The Björk-Shiley valve prosthesis

Surgery for Acquired Heart Disease

The Bjork-Shiley valve prosthesis Analysis of long-term evolution The long-term follow-up of 831 patients who underwent valve replacement with Bjork-Shiley Delrin and standard Pyrolyte prostheses (341 aortic, 345 mitral, and 145 mitroaortic) between 1971 and 1980 is reviewed. The follow-up concluded in 1985. Cumulative follow-up amounted to 4724 years, with a mean of 5.68 years per patient. Data on survival were obtained in 754 patients (complete follow-up in 90.8 % of cases). Perivalvular leak was the most frequent complication in aortic valve replacement, whether isolated or combined, with values significantly higher than those registered in isolated mitral replacement (p < 0.001 in both cases). No correlation was found between this complication and valve calcification, but it was statistically correlated with the size 19 model (p < 0.05). Prosthetic stenosis was more common in mitral than in aortic replacement (p < 0.001~ and of the former, size 23 was that most often affected (p < 0.001). The earliest case of mitral pannus was diagnosed 20 months postsurgery, and from 45 months on this pathology was the cause of every case of stenosis. The risk of thromboembolism was similar in aortic, mitral, and double prostheses, while it was the single most frequent complication in isolated mitral replacement. Prosthetic thrombosis was triggered in all cases in which it occurred by discontinuance of anticoagulant therapy. Anticoagulant-induced hemorrhages were more frequent in double replacement than in mitral (p < 0.05) and aortic valve replacement. Endocarditis was the complication that produced the highest mortality rate in all groups; the frequency of this infection was higher among patients with double prostheses when compared with either aortic replacement (p < 0.05) or mitral replacement (p < 0.001). The risk of suffering endocarditis was correlated with the existence of active preoperative infection in patients with mitral prostheses and double prostheses (p < 0.001 in both cases). Overall morbidity was higher in the double replacement group with respect to the mitral group (p < 0.01). The rate of mortality was also higher among the double valve replacement patients when compared with both the aortic (p = 0.0002) and mitral (p = 0.006) groups. (J 'fHORAC CARDIOVASC SURG 1992;104:1249-58)

L. Alvarez, PhD, FCCP, C. Escudero, PhD, D. Figuera, PhD, FCCP, and 1. L. Castillo-Olivares, PhD, Madrid, Spain

From the Services of Experimental Surgery and Cardiovascular Surgery, Hospital Puerta de Hierro, Madrid, Spain. This study was financed by grant No. 376/81 from the Fondode Investigaciones Sanitarias de la Seguridad Social and by grant No. PB87-0044 from the CICYT, Spain. Received for publication Oct. 14, 1991. Accepted for publication May 12, 1992. Address for reprints: Dr. Lourdes Alvarez, Rafael Bergamin, 9, IO-B, 28043 Madrid, Spain.

12/1/40078

Athough the long-term results of isolated or multiple valve replacement are well known, we consider that an analysis of the morbidity and mortality statistics themselvesis necessary, as it permits us to exercise the required control over the quality of health care in the specific geographic region in which a certain treatment is applied. The Bjork-Shiley prosthesis (Shiley, Inc., Irvine, Calif.), made commercially available at the end of the 60s, 1-3 was first implanted in patients in the Cardiovascu-

1249

12 5

° Alvarez et al.

The Journal of Thoracic and Cardiovascular Surgery

100 90 80 ...J

oct

s>a:

70 60

::::>

50

'0

40

CJl

;,l! 0

379 341 411 345 lB7 145

oJ

0

317 312 121

304 297 111

2

2BO 275 104

3

225 231 B6

4

199 19B 7B

5

160 170 6B

125 122 59

7

8

27 31 13

55 54 19

90 B7 37

9

10

... AVR -e- MVR . . DVR

11

12

13

YEARS

Fig. 1. Actuarial survival curves after aortic (AVR), mitral (MVR), and double valve replacement (DVR) with Bjork-Shiley prostheses. Perioperative mortality and the number of patients at risk at each interval of follow-up are included. lar Surgery Service of Hospital Puerta de Hierro in 1971. Initially, the Delfin model was used until the end of 1972, when the carbonf'yrolyte disc model, used until 1980,was substituted for it. This study focuses on the retrospective analysis of the long-term complications of patients subjected to aortic or mitral valve replacement with these models.

