Aortic Valve Repair of Congenital Stenosis WithBovine PericardiumMichael J. Tolan, FRCS(I), Piers E. Daubeney, MB, BS, Zdenek Slavik, MD,Barry R. Keeton, FRCP, Anthony P. Salmon, FRCP, and James L. Monro, FRCSWessex Cardiothoracic Centre, General Hospital Southampton, Southampton, England

Background. Conservative surgical options in the treat-ment of congenital aortic stenosis are limited. To relievethe obstruction necessitates full incision of the raphe ofthe larger valve leaflet, but this inevitably causes pro-lapse.Methods. We performed aortic valve repair in 6 chil-

dren, aged 14 months to 17 years, with congenital aorticstenosis, 2 having had aortic valvotomy as infants. Therepair consisted of suturing the base of a triangular pieceof bovine pericardium, with a simple vertical fold, to thefree edges of the incised raphe. The pericardial fold wasthen sutured vertically to the aortic wall.Results. At follow-up of 2 to 60 months, the mean peak

systolic Doppler gradients had decreased from 80 6

15 mm Hg to 26 6 9 mm Hg. The effective valvular orificearea increased from 33% 6 6% to 64% 6 3%, allowingblood flow to increase by a factor of 3.76. Two patientshave mild and 2 have mild-to-moderate aortic regurgita-tion.Conclusions. The described conservative repair renders

the valve tricuspid and trisinusoidal, and the deficientinterleaflet triangle is recreated, preventing cusp pro-lapse. Longer follow-up is required to assess the durabil-ity of unstented pericardium in the aortic position, butthe early results are encouraging.

(Ann Thorac Surg 1997;63:465–9)© 1997 by The Society of Thoracic Surgeons

The optimal surgical treatment of children with con-genital aortic stenosis remains unresolved in many

instances [1, 2]. It is generally accepted that cardiac valverepair is preferable to replacement [3], although recentliterature concerning aortic valve replacement with pul-monary autografts is impressive [4–6]. Some authorsrecommend pulmonary autograft replacement of theaortic valve in the presence of a nontricuspid aortic valveand after failed surgical valvuloplasty [7]. The aorticvalve in congenital bicuspid stenosis has unique patho-logic features [8], which must be considered in anyattempt at repair. In a study of bicuspid valves [9], thecircumferential amplitude of the smaller leaflet rangedbetween 120 and 180 degrees, one leaflet being larger in85% of valves. A raphe, or as it is otherwise called therudimentary commissure, was present in 84%, defined asa fibrous ridge at a right angle to the circumference of theleaflet dividing the leaflet into two component parts.Incising the raphe, recommended by some, provides alarge central opening and a better blood flow pattern butis usually associated with severe regurgitation. The shal-low, unsupported cusps tend to prolapse into the ven-tricular cavity, and there is poor coaptation with theopposite cusp. With time, lengthening and redundancyof the valve occurs, leading to prolapse and furtherinsufficiency not only at the commissure but alsothroughout the whole valve.It has been suggested that the sinuses of Valsalva play

a more active part in the function of the valve than waspreviously thought [10]. In 57 of 64 valves studied withtwo leaflets, examination of the ventricular aspect re-vealed evidence of three sinuses and three interleaflettriangles [9], one of the triangles being deficient in height.Also, a tricuspid valve morphology is preferable as itprovides a more effective central flow opening for a givenring size than is possible with a bicuspid valve.We describe a technique of valvuloplasty that involves

incision of the raphe and suspension of the cusps withbovine pericardium. This renders the valve tricuspid andtrisinusoidal, and the deficient interleaflet triangle isrecreated, resulting in a greatly increased cross-sectionalarea of the valve orifice.

