J Thorac Cardiovasc Surg 2006;132:595-601
© 2006 The American Association for Thoracic Surgery
Surgery for Acquired Cardiovascular Disease |
Clinical results of Hancock II versus Hancock Standard at long-term follow-up
Carlo Valfrè, MDa,
Giulio Rizzoli, MD, FETCSb,*,
Claudio Zussa, MDc,
Paolo Ius, MDa,
Elvio Polesel, MDc,
Salvatore Mirone, MDa,
Tomaso Bottio, MDb,
Gino Gerosa, MDb
a Cardiac Surgery Unit, Cà Foncello Hospital of Treviso, Treviso, Italy
b Cardiac Surgery Unit, University of Padova, Padova, Italy
c Cardiac Surgery Unit, Umberto I° Hospital of Venice, Venice, Italy.
Received for publication October 28, 2005; revisions received February 10, 2006; accepted for publication March 21, 2006.
* Address for reprints: Giulio Rizzoli, MD, FETCS, Cardiochirurgia, Via Giustiniani 2 35128, Padova, Italy (Email: giulio.rizzoli{at}unipd.it).
 |
Abstract
|
|---|
OBJECTIVE: We performed a multi-institutional study to compare the long-term structural valve deterioration of isolated Hancock Standard versus Hancock II bioprostheses.
METHODS: From 1983 to 2002, 714 Hancock Standard and 1293 Hancock II bioprostheses were implanted at hospitals of the Venetian territory (Padova, Treviso, and Venice). Follow-up on January 1, 2003, included 14,749 patient-years with a median of 12 years and was 96% complete: 115 Hancock Standard and 53 Hancock II bioprostheses were at risk at 15 years. The 2 series were nonconcomitant, and many covariates differed (Table 1). Survival was analyzed with Cox analysis, and durability was analyzed with Weibull analysis. Balancing analysis with the logistic propensity score model was performed.
RESULTS: Perioperative mortality was 6% in Hancock II and 12% in Hancock Standard operations. The overall unadjusted 15-year survival was identical (39.7% ± 2.3% vs 39.9% ± 2.4%, respectively), but age-adjusted survival at 15 years was 46% versus 25% (P < .001). Late survival was unrelated to the prosthetic model, whereas it was adversely affected by older age, previous operations, aortic regurgitation, male sex, higher New York Heart Association class, atrial fibrillation, and coronary artery bypass grafting. In Hancock II patients aged 65 years and older, the cumulative hazard of structural valve deterioration at 15 years was 6%, versus 17.5% in Hancock Standard patients. In younger patients, it was 18% and 37%, respectively. Analysis of 541 propensity-balanced patients showed a hazard ratio of the Hancock Standard prosthesis of 2 and a risk reduction of older age of approximately 10% every 10 years.
CONCLUSION: After balancing risk factors and calibrating age effects, Hancock II propensity-matched bioprostheses showed similar survival but definitely increased durability.
Abbreviations and Acronyms AF = atrial fibrillation; HII = Hancock II; HST = Hancock Standard; NYHA = New York Heart Association; SVD = structural valve deterioration
|
Introduction
|
|---|
| |
Drs Valfrè, Mirone, Rizzoli, Gerosa, Bottio, and Salvador (left to right)
|
|
The Hancock Standard (HST) porcine prosthesis was one the most commonly used biologic valve substitutes in the 1970s. Like other first-generation bioprostheses, it had the disadvantage of limited durability. In 1982, the Hancock II (HII) valve was introduced with a lower implant profile, a 2-stage fixation process (initial low pressure and late high pressure), and a calcium-retarding agent (sodium dodecylsulfate). Contemporarily, surgeons revised their clinical indications and therefore limited bioprosthetic use to older patients (
65 or 70 years) with a lower life expectancy and, presumably, insufficient time to develop structural valve deterioration (SVD). Age-related natural death, premature death, and valve replacement for other causes confound the estimate of improvement of durability.
