Substituting proxy ratings for patient ratings in cancer clinical trials: an analysis based on a Southwest Oncology Group trial in patients with brain metastases.

In studies of the effect of cancer treatment in the advanced disease setting, researchers have attempted to avoid missing data for quality of life (QOL) assessments by either substituting proxy for patient assessments from the outset or by interspersing proxy measures when patients are unable to respond. Although poor agreement between patient and proxy assessments has been amply demonstrated in the literature, interest in using proxy measures persists. Completion of the Spitzer QL-Index by a small sample of patients with brain metastases and family member proxies provided data for evaluating the ability to substitute proxy for patient QOL assessments. These data cannot address treatment efficacy due to the modest sample size. Rather, the analyses serve to alert researchers to the important distinction (in a clinical trial setting) between agreement and the use of the proxy as a surrogate. We present several methods for evaluating the accuracy of proxy measures and for identifying other sources of error and bias that may vary with time or with treatment arm. Lin's concordance correlation coefficient suggests that proxies are generally a poor substitute for capturing a patient's perspective of his/her QOL. A longitudinal analysis suggests that the use of proxy rather than patient responses could lead to different conclusions concerning radiation therapy's effect on QOL.

Evaluation of expanded polytetrafluoroethylene arteriovenous access grafts onto which microvessel-derived cells were transplanted to "improve" graft performance: preliminary results.

The purpose of this study was to implement and evaluate a clinical protocol for following longitudinally the luminal responses of microvessel cell seeded expanded polytetrafluoroethylene (ePTFE) vascular grafts implanted for hemodialysis access. Half of the patients enrolled in the study were randomized to receive grafts that were "seeded" with transplanted microvessel cells derived from autologous subcutaneous fat; the other half of the patients received nonseeded grafts. The patients agreed to scheduled biopsies of their grafts at three postoperative times. All biopsy samples were evaluated by routine histologic and electron microscopy techniques. Three men and six women were enrolled in the study. All operative procedures were tolerated well. However, only two of the nine patients agreed to 1-year postimplantation biopsies; one of these patients had been randomized to receive a "nonseeded" ePTFE graft and one randomized to receive a "seeded" graft. The "seeded" graft at 3 months showed endothelial cells on the luminal surface as well as some intimal thickening. By 20 months, the same "seeded" graft showed significant concentric intimal thickening and by 24 months, this "seeded" graft thrombosed. The "nonseeded" graft at 16 months had irregular areas of intimal thickening which were quite patchy in nature. The flow contacting surface of the "nonseeded" graft remained thin. The intima of the "seeded" graft was twice as thick as that of the "nonseeded" graft. The methodologies implemented in the study design were appropriate. Biopsy samples were obtained without complication and were easily processed for analysis. Patient compliance with the biopsy protocol was problematic however. The study was terminated because of the development of significant concentric intimal hyperplasia in a "seeded" graft.

Controlled delivery of vascular endothelial growth factor promotes neovascularization and maintains limb function in a rabbit model of ischemia.

