| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 1999;16:403-413
© 1999 Elsevier Science NL
a Department of Thoracic and Cardiovascular Surgery, Aarhus University Hospital in Skejby, Aarhus, Denmark
b Department of Cardiology, Aarhus University Hospital in Skejby, Aarhus, Denmark
Corresponding author. Department of Cardio-Thoracic Surgery, Copenhagen University Hospital, Gentofte, Niels Andersens Vej 65, DK-2900 Hellerup, Denmark. Tel.: +45-3977-3825; fax: +45-3977-7644
e-mail: olund{at}thorax.dk
 |
Abstract |
|---|
 Top Abstract 1. Introduction2. Materials and methods3. Results4. DiscussionReferences |
|---|
13 mm/m2. Conclusions: The Starr-Edwards aortic ball valves, not least the currently available silastic ball valve, are durable through the remaining life time of the patients and able to secure near normal age and sex specific survival provided valve and patient size mismatch is avoided.
Key Words: Heart valve replacement • Prosthetic heart valves • Survival • Valve related complications • Thromboembolism • Multivarate statistical analysis
|
1. Introduction |
|---|
|
TopAbstract1. Introduction2. Materials and methods3. Results4. DiscussionReferences |
|---|
Clinical trials with the Starr-Edwards cloth covered aortic valve began in 1967 [1]. Aiming to reduce thrombogenicity, the valve base and cage struts were covered with teflon while sporting a hollow metal ball [1]. However, wear of the cloth covering the struts, among other complications giving rise to intravascular haemolysis, prompted development of the track valve: it featured a narrow inner metal track on the cage struts to avoid contact between metal ball and teflon [1]. Clinical use of the track valve began in 1972 [1], but production was discontinued some 10 years later because it was comparatively expensive and noisy and because it eventually showed no long-term advantages compared with the silastic ball valve [3].
The Starr-Edwards ball valves were for many years the valves of choice in many centres world-wide and a significant number of patients with one of these valves are currently alive. A thorough account of the long term performance profile of these three Starr-Edwards ball valve types from a quantitatively large mono-institutional series could thus serve three purposes: first, a detailed knowledge of the complication peculiarities of these valves may give valuable guidance in managing the patients who are still alive; second, with a follow-up longer than 30 years and thus spanning the potential remaining life time of the average patient, the complete survival characteristics or ‘natural history’ of life after AVR in general with a mechanical valve may become apparent; and third, the very long term performance of the silastic ball valve will set a record yet to be matched by modern disc valves.
From the first AVR with a pre-silastic ball valve at the present centre in 1964, the Starr-Edwards ball valve remained our valve of choice until the mid 1980’s. Featuring a total of 717 patients with a silastic ball, cloth covered or track valve implanted in the aortic position during 1965 through 1993 and with a complete follow-up conducted during late 1997 through January 1999 we set out to pursue the three purposes stipulated in the above.
|
2. Materials and methods |
|---|
|
TopAbstract1. Introduction2. Materials and methods3. Results4. DiscussionReferences |
|---|
Using the above inclusion and exclusion criteria, a total of 1541 patients underwent a first AVR during 1965–1993 at the present centre. From the early 1970’s a disc valve was generally chosen in case of a small aortic root and in young patients. A bi-leaflet disc valve is currently the valve of choice. The present patients with a Starr-Edwards ball valve constituted 96% of AVR patients during 1965–1974 (N=204), 72% during 1975–1984 (N=498), 43% during 1985–1989 (N=369), and 4% during 1990–1993 (N=470). Patient age and prevalence of New York Heart Association functional classes I–II increased progressively for the entire AVR population during 1965–1993. With regard to these factors, the present 355 patients who received a silastic ball valve did not differ from 778 who received a disc valve in contrast to the present patients who received a cloth covered valve or a track valve (Table 1).
|
2.2. Operation and anticoagulant treatment
Previously described standard methods [4–6] involving complete extracorporeal circulation, general and topical hypothermia, and crystalloid cardioplegic cardiac arrest were used in all patients. Coronary artery bypass grafting was performed in 30 patients while six had the ascending aorta replaced (completely with reimplantation of the coronary ostia in four).
All patients were started on a regimen of life-long anticoagulant treatment with vitamin K antagonists aiming at maintaining Owren's prothrombin-proconvertin index within the 10–20% range. For our laboratory this translates into an INR corridor of 2.5–5.0. Apart from short pauses due to other surgical procedures, only 11 patients (three with a silastic ball valve, five with a cloth covered valve and three with a track valve) were shifted to low dose acetyl salisylic acid and dipyridamol treatment due to bleeding complications and difficulties in maintaining the INR values within the desired interval.