Patients and Methods

PatientsBetween 1991 and 1996, six children underwent an aorticvalvuloplasty using bovine pericardium. The median ageof the children was 5.9 years (range, 14 months to 17years), and the mean weight was 20.0 kg (range, 11.4 to71 kg). An aortic systolic murmur was detected in allpatients during the first week of life. Two patients had aprevious aortic valvotomy. One of these had mild aorticvalve stenosis at birth (gradient, 20 mm Hg) associatedwith moderately severe coarctation of the aorta (gradient,40 mm Hg), multiple ventricular septal defects, andpatent ductus arteriosus. He required a subclavian flaprepair of the coarctation of the aorta, banding of thepulmonary artery for multiple ventricular septal defects,and ligation of the patent ductus arteriosus in infancy,

Accepted for publication Sep 7, 1996.

Address reprint requests to Mr Monro, Wessex Cardiothoracic Centre,University Hospital, Tremona Rd, Southampton, England.

© 1997 by The Society of Thoracic Surgeons 0003-4975/97/$17.00Published by Elsevier Science Inc PII S0003-4975(96)01231-3

and debanding of the pulmonary artery with conven-tional aortic valvotomy at 16 months. The other patienthad critical aortic stenosis associated with severe mitralregurgitation requiring aortic valvotomy at 3 weeks oflife.All 6 patients were followed up in the ensuing years

using two-dimensional echocardiography with Dopplerstudies. The gradient across the aortic valve was esti-mated by measuring peak systolic velocity from Dopplerstudies, and aortic regurgitation was assessed. Threepatients had cardiac catheterization before the aorticvalvotomy with bovine pericardium suspension.Indications for operation were transaortic pressure

gradients of 80 to 96 mm Hg (mean, 85 mm Hg) in 4patients, 3 of whom were asymptomatic (2 had previousvalvotomy 3.6 and 5.8 years previously); two episodes ofsyncope in 1 patient whose gradient was 65 mm Hg; andsymptoms of breathlessness in 1 patient who had agradient of 60 mm Hg. There was electrocardiographicevidence of left ventricular strain in 2 patients. Fourpatients had aortic regurgitation preoperatively. Appear-ances of the valves suggested bicuspid morphology in allcases.

Surgical TechniqueModerate hypothermia with core cooling to 25° to 28°C,single or bicaval venous cannulation, and a left ventric-

ular vent were used. Myocardial protection consisted oftopical hypothermia and hyperkalemic cardioplegic ar-rest. If significant aortic regurgitation was present, thecardioplegic solution was directly infused into the coro-nary ostia after the aortic root was opened. Repeateddoses of cardioplegia were given at intervals throughoutthe operation if required.The ascending aorta was incised and retracted, and the

valve area was inspected. In all our patients bicuspidmorphology was confirmed (Fig 1A). Any cuspal fusionwas incised in a conventional manner. The raphe be-tween the other nonsupported cusp was then incised asfar as the aortic wall (Fig 1B). A measurement of thelength of this raphe was made, and note was taken of theheight of the sinutubular ridge. An isosceles triangle ofbovine pericardium was cut with dimensions such thatthe base was twice the length of the incised raphe and thevertical height was slightly greater than the height of thesinutubular ridge (Fig 1C). The pericardium was foldedalong its vertical axis and oriented in the aorta such thatthe two edges of the base were sutured to the two freeedges of the incised raphe, and the fold in the pericar-dium was sutured vertically to the aortic wall to a levelabove the sinutubular ridge (see Fig 1C). Thus the valvewas rendered tricuspid and trisinusoidal, and the defi-cient interleaflet triangle was recreated, preventing cuspprolapse. One patient initially underwent a commissur-

Fig 1. (A) A view of congenital bi-cuspid aortic valve showing onelarger cusp with a median raphe. (B)Incision of the median raphe. (C) Apericardial triangle (inset) is foldedand sutured along both edges of thedivided raphe and vertically to theaortic wall to provide support in di-astole.

