81Ann Thorac Cardiovasc Surg Vol. 14, No. 2 (2008)

OriginalArticle

Comparison of the Effects of Aortic Valve ReplacementUsing 19-mm Carpentier-Edwards PerimountBioprosthesis and 19-mm Medtronic Mosaic

Bioprosthesis

From Cardiovascular Surgery, Nagasaki University School ofMedicine, Nagasaki, Japan

Received April 13, 2007; accepted for publication June 28, 2007Address reprint requests to Kazuyoshi Tanigawa, MD: Cardiovas-cular Surgery, Nagasaki University School of Medicine, 1–7–1Sakamoto, Nagasaki 852-8501, Japan.

Objective: Short (≤3 months)- and middle (≥4 months)-term results of aortic valve replacement(AVR) using 19-mm Carpentier-Edwards Perimount (CEP) bioprosthetic valves and 19-mmMedtronic Mosaic (MM) bioprosthetic valves in patients with small aortic annulus were com-pared.Patients and Methods: At our facility, AVR was performed using bioprostheses in 110 patientsfrom April 1999 to March 2006. Of these patients, 40 were treated using 19-mm CEP (Group C),and 9 using 19-mm MM (Group M). Evaluation by inquiry, physical examination, andechocardiography was performed before, a short term after, and a middle term after surgery,and the effects of AVR were compared.Results: The New York Heart Association (NYHA) functional class grade showed improve-ments in both groups. The aortic valve peak pressure gradient was 29.8 ± 10.1 mmHg in GroupC and 53.8 ± 17.3 mmHg in Group M, being higher in Group M, a middle term after surgery.However, the left ventricular mass index (LVMI) showed improvements in both groups com-pared with the values before surgery, and the left ventricular ejection fraction (LVEF) was main-tained. During the middle term after surgery, the frequency of cardiac events showed no signifi-cant difference between the two groups.Conclusions: In the patients treated with 19-mm MM, the aortic valve peak pressure gradientwas higher than in those treated with 19-mm CEP, but acceptable improvements in the LVMI,maintenance of the LVEF, and avoidance of cardiac events were observed in both groups. (AnnThorac Cardiovasc Surg 2008; 14: 81–87)

Key words: aortic valve replacement, bioprosthetic valves, small aortic annulus

Kazuyoshi Tanigawa, MD, Kiyoyuki Eishi, MD, Shiro Yamachika, MD,

Koji Hashizume, MD, Seiichi Tada, MD, Kentaro Yamane, MD, Kenta Izumi, MD,

Hideaki Takai, MD, Takashi Miura, MD, and Shun Nakaji, MD

Introduction

With the recent aging of society, elderly patients receiv-ing aortic valve replacement (AVR) are increasing, and

the selection of bioprosthetic valves for this proceduretakes into consideration their good durability and low in-cidence of valve-related complications.1) Patient-prosthe-sis mismatch (PPM), first described by Rahimtoola2)

(“Mismatch can be considered to be present when theeffective prosthetic valve area, after insertion into thepatient, is less than that of the normal valve”), is consid-ered to be unavoidable as long as artificial stented valvesare used because of their structure. Thereafter, Pibarot etal.3,4) further evaluated the effects on the hemodynamics,complications, and mortality rate, and many studies are

82

Tanigawa et al.

Ann Thorac Cardiovasc Surg Vol. 14, No. 2 (2008)

also conducted today.The Carpentier-Edwards Perimount (CEP) bioprosthesis(Baxter Healthcare Corp., Edwards Division, Santa Ana,Calif.) and Medtronic Mosaic (MM) bioprosthesis(Medtronic, Inc., Minneapolis, Minn.) are stentedbioprostheses that are available in Japan among thosewidely used throughout the world. Because of the gener-ally small physique of the Japanese, many Japanese pa-tients have as mall aortic annulus, and small-sizedbioprostheses must be selected with a possible increasein the risk of PPM.

Since in many of our patients we noted a relatively highaortic valve pressure gradient after AVR using a bioprosthesis,we reviewed short (≤3 months)-term and middle (≥4months)-term results of AVR using 19-mm bioprostheses atour facility to evaluate the effects of the surgery.