With the aim of verifying the durability improvement of this second-generation tissue valve and its relationship with operative age and valve position, the clinical results of HII prostheses inserted at Padova, Treviso, and Venice were propensity-matched and compared with the results of a historic series of first-generation HST prostheses implanted in the 1970s at the University of Padova.
|
Patients and Methods
|
|---|
Study Group
The study group included 1931 patients who received 2007 isolated aortic (n = 998) or mitral (n = 1009) prostheses. Patients with tricuspid plastic repairs were included. There were 1293 HII and 714 HST valves. Twenty-seven HST patients received a second HST, and 30 received an HII prosthesis. Seventeen HII patients received a second HII, and 2 patients received a third one. Treviso contributed 984 records, Venezia 129, and Padova 894. All the HST valves were inserted in Padova, and 77% of the HII valves were inserted in Treviso; 62% of the mitral valves were operated on in Padova, and 65% of the aortic valves were operated on in Treviso. HST valves were implanted from 1970 to 1984, and HII valves were implanted from 1983 to 2002. The mean age was 48 ± 12 years in HST patients (median, 49 years) and 67 ± 8 years in the HII group (median, 69 years). Covariate distribution is summarized in Table 1.
Follow-up
Follow-up was conducted trough mail or telephone interview, was closed on January 1, 2003, and was 96% complete. The total follow-up included 14,749 patient-years: HII patients had a total follow-up of 8520 patient-years, and HST patients had a total follow-up of 6229 patient-years, with a median of 12 years in both series. Data analysis is reported at 15 years, with 53 HII and 115 HST patients at risk.
Operative Technique
Implantation used pledgeted sutures in a subannular position. The choice of prosthetic size was according to patient size. A single 19 mm prosthesis and 29 label 21 mm prostheses were used. The median valve label size was 23 mm in aortic HII and 25 mm in HST (P = .017), and it was 29 mm in mitral HII and 31 mm in HST (P = .08). The HST mean clamp time was 45 ± 19 minutes, and cardiopulmonary bypass time was 72 ± 33 minutes—significantly shorter (P < .001) than in the HII series (79 ± 29 minutes and 112 ± 41 minutes, respectively).
Outcomes
The focus of the study was mortality, reoperation, and SVD, recorded according to guidelines set forth by the American Association for Thoracic Surgery.1
Other valve-related events were considered secondary.
Statistical Analysis
Patients were censored when the valve was replaced, unless they died at reoperation. Stata for Windows was used for statistical analyses (Stata Corp, College Station, Tex).
Univariate Analysis
Differences in prognostic variables between HST and HII valves were evaluated by t tests for continuous variables and by the Fisher exact test for categorical variables. Distributional differences were evaluated by nonparametric bootstrapped quantile regression analysis. Time-related events were analyzed with the Kaplan-Meier method. Secondary valve-related events are reported as linearized rates for practical reasons.
Actual freedom from SVD and from reoperation was evaluated with the method previously described by Grunkemeier and colleagues.2 A simultaneous decrement analysis of the events survival with the original prosthesis, death, and SVD was performed by using the algorithm of Anderson.3
Multivariable Analysis
The preoperative variables tested for their univariate and multivariable association with the outcomes were valve position (mitral or aortic), demographics (age and sex), valve pathology (insufficiency, mixed lesion, stenosis, active endocarditis, healed endocarditis, rheumatic, degenerative, ischemic, prosthetic pathology, other pathology, precedent SVD event, and precedent Hancock valve insertion), clinical (New York Heart Association [NYHA] class, atrial fibrillation, and pacemaker), operative (operative date, ascending aorta replacement, coronary artery bypass grafting, left internal thoracic artery use, emergency operation, and operative center), and associated risk factors (hypertension, chronic renal failure, diabetes, previous cardiac operation, coronary artery disease, and critical coronary disease).
Late survival was evaluated by means of a Cox semiparametric model, and SVD was evaluated by means of Weibull parametric model. Variable choice, by means of a stepwise backward and forward algorithm (P to enter, .10; P to retain, .20), was supported by bootstrap analysis with 1000 repetitions of randomly selected samples with replacement. Calibration of the operative age variable was obtained by plotting the coefficients of SVD probability in 6 age groups of approximately equal size versus several age transformations and choosing the one with the best linearizing property. Propensity score analysis4 was performed to identify blocks of HII and HST valves with the same mean score and with balanced prevalence of all covariates (P > .01).
|
Results
|
|---|
Thirty-day Mortality
A total of 89 HST and 76 HII patients died, with a surgical risk of 12.5% (95% confidence interval, 10%-15%) versus 6% (95% confidence interval, 4.6%-7.3%).