Material and methods This retrospective study deals with 831 consecutive patients who underwent surgery between 1971 and 1980 in the Cardiovascular Surgery Service of the Hospital Puerta de Hierro for valve replacement (341 aortic, 345 mitral, and 145 mitral/aortic) with Bjork-Shiley prostheses (Delrin or standard Pyrolyte carbon models). Follow-up commenced from the moment of the patients' discharge from the hospital and concluded in 1985. The method employed for follow-up was direct examination of the patients in 50% of cases, and telephone or mail contact in the remaining 50%. Those patients for whom it was not possible to attain data for the preceding 12 months were considered to be lost to follow-up. Hospital morbidity was excluded, as the study was designed for the analysis of long-term evolution, but perioperative mortality is given in Fig. I. Aside from the general descriptive data of the series (age, sex, functional class according to New York Heart Association, type of valvular lesion corrected, hemodynamic situation, etc.), the following aspects were basically taken into account: position and size of the implanted prosthesis, complications suffered during the postoperative follow-up, and death and its causes. To define the complications, we have adopted the recommendations of the Ad Hoc Liaison Committee for Standardizing Definitions of Prosthetic Heart Valve Morbidity of The American Association for Thoracic Surgery and The Society of Thoracic Surgeons": Structural deterioration (any change in valve function result-

ing from an intrinsic abnormality causing stenosis or regurgitation) such as wear, stress, fracture, and disc escape. Nonstructural dysfunction (any abnormality resulting in stenosis or regurgitation occurring at the valve, but not intrinsic to the valve itself). Perivalvular leak, diagnosed in the absence of infection, could be determined in all cases by clinical exploration (new cardiac murmurs, increased cardiomegaly, hemolytic anemia, echocardiography, and angiocardiography). In the analysis of this complication, the existence of valvular calcification was considered, as well as the position and size of the prosthesis and the suture technique used in its placement. Entrapment by pannus and inappropriate sizing were grouped within the complication referred to as prosthetic stenosis, since the difference between the two could only be established with certainty in reoperated patients but not in nonreoperated valve recipients. Thromboembolism (any valve thrombosis or embolus in the absence of infection) included any new permanent or transient, focal, or global neurologic deficit, and any peripheral arterial embolus. Prosthetic thrombosis, included in the group of thromboembolic complications, was diagnosed by clinical workup (symptoms, signs, and echocardiography), with no need to resort to further explorations in any case. In the analysis of this complication, it was taken into account whether or not the anticoagulation was correct. Acute myocardial infarction occurring after the operation was considered a separate category, since in most patients there was no proof that the coronary arteries were normal. Anticoagulant-related hemorrhage (any episode of internal or external bleeding that caused death, stroke, operation, hospitalization, or required transfusion). All our patients received anticoagulant medication with dicumarol; the doses were adjusted according to each individual, all of whom were subjected to periodic control (at least once a month) of the prothrombin activity. As in the preceding section, for the analysis of this complication it was taken into consideration whether or not the anticoagulation was correct.

Volume 104 Number 5 November 1992

Bjork-Shiley prostheses

12 5 1

Table I. General characteristics DVR

MVR

AVR Surgical procedure No. of patients Hospital mortality (%)

379 10.0

411 16.0

187 22.4

341 204 (59.8) 137 (40.2)

345 142 (41.2) 203 (58.8)

145

40.5 ± 12.1 10-68

40.8 ± I\.6 5-79

39.4 ± 12.4 9-77

Follow-up No. of patients Male (No. & %) Female (No. & %) Age Mean ± SD Range

77 (53.1) 68 (46.9)

AVR, Aortic valve replacement; MVR, mitral valve replacement; DVR, double valve replacement.

Table II. Prevalence ofperivalvular leak and stenosis according to prosthesis size in AVR Single

Double

Size (mm)

Total

Leak

Stenosis

Total

Leak

Stenosis

19 21 23 25 27 29 31

5 79 114 86 50 6 I

2 12 17 7 6 0 I

0 2 0

7 47 63 26 2

2 6 5 3 I

0 0 0 0 0

I I 0 0

Abbreviations as in Table I.

Prosthetic valve endocarditis (any infectioninvolving a heart valvereplacement) was diagnosedon the basis of the usual clinical signs, positive blood cultures, or histologicconfirmation at reoperation. In the analysis of this complication, it was taken into account whether or not the subjects were acutely infected at the time of valve replacement. With respect to the causes of death, the concept "other" refers to those causes (accident, suicide, neoplasia, etc.) not relatedto cardiac diseaseor to the prosthesis.Nevertheless,their mentionhas not been omitted because we have not been able to ascertain the role that possiblepsychologic disorders and anticoagulation therapy may have played in these deaths. In the treatment of the data using general statistical programs.Y' the total number of observations, the mean, the standard deviation, the standard error of the mean, and the maximum and minimum values were recorded for each variable, as well as the frequency distribution of each of the codes or strata established for each variable. In the comparisonofthe frequenciesofthedifferent variables, the chi square contrast was used, and those differences with a p value of less than 0.05 were considered significant. The probability of remaining free of a given complication duringfollow-up was calculated using the actuarial method for large samples and the Kaplan-Meier method for short series. Forcomparisons betweencurves,the Mantel-Cox, Breslow,and log-rank regression models were applied.

