homograft aortic replacement in aortic incompetence and ... · homograft aortic valve replacement...

Thorax (1964), 19, 131

Homograft aortic valve replacement inaortic incompetence and stenosis

B. G. BARRATT-BOYES

From the Cardio-thoracic Surgical Unit, Green Lane Hospital, Auckland, New Zealand

The surgery of aortic incompetence and calcificaortic stenosis has made disappointing progress,chiefly because of the need for partial or totalvalve replacement in many of these patients andthe technical problems inherent in designing anartificial valve. In rheumatic aortic incompetenceit is now fairly clear that total replacement of thevalve is almost always required (McGoon, Man-kin, and Kirklin, 1963), and in calcific aorticstenosis the lesion is by no means always ade-quately relieved by decalcifying the cusps, so thatpartial or total valve replacement is necessary ina proportion of these cases also.

Dissatisfaction with the available plastic aorticvalve prostheses led to a review of the experi-mental and clinical experience with homograftaortic valves. The experimental evidence (Murray,1956; Lam, Aram, and Munnell, 1952; Brewin,1956; Beall, Morris, Cooley, and De Bakey, 1961),while a little conflicting, suggests that with thehost valve rendered incompetent a homograftvalve will function in the dog. Clinical experiencefrom Toronto (Murray, 1956; Kerwin, Lenkei,and Wilson, 1962), where homograft aortic valveshave been inserted in the descending thoracic aortain patients with severe aortic incompetence, showsthat these valves will function normally for upto six years. The evidence seemed sufficient tojustify attempts to insert homograft valves in asubcoronary position in anticipation that theywould function better and for longer periods thanplastic prostheses.The technique used for inserting the valve was

based on experience gained in approximately 100open operations on the aortic valve and was firstused in August 1962 before receiving the reportof a similar procedure performed by Ross (1962)at about the same time. The technique has provedvery satisfactory and the valves have all con-tinued to function well. Indeed the experience pre-sented here, including necropsy examination oftwo valves three and a half and four and a halfmonths after insertion, is most encouraging and

suggests that homograft aortic valves are worthyof serious consideration for aortic valve replace-ment.

MATERIAL

During the 16 months from August 1962 to Novem-ber 1963 inclusive, the aortic valve has been com-pletely excised and replaced by a homograft aorticvalve in 21 patients with aortic incompetence, 17 withcalcific aortic stenosis, and six with multivalvulardisease. Clinical data obtained before and after opera-tion in these 44 patients and various operative findingsare listed in the Table. In this same period one furtherpatient with aortic incompetence underwent operation,and it was intended to replace the valve with a homo-graft valve. Death occurred, however, immediatelybypass was started from retroperitoneal rupture of anaortic dissection which began at the site of cannula-tion of the external iliac artery and ended at the pointof rupture. The aortic wall showed extensive medio-necrosis at necropsy. Although this was a 'perfusion 'death, it does not reflect directly on homograft valvereplacement and will not be considered further.The ages varied from 14 to 66 years, and 20 patients

were over the age of 50. Symptoms have been dividedinto four grades of severity: grade 4 if there was ahistory of congestive heart failure, paroxysmal noc-turnal dyspnoea, orthopnoea or angina decubitus;grade 3 if there was angina of effort or syncope inaddition to effort dyspnoea; and grade 2 if there wasmoderate effort dyspnoea only. On this basis 28 ofthe 44 patients had grade 4 symptoms, 11 grade 3, andfive grade 2. The electrocardiogram showed a completeleft bundle branch block in cases 22 and 31 and acomplete right bundle branch block in case 21. Leftventricular surface leads and their equivalents showedinverted T waves with or without ST depression, indi-cating severe left ventricular hypertrophy in all butthree of the remaining patients, and in these (cases 16.19. and 35) the changes were moderate with T wavesof low voltage.

PREVIOUS SURGERY Four patients had had previousheart-lung bypass surgery, three for calcific aorticstenosis (cases 12, 17, and 22) and one (case 15) forventricular septal defect with severe aortic incom-petence. In case 12 the surgery was undertaken else-

131

on June 12, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.19.2.131 on 1 March 1964. D

ownloaded from

http://thorax.bmj.com/

B. G. Barratt-Boyes

TABLE

GradientPerfusion Data Across

Homo-

Coronary Haemo- graftB.P. By-pass Perfusion lysis Valve 'Post-operative

(mm.Hg) (min.) (min.) (mg. %) (mm.Hg) Complications

0

0

8

3

4

0

10

0

3

0

14

Stenosis4

36

8

6

0

0

0

0

0

0

0

Nil

Nil

Nil

Nil

Bleeding, Died

Positive bloodculture

Nil

Nil

Nil

Nil

Nil

Nil

Nil

Post-perfusicnsyndrome

Nil

Nil

Nil

Nil

Nil

Arrhythmiacardiac arrest

Nil

Follc

B.P.(mm.Hg)

105 !70

115,90

165!80150 90

135185145 85

140 85

140,'85130170160190130/70130 80

1 10'80

140 75

135, 80150 80

130,85

130,85

160/100

140i90

Nil 125/80Pulmonary embolus 135/80Nil 110170

Pulmonary embolus 110/70Nil 1301!85

Dissection leftcoronary Died

Pulmonary embolus 130,'80Heart block 130'80

(temporary)Nil 115/70Post-perfusion 130/70syndrome

Nil 140/80Post-cardiotomy 130/85

syndromneNil 135/85Nil 125/70

Heart block 130,'80(temporary)

Nil 130,175Nil 115'70

)w-up (December 1963)

Murmur' Status

M.S.M. 0-1 GoodE.D.M. 0-1M.S.M. 1-2 Poor

(Cardio-myopathy)

M.S.M. I Good

M.S.M. 1-2 GoodE.D.M. 1-2

M.S.M. 1-2 GoodE.D.M. 1-2M.S.M. 1-2 DiedE.D.M. 1-2 (June 1963)M.S.M. 0-1 Died

(July 1963)M.S.M. 1-2 GoodE.D.M. 0-1M.S.M. 1-2 Good

M.S.M. 1 GoodP.S.M. 1-2M.S.M. 1-2 Good

M.S.M. 1-2 GoodE.D.M. 1-2M.S.M. 2 GoodE.D.M. 1-2M.S.M. 2 GoodE.D.M. 2M.S.M. 1-2 GoodE.D.M. IM.S.M. I GoodE.D.M. 1-2M.S.M. 1-2 GoodE.D.M. 0-1M.S.M. I Good

M.S.M. 1-2 GoodE.D.M. 1-2M.S.M. 1-2 Good

M.S.M. 1-2E.D.M. I

M.S.M. 1-2

M.S.M. 1

M.S.M. 1

M.S.M. 1-2E.D.M. 1-2

Good

Good

Good

Good

Good

M.S.M. 0-1 Good

M.S.M. 1-2 Good

M.S.M. 1-2 Good

M.S.M. 1-2 GoodE.D.M. 1-2M.S.M. 1-2 Good

M.S.M. 1-2 GoodE.D.M. 0-1M.S.M. 1 Good

M.S.M. 1 GoodE.D.M. 1

M.S.M. 1 Good

M.S.M. 1 Good

M.S.M. I Good

(continued)

