Does Banding the Pulmonary Artery AffectPulmonary Valve Function After theDamus-Kaye-Stansel Operation?Zahid Amin, MD, Carl L. Backer, MD, C. Elise Duffy, MD, andConstantine Mavroudis, MDDivisions of Cardiovascular-Thoracic Surgery and Cardiology, Children’s Memorial Medical Center and Departments of Surgeryand Pediatrics, Northwestern University Medical School, Chicago, Illinois

Background. The Damus-Kaye-Stansel (DKS) opera-tion can be an effective palliation in patients who havesingle-ventricle physiology and systemic outflow ob-struction. Pulmonary artery banding (PAB) may be usedas a preliminary procedure in these patients to limitoverperfusion of the pulmonary circulation. In someseries, the DKS operation has been associated withpulmonary insufficiency (PI). We retrospectively ana-lyzed medical records of our patients who had PAB andlater DKS to determine the incidence of PI in thesepatients.

Methods. Between 1982 and 1996, 15 patients under-went PAB before DKS. Median age at PAB placementwas 7 days and median duration of PAB was 7 months.

Echocardiograms obtained before PAB, before DKS, andat the most recent post-DKS follow-up were reviewed.

Results. Follow-up ranged from 1 to 15 years (meanfollow-up, 7.5 years). One patient had trivial PI beforePAB, which progressed to moderate PI at the last follow-up. Only 1 other patient had mild PI, but only at the lastfollow-up after DKS.

Conclusions. These findings suggest that prior PABdoes not appear to cause significant PI in patients slatedfor DKS, and the incidence of significant PI after theDKS operation is relatively low.

(Ann Thorac Surg 1998;66:836–41)© 1998 by The Society of Thoracic Surgeons

The Damus-Kaye-Stansel (DKS) operation was ini-tially described for repair of transposition of the

great arteries (TGA) to avoid coronary artery transloca-tion [1–3]. Since the original description the spectrum ofanomalies for which the DKS operation is used hasexpanded to other complex univentricular or biventricu-lar cardiac lesions, such as double-inlet left ventricle,tricuspid atresia with TGA {S,D,D}, Taussig-Bing malfor-mation, and some patients with aortic arch interruptionwith a malalignment-type ventricular septal defect (VSD)[4–7]. The DKS operation has been associated with a highmortality in the neonatal period in some series [4, 5, 8, 9].Pulmonary artery banding (PAB) is an attractive option inpatients with overperfusion of the pulmonary circulationbut it does seem to have some drawbacks. Subaorticstenosis or bulboventricular foramen narrowing occursnaturally and sometimes secondary to ventricular hyper-trophy from PAB [10–13] or a change in the ventriculargeometry [14]. There is always a chance of damaging thepulmonary valve with an improper PAB, thus decreasingthe availability of the valve for future operations should aneed arise to incorporate the valve in the systemiccirculation [15]. There are some reports that suggest

abandoning PAB in favor of a more aggressive Norwood-type initial procedure because of complications attrib-uted to PAB [9, 11, 16]. Some authors have reported thatpulmonary regurgitation (PR) is common after the DKSoperation; this has been attributed to distortion of thevalve or prior PAB [5, 8, 16, 17]. Other authors claim thatthe incidence of PR after incorporating the pulmonaryvalve in the systemic circulation is low [18]. The purposeof this report was to review the medical records of ourpatients and determine whether prior PAB increases therisk of PR in patients who have subsequently undergonethe DKS operation.

Material and Methods

PatientsFrom 1982 through 1996, 15 patients had a DKS operationafter initial PAB. All patients who had the DKS operationas the primary procedure were excluded. The medicalrecords of these patients were reviewed. Particular atten-tion was paid to preoperative echocardiograms, opera-tive procedure, and postoperative echocardiogram atthe most recent follow-up. There were 10 male and 5female patients. The diagnosis included double-inlet leftventricle (n 5 6), TGA with VSD (n 5 4), Taussig-Binganomaly (n 5 2), tricuspid atresia (n 5 2), and correctedTGA (n 5 1). All associated cardiac anomalies are out-lined in Table 1.