PURPOSE: Vascular endothelial growth factor (VEGF) modulates new blood vessel development and growth and has been suggested as a potential therapeutic agent that could alleviate debilitating claudication in patients. The objective of this study was to determine whether controlled, local delivery of a low dose of VEGF from an osmotic pump could promote neovascularization, limb perfusion, and functional improvements in the hind limbs of rabbits rendered partially ischemic by surgery. The effects of VEGF were compared with those of the vasodilator nitroglycerin (NTG) and to saline administered similarly. METHODS: Thirty rabbits were randomly assigned to either VEGF (n = 10), NTG (n = 10), or saline (n = 10) treatment groups. Partial ischemia was induced in each left hind limb by surgical ligation of the common and superficial femoral arteries, leaving the internal iliac artery intact. The right limb of each animal served as a nonischemic control. Immediately after vessel ligations, a 28-day osmotic pump was implanted to deliver VEGF (0.22 microg/kg/day), NTG (17.8 microg/kg/day), or saline solution into the common iliac artery just proximal to the ligation site. Comparative vascularity between ischemic and nonischemic limbs within treatment groups and between groups was evaluated by (1) capillary counts from representative fields of hematoxylin and eosin stained muscle tissue taken from hind limbs at day 40; (2) digitized arteriograms of ischemic legs at day 40, which were used to quantify the complexity of vascular branching (fractal dimension index) and the total extent of vascularization (vascular density index); (3) measuring capillary refill times in ischemic limbs; and (4) observations of functional and trophic changes in ischemic limbs. Statistical differences between treatment groups were evaluated by one-way ANOVA. RESULTS: Complexity of vascular branching and vascular density were significantly greater (p < 0.001) in VEGF-treated ischemic limbs compared with NTG- and saline-treated ischemic limbs. By postoperation day 14, all VEGF-treated ischemic limbs had restored capillary refill (p < 0.001), new hair growth, and greatly improved limb function and appearance. Saline-treated limbs exhibited ischemic changes, with poor capillary refill and negligible limb function. Capillary refill in NTG-treated ischemic limbs did not differ significantly from saline-treated limbs. Ischemic VEGF-treated limbs had significantly more capillaries compared with both ischemic and nonischemic limbs in saline-treated animals (p < 0.05). Ischemic NTG-treated limbs also had significantly more capillaries compared with ischemic limbs in saline-treated animals (p < 0.05). Because of high variability, however, capillary counts in VEGF-treated ischemic limbs did not differ significantly from those of contralateral nonischemic limbs, or from capillary counts in either ischemic or nonischemic limbs of NTG-treated rabbits. CONCLUSIONS: Controlled release of microgram quantities of VEGF significantly enhanced neovascularization and vascular perfusion in ischemic limbs compared with controls in this rabbit model of partial ischemia. In addition, VEGF-treated ischemic limbs demonstrated near-normal function and appearance, whereas NTG- and saline-treated ischemic controls remained noticeably impaired. This novel approach of VEGF delivery may prove clinically useful either alone or combined with revascularization procedures.

In vitro evaluation of electrostatic endothelial cell transplantation onto 4 mm interior diameter expanded polytetrafluoroethylene grafts.

PURPOSE: To perform an in vitro evaluation of electrostatic endothelial cell transplantation of human umbilical vein endothelial cells (HUVEC) onto segments of 4 mm internal diameter expanded polytetrafluoroethylene (ePTFE) vascular prostheses. METHODS: This evaluation consisted of exposing vascular graft segments that had been subjected to either electrostatic or gravitation transplantation with HUVEC to a physiologic shear stress (15 dynes/cm2) under steady flow conditions within a flow loop system. Biochemical assays were performed on freshly transplanted grafts by means of radioimmunoassay for prostacyclin and thromboxane A2. RESULTS: There was a 30% loss of HUVEC after 30 minutes of shear stress exposure from the grafts subjected to gravitational transplantation with no additional significant (alpha = 0.05) loss after 120 minutes. Grafts subjected to electrostatic transplantation had no significant (alpha = 0.05) loss of HUVEC during exposure to physiologic shear stress. Furthermore, after 120 minutes of shear-stress exposure, the grafts subjected to electrostatic transplantation (78,420 +/- 6274 HUVEC/cm2) retained 2.3 times more HUVEC than the counterparts subjected to gravitational transplantation (34,427 +/- 4637 HUVEC/cm2). The biochemical assay results indicated no significant (alpha = 0.05) production of prostacyclin or thromboxane A2 regardless of the method of cell transplantation. CONCLUSIONS: (1) The electrostatic transplantation technique was superior to the gravitational transplantation technique in terms of cellular retention when the ePTFE grafts were exposed to physiologic shear stress. (2) Production of prostacyclin and thromboxane A2 did not differ between transplanted HUVEC subjected to gravitational or electrostatic procedures.

Quantitation of angiogenesis in the chick chorioallantoic membrane model using fractal analysis.

