Freehand allograft aortic valve replacement and aortic root replacement

J

THoRAc CARDIOVASC SURG

1991;101:545-54

Freehand allograft aortic valve replacement and aortic root replacement Utility of intraoperative echocardiography and Doppler color flow mapping Seventeen cOlRCurlve patients undergoing 20 planned aortic valve replacements with aUograft valves at Stanford University Medical Center were studied with intraoperative epicardial echocardiography and Doppler color flow mapping before and after cardiopulmonary bypass. Native aortic valves were replaced in 12 of the 20 patients, and eight patients underwent second aortic valve procedures. In 17 of 20 patients aUograft selection was guided by prebypass echocardiographic estimates of annular diameter and/or length of aUograft aortic root required. Other prebypass findings included unanticipated severe mitral regurgitation in one patient (which precluded aUograft aortic valve replacement), leftto-right shunts in five patients, ascending aortic dissection in one, and aortic root disease necessitating coronary reimplantation or bypass in two. Postbypass echocardiography demoMtrated acceptable competency of 18 of 19 aUograft valves (mild or no aortic insufficiency). Postbypass echocardiography also documented successful repair of four of five shunts and mild mitral regurgitation in 15 of 19 patients (versus 11 of 19 before bypass). Conclusions: Intraoperative echocardiography-Doppler mapping is a .useful adjunct for aUograft aortic valve or aortic root replacement; it aUows confident selection or' appropriate tissue size before aortic crossclamping, which minimizes delay from aUograft thawing procedures. It also provides helpful information about the extent of aortic root disease and coronary ostial anatomy before bypass, confinns aUograft competency after bypass, and detects accompanying valvular and other hemodynamic lesions before and after aUograft valve replacement.

Thomas Bartzokis, MD, Frederick St. Goar, MD, Aria DiBiase, MD, D. Craig Miller, MD, and Ann F. Bolger, MD, Stanford. Calif.

h e current resurgence of interest in aortic valve replacement (AVR) with an allograft follows substantial recent improvements in allograft procurement and storage capabilities. The freehand insertion technique of these unstented valves is technically demanding, and the combination of allograft AVR with aortic root replacement or reconstruction in some patients further complicates the From the Division of Cardiology and the Department of Cardiovascular Surgery, Stanford University School of Medicine, Stanford, Calif. Read at the Fifteenth Annual Meeting of The Western Thoracic Surgical Association, Monterey, Calif., June 21-25, 1989. Address for reprints: Ann F. Bolger, MD, Divisionof Cardiology, Stanford University School of Medicine Stanford, CA 94305-5246.

12/6/20186

surgical technique.t! Early experience with freehand allograft AVR suggested that long-term success depended on careful placement and matching of the donor valve to the recipient anulus." Intraoperative two-dimensional echocardiography in combination with Doppler color flow mapping provides high-quality images of cardiac anatomy and flow and has been of great value in assessing many complex cardiac surgical procedures.' Its role during allograft AVR has not been described to date, although a tecent report described the utility of postoperative transthoracic Doppler echocardiography." In this study we discuss our experience at Stanford University using intraoperative echocardiography as an aid in allograft selection, allograft placement, and confirmation of the surgical result during 20 freehand allograft AVR procedures performed in 17 patients.

545

The Journal of Thoracic and Cardiovascular Surgery

546 Bartzokis et at.

Table I. Baseline characteristics Patient No.

Sex

Age (yr)

Female

23

2 3

Male Male

66 37

4

Male

37

5

Female

53

6

Female

53

7

Female

22

8

Male

28

9

14

Male Male Male Male Male Male

42 56 27 51 20 26

15 16

Male Male

42 55

17

Female

50

18

Male

26

AR, primary valvular disease AS, bicuspid aortic valve AR, bioprosthesis degeneration AR, bicuspid aortic valve AS, congenital unicuspid valve AR, pulmonary allograft degeneration AR, idiopathic AS, pseudo-Hurler's syndrome, dwarf AS JAR, bioprosthesis degeneration, periprosthetic leak, VSD, healed endocarditis AR, healed endocarditis