Presented at the Thirty-fourth Annual Meeting of The Society of ThoracicSurgeons, New Orleans, LA, Jan 26–28, 1998.

Address reprint requests to Dr Backer, Division of Cardiovascular Sur-gery, Children’s Memorial Medical Center, 2300 Children’s Plaza, Box 22,Chicago, IL 60614 (e-mail: c-backer@nwu.edu).

© 1998 by The Society of Thoracic Surgeons 0003-4975/98/$19.00Published by Elsevier Science Inc PII S0003-4975(98)00608-0

The median age at the time of placement of PAB was 7days (range, 1 day to 4 months). The median duration forwhich the band was placed was 7 months (range, 4 to 84months). All associated procedures are outlined in Table1. The decision to take down the PAB was based onmultiple factors: increasing cyanosis, increase in bandpressure gradient, age appropriate for Fontan, and in-crease in pressure gradient across subaortic area. Theband pressure gradient before taking down the band wasalso measured.

EchocardiographyEchocardiograms obtained before the PAB, before theDKS operation, and at the most recent follow-up afterDKS were reviewed. The pulmonary valve, the PABpressure gradient, the main pulmonary artery to aortaanastomosis, and the subpulmonary area were analyzedfrom subcostal and parasternal windows. Pulmonaryregurgitation was classified as mild if the jet widthdivided by the width of the immediate subpulmonaryvalve region was less than 0.4, moderate if between 0.4and 0.6, and severe if more than 0.6 [18]. The subaorticpressure gradient was obtained from the first echocar-diogram and before taking down the PAB. It was definedas mild (,20 mm Hg), moderate (20 to 40 mm Hg), orsevere (.40 mm Hg).

Operative TechniquePULMONARY ARTERY BANDING. The PAB was performed ei-ther through a median sternotomy (n 5 5) or left thora-cotomy (n 5 10) concomitant with repair of coarctation(n 5 9) or interruption of the aortic arch (n 5 1). The band

was placed as far away from the pulmonary valve aspossible without jeopardizing the branch pulmonaryarteries (PA) (Fig 1). The target distal PA pressure washalf the systemic pressure with an increase in meanarterial pressure of 15 mm Hg. The band was anchored tothe pulmonary trunk with two Prolene (Ethicon, Somer-ville, NJ) mattress sutures. All bandings were performedwith a 4 mm Teflon-impregnated Dacron (n 5 11) or, inour earlier experience, Mersiline (n 5 4) band.

DAMUS-KAYE-STANSEL OPERATION. The DKS operation wasperformed using bicaval hypothermic cardiopulmonarybypass in 13 patients and circulatory arrest in 2 patients.The ascending aorta was cannulated close to the innom-inate artery. The aorta was cross-clamped immediatelyproximal to the innominate artery to provide perfusion tothe arch vessels during the repair. Cardioplegic solutionwas introduced through the aortic root. Topical myocar-dial cooling was also used. The main PA was transectedat the site of the PAB. The initial incision on the medialaspect of the aorta was begun just above the commissuralposts of the aortic valve. The approximate length of theaortic opening was equal to the diameter of the pulmo-nary trunk. It is critical to avoid distortion of the proximalpulmonary trunk and aorta as this will result in pulmo-nary valve regurgitation. In 14 of 15 patients, the DKSoperation was facilitated by the use of a prosthetic patch.A triangular-shaped patch of expanded polytetrafluoro-ethylene (W. L. Gore, Flagstaff, AZ) was used in 8patients, a Hemashield patch (Hemashield microvel dou-ble velour graft; Meadox Medicals, Inc, Oakland, NJ) in 4patients, and a pulmonary homograft in 2 patients was

Table 1. Patient Population

PatientNo. Diagnosis

AssociatedDiagnosis

Age at PAB(mo)

Age at DKS(mo)