2.3. Follow-up and valve related complications
The Aarhus University Hospital AVR database was created in 1983–1984 involving all patients operated during 1965–1973 [5]. Including all patients operated during 1965–1986, the second database version with follow-up during the second half of 1987 was ready in early 1988 [4,5,7,8]. The present third version with follow-up conducted during July 1997 through January 1999 included operations until December 31, 1993. A detailed account of database construction, definitions of variables and of follow-up involving direct contact to all patients who were still alive, to the general practitioner of all patients and to all hospitals where patients had been admitted throughout follow-up has been given previously [5]. Other AVR databases have been constructed in a similar manner [9].
All of the present 717 patients were accounted for at the end-of-study: 63 died during the first 30 days after AVR (early mortality) while 442 died late and 212 were still alive. Death certificates were available for all deaths while autopsies was performed in 54 (86%) of early deaths and in 184 (42%) of late deaths. Maximum (mean) follow-up was 31.1 (8.8) years for the silastic ball valves, 28.7 (11.3) years for the cloth covered valves, and 22.1 (11.5) years for the track valves with a total of 3130, 1851, and 2273 patient-years at risk, respectively.
Registration of valve related complications followed internationally accepted guidelines [10] as previously detailed [5,7,9]. Embolism was recorded if a systemic vascular event could not be proven to be thrombotic or hemorrhagic. Embolic and hemorrhagic events were recorded as minor if symptoms subsided within 48 h or as major if they did not. Valve thrombosis and paravalvular leaks were recorded in the absence of infective endocarditis, and haemolysis if B-hemoglobin in a patient with serologic signs of chronic haemolysis [6,11] could not be kept normal with iron, vitamin B-12 and folic acid medication.
2.4. Statistical analysis
All analyses were computerised by means of the BMDP Dynamic version 7.0 software package [12]. Simple comparisons were done using a one-way analysis of variance or a Pearson 
2-test. Cumulative survival and complication freedom curves were made using the Kaplan–Meier product-limit method and differences between curves tested with a log-rank test and a Gehan test. Multivariable analyses of long-term survival and valve related complications were done using a Cox proportional hazard regression analysis employing a comprehensive formalised analysis sequence as previously described [4,5]. The odds ratio (multiplication factor for risk increase relative to a basic risk of 1) was calculated [5] for each independent risk factor. The Cox analyses were performed on two patient groups: the 573 patients with aortic stenosis and the 144 with regurgitation including those with combined stenosis and regurgitation (Table 1). Relative survival for each postoperative year was calculated as the ratio between conditional survival probability of the patients and of a Danish background population which was selectively matched for sex, age and operative year to each valve type group separately [5,13]. Survival, complication freedoms, relative survival and linearised incidence rates are given together with 95% confidence intervals [5,12], while quantitative data are given with±one standard deviation. The level of statistical significance was 0.05.
|
3. Results |
|---|
|
TopAbstract1. Introduction2. Materials and methods3. Results4. DiscussionReferences |
|---|
The etiology of the aortic valve disease was tricuspid sclerotic in 410 patients, rheumatic in 155, bicuspide in 60, congenital stenosis in 34, endocarditis in 42 and idiopathic media necrosis, lues, Marfan's syndrome or other systemic disease in the remaining 16 without differences between the valve groups. All but 5% (N=39) had sinus rhythm preoperatively, 4% (N=31) had atrio-ventricular block grade 1–3, 6% (N=43) had ventricular ectopic beats in more than 10% of beats in resting electrocardiogram, and 9% (N=68) had left or right bundle branch block with no difference between the three valve groups.
The causes of the 63 early deaths are given in Table 2. Reflecting the early operative era, early mortality for the model 1200 valve was 22.9% (N=11) while it was 6.8% (N=21) for the model 1260 valve, 7.9% (N=l3) for the cloth covered valves, and 9.l% (N=l8) for the track valve (P<0.01). Early mortality dropped from 25.5% during 1965–1969 to 3.6% for 1984–1993 or to 1.5% when only isolated AVR for chronic valve disease was considered.
|
|
|
|
|
|
|
Relative survival was analysed for the patients who were alive 30 days after the operation and with the model 1200 valve excluded (Fig. 2). Upper 95% confidence limits of relative survival was less than l (indicating excess mortality relative to background population) during the 1st and 13th year for the model 1260 silastic ball valve, during the 1st, 3rd, 13th, 15th and 20th year for the cloth covered valve, and during the 1st, 10th and 20th year for the track valve.