Results The cumulative follow-up of the 831 patients included in this series was 4724 years, which results in a mean fol-

low-up of 5.68 years per patient. Follow-up was completed in 754 cases (90.8%) and was interrupted in 77 (9.2%). The data concerning the position of the implanted prostheses, age, and sex of the patients in this series are reflected in Table I, and the survival curves, calculated according to the actuarial method, including the perioperative mortality and the number of patients at risk at each interval, appear in Fig. 1. With respect to perioperative mortality, it should be pointed out that our group did not begin to use cardioplegic solutions for myocardial protection until 1980. The Delrin prosthesis was used in 21 aortic valve replacements (AVR), in 33 mitral valve replacements (MVR), and in 23 double (aortic and mitral) valve replacements (DVR). In the remainder, the standard carbon Pyrolyte model was implanted. There were no significant differences in survival or incidence of complications between the two prosthetic models. The size and position of the valve replacements are shown in Tables II and III. With respect to the complications detected during follow-up, their actuarial incidence appears in Figs. 2 to 6. However, the particulars of each complication require more ample description. Structural dysfunction. No cases of structural dysfunction were detected.

12 52

The Journal of Thoracic and Cardiovascular Surgery

Alvarez et al.

%

100 90 80 70 60

LEAK (L)

50

% EPISODES A % EPISODES M 8 % EPISODES AM . . AWITHOUTL -0- M WITHOUT L . . AM WITHOUT L •

o

40 30 20 10 0 2

3

4

5

6

7

8

9

10

11

12

13

YEARS

Fig. 2. Actuarial curves of patients free of perivalvular leak and the percentage of episodes diagnosed at each interval of follow-up.

Table III. Incidence ofperivalvular leak and stenosis according to prosthesis size in MVR Single

Double

Size (mm)

Total

Leak

Stenosis

Total

Leak

Stenosis

23 25

2 62 158 97 26

0

2 3 8 2 2

2 28 68 41 6

0 0

0 0 0 0 0

27

29 31

3 5 6 3

2 0 0

Abbreviations as in Table 1.

Perivalvular leak. All the prostheses were implanted with interrupted 2-0 sutures. When the aspect of the valve ring warranted, the stitches were anchored in Teflon patches, but this information could not be reconstructed for each separate case, nor were the number of stitches used for each valve implant available. All the patients presented varying degrees of hemolysis, but this did not influence per se the surgical indication in any case. In 45 of the patients subjected to AVR there were 47 episodes of perivalvular leak. Ten underwent reoperation for this complication, and in another five patients it was the cause of death. No relationship was detected between valvular calcification and perivalvular leak, but it was observed that this complication was more frequent in size 19 prostheses (p < 0.001) (Table 11). In 17 of the patients with MVR, 18 episodes of perivalvular leak were diagnosed. This resulted in reoperation in 13 cases and in death in four. There was no relationship between this complication and valvular calcification or prosthesis size (Table III). In 19 of the patients with DVR, 17 episodes of perivalvular leak were detected in the aortic prosthesis and two

episodes were seen with the mitral prosthesis. This required reoperation in four cases and was the cause of death in two. This complication was not correlated with valvular calcification. While it was more frequent in size 19 aortic prostheses (p < 0.05), size had no bearing on perivalvular leak in the mitral model (Tables II and III). The risk of developing perivalvular leak was significantly greater in AVR and DVR when each was compared with MVR (p < 0.001 in both cases), but no significant differences were found between the first two groups. With respect to the time in which diagnosis was established, there is a preferential distribution of onset that peaks in the second year for mitral valve leak and in the fourth year for leakage occurring in the AVR; although the frequency later diminishes, perivalvular leakage continues to appear throughout the entire follow-up period (Fig. 2). Prosthetic stenosis. Among the patients with AVR, four exhibited residual stenosis of their prostheses, one during the first year of follow-up and the remaining three from year 5 on (Fig. 3). No patient required reimplanta-

Bjork-Shiley prostheses

12 5 3

STENOSIS (S) •

% EPISODES A % EPISODES M 8 % EPISODES AM . . AWITHOUTS -0- M WITHOUT S . . AM WITHOUT S

o

5

6

7

8

9

10

11

12

13

YEARS

Fig. 3. Actuarial curves of patients free of stenosis and the percentage of episodes diagnosed at each interval of follow-up.

tion or died as a result of residual stenosis. No relationship whatsoever to prosthesis size was detected (Table II). Residual stenosis was found in 17 patients with MVR, being the motive for reoperation in 10. In two cases the valvesize was unsuitable and in the remaining eight pannus was present (Table IV). This complication was the cause of death in two patients. The two size 23 prostheses were stenotic (p < 0.001) (Table III), although the patients were not subjected to reimplantation because the hemodynamic disorder produced by the stenoses was consideredto be mild or moderate. This complication was diagnosed regularly throughout the entire follow-up period (Fig. 3). It is noteworthy that the earliest case of pannus did not appear until 20 months after surgery, and from 45 months on all cases of stenosis were the result of this pathologic condition (Table IV). The risk of prosthetic stenosis was significantly greater in MVR than in AVR (p < 0.001), while in DVR not a single case was detected. Thromboembolism. In the AVR group, 16 subjects suffered a total of 20 thromboembolic episodes, these being the cause of repeat surgery in six and of death in one. Of these episodes, five were found to be prosthetic thromboses. Among the patients with MVR, 26 experienced 29 thromboembolic episodes, resulting in reoperation in eight and death in three. Nine of these episodes were prosthetic thromboses. In the group with DVR, six patients presented seven episodes of thromboembolism, one of which required reoperation and another of which was the cause of death. Only one of them was a prosthesis-related thrombosis.