Overall Mortality
HST mortality was 392 patients, and HII mortality was 477 patients. Unadjusted survival of HST versus HII was 73% ± 1.7% versus 79% ± 1.2% at 5 years, 59% ± 1.9% versus 58% ± 1.7% at 10 years, and 40% ± 2.3% versus 39.7% ± 2.4% at 15 years (P = .46). Age-adjusted HST survival at 15 years was 25% versus 46% (P < .001), but the difference was exclusively related to the 30-day mortality, as shown in Figure E1.
Multivariate Survival Analysis
Bootstrapped multivariable Cox analysis of survival determinants of the perioperative survivors (Table E1) was identified 99.9% of the time (bootstrap drawings). It detailed a 63% decreased risk for patients up to 45 years, a decreased risk of 29% up to 60 years, a 26% increased risk up to 70 years, and an 84% increased risk for more than 70 years. The incremental risk of NYHA class was identified 100% of the times, with a hazard ratio of 2 for NYHA III, a hazard of 3.5 for NYHA IV, and a hazard of 3.7 if cardiogenic shock was also present. Endocarditis was selected in 97% of drawings, with a hazard ratio of 3. Male sex, with a 1.5 hazard ratio; coronary artery disease, with a 1.4 hazard ratio; and atrial fibrillation, with a 1.3 hazard ratio, were selected in 99%, 84%, and 53% of the drawings, respectively. A more recent operative year decreased the risk of death of 3% per year in 56% of the drawings.
Valve-related Adverse Events
HST valves had 91 embolic episodes, versus 123 for the HII; 41 hemorrhagic events, versus 49; 25 endocarditis events, versus 37; 4 thrombosis events, versus 4; and 19 perivalvular leaks, versus 5. The severity of HII events is detailed in this issue.5 Linearized rates are summarized in Table E2
and, with the exception of leaks, were similar between models and were within the limits reported by the review of the literature.6 Nonetheless, valve-related mortality and morbidity were significantly higher in the HST group as a result of a significantly larger number of reoperations.
Reoperation
Eighty HII and 340 HST patients underwent reoperation. SVD was the first cause of reoperation in 51 HII cases and 293 HST cases. The second most common cause was endocarditis (21 HII and 21 HST cases). Overall actuarial freedom from reoperation of the HII was 97.7% ± 0.5% at 5 years, 92.9% ± 1.1% at 10 years, and 76.3% ± 3.3% at 15 years. In HST patients, it was 93.8% ± 1.0%, 69.4% ± 2.1%, and 35.9% ± 2.4%, respectively (P < .001).
Structural Valve Deterioration
Bioprosthetic valve dysfunction due to structural valve deterioration (SVD) occurred in 57 HII (21 aortic and 36 mitral) and 313 HST (80 aortic and 233 mitral) patients. For HII in the mitral position, there were 27 events in patients younger than 65 years and 9 SVD episodes in older patients. In the aortic position, there were 15 events in patients younger than 65 years and 6 events in older patients. For HST in the mitral position, there were 231 events in patients younger than 65 years and 2 events in older patients. In the aortic position, there were 77 and 3 events, respectively. The relationship of age versus the probability of SVD, obtained by stratifying patients in 6 groups of equal size, was curvilinear, with a lower probability of the event as the operative age increased, and was linearized by an exponential transformation (Figure 1).
View larger version (17K):
[in this window]
[in a new window]
|
Figure 1. Continuous age calibration: operative age was stratified into 6 groups of equal size, and the SVD risk coefficient (Cox analysis) was plotted against the median age of each group (circles). The best age transformation that linearized the circle distribution was the exponential of age. SVD, structural valve deterioration.
|
|
The overall actuarial freedom from SVD of HII versus HST is shown in Figure 2, and freedom of patients from SVD younger than 65 years or 65 years or older is shown in Figures 3 and 4.