Objective quantitation of new vascular growth in angiogenesis assays has been difficult to achieve. The aim of this study was to test the potential applicability of fractal geometry to objectively quantify angiogenesis in the chick chorioallantoic membrane (CAM). To accomplish this, vascular growth in normal (NC) and tumor grafted (TGC) CAM was compared. Samples of human tumors of various origins acquired from biopsy specimens were grafted to CAMs at Days 6-12 of incubation. Images of 6- to 16-day NC and 10- to 17-day TGC were acquired and the box-counting method was used to estimate fractal dimension (Df). Vascular density (rho v) was derived from box counts. For NC, Df and rho v increased steadily from Days 6 to 14, peaked, and began to decrease by Day 15. Results from TGC showed that Df and rho v surrounding tumor grafts varied from those of NC. Complexity and density indices were calculated as amount of change in Df and rho v between TGC and NC of the same age, respectively. Based on these indices, differences in the angiogenic response elicited by various tumors were readily detected. Our results suggest that differences in Df and rho v, in both the temporal and spatial senses, are accurate and repeatable measures of vascular growth and may be useful for objective quantitation of the angiogenic response.

Pharmacologic modulation of intimal hyperplasia in canine vein interposition grafts.

PURPOSE: The purpose of this study was to evaluate the efficacy of three drugs (cilazapril, cyclosporine, and aspirin) in modulating the progression of intimal hyperplasia during short postoperative times in short-segment, autogenous vein bypass grafts in a canine model. The relative effects of the drugs on the progression of intimal hyperplasia were compared with the Gilman parameter, a measure used extensively as a wound healing descriptor. To our knowledge this is the first use of the Gilman parameter in assessing vascular disease. METHODS: Seventy-two conditioned mongrel dogs were randomly and equally divided according to a three-factor analysis of variance. The factors included (1) drug treatments (cilazapril [10 mg/kg/day], cyclosporine [4 mg/kg/day], aspirin [325 mg/day], and control [nonmedicated]), (2) implantation sites (femoral and carotid arteries), and (3) postoperative times of graft harvest (1, 3, and 6 weeks). Each dog had 2 cm segments of autogenous jugular vein interpositioned bilaterally into each of the paired carotid and femoral arteries. Quantitative data on luminal narrowing over time from intimal hyperplasia were compared from calculated Gilman parameters after image analysis of retrieved, histologically processed graft sections. RESULTS: The observed variability in the data was attributed to drug treatments and time. At 1 week after operation the mean Gilman parameters did not differ significantly among the treatment groups in either midgraft or distal graft segments. At 3 weeks the mean Gilman parameters of midgraft and distal graft sections of cyclosporine-treated dogs differed significantly (p < 0.05) from those of the control group and the cilazapril and aspirin-treated groups, which did not differ from each other. At 6 weeks after operation, mean Gilman parameters from aspirin- and cyclosporine-treated dogs differed statistically from control and cilazapril-medicated dogs and from each other (p < 0.001). CONCLUSIONS: These data support the efficacy of aspirin and cyclosporine in reducing intimal hyperplasia in short-segment arterialized vein grafts during short postoperative periods. Additional studies are required to ascertain whether the beneficial effects of aspirin and cyclosporine persist long-term.

Bovine myocardial epithelial inclusions.