Table IV. Patients subjected to surgery for stenosis of mitral valve prostheses Prosthesis size (mm)

TimeSurgDiag(mo)

Findings at reoperation

31

5 20 25 43 45 63 95 103 108 115

Inappropriate size Pannus Pannus Inappropriate size Pannus Pannus Pannus Pannus Pannus Pannus

27

27 27 27 31 25 25 29 29 Surg, Surgery; Diag, diagnosis.

In 53 patients, control of anticoagulation with dicumarol was inadequate; in fact, in all patients in whom prosthetic thrombosis arose this treatment was discontinued, whether arbitrarily or for gynecologic indications. When there were gynecologic reasons for not using dicumarol, sodium heparin was substituted. In the remainder of the thromboembolic episodes, no correlation with inadequate anticoagulation was found. There are no significant differences among the three groups with respect to the risk of thromboembolism (Fig. 4). Thromboembolic complications were more frequent during the first 4 years of follow-up, and then remained constant until the tenth year, after which no further episodes were detected (Fig. 4). Acute myocardial infarction occurred in 13 patients in

The Journal of Thoracic and Cardiovascular Surgery

1 2 5 4 Alvarez et al.

THROMBOEMBOLISM (T) % EPISODES A % EPISODES M ~ % EPISODES AM ... AWITHOUTT -0- M WITHOUT T . . AM WITHOUT T •

o

3

4

5

678

9

10

11

12

13

YEARS

Fig. 4. Actuarial curves of patients free of thromboembolism and the percentage of episodes diagnosed at each interval of follow-up.

%

100

90

80

70 60

HEMORRHAGE (H)

50

•

o

% EPISODES A

% EPISODES M EI % EPISODES AM ... AWITHOUTH -0- M WITHOUT H . . AM WITHOUT H

40 30 20 10

o-+-....I::::I.......I:::l,IR...I::i,IIL.J::~U::::I......J::::1.,a...Jt:::I.,IL.J:::l.,IIL.J:::L,--...-'-J.......,.-~-~ 2

3

4

5

6

7

8

9

10

11

12

13

YEARS

Fig. 5. Actuarial curves of patients free of anticoagulant-induced hemorrhage and the percentage of episodesdiagnosed at each interval of follow-up.

the AVR group, in eight among those with MVR, and in two in the DVR group. It was the cause of death in one patient with MVR. No statistically significant relationship was found between this complication and the variables being considered here. Anticoagulant-induced hemorrhage. In the AVR group, 28 patients suffered 29 hemorrhagic episodes, which was the cause of death in one of them. Among the MVR group, 17 patients had 18 episodes of hemorrhage, resulting in the death of four. In the DVR group, 16 patients suffered 17 episodes, which caused death in three cases. As in the study of thromboembolism, no correlation

could be established between anticoagulation therapy and the risk of hemorrhage. However, comparison of the groups with one another (Fig. 5) revealed that the risk of hemorrhage was greater among patients with DVR compared with those with MVR (p < 0.05), possibly because among the former the degree of anticoagulation had been maintained in the higher limits of the therapeutic range. The risk of hemorrhage was constant throughout the entire follow-up period (Fig. 5). Prosthetic valve endocarditis. Of the patients who underwent AVR, in eight cases an emergency operation was necessary because of active endocarditis. During follow-up, 16 patients presented endocarditis infection,

Volume 104 Number 5 November 1992

Bjork-Shiley prostheses

%

I 255

100 90 80 70 60

ENDOCARDITIS (E)

50

% EPISODES A % EPISODES M § % EPISODES AM . . AWITHOUTE -0- M WITHOUT E ... AM WITHOUT E •

D

40 30 20 10 0 2

3

4

5

6

7

8

9

10

11

12

13

YEARS

Fig. 6. Actuarialcurves of patientsfree of endocarditis and the percentage of episodes diagnosed at each interval of follow-up.