Actual freedom from SVD is summarized in Table 2.
View larger version (20K):
[in this window]
[in a new window]
|
Figure 2. Actuarial freedom from SVD for the Hancock II and Standard valves with 95% confidence intervals (C.I). The Weibull predicted probability curves, adjusted for operative age and other covariates in Table 3, are overlaid. SVD, structural valve deterioration.
|
|
View larger version (20K):
[in this window]
[in a new window]
|
Figure 3. Actuarial freedom from SVD for the Hancock (Hck) II and Standard valves in patients operated on at ages younger than 65 years. SVD, structural valve deterioration.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Figure 4. Actuarial freedom from SVD for the Hancock (Hck) II and Standard valves in patients operated on at ages 65 years or older. SVD, structural valve deterioration.
|
|
Bootstrapped multivariable Weibull analysis in 100% of draws identified older age as a powerful protective factor, with a risk reduction of 8.2% every 10 years. Given the exponential relationship of SVD and age, the SVD hazard ratio was 3 times higher for patients younger than 45 years of age and 2.4 times higher for patients younger than 60 years. It was 0.62 to 70 years and 0.28 above 70 years.
Critical coronary artery disease and degenerative valve pathology were decremental risk factors identified in 89% and 81% of the bootstraps, respectively, with a hazard reduction of 74% and 33%, respectively. Conversely, HST was an incremental risk factor in 100% of the bootstraps (hazard ratio, 2.2), with a predicted median duration of 12 years, versus 19 years for the HII. Hypertension (hazard ratio, 2.3) and mixed lesions (hazard ratio, 1.27) were incremental risk factors selected in 74% and 81% of the bootstraps, respectively (Table 3).
Multiple decremental analysis of 65-year-old or older patients operated on with the HST valve showed that after 15 years, their actual probability of being alive with the original prosthesis was 6%, their death probability was 76%, and the SVD probability was 18%. In 65-year-old or older patients operated on with HII valves, the 15-year probability of being alive with the original prosthesis was 30%, the death probability was 64%, and the SVD probability was 6%.
Propensity score analysis identified 5 blocks of HII and HST valves with the same mean propensity score and balancing of all covariates. These blocks add up to 541 valves (255 HST and 286 HII), with 81 events in the HST group and 26 in the HII group. The ratio of HST to HII was 169:165 in the mitral position and 86:121 in the aortic position. Bootstrapped Weibull analysis confirmed the incremental SVD risk of HST, with a hazard ratio of 2. Older age was a significant decremental risk factor, with an approximately 10% decrement every 10 years (Table 4).
|
Discussion
|
|---|
Porcine prostheses were introduced into clinical use by Carpentier and colleagues7 in the 1960s, and the commercially available HST valve was immediately accepted in Padova as an ideal substitute in young or middle-aged patients and fertile women.8 Many valves (n = 583) had already been inserted in all positions before the disquieting reports of early prosthetic degeneration in children by Geha and associates9 and Silver and colleagues.10 The actuarial SVD freedom of our HST valves, inserted in patients with a mean operative age of 48 years, was 38% at 15 years. This is similar to the results reported by several groups.11–13 It is therefore representative of the average durability of this prosthesis, and we consider this historical experience very useful to assess the advantage of the modifications of the second-generation HII prosthesis.
Tissue valve failures were summarized by Valente and associates14 as primary failure (due to dystrophic calcification, thrombosis, fibrous tissue overgrowth, primary tears, cuspal hematomas, and stent-post bending) or secondary failure (due to endocarditis and paravalvular leak). Calcification was the main factor (88%) that influenced long-term durability; therefore, HII improvement was directed toward maintenance of the natural collagen crimps, with low or zero pressure fixation, the addition of a dodecylsulfate (T6) calcium-retarding agent, and a stent with a lower implant profile.
