Porcine Valve Durability in Children

Transcription

1 Porcine Valve Durability in Children Jeffrey M. Dunn, M.D.* ABSTRACT Calcification of porcine heterograft valves may be greater in the pediatric than in the adult population. This multicenter study evaluates the long-term actuarially determined durability of porcine valves in children less than 21 years old. We evaluated 227 late survivors of porcine valve implantation with 47 aortic valves, 67 mitral valves, 14 tricuspid valves, 14 pulmonary valves, 77 pulmonary conduits, and 8 apicoaortic conduits. In this series, 20 porcine valves degenerated, most in the aortic and mitral positions. At 5 years 40% of aortic, 40% of mitral, 100% of tricuspid, 100% of pulmonary, and 89% of pulmonary conduits remained functional. In this series, calcification and dysfunction occurred significantly faster (p < 0.05) in younger compared with older children and in left-sided or systemic positions (aortic and mitral) compared with right-sided positions (tricuspid, pulmonary, and pulmonary conduit). Improvements in porcine bioprosthetic valves, including glutaraldehyde stabilization [l] and flexible valve stents [21, have made these heterografts attractive prostheses. The greatest experience has been in the adult population [3], where durability has remained excellent. The advantages of central flow, no need for anticoagulation, and relatively low profile make the porcine bioprosthetic valve appear to be an excellent prosthesis in the pediatric population [4-71. Sporadic reports of degeneration of porcine bioprosthetic valves in children have appeared since 1977 [ Geha and associates [12] reported on 5 such dysfunctions in the pediatric population, and suggested that heterograft fail- From the Section of Pediatric Cardiac Surgery, St. Christopher's Hospital for Children, Temple University, 5th and Lehigh Sts, Philadelphia, PA Presented at the Seventeenth Annual Meeting of The Society of Thoracic Surgeons, Jan 26-28,1981, Los Angeles, CA. *With the collaboration of the Pediatric Cardiac Surgery Study Group and the statistical assistance of Mrs. Chari Otis. ure rate is greater in children than in adults. Because the experience with porcine bioprosthetic valves is limited at any one pediatric cardiac center, this multicenter study was undertaken to evaluate the late actuarially determined durability of these valves in children. Methods The experience with porcine bioprosthetic valves at twelve pediatric cardiac centers was collated. The participating surgeons and their respective cardiac centers are listed on pages All data were collected on a standard questionnaire, with each surgeon responsible for collecting the data at his center. The patients were divided into six groups based on the site of valve insertion: Group 1, aortic (47 patients); Group 2, mitral(67 patients); Group 3, tricuspid (14 patients); Group 4, pulmonary (14 patients); Group 5, pulmonary conduits (77 patients); and Group 6, apicoaortic conduits (8 patients). For statistical purposes, all left-sided or systemic atrioventricular valves were placed in Group 2, the mitral valve group. Thus, two patients with lev0 (L) transposition of the great vessels and tricuspid valve replacement are included in the mitral group since the tricuspid position functions as a systemic atrioventricular valve in this congenital defect. The total experience with porcine bioprosthetic valves in patients 21 years of age and less was reported by each center with the exception of 'Columbia-Presbyterian, which excluded the experience with the pulmonary conduit. Since we are primarily concerned with the long-term durability of porcine valves, only operative survivors are included in this series. We have defined the operative survivors as patients who were discharged from the hospital or survived at least 30 days postoperatively. To eliminate confusion between valve degeneration secondary to infection and chronic intrinsic degeneration, all patients with documented by The Society of Thoracic Surgeons

2 358 The Annals of Thoracic Surgery Vol 32 No 4 October 1981 bacterial endocarditis of the implanted heterograft valve were also eliminated from this series. Actuarial analysis and life tables were developed employing methods outlined by Grunkemeier and Starr [17]. A t test was employed to statistically compare groups and subgroups [MI: t = (Pk, - PkJd(Sk,2 + Pkz2)/(N'k, + N'kz) where Pk, = kth year survival rate, i.e., the fraction of those entering the study expected to survive at least the kth year; N'k = effective number of valves exposed to risk during the k th year; and Sk = standard deviation of survival rate for kth year. For actuarial analysis, valves removed for reasons other than degeneration and valves in patients who died for reasons other than valve dysfunction were treated as "removed" from evaluation during their last "at risk" year. Results Group 1 PATIENT DATA. Forty-seven patients were longterm survivors of aortic valve replacement. One patient had simultaneous mitral valve replacement. The age range was 3.8 to 21.8 years, and the mean and median ages at implant were 14.1 years and 16.3 years, respectively. The distribution of valvular disease is summarized in Table 1. The valve size implanted ranged from 19 to 29 mm with 23 and 25 mm being the sizes most commonly employed. Twenty-eight