Light microscopic, histochemical, immunohistochemical, and ultrastructural methods were used to examine myocardial epithelial masses in the hearts of ten cattle. The tissues consisted of paraffin-embedded or formalin-fixed samples from eight hearts that were being inspected in slaughter houses and from two hearts from calves that died of septicemia. The ages of the cattle ranged from 4 days to 12 years; the breeds were unspecified for all but one Hereford female and the two Holstein calves; and there were three males, four females, and three steers. The masses in these cases were compared with similar appearing lesions found in other animal species. The lesions in the bovine hearts were single to multiple, well circumscribed, found in the left ventricle wall, and composed of squamous to cuboidal epithelial cells that formed tubular, ductular, and acinar structures with lumens that were void or filled with amorphous protein globules. Electron microscopic examination revealed epithelial cells that had sparse apical microvilli, tight apical intercellular junctions, perinuclear bundles of filaments, and rare cilia. Almost half of the bovine epithelial masses (4/9) had occasional diastase-resistant periodic acid-Schiff-positive granules in their cytoplasm, and few had hyaluronidase-resistant alcian blue-positive granules (2/9) or colloidal iron-positive granules (1/9). All myocardial masses had abundant collagen surrounding the tubular and acinar structures, and 2/9 had elastin fibers as well. None of the myocardial masses had Churukian-Schenk or Fontana Masson's silver staining granules in epithelial cells. Immunohistochemically, all bovine myocardial tumors stained positively for cytokeratin (8/8), and occasional masses stained positively for vimentin (3/8) or carcinoembryonic antigen (3/8). None of the masses stained positively for desmin. The myocardial epithelial tumors most likely represent endodermal rests of tissue misplaced during organogenesis.

The splenic snood: an improved approach for the management of the wandering spleen.

A wandering spleen is in constant danger of torsion and infarction. Splenectomy, the traditional treatment, leaves children in danger of postsplenectomy sepsis. Three children with wandering spleens were treated by a new splenopexy technique, the splenic snood. After detorsion, the spleens were wrapped in polyglycolic mesh and anchored by the mesh subdiaphragmatically in the left upper quadrant. All have retained their spleens which have remained where anchored up to a 4-year follow-up. The simplicity and technical ease of the splenic snood operation recommend it as an improved method to avoid splenectomy and safely normalize intraabdominal anatomy in the management of the wandering spleen.

Pathologic features of dogs inoculated with North American Trypanosoma cruzi isolates.

Twenty-three clinically normal Beagles were inoculated with North American Trypanosoma cruzi isolates from an opossum (Tc-O), an armadillo (Tc-A), or a dog (Tc-D). The dogs were grouped according to the clinical outcome of inoculation. Group 1 consisted of 7 dogs inoculated with Tc-O or Tc-A that died or were euthanatized during acute stages of disease. Group 2 consisted of 5 dogs inoculated with Tc-O or Tc-A, that also developed acute disease, but survived to develop chronic disease. Group 3 consisted of 7 dogs inoculated with Tc-D neither developed acute nor chronic disease. Group 4 consisted of 4 dogs and served as noninoculated controls. In group 1, the gross lesions were diffusely pale myocardiums with right ventricular enlargement, hepatomegaly, and a moderate amount of modified transudate in the abdominal cavity. Severe diffuse granulomatous myocarditis with large numbers of pseudocysts and minimal fibrosis characterized the tissues from all cardiac chambers and septum. The lesions were most severe in the right atrium and ventricle. Mild multifocal myositis and pseudocysts were observed in skeletal muscles and smooth muscles of the urinary bladder and small intestine. Multifocal encephalitis and pseudocysts were in the cerebral cortex, cerebellum, and brain stem. In group 2, the gross lesions were biventricular enlargement and thinning of the ventricular free walls. The right ventricle contained the most severe microscopic changes. There were mild multifocal interstitial lymphohistocytic cellular infiltrates, perivasculitis, and marked fibrosis in all areas of the myocardium. Mild myositis and multifocal encephalitis were seen in the skeletal muscles and brains. Pseudocysts were not observed in any tissues.(ABSTRACT TRUNCATED AT 250 WORDS)

Material and structural characterization of human saphenous vein.