19 20

Female Male

26 20

AS (congenital), VSD AR, bicuspid aortic valve

10 II

12 13

Indication AR, Takasayu's aortitis, sinus of Valsalva aneurysm AR, healed endocarditis AR, porcine bioprosthesis, healed endocarditis AR, allograft active fungal endocarditis AS, mechanical prosthesis AR, allograft active bacterial endocarditis, MR, VSD ASjAR, Marfan's syndrome, ascending aortic aneurysm AR, porcine active endocarditis

Procedure Freehand AVR, root replacement Freehand AVR, root replacement Freehand AVR, root replacement Freehand AVR, root replacement* Freehand AVR, root replacement, Manouguian procedure Freehand AVR, root replacement, * VSD repair Freehand AVR, root replacement Freehand AVR, root reconstruction, VSD repair, CABG Freehand AVR Freehand AVR Freehand AVR Freehand AVR Freehand AVR Freehand AVR Freehand AVR Freehand AVR, CABG Freehand AVR, VSD repair

Freehand AVR (pulmonary), mitral repair Freehand AVR, VSD repair Starr-Edwards AV and MV

AR, Aortic regurgitation; MR. mitral regurgitation; VSD, ventricular septal defect; AS, aortic stenosis; CABG, coronary artery bypass grafting; AV, aortic valve; MV, mitral valve. 'Repeat allograft AVR.

Methods Patient population. Twenty allograft AVR procedures were performed in 17 patients (five female and 12 male, aged 18 to 66 years) at Stanford University between April 1988 and May 1989 (Table I). The indication for valve replacement was aortic regurgitation in 16 of 20 and aortic stenosis in the other four; Native aortic valves were replaced in 12 patients and previously inserted prosthetic aortic valves were replaced in eight (four porcine, one Smeloff-Cutter [Cutter Biological, Berkeley, Calif.], and three allograft valves). In three patients allograft AVR was performed within 8 weeks of prior AVR because of active prosthetic valve endocarditis. Two of these patients had bacterial endocarditis (one porcine prosthesis and one allograft); one had fungal endocarditis (Candidaalbicans) complicating previousallograft implantation. These three patients died postoperatively of left ventricular and multisystem failure. All remaining patients except one (suicide) are currently alive and well. Allograft AVR was performed in combination with aortic root replacement in seven patients. In these patients, the

allograft aortic valve was implanted with variable amounts of retained proximal aorta. This required excision of the recipient proximal aorta and reimplantation of the native coronary arteries.s 3 Allograft AVR without root replacement was performed in 12 patients. Concomitant cardiac procedures were performed in seven patients (see Table I). A single patient underwent AVR with a pulmonary allograft with concomitant mitral valve repair. In one patient with severe mitral regurgitation, aortic and mitral valves were replaced with Starr-Edwards prostheses (Baxter Healthcare Corp., Edwards Division,Santa Ana, Calif.) rather than the planned allografts. Allograft valves. Allograft valves were supplied by Cryolife, Inc., Marietta, Georgia, and stored at a temperature of -196 0 C. The complete allograft thawing process required approximately 20 minutes. Allograft valves were inserted freehand (unstented) with varying amounts of proximal aortic root as dictated by the local pathologic anatomy. Aortic allograft sizes were derived from the allograft internal diameter; the manufacturer-reported difference between internal and external allograft diameters is 3 to 5 mm depending on size.

Volume 101 Number 3 March 1991

Allograft A VR

547

Fig. 1. Example of annular sizing from an intraoperative epicardial two-dimensional echocardiogram. Dotted line shows annular measurement taken at the base of the aortic leaflets. Closedarrow shows the anterior mitral leaflet; openarrow points to an aortic leaflet. (Systolic frame, long-axis view; LA. left atrium; LV, left ventricle, Ao. aorta.)

