Polymeric sustained local drug delivery system for the prevention of vascular intimal hyperplasia Deenu Kanjickal,1 Stephanie Lopina,1 M. Michelle Evancho-Chapman,2 Steven Schmidt,2 Duane Donovan,2 Sara Springhetti2 1 Department of Chemical Engineering, University of Akron, Akron, Ohio 44325 2 Falor Division of Surgical Research, Summa Health System, Akron, Ohio 44304 Received 1 July 2003; revised 3 September 2003; accepted 9 September 2003
Abstract: Anastomotic intimal hyperplasia (IH) is a major cause of both autologous vein and synthetic vascular graft failure. We have previously published data suggesting that cyclosporin may reduce the development of IH in a canine model. However, systemic administration of cyclosporin could create serious adverse effects. Therefore, it is our long-term goal to test the hypothesis that the controlled local release of cyclosporin from a polymeric vascular wrap will prevent the development of IH. To test this hypothesis, we developed a controlled release vascular wrap (sheet/ring) using a poly(ethylene glycol) (PEG) hydrogel. Sterilization of the polymers was performed using the ethylene oxide and hydrogen peroxide sterilization methods. It was found that except for one combination (8000 molecular weight and 1:1
INTRODUCTION Pathology The primary mechanism of failure for vascular surgical procedures listed in Table I is intimal hyperplasia (IH). IH is a chronic structural change of the vascular wall characterized by diffuse thickening of the intimal layer secondary to the migration and proliferation of medial smooth muscle cells. The innermost layer of a vessel is the intima, which is composed of endothelial cells. Next is the internal elastic lamina and then the media made of smooth muscle cells. The outer two layers are the external elastic lamina and finally the adventitia. Our drug delivery device is designed to be wrapped around the adventitial layer. IH is an exaggeration of the normal vascular wall healing response that may be induced by graft compliance mismatch, flow turbulence, shear stress, vessel stretch, surgical Correspondence to: S. Lopina; e-mail:
[email protected] Contract grant sponsor: Summa Health System Foundation © 2003 Wiley Periodicals, Inc.
crosslinking ratio), the differences in the swelling ratios for the sterilized and unsterilized hydrogels were not statistically significant. Release studies from unsterilized and ethylene oxide-sterilized PEG hydrogels were conducted. It was found that release lasted for approximately 50 h for sterilized as well as unsterilized PEG hydrogels. Acute animal studies, to test the deployment of both the polymeric sheets and rings to the adventitial surface of native arteries and veins, were completed successfully. © 2003 Wiley Periodicals, Inc. J Biomed Mater Res 68A: 489 – 495, 2004 Key words: intimal hyperplasia; drug delivery; cyclosporin; hydrogel; poly(ethylene glycol) (PEG)
trauma, mural ischemia, and luminal accumulation of various biochemical factors released from deposited fibrin and platelets.1 These factors then change the phenotype of the smooth muscle cells from a normal contractile or differentiated state to a synthetic form resulting in cell proliferation.2 The vessel wall then increases in size due to the increased number of cells, subsequently reducing the vessel’s luminal size. This decreased luminal size can result in decreased blood flow, vessel thrombosis, and finally graft failure. In a previous study,3 the systemic delivery of cyclosporin A (CyA) was able to inhibit IH in a canine model of venous IH. CyA is a potent immunosuppressive agent, usually administered orally or intravenously, that has been used to prevent graft rejection in organ transplant recipients.4 When administered orally, it has a low and variable bioavailability and high doses are believed to lead to nephrotoxicity.5,6 Therefore, alternative routes of drug administration are under study that would lead to better drug efficacy. A controlled and site-specific drug therapy would be advantageous over traditional drug delivery mechanisms by improving the efficacy and reducing
490
KANJICKAL ET AL.