Use of second-generation prostheses was prudentially restricted to older ages because of the wisdom that the older the patient, the lower the risk of having a reoperation,15 and because the "worldwide surgical pendulum had swung back toward favouring mechanical valves in the majority of patients."16 It is obvious that to verify the claimed prosthetic improvements, their use should be liberalized to relatively younger patients to permit age-matched comparisons, but this cannot be done, for ethical reasons. It is also obvious that actual probabilities must be calculated, because actuarial estimates are severely biased by the censoring of dead patients.2 Despite the reversed bias due to the introduction of sensible and specific echocardiography, HII valves showed 97% actual freedom from SVD at 10 years and 88% freedom at 15 years, with a significant improvement compared with the 64% actual SVD freedom of our HST series. The actual 15-year freedom from SVD of our HII series5 is similar to those recently reported by two large series of HII valves17,18 and seems superior to two large series of pericardial valves,19,20 being representative of the average performance of second-generation porcine prostheses.
The HII and HST series are not concomitant, and patient characteristics differ largely. In a time span of 3 decades, rheumatic mitral disease has been dramatically reduced, and there has been an increase of degenerative aortic stenosis of the elderly and of concomitant coronary artery disease21 and coronary artery bypass grafting. Operative techniques have improved in many details, with a decrease of operative mortality from 13% to less than 6%. The early surgeon had to use skill and surgical dexterity to offset the limitations imposed by cardiac ischemia, as shown by the significantly shorter clamp times of the HST. Adjusting for age differences, late mortality, valve-related mortality, and complications, rates were identical in the compared prostheses. The only differences were a significant reduction of the dehiscence rate and the SVD-related reoperation rate.
In the presence of a moving target, despite multivariable analysis adjustment, it is difficult to resist the accusation that we are comparing apples with oranges22; therefore, a propensity analysis was applied to balance the patients of the two series. Six groups of patients with equal propensity scores and balanced characteristics were identified, for a total of 541 patients with an almost equal proportion of HII and HST valves.
The principal confounder in the analysis of SVD events is the patient's age at operation, so the analysis must convince the reader that the SVD reduction is due to valve improvement and not to age differences between series. In this respect, we had to cope with a 20-year age difference between series, which decreased to 5 years by propensity matching, and with the nonlinear relationship of age and SVD probability. This required a calibration (linearizing transformation) of the age effect to avoid fitting of the residuals from any age-related variable, such as the prosthetic type.
The durability of the HST was similar in the aortic and mitral (63% vs 64%) positions. Conversely, the durability of the HII prosthesis was apparently better in the aortic position (92% vs 85%). This difference was observed by David and associates17 but was not significant in our multivariable analysis of the entire series or in propensity-matched patients. In patients 65 years or older, the actual probability of valve degeneration at 15 years was only 6%, and the probability of being alive with the original prosthesis was 30%.
We conclude that there is indeed a durability improvement of the HII valve, not entirely explained by use in an elderly population, that may be attributed to its technologic advancements over the first-generation porcine valves. This consideration would encourage relaxation of age limits for its use, particularly in the aortic position.5
|
Footnotes
|
|---|
Supported by Ministry of University and Scientific Research, 2003 (project "Meta-analisi degli studi comparativi dei risultati a distanza delle protesi biologiche e meccaniche" [Meta-analysis of the studies comparing long-term results of biologic and mechanical prostheses]).
|
References
|
|---|
- Edmunds LHJ, Clark RE, Cohn LH, Grunkemeier GL, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. J Thorac Cardiovasc Surg 1996;112:708-711.[Free Full Text]
- Grunkemeier GL, Anderson RP, Miller DC, Starr A. Time-related analysis of nonfatal heart valve complications. cumulative incidence (actual) versus Kaplan-Meier (actuarial). Circulation 1997;96:II70-II75.
- Anderson WN. Algorithms for competing risks analysis. Available at: http://lib.stat.cmu.edu/general/Markov.sas Accessed 6/17/2006..
- Becker SO, Ichino A. Estimation of average treatment effect based on propensity score. Stata J 2002;2:358-372.