which required reoperation in seven cases and resulted in death in eight. No relationship was found between the risk of prosthetic valveendocarditis and the previous existence of active infection. In the MVR group, four patients underwent emergency operation because of active endocarditis. During follow-up, endocarditis was detected in six patients, resulting in reoperation in three and in death in four. The existence of preoperative active endocarditis was statistically correlated with the risk of its recurrence as a postimplant complication (p < 0.001). Among those individuals with DVR, II had active endocarditis and required emergency surgery. In the long-term evolution of this group, 14 patients presented with endocarditis that affected one or both prostheses, a complication that made reoperation imperative in II patients and was the cause of death in another II. A background of preoperative active endocarditis was correlated with the risk of prosthetic valve endocarditis arising after valve replacement (p < 0.001). The risk of endocarditis (Fig. 6) was similar in the AVR and MVR groups. However, the probability of endocarditis was greater in DVR patients when compared both with AVR (p < 0.05) and with MVR patients (p < 0.001). With respect to the time of diagnosis, the majority of episodes were grouped in the first 3 years of follow-up, and from the seventh year on further cases of endocarditis were detected only in those with DVR (Fig. 6). Considering the complications as a whole, in AVR 102 patients (29.9%) suffered a total of 129 episodes. In MVR, 84 patients (24.3%) presented 96 episodes of

complications. Finally, in the DVR group, 53 patients (36.5%) suffered 59 episodes of the aforementioned complications. The risk of each of these complications was greater among patients with DVR compared with those with MVR (p < 0.01); however, there were no significant differences between those with DVR and those with AVR. Causes of death. In the AVR group, aside from the 15 patients who died as a result of the complications described above, in another 27 cases death was produced by cardiac insufficiency in eight, was sudden in six, was of unknown cause in seven, and was the result of other causes in six. Of the MVR patients, 18 died of these complications and another 38 succumbed because of cardiac insufficiency in four, sudden death in 12, of an unknown cause in 15, and of other causes in seven. Among those with DVR, aside from the 17 patients in whom the complications described above resulted in death, 19 died of cardiac insufficiency (four cases), sudden death (six cases), an unknown cause (seven cases), and of other causes (two cases). All of the deaths, regardless of the cause, are included in the survival curves. With respect to survival (Fig. 1), there were no significant differences between AVR and MVR (p = 0.2). Nevertheless, the risk of death was higher in DVR when compared with either AVR (p = 0.0002) or MVR (p = 0.006). Discussion In evaluating this study, we consider that its main limitation may be that most of the data were gathered ret-

12 5 6 Alvarez et al.

rospectively. However, aside from the size of the samples analyzed and the mean postoperative follow-up times, other data may be indicative of the quality control to which the patients were subjected. First among these might be the high percentage of patients with complete follow-up (90.8%). It is worthy of mention that in most publications the percentage of patients lost to follow-up is only explicitly cited when it is less than 10%.2,3,8-17 In our geographical area, only Martinell and associates'f assessed the proportion of patients lost to follow-up in retrospective studies as being between 19% and 22%, a level much higher than that recorded by us, Another evidence of the rigorous control exercised in our study is the distribution of the complications throughout postoperative follow-up, These are not exclusively grouped in the early years (which would indicate that the quality control of patients had diminished with the passage of time) but were detected throughout the entire duration of the follow-up period. These complications were of a type in which the risk supposedly remains constant, as is the case with anticoagulant-induced hemorrhage. These aspects led us to consider that the results obtained from the series analyzed were reasonably reliable. Although the Bjork-Shiley prosthetic models implanted in the patients dealt with in this study can be considered historical, as they are no longer commercially available, we consider that the long-term analysis of the complications may be of interest because there are still patients with these prostheses and because the risk of some of these complications does not depend strictly on the model implanted. It must be pointed outthat, as in other series.v 3,13,19,20 we have detected no cases of primary dysfunction of the Bjork prosthesis, either in the Delrin model or in the standard Pyrolyte model. In addition, there has been no significant hemolysis in the absence of prosthetic dysfunction. 20-23 Perivalvular leak was the most frequent complication among those with AVR and DVR (in the latter because of aortic involvement) and the primary cause of reoperation in the AVR and MVR groups. These data agree with those reported in other series.21, 24-27 In our experience, in contrast with other investigations" no relationship was detected between valve ring calcification and the incidence of perivalvular leak, possibly because the surgeon took additional precautions such as anchoring the stitches on Teflon patches. On the other hand, in this study, perivalvular leak was found to be especially frequent in size 19 prostheses. As in other series,28 this complication was diagnosed principally in the initial follow-up