Transtboracic and intraoperative ecbocardiograpby and Doppler color ftow mapping. All patients were studied intraoperatively' .with echocardiography and Doppler color flow mapping (IDE). A sterilely draped 5 MHz transducer (model 77020AC, Hewlett-Packard Co., Andover, Mass.) was placed directly on the anterior epicardial surface and images were recorded from multiple views on half-inch videotape. Imaging was performed before cardiopulmonary bypass (pre-CPB) and early after removal of the aortic crossclamp and restoration of native circulation (post-CPB). The post-CPB examination was performed as early as practical such that if a technical flaw was identified, the patient would still be hypothermic, and recooling and total CPB time would be reduced. Importantly, this approach would minimize the amount of time the patient's left ventricle would be exposed to adverse hemodynamic consequences (e.g., severe aortic regurgitation or stenosis) if a technical problem was present. This plan also reduced the length of time the echocardiography equipment had to be in the operating room and unavailable for other clinical use. If later studies to assess the completeness of "de-airing" the left-sided cardiac chambers were requested, these were also performed. Aortic regurgitation was quantitated by means of color flow mapping by regurgitant jet width immediately below the aortic valve;a regurgitant jet occupying less than one fourth of the left ventricular outflow tract was defined as mild aortic regurgitation, one fourth to one half as moderate, and more than one half as severe?' 8 Mitral regurgitation was also quantitated by color flow mapping. A regurgitant jet occupying less than one third of the atrium was classified as mild, one to two thirds as moderate, and more than two thirds as severe.? Intraoperative imaging procedures were performed with careful attention to sterility. In this series of patients, there was one case of mediastinitis that developed I month after allograft AVR and was attributed to sternal necrosis, diabetes, and obesity. No other wound infections occurred. Transthoracic two-dimensional Doppler echocardiography and color flow mapping (TIE) were performed within 4 weeks

before the operation in 14 cases and within 4 weeks after the operation in 16 cases. Postoperative TIE was not available because of perioperative deaths in three patients and nonallograft replacement of the aortic and mitral valves in one patient. Measurements of aortic valve diameter. The aortic valve and anulus were inspected during transthoracic and intraoperative studies from multiple views including parasternal longand short-axis planes. Annular diameter was measured from frozen systolic frames in the parasternal long-axis view. Annular measurements were taken immediately inferior to the base of the aortic valve leaflets (Fig . I). This measurement was reported as the mean of three measurements from different systolic frames . An allograft roughly 3 to 5 mm smaller than the IDE annular diameter was then selected and thawed before the institution of CPB. During the operation, aortic valve diameter was measured at the anulus with a calibrated dilator. In cases of a prosthetic aortic valve, the reported external diameter was recorded. In two patients, surgical aortic annular diameter was determined by pathologic examination of the excised intact native valve. Statistical metbods. Results from different tests were correlated by linear regression analysis. The data are presented as scatter graphs with corresponding correlation coefficients (r) and 90% confidence intervals .

Results Measurements of annular diameter. IOE measurements of annular diameter were technically adequate in all patients and guided allograft selection in 17 of 20 eases. The correlation between IOE and surgical annular diameter was high; r = 0.78 (Fig. 2B). Allograft sizes 3 to 5 mm less than the IOE annular diameters were selected (average, 4.0 ± 1.8 mm) whenever possible, although limited availability of allograft sizes, especially

The Journal of

5 4 8 Bartzokis et al.

Thoracic and Cardiovascular Surgery

C.

30

A.

r = 0.92 P = 0.0001

34 32

28

26

E E

30

",'

N

28

..,

26

r = 0.55 p=0.015


~

'" L

E 24 E

'"

c:

c:

18

20

22

24

26

28

Fig. 2A. Comparison of annular diameter measured with TIE preoperatively and epicardial IOE.

34

r=0.78 p = 0.0001

o

32 30

/

/

/

/

/

28

0 0 / /

26 /

9/ 22

/6

c;

WJ

20

/

o

/

/

/

/

/

24

L

«

/

/

/

~

c:

22

24

26

28.30

32

34

30

Preop Echo Annular Diameter, mm

::>

20

Fig. 2C. Comparison of intraoperative annular diameter and selected allograft size. 16

'"

18

10E Annular Diameter, mm 16

'" ~

18

16

18

~

L

20

0

16

::>

E E

u


'" 0;

~ 20

B.