466 TOLAN ET AL Ann Thorac SurgAORTIC VALVE REPAIR IN CHILDREN 1997;63:465–9

otomy of about 5 mm between the right and noncoronarycusps. There was no fusion between the left and non-coronary cusps. After cardiopulmonary bypass was dis-continued, the aortic gradient was still measured at80 mm Hg. Cardiopulmonary bypass was therefore rein-stituted and repair of the valve effected using the tech-nique described with a triangular patch of pericardium.A further patient, aged 17 years, had extensive decalcifi-cation of the annulus and thinning of the larger cuspbefore repair with the calf pericardium. One patient alsohad an associated mitral valve repair by chordal short-ening and annuloplasty.Intraoperative transesophageal echocardiography was

used to validate early competence of the repaired valve.Postoperative gradients were determined by Dopplerechocardiography.

Results

Follow-up has ranged from 2 to 60 months (median, 34months). There was no mortality or morbidity associatedwith the operation. All operations were of short duration:the maximum cardiopulmonary bypass time was 75 min-utes in the patient who had concomitant mitral valverepair. Intraoperative echocardiography demonstratedsuccessful initial repair in all patients, with good coapta-tion of cusps and no further modification required. Mea-sured gradients were much reduced from preoperativelevels (Table 1). All patients were discharged from thehospital within 8 days, and to date at follow-up all 6remain symptom free. Peak systolic Doppler gradientsdecreased significantly from 80 6 15 mm Hg preopera-tively to 26 6 9 mm Hg postoperatively. Mild aorticregurgitation was described in 4 patients preoperativelyand 2 patients immediately postoperatively (see Table 1).

Serial planimetric evaluation using the largest area dur-ing systole from two-dimensional echocardiography al-lowed estimation of the effective valvular orifice. Thiswas almost doubled compared with the preoperativevalues, an increase from 33% 6 6% of total aortic areapreoperatively to 64% 6 3% postoperatively, thus allow-ing blood flow to increase 3.76 times that of the preoper-ative state.Preoperative echocardiographic pictures in diastole

(Fig 2A) and systole (Fig 2B) demonstrate the bicuspidvalve with the large leaflet inhibiting wide opening of thevalve. Postoperatively, the now tricuspid morphology ofthe valve in diastole can be seen (Fig 3A), and in systolethere is a much greater valvular orifice area (Fig 3B).Electrocardiographic evidence of left ventricular strainreverted in the 2 patients in whom it was present preop-eratively. Late follow-up echocardiography and Dopplerstudies showed mobile cusps with the pericardial exten-sion maintaining its flexibility, resulting in a widelypatent orifice in systole (see Fig 3A), and a tricuspidconfiguration in closure in diastole (see Fig 3B).Measurement of gradients demonstrated no increase

in stenosis over the period of follow-up. However, therehas been some increase in the amount of valve regurgi-tation over the years postoperatively (see Table 1).

Comment

Attempts to preserve the aortic valve with congenitalstenosis have included thinning of valve leaflets, aug-mentation of scarred and retracted leaflets with autolo-gous pericardium, incision of the raphe with release ofthe rudimentary commissure from the aortic wall, andextension of the valvotomy incision into the aortic walland on both sides of the commissure. Using a combina-

Table 1. Patient Characteristics

Variable

Patient Number

1 2 3 4 5 6

Age at operation (y) 3.8 9.5 5.9 10.8 1.2 17Weight at operation (kg) 13.0 33.0 20.0 34.5 11.8 71Diagnosis AS/AR AS/AR AS/AR/MR AS AS AS/ARSymptoms 2 2 2 1 2 1

Previous aortic operation 1 2 1 2 2 2

Follow-up (mo) 60 40 34 30 5 2EchocardiographyGradient (mm Hg)Preop 96 80 80 65 80 65Postop 30 30 25 15 80/57a 30Current 20 30 25 25 35 30

RegurgitationPreop Mild Mild Severe Mild Absent ModPostop Absent Absent Mild Mild Absent AbsentCurrent Mild/mod Mild Mild Mild/mod Trivial Absent

a The first gradient of 80 mm Hg was after an initial conventional commisurotomy; the second gradient of 57 was after the valve was repaired withpericardium.

AR 5 aortic regurgitation; AS 5 aortic stenosis; Mod 5 moderate; MR 5 mitral regurgitation.