The Journal of Thoracic and Cardiovascular Surgery

periods, but the greater frequency of perivalvular leak in cases with AVR when compared with MVR and the late presentation of this complication in a certain number of patients, leads to a suspicion that other factors, aside from those inherent in the surgical procedure, playa role in the production of perivalvular leak, Stenosis may originate from two different mechanisms: (1) because of inappropriate size of the prosthesis.P 29-34 in which case its detection is usually early and (2) because of the formation of overgrowth tissue (pannus) on the sewing ring.35-38 It has been reported that it takes at least 10 months from the prosthesis implantation for the production of this pannus.F' and its etiology has been related to the hydrodynamic design of Bjork valve prostheses.P In this study, the importance of the time factor was confirmed, as was the greater frequency of stenosis among patients with MVR. The incidence of thromboembolism in our series remained within the range described in the literature. * It was the most frequent complication in the MVR group, a fact that can feasibly be related to hemodynamic factors. 21 Prosthetic thrombosis was also more frequent in this group, and in our experience the determining factor for its production was the discontinuation of dicumarol, even in those cases in which this agent was replaced by heparin. Anticoagulation-induced hemorrhages, the incidence of which was similar to that recorded by other groups,t occurred regularly throughout the entire follow-up period, a fact that could be foreseen, since at least theoretically the risk is constant in these patients. These hemorrhages were more frequent among those with DVR, possibly because in this group as a result of the implantation of two valves, anticoagulation was maintained in the upper limits of the therapeutic range. This treatment counteracts the benefit of prophylaxis from thromboembolic accidents. The frequency of bacterial endocarditis among our subjects was very similar to that registered in similar series.f and the risk of its occurrence was greater among MVR and DVR recipients in whom implantation coincided with active infection of their native valve-': 52; however, this was not the case with the AVR recipients. This is the most serious complication threatening heart valve prosthesis recipients. In our patients, it resulted in a mortality of 53% in AVR, 66% in MVR, and 78% in DVR, percentages that fall within the range of those reported in the literature.P- 43, 52-54 Among DVR patients, bacterial endocarditis was also the primary cause for reoperation. *References8, 10-14, 17,21,22,24,25,35,39-48. [References 12, 17,21,22,24,29,42,48. tReferences 9-11, 18,24,25,39,48-50.

Volume 104 Number 5 November 1992

It was found to be more frequent in the first stages offollow-up, 17,52 while from the seventh year on it continued to be detected only in the DVR group. The fact that there are no significant differences in the overall morbidity of the DVR group and that of the group with A VR is a result of the frequency with which a perivalvular leak arises. It was the first complication in both populations. However, it was made clear that the prognosis for morbidity as well as for mortality is better in groups with single valve replacement than in those with

DVR. We thank Drs. R. Martinez, I. Milhin, and M. Alonso of the Biostatistics Section for the statistical study; all Services of Hospital Puerta de Hierro for making available the clinical histories of their patients; and Ms. M. Messman for her translation of the article. REFERENCES I. Goldfaden DM, Silverman NA, Levitski S. Selection of a cardiac valve prosthesis. Cardiovasc Perspect 1986;1:3-6. 2. Montero CG, Rufilanchas JJ, Juffe A, Burgos R, Ugarte J, Figuera D. Long-term results of cardiac valve replacement with the Delrin-disc model of the Bjork-Shiley valve prosthesis. Ann Thorac Surg 1984;37:328-36. 3. Lindblom D, Bjork VO, Semb BKH. Mechanical failure of the Bjork-Shiley valve. Incidence, clinical presentation, and management. J THORAC CARDlOVASC SURG 1986;92:894907. 4. Edmunds LH, Clark RE, Cohn LH, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. J THORAC CARDlOVASC SURG 1988;96:351-3. 5. Benedetti J, Yaen K, Young L. Life tables and survival functions. In: BMDP manual. Berkeley: University of California Press, 1985:557-75. 6. Lee ET. Statistical methods for survival data analysis. Belmont, Calif.: Lifetime learning publications, 1980. 7. Siegel S. Estadistica no parametrica. Mexico City: Ed. Trillas, 1978. 8. Bjork VO, Henze A. Results five to seven years after aortic valve replacement with the original Delrin disc model Bjork-Shiley prosthesis. Scand J Thorac Cardiovasc Surg 1977;11:177-80. 9. Borkon AM, Soule L, Baughman KL, et al. Ten-year analysis of the Bjork-Shiley standard aortic valve. Ann Thorac Surg 1987;43:39-51. 10. Cohn LH, Allred EN, Cohn LA, et al. Early and late risk of mitral valve replacement. A 12-year concomitant comparison of the porcine bioprosthetic and prosthetic disc mitral valves. J THoRAc CARDlOVASC SURG 1985;90:87281. II. Cheung D, Flemma RJ, Mullen DC, Lepley D Jr, Anderson AJ, Weirauch E. Ten-year follow-up in aortic valve replacement using the Bjork-Shiley prosthesis. Ann Thorac Surg 1981;32:138-45.