ConcomitantProcedure Additional Procedures

1 DILV/TGA CoA 0.1 7 SP shunt Bi-di Glenn, Fontan2 DORV/TB CoA 0.1 6 Conduit Conduit replacement, closure

aortic valve3 DILV/TGA CoA 0.23 12 SP shunt Modified Fontan4 DILV/TGA 4 88 Fontan Extracardiac Fontan5 CTGA CoA 0.23 12 Glenn Died after 2 mo6 TGA/VSD CoA 0.26 7 Conduit Conduit replacement7 DORV/TB IAA 0.23 7 Conduit Conduit replacement8 DILV/TGA CoA 0.2 6 SP shunt Glenn shunt and Fontan9 TGA/MVSD DC 0.7 34 Conduit Conduit replacement, closure

aortic valve10 TGA/VSD CoA 0.16 10 Conduit Conduit replacement, closure

aortic valve11 TGA/MVSD VSD 0.36 7 Conduit Replacement of conduit12 DILV/TGA 2 7 SP shunt Glenn, modified Fontan13 TGA/VSD SCA 0.23 11 Conduit Conduit replacement14 DILV/TGA CoA 0.23 9 SP shunt Bi-di Glenn, Fontan15 TA/TGA CoA 0.23 7 SP shunt Glenn, Fontan

Bi-di Glenn 5 bidirectional Glenn; CoA 5 coarctation of aorta; CTGA 5 corrected transposition of the great arteries; DC 5 dextrocardia;DILV 5 double-inlet left ventricle; DKS 5 Damus-Kaye-Stansel operation; DORV 5 double-outlet right ventricle; IAA 5 interrupted aorticarch; MVSD 5 muscular ventricular septal defects; PAB 5 pulmonary artery banding; SCA 5 single coronary artery; SP shunt 5systemic-to-pulmonary artery shunt; TA 5 tricuspid atresia; TB 5 Taussig-Bing anomaly; TGA 5 transposition of the great arteries;VSD 5 ventricular septal defect.

837Ann Thorac Surg AMIN ET AL1998;66:836–41 DOES PAB BEFORE DKS CAUSE PI?

used as a hood for the DKS procedure (Fig 2A). The hoodtechnique facilitates construction of a tension-free anas-tomosis with no distortion of the pulmonary valve orcoronary arteries. In only 1 patient was a direct main PAto ascending aorta anastomosis constructed. This maydistort the pulmonary trunk, cause tension on the pul-monary valve, and result in pulmonary valve regurgita-tion (Fig 2B).

Results

There was no operative mortality during either PAB orDKS operation. There was one late death with resultantoverall mortality of 6.7% (95% confidence interval [CI],0.017 to 0.32). This patient had corrected transpositionwith subaortic stenosis and had a prior systemic to PAshunt in addition to a Glenn procedure because of lowoxygen saturations. He died 2 months after the DKSoperation of unrecognized conduit obstruction.

The Kaplan-Meier estimate of freedom from pulmo-nary insufficiency after PAB and later DKS is shown inFig 3. The prevalence of pulmonary insufficiency as of thelast follow-up is 13% (95% CI, 0.017 to 0.405). On theinitial pre-PAB echocardiograms, trivial PR was presentin 1 patient. The PR progressed to mild by echocardio-gram obtained before taking down the PAB in the same

patient and to moderate PR at the most recent follow-up.This patient had the PAB placed for 84 months and wastemporarily lost to follow-up primarily because of theparent’s reluctance to permit an operation. Mild PR haddeveloped in 1 other patient at the last follow-up. Thispatient had his PAB placed at another institution. At thetime of band take down, the band was found abutting thepulmonary valve leaflets, causing mild adhesions (whichhad to be separated) between the leaflets. There was noPR in the remaining patients at the last follow-up byechocardiogram or by physical exam. The PAB pressuregradient before the take down ranged from 35 mm Hg to100 mm Hg (mean, 47 mm Hg).

Interestingly, 4 patients in our series had aortic valveregurgitation, 1 patient with Taussig-Bing anomaly, 2with TGA, and 1 patient with double-inlet left ventricle.

Fig 1. (A) Placement of the pulmonary artery band away from thepulmonary valve to avoid pulmonary valve distortion. (B) Proximalplacement of the pulmonary artery band will distort the pulmonaryvalve. (© 1998 by Rachid F. Idriss. Reprinted with permission.)