Patients and Methods

From April 1999 to March 2006, 110 patients underwentAVR using bioprostheses among those who required AVRfor aortic valve stenosis (AS), aortic valve regurgitation(AR), aortic valve stenosis and regurgitation (ASR), orthoracic aortic disorders that necessitated AVR (includ-ing those who underwent this procedure as part of com-plex surgery for disorders that included angina pectorisand combined valvular disease). Of these patients, 40 weretreated using 19-mm CEP (Group C), and 9 with 19-mmMM (Group M). The preoperative data of each group aresummarized in Table 1.In Groups C and M, the following items were evaluatedby inquiry, physical examination, and echocardiographyat our facility before, a short term after, and a middleterm after surgery. Concerning patients who could not beexamined at our facility, we adopted data collected frominterviews with the patients and their families or fromthe results of evaluations by their family doctors.

Evaluation items: New York Heart Association(NYHA) functional class, Echocardiographic Measure-ment (left ventricular morphology and function, aorticvalve peak pressure gradient, left ventricular mass(LVM),5) left ventricular mass index (LVMI) = LVM/bodysurface area, cardiac event, and survival rate.

LVM (grams) = 0.8 × {1.04 × [(LVIDd+IVSd+PWd)3 –(LVIDd)3]} + 0.6, LVIDd = left ventricular internal dimen-sion at end-diastole, IVSd = interventricular septal thick-ness at end-diastole, and PWd = posterior wall thicknessat end-diastole.

Implanted bovine and porcine bioprosthesesThe CEP is a bovine stented heart valve fixed with glut-araldehyde at low pressure, and the cusps are treated witha surfactant to restrict calcification. Clinical investigationbegan in 1981 in the United States and has been in clini-cal use in Japan since 1985 and in the U.S. and Europesince 1991.

MM is a porcine stented heart valve fixed with glut-araldehyde at zero pressure and treated with α-amino-oleic acid to reduce tissue calcification. It has been inclinical use in Europe since 1994, in Japan since 1999,and in the U.S. since 2000.

Surgical techniqueAVR was performed under normal-temperature extracor-poreal circulation and cold crystalloid cardioplegic solu-tion. After the aortic valve was removed, calcificationsobserved at the aortic root or aortic annulus were removedas much as possible, and the size of the aortic annuluswas measured with a sizer for each bioprosthesis. If valvereplacement in the intra-annular position was possible,the prosthesis was sutured by everting mattress suture,but if not, it was sutured to the supra-annular position byinterrupted suture. The operative profile of each group issummarized in Table 2.

Statistical analysisThe data were expressed as the mean ± standard devia-tion. The Student’s t-test was performed to examine dif-ferences between the two independent groups. Differencesbetween the two groups concerning each intraindividualor interindividual factor were examined by repeated mea-sures ANOVA. Further, the survival rate was calculatedby the Kaplan-Meyer method, and differences in the sur-vival rate between the two groups were examined by thelog-rank test. Differences were considered significant atp < 0.05.

Results

NYHA functional classPreoperative and latest postoperative subjective symptomswere evaluated according to the NYHA. The NYHA im-proved from a preoperative mean of 2.8 to a postopera-tive mean of 1.3 in Group C and from 3.3 to 1.9 in GroupM. Although a higher percentage of patients in Group Cwere rated NYHA II, the differences between the twogroups were insignificant (Fig. 1).

Effects of Aortic Valve Replacement

Ann Thorac Cardiovasc Surg Vol. 14, No. 2 (2008) 83

Aortic valve peak pressure gradientIn Group C, the aortic valve peak pressure gradients were29.7 ± 10.9 mmHg a short term after surgery and 29.8 ±10.1 mmHg a middle term after surgery. In group M,they were 45.4 ± 12.6 mmHg and 53.8 ± 17.3 mmHg,respectively. Compared with the preoperative pressuregradients calculated by excluding AR patients, in whomthe aortic pressure gradient was not high (90.9 ± 27.2

Table 1. Preoperative data

NYHA, New York Heart Association; NS, not significant.