patients received Carpentier-Edwards porcine heterografts, and 17 patients received Hancock heterografts. In 2 patients, the graft manufacturer could not be ascertained. ACTUARIAL ANALYSIS. Follow-up was from five months after implantation through the sixth year. Three late deaths occurred in the aortic valve group. An 11-year-old patient with truncus arteriosus, type I, died of progressive pulmonary vascular obstructive disease 2 years after repair. A second patient, 11 years old and with truncus arteriosus, also died of pulmonary vascular obstructive disease a year after repair. In each of these patients, postmortem examination demonstrated a normal heterograft valve. The third death was in a 17-year-old who died Table 1. Distribution of Diagnoses in Group 1 (Aortic Valve Replacement) Diagnosis No. of Patients Aortic insufficiency 34 Congenital 15 Rheumatic 12 And VSD 5 And sinus of Valsalva aneurysm 2 Congenital aortic stenosis 9 Aortic stenosis and insufficiency 4 Total 47 VSD = ventricular septa1 defect. suddenly 2 years after valve replacement for congenital aortic stenosis. No postmortem examination was obtained. Five porcine bioprosthetic valves were removed. Four of the removals were necessary because of stenosis due to degeneration. Each of these valves had disrupted leaflets and calcification of leaflets and stent. One patient had removal of a normal porcine bioprosthetic valve at reoperation for subacute bacterial endocarditis of the aortic needle vent site. No patient died as a result of valve removal and replacement in the aortic position. No patient required replacement because he had outgrown the valve. Degeneration, calcification, and dysfunction of the porcine bioprosthetic valve occurred in 3 other patients who are awaiting valve replacement following diagnosis of dysfunction by echocardiography or angiography or both. Heterograft degeneration in these 7 valves was documented at 1.4, 2.2, 2.2, 2.6, 3.0, 3.2, and 3.3 years after implantation. The durability of porcine valves in the aortic position is represented in Figure 1. Degeneration of the porcine bioprosthetic valve first occurred during the second year after implantation. By 5 years after implantation, only 40% of the "at risk' valves remained free from calcification or dysfunction. Group 2 PATIENT DATA. Sixty-seven patients underwent mitral valve replacement with a porcine bio-

3 359 Dunn: Porcine Valve Durability in Children ; AORTIC VALVE (11~471 Fig 1. Actuarial durability versus time for porcine valves in the aortic position (Group 1). Numbers in parentheses represent number of patients at risk during that year. Bars represent 1 standard deviation. Table 2. Distribution of Diagnoses in Group 2 (Mitral Valve Replacement) Diagnosis No. of Patients Mitral regurgitation 61 Rheumatic 21 Congenital 18 Endocardia1 cushion 10 SBE 6 Systemic lupus erythematosus 2 Ebstein s L-TGA 2 Marfan s syndrome 1 Status after VSD repair 1 Congenital mitral stenosis 4 Cardiac tumors 2 Total 67 SBE = subacute bacterial endocarditis; L-TGA = levotransposition of the great arteries; VSD = ventricular septa1 defect. prosthetic valve. One of these patients had simultaneous aortic valve replacement. At the time of mitral valve replacement, the patients ranged from 10 months to 19.2 years old. The mean and median ages at implant were 11.8 and 13 years, respectively. The distribution of diseases requiring mitral valve replacement is summarized in Table 2. Included in the mitral valve group are 2 porcine valves placed in the tricuspid position for Ebstein s tricuspid malformation with L-transposition of the great arteries. These valves were included since the tricuspid position in both instances represents the systemic atrioventricular valve, thereby placing these porcine valves under the same physiological and hemodynamic stresses as porcine bioprosthetic valves in the mitral position in patients without transposed great vessels. The size range of implanted valves was 19 through 35 mm, with 29 and 31 mm being the sizes most frequently employed. Carpentier- Edwards valves were implanted in 15 patients and Hancock valves in 52 patients. ACTUARIAL ANALYSIS. Follow-up was from three months after implantation through the ninth year. There were 5 late deaths in the mitral group. The causes of death were as follows: 1. Sepsis three months after implantation in a patient with endocardia1 cushion defect 2. Congestive heart failure secondary to a perivalvular leak and prosthetic obstruction of the left ventricular outflow tract in a patient with congenital mitral stenosis six months after operation 3. Congestive heart failure secondary to prosthetic valve stenosis (not documented as a degenerated valve) in a patient with mitral stenosis and Hurler-Scheie syndrome 2.4 years after implantation 4. A drug overdose in a teenager with rheumatic mitral insufficiency 6 years after implantation 5. Valve calcification in a patient with congenital mitral insufficiency undergoing surgical removal of the calcified prosthesis 6.75 years after original implantation. Ten mitral prostheses calcified with stenotic dysfunction during the third through eighth years after implantation. Seven of these calcified valves have been removed with 1 operative death. The patient who died was an 11-year-old girl. She was seen with cardiogenic shock secondary to prosthetic valve stenosis and could not be weaned from bypass when the prosthetic valve was removed and replaced. Three patients have not yet undergone replacement of the dysfunctioning prosthetic mitral valve. No patient required valve replacement for size as he outgrew the valve. The durability of porcine valves in the mitral position is represented in Figure 2. Actuarially determined valve survival is 96.5% at 3 years,