TABLE I Percentage of Failure of Various Surgical Procedures in the United States26 –29 Surgical Procedures
Cases/Year
Failure %
Dialysis access grafts Peripheral arterial bypass surgery Coronary artery bypass Angioplasty with or without stent Carotid endarterectomy
150,000
60–80 at 1 year
200,000 220,000
30–60 at 5 years 20–30 at 2 years
500,000 120,000
20–40 at 5 years 5–10 at 5 years
the side effects of CyA. Local delivery to the injury site could achieve the desired drug concentration using relatively lower doses, thus avoiding nephrotoxicity. A number of vehicle systems have been developed for the perivascular delivery of drugs.7–9 For example, adventitial delivery of heparin from a polymeric matrix implanted adjacent to the injured artery in a rat model of restenosis has demonstrated a reduction in neointimal formation.10 Simons et al.11,12 have also reported inhibition of restenosis using a pluronic gel as an adventitial implant to deliver antisense proliferating cell nuclear antigen oligonucleotides. The advantages of a perivascular system of delivery include relative simplicity, generally large drug-loading capacities, and high reproducibility of release profiles.
Poly(ethylene glycol) (PEG) hydrogel PEG is a water-soluble polymer having desirable properties for biomedical applications including nonimmunogenicity and protein resistance, therefore resisting recognition by the immune system.13 PEG is nontoxic and has been approved by the United States Food and Drug Administration for internal consumption.14 –16 Intravenously administered PEG clears rapidly through the kidney.17 A number of PEG-based biomaterials have achieved clinical applications including Vigilon™ and Hypol™ and PEG hydrogels have been used extensively as drug delivery devices in the past decade. PEG has been used for a variety of biomedical applications such as biological purifications, pegylated proteins and peptides for medical applications, and synthesis of artificial enzymes.18 The two most common methods of PEG hydrogel preparation are radiation crosslinking19 –21 and chemical crosslinking.22,23 In the radiation crosslinking technique, the crosslinking between PEG chains is random; hence, controlling the structure of the gels is difficult using this technique. In contrast, the chemical crosslinking technique offers exact crosslinking ratios, and the chemical reactions do not involve elaborate procedures or equipment. The chemical crosslinking agent used is triphenyl methane triisocyanate (Desmo-
dur RE) and the choice was based on familiarity and availability.
Goals and objectives The objectives of this study were to synthesize and evaluate the PEG hydrogel sheets/rings for use as an adventitial vascular wrap. To evaluate the physical properties of the devices, studies were conducted to: 1. determine the swelling ratios of the polymers with different crosslinking ratios, 2. determine the effects of sterilization [ethylene oxide (EtO) and hydrogen peroxide] on the PEG hydrogels, 3. determine the length of drug delivery, and 4. test the applicability of the polymeric configurations in terms of their physical properties for deployment around a native vessel in an acute canine model.
MATERIALS AND METHODS Materials PEG with 3350 and 8000 average molecular weights (MWs) was purchased from Sigma (St. Louis, MO). Desmodur RE was received as a gift from Bayer (Pittsburgh, PA). Dichloromethane was bought from Aldrich (Milwaukee, WI) and was used as received, without further purification. An oral formulation of cyclosporin, Neoral威, was obtained from Novartis (East Hanover, NJ). Neoral威 oral solution is composed of 100 mg/mL cyclosporin, 11.9% v/v USP dehydrated alcohol, and the inactive ingredients corn oilmono-di-triglycerides, polyoxyl 40 hydrogenated castor oil NF, DL-␣-tocopherol USP, and propylene glycol USP.
Hydrogel preparation The hydrogels were synthesized by chemical crosslinking of PEG with triphenyl methane triisocyanate (Desmodur RE) as described by Yoshikawa et al.24 Briefly, 25% (weight) dichloromethane solution of Desmodur RE was added to 25% (weight) solutions of different MWs of PEG. The molar ratio of the hydroxyl group and the isocyanate moiety was adjusted to get the desired crosslinking ratio as defined by the ratio of the number of the isocyanate moieties to the number of hydroxyl groups. The solution was continuously stirred at a temperature of 70°C until the solution became fairly viscous. The mixture was then transferred to a flat dish for a thin vascular sheet of polymer or to an aluminum dish for the block polymer, which was then kept in a constant temperature water bath at 70°C for 24 h. The polymeric sheet/block polymer was removed from the dish by swelling it with deionized water. The sheets were then dried in a vacuum oven at 40°C for 24 h and were reswelled and dried 2–3 more times to remove unreacted polymer chains. The
POLYMERIC SUSTAINED LOCAL DRUG DELIVERY SYSTEM
rings of polymer were extracted from the block polymer by using a cork borer.
Sterilization The EtO sterilization and hydrogen peroxide sterilization were done in an Amsco Eagle series 3058 gas sterilizer and a Sterrad威 100S sterilizer (Johnson and Johnson), respectively.