Saphenous vein patch rupture after carotid endarterectomy is an infrequent but devastating complication. This study was undertaken to evaluate the material and structural properties of fresh human saphenous veins to understand the causes of this complication. Segments of saphenous veins were obtained from 22 patients from vein harvested during coronary artery bypass surgery. Ninety-three specimens, oriented in both circumferential (n = 45) and longitudinal (n = 48) directions, were prepared from the available vein segments for testing. Specimens were mounted on specially designed grips and then subjected to uniaxial tension testing. For each specimen the following material and structural parameters were determined: vessel diameter, tensile stiffness, failure and ultimate forces, and tensile modulus, failure stress, and strain. The physical properties of specimens evaluated in longitudinal orientations and thus limit the inherent strength of the vein. The physical properties of circumferentially tested vein specimens were negatively correlated to age, female gender, diabetes, and hypertension. The data obtained in this investigation suggest that age, hypertension, as well as diabetes and gender may adversely influence the circumferential tensile strength of human saphenous veins used as patch grafts.

The effects of heparin upon atherogenesis in the Watanabe heritable hyperlipidemic rabbit.

Heparin is routinely added to vascular cell cultures to inhibit smooth muscle cell proliferation. The in vivo effects of heparin upon atherogenesis have remained controversial, however. In this experiment 13 four-month-old, heritably hyperlipidemic, Watanabe rabbits were randomly assigned to three treatment groups: 1500 U heparin, 500 U heparin, or saline. An injection protocol was followed for nine months. The rabbits were bled monthly from their marginal ear veins and plasma from these bleeds was analyzed for cholesterol and triglyceride content. At sacrifice each rabbit's aorta was sampled at 10 specific locations from the arch to the iliac bifurcation. The reduction in the area of the aorta by atherosclerotic plaque at each site was calculated using a video imaging system. Heparin injections significantly reduced mean plasma levels of cholesterol and triglycerides compared to pretreatment levels and saline injected controls. However, the magnitude of the reduction was not dose-related. The aortic luminal area occupied by atherosclerotic plaque in rabbits injected with 1500 U of heparin was significantly less (p less than .05) than in rabbits injected with saline or 500 U of heparin. Plaques from heparin-treated animals contained more fatty deposits and foam cells compared to the plaques from saline-treated rabbits which were more fibromuscular in organization. We conclude that heparin modulates the occurrence and composition of atherosclerotic plaques in this animal model of naturally occurring atherogenesis.

A canine model of intimal hyperplasia (IH) in autogenous vein grafting: a preliminary report.

High failure rates (10-40% at 1 year and 2-6% per year thereafter) of autologous vein grafts in peripheral bypass surgery due to progressive intimal hyperplasia (IH) have prompted researchers to search for an animal model that develops IH in a relatively short period of time. This study summarizes our experiences in promoting IH in a canine model. Eight to ten centimeters of both jugular veins were exposed in 40 adult mongrel conditioned dogs. After division into 4-5-cm lengths, the segment of jugular vein most proximal was ballooned at 800-900 mm Hg with a modified 8F Fogarty catheter to induce intimal and medial layer injury. The distal segment was left nonballooned. Segments of these autologous vein grafts, 1.5 cm in length, both ballooned and nonballooned, were then anastomosed, end to end, into the carotid and femoral circulations. Six weeks postoperatively the grafts were perfusion-fixed, harvested, and histologically processed, and the amount of the lumen in midgraft sections occupied by IH was determined by image analysis. In all dogs, the degree of IH was significantly greater in the balloon catheterized vs noncatheterized graft segments. IH was more severe in the femoral than in the carotid arteries. The grafts that developed the most severe intimal hyperplasia were femoral grafts that had been balloon catheterized. We conclude that these protocols are effective in inducing IH in a canine model in short postoperative times.

Derivation of human microvascular endothelial cells for prosthetic vascular graft seeding.

This report summarizes our techniques and experiences deriving microvascular endothelial cells from fat for vascular graft seeding in 17 patients. Microvascular endothelial cells were derived from abdominal wall fat by collagenase incubation. The mean number of cells obtained using the described procedure was 6.83 x 10(5) cells/gram of fat processed. Histologic evaluation of the harvested cells revealed significant numbers of contaminating cell types in addition to Factor VIII-positive microvascular endothelial cells. These cells were seeded onto 6 mm ID PTFE vascular grafts in patients undergoing peripheral vascular arterial revascularization. The mean number of seeded cells was 8.04 x 10(6) cells/graft. Approximately 90 minutes were required to harvest and isolate the microvascular cells from the fat samples. We feel there are significant technical advantages to deriving endothelial cells from microvessels of human fat for vascular graft seeding.