4 360 The Annals of Thoracic Surgery Vol 32 No 4 October > 80- t - I a d 20 - L (31 (1) 0 I. I Fig 2. Actuarial durability versus time for porcine valves in the mitral position (Group 2). Numbers in parentheses represent number of patients at risk during that year. Bars represent 1 standard deviation. > 80- t - I; 60-3 W d 20-0 TRICUSPID VALVE (n* 14) Fig 3. Actuarial durability versus time for porcine valves in the tricuspid position (Group 3). Numbers in parentheses represent number of patients at risk during that year. Bars represent 1 standard deviation. Table 3. Distribution of Diagnoses in Group 3 (Tricuspid Valve Replacement) Diagnosis No. of Patients Tricuspid insufficiency 14 Ebstein s malformation 9 Status after tetralogy of Fallot 2 Rheumatic valve and SBE 1 Status after VSD repair 1 TGA status after Rastelli repair 1 Total 14 SBE = subacute bacterial endocarditis; VSD = ventricular septa1 defect; TGA = transposition of great arteries. 40% at 5 years, and 0 at 8 years. Although few patients were at risk beyond the sixth year after implantation, the high degeneration rate appears ominous. Group 3 PATIENT DATA. Fourteen patients had replacement of the tricuspid valve with a porcine bioprosthetic valve. This excludes the 2 patients already discussed in whom the tricuspid position represents the systemic atrioventricular valve. The patients in this group ranged from 6 through 18 years old. The mean and median ages at implantation were 11.8 and 13.0 years, respectively. Table 3 summarizes the underlying etiology necessitating valve replacement. All 14 patients required valve replacement for tricuspid insufficiency. Twelve valves were replaced with Hancock heterografts and 2 valves with Carpentier-Edwards heterografts. Prosthetic valves ranged in size from 25 to 35 mm, with the majority being 33 and 35 mm valves. ACTUARIAL ANALYSIS. Follow-up ranged from eight months to 9 years. No late deaths occurred in the tricuspid valve series. One valve was removed 7.6 years after implantation because the patient had outgrown the prosthesis. The removed Hancock valve was grossly normal. No dysfunction, calcification, or degeneration of the porcine bioprosthetic valve occurred in the tricuspid position, so valve survival was 100% throughout the 9 years of follow-up study (Fig 3). Group 4 PATIENT DATA. Fourteen valves were placed directly in the preexisting pulmonary annulus, which was enlarged with a patch in most patients. The patients ranged from 1.6 to 17.0 years old with a mean age of 11.2 years and median age of 12.2 years. Table 4 summarizes the distribution of diseases requiring pulmonary valve insertion. The majority of patients had tetralogy of Fallot with absence of the pulmonary valve or with patch enlargement of the pulmonary annulus. The valves used ranged from 21 to 29 mm, with the 25 mm size being most often employed. Nine Hancock and 5 Carpentier-Edwards valves were implanted. Both aortic model prostheses and heterografts fashioned from Hancock conduits were implanted. ACTUARIAL ANALYSIS. The patients in Group 4 were followed postoperatively from two

5 361 Dunn: Porcine Valve Durability in Children Table 4. Distribution of Diagnosis in Group 4 (Pulmonary Valve Replacement) Diagnosis Tetralogy of Fallot (repaired with 7 annular patch) Tetralogy of Fallot with absent 4 pulmonary valve VSD and pulmonary stenosis 1 VSD with absent pulmonary valve 1 AV canal, DORV, and pulmonary 1 stenosis Total k - d e 0 - $ 40- s - ae 20-0 No. of Patients VSD = ventricular septal defect; AV = atrioventricular; DORV = double-outlet right ventricle.. = = = = = = = * I ) (4) (41 (3) 13) (31 13) (I) PULHONIC VALVE (n*14) months to 9 years. One late death occurred in a patient with atrioventricular canal, doubleoutlet right ventricle, and pulmonary stenosis. He died five months after implantation of sepsis and a residual ventricular septal defect. The pulmonary prosthesis was normal at postmortem examination. The longest follow-up period was in a patient with tetralogy of Fallot in whom the original repair, at 8 years old, included a transannular outflow patch. The 21 mm Hancock valve was explanted 9 years later because it had become too small. It had small, yellowish plaques on the leaflet surfaces, but no calcification, and had functioned well prior to removal. No valves in the pulmonary position calcified. Thus the actuarial valve durability remained a perfect 100% at 9 years (Fig 4). I Table 5. Distribution of Diagnosis in Group 5 (Pulmonary Conduits) Diagnosis Right ventricle-pulmonary conduit Tetralogy of Fallot Pulmonary atresia Truncus arteriosus TGA, VSD, and LVOTO Double-outlet right ventricle AV canal and pulmonary stenosis Right atrium-pulmonary artery conduit Tricuspid atresia Single ventricle Left ventricle-pulmonary artery conduit L-TGA and pulmonary stenosis L-TGA, VSD, and pulmonary stenosis Total No. of Patients TGA = transposition of great arteries; VSD = ventricular septal defect; LVOTO = left ventricular outflow tract obstruction; AV = atrioventricular. Group 5 PATIENT DATA. Seventy-seven patients were in the pulmonary conduit group. The age distribution ranged from 0.8 to 21.3 years, with mean and median ages of 8.3 and 7.9 years, respectively. The conduits coursed from the right ventricle (62 patients), right atrium (13 patients), or left ventricle (2 patients) to the pulmonary artery. The diagnoses are summarized in Table 5. Tetralogy of Fallot was the largest single diagnosis. Conduits were employed in this condition because of (1) a small main pulmonary artery (9 patients), (2) pulmonary insufficiency secondary to a transannular patch (5 patients), (3) absent pulmonary valve (4 patients), and (4) anomalous coronary artery across the pulmonary outflow tract (1 patient). Seventy-five Hancock conduits and 2 Carpentier-Edwards conduits were implanted. The sizes ranged from 12 to 25 mm, with 18 through 22 mm conduits being utilized most often. ACTUARIAL ANALYSIS. The patient follow-up ranged from two months through the ninth year after implantation. There were 3 late deaths in this series. A 4-year-old child with tricuspid

6 362 The Annals of Thoracic Surgery Vol 32 No 4 October 1981 t PULMONARY CONDUITS (n.77) Fig 5. Actuarial durability versus time for porcine valves in pulmonary conduits (Group 5). Numbers in parentheses represent number of patients at risk during that year. Bars represent 1 standard deviation. atresia died of right heart failure 1 year after a modified Fontan repair; a 12-year-old child died of severe pulmonary obstructive vascular disease 1 year after truncus arteriosus repair; and a 7-year-old child died of subacute bacterial endocarditis a year after repair of pulmonary atresia. None of these deaths was attributable to conduit failure. Two valves calcified and malfunctioned, both after repair of truncus arteriosus. These conduits were successfully replaced 4.7 and 5.4 years following initial implantation. An additional 18 mm conduit with a normally functioning porcine bioprosthetic valve was replaced successfully ten months after a Rastelli repair for dextrotransposition of the great arteries with ventricular septal defect and left ventricular outflow tract obstruction. After the initial repair, this child experienced congestive heart failure due to right ventricular conduit stenosis and subaortic obstruction due to the ventricular septal defect patch configuration. The conduit was replaced, and obstructions were alleviated at the second operation. Durability of conduits is depicted in Figure 5. Calcification was first noted in 1 valve during the fifth follow-up year, giving an actuarially predicted valve survival of 89% at 5 years. The only other conduit valve dysfunction occurred during the sixth year. Valve survival was 73% at 6 years with no further attrition through 9 years. Table 6. Distribution of Diagnosis in Group 6 (Apicoaortic Conduit Implantation) Diagnosis Aortic stenosis and small annulus 4 Subvalvular stenosis 2 Supravalvular stenosis 1 Aortic insufficiency and small 1 annulus Total 8 Group 6 No. of Patients PATIENT DATA. Eight patients had apical left ventricular-descending aortic conduits placed for various forms of left ventricular outflow tract obstruction (Table 6). They ranged from 1.9 to 18.8 years old. The mean and median ages at operation were 11.5 and 11.9 years, respectively. All conduits contained Hancock valves, with a size distribution of 16 through 22 mm. ACTUARIAL ANALYSIS. Follow-up was six months to 3 years after insertion. The valves and conduits were removed in 4 patients (50%). In 2 patients, the conduit was removed a year after insertion because of recurrent systemic emboli. Emboli were thought to arise at the ventricular stent site. One of these patients died at the time of conduit removal. A second patient died seven days after removal of a conduit because of candida abscess formation at the conduit-abdominal aorta anastomosis. Death was attributed to anastomotic dehiscence at the abscess site. In each of the three instances of conduit removal, the removed Hancock porcine heterograft appeared grossly normal. Removal of a fourth conduit was required because of valve calcification and stenosis 3 years after insertion. Actuarial analysis is difficult because of the small patient population in Group 6. The single valve failure occurred in the only valve in place during the third postoperative year. Thus, the actuarial analysis indicates a 100% valve durability for the first and second "at risk" years, and 0% during the third "at risk" year (Fig 6). Again, it should be pointed out that the small patient population-especially at 3 years-