|
13 mm/m2 did.
|
3.2.1. Thromboembolism
A total of 108 patients suffered 142 thromboembolic events (Tables 3 and 4, Fig. 3): 82% (N=89) of these patients suffered at least one cerebral embolism, 10% (N=11) a central retinal artery embolism, 5% (N=5) a thrombosed valve while the remainder had extremity (N=l), mesentery (N=l) or coronary (N=l) embolism. Fifty-one percent (N=70) of all 137 embolic events were major (38% non-fatal, 13% fatal; Table 3). The five valve thromboses (Table 3) occurred between 12 and 17 years after AVR in two patients with a silastic ball valve (24 and 26 mm) and in two with a track valve (26 and 27 mm) but after 4 months in one with a 29 mm cloth covered valve; the latter case and one with a track valve were diagnosed at autopsy, two patients died 5 and 6 months after the diagnosis (one during redo AVR) and one underwent a successful redo AVR 7 months after the diagnosis. Counting only the first event, the rate of thromboembolism dropped from 1.61%/patient-year during the first 10 postoperative years to 1.08%/patient-year after the 10th year for the silastic ball valves, while the rate increased from 1.08 to 1.55%/patient-year for the cloth covered valves and remained unchanged for the track valve. The independent risk factors are shown in Table 6. Notably, valve type was of no influence, but valve size index >17 mm/m2 increased the risk of thromboembolism in the regurgitation group; valve size index >17 mm/m2 was present in 29% of 24 patients with idiopathic media necrosis, lues, Marfan's syndrome or other systemic diseases compared to 6% of 120 with other etiologies in the regurgitation group (P<0.05).
|
13 mm/m2 was an independent risk factor which doubled the risk of bleeding in the stenosis group.
3.2.3. Endocarditis
The rate of endocarditis tended to be lower for the silastic ball valve than for the other two valve groups (Tables 3 and 4). No independent risk factors could be identified. Twenty-three patients were treated conservatively while 14 underwent redo AVR; 70% (N=16) versus 2l % (N=3; P<0.05), respectively, died within 3 months after the start of treatment.
3.2.4. Paravalvular leak, haemolysis and structural failure
The incidence rates of paravalvular leak and haemolysis did not differ between the valve groups (Table 3). Out of the 14 cases with paravalvular leak, 10 also had haemolysis while the one case amongst the cloth covered valves actually was a partial dehiscence of the prosthesis noted at autopsy 2 months after AVR. The cases of haemolysis without paravalvular leak was noted 3 months, 3 years and 10 years after AVR in three patients with a silastic ball valve (24, 26, and 27 mm), after 14–18 years in three with a cloth covered valve (24, 26, and 27 mm), and after 5, 13, and 17 years in three with a track valve (27, 29, and 29 mm); eight of these nine patients died 0.5–3 years after the diagnosis of haemolysis from cardiac failure and one underwent a successful redo AVR. There were no cases of structural or mechanical failure except for wear of the cloth covering the cage of the cloth covered valves, which was noted in virtually all of the 64 late deaths amongst the cloth covered valves where an autopsy was performed (Table 2).
3.2.5. Composite complication rates and redo AVR
The silastic ball valves tended to have a higher rate of all the valve related complications than the other two valve groups (Tables 3 and 4, Fig. 5), mainly due to the tendency for a higher bleeding rate (Fig. 5). However, the 25-year freedom for the silastic ball valves was as high as 51%. There was a tendency for an increase of the incidence of all complications after the 10th postoperative year for both the cloth covered and the track valve. A small valve size index was an incremental risk factor in the stenosis group while a large valve size index doubled the risk in the regurgitation group (Table 6). Redo AVR was performed mainly due to endocarditis or paravalvular leak with a 30-day mortality of 18.5% (5 out of 27; Table 3). The 27 replacement valves were silastic ball valves (N=6), disc valves (N=18) or stented xenografts (N=3). The freedom from redo AVR was quite high in all three valve groups and as high as 95% at 25 and 30 years for the silastic ball valve (Table 4). The incidence tended to decrease after 10 years for the silastic ball valves (from 0.40 to 0.16%/patient-year) but tended to increase for the cloth covered valves (0.17–0.60) and track valves (0.33–0.65). There were no independent risk factors in the stenosis group while preoperative coronary artery disease (odds ratio 9.97, P<0.01) and endocarditis (odds ratio 7.61, P<0.05) were risk factors in the regurgitation group. Serious complications (Table 3) occurred at comparable rates in the three valve groups (Table 4) with a 25-year freedom of 75% in the silastic ball valve group. The independent risk factors are given in Table 6. Valve related deaths tended to occur at a lower rate in the silastic ball valve group than in the other two groups (Tables 3 and 4). The independent risk factors are given in Table 6.