Microvascular endothelial cell seeding of small-diameter Dacron vascular grafts.

One obstacle to the clinical implementation of endothelial cell seeding of vascular prostheses is the difficulty in derivation of large numbers of autologous endothelial cells from blood vessels of patients requiring vascular grafting. Capillary endothelial cells obtained from fat have been suggested as an abundant alternative to large-vessel endothelium for graft seeding. The object of this study was to evaluate the performance of 4-mm internal diameter (ID) Dacron Microvel grafts seeded with omentally derived microvascular endothelial cells. Six-cm lengths of the test grafts were implanted bilaterally into canine carotid arteries. One of each pair of grafts was seeded with endothelial cells (means = 8.4 x 10(6)) derived from collagenase digestion of autologous omental fat samples. The contralateral graft of each pair was nonseeded. At 5 weeks postoperatively, the grafts were harvested and evaluated. The mean patencies of both the seeded and nonseeded grafts were 89 percent. The mean thrombus-free surface area for seeded grafts was 95 +/- 11 percent. This value was significantly different statistically from the mean thrombus-free surface area of nonseeded grafts, which was 43 +/- 19 percent (P less than .05). Histologically, midgraft regions of seeded grafts were cellular, stained positive for collagen, and were characterized by inner capsules ranging in thickness between 35-94 microns. Luminal cells were identified as endothelial by peroxidase antiperoxidase staining techniques. Midgraft regions of nonseeded grafts demonstrated thrombus accumulation, limited cellularity, and inner capsules between 59-194 microns thick. Scanning electron microscopy of seeded grafts revealed smooth luminal surfaces with tight junctions between adjacent cells; surface cells were not present on midgraft regions of nonseeded grafts. In conclusion, endothelial cells derived from omental fat successfully surfaced on Dacron grafts and imparted characteristics to the graft that would predict long-term graft success.

Synthesis and 5-lipoxygenase inhibitory activity of 5-hydroperoxy-6,8,11,14-eicosatetraenoic acid analogues.

A series of eicosatetraenes (2-24) were designed, synthesized, and evaluated in vitro for inhibitory activity against 5-lipoxygenase (20000g supernatant from homogenized rat basophilic leukemia cells). All compounds were found to be active with the potencies (IC50's) ranging from 0.19 to 97 microM. Compounds containing the hydroxamic acid functionality (10-12) exhibited the best activity (IC50 = 0.19-2.8 microM). The most potent inhibitor was 5-[(hydroxyamino)carbonyl]methyl]-6,8,11,14-eicosatetraenoic acid (11), which was 10 times more active than the C-1 hydroxamates of arachidonic acid or 5-HETE. Cyclization of the linear eicosanoids 2 and 14 in the C-1 to C-5 region produced compounds (21 and 24, respectively) with several-fold greater potency.

The effects of different fixatives for immunofluorescent and immunoperoxidase localization of factor VIII-related antigen in canine carotid artery and vascular prostheses.

The evaluation of vascular grafts seeded with autologous endothelial cells requires a reliable method of endothelial cell identification. Previous attempts to identify positively tissue Factor VIII-related antigen, found in relatively large amounts in vascular endothelial cells, proved to be inconsistent when immunoperoxidase and immunofluorescent staining techniques were tried. Because the Factor VIII antigen is very labile, this study was performed to determine an optimal fixation technique for demonstrating this antigen in frozen sections of endothelial tissue. Unfixed, acetone-fixed, and formalin-fixed sections of canine carotid artery as well as vascular grafts fixed in 1-ethyl-3-(3-diaminopropyl)-carbodiimide were examined by an indirect immunofluorescence technique. Also, the immunoperoxidase method of Factor VIII identification was applied to unfixed, acetone-fixed, and carbodiimide-fixed endothelial cell seeded vascular grafts. Acetone was the preferred fixative, resulting in excellent antigen preservation with minimal background staining. The immunoperoxidase technique of Factor VIII-related antigen identification was found to be the method of choice because of its sensitivity.