132

Aortic Incompetence88 73 0

93 52

CaseNo.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

Gradeof

Symp-toms

4

4

4

3

3

2

4

3

4

4

2

4

4

4

4

2

4

4

2

4

4

4

3

4

3

4

3

3

3

4

3

3

4

2

3

4

4

4

Age,'Sex/Operation

14 F23.8.6221 M9.10.62

54 M6.11.6236 M18. 12.6232 F15.1.6327 M29.1.6362 F28.2.6345 M14.3.6355 M18.4.6352 M2.5.6323 M8.5.6360 F12.6.6317 F

27.6.6327 M3.7.6316 M

3 1.7.6323 M17.10.6360 M29.10.6363 M6.11.6318 M7.11.6341 M12.11.6348 M21.11.63

47 M13.12.6253 M

27.2.6359 M11.6.6354 M9.7.6345 M10.7.6351 M18.7.6354 M23.7.6350 F6.8.6352 M8.8.6356 M15.8.6350 F

22.8.6357 M4.9.6352 M16.10.6336 M30.10.6366 M31.10.6359 M13.11.6359 M14.11.63

140, 35 0

135,30-0

180J'40

170 40-0

160,,20-0

170i60

140,15018040-0

2001t60

165/35

165,35-0

170'40

150 0

130/30145'0

140'55

170/60130/40140i0205,50-0

220,150-0

1 10!80

1 10/80

110'85

105 70

110l50

120'75

120175120180

110150135/80

125,'85115,175

140'90

110/85

160/115

120165

140170

101

114

108

143

112

117

100

105

116

179

104

95

105

127

101

150

96

128

91

132

117

132

133

143

121

125

128

121

130

109

134

111

117

111

148

128

137

105

83

101

88

82

78

84

81

153

82

77

85

96

67U/

121

78

103

71

102

90

Calcific103

113

106

94

95

104

99

95

72

107

78

81

83

123

83

80

81

144

146

88

40

39

56

29

59

28

46

47

24

30

57

30

15

47

150

36

Aortic304

30

40

15

33

43

27

56

22

69

27

19

41

56

45

20

65

..i1.I



/T


ownloaded from


Homograft aortic valve replacement in aortic incompetence and stenosis

TABLE-(continued)

GradientPerfusion Data Across

Grade Homo- Follow-up (December 1963)of Coronary Haemo- graft

Case Age/Sex! Symp- B.P. By-pass Perfusion lysis Valve Post-operative B.P.No. Operation toms (mm.Hg) (min.) (min.) (mg.%) (mm.Hg) Complications (mm.Hg) Murmur1 Status

Multivalvular Disease39 54 M 4 155'65 137 110 33 -_ Heart block 115/80 M.S.M. 1 Fair

A.I.andl 25.6.63 (permanent) P.S.M. 2M.I.I40 43 F 4 155/80 164 140 40 - Nil 110/70 M.S.M. 2 GoodA.I. 25.7.63AS.andM.S.

41 36 F 4 120/50 117 83 29 0 Nil 115,80 M.S.M. I GoodA.l. 12.9.63 E.D.M. 1A.S.andM.S.42 22 M 4 150240-0 145 102 55 _ Heart block 135l85 M.S.M. 1 Good

AlI. 24.9.63 (temporary) E.D.M. 2and P.S.M. 1-2

43 35 M 4 190,80 173 150 54 - Ploay108 ...12 GoA.]. 3.10.63 complications E.D.M.iandM.I.44 20 M 4 120,50 195 155 56 - Heart failure

AlI. 15.10.63 DiedM.S.andT.I.

1Murmurs graded 1 to 4. M.S.M. -mid-systolic (aortic) E.D.M. early diastolic (aortic) ;P.S.M. pan stystolic (mitral).A.I. = aortic incompetence; M.I. = mitral incompetence TI. - tricuspid incompetence; A.S. = aortic stenosis; M.S. = mitralstenosis.

where in October 1958, when one cusp of a bicuspidvalve was replaced with an artificial leaflet and theother leaflet partially decalcified. This operation was

followed by renal shutdown which required dialysistwice before function returned, but the patient was

moderately improved for a year or so, when anginareturned: later congestive heart failure appeared.Aortic incompetence was then evident, although pres-

sures recorded at the second operation showed a resi-dual gradient of 60 mm. Hg across the valve. The arti-ficial leaflet had shrivelled and its remnant was

completely calcified. Pre-operatively renal functionshowed a reduced creatinine clearance of 63 ml./min.and a raised blood urea of 62 mg./ 100 ml. This patienttolerated the second perfusion and homograft valvereplacement extremely well without evidence of furtherrenal damage.

In case 17 the first operation was performed atGreen Lane Hospital in February 1963. Two of thecalcified cusps disintegrated and were replaced byteflon leaflets (Bahnson, Spencer, Busse and Davis,1960), with a good immediate result but with theappearance within three weeks of aortic incompetencewhich gradually became very severe. At the secondoperation the incompetence was found to be due totearing of one of the teflon cusps near its attachmentto the aorta.

In case 22, surgery had been undertaken at GreenLane Hospital in 1961 when myocardial cooling with-out continuous coronary perfusion was in use. De-calcification of the valve was attempted but in the

time available the cusps were not fully mobilized.Symptoms recurred within eight months and pro-gressed to effort dizziness and paroxysmal nocturnaldyspnoea, and physical signs indicated a return ofsevere aortic stenosis.

In case 15 the initial operation was in May 1963,at which time the ventricular septal defect was closedand the prolapsing right aortic cusp was replaced witha teflon leaflet with an excellent immediate result.Aortic incompetence soon re-appeared, however, andrapidly progressed until the boy was in uncontrollableheart failure. In addition, a continuous fever waspresent and splinter haemorrhages kept appearing inthe nail beds, indicating infection, although this wasnever proven. At the second operation the teflonleaflet still appeared to be functioning well, and theincompetence was due to partial detachment of thenon-coronary cusp where it met the teflon leaflet.

ASSOCIATED LESIONS Clinical assessment suggestedthat two patients in the aortic incompetence group(cases 11 and 14) also had mitral incompetence thatwas too mild to require surgical attention. This wasborne out by pressure measurements taken at opera-tion and by the post-operative course, because aftercomplete correction of the aortic leak the mitralincompetence became inconspicuous.Two patients had large intrapericardial ascending

arch aneurysms in association with aortic incom-petence (cases 10 and 18). In both, in addition toinserting a homograft aortic valve, the aneurysm was

133



/T


ownloaded from


B. G. Barratt-Boyes

completely excised and the aorta reconstituted bydirect end-to-end anastomosis just above the newvalve. In case 10 the aneurysm was syphilitic andcommunicated with the right pulmonary artery. Thisacquired aorto-pulmonary window was not suspectedfor the murmur was not continuous, presumablybecause the valvular incompetence was severe enoughto lower aortic diastolic pressure and prevent a largeleft to right shunt throughout diastole. In case 18the aneurysm was probably the result of a localizedlong-standing aortic dissection.One patient (case 30) was thought to have a

bronchogenic carcinoma producing marked stenosis ofthe right upper lobe bronchus and a right hilar mass.Bronchoscopic biopsy and sputum cytology wererepeatedly negative, and, as the aortic stenosis wassevere, the lung lesion was treated by radiotherapyalone, using the linear accelerator. There was con-siderable clearing of the hilar shadows and improve-ment in the bronchoscopic appearances. His cardiacstatus continued to deteriorate until frank heart failureappeared, and at this stage-10 months after com-pletion of radiotherapy-aortic valve surgery wasperformed without complications. There has been norecurrence to date of pulmonary disease.