- Rizzoli G, Mirone S, Ius P, Polesel E, Bottio T, Salvador L, et al. Fifteen-year results of the Hancock II valve. a multicenter experience. J Thorac Cardiovasc Surg 2006;132:602-609.[Abstract/Free Full Text]
- Grunkemeier GL, Li HH, Naftel DC, Starr A, Rahimtoola SH. Long term performance of heart valve prostheses. Curr Probl Cardiol 2000;25:73-154.[Medline]
- Carpentier A, Guilmet D, Dubost C. Aortic mitral and tricuspid valve replacement with frame mounted aortic heterografts. Thoraxchir Vask Chir 1968;16:615-618.[Medline]
- Cevese PG, Gallucci V, Morea M, Dalla Volta S, Fasoli G, Casarotto D. Heart valve replacement with the Hancock bioprosthesis. Analysis of long-term results. Circulation 1977;56(3 suppl):II111-II116.
- Geha AS, Laks H, Stansel Jr HC, Cornhill JF, Kilman JW, Buckley MJ, et al. Late failure of porcine valve heterografts in children. J Thorac Cardiovasc Surg 1979;78:351-364.[Abstract]
- Silver MM, Pollock J, Siver M, Williams WG, Trusler GA. Calcification in porcine xenograft valves in children. Am J Cardiol 1980;45:685-689.[Medline]
- Fann JI, Miller CD, Moore KA, Mitchell RS, Oyer PE, Stinson EB, et al. Twenty-year clinical experience with porcine bioprostheses. Ann Thorac Surg 1996;62:1301-1312.[Abstract/Free Full Text]
- Cohn LH, Collins Jr JJ, DiSesa VJ, Couper GS, Peigh PS, Kowalker W, et al. Fifteen-year experience with 1678 Hancock porcine bioprosthetic heart valve replacements. Ann Surg 1989;210:435-442.[Medline]
- Santini F, Luciani GB, Restivo S, Casali G, Pessotto R, Bertolini P, et al. Over twenty-year follow-up of the Standard Hancock porcine bioprosthesis implanted in the mitral position. Ann Thorac Surg 2001;71(suppl 5):S232-S235.[Abstract/Free Full Text]
- Valente M, Minarini M, Thiene G, Bortolotti U, Milano A, Talenti E, et al. The pathology of Hancock Standard porcine valve prosthesis. a 20-year span of experience. J Card Surg 1990;5:328-335.[Medline]
- Jamieson WR, Rosado LJ, Munro AI, Gerein AN, Burr LH, Miyagishima RT, et al. Carpentier-Edwards Standard porcine bioprosthesis. primary tissue failure (structural valve deterioration) by age groups. Ann Thorac Surg 1988;46:155-162.[Abstract]
- Burdon TA, Miller DC, Oyer PE, Mitchell RS, Stinson EB, Starnes VA, et al. Durability of porcine valves at fifteen years in a representative North American patient population. J Thorac Cardiovasc Surg 1992;103:238-252.[Abstract]
- David T, Ivanov J, Armstrong S, Feindel CM, Cohen G. Late results of heart valve replacement with the HII bioprosthesis. J Thorac Cardiovasc Surg 2001;121:268-278.
- Medtronic Hancock II bioprosthesis clinical compendium. Available at: http://www.ctsnet.org/file/vendors/598/pdf/199802821d_EN-74.pdf Accessed and verified 6/17/2006..
- Banbury MK, Cosgrove DM, White JA, Blackstone EH, Frater RWM, Okies JE. Age and valve size effect on the long term durability of the Carpentier–Edwards aortic pericardial bioprosthesis. Ann Thorac Surg 2001;72:753-757.[Abstract/Free Full Text]
- Dellgren G, David TE, Raanani E, Armstrong S, Ivanov J, Rakowski H. Late hemodynamic and clinical outcomes of aortic valve replacement with the Carpentier-Edwards Perimount pericardial bioprosthesis. J Thorac Cardiovasc Surg 2002;124:146-154.[Abstract/Free Full Text]
- Otto CM, Lind B, Kitzman DW, Gersh BJ, Siskovick DS. Association of aortic valve sclerosis with cardiovascular mortality and morbidity in the elderly. N Engl J Med 1999;341:142-147.[Abstract/Free Full Text]
- Blackstone EH. Comparing apples and oranges. J Thorac Cardiovasc Surg 2002;123:8-15.[Free Full Text]
Top
Abstract
Introduction