Group C Group M p valueNo. of patients 40 9Age on implantation

Mean 75.8 ± 4.5 75.2 ± 3.7 NSRange 72–83 61–84

Sex NSMale 8 1Female 32 8

Body surface area (m2) 1.41 ± 0.16 1.43 ± 0.20 NSCardiac rhythm

Sinus rhythm 35 9Atrial fibrillation 4 0Heart block 0 0Paced rhythm 1 0

NYHA calssificationClass I 0 0Class II 20 0Class III 9 6Class IV 11 3

Aortic valve lesionIsolated stenosis 25 7Stenosis and regurgitaton 8 1Isolated regurgitaton 7 1

Left ventricular mass index (g/m2) 171.4 ± 42.2 170.7 ± 15.2 NS

Table 2. Operative profile

AVR, aortic valve replacement; CABG, coronary arterybypass grafting; MVR, mitral valve replacement; MP, mi-tral valve plasty; MAZE, maze procedure.

Group C Group MBioprosthesis position

Intra-annular 28 3Supra-annular 12 5

Concomitant proceduresAVR 24 6AVR + CABG 6 1AVR + MVR 5 (+ MAZE 1) 1AVR + MP 3 1 (+ CABG)AVR + MAZE 1 0AVR+Total Arch Replacement 1 0

Fig. 1. Change of New York Heart Association classification.

mmHg in Group C, 77.0 ± 30.8 mmHg in Group M),the pressure gradient was significantly decreased in bothgroups in the short and middle terms after surgery(p<0.001). However, the postoperative pressure gradi-ent was slightly higher in Group M than in Group C inboth the short and middle terms after surgery (p=0.08)(Table 3, Fig. 2).

84

Tanigawa et al.

Ann Thorac Cardiovasc Surg Vol. 14, No. 2 (2008)

LVMIIn Group C, the LVMI was 171.4 ± 42.2 g/m2 before sur-gery, 139.3 ± 41.4 g/m2 a short term after surgery, and118.9 ± 47.4 g/m2 a middle term after surgery. In groupM, it was 170.7 ± 15.2 g/m2, 156.0 ± 34.6 g/m2, and 136.4± 14.4 g/m2, respectively. Although significant decreases(p<0.001) were observed in both groups during both theshort and middle terms after surgery, the improvementtended to be greater in Group C (p=0.40) (Table 3, Fig.3).

Left ventricular ejection fraction (LVEF)In group C, the LVEF was 65.5 ± 11.9% before surgery,65.2 ± 12.6% a short term after surgery, and 70.0 ± 12.2%a middle term after surgery. In group M, it was 64.7 ±16.0%, 60.6 ± 16.5%, and 67.4 ± 17.5%, respectively.

In both groups, the LVEF was similar to or better thanthe values before surgery, indicating satisfactory mainte-nance of the left ventricular function (p=0.01). However,

Table 3. Short-and middle-term results of transvalvular pressure gradient and LVMI and EF

LVMI, left ventricular mass index; EF, ejection fraction. † Values calculated by exclusion of pa-tients with aortic regurgitation.

Preoperative Short term Middle termTransvalvular pressure gradient (mmHg)

Group C (90.9 ± 27.2)† 29.7 ± 10.9 29.8 ± 10.1Group M (77.0 ± 30.8)† 45.4 ± 12.6 53.8 ± 17.3

LVMI (g/m2)Group C 171.4 ± 42.2 139.3 ± 41.4 118.9 ± 47.4Group M 170.7 ± 15.2 156.0 ± 34.6 136.4 ± 14.4

EF (%)Group C 65.5 ± 11.9 65.2 ± 12.6 70.0 ± 12.2Group M 64.7 ± 16.0 60.6 ± 16.5 67.4 ± 17.5

Fig. 2. Comparison of aortic valve peak pressure gradient. Fig. 3. Comparison of left ventricle mass index.

no significant difference was observed between the twogroups (p=0.56) (Table 3, Fig. 4).

Actuarial freedom from cardiac eventIn Group C, heart failure was noted in 1 patient 1 monthsafter surgery and in another 17 months after; hemolyticanemia was observed in 1 patient 34 months after sur-gery. In Group M, 1 patient developed cardiogenic cere-bral infarction associated with paroxysmal atrial fibrilla-tion 41 months after surgery. The actuarial freedom fromcardiac event was 88.6% in Group C and 75.0% in GroupM 53 months after surgery. No significant difference wasnoted between the two groups (p=0.79) (Fig. 5).