7 363 Dunn: Porcine Valve Durability in Children loo 1 > 80- t W s - ar 20 - APICOAORTIC CONDUITS (n.8) (11 0 I Fig 6. Actuarial durability versus time for porcine valves in apicoaortic conduits (Group 6). Numbers in parentheses represent number of patients at risk during that year. Bars represent 1 standard deviation. -PULMONARY CONDUITS (n.77).---.aortic VALVE (n~ c - d P - ar PULMONARY CONDUITS (11.77) O--+ MITRALVALVE (n* I I I I I I h P P P P P Fig 8. Actuarial durability of porcine valves in pulmonary conduits compared with those in the mitral position (Group 5 versus Group 2). Bars represent 1 standard deviation. At 3, 4, 5, 6, and 7 years, the p value represents a significant difference (p < 0.05). loo 1 s 2o 0 P P P P Fig 7. Actuarial durability of porcine valves in pulmonary conduits compared with those in the aortic position (Group 5 versus Group 1). Bars represent 1 standard deviation. At 2, 3,4, and 5 years, the p value represents a significant difference (p < 0.05). makes statistical reliance on these data impossible. Porcine Valve Durability Related to Location Comparing the degeneration of porcine bioprosthetic valves in left-sided positions (aortic and mitral positions) with right-sided positions (tricuspid, pulmonary, and pulmonary conduit) demonstrates statistically significant differences. The actuarial durability curves for the aortic position (Group 1) versus the pulmonary conduit position (Group 5) are depicted in Figure 7. The bars represent 1 standard deviation from the mean at each year of follow-up and p represents years at which durability of valves in the two groups are statistically different ( p < 0.05). The durability in the aortic position is consistently less than that of the pulmonary conduit position for all years they are compared (1 to 6 years). This is statistically significant for the second through fifth "at risk" years. Statistical significance is lost during the sixth year of comparison despite the absolute valve durability of 73% in pulmonary conduits and only 40% in the aortic position. This loss of statistical significance is probably because of the small population of the aortic group at 6 years. Likewise, a comparison of heterografts in the mitral versus conduit position (Fig 8) demonstrates a consistently greater failure rate in the mitral position. This difference is statistically significant from the third "at risk" year through the seventh year. Again, the small mitral population beyond 7 years defies statistical evaluation. Figure 9 demonstrates a slightly better durability in the mitral position compared with aortic position. This superiority is, however, lost by the fifth follow-up year. Porcine Valve Durability Related to Age of Patient An increased propensity is clearly evident for porcine valves to degenerate in the pediatric population compared with the adult population. The role of age was evaluated to see if this was a factor in valve degeneration within the pediatric population.

8 364 The Annals of Thoracic Surgery Vol 32 No 4 October t - d $ 60- K 2 - ; 40- s AORTIC VALVE 0--o>15OFAGE(n=27) o--+c15 OFAGEIn=20) P P P Fig 9. Actuarial durability of porcine valves in the mitral position compared with the aortic position (Group 2 versus Group 1). Bars represent 1 standard deviation. At 2, 3, and 4 years, the p value represents a significant difference (p < 0.05). (M = mitral; A = aortic.) MITRALVALVE 0 P P Fig 11. Actuarial durability of porcine valves in the aortic position in patients more than 15 years old compared with those less than 15 years old. Bars represent 1 standard deviation. At 2 and 3 years, the p value represents a significant difference (p < 0.05). risk" year and continued throughout the 5 years of comparison (Fig 11). 0' L I 7 I " 1 P P Fig 10. Actuarial durability of porcine valves in the mitral position in patients more than 10 years old compared with those less than 10 years old. Bars represent 1 standard deviation. At 3 and 4 years, the p value represents a significant difference (p < 0.05). MITRAL VALVES. The 67 patients in this group were divided into two subsets: less than 10 years old (n = 22) and more than 10 years old (n = 45). The actuarial valve durability of these two subsets of mitral valve replacement is depicted in Figure 10. Valve degeneration and dysfunction are consistently worse in the younger age group starting with the third "at risk" year ( p < 0.05). AORTIC VALVES. The aortic valve patients were divided into two subsets: less than 15 years old (n = 20), and more than 15 years old (n = 27). Fifteen years of age was chosen as the cutoff for statistical purposes to attempt to equalize the size of each subset. As in the mitral group, valve durability was greater in the older subset. This trend started with the second "at Comment Commercially available porcine bioprosthetic valves have been employed extensively since the early 1970s [3]. Incorporating technical advances including glutaraldehyde tissue fixation [l] and the flexible mounting stent [2], these valves have proved themselves to be reasonably durable and relatively inert in the adult population [l, 19, 201. Valve replacement in children is a far less frequent occurrence than in adults, yet case reports and limited series of children would indicate a high incidence of porcine bioprosthetic valve failure in the younger population [S-161. Since 1977, 25 cases of degeneration of porcine heterografts have been reported in children [S-161. This in itself would seem to represent an accelerated rate of porcine valve degeneration. A large series from Children's Hospital Medical Center, Boston [14], demonstrated that only 50% of xenografts in children remained functional after forty-five months. Despite this alarmingly high failure rate in children, porcine valves offer certain advantages in the pediatric population [4, 5, 7, Anticoagulation is not required in patients with regular atrial rhythms.