Fig 2. (A) The Damus-Kaye-Stansel operation constructed with thehood technique keeps the pulmonary trunk free of distortion andtension on the valve. (B) The Damus-Kaye-Stansel operation con-structed without the hood may cause tension on the valve and dis-tortion of the trunk and the valve with resultant valve regurgitation.(© 1998 by Rachid F. Idriss. Reprinted with permission.)

838 AMIN ET AL Ann Thorac SurgDOES PAB BEFORE DKS CAUSE PI? 1998;66:836–41

Three of them required closure of the aortic valve. Theremaining patient has mild aortic regurgitation and re-mains stable. Four patients in this series had complete heartblock, 2 during the initial DKS and 2 after revision of aFontan operation. The incidence of atrial arrhythmia aftercompletion to Fontan in this group was almost 40% (3 of 8).Two patients have undergone Fontan conversion to extra-cardiac Fontan, with intraoperative cryoablation guided byelectrophysiologic mapping of the tachycardia circuit.

A mild VSD or bulboventricular pressure gradient waspresent in 2 patients before the PAB and developed in 2patients subsequent to the PAB placement. Of the 2patients who had a VSD pressure gradient after the PAB,one had PAB placement at an outside institution. At thetime of DKS operation the PAB pressure gradient was 80to 100 mm Hg, and the VSD pressure gradient was 35 mmHg. The other patient had the band placed for 84 months,his PAB pressure gradient was 90 mm Hg, and the VSDpressure gradient was 20 mm Hg. Follow-up ranged from2 to 15 years (mean, 7.5 years). Three patients were lost tofollow-up after 5, 6, and 7 years, respectively.

We believe that 2 patients had PAB-related complica-

tions. One patient had bilateral branch PA stenosis andthe other patient had atrioventricular valve regurgitation.The first patient required pulmonary angioplasty at thetime of modified Fontan and the other had a DeVegavalvuloplasty at the time of DKS operation.

Comment

The recent trend toward early primary DKS operation isappealing and saves that patient from multiple opera-tions [19]. However, there are several reports in whichprimary DKS operation was associated with a high mor-tality [5, 9, 20] (Table 2). This is balanced by more recentreports showing a considerable improvement in survivalwith primary DKS operation [19, 21]. The purpose of thisreview is not to support a staged approach with PAB forall patients with single-ventricle morphology and sub-aortic obstruction, but rather to demonstrate that inselected patients PAB is not necessarily to be condemnedas always leading to PR. In fact, PAB in single-ventriclephysiology is a very controversial issue and there arenumerous articles in the literature for and against thisprocedure [4, 17, 19–21].

The fate of the pulmonary valve in the systemic circu-lation in the context of several other different proceduresappears to be satisfactory. Pulmonary valve function afterarterial switch, Norwood-type repair, and the Ross pro-cedure has been durable with a very low incidence ofvalve regurgitation [23]. On the other hand, the incidenceof PR after DKS operation has been significant in certainseries [5, 8, 17]. The main cause is thought to be distortionof the pulmonary trunk by the DKS and the prior PAB.Additionally, there are the other PAB-related complica-tions like ventricular hypertrophy [24], increase in out-flow tract gradient, atrioventricular valve regurgitation[16], and diastolic dysfunction [20]; all of these subse-quently may have an impact on Fontan survival [25].However our experience suggests that a PAB performedmeticulously without encroachment on the pulmonaryvalve can be successfully used as a short-term palliation

Fig 3. Kaplan-Meier curve of freedom from pulmonary valve insuf-ficiency after pulmonary artery banding followed by Damus-Kaye-Stansel operation.

Table 2. Pulmonary Regurgitation and Mortality After Damus-Kaye-Stansel Operation

Year AuthorNumber of

PatientsPulmonary

Regurgitation (%)Mortality

(%)

Preceded by PAB1991 di Carlo et al [8] 13 28 421993 Huddleston et al [17] 9 58 221993 Lui et al [5] 47 52 251994 Carter et al [7] 23 Not reported 131995 Serraf et al [9] 15 Not reported 501996 Jensen et al [12] 26 Not reported 191998 Amin et al (this report) 15 13 7

Primary DKS1995 Brawn [22] 24 Not reported 211997 Mosca [21] 38 3 211997 McElhinney et al [19] 21 Not reported 17

PAB 5 pulmonary artery banding.

