid citrate dextrose (National Institutes of Health) formula A [ACD-A]) as anticoagulant. The anticoagulated blood was transferred into a 30 ml screwcap polycarbonate tube (Nalge Co., Rochester, N.Y.). Platelet-rich plasma (PRP) was obtained by centrifugation of the collected blood at 260 × g for 15 minutes. PRP was then transferred into another 30 mi polycarbonate tube, and the pH was adjusted from 6.3 to 6.5, by adding 1 ml of ACD-A for each 10 ml of PRP. The platelets were centrifuged at 800 × g for 10 minutes. The supernatant plateletpoor plasma (PPP) was removed for later use. The platelet pellet was washed with 5 ml modified Tyrode's solution (MTS), at 37 ° C. ~6The platelets were resuspended in 1 ml MTS and the required amount of indium 111-oxine (Amersham Int. Inc., Amersham, U.K.) was added drop by drop to the suspension. After being incubated with the isotope for 2 minutes at 37 ° C, 2 rnl of PPP was added, and the mixture was centrifuged at 800 × g for 10 minutes. The supernatant was removed and the platelet pellet was finally resuspended in 5 ml of PPP. To calculate labeling efficiency, the platelet-bound radioactivity and the radioactivity in the supernatant were measured in a dose calibrator (Radioisotope Calibrator CRC-5RB, Capintec, Inc., Ramsey, N.J.). Two hundred microcuries was injected into the patients. All procedures were carried out aseptically at room temperature, if not otherwise stated. Platelet survival studies. One hundred and 200 Ixl of the labeled platelet suspension were used as standards. The total blood volume was calculated from body weight and length according to Nadler, Hidalgo, and Block. ~7 Ten milliliters of blood was collected into EDTA 30 minutes after infusion of the labeled platelets and subsequently at 24-hour intervals for 7 to 10 days. All blood samples were drawn after the patients had rested for 15 minutes in the supine position. The platelet count and the packed

red cell volume were determined in association with each sampling. Two milliliters of whole blood and 2 ml of PPP (obtained by centrifugation at 2000 × g for 20 minutes) were counted in a crystal scintillation counter, with a spectrometer setting of 200 to 260 keV (Selektronik AS, Copenhagen, Denmark). Correction was made for plasma-bound radioactivity, and, by computer analysis, platelet mean life span (PMLS) was calculated by fitting a multiple hit gamma function to the experimental data. Platelet recovery was obtained by extrapolation of the zerotime platelet-bound radioactivity. The software for these computer analyses was obtained from the office of the International Committee for Standardisation in Haematology.~S Graft-platelet interaction. In association with each platelet survival study anterior scanning of the abdomen was performed at 24, 48, 72, and 96 hours after the labeled platelets were injected, by means of a scintillation camera (MaxiCamera II, General Electric) interfaced with a computer (PDP 11-34, Digital Equipment, Marlboro, Mass.). Both photo peaks emitted by indium 111 (173 and 247 keV) were collected through a medium energy (maximum 370 keV) parallel hole collimator, with a 15% energy window setting, during six consecutive 5 minute periods each time. For visual assessment images were recorded on Polaroid film. The images were also stored on computer disk for data processing, with a 64 × 64 matrix system, dividing the image into 4096 pixels (software: Gamma 11). These computergenerated gamma images were assessed by a nuclear medicine specialist, by means of a color-coded scale for platelet deposition intensity (Fig. 1). To correct the image for contributing radiation from the pelvic and vertebral bone marrow the background activity from this area was subtracted, s,~9By use of the matrix system any region of interest (ROI) in the image could be analyzed with respect to the accumulated number of counts obtained from that area during the 30-minute imaging period. Dividing by the actual

Volumc 6 Number 1 July 1987

number of pixels necessary to cover the ROI, an average count per square unit of the R O I was obtained. Dividing this average count by the number of counts obtained from 0.1 ml of whole blood, collected at the same time and counted for 200 seconds in a crystal scintillation counter, gave a graft blood ratio corrected for both isotope decay and platelet senescence with the formula:

Human endothelial seeding 21

TI _ "T

Graft/blood ratio = 100 × counts in ROI - back~;round counts (number of pixels in ROI) × (whole blood counts i This ratio, choscn as an index of graft-platelet interaction, was designated the thrombogenicity index (TI). The ROIs were kept identical between different imaging times. Whenever possible, scans were viewed in a blinded fashion regarding the seeded or unseeded status of graft limbs. However, complcte blinding was not possible since detailed knowledge of the operative anatomy was needed to discriminate graft limbs f~om excluded vcssd segments, which also accumulated platdets. To correct for variations in distance between the scintillation camera and the graft limbs, all calculations o f TI were performed on equal areas of limbs to be compared. Follow-up. In association with platelet survival studies and VP scanning at 1 and 4 months after operation, the patients were also assessed by clinical and femoral Doppler examination. If these resuks indicated any postoperative complication, additional angiograms were performed. Statistics. Statistical analysis was performed with paired Student's t test to compare intraindividual differences in TI at the same imaging time and unpaired t test to compare intraindividual changes of TI between different imaging times. The coefficient of variation was calculated for the assessment ofintraobserver reliability. All values are given as mean ± one standard deviation. RESULTS Cell harvesting and cell culture An average of 27 ± 14 × 10 a cells were harvested from donor veins with a mean area of 5.2 ± 0.6 cm 2. The mean VP area seeded was 41 - 13 cm 2, resulting in a mean vein-to-VP area ratio of 0.13 ± 0.05. The mean number of viable cells seeded was 17 ± 8 × 103, resulting in 440 ± 186 viable cells seeded per square centimeter of VP. In four cases we were able to establish cell cultures that morphologically resembled ECs and stained positive for factor VIII-related antigen. Two of these cultures could subsequently be "passaged." The ad-

DAYS •

•

IE

•

1 2 3 4 Fig. 2. Platelet accumulation (77) over left (dotted line) and right (solid line) common and external iliac arteries in control subjects (N = 5). No significant difference was seen between sides.

mixture of other cell types in these cultures was judged to be less than 5% from visual assessment and grid counting. In all cases where we could establish an EC culture, differences in TI in favor of the seeded limb was noted. Radioisotopic studies

Platelet data. Platelet-bound radioactivity compared with that of plasma always exceeded 95%. The results for labeling efficiency, platelet recovery rate, and PMLS are given in Table II. A slight reduction in PMLS was noted 1 month after operation. This difference was not statistically significant compared with healthy volunteers. PMLS at 4 months after surgery was normalized. Graft-platelet interactions. ROIs studied were (1) the whole VP, (2) the bifurcation region of the VP, (3) the region of the abdominal aortic remnant, and (4) equal portions of the two graft limbs. These ROIs were kept identical between different imaging times, and TIs were calculated for each R O I at each imaging time. To assess intraobserver reproducibility of the method (i.e., the standard error of single determination), duplicate estimates o f the TI for all ROIs were performed with a time interval of at least 3 weeks. The coefficient of variation was 9%. Scans of control subjects (N = 5). Five healthy

Journal of VASCULAR SURGERY

22 Ortenwall et al.

TI 15. VP

B

A

T

! I

10-

L .

DAYS .

.

1234

.

.

.

.

.

.

1234

Fig. 3. Platelet accumulation (T/) over bifurcation area (B), whole vascular prostheses (VP), and abdominal aortic remnant (A). Solid line = 1 month after surgery (N = 9), thin dotted line = 4 months after surgery (N = 6); thick dotted line = control subjects (N = 5).

male volunteers (mean age 32.5 + 8.5 years) were scanned in an identical fashion, as described earlier. TIs of all ROIs showed a horizontal or slightly dedining curve during the four consecutive days of imaging. No significant difference could be detected for the ROI of the left or right common and external iliac artery, respectively (Fig. 2). Scans 1 month after surgery (N = 9). Platelets accumulated over the whole VP during the consecutive four days. For the bifurcation area and the region of the aortic remnant, maximum radioactivity was reached on the third day (Fig. 3). When the seeded and the control graft limbs were visually compared, less radioactivity accumulated on the seeded graft limb in seven of nine patients. Comparison of computer-generated TI with the paired t test resuited in significant differences at all imaging times (0.026 < p < 0.040) (Fig. 4). Scans 4 months after surgery ( N = 6). The accumulation ofplatelets on the region of the aortic remnant was unchanged. However, the TI for the whole VP had declined at all imaging times, and the difference was significant (p < 0.05) at 96 hours corn-