|
4. Discussion |
|---|
|
TopAbstract1. Introduction2. Materials and methods3. Results4. DiscussionReferences |
|---|
In the present series, the currently available silastic ball valve has passed the 30-year survival mark. The Starr-Edwards valve was our valve-of-choice until the mid 1980s and our silastic ball valve patients were generally comparable to our large group of patients who received a disc valve, the majority after 1985. These facts indicate that the present series covers the spectrum of patients who also require AVR today. Seen in this light, it is quite remarkable to note that the model 1260 silastic ball valve in particular, seemed to secure our patients near normal survival beyond 30 years when compared with a matched Danish background population. Strengthened by current era prosthesis-type-independent low operative mortality and complete freedom from structural and mechanical failure, such a result argues for continued use of the silastic ball valve.
A just comparison with other valve series craves that crucial factors such as time period of operation and age of patients are comparable. Also the length of follow-up may play a role since incidence rates of certain complication modalities may change over time. Concentrating on currently available aortic valve prostheses, we have been able to identify two porcine xenograft AVR series [16,17] and two silastic ball valve series [18,19] with a follow-up to 20 years or more which were comparable with the present series as regards patient age and time period. Long-term survival of these series and of the present model 1260 silastic ball valve group were comparable. Mainly due to degenerative valve failure, however, the 15-year freedom of all complications was only 31 [16] or 40% [17] for the porcine xenografts compared with 51% also at 25 years for the present silastic ball valve. Needless to say, the incidence of redo AVR was considerably higher for the xenografts [16,17] than for the present Starr-Edwards ball valves. In a series with follow-up exceeding 25 years after AVR with an allograft [9], long-term survival was better than in the present report, but 20-year freedom from degenerative valve failure was only 18% and the patients were on average 6–9 years younger than the patients of the present three ball valve groups. In two age comparable AVR series with up to 17 years of follow-up, the most widely used mechanical valve today, the St. Jude bi-leaflet disc valve introduced in 1978, had similar or better long-term survival than the present model 1260 silastic ball valve, mainly due to operative era specific lower early mortality [20,21]. However, freedoms from all complications of 47–60% at 13–14 years [20,21] seem to be comparable to the present results. In a previous comparison between aortic silastic ball valves and St. Jude valves implanted during 1980–1986 we were unable to pinpoint differences in medium-term performance [22].
Thromboembolism and bleeding are the dominant complications of any mechanical valve series and constituted 82% of all complication events of the present patients. Pooled data indicate that the incidence of anticoagulant related bleeding does not relate to prosthesis type for a given target INR [23]. With a target INR corridor comparable to the present one, incidences of embolism, on the other hand, seems to be a little lower for contemporary disc valves compared with the silastic ball valve [23], in accordance with a large meta analysis [24]. However, the present incidences of thromboembolism of our three ball valve groups seem to lie well within the spectrum of reported incidences for contemporary disc valves [23]. Nonetheless, the current trend, to lower target INR for patients with a modern disc valve [25], is probably not advisable for patients with a ball valve. Whether the aim of lowering target INR in disc valve patients is fulfilled, a reduced rate of bleeding at the expense of only a marginal increase of the incidence of embolism [25], remains to be proven. Furthermore, according to the present distribution of major and minor complications, a lowering of target INR would reduce the rate of a rather benign complication (nearly 80% of bleeding episodes were ‘minor’) at the potential price of increasing the rate of a more ‘malignant’ complication. Another result should be brought to mind in this connection: despite increasing age during follow-up, the incidence of embolism of our silastic ball valve patients dropped after the first 10 years but the incidence of bleeding did not increase. Accordingly, we have previously shown through intermediate- [22] and long-term studies [8] that the incidence of thromboembolism of the silastic ball valve seemed to drop appreciably after the 6th to the 7th year of implantation with no such change for the cloth covered or track valve [8].