Six patients had pre-operative renal damage. Thiswas most marked in case 6, as already detailed, butthe creatinine clearance was reduced to between 60and 70 ml./min. in cases 18, 36, and 37 and toapproximately 80 ml./min. in cases 3 and 41, and theblood urea was also at the upper limits of normal inthese five patients.The six patients in the multivalvular disease group

all had aortic incompetence and mitral valve diseasewith advanced cardiac decompensation. In cases 40 and41 there was some degree of aortic stenosis but withenough incompetence to make valve replacementmandatory. Mitral incompetence was moderate incase 39 and severe in cases 42 and 43, and in two ofthese patients it was the result of bacterial endocarditiswith chordal rupture. Mitral stenosis was severe incases 40, 41, and 44 and was associated with severeorganic tricuspid incompetence in case 44.

CATHETERIZATION In eight patients preliminary cardiaccatheterization was undertaken for a variety ofreasons: in cases 16 and 17 to exclude associatedmitral stenosis; in cases 3, 27, and 35 to confirm thatthe aortic lesion was severe; in three of the patientswith multivalvular disease (cases 40, 41, and 44) in anattempt to quantitate the severity of each lesion.Coronary angiography was not done as associatedcoronary artery disease was not considered a contra-indication to operation for severe aortic valve disease.

TECHNIQUE

Homograft aortic valves were removed within 15hours of death. The age of the donors was limitedat first to less than 45 years, but some more recentvalves, for use in older patients, were taken from

subjects up to the age of 55 years. A full sterile tech-nique was used in the majority of cases, but latterlya number of valves have been taken by the pathologistunsterile and immediately sterilized using beta-propiolactone (Rains, Crawford, Sharpe, Shrewsbury,and Barson, 1956). The valves were then placed innutrient medium similar to that described by Gross,Bill, and Peirce (1949), to which penicillin and strepto-mycin were added, and stored at 40 C. for up to aweek before freeze-drying. They were cultured justbefore being placed in the medium and again onremoval before freeze-drying: if positive they werenot used. After freeze-drying, the outside of the sealedglass vacuum tube was sterilized in an ethylene oxidegas chamber at body temperature. Freeze-dried valveswere reconstituted by placing them in distilled waterfor 30 minutes and then in isotonic saline.The valves were inserted with the aid of a heart-lung

machine of either Melrose or Kay Cross design withmodifications to the venous side of the oxygenatorcircuit (Barratt-Boyes and Yarrow, 1961). Themachine was primed with a mixture of two parts ofheparinized blood (drawn on the morning of opera-tion) to one part of 5% glucose containing 5 g. ofserum albumin per 100 ml. The mediastinum wasopened through a vertical sternal-splitting incision,avoiding entry into either pleural cavity. The arterialcannula was inserted into the external iliac artery bothto ensure as low an arterial line pressure as possibleand to avoid any possibility of contaminating thethoracic field from a groin incision. A single rightatrial cannula was used, except in cases 10 and 18,where both cavae were taped and two caval linesinserted: this was because previous experience hadshown that a single right atrial line is dangerous inascending aortic arch aneurysm where there is a pos-sibility of holing the pulmonary artery. In case 10 itwas fortunate that both caval tapes were in positionas there was an unsuspected aorto-pulmonarycommunication.The patients were cooled to a nasopharyngeal

temperature of 300 C., and perfusion flow was fixedat 2 L./m.2/min. during the cooling phase to allowthe heart to contribute to the circulation in theexpectation that body cooling would be more uniform,and in patients with severe aortic incompetence, toprevent overdistension of the left ventricle fromexcessive retrograde aortic flow during the coolingphase. The aorta was cross-clamped in cases withincompetence immediately there was any suggestionof overdistension of the left ventricle, and in othersat 300 C.; it was then opened transversely about5 mm. above the origin of the right coronary arteryaround at least three-quarters of its circumference.Just before cross-clamping the aorta, a metal-tippedsuction line was inserted through the apex of the leftventricle to empty the left heart. Suction on this linewas maintained by a Sigmamotor pump, and a secondopen-heart line with a separate pump was also avail-able. Because of the importance in long perfusions ofpreventing significant haemolysis, the open-heart

134



/T


ownloaded from



pumps were repeatedly adjusted to avoid excessive airin the lines. Moreover the perfusion flow was main-tained at 2 l./m.2/min. until towards the end of theprocedure when rewarming had been started and thenasopharyngeal temperature had reached 320 C., atwhich time flows were increased to 2.4 l./m.2/min.In addition to keeping flows fixed at 2 L./m.2/min., assoon as the aorta was cross-clamped blood was takenfrom the patient into the machine until right atrialpressures were negative. This manceuvre helped todecompress the heart and lungs, to lessen open-heartreturn, and thus to improve operating conditions.When a single right atrial cannula is used and flowsare fixed at 2 l./m.'/min., it is of course imperativeto have a low venous pressure if the heart is stillbeating, otherwise some of the right atrial return willbe pumped through the lungs and will appear in theleft heart field; in about half the cases the heartcontinued to beat throughout the procedure. Thepatient's arterial pressure was monitored continuously,and in the few cases in which mean arterial pressurefell below 70 mm. Hg during perfusion, blood wasreturned to the patient and flows increased. Otherwiseblood was not returned until the rewarming phase.As soon as the left heart was emptied of blood,

the left and right coronary arteries were cannulated.The blood for coronary perfusion was taken by aseparate line from the arterial end of the oxygenatorand was not separately filtered or cooled. The myo-cardium was thus maintained at a temperature ofapproximately 300 C. Plastic coronary artery lineswere used with metal tips fitted with self-inflatingrubber cuffs of Mayo Clinic design (Fig. 1). Carewas taken to avoid advancing the left coronary can-nula tip too far and so perfusing only one majorbranch of this vessel. These cannulae proved verysatisfactory as blood rarely leaked around them.During subsequent surgical manipulations, however,

4 -.rX A- < 2, *II,

O 1 2~~~~~~~~~~~~...--I

they were liable to displacement unless furthersecured, and this was satisfactorily achieved by a singlestitch through the aortic wall close to each coronaryorifice (Fig. 2). The two ends of the stitch were passedthrough a length of fine rubber tubing and, aftersnuggling the end of this tubing down on to the outeraortic wall, the plastic tubing of the coronary linewas held firmly against the inner aortic wall. The twostitches, one for each coronary line, were left inplace until aortic closure was almost completed. Bothcoronary arteries were separately perfused with aSigmamotor pump at a flow of 150 to 200 ml./min.,with monitoring of line pressures just beyond thepump heads. This pressure averaged 150 to 200 mm.Hg, and the coronary tips were adjusted whenevernecessary to prevent a rise above 300 mm. Hg. Ineight patients the right coronary artery was not per-fused. In case 1, aged 14, the orifice was too small;in cases 12 and 29 it was absent; in case 21 it waspin hole in size due to extensive fibrous thickeningand calcification within the aortic wall from syphiliticaortitis. In the other four patients (cases 17, 18, 36,and 37), with an average age of 62 years, it was pos-sible to cannulate the right coronary orifice acouratelybut not to perfuse the artery as line pressures imme-diately rose precipitately. It is almost certain thatthese arteries were blocked close to their origins. Inthese eight patients flow through the left coronarycannula was increased and blood temperature wasdropped to approximately 27° C.

Placement of the coronary cannulae in the waydescribed took only a few minutes, and further atten-tion to them was seldom required. The aortic valvewas then excised. When fibrous, a peripheral cuspremnant was left: if extensively calcified, the-calciumwas nibbled away carefully from the aortic wall andventricular outflow tract with complete removal of thecusps. The homograft aortic valve was then sutured

3 4

A

xIG. 1. (A) Metal coronary artery perfusion tips. The left coronary tips are made in three sizes with a mushroom-shaped end: the right in two sizes with a basket tip. (B) Large left and small right coronary artery perfusioncannulae. The rubber cuffs, one of which is shown separately in the centre of the photograph, are tied over the metaltips and are automatically inflated, once the tip lies in the coronary orifice, when blood passes through the side holebeneath the cuff.