pared with the TI at 1 month after operation (Fig. 3). The pattern of accumulation had also changed, now reaching a maximum at 72 hours. The TI for the bifurcation region had also decreased but without reaching significance. A reduction of TI for the graft limbs had occurred (seeded: p < 0.05; control: p < 0.01). This reduction in TI was greater for the control limb than the seeded limb and resulted in loss of the statistically significant difference between the graft limbs, which was noted ! month after operation. FoUow-up All patients have been followed up for at least 3 months. One patient had reversible renal failure in the immediate postoperative period, which necessitated temporary hemofiltration. Two patients were operated on for bowel obstruction because of adhesions at 3 weeks and 3 months, respectively. The latter patient was admitted to the hospital 1 week after the onset of abdominal pains and vomiting. Laparotomy revealed small bowel strangulation and resection was performed. The postoperative course

Volume 6 Number 1 July 1987

Human endothelial seedin~¢ 23

TI

I

m

_

lmont~

4months

I /

m

ml

I

U

m

II

days

Fig. 4. Platelet accumulation (77) on seeded (dotted line) and control limbs (solid line) at 1 month (N = 9) and 4 months (N = 6) after operation. Differences in TI at 1 month between seeded and control limbs were statistically significant at day 1, 2, 3, and 4 at p < 0.028, p < 0.026, p < 0.040, and p < 0.031, respectively. At 4 months no statistically significant differences were noted. was complicated by severe adult respiratory distress syndrome, with pneumothorax and septicemia; the patient died after 3 weeks in the intensive care unit. One patient had a seroma in the groin of the seeded graft limb that healed spontaneously after 2 weeks. No infections have been recorded, nor have we experienced any adverse reaction's from the plateletlabeling procedure or the platelet scans. All graft limbs are patent, and only one repeat angiogram has been performed because of suspicion of a false aneurysm developing at a distal anastomosis. DISCUSSION Our data suggest that a significant reduction in measurable platelet-VP interaction can occur 1 month after operation in aortobifemoral knitted Dacron grafts treated with AES. Since the patients' control graft limbs were shamseeded with medium 199, this difference in platelet

reactivity was unlikely to arise solely from the effects of mixing culture medium with blood during the preclotting procedure. Scans performed 4 months after operation revealed an unchanged platelet accumulation on the abdominal aortic renmant, whereas the accumulation on the whole VP had diminished, with disappearance of the differences between seeded and control graft limbs. This decrease in platelet accumulation on Dacron grafts with time has previously been designated as "maturation" of the VP. 2° By the use of gamma camera scintigraphy, deposition of radiolabeled platelets in an R O I means than platelets have accumulated excessively compared with a reference point (whole blood). The decrease in the count of radiolabeled platelets in whole blood over time is a resuk of physical decay of the isotope and removal of radiolabeled platelets from the circulation. If no deposition of platelets occurs on the

24

Ortenwall et al.

VP, the recorded activity over the graft area will strictly follow the radioactivity in whole blood, and the ratio between them will be constant over time. Hence, the divergence of activity in the graft area and the radioactivity in whole blood is an estimate of platelet deposition. Like other investigators5 we have expressed the platelet deposition as a graft~blood ratio. Because of interindividual differences (e.g., in PMLS and attenuation of radiation), the obtained values of the graft/blood ratio are spechic for the patient, and the ratios should not be compared between different persons. Variation in platelet deposition because of differences in flow rate is unlikely in the present study since postoperative ABIs for seeded and unseeded limbs were similar, indicating equal runoff conditions on both sides. No patency differences between seeded and unseeded graft limbs were detected at the time of this report, but patency rates in high-flow, large-diameter VPs, such as aortofemoral bifurcation grafts, are unlikely to show strong effects from endothelial coverage in any case. The principal advantage of the use of aortofemoral prostheses for this study was the ability of each patient to serve as his or her own control. The present data do not allow us to state whether the observed differences in TI were specifically due to endothelial coverage of the seeded graft limbs. For obvious reasons the demonstration of EC coverage in a clinical study has to be indirect by scintigraphic imaging. The feasibility o f AES in human beings was demonstrated in a case report by Herring et al.,~2 in which ECs were seen in a biopsy specimen from a seeded graft. Our findings of reduced platelet-VP interactions in AES-treated limbs were consistent with the hypothesis that AES in humans can enhance the rate and extent of EC coverage of VPs in a manner similar to that observed in dogs. Estimates of the density of ECs seeded suggest that this was accomplished with EC seeding densities much lower than those usually required for EC coverage in canine studies of AES& '2' In particular, our failure to identify any difference in platelet accumulation in two patients in whom ECs failed to grow in culture suggests that the efficiency of our present methods of cell harvest may have to be improved to make AES a clinically reliable procedure. It is possible that such improvements could lead to more consistent and durable differences in platelet-VP interactions than those observed in the present study. It can only be hypothesized whether the low seeding density observed in