22

s

E

WJ

:;;: -o

22

0'" -c

24

~

'-

~

0>

P

/

o

/

0

0

/

/

~

... o

16 16

18

20

22

24

26

28

30

32

34

Surgical Aortic Valve Diameter, mm

Fig. 2B. Comparison of internal aortic valve diameter measured surgically and annular diameter measured with IOE.

larger sizes, did not allow this in all cases(Table II, Fig. 2e). This exigency explains the relatively weak correlation between IOE-determined diameter and size of allograftvalve selected (Fig. 2e). Preoperative TIE measurements of annular diameter wereperformed in 14patientsand considered technically adequateto assess annulardiameterin 12(86%). In these patients, annular diameter measurements from IOE and

TIE correlated highly with the correlation coefficient; r = 0.92 (Fig. 2A). Allograft valve function. Immediate post-CPB assessment of the allograft by IOE revealed normal leaflet motion, unrestricted outflow tract, and properpositioning of the aortic valve in all cases. Postoperative TTE demonstrated normal Doppler forward flow velocities across all the allograftvalves and thus confirmed the absence of stenosis. Acceptable allograft competencywasdocumented by IOE in 18 of 19 allograft recipients. No aortic regurgitation was seen in 10 of 19,and mildtransvalvular aortic regurgitation was seenin seven of 19 (Fig. 3). A small perivalvular leak causing mild aortic regurgitation wasdetected in one patient.The single patient with significant aortic regurgitation after CPB diedpostoperatively of severe left ventricular dysfunction. The severity of aortic regurgitation did not changeby more than one grade in any case during postoperative TIE evaluation (Fig.3).Thesingle patientwiththesmall perivalvular leakdid not haveanychangein aorticregurgitationgrade,and nonewcasesof perivalvular leakwere identified. In the onlypatienthaving AVR witha pulmonaryallograft, aorticregurgitation increased frommildto moderate postoperatively; progressive aortic regurgitation over the following 10 months prompted another allograft AVR. Intracardiac and aortic disease. IOE identified intracardiacabnormalities in 10 patients; in threeof them, abnormalities werenotanticipated preoperatively. In one patient, IOE detection of severe, previously underestimated mitral regurgitation resulted in abandonment of planned allograftAVR in favor of combined insertion of mitral and aortic mechanical prostheses. Intracardiac left-to-right shunts were detected by IOE in five cases:

Volume 101 Number 3 March 1991

Allograft AVR 5 4 9

Severe Moderate AR

Severity

Mild None Pre-CPB

Post-CPB

Post-Dp

Echocardi ogram Fig. 3. Severity of aortic regurgitation (AR) by echocardiography at pre-CPB, post-CPB, and postoperative examination.

three small interventricular communications at the level of the membranous septum and two aorta-to-right ventricular fistulas causedby endocarditis (Fig. 4, A and B). Each of 'these shunts was suspected preoperatively and confirmed during surgical exploration, although one of the aorta-right ventricular fistulas was proved to be a large ventricular septal defect located immediately inferior to the anulus. Post-CPB IOE confirmed complete repair of the intracardiac shunts in four of five patients. Aortic dissection was identified by IOE and confirmed surgically in a patient with Marfan's syndrome. This had beensuspected preoperatively by computedtomographic scan but not definitively identified. Abnormal, low-lying coronary ostia were detected duringpre-CPBexamination in one patient;this wasrecognized onlyin retrospect. Because the allograftroot had been trimmed beforethe need for coronary reimplantation was recognized, coronary bypass was performed rather than the preferableallograftrootreplacementwith coronary reimplantation. On all subsequent pre-CPB studies, the coronary ostia were identified. In another patient,pre-CPBIOE identified involvement of the right coronaryostium in a perivalvular abscess, which alerted the surgicalteam to the potentialneed for right coronary artery bypass (the right internal mammary artery was dissected free before CPB was instituted). In a patient with active fungal endocarditis, new hypokinesia of the anterior wall on post-CPB IOE suggested embolization of the left anterior descending coronary artery; a vegetation was subsequently removed from the coronary artery with a Fogarty embolectomy catheter with subsequent improvement in wall motion.