Actuarial survival rateNo death was observed in Group C. In Group M, 1 pa-tient died as a result of cardiogenic cerebral infarction 42months after surgery, and another died from pneumonia43 months after surgery. The actuarial survival rate was

Effects of Aortic Valve Replacement

Ann Thorac Cardiovasc Surg Vol. 14, No. 2 (2008) 85

100% in Group C and 100% in Group M 24 months aftersurgery, but 100% in Group C and 50% in Group M 53months after surgery, with a significant difference betweenthe two groups (p=0.007) (Fig. 6).

Discussion

The minimum label size of bioprosthetic cardiac valvesavailable in Japan is 19 mm for both CEP and MM. Theyare used frequently for AVR with Japanese patients whoexhibit a small aortic annulus. However, since the smallorifice area of the 19-mm valve is very likely to cause thepersistence of a high-pressure gradient, the use of a largerprosthesis with an enlargement of the aortic root6) or astentless bioprosthesis7) has been suggested to be useful.Annular enlargement surgery has been reported by Nickset al.,8) Konno et al.,9) and Manouguian et al.,10) but theseprocedures have also been reported to be more invasive.11)

Each one requires slightly more complex surgical ma-nipulations, a longer use of a cardiopulmonary bypass,and a longer aortic clamp time. In consideration of theprogressive aging of candidates for these procedures, thesafety of valve replacement should be improved further.

In AVR, CEP has shown excellent durability and a lowincidence of complications over a period of 17 years sinceits introduction.12,13) MM is a new-generation bioprosthesisthat has been used since 1994 in Europe, and excellentdurability and a low incidence of complications have alsobeen reported as midterm results.14) However, Kirsch etal.15) reported that relatively high pressure gradients per-sisted after AVR using 19-mm MM, i.e., a peak pressuregradient of 40.8 ± 12.1 mmHg and a mean pressure gra-dient of 23.4 ± 6.9 mmHg. Takakura et al.16) reported that

the peak and mean pressure gradients after AVR using a19-mm CEP were 27.7 ± 9.5 mmHg and 12.3 ± 4.8mmHg, respectively, at rest and 50.4 ± 11.5 mmHg and22.2 ± 4.8 mmHg, respectively, even under dobutaminestress (10 µg/kg/min), and were acceptable. Eichinger etal.17,18) compared mean aortic valve pressure gradient us-ing 19-mm CEP and 19-mm MM and reported 11.4 ± 4.1mmHg for CEP and 14.6 ± 5.0 mmHg for MM at rest and20.7 ± 4.4 mmHg and 28.4 ± 9.0 mmHg, respectively,during 25 W exercise. Although the number of patientswas small and no statistical evaluation was performed,the results with CEP appear to be better. Marquez et al.19)

performed a hydrodynamic evaluation of variousbioprostheses, including CEP and MM in vitro, andshowed that the valve pressure gradient increases withthe cardiac output and that the mean effective valve ori-fice area was associated with the valve pressure gradient.Moreover, they reported that the valve pressure gradient

Fig. 4. Comparison of left ventricle ejection fraction.

Fig. 5. Actuarial freedom from cardiac event.

Fig. 6. Actuarial survival rate.

86

Tanigawa et al.

Ann Thorac Cardiovasc Surg Vol. 14, No. 2 (2008)

was lowest for CEP.Various opinions about the importance of PPM are still

being presented.20–23) According to Walther et al.24), whoevaluated 1,856 patients who received mechanical pros-theses and 2,275 who received bioprostheses, moderatePPM was noted in 26.7% of all patients, but its frequencywas higher in the bioprosthesis group, and it was an inde-pendent predictive factor for the short-term and long-termmortality rates. However, N.J. Howell et al.25) reportedthat even severe PPM with an indexed effective orificearea (EOA) < 0.6 cm2/m2 did not affect the hospital mor-tality rate or middle-term survival rate. If so, the use ofbioprostheses with a small orifice area should not neces-sarily be avoided.