Bjork-Shiley prostheses

I 257

12. Daenen V, Nevelsteen A, Van Cauwelaert P, De Haesschalk E, Willems J, Stalpaert G. Nine years' experience with the Bjork-Shiley prosthetic valve: early and late results of 932 valve replacements. Ann Thorac Surg 1983;35:65163. 13. Joyce LD, Nelson RM. Comparison of porcine valve xenografts with mechanical prostheses. A 7lf2-yearexperience. J THORAC CARDlOVASC SURG 1984;88:102-13. 14. Mattingly WT Jr, O'Connor W, Zeok JV, Todd EP. Thrombotic catastrophe in the patient with multiple BjorkShiley prostheses. Ann Thorac Surg 1983;35:253-6. 15. Murphy DA, Levine FH, Buckley MJ, et al. A comparative analysis of the Starr-Edwards and Bjork-Shiley prostheses. J THoRAc CARDlOVASC SURG 1982;86:746-52. 16. Heck HA, Wright CB, Doty DB, Rossi NP, Ehrenhaft JL. Combined multiple-valve procedures: a five-year experience with 125 patients. Ann Thorac Surg 1979;27:320-7. 17. Lindblom D. Long-term clinical results after mitral valve replacement with Bjork-Shiley prosthesis. J THORAC CARDlOVASC SURG 1988;95:321-33. 18. Martinell J, Fraile J, Artiz V, Moreno J, Rabago G. Longterm comparative analysis of the Bjork-Shiley and Hancock valves implanted in 1975. J THORAC CARDlOVASC SURG 1985;90:741-9. 19. Marbarger JP Jr, Clark RE. The clinical life history of explanted prosthetic heart valves. Ann Thorac Surg 1982; 34:22-33. 20. Roberts WC, Hammer WJ. Cardiac pathology after valve replacement with a tilting disc prosthesis (Bjork-Shiley type). A study of 46 necropsy patients and 49 Bjork-Shiley prostheses. Am J Cardiol 1976;37:1024-33. 21. Kirklin JW, Barratt-Boyes BG. Acquired valvular heart disease. In: Kirklin JW, Barratt-Boyes BG, eds. Cardiac surgery. New York: John Wiley & Sons, 1986:323-459. 22. Lefrak EA, Starr A. Bjork-Shiley valve. In: Lefrak EA, Starr A, eds. Cardiac valve prostheses. New York: Appleton-Century-Crofts, 1979:215-45. 23. Bjork VO, Henze A, Carlstrom A. Haematological evaluation of the Bjork-Shiley tilting disc valve prosthesis in isolated aortic valvular disease. Scand J Thorac Cardiovasc Surg 1974;8:12-22. 24. Horstkotte D, Korfer R, Seipel L, Bircks W, Loogen F. Late complications in patients with Bjork-Shiley and St. Jude Medical heart valve replacement. Circulation 1983; 68(suppl II): 175-84. 25. NashefSAM, Sethia B, Turner MA, Davidson KG, Lewis S, Bain WHo Bjork-Shiley and Carpentier-Edwards valves. A comparative analysis. J THORAC CARDlOVASC SURG 1987;93:394-404. 26. Figuera D, Montero CG. Techniques in mitral valve implantation. In: Rabago G, Cooley DA, eds. Heart valve replacement: current status and future trends. Mount Kisco, NY: Futura Publishing Co, 1987:55-67. 27. Maroiias JM, Rufilanchas JJ, Villagra F, Tellez G, Agosti J, Figuera D. Reoperation for dysfunction of Bjork-Shiley mitral disc prosthesis. Report of eight cases. Am Heart J 1977;93:316-20. 28. Orszulak TA, Schaff HV, Danielson GK, Pluth JR, Puga

1 2 5 8 Alvarez et al.