839Ann Thorac Surg AMIN ET AL1998;66:836–41 DOES PAB BEFORE DKS CAUSE PI?

in selected patients. Surgeons recommending PAB as aninitial palliation for these patients need to emphasize thatthe DKS operation should be undertaken sooner ratherthan later (4 to 8 months after PAB) [12, 17].

We believe that there is a subset of patients who canbenefit from a PAB rather than a primary DKS operationwithout affecting the ultimate survival and function of thepulmonary valve. The final decision as to the type ofoperation depends on the surgeon and the individualinstitutional experience. The advantage of the thoracot-omy approach is that there are minimal adhesions to dealwith at the time of DKS. In addition, this is our standardapproach to coarctation repair (without cardiopulmonarybypass), which was required in 9 of 10 patients who hada thoracotomy. The sternotomy approach is our prefer-ence, however, if arch repair is not needed as it allows formore precise band placement and anchoring.

Although 6 patients in this series had a form of TGAwith ventricular septal defect treated by DKS and rightventricular-to-pulmonary artery conduit, since 1990 allpatients with Taussig-Bing anomaly or complex TGAhave undergone either arterial switch operation or Ka-washima repair [26]. These patients are no longer con-sidered candidates for a DKS and most have completerepair in the neonatal period.

In our experience, we believe that the use of a hood ofGore-tex or homograft in completing the anastomosis ofthe DKS operation has helped prevent late pulmonaryvalve regurgitation. Without the hood the wall of thepulmonary trunk away from the aorta is under traction,which can distort the pulmonary trunk and hence thepulmonary valve (Fig 2B). We speculate that this contrib-uted to post-DKS PR. In addition, the impact of theejection jet on one side of the valve (caused by pulmo-nary valve distortion) may cause turbulent flow in thepulmonary trunk that further increases regurgitationwith passage of time. The Kaplan-Meier estimate offreedom from pulmonary insufficiency at 8 years afterDKS in our patients was 0.75. In all univentricular con-nections with restrictive ventricular septal defects, thepulmonary valve is essentially the “good” systemic valve.In our opinion these valves can be used as part of theeventual repair. Other authors have described modifica-tions of the original DKS procedure that minimize ten-sion on the anastomosis [7, 27].

In conclusion, we have demonstrated that in selectedpatients PAB followed by DKS does not necessarily resultin pulmonary valve regurgitation. We feel that PAB is asafe and effective means of palliation for many infantswith single ventricle and heart failure secondary tounobstructed pulmonary blood flow. The initial PABshould be placed distal enough to not abut the pulmo-nary valve, and DKS should be constructed with a pros-thetic hood to prevent pulmonary valve distortion.

References

1. Damus PS. Correspondence. Ann Thorac Surg 1975;20:724–5.2. Kaye MP. Anatomic correction of transposition of great

arteries. Mayo Clin Proc 1975;50:638–40.

3. Stansel SC. A new operation for d-loop transposition of thegreat arteries. Ann Thorac Surg 1975;19:565–7.

4. Ceitham EL, Puga FJ, Danielson GK, McGoon DC, Ritter DG.Results of the Damus-Stansel-Kaye procedure for transpo-sition of great arteries and for double-outlet right ventriclewith subpulmonary ventricular septal defect. Ann ThoracSurg 1984;38:433–7.

5. Lui RC, Williams WG, Trusler GA, et al. Experience with theDamus-Kaye-Stansel procedure for children with Taussig-Bing or univentricular hearts with subaortic stenosis. Circu-lation 1993;88(Part 2):170–6.

6. DeLeon SY, Ilbawi MN, Tubeszewski K, Wilson WR, IdrissFS. The Damus-Stansel-Kaye procedure: anatomical deter-minants and modifications. Ann Thorac Surg 1991;52:680–7.

7. Carter TL, Mainwaring RD, Lamberti JJ. Damus-Kaye-Stansel procedure: midterm follow up and technical consid-erations. Ann Thorac Surg 1994;58:1603–8.