Post-CPBIOE revealed mildmitral regurgitationin 15 studiesand severe mitral regurgitationin one.The patient withsevere mitral regurgitationhad had onlymild regurgitation before CPB; the cause of this increase was a technically inadequate repeat Manouguian procedure'? in a patientwithactiveendocarditis. The severity of mitral regurgitationdid not change significantly from the intraoperative post-CPB period to the postoperative study (Fig. 5). Discussion Freehand aortic allograft insertion is gaining renewed popularity, but it remains a complex technique that dependson detailed and accurate knowledge of the proximal aortic,annular, and coronaryanatomy. Echocardiographic imaging from the epicardium provides highqualityimageswithexcellent resolution of theseanatomic structures and intracardiac flow. The multiple imaging planes accessible from the epicardium allow careful delineation of structures in three dimensions. This informationcan beobtainedquicklywith minimaldelayto the surgicalteam and with no prolongation of CPB or aortic crossclamp time. The advantagesof IOE have been recognized at many centerswhere IOE has beenusefulduring other complex cardiac surgicalprocedures, including mitral valve repair and pediatric cardiovascular reconstructions.S II. 12 In this study we have shown that IOE is a similarly important adjunct during allograft AVR and aortic root replacement. Direct surgicalsizingof the aortic anulus after aortic crossclamping prolongs the ischemic time by the 20 minutes required for allograft thawing. A reliable pre-CPB

5 5 0 Bartzokis et al.

The Journaf of Thoracic and Cardiovascular Surgery

Fig. 4. A, Two-dimensional IOE at the level of the aortic anulus in a patient with a fistulous connection between the aorta and the right ventricle. (Short axis view; LA, left atrium; RA, right atrium; RV, right ventricle; Ao, aorta.) B, Superimposition of color Doppler information shows the systolic jet through the fistula from aorta to right ventricle (arrow).

estimate of annular size from IOE enables the surgeon to select allograft size before direct valve inspection. This allows adequate lead time for thawing of the stored, frozen allograft before institution of CPB. In this series we have shown that IOE estimates of annular size reliably correlate with surgical estimates of aortic valve diameter and can be used as a pre-CPB estimate of allograft size, therefore minimizing aortic crossclamp time and CPB time and optimizing .allograft-recipient anulus match . Short-term and long-term aortic allograft patency rates may be improved by optimizing the match of allograft valve to annular size." We also found an excellent correlation between transthoracic and intraoperative estimates of annular diameter when measured by experienced echocardiographers

familiar with the relevant surgical landmarks. However, in 14%of the TTE studies, image quality was too poor to provide a reliable annular size. In the setting of goodquality preoperative studies, required allograft sizecan be estimated if careful attention is paid to obtaining reproducible parasternal long-axis views, consistent measurement techniques, and discipline in rejecting studies of borderline quality or nonstandard views.Because annular sizing is not a conventional or rigidly standardized echocardiographic measurement, reliable use of this information requires ongoing cooperation and dialogue between the echocardiographer and the surgical team, and each clinical laboratory must document the quality of the measurements by correlation with IOE and surgical data.

Volume 101 Number 3 March 1991

Allograft AVR 5 5 1

Severe Moderate

Mild

None

Prepump

Postpump

Peste pe rati ve

Echocardiogram Fig. 5. Severity of mitral regurgitation (MR) by echocardiography at pre-CPB, post-CPB, and postoperative examination.

IOE wasinstrumental in defining previously unrecognized intracardiac and aortic root abnormalities, which wasimportant in planning the surgical approach. In the seven cases in which allograft aortic valve with root replacement wasperformed, the IOE waspredominantly instrumental in assessing the amount of proximal allograft root necessary to incorporate in the repair and the need for coronary reimplantation or bypass. In 13 cases in which isolated freehand AVR without rootreplacement wasanticipated, the IOE provided new additional information that altered the surgicalapproach in threecases. Theseincluded onepatient in whomsevere mitral regurgitation had been underestimated by preoperative evaluation. As a result,this patient received combined mitral and aortic valve replacement with mechanical valves rather than an allograft aortic valve. In the second patientthe pre-CPB10E identified involvement of the right coronary ostiumin a perivalvular abscess, alerting the surgical team to the potentialneed for right coronary artery bypass (the right internal mammary artery wasdissected freebeforeCPB wasinstituted). In the third patient, low-lying coronaryostia were identified only in retrospect. Because the allograft root had been trimmed before the need for coronary reimplantation was recognized, coronary bypass was performed rather than the preferable allograftrootreplacement withcoronaryreimplantation. On all subsequent pre-CPB studiesthe coronary ostia were identified.