A point that requires particular attention is that the la-bel sizes of both CEP and MM used in this study were 19mm, but the internal orifice diameter, which is a factorthat affects the EOA, was 18.0 mm in CEP and 16.5 mmin MM. Also, the external sewing ring diameter, which isconsidered to be the maximum diameter of abioprosthesis, was 26 mm in CEP and 25 mm in MM, sothat the 19-mm MM is considered to be smaller than 19-mm CEP (values are those announced by the manufac-turers).26) When the sizers provided by the manufacturerswere compared, the diameter of the sizer part insertedinto the annulus at suturing to the supra-annular positionwas 19.0 mm in the 19-mm CEP, but it was clearly smallerat 16.6 mm in the 19-mm MM (values measured usingslide calipers) (Fig. 7). Therefore 19-mm MM may beusable for a small aortic annulus, which the sizer of a 19-mm CEP does not fit.15,17)

In 19-mm MM, a peak pressure gradient of 45.4 ± 12.6mmHg remained early after surgery. The postoperativepressure gradient tended to be higher in Group M than inGroup C, but it showed improvements compared with thepreoperative values, and improvements in the LVMI andsubjective symptoms and a sufficiently acceptable inci-dence of complications were obtained. The 19-mm MMmay be underestimated in comparisons with otherbioprostheses because of its label size. Beyond the per-sistence of postoperative high -pressure gradient, 19-mmMM has a smaller profile and is capable of fitting a smalleraortic annulus.

Limitations of studyIn this study, the number of patients was small. And wedid not evaluate PPM because the EOA was not mea-sured in many patients by postoperative echocardiography.

And a direct comparison of the pressure gradient be-

tween before and after surgery was difficult because thesubjects included patients with AR, in whom the preop-erative aortic valve pressure gradient was not high. There-fore the preoperative pressure gradient in patients withAS and ASR and the postoperative pressure gradient ofall patients were compared as a reference.

Conclusions

Short-term and middle-term results of AVR using 19-mmCEP and 19-mm MM were compared. Although the post-operative aortic valve pressure gradient was higher andsurvival rate was lower in 19-mm MM than in 19-mmCEP, the improvement in the LVMI, maintenance of theLVEF, and incidence of cardiac event were acceptable inboth groups.

References

1. Khan SS, Trento A, DeRobertis M, Kass RM, SandhuM, et al. Twenty-year comparison of tissue and me-chanical valve replacement. J Thorac Cardiovasc Surg2001; 122: 257–69.

2. Rahimtoola SH. The problem of valve prosthesis-pa-

Fig. 7. The sizers provided by manufacturers.White arrows show the diameter of the sizer part inserted intothe annulus at suturing to the supra-annular position. The di-ameter in the 19-mm Medtronic Mosaic (MM) valve is clearlysmaller than in the 19-mm Carpentier-Edwards Perimount(CEP) valve.

Effects of Aortic Valve Replacement

Ann Thorac Cardiovasc Surg Vol. 14, No. 2 (2008) 87

tient mismatch. Circulation 1978; 58: 20–4.3. Pibarot P, Dumesnil JG, Lemieux M, Cartier P, Métras

J, et al. Impact of prosthesis-patient mismatch on he-modynamic and symptomatic status, morbidity andmortality after aortic valve replacement with abioprosthetic heart valve. J Heart Valve Dis 1998; 7:211–8.

4. Pibarot P, Dumesnil JG. Hemodynamic and clinicalimpact of prosthesis-patient mismatch in the aorticvalve position and its prevention. J Am Coll Cardiol2000; 36: 1131–41.

5. Guidelines from the Canadian Cardiovascular Societyand the Canadian Hypertension Society on theechocardiographic determination of left ventricularmass. Task Force of the Echocardiography Section. CanJ Cardiol 1995; 11: 391–5.

6. Castro LJ, Arcidi JM Jr, Fisher AL, Gaudiani VA. Rou-tine enlargement of the small aortic root: a preventivestrategy to minimize mismatch. Ann Thorac Surg 2002;74: 31–6.

7. Borger MA, Carson SM, Ivanov J, Rao V, Scully HE,et al. Stentless aortic valves are hemodynamically su-perior to stented valves during mid-term follow-up: alarge retrospective study. Ann Thorac Surg 2005; 80:2180–5.

8. Nicks R, Cartmill T, Bernstein L. Hypoplasia of theaortic root. The problem of aortic valve replacement.Thorax 1970; 25: 339–46.

9. Konno S, Imai Y, Iida Y, Nakajima M, Tatsuno K. Anew method for prosthetic valve replacement in con-genital aortic stenosis associated with hypoplasia ofthe aortic valve ring. J Thorac Cardiovasc Surg 1975;70: 909–17.

10. Manouguian S, Seybold-Epting W. Patch enlargementof the aortic valve ring by extending the aortic incisioninto the anterior mitral leaflet. New operative technique.J Thorac Cardiovasc Surg 1979; 78: 402–12.