In comparing the results of the present three ball valve types it should be emphasised that the silastic ball valve patients were older and had a higher preoperative prevalences of kidney failure, and endocarditis. Furthermore, the rather small model 1200 silastic ball valve group included the majority of small valves (23 mm). Kidney failure is a well-known risk factor for early mortality after AVR [5,7] but also for thromboembolism [7] perhaps due to a relation to concomitant coronary artery disease [5,7], the latter having a direct relation with both early [26] and late mortality [26] and thromboembolism [5,7]. The present strong influence of high age and factors indicating advanced heart disease preoperatively on both late mortality, thromboembolism, bleeding and complications in general accords well with previous studies [4,5,7,9]. We have previously shown that a prosthetic valve orifice (inner opening) diameter of 15 mm or less was a significant risk factor for late mortality after AVR [4]. An orifice diameter of 15 mm corresponds to a 23 mm silastic ball valve and a 19 mm St. Jude valve [4]. A significant trans-prosthetic pressure gradient and maintenance of left ventricular hypertrophy after AVR springs to mind. In the present analysis the influence of valve size turned out to be more differentiated. In the aortic stenosis group a valve size index of 13 mm/m2 or less was an incremental risk factor both for late mortality, bleeding and all complications. Thus, small valves in a large patient is the factor to avoid. We have previously been able to relate both small valve size index and a dilated left ventricle to a high trans-prosthetic gradient after AVR [6]. As these factors were also related to increased intravascular haemolysis [6,11], we may have an explanation of the present influence of small valve size index on tendency to bleeding complications: increased turbulent flow with shear forces across the patient (and heart) size mismatched too small prosthesis may lead to platelet damage and bleeding proneness in an anticoagulated patient [6].
A large valve size index was a risk factor in the aortic regurgitation group. However, a large valve size index was related to diseases (idiopathic media necrosis, Marfan's syndrome, lues, other systemic diseases) causing a well known dilatation of the aortic root. Accordingly, it has previously been shown that preoperative aortic regurgitation was an incremental risk factor for embolism in a large allograft AVR series where the diseases just mentioned constituted a large fraction of the patients with regurgitation [9].
A special feature of the cloth covered valves is a tendency for increasing risk of embolism [1,8], valve thrombosis [1] and haemolysis [1,11] due to progressive wear of the cloth covering the cage [1]. The present result regarding embolism accords well with these findings but our rates of thrombosis and haemolysis were low. Our cloth covered valve had a rate of endocarditis which were twice that of our silastic ball valves, perhaps related to non-endothelial covered worn teflon. In general, we noted tendencies for increasing rates of all complications and redo AVR after 10 years for both the cloth covered and the track valves, unlike the silastic ball valve. Apart from optimal management of anticoagulant treatment, available remedies in patients with direct valve related complications such as recurrent embolism, valve thrombosis and uncontrolable haemolysis are narrowed down to replacing the valve. Since results after conservative treatment are inexorably poor [27], prosthetic valve endocarditis should also be considered a surgical disease, not least when involving a cloth covered valve with cloth wear. Haemolysis, although rare, seemed to be a serious complication since the patients died from congestive heart failure within a few years. However, the haemolysis may be a secondary factor. Impaired left ventricular function significantly increases the degree of intravascular haemolysis associated with ball valves [11] as well as with disc valves [6]. Furthermore, the silastic ball valve seemed to be involved with the very rare instances of valve thrombosis and haemolysis more than a decennium after AVR. In general, old age and poor condition may preclude redo AVR.
In conclusion, the Starr-Edwards aortic ball valves in general have offered the patients excellent palliation through three decades of life. With a clinical follow-up extending to 31 years, the currently available silastic ball valve is completely durable through the remaining life-time of the average patient and able to secure near normal age and sex specific survival provided valve and patient size mismatch is avoided.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstract1. Introduction2. Materials and methods3. Results4. DiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Shiono, Y. Sezai, A. Sezai, M. Hata, M. Iida, and N. Negishi Long-Term Results of the Cloth-Covered Starr-Edwards Ball Valve Ann. Thorac. Surg., July 1, 2005; 80(1): 204 - 209. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Naito, M. Nakajima, H. Inoue, N. Hibino, E. Mizutani, and K. Tsuchiya Unexpected durability of Smeloff-Cutter aortic ball valve prosthesis Ann. Thorac. Surg., May 1, 2003; 75(5): 1633 - 1635. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. P. Casselman, M. L. Bots, W. Van Lommel, P. J. Knaepen, R. Lensen, and F. E.E. Vermeulen Repeated thromboembolic and bleeding events after mechanical aortic valve replacement Ann. Thorac. Surg., April 1, 2001; 71(4): 1172 - 1180. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Lund, S. L. Nielsen, H. Arildsen, L. B. Ilkjaer, and H. K. Pilegaard Standard aortic St. Jude valve at 18 years: performance profile and determinants of outcome Ann. Thorac. Surg., May 1, 2000; 69(5): 1459 - 1465. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