491
These devices were tested in two acute canines in the Applied Surgical Research Lab, Summa Health System. The devices were placed on native canine jugular vein, femoral vein, and femoral artery for up to 4 h. All procedures were reviewed and approved by Summa Health System’s Institutional Animal Care and Use Committee. The study adhered to standards of animal care and use published by the National Institutes of Health in The Guide for the Care and Use of Laboratory Animals and were conducted in an AAALAC accredited facility (AAALAC—Association for the Assessment and Accreditation of Laboratory Animal Care, International).
Drug/dye loading RESULTS AND DISCUSSION PEG hydrogels were loaded with rhodamine b by immersing them in a saturated solution of the dye in deionized water for a period of 5 days. Neoral威 is an oral formulation of cyclosporin that immediately forms a microemulsion in an aqueous environment. A 20% (by volume) solution of Neoral威 in deionized water was used for drug loading. The dried PEG hydrogels were diffusion loaded with cyclosporin by immersing them in the Neoral威 solution for a period of 5 days and the drug loading percentage achieved for Neoral威 was 1.66. The drug loading was calculated as a ratio of the drug taken up upon loading to the wet weight of the hydrogel.
Drug/dye diffusion After the PEG hydrogel samples were loaded with the drug/dye, they were transferred to bottles (amber teflon bottles for cyclosporin and amber glass bottles for rhodamine b) containing phosphate-buffered saline. These bottles were kept in a shaker bath (Innova 2000 at 60 rpm), which was placed in an incubator (Forma Scientific) maintained at 37°C. Samples were taken out at regular intervals and replaced with fresh buffer. Release of cyclosporin was monitored continuously with a TDxFLx威 system (Abbott Laboratories).25 This system uses a fluorescence polarization immunoassay technology to determine the concentration of CyA.25 Rhodamine concentrations were monitored using a UV spectrophotometer (Beckman) at a wavelength of 553 nm.
Animal studies Two PEG hydrogel controlled release delivery systems were developed for animal studies to test the applicability of the polymers in terms of their physical properties for deployment around a native vessel. The first delivery system consisted of thin PEG hydrogel sheets (approximately 6.5 ⫻ 4 ⫻ 0.01 cm), which were wrapped around the vessel. The second delivery system used PEG hydrogel blocks cut into rings of various sizes and diameters that could be snapped into place around a vessel.
Swelling behavior of sterilized and unsterilized PEG hydrogels Because the PEG hydrogels will have to undergo a sterilization procedure before application in vivo, we investigated the differences in swelling characteristics of unsterilized, EtO-sterilized, and hydrogen peroxide-sterilized PEG hydrogels. Tests were performed on hydrogels of different crosslinking ratios. The hydrogels were cut into disks and then placed in deionized water at 37°C. The weights of the swelled gels were taken after 24 h. The weight swelling ratio was determined by the following equation: G⫽
共W ⫺ W 0 兲 W0
W ⫽ hydrogel weight after 24 h of swelling and W0 ⫽ xerogel weight. The swelling studies were done in triplicate. A three-way analysis of variance using Minitab威 software was used for comparing the swelling ratios for the unsterilized and the sterilized polymers. A completely randomized 2 ⫻ 3 ⫻ 4 factorial design with three replicates per combination was used. The swelling ratios obtained are an average of three readings (Table II). The analysis does not show evidence of differences in sterilization except for the 8000 MW 1:1 crosslinking ratio polymer. For this particular combination, the EtO-sterilized polymer had a statistically significant reduction in swelling (at a level of 0.05) as compared with the unsterilized (p ⫽ 0.0002) and the hydrogen peroxide-sterilized (p ⫽ 0.0000) polymer. Also, the hydrogen peroxide-sterilized hydrogel had a statistically significant increase in the swelling ratio as compared with the unsterilized hydrogel (p ⫽ 0.003). The Fisher’s method was used for the comparison tests. We believe that these sterilization procedures may induce the formation of free radicals that affect the properties of the polymer. In our case, the EtO steril-
492
KANJICKAL ET AL.