Table II JOE annular diameter (mm)

Allograft diameter (mm)

26

21 19 18 21 18 18 20 19 19 18 18 21 20

23 24

28 22 23

21 23

21 19 19 26 26 26 22 32

21 28 24

22

18 24 25

21 21

IOE was highly sensitive in identifying left-to-right shunts at the ventricularseptum or proximalaorta. Preoperative TIE had identified theselesions as well. Therefore, the greatest benefit of IOE in this setting was the post-CPB assessment of the adequacy of shunt repair. Most allograft valves were competent after insertion. IOE demonstrated that aorticregurgitation wasabsentor

The Journa l of

5 5 2 Bartzokis et al.

Thoracic and Cardiovascular Surgery

Fig. 6. Two-dimensional IOE performed in the early post-CPB period shows multiple smaIl air bubbles in the leftsided cardiac chambers. (Long-axis view; LA, left atrium; LV, left ventricle; Ao, aorta.)

mild in all but one case. The patient who underwent AVR with a pulmonary allograft was noted to have an early postoperative increase in aortic regurgitation that progressed over 10 months. There were no cases of structural valve deterioration" among the patients having AVR with an aortic allograft. Because of the paucity of cases of late structural failure during this follow-up period, it is not possible to comment on the JOE predictors of late allograft failure. Our study suggests that immediate post-CPB severity of aortic insufficiency is representative of early postoperative findings. With the surgical technique and sizing procedure used, there were no early postoperative aortic allograft valve failures and no unsuspected findings on follow-up echocardiographic studies. Additional information available from the post-CPB JOE study included assessment of regional and global left ventricular wall motion and detection of residual intracardiac air. Small "bubbles" were normally seen in both left and right ventricles during the immediate post-CPB study; repeat JOE assessment later after completion ofall "de-airing" maneuvers can confirm the completeness of this step (Fig. 6). JOE was also helpful in assessing concomitant postCPB mitral insufficiency. Post-CPB JOE identified mild mitral regurgitation in 15 cases and severe regurgitation in one. Grade of mitral regurgitation did not increase on subsequent postoperative TIE studies in any case. Therefore, the presence of mild mitral regurgitation on the immediate post-CPB JOE is common and does not appear to identify patients with high risk for early progression of mitral regurgitation.

a

Conclusions JOE providesvaluable information about anatomy and flow that is useful in planning and assessing allograft AVR. The time required for imaging is minimal and can result in important savings in crossclamp and CPB time. Echocardiographic information aids allograft sizing, identification of associated intracardiac abnormalities, assessment of allograft aortic and other valvular competency, and evaluation of the overall success of the operative procedures. The early postoperative success of this procedure at our institution is encouraging; ongoing follow-up of these patients will be required to determine if the improved match of allograft size to recipient anulus and better allograft preservation will combine to improve the long-term success of freehand allograft AVR and aortic root replacement.

REFERE NCES 1. Moreno-Cabral CE, Miller DC , Shumway NE. A simple technique for aortic replacement using freehand allografts. J Cardiac Surg 1988;3:69-76. 2. Okita Y, Franciosi G, Matsuki 0, Robles A, Ross D. Early and late results of aortic root replacement with antibioticsterilized aortic homograft. J THORAC CARDIOVASC SURG 1988;95:696-704. 3. Saldanha RF, Raman J , Feneley M, Farnsworth AE. Homograft aortic root replacement to correct infective endocarditis requiring seven open cardiac procedures. Ann Thorac Surg 1989;47:300-1. 4. Barratt-Boyes BG, Roche HG, Subramanian R, Pemberton JR, Whitlock RML. Long-term foIlow-up of patients

Volume 101 Number 3 March 1991

with the antibioticsterilized aortic homograftvalve inserted freehand in the aorticposition. Circulation1987;75:76877.