L

MM.I . ..

135



/T


ownloaded from


B. G. Barratt-Boyes

FIG. 2. Coronary artery cannulae in position at operation. The transverse aortic incision is shown. The retainingsuture around the right coronary cannula is in position anteriorly, but the rubber sling through which the ends ofthe suture have been passed is not yet tightened down onto the outside of the aorta. The left coronary cannula is alsopositioned with its retaining suture snugged tight. The aortic valve has been excised, giving a clear view of theanterior mitral leaflet (m).

B

A NTERIOR

\ 1e a /-if**21sS'U.N/MfNiPUrli i' Th /c*p

hu./ficah(iwhlrt' Ah ch|a..h ' '(tf r l If ca1, i a,

HOST

C POSTERIOR

136



/T


ownloaded from



r.

FIG. 4. Technique of insertion of valve. (A) The first stay suture is in place; (B) the valve is inverteddown into the left ventricle and the three stay sutures are tied. The lower suture line is begun; (C) thevalve is in itsfinalposition and the three highest points of the graft have been attached with the uppersuture line partly completed. The insert shows how the lower suture line becomes buried when thevalve is pulled upwards.

into position below the coronary ostia. The aorticwall was trimmed from the homograft by cuttingaway the aortic sinuses above (Fig. 3A), leaving theaortic ring intact just below the cusps (Fig. 3B). Twocontinuous suture lines were used, one through theaortic ring below the cusps and the second followingthe line of the aortic remnant above each cusp. Un-fortunately, the aortic ring below the cusps, whilefibrous posteriorly, is largely friable muscular tissueanteriorly: for this reason the homograft valve wasrotated through 180° (Fig. 3C) so that the moredifficult anterior sutures, which bite into musculartissue of the patient's septum, passed through thefibrous aortic ring of the graft. Suturing was com-menced by inserting three stay sutures through theaortic ring of the graft beneath the most dependentportion of each cusp (Fig. 4A). These sutures werepassed into the left ventricular outflow region throughcorresponding points below the most dependent por-tions of each of the patient's excised cusps. The valvewas then lowered into position and inverted down intothe cavity of the left ventricle (Fig. 4B) so that theaortic ring presented as the most superficial part ofthe prosthesis; and the three sutures were tied. Itwas then a relatively simple matter to complete sutur-ing the aortic ring of the graft to the host, using the

three stay sutures. In case 39 this suture line wascompleted before it was realized that the graft hadbeen turned inside out. The sutures had to be cutout and inserted again with the graft correctly placed;but on this second occasion the suture line presumablypassed too close to membranous ventricular septaledge for a permanent complete heart block resultedin this patient.When the proximal suture line was completed the

graft was pulled upwards into its correct position(Fig. 4C) and the uppermost points of each cuspattachment were sutured to the aortic wall just abovethe patient's cusp remnant. The upper sutures were thencompleted between these three points as continuoussuture lines, picking up the aortic wall of the graftat the base cf each cusp and, ideally, passing'throughthe patient's cusp remnant. It was not always possibleto use this remnant, because either the homograftvalve tended to lie better a fraction higher than.this,or the line of the remnant was distorted or even bi-cuspid, or in calcified cases the valve had alreadybeen completely removed. In these cases the uppersutures were placed through the aortic wall above thecusp remnant usually at the lowest extension of eachsinus of Valsalva, as shown in Fig. 4C; although,when a bicuspid valve had been excised (as was the

137



/T


ownloaded from


B. G. Barratt-Boyes

case in 10 patients), the homograft valve required acompletely new line of attachment dictated by the lieof its three cusps. With valve placement completed,the aortic incision was closed with a running suturestarted at each angle and carried anteriorly. Providedthis single suture line was kept taut, it seldom leaked,and additional sutures were rarely required. Thecoronary cannulae were left in position until aorticclosure was almost completed and the left heart wasfilled with blood.The stage at which the heart was electrically de-

fibrillated depended upon the excellence of thecoronary perfusion. When both coronary arteries hadbeen perfused, defibrillation was readily accomplishedin every instance before removing the coronary can-nulae and completing aortic closure, since by thisstage the nasopharyngeal and myocardial temperatur&had been returned to 35° C. In these patients, oncethe heart was stable and blood was being ejectedthrough the unclosed portion of the aortic incision,coronary perfusion was stopped, and the cannulaewere removed after releasing the stay sutures aroundthem. Aortic closure was then rapidly completed andthe aortic cross-clamp was released. A large-borehypodermic needle was passed through the front ofthe suture line to release any air trapped above thevalve, and the left ventricular suction line was thenremoved. When, however, only the left coronary arteryhad been perfused, rewarming above 32°C. wasdelayed until aortic closure was completed, and theaortic clamp was released with the heart still fibrillat-ing. This allowed optimum coronary perfusion fromthe machine while rewarming was continued to about34°C. when defibrillation was attempted and wasimmediately successful in all but case 37.In this patient, although five to 10 co-ordinatedbeats followed each attempt at defibrillation,ventricular fibrillation kept recurring over a periodof 45 minutes. Fortunately, with the aid of variousdrugs, including procaine amide, electrical defibrilla-tion was finally successful and a stable sinus rhythmwas maintained thereafter.

Bypass was then stopped, adding blood from themachine until the venous pressure reached 12 to 15mm. Hg. After decannulation the pericardium wasclosed loosely over one soft rubber catheter, and asecond catheter was placed in the anteriormediastinum.

In the six patients with multivalvular disease, whilethe homograft valve was inserted as described, asternal-split chest incision was not satisfactorybecause it did not permit proper exposure of the mitralvalve. Thus in case 39, although the left atrium waswidely opened from the right side, it was a rathersmall chamber and the mitral valve could not beexposed. A ruptured chorda, clearly seen from theaortic approach, could not therefore be repaired. Incase 40, a right lateral thoracotomy in the fifth inter-costal space allowed easy cannulation of the coronaryarteries through the usual transverse aortotomy andexcellent exposure of the mitral valve, but suturing of

the homograft aortic valve was extremely difficult.In cases 41, 42, and 43, a bilateral anterior thoraco-tomy was used and allowed good exposure of bothvalves. This incision is not favoured, however, inpatients with advanced heart disease because of thegreater blood loss and greater difficulty with ventila-tion post-operatively. In case 44, a new incision wastried through the anterior half of the right fifth inter-costal space, across the sternum at this level and verti-cally upwards along the left sternal edge, inside theinternal mammary vessels, where the second, third,and fourth left costal cartilages were divided withoutopening the left pleura. This right antero-lateralapproach, with sternal division and the patient posi-tioned slightly obliquely, gives surprisingly goodexposure of aortic, mitral, and tricuspid valves withoutthe disadvantages inherent in a bilateral anteriorthoracotomy and is currently the approach of choicein multivalvular disease.