Journal of VASCULAR SURGERY

this study resulted in attachment of sparsely scattered ECs to the graft limb. The ability of these cells to synthesize prostacyclin, for example, might explain the observed difference in thrombogenicity, observed 1 month after operation. The diminished platelet accumulation observed in Dacron VPs with time affected the thrombogenicity of the VP more than the few ECs, which might explain the loss ofsignhicance of T! differences between seeded and unseeded graft limbs, observed between scans performed 1 and 4 months after surgery. This general reduction in thrombogenicity with time is now an unexplained phenomenon. It is not known whether a similar process occurs in VPs placed in low flow areas, such as the femoropopliteal region or in grafts of other fabrics (e.g., expanded PTFE). There was a slight but not significant reduction in PMLS in the patients who were operated on compared with that in healthy male volunteers. However, these data are in accordance with those obtained by other groups, 2° and the lack of significance is probably due to the limited number of subjects studied. The trypan blue exclusion dye test is a suboptimal measure of viability. A calculation of plating efficiency would have resulted in a more accurate estimation of EC viability. However, the EC suspensions had to be transported after operation to our cell culture laboratory at another hospital. This time delay (mean 3 hours) might have affected cell viability and contributed to the low success rate in establishing cell cultures in vitro. EC attachment on Dacron VPs has been estimated to be 80%. 23 At the seeding density reported in the present study, this would result in 350 cells attached per square centimeter of seeded graft limb. Assuming a normal EC density of 130 × 10 s cells/cm2, this would mean that the maximum area of the VP covered with cells at restoration of blood flow is just 0.27%. A considerable loss of initially attached ECs may also occur during the first 24 hours? 42~ If the cell loss is assumed to be 50%, and all ECs are capable of 60 population doublings, this would at actual seeding density never result in EC coverage exceeding 10% of the VP seeded. However, the actual number of ECs required to seed a graft in human subjects is unknown. Furthermore, it is uncertain whether 100% EC coverage of a graft is needed to reduce graft occlusion in low flow situations. The present results support the concept that AES may be feasible in human beings even at very low seeding densities, resulting in a reduced VP-platelet interaction in the early postoperative phase. However, this report is

Vokune 6 Ntwnber 1 July 1987

based on a limited number of patients. The durability of the benefit from AES has to be documented in further studies. In addition, improvements in EC harvesting efficiency and EC-VP attachment are needed to establish whether this procedure is worthwhile for use in distal reconstructions in low flow situations when autogenous vein is not available. We are greatly indebted to Dr. John B. Sharetkin, Uniformed Services University o f the Health Services, Bethesda, Maryland, for most valuable and stimulating discussions. REFERENCES I. Berger K, Sauvage LB, Rao AM, Wood SJ. Healing of arterial prostheses in man: its incompleteness. Ann Surg 1972;175: 118-27. 2. Sanvage LR, Berger K, Beilin LB, Smith JC, Wood SJ, Mansfield PB. Presence of endothelium in an axiUary-femoral graft of knitted Dacron with an external velour surface. Ann Surg 1975; 182:749-53. 3. Harker LA, Slichter SJ, Sauvage LR. Platelet consumption by arterial prostheses: the effects of endothelialization and pharmacologic inhibition of platelet function. Ann Surg 1977;186:594-601. 4. Goldman M, Norcott HC, Hawker KJ, Drole Z, McCollum CN. Platelet accumulation on mature Dacron grafts in man. Br J Surg 1982;69(Suppl):S38-40. 5. Stratton JR, Thiele BL, Ritchie JL. Platelet deposition on Dacron aortic bifurcation grafts in man: quantitation with indium- 111 platelet imaging. Circulation 1982;66:1287-93. 6. Herring M, Dilley R, Jersild RA Jr, Boxer L, Gardner A, Glover I. Seeding arterial prostheses with vascular endothelium. The nature of the lining. Ann Surg 1979;190:84-90. 7. Graham LM, Burkel WE, Ford JW, Vinter DW, Kahn RH, Stanley JC. Immediate seeding of enzymatically derived endothelium in Dacron vascular grafts. Early experimental results with autologous canine cells. Arch Surg 1980;115: 1289-94. 8. Sharefldn JB, Larker C, Smith M, Cruess D, Clagett GP, Rich NM. Early normalization of platelet survival by endothelial seeding of Dacron arterial prostheses in dogs. Surgery 1982;92:385-93. 9. Schmidt SP, Hunter TJ, Sharp WV, Malindzak GS, Evancho MM. Endothelial-seeded four-millimeter Dacron vascular grafts: effects on blood flow manipulation through the grafts. J VASCSURG 1984;1:434-41. 10. Allen BT, Long JA, Clark RE, Sicard GA, Hopkins KT, Welch MJ. Influence of endothelial cell seeding on platelet deposition and patency in small-diameter Dacron arterial grafts. J VASe SURG 1984;1:224-33.