5. Maurer G, Czer LSC, Chaux A, et al. Intraoperative Doppler colorflow mappingfor assessment of valve repair for mitral regurgitation. Am J CardioI1987;60:333-7. 6. Jaffe WM, Coverdale HA, Roche AHG, et al. Doppler echocardiography in the assessment of the homograftaortic valve. Am J CardioI1989;63:1466-70. 7. SwitzerDF, Yoganathan AP, Nanda NC, et al. Calibrationof colorDoppler flow mappingduring extreme hemodynamic conditions in vitro: a foundation for a reliable quantitative grading systemfor aortic incompetence. Circulation 1987;75:837-46. 8. Perry GL, Helmcke F, Nanda NC, et al. Evaluation of aortic insufficiency by Dopplercolor flow mapping. J Am ColI CardioI1987;9:952-9. 9. Helmcke F, Nanda NC, Hsiung MC, et al. Color Doppler assessment of mitral regurgitation with orthogonal planes. Circulation 1987;75:175-83. 10. Manouguian S, Seybold-Epting W. Patch enlargement of theaorticvalve ringbyextending the aorticincision intothe anteriormitral leaflet: newoperative technique. J THORAC CARDIOVASC SURG 1979;78:402-12. 11. Ungerleider RM, Greeley WJ, Sheikh, et al. The use of intraoPerative Doppler with color flow imaging to predict outcomes following repair of congenital cardiac defects. Ann Surg [In press]. 12. Maurer G, Czer LSC, Chaux A, MatloffJM. Intraoperativecolorflow Doppler in evaluatingvalvuloplasty and correction of congenital heart disease. Am J Cardiac Imaging 1987;3:234-41. 13. Edmunds LH, Clark RE, Cohn LH, Miller DC, Weisel RD. Guidelines for reporting morbidityand mortalityafter cardiacvalvular operations. J THORAC CARDIOVASC SURG 1988;96:351-3.

Discussion Dr. David R. Clarke (Denver, Colo.). Although I do not question at all the effectiveness of this expensive modality in evaluating cardiacanatomy and physiology and, in particular, evaluating the sizeof the aortic anulus, I wonderwhether it is routinely necessary. I have these thoughts for several reasons. First,in eight of the 20 casesin this series, aortic root replacementwasthe procedure ofchoice. It isclear to me that in aortic rootreplacement a relatively widerangeofallograftsizescan be used in a particularheart. The degreeof variability is such that precise sizing is not necessary becausethe competence of the aortic valve depends on the intrinsic structure of the allograft and does not depend at all on the annular match between the allograftand the recipient. Second, a technique has been described using caliper measurement of the aortic root at the level of the top of the sinuses ofValsalva, which correlates verywellwith the sizeof the anuIus. If this measurement is made externally early in the procedure,and approximately 6 to 8 mm subtracted from this mea-