In all six patients the mitral valve was approachedfrom the right side and, when stenotic, was openedwith a dilator and a knife. In case 44, because ofvalvular calcification and fibrosis, incompetence wasproduced, and a Starr-Edwards ball valve (1961) wastherefore inserted in the mitral orifice. The tricuspidincompetence was corrected by a tricuspid annulo-plasty. Of the three patients with mitral incompetence,correction was not practicable, as already noted, incase 39; in case 42, complete correction was achievedby repairing a ruptured chorda in a manner describedpreviously (Barratt-Boyes, 1963) and partially closingthe adjacent commissure; and in case 43, a Starr-Edwards ball valve was inserted.The duration of heart-lung bypass and of coronary

perfusion in each patient is listed in the Table. Case 10required a three-hour bypass and coronary perfusionfor 153 minutes, but an ascending arch aneurysmwas also resected in this patient. In most cases withaortic incompetence alone, bypass time was just undertwo hours; in cases with calcific stenosis it averagedslightly more than two hours ; and in the multivalvulargroup a two-and-a-half- to three-hour perfusion wasthe rule. The serum haemoglobin level was measuredspectrophotometrically at half-hourly intervalsthroughout the perfusion. The level in machine bloodafter perfusion was significantly raised in only fourpatients: to approximately 150 mg. in cases 3, 4, and20, and to 304 mg. in case 22.

Post-operative management did not differ from thatfor other bypass cases. Venous and arterial pressureswere monitored continuously for the first eight to12 hours, and moderate elevation of the venous pres-sure was maintained. Oxygen tents were seldom used,and intermittent positive pressure respiration was neverrequired. Additional base was not given to any patientas serial pH and carbon dioxide values were satis-factory. Digitalis therapy was commenced or continuedwithin 12 hours of warding. The administration ofcrystalline penicillin, one million units six-hourly, wascontinued for two weeks, and of streptomycin, 1 g.daily, for four days. Anticoagulants were not used.

133



/T


ownloaded from



Intravenous mannitol was given in eight cases as aprophylactic measure against renal failure, andoliguria did not occur in these or in any other patient.In cases 4 and 22 the only indication for its use wasa raised serum haemoglobin level above 100 mg. atthe end of bypass (Porter, Sutherland, McCord, Starr,Griswold, and Kimsey, 1963), while in the remainingsix there was evidence, already detailed, for pre-operative renal damage.

RESULTS

Forty-one of the 44 patients survived operationwith an apparently normally functioning homo-graft valve. One death occurred in each of thethree groups.Case 5 with aortic incompetence died within six

hours of warding. In this patient, three pints ofblood drained from the mediastinum during thefirst three hours, and there was a large haematomabeneath the right iliac fossa incision for insertionof the arterial cannula. The wounds were reopenedand the bleeding points were controlled, but anhour or so later, when the blood pressure andother vital signs were satisfactory and bleedinghad stopped, she suddenly vomited blood, inhaleda considerable quantity, and died soon after.Necropsy showed bl0od clot in the trachea andright main bronchus but no blood in the peripheralbronchi, and there was total collapse of the rightupper lobe and partial lung collapse elsewhere.The stomach contained a small acute ulcer andblood, and it is assumed that this 'stress' ulcerwas the source of the vomited blood. The homo-graft aortic valve was accurately placed and com-petent, and the other heart valves were normal.

Case 28, with calcific aortic stenosis, died onthe third post-operative day from a large full-thickness infarction of the left ventricular walldue to compression of the left coronary artery andits two main branches by haemorrhage which layoutside the media. The mechanism of this bleed-ing is obscure as there was no medionecrosis ofthe wall or separation of the media, and the condi-tion was therefore not a 'dissecting aneurysm';nor could any communication be found betweenthis blood and the vessel lumen.Case 44, with aortic and tricuspid incompetence

and mitral stenosis, died on the evening of theday of surgery. This patient had very severe diseasewith gross cardiomegaly (cardio-thoracic ratio75%). After replacement of the aortic and mitralvalves and tricuspid annuloplasty the only valvulardisease was slight tricuspid incompetence. Hiscourse was quite satisfactory with a normal bloodpressure until 10 to 15 minutes before death when

the direct recorder showed a rapid rise in venouspressure followed by ventricular fibrillation.Necropsy showed that the new valves were in anaccurate position and the annuloplasty was intact;it did not help to elucidate the cause of death.

Complications in the surviving patients are listedin the Table. Pulmonary embolism occurred threetimes and was treated successfully by severalweeks' administration of anticoagulants. Twopatients had a typical post-perfusion syndrome(Kreel, Zaroff, Canter, Krasna, and Baronofsky,1960) and one a post-cardiotomy syndrome(Robinson and Brigden, 1963), and in case 43, whohad bilateral anterior thoracotomies, a tracheo-stomy was required for pulmonary complications.Atrial fibrillation appeared post-operatively forthe first time in about one-fifth of the patientsbut usually reverted spontaneously to sinusrhythm. One patient (case 20) developed ventri-cular tachycardia followed by ventricular fibrilla-tion four days post-operatively. This responded toprompt external cardiac massage and defibrillationwithout any brain damage or other complication.

Significant post-operative morbidity occurredonly twice. In the first patient (case 6), bacterialendocarditis developed, the temperature showinga progressive rise from the ninth to the twentiethday when a coagulase-negative Staphylococcusaureus was cultured from the blood. Fortunatelythis organism was fully sensitive to penicillin;treatment was recommenced and continued forfour weeks in a dose of one million units four-hourly; sodium methicillin, 1 g. four-hourly, wasalso given. The temperature fell quite quickly andthere were no further positive blood cultures.During this period the patient had one attack ofblindness in the right eye, lasting for about oneminute, and subsequently had three similar epi-sodes, the last occurring three months after opera-tion but without any permanent visual field defect.As the temperature was then normal and bloodcultures were consistently negative it was assumedthat these were platelet emboli, and he was treatedwith anticoagulants for one month. An intervalof seven months has now passed without furthereye symptoms and it is nine months since anti-biotics were stopped. The homograft valve con-tinues to function normally and the patient isasymptomatic. One further patient (case 35) hashad two momentary attacks of blindness whichmay be due to platelet emboli, for in this patientthere has been no evidence of infection. Thesecond patient with significant morbidity is case

39 who, in addition to uncorrected mitral incom-petence, has a persistent surgically induced com-

139



/T


ownloaded from


B. G. Barratt-Boyes

A BFIG. 5. Case 1. Postero-anterior radiographs of the chest (A) immediately before operation and (B) 11 months later.

plete heart block, but despite these features hispost-operative course has been smooth. An externalcardiac pacemaker was used almost continuouslyin the early stages and was later replaced by animplanted unit. Three other patients (cases 29,36, and 42) had temporary heart block which wasprobably unrelated to the suturing of the valve.The follow-up period is relatively short but is

now 12 months or more for the first five patients(cases 1 to 4 and 22). The aortic valve homograftis functioning normally in all these patients, andcases 1, 3, 4, and 22 are asymptomatic. The pre-operative chest radiograph and electrocardiogramin case 1 may be compared with those takenapproximately one year later (Figs. 5 and 6). Atthe time of operation this 14-year-old girl was insevere congestive heart failure from rapidly pro-gressive aortic incompetence which followed bac-terial endocarditis. One year later there was a faintejection systolic murmur, a faint short early dia-stolic murmur, and normal splitting of the secondheart sound; the chest radiograph and electro-cardiogram were normal and the blood pressurewas 105/70 mm. Hg. These findings are representa-tive of all the surviving patients with a sufficientperiod of follow-up except case 2. In these otherpatients the symptoms have rapidly lessened, andin the aortic incompetence group in particularthe heart size has rapidly diminished. About halfthe patients have a faint short early diastolicmurmur maximal along the left sternal edge. This

murmur has seldom been heard until about oneweek after operation and in some cases has sub-sequently tended to diminish in intensity. In nopatient has there been any other clinical evidenceof residual aortic incompetence. An ejectionsystolic murmur of mild to moderate intensity hasinvariably been present from the day of operationbut has also tended to disappear. In case 23 thismurmur was associated with a gradient of 36 mm.Hg across the homograft valve (Table). Thispatient had a rather small left ventricular outflowand aortic ring, and the homograft valve was alittle large. Despite this, however, progress hasbeen very favourable with complete loss of symp-toms of dyspnoea, angina, and syncope, and witha rapid return of the heart size and electro-cardiogram towards normal (Fig. 7). Case 2requires special mention because, although theaortic incompetence has been cured (B.P. 115/90mm. Hg), 10 months later there is still no reduc-tion in cardiomegaly or left ventricular hyper-trophy, and the patient has recently been admittedto hospital because of recurrent congestive heartfailuire. In the absence of evidence for residualvalvular disease or active rheumatic fever, myo-cardial disease seems the probable cause of hisdisappointing progress.There have been two unexpected deaths about

four months after operation. Case 7, a 62-year-oldwoman, had a history of severe angina decubitus,nocturnal dyspnoea, and orthopnoea. The aortic