Prostaglandin biochemistry of seeded endothelial cells on Dacron prostheses.

The beneficial effect of seeding endothelial cells on synthetic vascular conduits has been well established. The biochemical production and interaction of the prostaglandins, prostacyclin (PGI2) and thromboxane (TxA2), were studied on Dacron vascular grafts that were seeded with autogenous venous endothelial cells. Seventy-three seeded and nonseeded grafts were implanted into the carotid arteries of dogs. Animals were medicated with either cyclooxygenase inhibitors (aspirin and dipyridamole, or ibuprofen, or U-53,059), or dipyridamole alone, or a thromboxane synthase inhibitor, U-63557A. All animals were killed at 5 weeks and analyzed for patency, thrombus-free surface (TFS), and PGI2 and TxA2 production from mid-graft punch biopsies. PGI2 and TxA2 identifications were made by radioimmunoassay determination of 6-keto PGF1 alpha and TxB2, respectively. Results of the study demonstrated in nonmedicated animals a slightly increased patency rate in seeded vs. nonseeded grafts (50% vs. 40%) and a more significant difference in TFS (49% vs. 24%). The addition of cyclooxygenase inhibitors or TxA2 synthase inhibitors significantly improved both patency (90% vs. 47%) and TFS (87% vs. 9%) in seeded vs. nonseeded grafts. PGI2 production was decreased in seeded grafts with the use of cyclooxygenase inhibitors in all cases. It is concluded that seeded endothelial cells on Dacron velour grafts can synthesize PGI2; these PGI2 levels are far less than PGI2 levels produced by endothelial cells from the adjacent carotid artery; and TxA2 synthase inhibitors best improve thromboresistance of seeded grafts without significant reduction in PGI2 production.

Effects of antiplatelet agents in combination with endothelial cell seeding on small-diameter Dacron vascular graft performance in the canine carotid artery model.

The purpose of this study was to assess the success of endothelial cell-seeded and non-seeded small-diameter vascular grafts in dogs medicated with antiplatelet agents. Eighty dogs underwent bilateral carotid artery replacements with 6 cm lengths of 4 mm I.D. double-velour Dacron grafts. In each dog one graft was seeded with enzymatically derived autologous endothelial cells; the contralateral graft was nonseeded. The following anti-platelet medications were administered beginning 4 days preoperatively: aspirin (5 grains every day); dipyridamole (50 mg twice a day); aspirin plus dipyridamole (5 grains each day plus 50 mg twice a day); aspirin (1.25 grains every other day); ibuprofen (10 mg/kg/day); U-53,059, a cyclooxygenase inhibitor (3 mg/kg/day); and U-63557A, a thromboxane synthase inhibitor (10 mg/kg/day). Grafts were harvested 5 weeks postoperatively. Graft success was evaluated by patency, thrombus-free surface area, area endothelialized, and graft production of prostacyclin. None of the medications prevented neoendothelialization of seeded grafts. Mean patencies of endothelial cell-seeded grafts from medicated dogs were significantly greater than mean patencies of endothelial cell-seeded grafts from nonmedicated dogs. The cyclooxygenase inhibitors best maintained patency in nonseeded grafts. Thrombus-free surface areas of endothelial cell-seeded grafts from medicated dogs were significantly greater than from nonseeded control grafts from the medicated dogs. All medications impaired prostacyclin synthesis. We conclude that the combination of endothelial cell seeding plus antiplatelet medication is most efficacious in small-vessel grafting success and that high levels of prostacyclin production by vascular grafts are not necessary to maintain patency in dogs medicated with antiplatelet agents.