TABLE II Swelling Fractions of Sterilized and Unsterilized PEG Hydrogels for Different Crosslinking Ratios and Molecular Weightsa Swelling Ratio Crosslinking Ratio 3350 MW PEG hydrogel 1:1 2:1 3:1 4:1 8000 MW PEG hydrogel 1:1 2:1 3:1 4:1
Unsterilized Samples
EtO-Sterilized Samples
Hydrogen PeroxideSterilized Samples
4.01 ⫾ 0.08 1.251 ⫾ 0.004 0.74 ⫾ 0.02 0.470 ⫾ 0.006
3.9 ⫾ 0.2 1.24 ⫾ 0.01 0.69 ⫾ .04 0.469 ⫾ 0.003
4.1 ⫾ 0.1 1.29 ⫾ 0.02 0.74 ⫾ .01 0.495 ⫾ 0.003
7.8* ⫾ 0.3 3.4 ⫾ 0.1 2.1 ⫾ 0.3 1.17 ⫾ 0.06
7.4** ⫾ 0.1 3.5 ⫾ 0.1 2.00 ⫾ 0.09 1.170 ⫾ 0.007
8.2*** ⫾ 0.5 3.50 ⫾ 0.02 2.2 ⫾ 0.1 1.18 ⫾ 0.02
a
Values obtained are an average over three measurements. *, **, ***Statistical difference as determined by Fisher’s method (p ⬍ 0.05).
ization leads to a higher crosslinking of the polymer and hence a reduction in the swelling ratio. However, the hydrogen peroxide sterilization breaks down the crosslinks in the polymer leading to an increase in the swelling ratio. The “dual effect” of the free radicals on the polymers may be attributed to the relative amount of free radicals that are produced upon sterilization. These effects will be more pronounced on hydrogels with large swelling ratios, increasing as the MW of the polymer increases and as the crosslinking density decreases. In this study, the differences in the sterilization methods were observed for only the 8000 MW 1:1 crosslinked polymer, the polymer with the highest swelling ratio as compared with the other hydrogels investigated.
Release from sterilized and unsterilized PEG hydrogels Solute transport was studied from sterilized and unsterilized PEG hydrogels produced from 3350 MW linear PEG. Rhodamine b is a dye that has been used as a model drug for release studies. The concentration of rhodamine b in phosphate-buffered saline was monitored continuously with the UV spectrophotometer at an exciting wavelength of 553 nm. The release of rhodamine b from a 1:1 crosslinked PEG hydrogel is shown in Figure 1(A). The release readings are an average over four samples. As can be seen from the figure, the release of rhodamine b from the PEG hydrogels, both sterilized and unsterilized, was quick, extending over approximately 8 h. Note that the release characteristic remained the same for the sterilized as well as the unsterilized PEG hydrogels, following similar trends. Figure 1(B) shows the release of rhodamine b from a 3:1 crosslinked polymer and here
also the release characteristics remain similar for the sterilized and unsterilized hydrogels. The duration of release is longer for the hydrogel with increased crosslinking ratio. The conclusion from the rhodamine b release studies is that the sterilization did not have a significant effect on the release of rhodamine from the PEG hydrogels. Cyclosporin is a stable compound if solutions are protected from light. However, cyclosporin tends to adsorb on container walls decreasing its concentration in solution; therefore, the release was performed in amber teflon bottles. The release was done in phosphate-buffered saline kept at 37°C. A TDxFLx威 fluorescent assay was used to determine the concentration of CyA. Figure 2 depicts the release of CyA from unsterilized, EtO-sterilized, and hydrogen peroxidesterilized PEG hydrogels. Release studies were done on 3:1 crosslinked polymer and the release continued for approximately 50 h. CyA is a hydrophobic molecule as compared with rhodamine b, which is hydrophilic. Also, rhodamine b is a small molecule with a MW of 479 Da as compared with 1202.6 Da of CyA. This might explain the differences in the release times for drug and the dye. The sterilization did not have a significant impact on the release characteristics of CyA from a 3:1 crosslinked PEG hydrogel.