140

.4

1. "" -, bk..:



/T


ownloaded from


Homograft aortic valve replacement in aortic incompetence and stenosis14

VR VL,..

ts-

VI 2 V3

VR

LM

VI V

V4

I

-NV

Vs6

Vt

is V6

FIGr. 6. Case 1. Electrocardiograms (A) before surgery, patient receiving digitalis; (B) I11months after operation. The ST and T changes, due to left ventricular hypertrophy anddigitalis effect, have disappeared.

ik

141

.j



/T


ownloaded from


B. G. Barratt-Boyes

VP

V7

iI

;004 too `14

vW / X

V...A

XI

N;.r ...

I

V4 V5

VF5

AV.- I

I

V

04 Emor, ll..00..

S'I

FIG. 7. Case 17. Electrocardiograms (A) before surgery, patient not receiving digitalis; (B) nine

months after operation. The ST and T changes reflecting left ventricular hypertrophy have

lessened considerably, although the voltage of T in V6 and VF has not yet returned to normal.

Vt

la

V V4 ~7T7\i V v 4 Ili _/

m

l~~~~~~~~~~~~~~~~~~~~~~~~~---

142

wjol.pmw%i

4-4I

..1

0%4.:: oq*4...ewvi

I

0%.j ^NO.1,

rw^-I

t

47

I

v C'

.,^ W%. .el'. .,.4



/T


ownloaded from



FIG. 8. Case 7. Postero-anterior radiographs of the chesteffusion; (B) three months later.

(A) two weeks before operation; note right basal pleural

-F v_ ,2

't.~~~~~~~~~~~~~~~~~~~~~~~~~~O

FIG. 9. Case 7. Myocardial mural artery showing eccentric fibro-elastic intimalthickening. Verhoeff elastic tissue stain plus Van Gieson x 125.

143



/T


ownloaded from


FIG. 10. Case 7. Homograft valve viewed from above.Aorta dividedjust above healed transverse aortotomy.

FIG. 11. Case 7. Sagittal sections through two of the cusps.(A) Posterior cusp with host aortic wall above and anterior

.*so_ .r t mitral leaflet (MV) below. This portion of the graft israther bulky, partly because it contains muscular tissuebelow the leaflet. Elastic tissue of the aortic portion of thegraft (ETG) lies inside host elastic tissue (ETH). A thickintimal sheath of young fibrous tissue (IFS) starts at thebase of the cusp and extends over the upper suture lineonto the aortic wall. A similar layer covers a calcificnodule (arrow) presenting on the ventricular surface of thethickened mitral leaflet. Verhoeff elastic tisgue stain plusvan Gieson x 2. (B) Anterior cusp. The graft is entirelyfibrous and less bulky at this point. The central defectseparates graft from host and is lined by young fibroustissue. A silk suture is seen cut transversely near the loweredge of the graft. Th( intimalfibrous sheath above the cuspis so thin here that it cannot be clearly seen. Haematoxylinand eosin x 2.

FIG. 10.

V*M i

1:~~~o .:..9.

p

A A4M:S~~~~~~~~~~~~~~~~~~~~~~~....M.l:>

,~~~ ~ ~ ~ ~ ~ ~ ~ ~.....V

FIG. 11.

144' B. G. Barratt-Boyes



/T


ownloaded from


FIG. 12. Case 7. Base of homograft cusp and acellular aortic wall of graft at right. The intimal fibroussheath (IFS) lies above the cusp and ends abruptly at its base. Verhoeff elastic tissue stain plus vanGieson x 45.

x45X . 4 , X

. f,V>t>0BvP ..

FIG. 13. Case 7. Upper suture line of graft under higher magnification. Black-staining host elastic tisskeat right, well preserved graft elastic tissue at left, and transversely cut silk suture between. The youngfibrous tissue of the intimalfibrous sheath lies above with the vessel lumen in the upper left corner.Cellular reaction, consisting of macrophages and less numerous lymphocytes and fibroblasts, is mostmarked around the suture material and is absent in the graft-host interface deep to the suture.Verhoeff elastic tissue stain plus van Gieson x 125.

.,Wogoe

TUNC,40,t.,

A "i,Of



/T


ownloaded from


B. G. Barratt-Boyes

valve was bicuspid, severely incompetent, andheavily calcified. Her post-operative course wasuncomplicated, and when last examined threemonths after operation there had been a dramaticreduction in the heart size (Fig. 8) and she wassymptom-free. The blood pressure was 135/85mm. Hg, and faint aortic systolic and diastolicmurmurs were heard. Death occurred two weekslater and was immediately preceded by anginal

pain which came on at rest. Necropsy showeddeath to be due to acute left ventricular failurewith marked pulmonary congestion and oedema.The left ventricle showed residual hypertrophy andan old antero-lateral infarction. The coronaryartery branches showed some atheromatousnarrowing but were macroscopically patent.Microscopically, however, the majority of thesmall mural arteries were narrowed by eccentric

_<g Wr...,...

4--- _,?:C lEsNs..:

-_ ?, :iW.U.1. | x |_ y M .

- X_u i p i ;S SS S a_s_= | |

ql L_i_ ._

l %

^llb _iii #

M E' s _ s g:| s ,\l \ _§:.: i.|#'_s l l _ ...... -:.:;P

>3 3 i ......... ^ #oF M __::;iiw: a '.'S:iiPi.

_ d'l,:.B l]<.n ]..j |'t- i.:tw'=] -.y.:. <s.

= s__ Pi..;s:it:SF

.S__ 'E._x____

^sr.___s___i_____maX

__iX_w_,$AB_

a_w__S_.{D___

F___w__:_

* _r

WE3t__Drs P.rs.. a1B i_. 04 :lL

='_IIYoSao_W..; :..t..

FIG. 14. Sagittal sections through the aortic valve of a 64-year-old woman.(A) Aortic wall lies above with elastic tissue tailing off well above the base of theaortic leaflet. Below is the anterior leaflet of the mitral valve. Verhoeff elastictissue stain plus van Gieson x 2. (B) Aortic cusp under higher magnification.In both photographs the ventricular aspect of the cusp with its major elastic tissuecomponent lies to the left. The aortic aspect is predominantly collagenous.Verhoeff elastic tissue stain plus van Gieson x 45.

146



/T


ownloaded from


1R_& II I.....s..

FIG. 15. Case 9. Homograft aortic valve laid open. The silk sutures above and below the graftand in the aortotomy above the graft are hidden and both coronary ostia are well seen. There aretwo small blind pockets beneath the lower suture line. The leaflets appear nonnal.