11. Sommers KE, David TE. Aortic valve replacement withpatch enlargement of the aortic annulus. Ann ThoracSurg 1997; 63: 1608–12.

12. Dellgren G, David TE, Raanani E, Armstrong S, IvanovJ, et al. Late hemodynamic and clinical outcomes ofaortic valve replacement with the Carpentier-EdwardsPerimount pericardial bioprosthesis. J ThoracCardiovasc Surg 2002; 124: 146–54.

13. Banbury MK, Cosgrove DM 3rd, Thomas JD,Blackstone EH, Rajeswaran J, et al. Hemodynamic sta-bility during 17 years of the Carpentier-Edwards aor-tic pericardial bioprosthesis. Ann Thorac Surg 2002;73: 1460–5.

14. Eichinger WB, Botzenhardt F, Gunzinger R, KemkesBM, Sosnowski A, et al. European experience with theMosaic bioprosthesis. J Thorac Cardiovasc Surg 2002;124: 333–9.

15. Kirsch ME, Tzvetkov B, Vermes E, Pouzet B, SauvatS, et al. Clinical and hemodynamic performance of the19-mm medtronic mosaic bioprosthesis. J Heart ValveDis 2005; 14: 433–9.

16. Takakura H, Sasaki T, Hashimoto K, Hachiya T,Onoguchi K, et al. Hemodynamic evaluation of 19-mm Carpentier-Edwards pericardial bioprosthesis inaortic position. Ann Thorac Surg 2001; 71: 609–13.

17. Eichinger WB, Botzenhardt F, Guenzinger R, BleizifferS, Keithahn A, et al. The effective orifice area/patientaortic annulus area ratio: a better way to compare dif-ferent bioprostheses? A prospective randomized com-parison of the Mosaic and Perimount bioprostheses inthe aortic position. J Heart Valve Dis 2004; 13: 382–8.

18. Eichinger WB, Botzenhardt F, Keithahn A, GuenzingerR, Bleiziffer S, et al. Exercise hemodynamics of bo-vine versus porcine bioprostheses: a prospective ran-domized comparison of the mosaic and perimount aor-tic valves. J Thorac Cardiovasc Surg 2005; 129: 1056–63.

19. Marquez S, Hon RT, Yoganathan AP. Comparative hy-drodynamic evaluation of bioprosthetic heart valves. JHeart Valve Dis 2001; 10: 802–11.

20. Ruel M, Al-Faleh H, Kulik A, Chan KL, Mesana TG,et al. Prosthesis-patient mismatch after aortic valvereplacement predominantly affects patients with pre-existing left ventricular dysfunction: effect on survival,freedom from heart failure, and left ventricular massregression. J Thorac Cardiovasc Surg 2006; 131: 1036–44.

21. Tasca G, Brunelli F, Cirillo M, DallaTomba M, MhagnaZ, et al. Impact of valve prosthesis-patient mismatchon left ventricular mass regression following aorticvalve replacement. Ann Thorac Surg 2005; 79: 505–10.

22. Flameng W, Meuris B, Herijgers P, Herregods MC.Prosthesis-patient mismatch is not clinically relevantin aortic valve replacement using the Carpentier-Edwards Perimount valve. Ann Thorac Surg 2006; 82:530–6.

23. Takai H, Yamachika S, Hazama S, Ariyoshi T, OdateT, et al. Consideration of prosthesis-patient mismatchand left ventricular mass regression after implantationof the Carpentier-Edwards pericardial valve in elderlyJapanese patients: body surface area may be irrelevant.J Artif Organs 2004; 7: 203–6.

24. Walther T, Rastan A, Falk V, Lehmann S, Garbade J, etal. Patient prosthesis mismatch affects short- and long-term outcomes after aortic valve replacement. Eur JCardiothorac Surg 2006; 30: 15–9.

25. Howell NJ, Keogh BE, Barnet V, Bonser RS, GrahamTR, et al. Patient-prosthesis mismatch does not affectsurvival following aortic valve replacement. Eur JCardiothorac Surg 2006; 30: 10–4.

26. Hashimoto K. Patient-prosthesis mismatch: the Japa-nese experience. Ann Thorac Cardiovasc Surg 2006;12: 159–65.