467Ann Thorac Surg TOLAN ET AL1997;63:465–9 AORTIC VALVE REPAIR IN CHILDREN

tion of techniques, Ankeney and associates [2] achieved a92.5% chance of avoiding reoperation for 10 years, but by15 years the chance fell to 73%. In their series themajority of patients were slightly older, between 11 and15 years. Van Son and associates [8] reviewed the cases of22 patients, with a median age of 5.7 years, who under-went aortic valvuloplasty; 5 required aortic valve replace-ment within the first week and a further 8 had eitherstenosis with a gradient of 40 to 60 mm Hg or severeregurgitation at a median follow-up of 16 months. Abicuspid morphology was present in 10 of the valvulo-plasty failures. Caspi and colleagues [11] described atechnique of extended aortic valvuloplasty as the initialapproach for management of congenital aortic stenosis.Incision of the fused commissures to the aortic wall andthen extension in a circumferential fashion in both direc-tions for 1 to 2 mm gave better relief of obstruction andless residual regurgitation than with standard valvulot-omy. The same procedure was applied in children withsignificant recurrent aortic valve stenosis with good re-sults at short-term follow-up. After use of their tech-nique, however, the valve remains bicuspid and thus

loses the advantages of the tricuspid morphology. Ourtechnique, by creating a tricuspid valve, achieves almosta doubling of the effective orifice area, which results in anexponential increase in blood flow. It is particularlyappropriate when there is very little fusion of the bicus-pid valve and therefore no commissurotomy is possible.Incision of the raphe of the fused larger leaflet allows thecusps to fold back onto the aortic wall in systole, thusallowing a much larger central orifice. In diastole they areprevented from prolapsing by the supporting pericardialinsert.Our technique also attempts to restore the trisinusoi-

dal morphology of the valve. Leonardo da Vinci firstdemonstrated the importance of the trisinusoidal config-uration and formation of vortices on preventing shearstress on the leaflets during valve closure [12]. Eddyformation behind the leaflets prevents their impaction onthe aortic wall on opening, and in diastole, by theoutward movement of the aortic walls, the sinuses as-sume an almost spheric shape and thus allow much ofthe load on the leaflets to be taken up by the sinus walls.

Fig 2. Preoperative echocardiography: (A) the aortic valve in dias-tole showing the bicuspid morphology and (B) the aortic valve insystole showing a restricted valve orifice.

Fig 3. Postoperative echocardiography: (A) the aortic valve in dia-stole showing the tricuspid closure lines of the reconstructed valveand (B) the aortic valve in systole demonstrating the widely patentvalve orifice of the reconstructed valve.

468 TOLAN ET AL Ann Thorac SurgAORTIC VALVE REPAIR IN CHILDREN 1997;63:465–9

This stress-sharing decreases the stress and the wear onthe leaflets, which is of major import when using un-stented pericardium in the aortic position.Pericardium has long been used as a valve substitute

and for valve repair [13]. A recent experience of un-stented porcine pericardium for valve replacement in theaortic position has been described by David and associ-ates [14], who reported on 123 patients with 6-yearsurvival at 91% and very uncommon valve related com-plications. Duran and colleagues [15] have reported fa-vorable results for cusp extension using glutaraldehyde-treated bovine or autologous pericardium, althoughfollow-up was just 30 months. Likewise, Al-Fagih andassociates [16] reported single cusp extension using bo-vine pericardium with excellent results but only 23months’ maximum follow-up. Batista and associates [17]reported the replacement of the aortic valve with amonopatch of bovine pericardium, but again follow-upwas short. The natural history of free pericardial patchesseems to be one of eventual thickening, fibrosis, andretraction. Under high pressure they may stretch andbecome aneurysmal. However, calcification eventuallyoccurs in glutaraldehyde-treated bovine tissue. Furtherfollow-up studies will determine the eventual role ofpericardium in valve reconstruction. Meanwhile, resultsjustify its continued use, and it has been encouraging thatin our series there has been no significant restenosis. Theslight increase in the amount of regurgitation in 3 of ourpatients may be due to some shortening of the pericardialtissue and is being kept under constant review.In conclusion, we have described a technique that

allows marked enlargement of a bicuspid valve withresuspension of the divided raphe to avoid prolapse byuse of a triangular bovine pericardial patch. Althoughour series is small and the follow-up is only to 5 years,this technique does provide a method of enlarging abicuspid valve orifice even when there is no commissuralfusion.