FIG. 16. Case 9. (A) There is extension of young intimal fibrous tissue (IFS) onto the ventricularaspect of this cusp. Splitting of the cusp is a processing artefact. Above the cusp the intimalfibroussheath (IFS) is thick but does not extend on to the cusp. (B) Homograft leaflet showingnormal architecture four and a half months after transplantation without any cellular reaction.Compare with Fig. 14. Verhoeff elastic tissue stain plus van Gieson x 45.



/T


ownloaded from


B. G. Barratt-Boyes

fibro-elastic intimal thickening (Fig. 9), similar tothat seen in pulmonary vessels in low flow statesand thrombo-embolic disease, and there werenumerous irregular fibrous tissue scars of varyingsize scattered throughout the septum and leftventricle. Coronary insufficiency and death werepresumed to result from these changes. The homo-graft valve could have been mistaken for normalas the silk suture lines were covered and smoothand the cusps were thin, semi-transparent, andpliable and easily approximated one another (Fig.10). The microscopic features are illustrated inFigs. 11, 12, and 13. The aortic wall and ventri-cular muscle of the homograft had lost theircellular elements, but the elastic and collagen fibreswere well preserved. Cellular reaction betweengraft and host was virtually confined to pointsaround the sutures and to the intimal surface ofthe graft particularly above the cusps. Here therewas a surprisingly thick sheath of young fibroustissue which smoothed over the crevices of theupper suture line and became continuous abovewith host intima. Below, this new tissue stoppedabruptly at the base of the cusps. The cusps werecompletely free of any reaction, and their archi-tecture was indistinguishable from that of a

normal valve from a patient of approximately thesame age (Fig. 14).The second late death was that of case 9, a 45-

year-old butcher, who developed aortic incom-petence after an attack of bacterial endocarditisin 1952. Six years later angina appeared and thenprogressed until it came on after only minor effort.The aortic valve was tricuspid and severely calci-fied. Post-operatively he developed atrial fibrilla-tion which converted spontaneously to sinusrhythm, but there were no other incidents. Hereturned to work within three months and was

free of symptoms. When last examined the bloodpressure was 135/85 mm. Hg, and there was a

faint aortic ejection murmur but no diastolicmurmur. He died suddenly four and a half monthsafter surgery while shovelling gravel into a con-

crete mixer. As in the previous patient necropsyshowed marked pulmonary oedema and residualleft ventricular hypertrophy, and again, while allcoronary arteries were macroscopically patent withunimportant atheroma, there was the same wide-spread fibro-elastic intimal thickening of the smallmural arteries and numerous scattered areas ofmyocardial fibrosis, indicating diffuse coronary in-sufficiency. In the absence of any other explana-tion this is presumed to have been the cause ofdeath. The homograft valve was obviously com-

petent and looked normal (Figs. 15 and 16) except

for two small blind pockets beneath the proximalsuture line where the suture was either incompleteor had torn away over a distance of 3 mm. Inthis patient the intimal sheath of young fibroustissue extended onto the under surface of onecusp for a short distance (Fig. 16).

DISCUSSION

This series of 44 patients is too small and thefollow-up too short to permit firm conclusions,but the results suggest that the homograft aorticvalve is a very satisfactory type of replacementand they have encouraged us to continue withthis operation. Over a period of four and a halfmonths homograft valve leaflets have shown nomicroscopic change in structure, and 15 monthsafter insertion they continue on clinical groundsto function satisfactorily without evidence of anyprogressive change. These findings strongly supportthe earlier beliefs of Gordon Murray (1956), whichare supported by the satisfactory six-year follow-up of his first patient in whom a homograft valvewas inserted in the descending thoracic aorta(Kerwin et al., 1962). The absence of any micro-scopic evidence of rejection of the valve cusps inour two patients leads to the hope that the elasticand collagenous tissue of the cusp, which isnourished entirely by the blood in contact with itssurface, will function permanently.

Selection of patients for homograft valvereplacement has differed a little in the variousdisease categories. In aortic incompetence thisoperation was considered the treatment of choiceafter the first three advanced cases made unevent-ful recoveries, and in rheumatic and syphilitic casesno other form of valve replacement has since beenused. Pending the results of further observation ofthe long-term behaviour of these valves, however,surgery has to date been reserved for those withgross incompetence. In calcific aortic stenosisdecalcification of the valve has as a routine beentried first, and homograft replacement has beenreserved for patients whose valves defied attemptsat decalcification. While we now regard homograftreplacement as the method of choice in any patientwhose valve cannot be decalcified, in the earlierstages of the period under review single or multiplecusp replacement with teflon leaflets (Bahnsonet al., 1960) was also used occasionally. In themultivalvular group surgery has been advised onlywhen the patient was not controllable medically.It should be emphasized that in selecting thepatients in this series none has been refused opera-tion because of the severity of aortic valve disease,

148



/T


ownloaded from



nor has a history suggestive of previous myocardialinfarction been considered a contra-indication.Moreover, age alone has not been viewed un-favourably provided the patient was well preserved.The oldest patient in this series was 66 years, butin the same period two others, aged 68 and 71years, have had their aortic stenosis relieved with-out the necessity for valve replacement, and allthree of these patients have made very satisfactoryprogress.The aseptic technique used to obtain valves in

the mortuary within 15 hours of death has beenexacting and time-consuming, and it is certainlyeasier when the pathologist collects the valveduring the necropsy examination and has it steril-ized by beta-propiolactone. While seven suchvalves have now been inserted in this series andhave functioned well for up to three months, itseems preferable at this stage to obtain the valvessterile in case the sterilization process alters thehost reaction to the graft. Once a sterile graft isobtained, freeze-drying has been accepted as thebest method of long-term storage; but unfor-tunately it takes 30 minutes to reconstitute afreeze-dried valve and, if the principle is adoptedof deciding to use a homograft only when decalci-fication has failed, this delay is unacceptable. Ifthis practice is followed, one valve of 26 to 28 mm.outside diameter, a size suitable for almost alladults, must always be kept stored in nutrientmedium at 4°C., so that it is immediately availablewhen the decision to use a valve is made.When inserting the valve, perfusion time and

aortic cross-clamping may both be prolonged asthe preliminary removal of a calcified valve takestime, and the suturing can be awkward becauseof the confined space. The perfusion techniqueused in these cases was designed to provide goodoperating conditions with an adequate flow anda minimum of haemolysis and has proved verysatisfactory. Coronary artery perfusion is certainlymandatory and should be continuous in bothcoronary arteries. This requires cannulae thatneither leak nor obstruct the field. The flexibleplastic coronary artery lines of Mayo Clinic designwith small self-inflating cuffed tips fulfil thesecriteria admirably, provided they are given addi-tional stability by special retaining sutures whichkeep them in place throughout the suturing of thevalve and until closure of the aortic incision isalmost completed. Occasionally there was slightleakage around the left cannula tip but this wasseldom severe enough to obscure the field. WVhenboth coronary arteries are continuously perfusedwithout any leakage or other technical problems,it has not seemed necessary to lower the myo-

cardial temperature below 30°C. Moreover, thisexperience seems to indicate that when the rightcoronary artery is pathologically narrowed or evenoccluded adequate perfusion of the left coronaryartery at a temperature of 27°C. is usually suffi-cient to protect the myocardium for periods aslong as 100 minutes. Complications arising fromcannulation of the coronary arteries in the waydescribed have been absent except in case 28 wheredeath was the result of bleeding around the leftcoronary artery. Even in this patient, however,there is no clear proof that this was caused by thecannula as there was no tear in the vessel wall.In some of the earlier cases the heart was keptbeating during insertion of the valve in the beliefthat coronary perfusion was more adequate thanwhen ventricular fibrillation was present, but, sincea contracting ventricle may increase the technicalproblem of suturing the valve, and since this groupdid no better post-operatively, this practice wasabandoned.The method used to suture the valve in position