9 365 Dunn: Porcine Valve Durability in Children There is a low incidence of valve thrombosis or thromboembolic episodes. The valve has a moderately low-profile configuration, important in a small ventricle or ascending aorta. Hemodynamics are that of central orifice flow with minimal regurgitation. Valve degeneration follows a slow, predictable course-amenable to surgical replacement-rather than rapid catastrophic failure. The disadvantages of the porcine valves in the pediatric population include the following: 1. Porcine valves have a poor orifice-toannulus ratio, causing hemodynamic obstruction, especially in smaller sizes. 2. The apparently accelerated valve degeneration demonstrated by this and other series seems to make early prosthetic removal and replacement almost inevitable. It is unclear whether the advantages of porcine valves outweigh the disadvantages in the pediatric population. It can be argued that most prosthetic valves in children would require replacement as the child outgrew the valve, even if the prosthesis did not degenerate. However, this series demonstrates that we can expect to replace valves for degeneration, not for size discrepancy as the child grows. Because most of the valves replaced in the series were for regurgitant lesions, the porcine bioprosthetic valves implanted were in the large or adult size range, decreasing the likelihood of valve removal because of size. Twenty-one of the porcine valves in our study required removal. Cf these, only 2 were removed because the recipient had outgrown the valve. These were in the tricuspid (Group 3) and pulmonary positions (Group 4). As we have demonstrated, however, the risk of valve degeneration is minimal in right-sided positions. In the mitral and aortic positions, the predominant reason for prosthetic removal is valve degeneration and dysfunction. In this series, 7 mitral-positioned porcine valves were removed because of calcific degeneration. In the aortic position, 4 of 5 removed valves were calcific. The fifth aortic prosthesis removed was for subacute bacterial endocarditis of the aortic needle vent site. No left-sided valve was removed because of size after growth of the recipient. Although initially thought to be inert, it is now accepted that the glutaraldehyde-fixed porcine valve demonstrates definite and progressive degenerative changes including macrophage infiltration, microscopic collagen disruption leading toward cusp disruption, calcification, and valve dysfunction [22]. The etiology of the accelerated degeneration documented in children is not understood. It may be related to the positive calcium balance of childhood leading to a "metastatic" heterograft calcification [14]. The relative freedom from calcification of right-sided valves suggests this is not the case. Most calcified porcine valves are in the aortic and mitral positions in this and other published series. The higher pressures of the systemic circulation may well place more mechanical stress on the mitral and aortic prostheses [23], resulting in a more rapid degeneration and ultimate dystrophic calcificationthat is, calcification that occurs in already damaged tissue. Furthermore, the more rapid degeneration and calcification of porcine bioprosthetic valves in children may be related to greater mechanical stress secondary to faster heart rate in children than in adults. Although the etiology of accelerated degeneration in children is obscure, it is clear that the problem is real. In this series, 20 porcine heterografts became calcified (10 mitral, 7 aortic, 2 pulmonary conduits, and 1 apicoaortic conduit). We have demonstrated that accelerated degeneration is most common in the systemic valve positions (aortic and mitral), 5-year valve survival being only 40% in both positions. Degeneration occurred more rapidly in the younger children of this series. Pediatric Cardiac Surgery Study Group John Brown, M.D., Indiana University School of Medicine, Indianapolis, IN David Clark, M.D., University of Colorado Medical Center, Denver, CO James Donahoo, M.D., The Johns Hopkins Hospital, Baltimore, MD