8. Di Carlo DC, Di Donato RM, Carotti A, Ballerina L, Marcel-liti C. Evaluation of the Damus-Kaye-Stansel operation ininfancy. Ann Thorac Surg 1991;52:1148–53.

9. Serraf A, Conte S, Lacour-Gayet F, et al. Systemic obstruc-tion in univentricular hearts: surgical options in the neo-nates. Ann Thorac Surg 1995;60:970–7.

10. Freedom RM. The dinosaur and banding of the main pul-monary trunk in the heart with functionally one ventricleand transposition of the great arteries: a saga of evolutionand caution. J Am Coll Cardiol 1987;10:427–9.

11. Franklin RG, Sullivan ID, Anderson RH, Shinebourne EA,Deanfield JE. Is banding of the pulmonary trunk obsolete forinfants with tricuspid atresia and double inlet ventricle witha discordant ventriculoarterial connection? Role of aorticarch obstruction and subaortic stenosis. J Am Coll Cardiol1990;16:1455–64.

12. Jensen RA, Williams RG, Laks H, Drinkwater D, Kaplan S.Usefulness of banding of the pulmonary trunk with singleventricle physiology at risk for subaortic obstruction. Am JCardiol 1996;77:1089–93.

13. Freedom RM, Benson LE, Smallhorn JF, Williams WG,Trussler GA, Rowe RD. Subaortic stenosis, the univentricu-lar heart, and banding of the pulmonary artery: an analysisof the course of 43 patients with univentricular heart palli-ated by pulmonary artery banding. Circulation 1986;73:758–64.

14. Donofrio MT, Jacobs ML, Norwood WI, Rychik J. Earlychanges in ventricular septal defect size and ventriculargeometry in the single left ventricle after volume unloadingsurgery. J Am Coll Cardiol 1995;26:1008–15.

15. Jonas RA. Intracardiac thrombus after the Fontan operation[Editorial]. J Thorac Cardiovasc Surg 1995;110:1502–3.

16. Malcic I, Sauer U, Stern H, et al. The influence of pulmonaryartery banding on outcome after the Fontan operation.J Thorac Cardiovasc Surg 1992;104:743–7.

17. Huddleston CB, Canter CE, Spray TL. Damus-Kaye-Stanselwith cavopulmonary connection for single ventricle andsubaortic obstruction. Ann Thorac Surg 1993;55:339–46.

18. Chin AJ, Barber G, Helton G, et al. Fate of the pulmonaryvalve after proximal pulmonary artery to ascending aortaanastomosis for aortic outflow obstruction. Am J Cardiol1988;62:435–8.

19. McElhinney DB, Reddy VM, Silverman NH, Hanley FL.Modified Damus-Kaye-Stansel procedure for single ventri-cle, subaortic stenosis and arch obstruction in neonates andinfants: midterm results and techniques for avoiding circu-latory arrest. J Thorac Cardiovasc Surg 1997;114:718–26.

20. Rothman A, Lang P, Lock JE, Jonas RA, Mayer JE, CastanedaAR. Surgical management of subaortic obstruction in singleleft ventricle and tricuspid atresia. J Am Coll Cardiol 1987;10:421–6.

21. Mosca RL, Hannein HA, Kulik TJ, et al. Modified Norwoodoperation for single left ventricle and ventriculoarterial dis-cordance: an improved surgical technique. Ann Thorac Surg1997;64:1126–32.

22. Brawn WJ, Sethia B, Jagtap R, et al. Univentricular heart with

840 AMIN ET AL Ann Thorac SurgDOES PAB BEFORE DKS CAUSE PI? 1998;66:836–41

systemic outflow obstruction: palliation by primary Damusprocedure. Ann Thorac Surg 1995;59:1441–7.

23. Chambers JC, Somerville J, Stone S, Ross D. Pulmonaryautograft procedure for aortic valve disease: long termresults of a pioneer series. Circulation 1997;96:2206–14.

24. Seliem M, Muster AJ, Paul MJ, Benson W. Relation betweenpreoperative left ventricular muscle mass and outcome ofthe Fontan procedure in patients with tricuspid atresia. J AmColl Cardiol 1989;14:750–5.