is based on the assumption that two suture linesare necessary for security. The proximal sutureline below the cusps is not very difficult to insertonce the graft is turned down into the ventricle;and it is an advantage that it is buried when thegraft is pulled upwards into its correct final posi-tion. The additional segment of aortic wall andventricular muscle which must be left attachedto the graft below the cusps adds slight extrabulk and may be the cause of a minor gradientacross the valve at the end of operation. Thisgradient has never proved significant, however,and as this portion of graft presumably atrophiesonce union is complete, it is of only temporaryimportance. This suture line lies just above theconduction bundle, which was permanentlydamaged in one patient. This error can be avoidedby keeping the sutures away from the lowermargin of the membranous interventricularseptum. The second and distal suture line liesabove the cusps and often passes partly or com-pletely through the host aortic wall rather thanthe cusp remnant. Both proximal and distal suturelines are continuous stitching, as this avoidsnumerous knots and free silk ends, which arepotential sites for infection, and also serves todistribute stresses evenly. This is particularlyimportant when the valve removed is bicuspid, forthen the upper suture line of the tricuspid pros-thesis cannot follow the line of the excised cuspsbut crosses the thin wall -of the sinuses of Valsalva.Our experience makes, it quite clear that this is notdangerous, at least when using homograft valves,and that teflon pledgets outside the aortic wall are

149



/T


ownloaded from


B. G. Barratt-Boyes

not necessary as there has never been leakagethrough the aortic wall and no graft has becomedetached.The three hospital deaths in these 44 patients,

two-thirds of whom had grade 4 symptoms andsigns of very advanced disease, indicate that therisk of operation is small. Although about half thepatients have a short faint aortic early diastolicmurmur as well as an insignificant ejection systolicmurmur, no patient appears to have significantaortic incompetence or stenosis on clinicalgrounds. In one patient a positive blood culturewas obtained about three weeks after operation,and it seems that this infection, due to a penicillin-sensitive Staph. aureus, has been completelyeradicated by one four-week course of antibiotics.When it is realized that infection on a plasticaortic prosthesis is virtually incurable, this is a

significant and most encouraging achievement.The two deaths approximately four months after

operation are disappointing, but neither can beattributed to aortic valve failure. In both patientsdeath was probably due to coronary artery diseasewith widespread fibro-elastic intimal thickening inthe smaller mural branches and secondary patchyfibrosis of hypertrophied myocardium. Theselesions were an unexpected finding and theirincidence and aetiology are being studied further.

SUMMARY

The aortic valve has been totally replaced by a

homograft valve in 21 patients with incompetence,17 with calcific stenosis, and six with multivalvulardisease in which two or three valves were dealtwith at the same operation. All the patients hadsevere valve lesions and 28 of the 44 patients hadsymptoms and signs of very advanced disease.Forty-one patients survived and have a normallyfunctioning aortic valve up to 15 months later.Two of the homograft valves have been

examined at necropsy three and a half and fourand a half months after insertion, and, in both,the architecture of the cusps was normal.

Significant morbidity has been confined to post-operative endocarditis, rapidly cured bya singlecourse of antibiotics in one patient, and topermanent complete heart block in another.The majority of the valves have been secured

sterile and stored in nutrient medium or by freeze-drying. Heart-lung bypass was conducted at30°C.with continuous perfusion of both coronaryarteries using plastic coronary cannulae with self-sealing metal tips. The method of inserting thevalve in a subcoronary position has been describedin detail.

The results suggest that homograft aortic valvereplacement is the method of choice in the treat-ment of severe aortic incompetence and of mostexamples of severe calcific aortic stenosis.My thanks are due to my cardiological colleague,

Dr. J. B. Lowe, for the medical management of thesepatients, and to Dr. Judith Smith for the pathology.

REFERENCES

Bahnson, H. T., Spencer, F. C., Busse, E. F. G.. and Davis, F.W. Jr.(1960). Cusp replacement and coronary artery perfusion inopen operations on the aortic valve. Ann. Surg., 152, 494.

Barratt-Boyes, B. G. (1963). Surgical correction of mitral incom-petence resulting from bacterial endocarditis. Brit. Heart J.,25, 415.

-, and Yarrow, S. (1961). A combined venous reservoir anddefoaming chamber for use in a heart-lung by-pass circuit.Aust. N.Z. J. Surg., 30, 260.

Beall, A. C. Jr., Morris, G. C. Jr., Cooley, D. A., and De Bakey,M. E. (1961). Homotransplantation of the aortic valve. J. thorac.Surg., 42, 497.

Brewin, E. G. (1956). The use of tissue transplants in the surgery ofcardiac valvar disease-an experimental study. Guy's Hosp. Rep.,105, 328.

Gross,R. E., Bill, A. H. Jr., and Peirce, E. C. (1949). Methods forpreservation and transplantation of arterial grafts. Observationson arterial grafts in dogs. Report of transplantation of preservedarterial grafts in 9 human cases. Surg. Gynec. Obstet., 88, 689.

Kerwin, A. J., Lenkei, S. C., andWilson, D. R. (1962). Aortic-valvehomograft in the treatment of aortic insufficiency: Report ofnine cases, with one followed for six years. New Engl. J. Med.,266, 852.

Kreel, I., Zaroff, L. I., Canter, J. W., Krasna, I., and Baronofsky,I. D. (1960). A syndrome following total body perfusion. Surg.Gynec. Obstet., 111, 317.

Lam, C. R., Aram, H. H., and Munnell, E. R.(1952). An experimentalstudy of aortic valve homografts. Ibid., 94, 129.

McGoon, D. C., Mankin, H. T., and Kirklin, J.W. (1963). Resultsof open-heart operation for acquired aortic valve disease. J.thorac. Surg., 45, 47.

Murray,G. (1956). Homologous aortic-valve segment transplants as

surgical treatment foraortic and mitral insufficiency. Angiology,7, 466.

Porter, G. A., Sutherland, D.W., McCord, C.W., Starr, A., Griswold,H. E., and Kimsey, J. (1963). Prevention of excess hemolysisduring cardiopulmonary bypass by the use of mannitol. Circu-lation, 27, 824.

Rains, A. J. H., Crawford, N., Sharpe, S. H., Shrewsbury, J. F. D.,and Barson, G. J. (1956). Management of an artery-graft bankwith special reference to sterilisation by beta-propiolactone.Lancet, 2, 830.

Robinson, J., andBrigden,W. (1963). Immunological studies in the

post-cardiotomy syndrome. Brit. med. J., 2, 706.Ross, D. N. (1962). Homograft replacement of the aortic valve.

Lancet,2, 487.Starr, A., and Edwards, M. L. (1961). Mitral replacement clinical

experience with a ball-valve prosthesis. Ann. Surg., 154, 726.

ADDENDUM

Since this paper was submitted case 2 has died.At necropsy the homograft aortic valve, which hadbeen inserted 14 months previously, looked similarto the spacimens obtained from cases 7 and 9. Thecusps remained competent, pliable, and completelyfree of calcium, and clinically there was never any

post-operative aortic diastolic murmur in thispatient. Death was undoubtedly due to extensivemyocardial fibrosis secondary to obliteration ofthe mural branches of the coronary arteries fromfibroelastic intimal thickening identical with thatdescribed in cases 7 and 9.A further 13 patients have now been operated

upon (four with aortic incompetence, eight withcalcific aortic stenosis, and one with multivalvulardisease), and one has died.

150



/T


ownloaded from


homograft aortic replacement in aortic incompetence and ... · homograft aortic valve replacement...

Documents