References

1. Wheller JJ, Hosier DM, Teske DW, Kilman JW. Results ofoperation for aortic valve stenosis in infants, children, andadolescents. J Thorac Cardiovasc Surg 1988;96:474–7.

2. Ankeney JL, Tzeng TS, Liebman J. Surgical therapy forcongenital aortic valvular stenosis. J Thorac Cardiovasc Surg1983;85:41–8.

3. Williams WG, Pollock JC, Geiss DM, Trusler GA, Fowler RS.Experience with aortic and mitral valve replacement inchildren. J Thorac Cardiovasc Surg 1981;81:326–33.

4. Elkins RC, Santangelo K, Randolph JD, et al. Pulmonaryautograft replacement in children. Ann Surg 1992;216:363–71.

5. Gerosa G, McKay R, Davies J, Ross DN. Comparison of theaortic homograft and the pulmonary autograft for aorticvalve or root replacement in children. J Thorac CardiovascSurg 1991;102:51–61.

6. Kouchoukos NT, Davilla-Roman VG, Spray TL, Murphy SF.Replacement of the aortic root with a pulmonary autograft inchildren and young adults with aortic valve disease. N EnglJ Med 1994;330:59–60.

7. Van Son JA, Reddy VM, Black MD, Rajasinghe H, Haas GS,Hanley FL. Morphologic determinants favoring surgical aor-tic valvuloplasty versus pulmonary autograft aortic valvereplacement in children. J Thorac Cardiovasc Surg 1996;111:1149–57.

8. McKay R, Smith A, Anderson RH. Morphology of the ven-triculoaortic junction in critical aortic stenosis. J ThoracCardiovasc Surg 1992;104:433–42.

9. Angelini A, Ho SY, Anderson RH, Becker AE. The morphol-ogy of the normal aortic valve as compared with the aorticvalve having two leaflets. J Thorac Cardiovasc Surg 1989;98:362–7.

10. Kunzelman KS, Grande J, David TE, Cochran RP. Aortic rootand valve relationships. J Thorac Cardiovasc Surg 1994;107:162–70.

11. Caspi J, Ilbawi MN, Roberson DA, Najafi H. Extended aorticvalvuloplasty for recurrent valvular stenosis and regurgita-tion in children. J Thorac Cardiovasc Surg 1994;107:1114–20.

12. Robicesek F. Leonardo da Vinci and the sinuses of Valsalva.Ann Thorac Surg 1991;52:328–35.

13. Duran CMG. Pericardium in valve operations. Ann ThoracSurg 1993;56:1–2.

14. David TE, Feindel CM, Bos J, Rakowski H. Aortic valvereplacement with stentless porcine aortic valves: a six yearexperience. J Heart Valve Dis 1992;1:244–8.

15. Duran C, Kumar N, Gometz B, Halees ZA. Indications andlimitations of aortic valve reconstruction. Ann Thorac Surg1991;52:447–54.

16. Al-Fagih M, Kasab SM, Ashmeg A. Aortic valve repair usingbovine pericardium for cusp extension. J Thorac CardiovascSurg 1988;96:760–4.

17. Batista RJV, Dobrianskij A, Comazzi M, Moreira R, OliveiraPF. Clinical experience with stentless pericardial aorticmonopatch for aortic valve replacement. J Thorac Cardio-vasc Surg 1987;93:19–26.

469Ann Thorac Surg TOLAN ET AL1997;63:465–9 AORTIC VALVE REPAIR IN CHILDREN