Human endothdiat seeding 25

11. Stanley JC, Burkel WE, Ford JW, et al. Enhanced patencff of small-diameter, externally supported Dacron iliofemoral grafts seeded with endothelial cells. Surgery, 1982;92:9941005. 12. Herring M, Banghrnan S, Glover J. Endothelium develops on seeded human arterial prosthesis: a brief clinical note. J VASe SURG 1985;2:727-30. 13. Watldns MT, Sharefldn JB, Zajtchuk R, et al. Adult human saphenous vein endothelial cells: assessment of their reproductive capacity for use in endothelial seeding of prostheses. J Surg Res 1984;36:588-96. 14. larrell B, Lcvine E, Shapiro S, et al. Human adult endothelial cell growth in culture. J VASCSURG 1984;1:757-64. 15. Baughman S, Herring M, Glover J, Dilley R. The peroxidase antiperoxidase staining of factor VIII-related antigen on cultured endothelial cells. J Biomed Mater Res 1984;18:561-6. 16. Thakur ML, Walsh L, Malech HL, Gottschalk A. Indium Ill-labeled human platelets: improved merh~xl, efficacy and evaluation. J Nucl Med 1981;22:381-5. 17. Nadler SB, Hidalgo JV, Bloch T. Prediction of blood volume in normal human adults. Surgery 1962;51:224-32. 18. The panel on diagnostic application of radioisotopes in Haematology, ICSH. Recommended methods fbr radioisotope platelet survival studies. Blood 1977; 16:1137-44. 19. Heyns ADuP, Lotter MG, Badenhorst PN, et al. Kinetics, distribution and sites of destruction of indium 111 labelled human platelets. Br J Haematol 1980;44:269-80. 20. McCollum CN, Kester RC, Rajah SM, Learovd. P, Pepper M. Arterial graft maturation: the duration of thrombotic activity in Dacron aortobifemoral grafts measured by platelet and fibrinogen kinetics. Br J Surg 1981z68:61-4. 21. Schmidt SP, Hunter TJ, Hirko M, et al. Small-diameter vascular prostheses: two designs of PTFE and endothelial cellseeded and nonseeded Dacron. J VASCSURG 1985;2:292-7. 22. Herring M, Dilley R, Cutlison T, Gardner A, Glover 1. Seeding endothelium on canine arterial prostheses--the size of the inoculum. J Surg Kes 1980;28:35-8. 23. Stanley JC, Burkel WE, Graham LM, Lindblad B. Endothelial cell seeding on synthetic vascular prostheses. Acta Chir Scand 1985;Suppl 529:17-28. 24. Sharefldn JB, Latker C, Smith M, Rich NM. Endothelial cell labeling with 111-indium-oxide as a marker of cell attachment to bioprosthetic surfaces. J Biomed Mater Res 1983;17: 345-57. 25. Kosenman JE, Kempczinski RF, Pearce WH, Silberstein EB. Kinetics of endothelial cell seeding. J V a ~ SUR(; 1985; 2:778-84. 26. Sentissi JM, Ramberg K, O'Donnell TF, Connolly RJ, Callow AD. The effect of flow on vascular endothelial cells grown in tissue culture on polytetrafluoroethylene grafts. Surgery. 1986;99: 337-43.