25. Kirklin JK, Blackstone EH, Kirklin JW, Pacifico AD, Mar-geron LM. The Fontan operation: ventricular hypertrophy,

age, and date of operation as risk factors. J Thorac Cardio-vasc Surg 1986;92:1049–64.

26. Mavroudis C, Backer CL, Muster AJ, Rocchini AP, Rees AH,Gevitz M. Taussig-Bing anomaly: arterial switch versusKawashima intraventricular repair. Ann Thorac Surg 1996;61:1330–8.

27. Van Son JAM, Reddy VM, Haas GS, Hanley FL. Modifiedsurgical techniques for relief of aortic obstruction in {S,L,L}hearts with rudimentary right ventricle and restrictive bul-boventricular foramen. J Thorac Cardiovasc Surg 1995;110:909–15.

DISCUSSION

DR JOHN W. BROWN (Indianapolis, IN): I am interested inknowing whether during this time period you attempted to dothe DKS at the time of initial repair, not doing a band at all, andjust converting the child to a central shunt like we would with aNorwood or a similar physiology. We have used that a numberof times because we have had some pulmonary insufficiency inpatients who we banded and then came back later to do a DKS.

DR AMIN: Since 1990 we have operated on 6 patients in whomwe did a primary DKS operation. One patient died and 1 had aprolonged hospitalization.

DR BROWN: In the survivors, what did their semilunar valveslook like?

DR AMIN: I know the follow-up in only 2 patients with primaryDKS on whom we operated over the last 18 months, and theyhad only trivial semilunar valve regurgitation.

DR CHRISTO I. TCHERVENKOV (Montreal, Que, Canada): Icongratulate you on your nice presentation. I think we are stilltrying to clarify the role of pulmonary artery banding in con-genital heart surgery, which most of us will agree is a limitedone. Some years ago, Bob Freedom wrote an editorial suggestingthat the band should go out with the dinosaurs. Nevertheless, Ithink that there probably are some very limited applications.

I have two comments. The main purpose of your review ofthese patients is to look at the incidence of pulmonary valveregurgitation. What about the incidence of stenosis? Did stenoticvalves develop in any of these patients at long-term follow-up?

DR AMIN: No. In our series there was no patient who hadstenosis of the pulmonic valve.

DR TCHERVENKOV: Second, I recently treated a patient witha double-inlet single ventricle who had severe subaortic ob-struction after pulmonary artery banding at another institution.

I intended to use the pulmonary valve for a Damus-Kaye-Stansel (DKS) operation, having been assured by the preopera-tive echocardiogram that there was no pulmonary valve regur-gitation. When I opened the pulmonary artery to do the DKS,there was no valve to be found; it was gone. There was noregurgitation because there was no valve. So sometimes theechocardiographic assessment may be misleading, perhaps be-cause of the very short distance between the pulmonary valveannulus and the placement of the band.

DR AMIN: Yes, I appreciate your comment. I agree with youthat there is a short distance between the pulmonary valve andthe anastomosis. However, it is very unusual to have completeobliteration of the valve, unless the band was too close to thevalve or eroded through the pulmonary trunk. Four patients inour series also had follow-up cardiac catheterization, and wewere able to see the two valves and to determine that they wereboth competent (the aortic and the pulmonic valve). In addition,when we reviewed our echocardiograms, we viewed the pul-monic valve from three or four different views, ensuring theaccurate evaluation of the pulmonic valve.

DR BROWN: When do you do primary DKS and when do youdo banding followed by subsequent DKS?

DR AMIN: If we have a patient with double-inlet left ventricle{S,L,L} and a mild subaortic pressure gradient, less than20 mm Hg, the patient should be an ideal candidate for pulmo-nary artery banding, especially in institutions that do not have agood experience with a primary DKS or Norwood type ofoperation. The point of this report is not to recommend pulmo-nary artery banding in all these patients. The idea was that thereare centers where primary DKS still has a high mortality. Whendeciding between primary DKS and preliminary pulmonaryartery banding, we believe that surgeons should know thatpulmonary artery banding (in our experience) will not necessar-ily result in significant late pulmonary insufficiency.

841Ann Thorac Surg AMIN ET AL1998;66:836–41 DOES PAB BEFORE DKS CAUSE PI?