Allograft AVR 5 5 3

surement,an appropriatesizedallograft can be thawed before aortic clamping. I would certainly applaud this approach. Third, Mr. DonaldRossfrom London,whohas probablyhad moreexperience withallograftvalvereplacementthan anyofus, sizes hisvalves the waywesizegolfshirts-small, medium,and large.Thereforehe must not think that exact size matching is important. Fourth,the measuredinternaldiameterofthe allograftsisnot as precise as we would liketo believe. Differenttechnicians can achieve as muchas 3 or 4 mm ofvariabilitywhenthey measure the internaldiameter of an aortic allograft,whichmakes exact recipient annular sizingmoot. Last, the same information apparently can be obtained by means of preoperative echocardiography and angiographic evaluation of anulus size.One of your figures (Fig. 2A) shows a highdegreeof correlation betweenthe preoperative and intraoperative echocardiographic measurements. I have no argument with the use of this techniquein selected caseswhenthe surgeonbelieves that the anatomy may be complex, when there is difficulty in weaning the patient from the pump,or when a problembecomes apparent after the start of the operation, but I am concerned about the routine use of an expensive modalitywhen we are all battling risinghealth care costs. I havea fewquestions. First,couldyoutellus the costof these studiesto the patientand the time the equipmentand the technician are tied up to do one of these studies? Dr. Bartzokis. The cost of an intraoperative study is approximately the costof a transthoracicstudy-about $500althoughat Stanford we don't actually charge for the intraoperativestudyif the patient has had a preoperative transthoracic study. In terms of technician time, we do not use a dedicated echocardiography machine in the operating room. In concert with the surgeons, the technician willgo downwhencalledfor, set thingsup, and then approximately 10 minutesbeforeimaging the cardiologists will arrive for the procedure. It takes approximately half an hour of setup time for each stage of the procedure: the pre-CPB and post-CPB stages. Dr. Clarke. Is the machineleft in the operatingroomduring the entire procedure? Dr. Bartzokis. If the machineis needed, it willbe removed and brought back for the second imaging. Dr. Clarke. Do youhaveany idea as to the costeffectiveness ofthis,eitherbyprevention ofveryexpensive therapythat might berequiredif it is notapplied, or byprevention of morbidityand mortality?In howmanyof thesepatientsdid the findings in the echocardiogram really modify the therapy that was accomplished and provide information that youcouldnot havelearned any other way? Dr. Bartzokis. The major benefitof the preoperative study is in sizing of the valve. Other modalities that can be used for sizing include preoperative echocardiography or left ventricular angiography. Manyofthecasesinthisstudywerecomplex, with perivalvular fistulas, and I think it was helpful for the surgeons to have an idea not onlyof annular size,but also of the extent of perivalvular involvement by abscess or infection which, at least in one case, modified the allograft size they used. In this case,ourannularmeasurement wasapproximately 21or 22mm but, becauseof a large perivalvular abscess, the surgeonschose a larger allograft. The postoperative studywasparticularlyhelpful. There were

5 5 4 Bartzokis et al.

four patients in whom the surgeons modified their approach. One patient required a right internal mammary artery bypass graft for a right coronary ostium that was involved in the perivalvular abscess. Another patient actually had very low-lying coronary ostia. The surgeons had hoped to put a simple freehand allograft aortic valve in place, but the position of the coronary ostia prevented that. One patient did not undergo an allograft replacement because of underestimated mitral regurgitation preoperatively. That patient underwent double mechanical valve replacement. In those three, there was certainly a change in treatment. In a fourth patient, who was not mentioned here, postoperative anterior wall motion was markedly abnormal, and he had had endocarditis with a previous allograft valve. That alerted the surgeons to the possibility of an embolus, and they subsequently removed a mycotic embolus from the left anterior descending coronary artery. Dr. Clarke. Another question related to echocardiography: It seems to me that epicardial echocardiography, although relatively effective, is somewhat cumbersome, consumes surgeon time and effort, and increases the possibility of introducing infection into the operative area. What is your opinion about the use of transesophageal echocardiography in this regard? Dr. Bartzokis. Transesophageal echocardiography has its advantages in the operating room. Invasion of the operative field by the transducer is not necessary, and the transducer can be left in place throughout the entire procedure. Its major limitations in our hands, at least in this type of operation, are that it has

The Journal oi Thoracic and Cardiovascular Surgery

provided us with suboptimal views of the aortic valve. Many of these patients have complex anatomic problems, and the epicardial approach offers multiple windows from which to sample the aortic valve and proximal aortic root. A biplane transesophageal echocardiograph is currently being introduced, which may increase the number of windows that are available, but it is a $40,000 piece of equipment and certainly raises the cost substantially. Dr. Clarke. I have one final question that is not related to echocardiography. You mentioned the one patient who had a pulmonary allograft chosen for AYR, and this patient subsequently required rereplacement because of progressive regurgitation. Would you say that this one patient should give us some cause to think that the pulmonary allograft may not be an adequate aortic valve substitute, or was there something specific about this patient from a technical standpoint or otherwise that caused this to happen? Dr. Bartzokis. My experience in this matter is limited. In terms of the initial results, the pulmonary allograft fit well. It was approximately 4 mm smaller than our annular diameter. He had trace aortic regurgitation immediately post-CPO and postoperatively, and subsequently more aortic regurgitation developed. I cannot answer the first part of your question because of my limited experience, but there was nothing unusual in the early echocardiographic studies to suggest that reoperation would be necessary.