10 366 The Annals of Thoracic Surgery Vol 32 No 4 October 1981 Jeffrey M. Dunn, M.D., St. Christopher s Hospital for Children, Philadelphia, PA Eric Foster, M.D., Albany Medical College, Albany, NY Hillel Laks, M.D., Yale University School of Medicine, New Haven, CT Sidney Levitsky, M.D., University of Illinois Medical Center, Chicago, IL Frank Midgley, M.D., Children s Hospital National Medical Center, Washington, DC Anthony Moulton, M.D., Columbia-Presbyterian Medical Center, New York, NY Robert Sade, M.D., Medical University of South Carolina, Charleston, SC Ralph Siewers, M.D., University of Pittsburgh School of Medicine, Pittsburgh, PA Robert Szamicki, M.D., Presbyterian Hospital, San Francisco, CA References 1. Carpentier A, Deloche A, Relland J, et al: Six year follow-up of glutaraldehyde preserved heterografts. J Thorac Cardiovasc Surg 68:771, Reis RL, Hancock WD, Yarborough JW, et al: The flexible stent: a new concept in the fabrication of tissue heart valve prostheses. J Thorac Cardiovasc Surg 62:683, Durability assessment of the Hancock porcine bioprosthesis: a multicenter retrospective analysis of patients operated prior to Hancock Labs Report, April, Gonzalez-Lavin L, Folger GM, Davila JC: Porcine xenografts in the treatment of mitral valvular disease in children. Mich Med, p 195, April, Roberts WC: Choosing a substitute cardiac valve: type, size, surgeon. Am J Cardiol 38:633, Crawford FA, Selby JH, Joransen JA: Mitral valve replacement with a porcine heterograft in an infant. J Thorac Cardiovasc Surg 75:705, Davila JC, Magilligan DJ Jr, Lewis JW Jr: Is the Hancock porcine valve the best cardiac valve substitute today? Ann Thorac Surg 26:303, Levitsky S: Discussion of Stinson EB, Griepp RB, Oyer PE, Shumway NE: Long-term experience with porcine aortic valve xenograft. J Thorac Cardiovasc Surg 73:54, Brown JW, Dunn JM, Spooner E, Kirsh MM: Late spontaneous disruption of a porcine xenograft mitral valve. J Thorac Cardiovasc Surg 75:600, Kutsche LM, Oyer PE, Shumway N, Baum D: An important complication of Hancock mitral valve replacement in children. Circulation 6O:Suppl 1:1-98, Rose AG, Forman R, Bowen RM: Calcification of glutaraldehyde fixed porcine xenograft. Thorax 33:111, Geha AS, Laks H, Stansel HC, et al: Late failure of porcine valve heterografts in children. J Thorac Cardiovasc Surg 78:351, Lamberti JJ, Wainer BH, Fisher KA, et al: Calcific stenosis of the porcine heterograft. Ann Thorac Surg 28:28, Sanders SP, Levy RJ, Freed MD, et al: Use of Hancock porcine xenografts in children and adolescents. Am J Cardiol 48:429, Clarkson S, Sade RM, Hohn A: Clinical hemodynamic results of extracardiac conduit reconstruction of the pulmonary artery. Clin Cardiol 3:42, Silva MM, Pollack J, Silva MD, et al: Calcification in porcine xenograft valves in children. Am J Cardiol 45:685, Grunkemeier GL, Starr A: Actuarial analysis of surgical results: rationale and method. Ann Thorac Surg 24:404, Cutler S, Ederer F: Maximum utilization of the life table method in analyzing survival. J Chronic Dis 8:699, Oyer PE, Stinson EB, Griepp RB, Shumway NE: Valve replacement with the Starr-Edwards and Hancock prosthesis: comparative analysis of late morbidity and mortality. Ann Surg 186:301, Hetzer R, Hill JD, Kerth WJ, et al: Thrombosis and degeneration of Hancock valves: clinical and pathological findings. Ann Thorac Surg 26:317, Sade RM, Ballinger JF, Hohn AR, et al: Cardiac valve replacement in children: comparison of tissue with mechanical prosthesis. J Thorac Cardiovasc Surg 78:123, Spray TL, Roberts WC: Structural changes in porcine xenografts used as substitute cardiac valves. Am J Cardiol40:319, Carpentier A, Lemaigre G, Robert L, et al: Biological factors affecting long-term results of valvular heterografts. J Thorac Cardiovasc Surg 58: Discussion DR. TIMOTHY J. GARDNER (Baltimore, MD): This report by Dr. Dunn is commendable, especially because it combines the experience of 12 pediatric cardiac surgical groups, with the result that 227 late surviving children and adolescents are available for analysis. I hope there will be more such combined series, especially in the area of congenital deformities. The fact that only 40% of aortic and mitral valve substitutes were functional at S years is even worse than might have been predicted from previous reports and is obviously the most alarming aspect of this series. The apparent durability of right-sided porcine valves, however, especially in the tricuspid position where the risk of valve thrombosis with mechanical prostheses is high, is encouraging. It is unfortunate that the actual number of tricuspid and pulmonary valve substitutes in the series is small.