Animal studies The PEG sheets and rings were successfully placed peri-adventitially around canine jugular veins, femoral veins, and femoral arteries. (Fig. 3) The PEG sheet configured drug delivery systems were wrapped around two canine jugular veins. The sheets were flexible and were readily placed in close contact with
POLYMERIC SUSTAINED LOCAL DRUG DELIVERY SYSTEM
493
studies. Swelling studies were conducted to investigate the effect of different sterilization procedures on PEG hydrogels. It was found that there is a statistically significant decrease in the swelling ratio for the EtOsterilized 8000 MW 1:1 crosslinked polymer whereas the hydrogen peroxide sterilization leads to a statistically significant increase in the swelling ratio. Release from these sterilized polymers was also measured. We were able to achieve release of CyA from a 3:1 crosslinked 3350 MW polymer for a span of approximately 50 h. Preliminary acute animal studies have shown that the polymers can be applied successfully to the site of potential IH occurrence. The ring config-
Figure 1. Percentage release of rhodamine b in phosphatebuffered saline at 37°C from PEG hydrogels having a crosslinking ratio of (A) 1:1 MW of 3350, and (B) 3:1 MW of 3350. Each point on the graph is an average over four measurements. Percentage released ⫽ Mt/M⬁ * 100; Mt ⫽ concentration of drug/dye at time t; and M⬁ ⫽ maximum concentration of drug/dye obtained in the release study.
the venous adventitia. However, it was difficult to secure the wrap into place. When we passed a ligature around the construct, the sheet would tear when the tie was tightened into place. The controlled drug delivery systems configured as PEG rings were easily snapped into place around both the canine femoral artery and vein. To assure close contact with the vessel adventitia, the internal diameters of the rings were matched to the diameter of the vessels studied. The wall thickness was kept at a standard thickness of 3.5 mm and was 7 mm in length. Multiple rings could be place on a vessel to optimize the area over which drug could be delivered. The design of the ring construct is such that no additional retention devices were needed to keep the ring in close contact with vessel adventitia. In addition, the ring construct provided a larger reservoir into which drug could be loaded. CONCLUSION PEG hydrogels were synthesized using a chemical crosslinking technique and characterized by swelling
Figure 2. Percentage release of cyclosporin in phosphatebuffered saline at 37°C from Neoral威 loaded (A) unsterilized, (B) EtO-sterilized, and (C) hydrogen peroxide-sterilized PEG hydrogels having a crosslinking ratio of 3:1 and an average MW of 3350. Each point on the graph is an average over four measurements.
494
KANJICKAL ET AL.
Figure 3. Acute animal studies. (A) PEG hydrogel sheet wrapped around a canine jugular vein. (B) PEG ring placed around a canine femoral vein. (C) Two PEG rings placed around a canine femoral artery.
uration can be manufactured in a variety of diameters and widths. In the animal model, the rings polymers performed better than the polymer sheets, because they adapted more readily to vessel size, shape, and anastomotic angles. The authors express their gratitude to Dr. Nicholas O’Donnell (Summa Health System, Toxicology Lab) and Dr. Richard Einsporn (University of Akron, Statistics Department) for helping with the TDxFLx威 system and the statistical analysis, respectively. They also thank Bayer for the donation of Desmodur RE.
References 1.
Chervu A, Moore W. An overview of intimal hyperplasia. Surg Gynecol Obstet 1990;171:433– 447.
2.
Clowes AW, Schwartz SM. Significance of quiescent smooth muscle migration in the injured rat carotid artery. Circ Res 1985;56:139 –145. 3. Hirko MK, McShannic JR, Schmidt SP, Sharp WV, Evancho MM, Sims RL, Siebert JD. Pharmacologic modulation of intimal hyperplasia in canine vein interposition grafts. J Vasc Surg 1993;17:877— 887. 4. Ptachcinski RJ, Venkataramanan R, Burckart GJ. Clinical pharmacokinetics of cyclosporin. Clin Pharmacokinet 1986;11:107–132. 5. Calne RY, White DJG, Thiru S, Evans DB, McMaster P, Dunn DC, Craddock GN, Pentlow BD, Rolles K. Cyclosporin A in patients receiving renal allografts from cadaver donors. Lancet 1978;2:1323–1327. 6. Calne RY. Cyclosporin. Nephron 1980;26:57– 63. 7. Golomb G, Mayberg M, Domb AJ. Polymeric perivascular delivery systems. In: Domb AJ, editor. Polymeric site-specific pharmacotherapy. New York: John Wiley & Sons; 1994. p 205–219. 8. Golomb G, Moscovitz D, Fishbein I, Mishaly D. Perivascular polymeric controlled-release therapy. In: Wise DL, editor. En-