FJ, Piehler JM. Results of reoperation for periprosthetic leakage. Ann Thorac Surg 1983;35:584-9. 29. Bjork va, Holmgren A, Olin C, Ovenfors CO. Clinical and haemodynamic results of aortic valve replacement with the Bjork-Shiley tilting disc valve prosthesis. Scand J Thorac Cardiovasc Surg 1971;5:177-91. 30. Bove EL, Marvasti MA, Potts JL, et al. Rest and exercise hemodynamics following aortic valve replacement. A comparison between 19 and 21 mm lonescu-Shiley pericardial and Carpentier-Edwards porcine valves. J THORAC CARDlOVASC SURG 1985;90:750-5. 31. Foster AH, Tracy CM, Greenberg GJ, McIntosh CL, Clark RE. Valve replacement in narrow aortic roots: serial hemodynamics and long-term clinical outcome. Ann Thorae Surg 1986;42:506-16. 32. Horstkotte D, Haerten K, Seipel L, et al. Central hemodynamics at rest and during exercise after mitral valve replacement with different prostheses. Circulation 1983; 68(suppl II):161-8. 33. Mori T, Kawashima Y, Kitamura S, Nakano S, Kawachi K, Nakata T. Results of aortic valve replacement in patients with a narrow aortic annulus: effects of enlargement of the aortic annulus. Ann Thorac Surg 1981;31:111-6. 34. Schaff HV, Borkon AM, Hughes C, et al. Clinical and hemodynamic evaluation of the 19 mm Bjork-Shiley aortic valve prosthesis. Ann Thorac Surg 1981;32:50-7. 35. Wright JO, Hiratzka LF, Brandt B III, Doty DB. Thrombosis of the Bjork-Shiley prosthesis. Illustrative cases and review of the literature. J THoRAc CAROIOVASC SURG 1982;84:138-44. 36. Cleveland JC, Lebenson 1M, Dague JR. Early postoperative development of aortic regurgitation related to pannus ingrowth causing incomplete disc seating of a Bjork-Shiley prosthesis. Ann Thorac Surg 1982;33:496-8. 37. Gosalbez F, Cofiiio JL, Naya JL, De Linera FA. Pannus ingrowth on a Bjork-Shiley prosthesis. Ann Thorac Surg 1983;36:121-2. 38. Yoganathan AR, Corcoran WH, Harrison EC, Carl JR. The Bjork-Shiley aortic prosthesis: flow characteristics, thrombus formation and tissue overgrowth. Circulation 1978;58:70-6. 39. Gersh BJ, Fisher LD, Schaff HV, et al. Issues concerning the clinical evaluation of new prosthetic valves. J THORAC CAROIOVASC SURG 1986;91:460-6. 40. Bjork va, Henze A. Isolated mitral valve replacement with the Bjork-Shiley tilting disc valve prosthesis. A six-year review and a comparison between the Delrin and the Pyrolytic carbon disc models. Scand J Thorac Cardiovasc Surg 1977;11:181-5. 41. Bjork va, Henze A. Ten years' experience with the BjorkShiley tilting disc valve. J THORAC CARDIOVASC SURG 1979;78:331-42. 42. Edmunds LH Jr. Thromboembolic complications of cur-

The Journal of Thoracic and Cardiovascular Surgery

rent cardiac valvular prostheses. Ann Thorac Surg 1982; 34:96-106. 43. Jones EL, Schwarzmann SW, Check WA, Hatcher CR Jr. Complications from cardiac prostheses. I. Infection, thrombosis and emboli associated with intracardiac prostheses. In: Sabiston DC Jr, Spencer FC, eds. Gibbon's surgery of the chest. Philadelphia: WB Saunders Company, 1983: 1253-66. 44. Karp RB, Cyrus RJ, Blackstone EH, Kirklin JW, Kouchoukos NT, Pacifico AD. The Bjork-Shiley valve. Intermediate-term follow-up. J THORAC CAROIOVASC SURG 1981;81:602-14. 45. Macmanus Q, Grunkemeier GL, Lambert LE, Teply FJ, Harlan BJ, Starr A. Year of operation as a risk factor in the late results of valve replacement. J THORAC CARDIOVASC SURG 1980;80:834-41. 46. de Teresa E, Vera A, Ortigosa J, Catalan F, Ferrer J, de Artaza M. Tromboembolismo y protesis de Bjork-Shiley. N uestra experiencia con un sistema propio de control de la anticoagulacion a largo plazo en 104 enfermos. Rev Esp Cardiol 1979;32:156-60. 47. Tandon AP, Sengupta SM, Lukacs L, lonescu MI. Longterm clinical and hemodynamic evaluation of the lonescuShiley pericardial xenograft and the Braunwald-Cutter and Bjork-Shiley prostheses in the mitral position. J THoRAc CAROIOVASC SURG 1978;76:763-70. 48. Vallejo JL. Estudio clinico, prospectivo y randomizado comparando las protesis de Bjork-Shiley, Lillehei-Kaster y Angell-Shiley en posicion mitral. Rev Esp Cardiol 1981; 34:457-63. 49. Cohn LH, Allred EN, DiSesa VJ, Sawtelle K, Shemin RJ, Collins JJ Jr. Early and late risk of aortic valve replacement. A 12-year concomitant comparison of the porcine bioprosthetic and tilting disc prosthetic aortic valves. J THORAC CAROIOVASC SURG 1984;88:695-705. 50. Rahimtoola SH. Valve replacement should not be performed in all asymptomatic patients with severe aortic incompetence. J THoRAc CAROIOVASC SURG 1980; 79:163-72. 51. D' Agostino RS, Miller DC, Stinson EB, et al. Valve replacement in patients with native valve endocarditis: what really determines operative outcome? Ann Thorac Surg 1985;40:429-38. 52. Ivert TSA, Dismukes WE, Cobbs CG, Blackstone EH, Kirklin JW, Bergdahl LAL. Prosthetic valve endocarditis. Circulation 1984;69:223-32. 53. Calderwood SB, Swinski LA, Karchmer A W, Waternaux CM, Buckley MJ. Prosthetic valve endocarditis. Analysis of factors affecting outcome of therapy. J THORAC CARDIOVASC SURG 1986;92:776-83. 54. Masur H, Johnson WD Jr. Prosthetic valve endocarditis. J THORAC CAROIOVASC SURG 1980;80:31-7.