11 367 Dunn: Porcine Valve Durability in Children Although the report includes nearly 80 patients with a valved conduit, conclusions about the longterm durability of this valve are difficult to make based on the medium-range follow-up now available. In the 2 patients with valve failure in this group, high pulmonary and right ventricular pressures presumably were present, as both patients had truncus deformities. This may explain the accelerated valve degeneration in these 2 patients. Although I think it is fair to say that most porcine valves in children, including those on the right side of the heart, will eventually fail, medium-term durability in the case of valved conduits may be adequate, and problems such as tissue ingrowth and the development of flow-to-conduit diameter discrepancy because of patient growth may necessitate conduit replacement prior to the occurrence of valve failure itself. I am curious about the potential usefulness of the bovine pericardial valve for children. In terms of the range of available sizes and the flow characteristics of the Ionescu-Shiley prosthesis, the valve is particularly attractive for use in children, especially in children with small aortic roots. I wonder if Dr. Dunn has any significant follow-up information regarding this valve in children. On the basis of this present report, the use of a porcine valve in a left-sided position in a child not yet out of the growth phase is difficult to justify. As discouraging as this is, since we would prefer to avoid warfarin treatment in children, we have been encouraged by our own long-term experience with mechanical valves in children, which was recently reviewed. Most of these children have been managed with warfarin, and we have seen very few thromboembolic and anticoagulant-related complications. In addition, we have several children with St. Jude prostheses, including 3 children who required mitral valve replacement during infancy and who for various reasons are not on a regimen of warfarin. These children are receiving aspirin alone and are doing well 1 to 2 years after valve placement. DR. GORDON K. DANIELSON (Rochester, MN): Dr. Dunn s fine presentation has clearly stated the problem of the biological valve in children. We have followed the 49 patients who have survived porcine valve replacement at the Mayo Clinic between 1973 and They ranged from 2 to 18 years old. Half of these patients had undergone a previous operation, such as repair of tetralogy of Fallot, atrioventricular canal, or transposition. In 32 patients concomitant procedures were performed, such as septation of the univentricular heart and the Fontan procedure. About half of the valves replaced were on the right side and half were on the left. Follow-up has been completed in 44 patients. Severe prosthetic valve dysfunction required reoperation in 7 patients a mean of only thirty-four months postoperatively. In addition, another 7 patients have clinical evidence of valvular dysfunction. Six patients died in the follow-up period, at least 1 of whom had evidence of prosthetic valve dysfunction. One example of extensive calcification of the aortic and mitral valves occurred in a patient who required urgent reoperation only fourteen months after the valves were implanted. This is an alarmingly high early failure rate and compares very unfavorably with our previous results with mechanical valvular prostheses. Based on these data and those of others, we discontinued the use of bioprostheses as isolated valve replacements in children several months ago. DR. w. G. WILLIAMS (Toronto, Ont, Canada): I compliment Dr. Dunn on organizing this multicenter study. The implied conclusion is that we should use mechanical rather than tissue valves. Our experience consists of 89 children with leftsided prosthetic valves, 39 in the aortic position and 50 in the mitral. There is a marked difference in survival between the two groups, with no mortality among the patients having aortic valve replacement and a 50% 5-year survival among those having mitral valve replacement. In spite of this difference in survival, the incidence of complications was virtually identical in these two groups. Actuarial analysis of complications in the 53 children with tissue valves as opposed to the 36 with mechanical prostheses reveals that problems occur with equal frequency with either valve. Fifty percent of the children had a complication within 5 years of operation. One-third of the complications were related to valve wear, and, in the porcine valves, the incidence of calcification beyond the first postoperative year is 18% and increasing. Unlike Dr. Dunn s group, we found that one-sixth of the complications were related to outgrowth of the valve. Only 2 patients have had a systemic embolus. One occurred in a patient with a tissue valve and one in a patient with a tilting-disc valve. Results with aortic or mitral valve prostheses whether mechanical or tissue are not good. Every reasonable effort should be taken to repair the child s natural valve. DR. DUNN: I thank the discussants for these confirming data. In reference to Dr. Gardner s comment, although the 2 patients who experienced calcified conduits had truncus arteriosus, neither had pulmonary hypertension postoperatively. We do not have a large experience with bovine pericardial valves. Most of the reports of valve failure in the literature were before these valves became commercially available. However, in direct communications with the company making these valves, it would appear that 3 or 4 children now have calcified valves. I think that bovine pericardial valves

12 368 The Annals of Thoracic Surgery Vol 32 No 4 October 1981 are not going to prove to be much better in this regard than porcine valves. Most of the data that we have been able to find on the risk of anticoagulation in children are anecdotal in nature. I was glad to hear from Dr. Gardner about a series of anticoagulated patients in the pediatric population without a high failure rate. I think you should know that the 12 authors of this paper did not all agree on whether the porcine valve should be used or whether it should be abandoned for mechanical valves. Although we have had 20 valves calcify and degenerate, only 1 patient died during reoperation. The others survived even though they required repeat valve replacement. Porcine valve degeneration is not as slow as rheumatic valvular degeneration and can occur over just a few months of time. Thus, it can be very severe and life threatening. Notice from the Southern Thoracic Surgical Association The Twenty-eighth Annual Meeting of the Southern Thoracic Surgical Association will be held at The Breakers, Palm Beach, FL, Nov 5-7, There will be a $100 registration fee for nonmember physicians except for guest speakers, authors and co-authors on the program, and residents. There will be a Postgraduate Program on Current Concepts in Thoracic and Cardiac Surgery. This meeting has been ap- proved for Category I, 14 hours CME credit. Manuscripts of papers accepted for the program must be submitted to The Annals of Thoracic Surgery by October 15, Hotel reservations are sent to members. Guests may correspond with The Breakers, Attn. Reservation Manager, PO Box 831, Palm Beach, FL 33480; phone (305)