POLYMERIC SUSTAINED LOCAL DRUG DELIVERY SYSTEM
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
cyclopedic handbook of biomaterials and bioengineering. New York: Marcel Dekker; 1995. p 1043–1055. Teomim D, Fishbein I, Golomb G, Orloff L, Mayberg M, Domb AJ. Perivascular delivery of heparin for the reduction of smooth muscle cell proliferation after endothelial injury. J Control Release 1999;60:129 –142. Edelman ER, Adams DH, Karnovsky MJ. Effect of controlled adventitial heparin delivery on smooth muscle cell proliferation following endothelial injury. Proc Natl Acad Sci USA 1990;87:3773–3777. Simons M, Edelman ER, DeKeyser JL, Langer R, Rosenberg RD. Antisense c-myb oligonucleotides inhibit intimal arterial smooth muscle cell accumulation in vivo. Nature 1992;359:67–70. Simons M, Edelman ER, Rosenberg RD. Antisense proliferating cell nuclear antigen oligonucleotides inhibit intimal hyperplasia in a rat carotid artery injury model. J Clin Invest 1994; 93:2351–2356. Kondo A, Kishimura M, Katoh S, Sada E. Improvement of proteolytic resistance of immunoadsorbents by chemical modification with polyethylene glycol. Biotechnol Bioeng 1989;34: 532–540. Herold DA, Keil KA, Bruns DE. Oxidation of polyethylene glycols by alcohol dehydrogenase. Biochem Pharmacol 1989; 38:73–76. Smyth HF Jr, Carpenter CP, Weil CS. The toxicology of the polyethylene glycols. J Am Pharm Assoc 1950;39:349 –354. Johnson AJ, Darpatkin MH, Newman J, Margaret H. Clinical investigation of intermediate and high-purity antihemophilic factor (factor VIII) concentrates. Br J Haematol 1971;21:21– 41. Shaffer CB, Critchfield FH. The absorption and excretion of the solid polyethylene glycols (carbowax compounds). J Am Pharm Assoc 1947;36:152–257. Harris JM. Poly(ethylene glycol) chemistry: Biotechnical and biomedical applications. New York: Plenum Press; 1992. p 1–14.
19.
20.
21.
22.
23. 24.
25.
26.
27.
28.
29.
495
Peppas NA, Keys KB, Torres-Lugo M, Lowman AM. Poly(ethylene glycol)-containing hydrogels in drug delivery. J Control Release 1999;62:81– 87. Stringer JL, Peppas NA. Diffusion of small molecular weight drugs in radiation-crosslinked poly(ethylene oxide) hydrogels. J Control Release 1996;42:195–202. Merrill EW, Dennison KA, Sung C. Partitioning and diffusion of solutes in hydrogels of poly(ethylene oxide). Biomaterials 1993;14(15):1117–1126. Iza M, Stoianovici G, Viora L, Grossiord JL, Couarraze G. Hydrogels of poly(ethylene glycol): Mechanical characterization and release of a model drug. J Control Release 1998;52:41– 51. Graham NB, McNeill ME. Hydrogels for controlled drug delivery. Biomaterials 1984;5:27–36. Yoshikawa M, Shiota A, Sanui K, Ogata N. Preparation of polyoxyethylene hydrogel membranes and their ultrafiltration performance. New Polymeric Mater 1989;1(3):223–234. TDx/TDxFLx cyclosporin monoclonal whole blood. Abbott Laboratories, Diagnostics Division, Abbott Park, IL 60064. List no. 9797-60, 69-4972/R6, July 2000. Miller PE, Carlton D, Deierhoi MH, Redden DT, Allon M. Natural history of arteriovenous grafts in hemodialysis patients. Am J Kidney Dis 2000;36:68 –74. Jackson MR, Belott TP, Dickason T, Kaiser WJ, Modrall JG, Valentine RJ, Clagett GP. The consequences of a failed femoropopliteal bypass grafting: Comparison of saphenous vein and PTFE grafts. J Vasc Surg 2000;32:498 –505. Johnson WC, Lee KK. A comparative evaluation of polytetrafluoroethylene, umbilical vein, and saphenous vein bypass grafts for femoral-popliteal above-knee revascularization: A prospective randomized Department of Veterans Affairs cooperative study. J Vasc Surg 2000;32:268 –277. American Heart Association. 2002 Heart and stroke statistical update. Dallas, TX: American Heart Association; 2001.
